Effects of interventions targeting the systemic inflammatory response to cardiac surgery on clinical outcomes in adults

Abstract

Background

Organ injury is a common and severe complication of cardiac surgery that contributes to the majority of deaths. There are no effective treatment or prevention strategies. It has been suggested that innate immune system activation may have a causal role in organ injury. A wide range of organ protection interventions targeting the innate immune response have been evaluated in randomised controlled trials (RCTs) in adult cardiac surgery patients, with inconsistent results in terms of effectiveness.

Objectives

The aim of the review was to summarise the results of RCTs of organ protection interventions targeting the innate immune response in adult cardiac surgery. The review considered whether the interventions had a treatment effect on inflammation, important clinical outcomes, or both.

Search methods

CENTRAL, MEDLINE, Embase, conference proceedings and two trial registers were searched on October 2022 together with reference checking to identify additional studies.

Selection criteria

RCTs comparing organ protection interventions targeting the innate immune response versus placebo or no treatment in adult patients undergoing cardiac surgery where the treatment effect on innate immune activation and on clinical outcomes of interest were reported.

Data collection and analysis

Searches, study selection, quality assessment, and data extractions were performed independently by pairs of authors. The primary inflammation outcomes were peak IL‐6 and IL‐8 concentrations in blood post‐surgery. The primary clinical outcome was in‐hospital or 30‐day mortality. Treatment effects were expressed as risk ratios (RR) and standardised mean difference (SMD) with 95% confidence intervals (CI). Meta‐analyses were performed using random effects models, and heterogeneity was assessed using I².

Main results

A total of 40,255 participants from 328 RCTs were included in the synthesis. The effects of treatments on IL‐6 (SMD ‐0.77, 95% CI ‐0.97 to ‐0.58, I² = 92%) and IL‐8 (SMD ‐0.92, 95% CI ‐1.20 to ‐0.65, I² = 91%) were unclear due to heterogeneity. Heterogeneity for inflammation outcomes persisted across multiple sensitivity and moderator analyses. The pooled treatment effect for in‐hospital or 30‐day mortality was RR 0.78, 95% CI 0.68 to 0.91, I² = 0%, suggesting a significant clinical benefit. There was little or no treatment effect on mortality when analyses were restricted to studies at low risk of bias. Post hoc analyses failed to demonstrate consistent treatment effects on inflammation and clinical outcomes. Levels of certainty for pooled treatment effects on the primary outcomes were very low.

Authors' conclusions

A systematic review of RCTs of organ protection interventions targeting innate immune system activation did not resolve uncertainty as to the effectiveness of these treatments, or the role of innate immunity in organ injury following cardiac surgery.

Plain language summary

Effects of interventions targeting the systemic inflammatory response to cardiac surgery on clinical outcomes in adults

Key messages

‐ We are unsure of the benefits of anti‐inflammatory interventions after cardiac surgery.

‐ Further research should explore the role of factors such as age and chronic conditions on the response to anti‐inflammatory interventions.

What is the relation between inflammatory response and organ damage after cardiac surgery?

Following cardiac surgery, different organs can be damaged, leading to serious complications, and sometimes death. We still do not clearly know how to prevent this. Some authors think that by reducing an excess of inflammatory response (SIRS, systemic inflammatory response syndrome) with which our body reacts to the surgery we might be able to reduce organ damage. This has been tested in several scientific studies, but results are not conclusive. SIRS is observed in over 80% of patients undergoing major surgery, and in a range between 28.3% to 96.2% in the specific case of heart surgery.

What did we want to find out?

We wanted to understand if reducing the inflammatory response after cardiac surgery had a clinical benefit.

What did we do?

We conducted this review to summarise the results from these studies, by searching databases containing reports of ongoing and completed studies from the last 30 years. We included only studies conducted in adult patients.

What did we find?

Results from 328 studies including 40,255 patients are hereby presented. The effect on inflammatory activity (measured via the level of the main signals circulated in our blood) was not clearly changed. Mortality, on the other hand, was reduced in these studies, although this effect could not be confirmed when we repeated our calculations in the studies conducted at the highest quality standard exclusively. Other signs of inflammation and organ damage were not clearly changed.

In conclusion, further studies are needed before we can understand whether, by targeting inflammation, we can prevent organ damage.

What are the limitations of the evidence?

Our work show inconsistencies in the conclusions of these studies and issues in some of the methods adopted that reduce the confidence in the result of our analysis.

How up‐to‐date is this evidence?

The evidence is updated to October 2022.

Summary of findings

Summary of findings 1. Summary of findings table for pooled treatment effects of the three categories of inflammatory intervention against control.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with control	Risk with anti‐inflammatory interventions
Peak IL‐6 (between the end of surgery and 24 hours post‐surgery)	‐	SMD 0.77 SD lower (0.97 lower to 0.58 lower)	‐	6485 (152 RCTs)	⨁◯◯◯ Very low a,b,c,d
Peak IL‐8 (between the end of surgery and 6 hours post‐surgery)	‐	SMD 0.92 SD lower (1.20 lower to 0.65 lower)	‐	2963 (72 RCTs)	⨁◯◯◯ Very low a,b,c,d
Hospital or 30‐day mortality	31 per 1000	25 per 1000 (21 to 28)	RR 0.78 (0.68 to 0.91)	24,817 (190 RCTs)	⨁◯◯◯ Very low a,d,e
Myocardial injury (hospital stay)	75 per 1000	74 per 1000 (67 to 81)	RR 1.04 (0.95 to 1.14)	23,118 (86 RCTs)	⨁◯◯◯ Very low a,f +
Kidney injury (safety outcome) (hospital stay)	54 per 1000	53 per 1000 (48 to 59)	RR 1.06 (0.96 to 1.16)	23,103 (79 RCTs)	⨁⨁◯◯ Low a, c
Surgical Site Infection (hospital stay)	79 per 1000	72 per 1000 (65 to 79)	RR 0.92 (0.83 to 1.01)	19,368 (40 RCTs)	⨁⨁◯◯ Low a, c
Duration of hospital length of stay	‐	MD 0.51 lower (0.69 lower to 0.34 lower)	‐	15,547 (148 RCTs)	⨁◯◯◯ Very low a, b ++
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group across the studies included in the review and the relative effect of the intervention (and its 95% CI). For peak IL‐6, peak IL‐8, and duration of hospital length of stay, SMD can be interpreted according to the report by Cohen (Cohen 1988) as representing a small effect (0.2), a medium effect (0.5), or a large effect (0.8).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations a. Downgraded for study limitations (risk of selection and detection bias) b. Downgraded for substantial inconsistency (with statistical heterogeneity above 50%) in study results that persisted after investigation of both prespecified and post hoc hypotheses that could have explained heterogeneity c. Downgraded for imprecision (wide confidence intervals for treatment effect) d. Downgraded for publication bias (high likelihood of publication bias from funnel plot inspections) e. Downgraded for inconsistency as both the sensitivity analysis and the post hoc analysis in trials at low risk of detection bias could not replicate the effect found in the primary analysis f. Downgraded for indirectness due to differences in the adopted definitions of the outcome across trials + Downgraded by two levels for very serious risk of bias, and by one level for indirectness ++ Downgraded by two levels for very serious risk of bias, and by one level for inconsistency

CI: confidence interval; IL: Interleukin; RR: risk ratio; SMD: standardised mean difference

Background

.

Description of the condition

The Systemic Inflammatory Response Syndrome (SIRS) is observed in over 80% of patients undergoing major surgery (Brun‐Buisson 2000), and in a range between 28.3% (Squiccimarro 2019) to 96.2% (MacCallum 2014) in the specific case of cardiac surgery, mainly based on the risk profile and the specific surgical operation performed in the cohort examined. SIRS has been defined clinically by the presence of at least two of the following criteria: fever higher than 38.0°C or hypothermia lower than 36.0°C, heart rate > 90 beats/minute, respiratory rate > 20 breaths/minute, leukocytosis (higher white blood cell count) > 12*10⁹/L or leucopenia (lower white blood cell count) < 4*10⁹/L (Bone 1992). Concentrations of innate immune system activation biomarkers in blood; interleukin (IL)‐8, IL‐6 and Tumour Necrosis Factor alpha (TNFα), are considered to be indicative of the severity of SIRS. In cardiac surgery patients, an increase in the number of positive criteria in the SIRS definition, as well as higher concentrations of SIRS biomarkers have been associated with organ injury, increased mortality, and the use of resources (Murphy 2004a). These observations have led to years of research evaluating the potential organ protective effects of interventions that target components of the SIRS pathway.

Acute kidney injury (AKI), pulmonary dysfunction, and low cardiac output are common complications of cardiac surgery affecting 30% to 50% of all patients and preceding the majority of deaths (Murphy 2015). Brain injury is another frequent complication of CPB (cardiopulmonary bypass), which commonly manifests as a spectrum of disorders including cognitive dysfunction, stroke, and seizure with an overall incidence between 30% and 70% (Jufar 2021). Supportive care for patients with organ injury also has significant resource implications; in one large high‐quality study, healthcare costs were 70% higher for patients with organ (kidney, lung, myocardial) injury (n = 717, mean cost from surgery to three months £24,539) compared to those without (n = 1291, £14,450) (Murphy 2015). Despite decades of research, effective prevention strategies for post‐cardiac surgery organ injury remain elusive (Bone 1992; Landis 2014; Landis 2015). Annually, cardiac surgery is performed on over 35,000 UK patients (Hickey 2012) and an estimated number above 1 million patients worldwide, with the proportion of patients at increased risk of post‐surgery organ injury increasing year‐on‐year (Hickey 2012). Reducing perioperative organ injury, therefore, presents an ever‐increasing challenge for clinicians and health services, and is a clinical research priority (PSP 2019). Morbidity and mortality (often a consequence of organ injury) might be reduced by diminishing the inflammatory response to cardiac surgery.

The host response to cardiac surgery is characterised by the simultaneous activation of multiple inflammatory pathways. These may occur as the result of blood activation by the cardiopulmonary bypass circuit or operative field, failure of autoregulatory processes leading to tissue oxygen supply and demand mismatch, systemic shock secondary to bleeding or low cardiac output, thromboemboli, or massive blood transfusion. These pathways are characterised by redundancy but ultimately result in the activation of vascular endothelial cells (the internal layer of blood vessels), as well as myelomonocytic cells (a subpopulation of white blood cells), and elevated levels of myelomonocytic cytokines (small signalling proteins) that promote cellular tissue sequestration and inflammation. This is characterised by refractory hypoxia, mitochondrial dysfunction, oxidative stress, high‐energy phosphate depletion in tissues, and ultimately, organ injury most commonly affecting the heart, lungs, brain, and kidneys (Murphy 2004a; Murphy 2004b).

Description of the intervention

The following types of interventions targeting the inflammatory response in cardiac surgery were included:

1. Interventions that attenuate haematological activation by the circuit or surgical field: Two types of interventions targeting this mechanism were included:

   1. Pharmacological ‐ kinin inhibitors, complement inhibitors, complement inhibitors, thrombin inhibitors, plasmin protease inhibitors or neutrophil elastase inhibitors; and

   2. Mechanical ‐ leukocyte‐depleting filters, aspiration of split blood, discarding of split blood, use of biocompatible circuit coating, minimised extracorporeal circuits or off‐pump coronary revascularisation.

2. Interventions that attenuate ischaemia reperfusion injury: Two types of interventions targeting this mechanism were included:

   1. Pharmacological ‐ use of inhalational anaesthetic agents, intravenous propofol, lignocaine, aminophylline/adenosine agonists or the targeting of cell survival signalling with Insulin‐Glucose‐Potassium therapy; and

   2. Adaptive/preventive ‐ remote ischaemic conditioning or avoidance of cardioplegic arrest during.

3. Interventions with nonspecific anti‐inflammatory activity: Two types of interventions targeting this mechanism were included:

   1. Pharmacological ‐ use of corticosteroids, statins, nitric oxide (NO) donors, phosphodiesterase 5 (PDE‐5) antagonists, soluble guanylate cyclases (sGC) agonists, atrial natriuretic peptides, erythropoietin type drugs, N‐acetyl cysteine; and

   2. Mechanical ‐ ultrafiltration in the circuit.

How the intervention might work

Interventions that attenuate haematological activation by the CPB circuit or surgical field:

Direct contact of blood with the plastic of the extracorporeal Cardio‐Pulmonary Bypass (CPB) circuit (the technology that allows blood oxygenation while it is diverted from the heart during cardiac surgery) causes activation of several plasma protease cascades including the complement system, kallikrein/kinin system, and coagulation and fibrinolytic system. These pathways result in the activation of pro‐inflammatory mediators that include, but are not restricted to, thrombin, plasmin, bradykinin, and activated complement, which in turn activate innate and humoral immune responses as well as endothelium and platelets (Despotis 2001). The CPB circuit also directly activates blood cells, including neutrophils and platelets. Activation of protease pathways in the operating field is also associated with the release of tissue factors.

Interventions that attenuate ischaemia reperfusion injury

During cardiac surgery with CPB, the heart is commonly isolated from the arterial tree, depriving it of oxygenated blood, and resulting in myocardial ischaemia. This is an absolute requirement for intracardiac (valve) procedures to enable surgery on a still, blood‐free, operating field, and is also used in most epicardial procedures (coronary artery bypass grafts). At the end of surgery, blood flow is restored to the heart, allowing it to beat and eventually allow weaning from the CPB circuit. This procedure results in a myocardial ischaemia reperfusion injury (Turer 2010). This injury results in the local activation of many of the cascades activated systemically by the CPB circuit, along with Damage‐Associated Molecular Patterns (DAMPs), and other activating factors released by damaged cells. The release of inflammatory mediators by the heart can activate other inflammatory cascades and contribute to the dysfunction of other organ systems. In addition to the myocardium, other organ systems also demonstrate evidence of metabolic stress and ischaemia. S100, an intracellular protein with DAMP activity associated with neuronal or brain injury, is commonly elevated following cardiac surgery. Gut permeability is directly altered by CPB resulting in elevated levels of Pathogen‐Associated Molecular Patterns (PAMPs), including lipopolysaccharide.

Interventions with nonspecific anti‐inflammatory activity

The inflammatory response is characterised by the activation of multiple parallel inflammatory pathways. Interventions that target individual pathways may fail due to redundancy. An alternative approach is to administer treatments that have nonspecific inhibitory effects on innate and humoral immune system activation. Ultrafiltration is a mechanical intervention whereby blood in the CPB circuit is passed through a haemofilter to reduce plasma cytokine, DAMP and PAMP concentrations.

Why it is important to do this review

Clinical progress towards the personalised prevention and treatment of post‐cardiac surgery organ injury is limited by our poor understanding of underlying mechanisms. Cardiopulmonary bypass (CPB) is considered an important cause of organ injury (Puskas 2015). Specifically, haematological activation (coagulation and other protease cascades, platelets, leucocytes) by the extracorporeal circuit results in Systemic Inflammatory Response Syndrome (SIRS), and endothelial dysfunction. Cardioplegic arrest during CPB also results in a myocardial‐specific ischaemia reperfusion injury (IRI). This can cause suboptimal blood supply to the organs with or without low cardiac output, which compounds the metabolic stress attributable to CPB (Jufar 2021; Murphy 2004a). However, these processes only partially explain organ injury and, although organ injury is associated with SIRS, it is unclear whether this is a causal relationship. Critically, interventions that attenuate haematological activation, ischaemia reperfusion and the systemic inflammatory response to surgery have not been shown to reduce AKI, acute lung injury (ALI), brain injury, or low cardiac output in clinical trials (Bone 1992; Landis 2014; Landis 2015). This disconnect between SIRS and clinical outcome has also been highlighted in sepsis, a clinical syndrome initially defined by the presence of SIRS in the setting of infection (Bone 1992). As in cardiac surgery, decades of research targeting inflammation in sepsis have failed to deliver clinical benefits (Landis 2014). This has led to a recent revision of the consensus sepsis definition to focus on organ failure as the result of dysregulation of the host response to infection, and the rejection of previous SIRS‐based definitions. Similarly, in cardiac surgery, organ injury is increasingly viewed as the product of the interaction between patients’ baseline clinical status, the severity of the surgical stress, and the nature of the inflammatory response (Dieleman 2017; Papachristofi 2016). Randomised trials of anti‐inflammatory interventions represent the best method to evaluate the contribution of inflammation to clinical outcomes. By reviewing all the trials that have targeted inflammation as an organ protection strategy, we will gain important new insights into this question. In addition, we suggest that the diversity of the interventions evaluated in these trials may provide important mechanistic insights into the processes underlying SIRS and organ failure.

Sources of heterogeneity

The redundancy of inflammatory mechanisms

The surgical trauma and haematological activation associated with cardiac surgery are severe and result in the activation of multiple inflammatory pathways. Our limited understanding of the causal relationship between each pathway and subsequent organ failure may contribute to heterogeneity and explain negative results in trials of interventions that target redundant, secondary, or regulatory inflammatory pathways. For example, targeting of pathways early in the activation cascade, such as complement or tissue factor release, may have little or no influence on other important mechanisms. In contrast, IL‐8 release and other indicators of myelomonocytic activation, or IL‐6 release, a marker of tissue inflammation, are thought to be later events in the inflammation cascade. These, in turn, may only represent the response to injury rather than the cause.

The heterogeneity of the cohorts being studied

The status of patients pre‐surgery largely determines the response to surgical stress (Hickey 2012). The nature of this interaction is complex due to the diverse pathology of chronic diseases. For example, elderly patients undergoing CPB exhibit less SIRS than younger patients but increased levels of organ injury relative to younger patients (Dieleman 2017). Patients who have had a recent myocardial infarction (MI), poor left ventricular function, or who have heart failure are also at increased risk of developing SIRS (Dieleman 2017). Intuitively, this supports the hypothesis that SIRS is linked to myocardial ischaemia reperfusion injury during cardiac surgery. However, SIRS is more common in cardiac surgery patients with BMI > 30 (Dieleman 2017), who are paradoxically at reduced risk of low cardiac output, and death relative to normal weight patients (Mariscalco 2017). A final observation is that high‐dose statin therapy, which attenuates post‐CPB inflammation in RCTs, actually increases the frequency of AKI post‐cardiac surgery (Zheng 2016). It follows that patients’ age, comorbidity, frailty and severity of underlying cardiac disease in trial participants may be critical determinants of responses to organ protection interventions. Heterogeneity in patient cohorts in trials of organ protection interventions may therefore contribute to negative or indeterminate trial results.

The methodological limitations of existing trials

Many trials have used measures of the inflammatory response as surrogate markers of clinical effectiveness as this reduces sample size and costs of running trials. These RCTs often have small cohorts and are poorly conducted or reported, and are subject to publication bias. We therefore anticipate that methodological limitations and small study effects may influence our results and we will address these in subgroup and sensitivity analyses.

Review Questions

Our overarching question is: Do interventions that target the inflammatory response to cardiac surgery have clinical benefits?

Secondary questions are:

1. Does attenuation of specific components of the inflammatory response have clinical benefits?

2. Is heterogeneity of effects attributable to the nature of the intervention including:

   1. Interventions that attenuate haematological activation such as off‐pump, mini CPB, heparin coating, serine protease inhibitors, neutrophil filters;

   2. Attenuation of myocardial ischaemia reperfusion injury by pharmacological or ischaemic pre‐conditioning; or

   3. Pharmacological interventions that have nonspecific anti‐inflammatory activity such as corticosteroids or statins.

3. Is heterogeneity of effects attributable to the patient characteristics at baseline such as age, sex, diabetes, COPD (chronic obstructive pulmonary disease)/asthma, renal impairment, heart failure, valve versus coronary disease, peripheral vascular disease or recent myocardial infarction?

We suggest that our analyses will provide new insights into the pathological basis of post‐cardiac surgery inflammation and organ injury and assist with the design of trials evaluating novel personalised organ protection interventions.

Objectives

Primary objective:

To assess the effects of interventions targeting the systemic inflammatory response to cardiac surgery on clinical outcomes in adults.

Secondary objective:

To evaluate whether interventions that target different drivers of inflammation during cardiac surgery influence the nature of the inflammatory response or clinical outcomes.

Methods

Criteria for considering studies for this review

Types of studies

We included parallel‐group and cluster‐randomised controlled trials (RCTs).

All publications related to a trial were identified and grouped under a single reference ID. We included studies published as full text, to avoid duplicated reports, and to enable comprehensive assessment of methodological quality and other potential sources of bias.

We included only high‐quality abstracts in this review. Concerns have been raised about discordance between the data presented in abstracts and published trials (Altwairgi 2012; Toma 2006). Adherence of abstracts to CONSORT has also been shown to be suboptimal in a recent review (Khan 2019), even in the specific case of cardiovascular journals, hence affecting their appraisal and potentially generating erroneous conclusions. Due to the high number of interventions included in our review, we anticipated that these shortcomings would be amplified, since abstracts may ultimately represent an extensive percentage of the available evidence (if not the majority) retrieved from our search. Despite the above listed limitations, we included reports that did not end up in a peer‐reviewed publication if they allowed a methodological quality assessment, allowing us to complete ratings of the risk of bias domains considered in the Cochrane Risk of Bias tool outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017), and if they conformed to the CONSORT statement (Hopewell 2008); moreover, we included a report from the ongoing studies identified in our search in clinical trial registers, ensuring publication status did not affect our review while preserving the quality of the evidence (De Angelis 2004).

Types of participants

Trials recruiting participants over 16 years of age undergoing cardiac surgery on the heart or great vessels (CABG, valve replacement or repair, aortic or all other cardiac surgery procedures, except for heart transplantation) were included.

Trials recruiting participants with sepsis were excluded as sepsis may involve tissue and inflammatory responses that differ from those in patients undergoing cardiopulmonary bypass. When only a subset of eligible participants was included in the study, extraction of the relevant data was performed, where possible. We contacted investigators or study sponsors in order to obtain missing numerical outcome data on eligible subsets, where possible.

Types of interventions

We included trials evaluating an intervention (or a combination of interventions) that attenuates the inflammatory response as listed below:

Interventions that attenuate haematological activation by the CPB circuit or surgical field:

1. Off‐pump coronary revascularisation;

2. Minimised extracorporeal circuits;

3. Minimally invasive cardiac surgery; mini‐sternotomy, mini‐thoracotomy, laparoscopic;

4. Biocompatible circuit coating;

5. Leukocyte‐depleting filters;

6. Ultrafiltration;

7. Pericardial blood washing and auto transfusion;

8. Discard of mediastinal blood;

9. Aprotinin;

10. Tranexamic acid;

11. Complement inhibitors;

12. Neutrophil elastase inhibitors;

13. Direct thrombin inhibitors;

14. Bradykinin inhibitors.

Interventions that attenuate ischaemia reperfusion injury:

1. Non‐cardioplegic arrest;

2. Insulin‐Glucose‐Potassium therapy;

3. Propofol anaesthesia;

4. Sevoflurane anaesthesia;

5. Lidocaine;

6. Aminophylline/adenosine;

7. Remote ischaemic preconditioning.

Interventions with nonspecific anti‐inflammatory activity:

1. Corticosteroids;

2. Statin therapy;

3. Upregulation of NO signalling; NO donors, PDE‐5 antagonists;

4. Atrial natriuretic peptides;

5. Haptoglobin;

6. N‐acetyl cysteine;

7. Erythropoietin.

All doses, durations, and modes of delivery of the interventions were eligible.

We considered the following comparisons:

• each of the three categories of intervention versus placebo or standard of care;

• each of the combinations of the three categories (interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion, interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity, interventions attenuating ischaemia reperfusion + interventions with nonspecific anti‐inflammatory activity, interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion + interventions with nonspecific anti‐inflammatory activity) versus placebo or standard of care

We excluded trials considering interventions that target established organ failure: weaning from ventilators, renal replacement therapy, implantation of ventricular assist devices, extra‐corporeal membrane oxygenation (ECMO), on the basis that the inflammatory response that determines organ recovery had to be distinct from that targeted by perioperative interventions.

For the purposes of this review, controls received standard care (no anti‐inflammatory intervention) or placebo. Trials that compared one anti‐inflammatory intervention against another (rather than against standard care) were excluded. Trials that compared one anti‐inflammatory intervention versus standard care or placebo, but where both groups received a concomitant anti‐inflammatory intervention such as tranexamic acid or coated circuits were included. This potential source of heterogeneity was assessed in subgroup analyses.

Types of outcome measures

To be included in the synthesis, trials had to measure biomarkers of the inflammatory response and clinical outcomes as listed below. Reporting one or more of the outcomes listed here in the review was not an inclusion criterion for the review. Where a published trial report did not appear to report one of these outcomes, we accessed the trial protocol and contacted the trial authors to ascertain whether the outcomes were measured but not reported. Relevant trials that measured these outcomes but did not report the data at all, or not in a usable format, were included in the review as part of the narrative. However, we excluded trials that did not measure or mention any data regarding the collection of the outcomes of interest.

Primary outcomes

Measures of treatment effect on inflammation

The co‐primary inflammation outcomes were IL‐8, the primary chemokine of the systemic inflammatory response, assessed as the highest measured value between the end of surgery and six hours post‐surgery, and IL‐6, a common biomarker of tissue inflammation, as the highest measured value between the end of surgery and 24 hours post‐surgery.

Measures of treatment effect on clinical outcomes

The primary clinical effectiveness outcome was in‐hospital or 30‐day mortality.

Secondary outcomes

Secondary outcomes

1. Clinical indicators of organ injury

a. Myocardial injury: low cardiac output as defined by the study authors;

b. Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome);

c. Lung injury: acute lung injury/respiratory distress syndrome or tracheostomy as described by study authors;

d. Brain injury: stroke as described by study authors;

e. Infection: surgical site infection (safety outcome);

f. Duration of hospital length of stay.

For all clinical indicators of organ injury, we will consider events recorded during the postoperative stay.

2. Biomarkers of the innate immune response

For analyses of inflammation, we collected the peak value reported within the prespecified time window for that biomarker. In every prespecified time window, in the case of multiple equal peaks, we recorded the measurement at the longest follow‐up time available. The expected peak was prespecified for each biomarker. For example, complement activation and neutrophil activation are expected to peak during surgery; IL‐8 and IL‐6 are expected to peak between 0 and six hours post‐surgery. Monocyte activation is expected to peak at 12 to 24 hours post‐surgery. Values outside these expected peak windows were not collected on the expectation that these values could be attributable to other unmeasured processes such as infection or shock. This approach increased the likelihood that extracted data represented the treatment effects of interventions on the inflammatory response to surgery.

1. Extracorporeal circuit activation pathways

• Complement factors C3a, C5a, or C5b9/terminal complement complex values up to six hours post‐surgery.

• Fibrinolysis: D‐dimers measured value up to six hours post‐surgery.

2. Cellular activation

• Blood leucocyte activation markers: neutrophil elastase up to six hours post‐surgery.

• Endothelial cell activation: ICAM, E‐Selectin highest measured value between 24 and 48 hours post‐surgery.

3. Cytokines

• Acute phase proteins: CRP, TNFα between 24 and 48 hours post‐surgery.

• Regulatory cytokines: IL‐10 up to six hours post‐surgery.

4. Oxidative stress

• Malondialdehyde between six and 24 hours post‐surgery.

• Myeloperoxidase between surgery and six hours post‐surgery.

Search methods for identification of studies

Electronic searches

We identified trials through systematic searches of the following bibliographic databases:

• Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (CENTRAL Issue 9 of 12, 2022, Cochrane Library);

• MEDLINE (Ovid, 1946 to October 8, 2022);

• Embase (Ovid, 1980 to 2022 week 40);

• Conference Proceedings Citation Index‐Science (CPCI‐S) on the Web of Science (Clarivate Analytics, 1990 to 08 October 2022).

The searches were designed and performed by a Cochrane Information Specialist.

The search was limited to 1993 until present to identify studies relevant to contemporary cardiac surgery, in which cytokine assessment was considered. The searches have been updated up to October 2022. The results from all databases have been de‐duplicated with each other.

There have been enormous advancements in CPB technology, as well as changes in patients undergoing cardiac surgery over the last 20 years (Reichenspurner 2016). Bypass circuits, tubing, filters and oxygenators are very different from those used 20 years ago. We suggest that studies over 20 years old will contribute little to our synthesis, as: 1. The finding will not be relevant to contemporary cardiac surgery; 2. The quality of the reporting of these studies is likely to be poor; 3. The techniques for measurement of inflammatory biomarkers will be significantly different; 4. These factors are likely to introduce heterogeneity in our primary analysis. We therefore did not see a balance of risks and benefits in favour of including these studies.

The preliminary search strategy for MEDLINE (Ovid) (Appendix 1) was adapted for use in the other databases. The Cochrane sensitivity and precision‐maximising RCT filter (Lefebvre 2011) was applied to MEDLINE (Ovid) and adaptations of it to other databases, except CENTRAL.

We also conducted a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the WHO International Clinical Trials Registry Platform (ICTRP) Search Portal (trialsearch.who.int) for ongoing or unpublished trials.

We imposed no restrictions on the language of publication.

We did not perform a separate search for adverse effects of drugs used for the prevention of postoperative inflammation. We considered adverse effects described in included studies only.

Retractions and errata were identified contextually with the retrieval of the full‐text publication for each of the references assessed.

Searching other resources

We checked reference lists of all included studies and any relevant systematic reviews identified for additional references to trials. We also examined any relevant retraction statements and errata for included studies.

Data collection and analysis

Selection of studies

Teams of two authors (from RGA, MR, AR, SP, JR, CL, GL, AB) independently screened batches of titles and abstracts for inclusion of all the potential studies we identified as a result of the search and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. If there were any disagreements, a third author was asked to arbitrate (GJM). We retrieved the full‐text study publication and teams of two authors (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) independently screened the full text and identified studies for inclusion, and identified and recorded reasons for exclusion of the ineligible studies. We resolved any disagreement through discussion or, if required, we consulted a third person (GJM). We identified and excluded duplicates and collated multiple reports of the same study so that each study rather than each report was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and Characteristics of excluded studies table (Liberati 2009).

Data extraction and management

We used a data collection form for study characteristics and outcome data which had been piloted on at least one study in the review. Teams of two authors (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) extracted study characteristics from included studies. We extracted the following study characteristics.

1. Methods: study design, total duration of study, details of any 'run‐in' period, number of study centres and location, study setting, and date of study.

2. Participants: N randomised, N lost to follow‐up/withdrawn, N analysed, reasons for withdrawals, mean age, gender, severity of condition (according to study authors), inclusion criteria, and exclusion criteria.

3. Interventions: intervention, comparison.

4. Outcomes: primary and secondary outcomes specified and collected, and time points reported.

5. Notes: funding for trial, and notable conflicts of interest of trial authors.

Teams of two authors (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) independently extracted outcome data from batches of the included studies. We resolved disagreements by consensus or by involving a third person (GJM). All authors (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) transferred batches of data into the RevMan Web (RevMan Web 2020) file. We double‐checked that data were entered correctly by comparing the data presented in the systematic review with the data extraction form. A second review author (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) spot‐checked study characteristics for accuracy against the trial report.

Assessment of risk of bias in included studies

Teams of two authors (from RGA, MR, AR, SP, JR, CL, ST, GL, AB) independently assessed the risk of bias for batches of the included studies using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We resolved any disagreements by discussion or by involving another author (GJM). We assessed the risk of bias according to the following domains.

1. Random sequence generation.

2. Allocation concealment.

3. Blinding of participants and personnel.

4. Blinding of outcome assessment.

5. Incomplete outcome data.

6. Selective outcome reporting.

7. Other bias.

We graded each potential source of bias as high, low or unclear and provided a quote from the study report when appropriate together with a justification for our judgement in the Risk of bias table. We summarised the risk of bias judgements across different studies for each of the domains listed. Where information on risk of bias related to unpublished data or correspondence with a trialist, we noted this in the Risk of bias table.

We rated trials at high risk of funding bias where industry support to the authors or to the study itself was disclosed and/or suspected.

When considering treatment effects, we took into account the risk of bias for the studies that contributed to that outcome.

Assessment of bias in conducting the systematic review

We conducted the review according to the published protocol and reported any deviations from it in the Differences between protocol and review section of the systematic review.

Measures of treatment effect

We analysed dichotomous data (in‐hospital or 30‐day mortality, myocardial injury, kidney injury, lung injury, brain injury, and infection) as risk ratios (RR) and continuous data (duration of hospital length of stay, peak IL‐6, peak IL‐8, biomarkers of the innate immune response), as the mean difference (MD) or standardised mean difference (SMD).

For peak IL‐6, peak IL‐8, and duration of hospital length of stay, SMD can be interpreted according to the report by Cohen (Cohen 1988) as representing a small effect (0.2), a medium effect (0.5), or a large effect (0.8).

If we combined data from rating scales in a meta‐analysis, we ensured they were entered with a consistent direction of effect (e.g. lower scores always indicated improvement). Treatment effects were reported along with their 95% Confidence Intervals (CIs).

Unit of analysis issues

We included both parallel‐design and cluster‐RCTs. When multiple treatment arms were reported, we took into account the highest dose of the arm with multiple combined interventions and compared it to the control group to avoid double‐counting in the meta‐analysis.

When repeated outcome measurements at different time points were reported, we considered only the time points we prespecified in the Types of outcome measures section.

For cluster‐randomised trials, we extracted the estimates of the required effect measure (for example, odds ratio and its confidence interval) that had properly accounted for the cluster design from the published papers. These effect estimates and their standard errors were meta‐analysed with those from the studies with a parallel design using the generic inverse‐variance method (Higgins 2019).

Cross‐over trials were excluded due to the type of interventions considered in our review.

Dealing with missing data

Where possible, we calculated missing standard deviations using other data from the trial, such as confidence intervals, based on methods outlined in Chapter 6: Choosing effect measures and computing estimates of effect of the Cochrane Handbook(Higgins 2019). Where this was not possible, and the missing data were thought to introduce serious bias, we explored the impact of including such studies in the overall assessment of results through a sensitivity analysis.

Assessment of heterogeneity

We inspected forest plots visually to consider the direction and magnitude of effects and the degree of overlap between confidence intervals. We used the I² statistic to measure statistical heterogeneity amongst the trials in each analysis, but acknowledged that there is substantial uncertainty in the value of I² when there are only a few studies. We also considered the P value from the Chi² test. If we identified substantial or considerable heterogeneity, we reported it and explored possible causes by prespecified subgroup analysis (for primary outcomes only). Post hoc analyses on selected outcomes were also performed, and were clearly indicated in the text.

We considered:

• I² 0% to 40%: might not be important;

• I² 30% to 60%: may represent moderate heterogeneity;

• I² 50% to 90%: may represent substantial heterogeneity;

• I² 75% to 100%: considerable heterogeneity.

Several pre‐planned subgroup analyses and meta‐regressions have been performed to identify potential causes of clinical and methodological heterogeneity.

Assessment of reporting biases

If we were able to pool more than 10 trials, we created and examined a funnel plot to explore possible small study biases for the primary outcomes. When sufficient studies for a funnel plot were available, we performed a formal statistical test for asymmetry (Egger 1997).

Data synthesis

We undertook meta‐analyses only where this was meaningful i.e. if the treatments, participants and the underlying clinical question were similar enough for pooling to make sense. We conducted meta‐analysis when the outcomes were reported in two or more studies. We reported the pooled estimates from the random‐effects models, since we anticipated a high level of heterogeneity, originating from multiple sources (preoperative patients’ characteristics, surgery type and technique, timing, type and dose of interventions).

Subgroup analysis and investigation of heterogeneity

Due to the large number of interventions and outcomes, we proposed to only explore causes of significant heterogeneity for our primary outcomes based on prespecified subgroup sources.

1. Surgical cohort defined by study authors as high risk versus low risk.

2. Type of intervention (interventions attenuating haematological activations, interventions attenuating ischaemia reperfusion, interventions with nonspecific anti‐inflammatory activity).

3. Type of surgery (exclusively CABG or other).

4. Source of funding (industry‐funded vs non‐industry funded).

In addition, we conducted meta‐regression for the co‐primary inflammation and the primary clinical outcomes to explore the relationship between study‐level characteristics (patient and surgery‐related factors including age, sex, diabetes, renal impairment, COPD, BMI, anaemia, heart failure, and ejection fraction) and the treatment effect size. The effect of anti‐inflammatory interventions might in fact be determined by patients and surgical procedures’ characteristics (surgical risk, type of surgery, and urgency), the type of anti‐inflammatory intervention itself, and by methodological factors intrinsic to the trial design (source of funding).

We tested whether there was statistical significance (P value less than 0.05) in a linear relationship between the moderator and the intervention effect, and we reported the regression coefficients (β) to describe how the intervention effect was changed with the increase or decrease in the potentially explanatory moderator.

Additional quantitative analysis outside RevMan was performed in R (Viechtbauer 2010).

We used the formal statistical test Chi² as an overall test for subgroup differences (not a statistical test for interactions) in Review Manager (Review Manager 2014). A P value less than 0.05 was considered to represent significant interactions.

Sensitivity analysis

We performed sensitivity analyses for the co‐primary inflammation and the primary clinical outcomes to test whether important methodological limitations had influenced the results of the primary analyses. One sensitivity analysis considered exclusively studies with a low risk of bias for allocation concealment. An additional post hoc sensitivity analysis excluded studies at high risk of detection bias.

Summary of findings and assessment of the certainty of the evidence

We created a Summary of findings table using the following outcomes (IL‐8, IL‐6, in‐hospital or 30‐day mortality, clinical markers of myocardial and renal injury, surgical site infection and hospital length of stay), for the combined effect of the three main categories of interventions pooled together. We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it related to the studies which contributed data to the meta‐analyses for the prespecified outcomes. We used methods and recommendations described in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2017) using GRADEpro software (GRADEpro GDT 2015). We justified all decisions to downgrade the quality of studies using footnotes, and we made comments to aid the reader's understanding of the review, where necessary.

Judgements about evidence quality were made by three review authors (RGA, ST, MR) working independently, with disagreements resolved by discussion or involving a fourth author (GJM). Judgements were justified, documented and incorporated into the reporting of results for each outcome.

We extracted study data, formatted our comparisons in data tables and prepared a Summary of findings table before writing the results and conclusions of our review.

Reaching conclusions

We based our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. We avoided making recommendations for practice and our implications for research will suggest priorities for future research and outline what the remaining uncertainties are in the area.

Results

Description of studies

Results of the search

Results from the search and study selection process are summarised in a PRISMA diagram in Figure 1. Of the 5144 records identified from literature searches and from reference lists of included systematic reviews, 397 full‐text studies were assessed for eligibility. A total of 328 studies, with 379 eligible comparisons between treatment groups, were included in the analysis. Thirteen studies are currently ongoing (Cardoso 2021; ChiCTR2000038585; ChiCTR2000041099; ChiCTR2100043950; Fiorentino 2015; ISRCTN15255199 2019; NCT04632095 2020; NCT04648540 2020; NCT05162742 2021; NCT02518087 2015; NCT02984111 2016; NCT03657225 2018; Zhang 2022).

1.

Prisma Flow Diagram

Studies were published between 1993 and 2022. The median number of participants was 40 (IQR 28 to 60).

Included studies

A total of 40,255 participants took part in the 328 trials included in this review. The median age of participants was 63.3 (IQR 60.8 to 65.8). Participants were mostly males (median female to male ratio was 0.26, IQR 0.16 to 0.4), with the frequency of important comorbid conditions as follows; diabetes (21%, IQR 5% to 33%), peripheral vascular disease (8%, IQR 0% to 14.3%), recent myocardial infarction (0%, IQR 0% to 10.3%), and chronic obstructive pulmonary disease (6%, IQR 0% to 13%). Median baseline haemoglobin and serum creatinine were respectively 13.2 g/dL (IQR 11.9 to 13.8) and 91.4 umol/L (IQR 88.2 to 98.6). Additional characteristics from the included studies are included in Table 2.

1. Study characteristics.

Characteristics of included studies
Publication year, median (range)	2007 (1986‐2022)
Study size, median (IQR)	40 (28‐60)
Participant age, median (IQR)	63.3 (60.8‐65.7)
Female to male ratio, median (IQR)	0.26 (0.16‐0.4)
Diabetic participants (%), median (IQR)	21% (5%‐33%)
Peripheral vascular disease participants (%), median (IQR)	8% (0%‐14.3%)
Recent myocardial infarction (%), median (IQR)	0% (0%‐10.3%)
Chronic obstructive pulmonary disease (%), median (IQR)	6% (0%‐13.0%)
Baseline haemoglobin (g/dL), median (IQR)	13.2 (11.9‐13.8)
Baseline serum creatinine (umol/L), median (IQR)	91.4 (88.2‐98.6)
Left ventricular ejection fraction (%), median (IQR)	55% (51%‐61%)
BMI, median (IQR)	27.4 (26.5‐28.5)
Type of urgency
Elective	334 (88.1%)
Urgent	5 (1.3%)
Emergent	1 (0.3%)
Mixed	8 (2.1%)
Unspecified	31 (8.2%)
Type of surgery
Exclusively Coronary Artery Bypass Graft	265 (69.9%)
Exclusively Valve surgery	33 (8.7%)
Mixed	81 (21.4%)
Type of intervention
Targeting haematological activation	247 (65%)
Targeting ischaemia reperfusion injury	41 (11%)
Nonspecific anti‐inflammatory activity	100 (26%)

Number of studies (% of column total) if not stated otherwise.
BMI: body mass index
IQR: Interquartile range

Excluded studies

Fifty‐three studies were excluded after reviewing the full text. Exclusions were due to the lack of an eligible control group in eight trials (Gott 1998; Gursu 2013; Krivoy 2008; NCT00484575; NCT00484575 2007; NCT02118025 2006; Permanyer 2020; Santarpino 2009), assessment of an intervention not listed in the protocol in six trials (Beaver 2018; Diab 2022; Kim 2012; Luciani 2009; Nader 2004; Wang 2009), no collection of prespecified outcomes of interest in 12 trials (Alexiou 2004; Bingol 2005; Elgebaly 2020; Eren 2003; Haase 2007; Orhan 2006; Ozaydin 2008; Ristikankare 2006; Sano 2003; Senay 2009; Von Spiegel 2001; Tosun 2013), and patient populations not relevant to those specified in the review protocol in five trials (Adabag 2008; Brie 2022; Burns 2005; Sisillo 2008; Wijeysundera 2007). Twenty‐one reports were related to a study type different from a randomised controlled trial (Akowuah 2017; Anastasiadis 2013; Baker 2009; Brown 2009; Bunenkov 2020; Casula 2022; Chen 2022; Diegeler 2000; Dieleman 2011; Dvirnik 2018; He 2018; Landis 2014; Later 2013; Loubser 1997; Morisaki 2013; Ng 2020; Schonebeck 2007; Yao 2020; Zakkar 2015; Zhang 2014; Zhen‐Han 2017). One of the studies was a duplicate reference for a trial already included (Volk 2003).

We could not obtain the full report for three studies (Kawamura 1999; Miura 1998; Zhou 2010) and these are therefore awaiting classification.

Risk of bias in included studies

Figure 2 and Figure 3 provide an overview of the risk of bias assessment for the included studies.

2.

Risk of bias graph

3.

Risk of bias summary

Allocation

Risk of bias due to the technique adopted for random sequence generation was low in 34% (130/379), uncertain in 65% (245/379), and high (inadequate randomisation technique ‐ quasi randomisation) in 1% (4/379) of the trials.

Risk of bias due to allocation concealment was low in 23% (88/379), uncertain in 72% (273/379), and high (insufficient concealment of allocation) in 5% (18/379) of the trials.

Blinding

Risk of bias due to blinding of participants and personnel was low in 33% (124/379), uncertain in 58% (218/379), and high (blinding of participants and personnel where potentially feasible but not implemented) in 10% (37/379) of the trials.

Risk of bias due to blinding during outcome assessment was low in 23% (89/379), uncertain in 73% (275/379), and high in 4% (15/379) of the trials.

Incomplete outcome data

Risk of bias due to incomplete outcome data was low in 81% (307/379), uncertain for 8% (29/379) of the trials and high (unjustified trial group changes, withdrawals, losses to follow‐up that might have impacted the final analysis) in 11% (43/379) of the trials.

Selective reporting

Risk of bias due to selective reporting was low in 75% (283/379), uncertain in 15% (56/379) and high (discrepancy between the reported outcomes and the analysis plan) in 10% (40/379) of the trials.

Inspections of funnel plots identified a potential risk of reporting bias for the co‐primary measures of inflammation IL‐6 (Figure 4A) and IL‐8 (Figure 4B), confirmed by Egger’s test (P < 0.0001 and P = 0.0002, respectively). Asymmetry for the reported primary clinical outcome, 30‐day mortality, in funnel plots was not statistically significant per Egger’s test, P = 0.257 (Figure 4C).

4.

Funnel plot investigating potential publication bias for trials reporting measures of IL‐6 (A), IL‐8 (B), and mortality (C)

Other potential sources of bias

Risk of funding bias was low in 64% (245/379), uncertain in 30% (110/379) and high (industry, commercial funding) in 6% (24/379) of the trials. Detection bias arising from the adoption of measures of efficacy different from current reporting standards (using the 99th percentile as a cut‐off to define outliers for interleukins) was suspected due to the very wide ranges of the raw data in the analysis of inflammation biomarkers. This was explored in a post hoc sensitivity analysis.

Effects of interventions

See: Table 1

The summary of findings, incorporating the quality of the evidence (GRADE) assessments is reported in (summary of findings Table 1).

A summary of the treatment effects of the intervention versus placebo or standard of care for primary and secondary outcomes is provided in Table 3.

2. Overview of results for studies comparing each category of anti‐inflammatory intervention against control.

Outcome	Studies	Participants	Treatment effect [95% confidence interval]	Heterogeneity (I2)
IL‐6	152	6485	SMD ‐0.77 [‐0.97, ‐0.58]	92%
IL‐8	72	2963	SMD ‐0.92 [‐1.20, ‐0.65]	91%
Hospital or 30‐day mortality	190	24,817	RR 0.78 [0.68, 0.91]	0%
C3a	28	1767	SMD ‐1.08 [‐1.46, ‐0.69]	91%
C5a	3	160	SMD ‐0.43 [‐0.75, ‐0.12]	0%
C5b9/Terminal Complement Complex	15	1153	SMD ‐0.53 [‐0.85, ‐0.20]	83%
D‐dimer	20	956	SMD ‐1.17 [‐1.61, ‐0.73]	88%
Neutrophil elastase	30	1413	SMD ‐0.96 [‐1.46, ‐0.47]	93%
ICAM	10	612	SMD ‐0.80 [‐1.42, ‐0.19]	90%
E‐selectin	4	155	SMD ‐0.41 [‐1.17, 0.34]	79%
C‐reactive protein	43	2500	SMD ‐0.24 [‐0.51, 0.04]	90%
Tumour necrosis factor‐α	61	2609	SMD ‐0.95 [‐1.29, ‐0.60]	93%
IL‐10	38	1808	SMD 0.19 [‐0.28, 0.65]	94%
Malondialdehyde	10	620	SMD ‐2.52 [‐3.43, ‐1.60]	94%
Myeloperoxidase	9	316	SMD 0.33 [‐0.02, 0.68]	46%
Myocardium injury	86	23,118	RR 1.04 [0.95, 1.14]	0%
Kidney injury	79	23,103	RR 1.06 [0.96, 1.16]	0%
Lung Injury	31	14779	RR 0.89 [0.77, 1.02]	9%
Brain Injury	86	23565	RR 0.84 [0.70 1.02]	0%
Infection	40	19368	RR 0.92 [0.83, 1.01]	0%
Duration of hospital length of stay	148	15547	MD ‐0.51 [‐0.69, ‐0.34]	74%

Overview of results of meta‐analyses for anti‐inflammatory interventions vs control

IL: interleukin
MD: mean difference
RR: risk ratio
SMD: standardised mean difference

Primary outcomes

In‐hospital or 30‐day mortality

Interventions targeting the systemic inflammatory response may reduce mortality. However, the certainty of the evidence was very low due to risk of allocation, detection and publication bias, and inconsistency from the sensitivity and post hoc analyses conducted in trials at low risk of detection bias (190 studies, RR 0.78, 95% CI 0.68 to 0.91, I² = 0%; very low‐certainty of the evidence; Analysis 24.1).

24.1. Analysis.

Comparison 24: In‐hospital or 30‐day mortality (pooled), Outcome 1: Each category and combination of intervention versus placebo or standard of care

On subgroup comparisons, different intervention categories had a similar risk in terms of mortality (P = 0.26) (Analysis 1.1). The pooled analyses of combinations of interventions included fewer than 10 trials and were not reported.

1.1. Analysis.

Comparison 1: In‐hospital or 30‐day mortality, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

A sensitivity analysis restricted to trials at low risk of allocation concealment did not demonstrate a reduction in mortality from interventions targeting the systemic inflammatory response (26 studies, RR 0.90, 95% CI 0.75 to 1.08, I² = 0%).

The moderator analysis did not identify any baseline variable associated with a difference in treatment effects on mortality; study publication year (P = 0.149), age (P = 0.435), female/male ratio (P = 0.690), percentage of diabetes (P = 0.250), BMI at baseline (P = 0.631), haemoglobin at baseline (P = 0.265), percentage of COPD (P = 0.686), percentage of renal impairment (P = 0.464), eGFR at baseline (P = 0.168), creatinine at baseline (P = 0.977), LVEF at baseline (P = 0.473), percentage of peripheral vascular disease (P = 0.293), percentage of recent myocardial infarction (P = 0.939).

No prespecified subgroup analysis identified significant treatment interactions with the outcome in‐hospital or 30‐day mortality; surgical risk (test for subgroup difference, P = 0.81), surgery type (test for subgroup difference, P = 0.294), funding bias (test for subgroup difference, P = 0.834), and urgency (test for subgroup difference, P = 0.58) (Figure 5 and Figure 6).

5.

Summary of the results of the meta‐analyses for the primary outcomes, with subgroup analyses by intervention category. Each comparison is for the intervention (or combination of different interventions) against control.
AS: aspecific
HA: haematological activation|
IR: ischaemic reperfusion
RR: risk ratio
SMD: standardised mean difference.

6.

Results from the subgroup analysis of the meta‐analyses for the primary outcomes.

CABG: coronary artery bypass graft
CI: confidence interval
RR: risk ratio
SMD: standardised mean difference

IL‐6

The certainty of the evidence supporting a treatment effect of interventions targeting the systemic inflammatory response for IL‐6 was very low due to inconsistency, imprecision, and risk for allocation, detection and publication bias (152 studies, SMD ‐0.77, 95% CI ‐0.97 to ‐0.58; I² = 92%; very low certainty of the evidence; Analysis 22.1).

22.1. Analysis.

Comparison 22: IL‐6 (pooled), Outcome 1: Each category and combination of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different effects in terms of IL‐6 (P < 0.00001) (Analysis 2.1). The pooled analyses of combinations of interventions included fewer than 10 trials and was not reported.

2.1. Analysis.

Comparison 2: IL‐6, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Sensitivity analysis restricted to trials at low risk of allocation concealment did not resolve uncertainty attributable to heterogeneity (IL‐6: 33 studies, SMD ‐0.76, 95% CI ‐1.10 to ‐0.41, I² = 90%).

In the moderator analysis (Table 4), older age (β = 0.05, 95% CI 0.02 to 0.08, P < 0.001) and recent myocardial infarctions (β = 2.25, 95% CI 0.07 to 4.43, P = 0.043) were associated with a larger treatment effect of anti‐inflammatory interventions on IL‐6, but could only partially explain the heterogeneity.

3. Results from the moderator analyses for the primary outcomes.

Moderator	IL‐6	IL‐8	Mortality
Study publication year	0.02 [‐0.01 to 0.05], P = 0.181	0.02 [‐0.02 to 0.06], P = 0.398	0.02 [‐0.01 to 0.05], P = 0.149
Age	0.05 [0.02 to 0.08], P = 0.001 *	0.05 [0.01 to 0.08], P = 0.013 *	0.01 [‐0.02 to 0.05], P = 0.435
Female/male ratio	‐0.28 [‐0.83 to 0.27], P = 0.326	‐0.56 [‐1.61 to 0.48], P = 0.292	‐0.08 [‐0.45 to 0.3], P = 0.69
Diabetes	0.91 [‐0.65 to 2.48], P = 0.254	‐0.74 [‐2.37 to 0.89], P = 0.373	1.44 [‐0.76 to 3.63], P = 0.2
BMI	0.05 [‐0.07 to 0.16], P = 0.433	0.12 [0 to 0.24], P = 0.049 *	‐0.06 [‐0.29 to 0.18], P = 0.631
Haemoglobin (baseline)	‐0.1 [‐0.32 to 0.12], P = 0.395	‐0.03 [‐0.34 to 0.28], P = 0.851	0.19 [‐0.14 to 0.52], P = 0.265
COPD	0.98 [‐0.9 to 2.86], P = 0.306	0.3 [‐1.27 to 1.88], P = 0.708	1.15 [‐4.41 to 6.7], P = 0.686
Renal impairment	5.98 [‐6.87 to 18.83], P = 0.362	‐44.46 [‐121.95 to 33.03], P = 0.261	1.69 [‐2.83 to 6.21], P = 0.464
eGFR	na	na	0.13 [‐0.05 to 0.32], P = 0.168
Creatinine	0 [‐0.02 to 0.03], P = 0.828	‐0.07 [‐0.14 to 0], P = 0.041 *	0 [‐0.04 to 0.03], P = 0.977
LVEF	‐0.8 [‐3.55 to 1.96], P = 0.569	‐2.39 [‐7.19 to 2.4], P = 0.328	1.37 [‐2.37 to 5.11], P = 0.473
PVD	3.88 [‐4.29 to 12.06], P = 0.352	0.99 [‐7.55 to 9.53], P = 0.82	3.24 [‐2.8 to 9.29], P = 0.293
Recent MI	2.25 [0.07 to 4.43], P = 0.043 *	3.72 [‐4.51 to 11.95], P = 0.376	0.12 [‐2.94 to 3.18], P = 0.939

Data are presented as regression coefficients with 95% confidence intervals, and the related P value for significance. 
BMI: Body Mass Index
COPD: Chronic Obstructive Pulmonary Disease
eGFR: estimated Glomerular Fraction Rate
LVEF: Left Ventricular Ejection Fraction
MI: Myocardial Infarction
na: not available
PVD: Peripheral Vascular Disease

Diabetes, BMI, gender, baseline haemoglobin, COPD, renal function, ventricular function, vascular disease, or surgery type (elective vs urgent vs emergency) were not significant moderators of the treatment response.

No prespecified subgroup analysis (Figure 5, Figure 6) identified significant treatment interactions: surgery type (P = 0.312), risk category (P = 0.342), urgency (P = 0.34), and funding (P = 0.160).

IL‐8

The certainty of the evidence supporting a treatment effect of interventions targeting the systemic inflammatory response for IL‐8 was very low due to inconsistency, imprecision, and risk for allocation, detection and publication bias (72 studies, SMD ‐0.92, 95% CI ‐1.20 to ‐0.65; I² = 91%; very low certainty of the evidence; Analysis 23.1).

23.1. Analysis.

Comparison 23: IL‐8 (pooled), Outcome 1: Each category and combination of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different effects in terms of IL‐8 (P = 0.010) (Analysis 3.1). The pooled treatment effects for interventions targeting ischaemia reperfusion or combinations of interventions were not reported as the subgroups included fewer than 10 trials.

3.1. Analysis.

Comparison 3: IL‐8, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Sensitivity analysis restricted to trials at low risk of allocation concealment did not resolve heterogeneity in the pooled analyses of trials evaluating all interventions targeting the systemic inflammatory response on IL‐8 (14 studies, SMD ‐0.94, 95% CI ‐1.54 to ‐0.33, I² = 93%).

In the moderator analysis (Table 4), older age (β = 0.05, 95% CI 0.01 to 0.08, P = 0.013) and higher BMI values (β = 0.12, 95% CI 0 to 0.24, P = 0.049) were associated with a larger treatment effect of anti‐inflammatory interventions on IL‐8, while baseline creatinine (β = ‐0.07, 95% CI ‐0.14 to 0, P = 0.041) was associated with a lower effect, but could only partially explain the heterogeneity.

Diabetes, gender, recent myocardial infarction, baseline haemoglobin, COPD, renal function, ventricular function, or vascular disease were not significant moderators of the treatment response.

There was no significant treatment interaction with the subgroup: surgery type (Figure 6, P = 0.79). Subgroup treatment interactions were not reported for urgency, funding bias, and surgical risk, as there were fewer than 10 studies in each.

Secondary Outcomes

Myocardial injury

There was little or no treatment effect of interventions targeting the systemic inflammatory response on myocardial injury (low cardiac output), although there was very low certainty for this estimate due to indirectness and risks of allocation and detection bias (86 studies, RR 1.04, 95% CI 0.95 to 1.14, I² = 0%; very low certainty of the evidence; Analysis 25.1).

25.1. Analysis.

Comparison 25: Myocardial injury: low cardiac output as defined by study authors (pooled), Outcome 1: Each category and combination of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different risk in terms of myocardial injury (P = 0.008) (Analysis 4.1). The pooled treatment effects for interventions targeting ischaemia reperfusion or combinations of interventions were not reported as the subgroups included fewer than 10 trials.

4.1. Analysis.

Comparison 4: Myocardial injury: low cardiac output as defined by study authors, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Kidney injury

There was little or no treatment effect of interventions targeting the systemic inflammatory response on acute kidney injury, although there was low certainty for this estimate due to imprecision, and risk of allocation and detection bias (79 studies, RR 1.06, 95% CI 0.96 to 1.16, I² = 0%; low certainty of the evidence; Analysis 26.1).

26.1. Analysis.

Comparison 26: Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome) (pooled), Outcome 1: Each category or combination of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had similar risk in terms of renal injury (P = 0.70) (Analysis 5.1). There were not enough trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

5.1. Analysis.

Comparison 5: Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome), Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Surgical site infection

Interventions targeting the systemic inflammatory response may reduce infection, although there was low certainty for this estimate due to imprecision, and risk of allocation and detection bias (40 studies, RR 0.92, 95% CI 0.83 to 1.01, I² = 0%; low certainty of the evidence; Analysis 27.1).

27.1. Analysis.

Comparison 27: Infection: surgical site infection (safety outcome) (pooled), Outcome 1: Each of categories or combinations of intervention versus placebo or standard of care

No differences between the subgroups were observed for surgical site infection (P = 0.44) (Analysis 8.1).

8.1. Analysis.

Comparison 8: Infection: surgical site infection (safety outcome), Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were not enough trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

Duration of hospital length of stay

Interventions targeting the systemic inflammatory response may reduce length of hospital stay, but the certainty of the evidence was very low because of inconsistency and risk of allocation and detection bias (148 studies, MD ‐0.51, 95% CI ‐0.69 to ‐0.34; I² = 74%; very low certainty of the evidence; Analysis 28.1).

28.1. Analysis.

Comparison 28: Duration of hospital length of stay (pooled), Outcome 1: Each of the categories or combinations of interventions versus placebo or standard of care

No differences between the subgroups were observed for duration of hospital length of stay (P = 0.21) (Analysis 9.1).

9.1. Analysis.

Comparison 9: Duration of hospital length of stay, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient trials to report analysis of combinations of different categories of interventions.

Lung injury

It is unclear whether interventions targeting the systemic inflammatory response reduce lung injury (31 studies, RR 0.89, 95% CI 0.77 to 1.02, I² = 9%; Analysis 6.1).

6.1. Analysis.

Comparison 6: Lung injury: acute lung injury/respiratory distress syndrome or tracheostomy as described by study authors, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

No differences between the subgroups were observed for lung injury (P = 0.23) (Analysis 6.1).

There were not enough trials to report results for interventions targeting ischaemia reperfusion or for combinations of different categories of interventions.

Brain injury

Interventions targeting the systemic inflammatory response may reduce risk of brain injury slightly (86 studies, RR 0.84, 95% CI 0.70 to 1.02, I² = 0%; Analysis 29.1).

29.1. Analysis.

Comparison 29: Brain injury: stroke as described by study authors (pooled), Outcome 1: Each category and combination of intervention versus placebo or standard of care

No differences between the subgroups were observed for brain injury (P = 0.21) (Analysis 7.1). There were insufficient trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

7.1. Analysis.

Comparison 7: Brain injury: stroke as described by study authors, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

C3a

Interventions targeting the systemic inflammatory response may reduce C3a, but the evidence is very uncertain due to high heterogeneity (28 studies, SMD ‐1.08, 95% CI ‐1.46 to ‐0.69; I² = 91%, Analysis 10.1).

10.1. Analysis.

Comparison 10: C3a, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Interventions targeting haematological activation might reduce C3a, although the evidence is very uncertain due to high heterogeneity (28 studies, SMD ‐ 1.08, 95% CI ‐1.46 to ‐0.69, I² = 91%; Analysis 10.1.1). There were insufficient numbers of trials available for the other subgroup comparisons.

C5a

Interventions targeting the systemic inflammatory response may reduce C5a (3 studies, SMD ‐0.43, 95% CI ‐0.75 to ‐0.12; I² = 0%, Analysis 11.1).

11.1. Analysis.

Comparison 11: C5a, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient numbers of trials for pooled analyses of the other subgroups.

C5b9/terminal complement complex

Interventions targeting the systemic inflammatory response may reduce C5b9/Terminal Complement Complex, although the evidence is very uncertain due to high heterogeneity (15 studies, SMD ‐0.53, 95% CI ‐0.85 to ‐0.20; I² = 83%; Analysis 12.1).

12.1. Analysis.

Comparison 12: C5b9/terminal complement complex, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

No differences between the subgroups were observed for C5b9/Terminal Complement Complex (P = 0.94) (Analysis 12.1). There were insufficient numbers of trials to report analyses of the other subgroups.

D‐dimer

Interventions targeting the systemic inflammatory response may reduce D‐dimer, but the evidence is very uncertain due to high heterogeneity (20 studies, SMD ‐1.17, 95% CI ‐1.61 to ‐0.73; I² = 88%, Analysis 13.1).

13.1. Analysis.

Comparison 13: D‐dimer, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

Interventions targeting haematological activation may reduce D‐dimer levels, but the evidence is very uncertain due to high heterogeneity (20 studies, SMD ‐ 1.17, 95% CI ‐1.61 to ‐0.73, I² = 88%; Analysis 13.1.1). There were insufficient numbers of trials for pooled analyses of the other subgroups.

Neutrophil elastase

Interventions targeting the systemic inflammatory response may reduce neutrophil elastase, but the evidence is very uncertain due to high heterogeneity (30 studies, SMD ‐0.96, 95% CI ‐1.46 to ‐0.47; I² = 93%, Analysis 14.1).

14.1. Analysis.

Comparison 14: Neutrophil elastase, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

No differences between the subgroups were observed for neutrophil elastase (P = 0.78) (Analysis 14.1). There were insufficient numbers of trials to report analyses of the other subgroups.

Intercellular adhesion molecule

Interventions targeting the systemic inflammatory response may reduce ICAM concentrations in blood, although the evidence is very uncertain due to high heterogeneity (10 studies, SMD ‐0.80, 95% CI ‐1.42 to ‐0.19; I² = 90%, Analysis 15.1).

15.1. Analysis.

Comparison 15: ICAM, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient numbers of trials to report analyses of the categories of interventions (or combinations of these) on ICAM.

E‐selectin

It is unclear whether interventions targeting the systemic inflammatory response reduce E‐selectin concentrations in blood (4 studies, SMD ‐0.41, 95% CI ‐1.17 to 0.34; I² = 79% Analysis 16.1).

16.1. Analysis.

Comparison 16: E‐Selectin, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient numbers of trials to report analyses of the categories of interventions (or combinations of these) on E‐selectin.

C‐Reactive Protein

Interventions targeting the systemic inflammatory response may lower CRP levels post‐surgery, although the evidence is very uncertain due to high heterogeneity and imprecision (43 studies, SMD ‐0.24, 95% CI ‐0.51 to 0.04; I² = 90%, Analysis 17.1).

17.1. Analysis.

Comparison 17: CRP, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different effects on CRP levels (P = 0.02) (Analysis 17.1). There were insufficient trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

Tumour Necrosis Factor‐α

Interventions targeting the systemic inflammatory response may reduce TNF‐α, although the evidence is very uncertain due to high heterogeneity (61 studies, SMD ‐0.95, 95% CI ‐1.29 to ‐0.60; I² = 93 %, Analysis 18.1).

18.1. Analysis.

Comparison 18: TNFα, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different effect on TNF‐α levels (P < 0.0001) (Analysis 18.1). There were insufficient trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

Interleukin‐10

Interventions targeting the systemic inflammatory response may have little to no effect on IL‐10, although the evidence is very uncertain due to high heterogeneity (38 studies, SMD 0.19, 95% CI ‐0.28 to 0.65; I² = 94%; Analysis 19.1).

19.1. Analysis.

Comparison 19: IL‐10, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

On subgroup comparisons, the three intervention categories had different effects on IL‐10 levels (P = 0.0005) (Analysis 19.1). There were insufficient trials to report results for interventions targeting ischaemia reperfusion and for combinations of different categories of interventions.

Malondialdehyde

Interventions targeting the systemic inflammatory response may reduce malondialdehyde, but the evidence is very uncertain due to high heterogeneity (10 studies, SMD ‐2.52, 95% CI ‐3.43 to ‐1.60; I² = 94%; Analysis 20.1).

20.1. Analysis.

Comparison 20: Malondialdehyde, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient numbers of trials to report analyses of the categories of interventions (or combinations of these) on malondialdehyde.

Myeloperoxidase

Interventions targeting the systemic inflammatory response may increase myeloperoxidase levels in blood, but the evidence is very uncertain due to inconsistency and imprecision (9 studies, SMD 0.33, 95% CI ‐0.02 to 0.68; I² = 46%; Analysis 21.1).

21.1. Analysis.

Comparison 21: Myeloperoxidase, Outcome 1: Each of the three categories of intervention versus placebo or standard of care

There were insufficient numbers of trials to report analyses of the categories of interventions (or combinations of these) on myeloperoxidase.

Post hoc analyses

Three additional post hoc analyses were performed to address the severe heterogeneity in the meta‐analytical results obtained for inflammatory cytokines.

We restricted the analysis to trials at low risk of detection bias (Table 5), by selecting exclusively reports that adopted consensus definitions of organ injury (KDIGO 2012; Thygesen 2018). No significant treatment effect was evident for in‐hospital or 30‐day mortality (11 studies, RR 0.89, 95% CI 0.74 to 1.07; I² = 0%). Treatment effects on inflammation outcomes were limited by high heterogeneity; IL‐6 (16 studies, SMD ‐0.56, 95% CI ‐1.1 to ‐0.03; I² = 93%), IL‐8 (9 studies, SMD ‐0.54, 95% CI ‐1.33 to 0.25; I² = 95%). There was no significant treatment effect on myocardial injury (15 studies, RR 0.89, 95% CI 0.66 to 1.21; I² = 28%), kidney injury (15 studies, RR 0.96, 95% CI 0.80 to 1.16; I² = 21%), brain injury (7 studies, RR 0.93, 95% CI 0.72 to 1.20; I² = 0%), and lung injury (7 studies, RR 0.92, 95% CI 0.71 to 1.19; I² = 60%).

4. Analyses of effectiveness and efficacy outcomes after exclusion of trials at high risk of detection bias.

	No clinical detection bias		No biomarker detection bias
Outcome	N of studies	Treatment effect	N of studies	Treatment effect
Mortality	11	RR [95% CI], I2, P value	87	RR [95% CI], I2, P value
		0.8892 [0.7359, 1.0745], 0%, 0.2240		0.7880 [0.6866, 0.9043], 0%, 0.0007
IL‐6	16	SMD [95% CI], I2, P value	168	SMD [95% CI], I2, P value
		‐0.5641 [‐1.1002, ‐0.0281], 93%, 0.0391		‐0.9006 [‐1.0895, ‐0.7116], 92%, < 0.0001
IL‐8	9	SMD [95% CI], I2, P value	83	SMD [95% CI], I2, P value
		‐0.5428 [‐1.3348, 0.2492], 95%, 0.1792		‐0.8930 [‐1.1450, ‐0.6411], 91%, < 0.0001
Myocardial Injury	15	RR [95% CI], I2, P value	73	RR [95% CI], I2, P value
		0.8935 [0.6578, 1.2138], 28%, 0.1506		1.0392 [0.9527, 1.1337], 0%, 0.3857
AKI	15	RR [95% CI], I2, P value	59	RR [95% CI], I2, P value
		0.9629 [0.7971, 1.1631,] 20.5%, 0.2251		1.0627 [0.9678, 1.1670], 0%, 0.2025
Brain injury	7	RR [95% CI], I2, P value	49	RR [95% CI], I2, P value
		0.9262 [0.7163, 1.1975], 0%, 0.5585		0.8419 [0.6966, 1.0176], 0%, 0.0751
Lung injury	7	RR [95% CI], I2, P value	28	RR [95% CI], I2, P value
		0.9221 [0.7131, 1.1924], 60%, 0.5366		0.8898 [0.7636, 1.0368], 12.9%, 0.1344

AKI: acute kidney injury
CI: confidence interval
IL: interleukin
RR: relative risk
SMD: standardised mean difference

We also excluded trials reporting measurements of IL‐6 beyond the 99th percentile of the mean. Treatment effect were as follows: in‐hospital or 30‐day mortality (87 studies, RR 0.79, 95% CI 0.69 to 0.90; I² = 0%), IL‐6 (168 studies, SMD ‐0.90, 95% CI ‐1.09 to ‐0.71; I² = 92%), IL‐8 (83 studies, SMD ‐0.89, 95% CI ‐1.15 to ‐0.64; I² = 91%), myocardial injury (73 studies, RR 1.04, 95% CI 0.95 to 1.13; I² = 0%), kidney injury (59 studies, RR 1.06, 95% CI 0.97 to 1.17; I² = 0%), brain injury (49 studies, RR 0.84, 95% CI 0.70 to 1.02; I² = 0%), and lung injury (28 studies, RR 0.89, 95% CI 0.76 to 1.04; I² = 13%).

Finally, to assess whether the large number of different interventions may have contributed to heterogeneity, we calculated pooled treatment effects for the primary outcomes and some secondary outcomes for each intervention (Table 6). Heterogeneity remained substantial for inflammation outcomes across all interventions. Moreover, treatment effects on inflammation outcomes and primary and secondary clinical outcomes were inconsistent across all individual interventions. Specifically, no intervention demonstrated consistent treatment effects on inflammation AND clinical outcomes.

5. Effect on clinical outcomes and cytokines categorised by intervention.

	IL‐6	IL‐8	Mortality	Myocardial injury	Kidney injury	Brain injury	Lung injury
Intervention
Ultrafiltration	10 studies, ‐0.75 (‐1.40 to ‐0.09) I2 = 0.85	3 studies, ‐0.2 (‐1.38 to 0.98) I2 = 0	4 studies, 0.97 (0.24 to 3.9) I2 = 0	2 studies, 1.01 (0.18 to 5.7) I2 = 0	2 studies, 0.52 (0.07 to 3.78) I2 = 0%		1 study, 0.61 (0.04 to 8.45) I2 = NA
Corticosteroids	29 studies, ‐2.36 (‐2.80 to ‐1.93) I2 = 0.92	12 studies, ‐2.2 (‐2.83 to ‐1.49) I2 = 0.94	12 studies, 0.9 (0.41 to 1.96) I2 = 0	11 study, 0.78 (0.35 to 1.75) I2 = 0	9 studies, 0.97 (0.38 to 2.44) I2 = 0	9 studies, 0.82 (0.33 to 2.03) I2 = 0	8 studies, 0.97 (0 to 0.97) I2 = 0
Minimised extracorporeal circuits	18 studies, 0.07 (‐0.43 to 0.56) I2 = 0.91	4 studies, ‐0.67 (‐1.69 to 0.35) I2 = 0.79	12 studies, 0.81 (0.32 to 2.04) I2 = 0	8 studies, 0.81 (0.27 to 2.37) I2 = 0	7 studies, 1.42 (0.51 to 3.98) I2 = 0	8 studies, 0.65 (0.21 to 2.06) I2 = 0	4 studies, 1.23 (0 to 1.23) I2 = 0
Neutrophil elastase inhibitors	3 studies, ‐1.61 (‐2.82 to ‐0.41) I2 = 0.97	4 studies, ‐0.84 (‐1.89 to 0.21) I2 = 0.9	1 study, 0.5 (0.02 to 14.73) I2 = NA	2 studies, 0.74 (0.12 to 4.57) I2 = 0.28	1 study, 0.1 (0 to 3.7) I2 = NA	1 study, 0.17 (0 to 7.08) I2 = NA	2 studies, 0.53 (0 to 0.53) I2 = 0.79
Insulin‐glucose‐potassium therapy	4 studies, ‐0.72 (‐1.75 to 0.32) I2 = 0.97	3 studies, 0.11 (‐1.08 to 1.31) I2 = 0	1 study, 0.67 (0.05 to 9.2) I2 = NA	2 studies, 0.61 (0.05 to 7.29) I2 = 0	1 study, 0.21 (0.01 to 7.4) I2 = NA	1 study, 3.15 (0.08 to 27.76) I2 = NA
Discard mediastinal blood	2 studies, ‐2.15 (‐4.11 to ‐0.20) I2 = 0.97	2 studies, ‐0.3 (‐1.79 to 1.18) I2 = 0.77	2 studies, 0.67 (0.06 to 7.99) I2 = 0	1 study, 3 (0.07 to 22.99) I2 = NA	2 studies, 0.38 (0.04 to 3.83) I2 = 0
Statin therapy	8 studies, ‐1.56 (‐2.32 to ‐0.81) I2 = 0.96	3 studies, ‐0.79 (‐2.01 to 0.43) I2 = 0.81	5 studies, 1.68 (0.36 to 7.90) I2 = 0	6 studies, 1.36 (0.39 to 4.75) I2 = 0.35	4 studies, 1.56 (0.38 to 6.38) I2 = 0	3 studies, 1.28 (0.26 to 6.39) I2 = 0	3 studies, 0.68 (0 to 0.68) I2 = 0.17
Off‐pump coronary revascularisation	8 studies, ‐0.82 (‐1.57 to ‐0.06) I2 = 0.96	8 studies, ‐1.25 (‐2.01 to ‐0.50) I2 = 0.93	7 studies, 0.7 (0.18 to 2.65) I2 = 0	7 studies, 0.86 (0.27 to 2.71) I2 = 0	2 studies, 0.55 (0.05 to 6.67) I2 = 0	3 studies, 0.6 (0.1 to 3.78) I2 = 0.13	1 study, 0.29 (0.13 to 0.29) I2 = NA
Propofol anaesthesia	4 studies, 0.82 (‐0.29 to 1.93) I2 = 0.96	2 studies, 0.66 (‐1.01 to 2.34) I2 = 0.99	1 study, 0.12 (0 to 4.73) I2 = NA
Tranexamic acid	4 studies, ‐0.64 (‐1.66 to 0.38) I2 = 0.83		4 studies, 0.6 (0.1 to 3.46) I2 = 0	2 studies, 0.27 (0.02 to 3.42) I2 = 0.03	2 studies, 0.43 (0.07 to 2.7) I2 = 0	2 studies, 1.56 (0.13 to 18.36) I2 = 0	1 study, 0.73 (0 to 0.73) I2 = NA
Aprotinin	12 studies, ‐0.75 (1.40 to ‐0.10) I2 = 0.77	8 studies, 0.74 (‐1.51 to 0.02) I2 = 0.75	3 studies, 1.42 (0.28 to 7.28) I2 = 0	4 studies, 0.35 (0.06 to 1.85) I2 = 0	4 studies, 1.67 (0.41 to 6.76) I2 = 0	3 studies, 0.83 (0.16 to 4.43) I2 = 0	1 study, 2.21 (0 to 2.21) I2 = NA
Leukocyte‐depleting filters	8 studies, ‐0.5 (‐1.23 to 0.24) I2 = 0.86	7 studies, ‐0.6 (1.39 to 0.19) I2 = 0.66	4 studies, 0.79 (0.17 to 3.69) I2 = 0.05	5 studies, 0.81 (0.25 to 2.69) I2 = 0	2 studies, 0.42 (0.04 to 4.5) I2 = 0	1 study, 0.33 (0.01 to 13.69) I2 = NA	2 studies, 0.64 (0 to 0.64) I2 = 0
Bradykinin inhibitors	2 studies, ‐0.65 (‐2.08 to 0.78) I2 = 0.96	1 study, 0.27 (‐1.75 to 2.28) I2 = NA			1 study, 0.81 (0.09 to 7.34) I2 = NA	1 study, 0.57 (0.07 to 4.75) I2 = NA
Upregulation of NO signalling	6 studies, ‐1.08 (‐1.99 to ‐1.18) I2 = 0.95	4 studies, ‐0.66 (‐1.77 to 0.46) I2 = 0.96	2 studies, 0.5 (0.08 to 3.27) I2 = 0	2 studies, 0.8 (0.14 to 4.53) I2 = 0	3 studies, 0.71 (0.17 to 3.04) I2 = 0	2 studies, 1.06 (0.1 to 11.4) I2 = 0	1 study, 1.16 (0 to 1.16) I2 = NA
Biocompatible circuit coating	26 studies, ‐0.6 (‐1.04 to ‐0.17) I2 = 0.86	12 studies, ‐0.82 (‐1.45 to ‐0.19) I2 = 0.78	16 studies, 0.51 (0.24 to 1.08) I2 = 0	4 studies, 1.08 (0.26 to 4.37) I2 = 0	7 studies, 0.83 (0.25 to 2.75) I2 = 0.02	9 studies, 0.46 (0.17 to 1.26) I2 = 0
Sevoflurane anaesthesia	3 studies, ‐1.22 (‐2.48 to 0.05) I2 = 0.9	3 studies, ‐0.99 (‐2.21 to 0.24) I2 = 0.85		1 study, 2.5 (0.22 to 29.06) I2 = NA	1 study, 5 (0.14 to 73.53) I2 = NA
Minimally invasive cardiac surgery	2 studies, 0.54 (‐1.00 to 2.07) I2 = 0
Aminophylline/adenosine	5 studies, ‐0.14 (‐1.06 to 0.77) I2 = 0.78		2 studies, 4.68 (0.36 to 60.24) I2 = 0	2 studies, 1.27 (0.13 to 12.77) I2 = 0.26			1 study, 0.33 (0 to 13.33) I2 = NA
Complement inhibitors			6 studies, 0.52 (0.19 to 1.42) I2 = 0	4 studies, 0.81 (0.27 to 2.39) I2 = 0.38	3 studies, 1.04 (0.31 to 3.53) I2 = 0	1 study, 1.1 (0.14 to 8.34) I2 = NA
Remote ischaemic preconditioning	5 studies, ‐0.25 (‐1.17 to 0.67) I2 = 0.84	3 studies, ‐0.16 (1.35 to 1.03) I2 = 0.86	3 studies, 0.72 (0.13 to 4.11) I2 = 0	4 studies, 1.01 (0.28 to 3.59) I2 = 0.22	5 studies, 1.09 (0.39 to 3.02) I2 = 0	1 study, 0.99 (0.03 to 29.28) I2 = NA	2 studies, 0.66 (0.14 to 3.02) I2 = 0
Pericardial blood washing and auto‐transfusion	5 studies, ‐0.3 (‐1.23 to 0.63) I2 = 0.62	3 studies, ‐1.08 (‐2.30 to 0.14) I2 = 0.94	2 studies, 1.01 (0.12 to 8.45) I2 = 0	1 study, 0.31 (0.01 to 12.47) I2 = NA	2 studies, 3.84 (0.29 to 50.17) I2 = 0	4 studies, 1.64 (0.27 to 10.03) I2 = 0
Erythropoietin	2 studies, 0.1 (‐1.35 to 1.54) I2 = 0			1 study, 0.8 (0.08 to 7.93) I2 = NA
N‐acetylcysteine	1 study, ‐4.44 (‐6.74 to ‐2.14) I2 = NA		1 study, 0.5 (0.02 to 10.97) I2 = NA	2 studies, 0.71 (0.11 to 4.48) I2 = 0	1 study, 2.25 (0.23 to 21.84) I2 = NA	1 study, 1 (0.03 to 29.53) I2 = NA
Non‐cardioplegic arrest				1 study, 0.34 (0.01 to 14.21) I2 = NA
Direct thrombin inhibitors	3 studies, ‐0.95 (‐2.16 to 0.26) I2 = 0.62	2 studies, ‐1.48 (‐2.96 to 0.01) I2 = 0	2 studies, 0.34 (0.04 to 2.56) I2 = 0	2 studies, 0.66 (0.1 to 4.52) I2 = 0			1 study, 2.33 (0.24 to 22.2) I2 = NA
Dexmedetomidine	6 studies, ‐1.87 (‐2.83 to ‐0.90) I2 = 0.95	1 study, ‐3.77 (‐5.90 to ‐1.64) I2 = NA		1 study, 1 (0.07 to 15.12) I2 = NA

IL: interleukin
NA: not available
NO: nitrous oxide

Discussion

Summary of main results

In this systematic review, we assessed 328 studies involving 40,255 patients evaluating organ protection interventions targeting the innate inflammatory response with any control in adult patients undergoing cardiac surgery.

The results suggest that anti‐inflammatory treatments may reduce mortality, although the certainty of the evidence for this finding was very low due to the risk of bias in included trials, and inconsistency across sensitivity and post hoc analyses in trials at low risk of bias. Treatment estimates also suggested a significant reduction in inflammatory cytokines, however, high heterogeneity and the imprecision in the results for the included studies reduced the certainty of this finding to very low. There was evidence of treatment effects (reductions) on key secondary outcomes recorded during participants' perioperative stay including infection, brain injury and hospital length of stay, although the certainty of these estimates was limited by the risk of bias and imprecision.

Prespecified analyses of interventions clustered by the intended mechanism of action demonstrated that interventions targeting haematological activation reduced serum IL‐8 concentrations, in‐hospital or 30‐day mortality, and myocardial and brain injury. In contrast, interventions with nonspecific anti‐inflammatory effects had little or no treatment effect on mortality and may increase myocardial injury. Estimated treatment effects of treatments targeting ischaemia reperfusion injury were limited by the small number of trials. These results were not subject to the same scrutiny as those included in the GRADE analyses, were likely subject to the same limitations as the overall cohort of trials, and should be considered hypothesis generating.

Moderator analyses demonstrated that age, BMI, and recent myocardial infarction can moderate treatment responses to anti‐inflammatory interventions in terms of cytokine (IL‐6 and IL‐8) production, but the role of these moderators was inconsistent. No patient‐ or surgery‐related characteristics significantly altered the treatment effect on mortality.

No subgroup, sensitivity, moderator, or post hoc analysis explained the heterogeneity for estimates of treatment effects on inflammation biomarkers. For this reason, the precision of all of these analyses is uncertain.

Overall completeness and applicability of evidence

Despite our effort to include all relevant anti‐inflammatory interventions, our review focused on the most common treatments currently adopted in cardiac surgery, and might not include more recent strategies. Future updates of this work will expand the list of interventions.

Moreover, we aimed at categorising the anti‐inflammatory interventions into three groups to simplify the management strategy as a whole, instead of focusing on each specific treatment. Applying the evidence presented requires clinicians to further consider the role of each intervention for the specific patient subgroup (we partially addressed this with post hoc analyses). Finally, we did not classify the trial results based on the dosage and specifics of the treatment considered. This would be better explored in reviews related to individual interventions on the list.

Quality of the evidence

We rated the quality of evidence (summary of findings Table 1) to be very low for IL‐6, IL‐8, in‐hospital or 30‐day mortality, myocardial injury, and duration of hospital stay, and low for kidney injury and surgical site infection. It is, therefore, very likely that the treatment effect estimates will change as further evidence accumulates.

All clinical and laboratory outcomes were equally affected by selection and detection bias, and inspections of the forest plots revealed a potential publication bias in the primary outcomes. Cytokine assessments and hospital length of stay were further downgraded because of substantial inconsistency. The estimates of treatment effects for inflammation biomarkers were limited by detection and reporting bias, and heterogeneity that could not be explained by subgroup, sensitivity, and post hoc analyses. The moderator analysis demonstrated that some baseline conditions influence the innate immune system activation and responses to treatment, providing one potential explanation for this heterogeneity. Over the three decades considered by this review, people referred for cardiac surgery became increasingly elderly with more long‐term conditions (Hickey 2013). We did not include year of publication in our moderator analysis, however, this was shown to be a significant moderator of anti‐inflammatory treatment effects in a previous study (Abbasciano 2020). In individual patient data analyses, we have also demonstrated that the associations between specific inflammatory responses (IL‐6, IL‐8) and organ injury demonstrated significant interactions with long‐term conditions such as anaemia, old age, diabetes, or obesity (Abbasciano 2020). These findings point towards heterogeneous processes underlying post‐cardiac surgery that could explain some of our current observations.

The methodological quality of the studies in the review was generally higher in trials testing interventions with nonspecific anti‐inflammatory effects, which included several large multi‐centre, placebo‐controlled, double‐blinded RCTs such as DECS 2012, SIRS 2015 and Zheng 2016. Evidence of increased myocardial injury in this group requires further investigation.

Lastly, cohort characteristics were under‐reported or presented in a way that prevented us from conducting a robust meta‐regression for key baseline moderators such as creatinine, myocardial injury, heart failure, eGFR and haemoglobin. Individual patient data meta‐analyses might be better suited to investigate the way these factors influence the response to anti‐inflammatory interventions.

Potential biases in the review process

One author (GJM) was the research lead and senior author for two of the included studies (Kumar 2020; Murphy 2011) and, therefore, was not involved in making study eligibility decisions about, extracting data from, carrying out the risk of bias assessment for, or performing GRADE assessments for these specific trials.

Although the categorisation we adopted for anti‐inflammatory interventions is arbitrary and perhaps simplistic, it was established to provide a helpful overview of a wide list of interventions. We mitigated this limitation by providing an additional post hoc analysis based on individual interventions.

Also, our protocol did not include potential moderators which could inform regarding the surgical trauma‐induced inflammation (such as CPB and cross‐clamp time), which were nonetheless part of an IPD meta‐analysis conducted previously by our group (Abbasciano 2020).

Moreover, trials assessing perioperative interventions rarely report long‐term outcomes. Quality of life as well is markedly under‐reported in these investigations, which focus mostly on clinical outcomes related to the perioperative period and on cytokine changes. We acknowledge the lack of these long‐term and patient‐centred outcomes as a limitation of this review, and we will include these in the future updates of this work. We also strongly support the inclusion of these outcomes in the trials investigating this particular intervention category.

Lastly, due to the number of interventions included in the present work, our search strategy included outcome terms, increasing therefore the likelihood of retrieving the number of studies with favourable intervention effects (Tsujimoto 2022). We tried to mitigate this limitation by conducting additional analyses exploring the effect of potential reporting biases.

Agreements and disagreements with other studies or reviews

The inconsistency in treatment effects highlighted by our work has been shown as well in four large studies recruiting more than 1000 patients, that were identified in our search and included in the review.

In the international multi‐centre randomised double‐blind parallel‐group placebo‐controlled PRIMO‐CABG trial, Verrier and colleagues (Verrier 2004) compared the effects of pexelizumab versus placebo in 3099 people undergoing CABG with or without valve surgery at 205 hospitals. No significant reduction was detected in the composite endpoint of death or myocardial infarction for patients who had undergone exclusively CABG.

The DECS trial (DECS 2012) was a multi‐centre, randomised, double‐blind, placebo‐controlled trial conducted over five years in eight cardiac surgery centres. A total of 4482 patients aged 18 years or older undergoing cardiac surgery with cardiopulmonary bypass were randomised to receive either a single intraoperative dose of 1 mg/kg dexamethasone or placebo. The composite primary endpoint of death, myocardial infarction, stroke, renal failure or respiratory failure was not different between the two groups, although postoperative infections and postoperative stay were reduced in the dexamethasone group.

In the SIRS trial (SIRS 2015), 7507 high‐risk patients from 18 cardiac surgery centres in 18 countries were randomised to either intravenous methylprednisolone (3755) or placebo (3752). There was no difference between the groups with respect to the primary outcome of 30‐day mortality and a composite of death and major morbidity within 30 days.

The STICS trial (Zheng 2016) was a multi‐centre, placebo‐controlled, randomised controlled trial conducted from September 2011 until October 2013, in which 1922 patients who were scheduled to undergo elective cardiac surgery received either perioperative rosuvastatin (a dose of 20 mg daily) or placebo. The primary outcomes (rate of postoperative atrial fibrillation and degree of myocardial injury within 120 hours) were unchanged.

A previous systematic review by this group adopted a different protocol to focus on the causal relationship between organ injury and inflammatory response after cardiac surgery (Abbasciano 2020). Although the scope of the review was smaller than the current one, the primary analysis demonstrated inconsistency in the direction of the treatment effects on efficacy outcomes, similarly to the present work.

Recent systematic reviews (Dvirnik 2018) focused on pooling the evidence related to specific anti‐inflammatory strategies such as corticosteroids. The authors showed how steroids determined little to no difference versus placebo in mortality, with the added risk of myocardial injury for the intervention group.

A recent narrative review (Squiccimarro 2022) also considered genomic factors as potentially affecting inflammatory response to triggers such as CPB and surgery, but these were outside the scope of this review.

To our knowledge, this work is the first attempt at summarising the evidence related to anti‐inflammatory interventions on both the inflammatory response and clinical outcomes for adult cardiac surgery patients. Our results highlight the gap in the knowledge of their mechanisms of action and the inconsistency surrounding the clinical implications of their use.

Authors' conclusions

Implications for practice.

The certainty of the evidence supporting clinical benefit from anti‐inflammatory interventions is currently very low, due to methodological limitations in included trials, inconsistency in the trial results, and imprecision in the assessment of effectiveness.

Implications for research.

Interventions targeting the systemic inflammatory response have been extensively tested in the setting of cardiac surgery but, when considering low risk of bias trials, their effect on patient‐important outcomes is uncertain, despite potential effects on physiologic markers of inflammation. Further research should be conducted into tailoring specific interventions to address recognised risk factors for mortality and morbidity after cardiac surgery, such as age and chronic conditions. Moreover, in addition to assessing clinical outcomes related to the perioperative period, these trials should focus on long‐term outcomes such as long‐term overall survival and quality of life. Our work supports the conduction of rigorous randomised controlled trials investigating the effects in terms of organ injury of anti‐inflammatory strategies compared to standard care on elderly and frail patients.

History

Protocol first published: Issue 4, 2020

Appendices

Appendix 1. Search strategy

Database searched	Date searched	Number of results	Date updated	Number of results
CENTRAL Issue 9 of 12, 2022 (Cochrane Library)	12 May 2020	2541	08 October 2022	2855
Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, MEDLINE Daily and MEDLINE (Ovid, 1946 to May 11, 2020)	12 May 2020	1608	08 October 2022	1755
Embase (Ovid, 1980 to 2020 week 19)	12 May 2020	1952	08 October 2022	2264
CPCI‐S Web of Science (Clarivate Analytics, 1990 to 12 May 2020)	12 May 2020	200	08 October 2022	214
Total		6301		7088
After de‐duplication		3656		5144

CENTRAL

#1 MeSH descriptor: [Thoracic Surgery] this term only

#2 MeSH descriptor: [Cardiac Surgical Procedures] explode all trees

#3 ((heart or cardiac or coronary or cardio* or atrial or atrium or auricular or valv*) NEAR/2 (surg* or procedure* or operat*))

#4 ((coronary or heart or cardio* or cardiac* or valve) NEAR/3 (graft* or bypass or plasty or replace* or repair))

#5 CABG

#6 #1 or #2 or #3 or #4 or #5

#7 MeSH descriptor: [Systemic Inflammatory Response Syndrome] explode all trees

#8 inflam*

#9 SIRS

#10 #7 or #8 or #9

#11 #6 and #10 with Publication Year from 1993 to 2020, in Trials

MEDLINE Ovid

1 Thoracic Surgery/ (12624)

2 exp Cardiac Surgical Procedures/ (215774)

3 ((heart or cardiac or coronary or cardio* or atrial or atrium or auricular or valv*) adj2 (surg* or procedure* or operat*)).tw. (104597)

4 ((coronary or heart or cardio* or cardiac* or valve) adj3 (graft* or bypass or plasty or replace* or repair)).tw. (118894)

5 CABG.tw. (17650)

6 1 or 2 or 3 or 4 or 5 (312935)

7 exp Systemic Inflammatory Response Syndrome/ (125913)

8 inflam*.tw. (921307)

9 SIRS.tw. (5318)

10 7 or 8 or 9 (1032376)

11 6 and 10 (10530)

12 randomized controlled trial.pt. (505337)

13 controlled clinical trial.pt. (93661)

14 randomized.ab. (478817)

15 placebo.ab. (207590)

16 clinical trials as topic.sh. (191096)

17 randomly.ab. (332724)

18 trial.ti. (217914)

19 12 or 13 or 14 or 15 or 16 or 17 or 18 (1286683)

20 exp animals/ not humans.sh. (4697257)

21 19 not 20 (1183921)

22 11 and 21 (1634)

23 limit 22 to yr="1993‐current" (1608)

Embase Ovid

1 thorax surgery/ (32152)

2 exp heart surgery/ (349621)

3 ((heart or cardiac or coronary or cardio* or atrial or atrium or auricular or valv*) adj2 (surg* or procedure* or operat*)).tw. (146761)

4 ((coronary or heart or cardio* or cardiac* or valve) adj3 (graft* or bypass or plasty or replace* or repair)).tw. (157958)

5 CABG.tw. (32053)

6 1 or 2 or 3 or 4 or 5 (459697)

7 exp systemic inflammatory response syndrome/ (261223)

8 inflam*.tw. (1304395)

9 SIRS.tw. (9431)

10 7 or 8 or 9 (1531928)

11 6 and 10 (25436)

12 random$.tw. (1514948)

13 factorial$.tw. (37193)

14 crossover$.tw. (73950)

15 cross over$.tw. (31164)

16 cross‐over$.tw. (31164)

17 placebo$.tw. (300707)

18 (doubl$ adj blind$).tw. (201436)

19 (singl$ adj blind$).tw. (24484)

20 assign$.tw. (387598)

21 allocat$.tw. (149740)

22 volunteer$.tw. (248765)

23 crossover procedure/ (62898)

24 double blind procedure/ (169001)

25 randomized controlled trial/ (597887)

26 single blind procedure/ (38754)

27 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 (2280310)

28 (animal/ or nonhuman/) not human/ (5570082)

29 27 not 28 (2017865)

30 11 and 29 (2916)

31 limit 30 to yr="1993‐current" (2886)

32 limit 31 to embase (1952)

CPCI‐S

# 10 #9 AND #8 Timespan=1993‐2020

# 9 TS=(random* or blind* or allocat* or assign* or trial* or placebo* or crossover* or cross‐over*)

# 8 #7 AND #4

# 7 #6 OR #5

# 6 TS=SIRS

# 5 TS=inflam*

# 4 #3 OR #2 OR #1

# 3 TS=CABG

# 2 TS=((coronary or heart or cardio* or cardiac* or valve) NEAR/3 (graft* or bypass or plasty or replace* or repair))

# 1 TS=((heart or cardiac or coronary or cardio* or atrial or atrium or auricular or valv*) NEAR/2 (surg* or procedure* or operat*))

The RCT filter for MEDLINE is the Cochrane sensitivity‐maximising RCT filter, and for Embase, terms as recommended in the Cochrane Handbook have been applied. For both the reference is

Lefebvre C, Glanville J, Briscoe S, Littlewood A, Marshall C, Metzendorf M‐I, Noel‐Storr A, Rader T, Shokraneh F, Thomas J, Wieland LS. Chapter 4: Searching for and selecting studies. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019). Cochrane, 2019. Available from: https://training.cochrane.org/handbook/archive/v6

For the other databases, except CENTRAL, an adaptation of the Cochrane RCT filter has been applied.

Data and analyses

Comparison 1. In‐hospital or 30‐day mortality.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Each of the three categories of intervention versus placebo or standard of care	188	24592	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.67, 0.90]
1.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	132	11904	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.54, 0.85]
1.1.2 Interventions that attenuate ischaemia reperfusion injury	17	1016	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.30, 2.98]
1.1.3 Interventions with nonspecific anti‐inflammatory activity	40	11672	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.71, 1.06]
1.2 Each of the combinations of the three categories versus placebo or standard of care	2	195	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.23, 4.90]
1.2.1 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	195	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.23, 4.90]

1.2. Analysis.

Comparison 1: In‐hospital or 30‐day mortality, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 2. IL‐6.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Each of the three categories of intervention versus placebo or standard of care	149	6312	Std. Mean Difference (IV, Random, 95% CI)	‐0.78 [‐0.99, ‐0.58]
2.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	89	3830	Std. Mean Difference (IV, Random, 95% CI)	‐0.46 [‐0.68, ‐0.23]
2.1.2 Interventions that attenuate ischaemia reperfusion injury	18	862	Std. Mean Difference (IV, Random, 95% CI)	‐0.29 [‐0.87, 0.28]
2.1.3 Interventions with nonspecific anti‐inflammatory activity	43	1620	Std. Mean Difference (IV, Random, 95% CI)	‐1.82 [‐2.32, ‐1.33]
2.2 Each of the combinations of the three categories versus placebo or standard of care	4	173	Std. Mean Difference (IV, Random, 95% CI)	‐0.60 [‐1.55, 0.34]
2.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Std. Mean Difference (IV, Random, 95% CI)	0.24 [‐0.39, 0.87]
2.2.2 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	74	Std. Mean Difference (IV, Random, 95% CI)	‐0.60 [‐2.18, 0.99]
2.2.3 interventions attenuating ischaemia reperfusion + interventions with nonspecific anti‐inflammatory activity	1	60	Std. Mean Difference (IV, Random, 95% CI)	‐1.45 [‐2.02, ‐0.88]

2.2. Analysis.

Comparison 2: IL‐6, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 3. IL‐8.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Each of the three categories of intervention versus placebo or standard of care	72	2899	Std. Mean Difference (IV, Random, 95% CI)	‐0.91 [‐1.19, ‐0.63]
3.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	47	1902	Std. Mean Difference (IV, Random, 95% CI)	‐0.72 [‐0.99, ‐0.46]
3.1.2 Interventions that attenuate ischaemia reperfusion injury	9	373	Std. Mean Difference (IV, Random, 95% CI)	‐0.14 [‐1.07, 0.78]
3.1.3 Interventions with nonspecific anti‐inflammatory activity	17	624	Std. Mean Difference (IV, Random, 95% CI)	‐2.04 [‐2.95, ‐1.12]
3.2 Each of the combinations of the three categories versus placebo or standard of care	2	64	Std. Mean Difference (IV, Random, 95% CI)	‐1.51 [‐3.04, 0.03]
3.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Std. Mean Difference (IV, Random, 95% CI)	‐2.32 [‐3.28, ‐1.37]
3.2.2 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	34	Std. Mean Difference (IV, Random, 95% CI)	‐0.76 [‐1.46, ‐0.06]

3.2. Analysis.

Comparison 3: IL‐8, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 4. Myocardial injury: low cardiac output as defined by study authors.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Each of the three categories of intervention versus placebo or standard of care	85	23088	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.95, 1.14]
4.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	51	7007	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.74, 1.00]
4.1.2 Interventions that attenuate ischaemia reperfusion injury	8	467	Risk Ratio (M‐H, Random, 95% CI)	1.39 [0.86, 2.25]
4.1.3 Interventions with nonspecific anti‐inflammatory activity	27	15614	Risk Ratio (M‐H, Random, 95% CI)	1.14 [1.02, 1.27]
4.2 Each of the combinations of the three categories versus placebo or standard of care	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]
4.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]

4.2. Analysis.

Comparison 4: Myocardial injury: low cardiac output as defined by study authors, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 5. Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Each of the three categories of intervention versus placebo or standard of care	78	23064	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.96, 1.16]
5.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	49	7156	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.71, 1.24]
5.1.2 Interventions that attenuate ischaemia reperfusion injury	7	539	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.91, 1.27]
5.1.3 Interventions with nonspecific anti‐inflammatory activity	22	15369	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.77, 1.31]
5.2 Each of the combinations of the three categories versus placebo or standard of care	1	39	Risk Ratio (M‐H, Random, 95% CI)	0.21 [0.01, 4.11]
5.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Risk Ratio (M‐H, Random, 95% CI)	0.21 [0.01, 4.11]

5.2. Analysis.

Comparison 5: Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome), Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 6. Lung injury: acute lung injury/respiratory distress syndrome or tracheostomy as described by study authors.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Each of the three categories of intervention versus placebo or standard of care	31	14779	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.77, 1.02]
6.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	17	1690	Risk Ratio (M‐H, Random, 95% CI)	0.69 [0.41, 1.17]
6.1.2 Interventions that attenuate ischaemia reperfusion injury	2	266	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.52, 0.97]
6.1.3 Interventions with nonspecific anti‐inflammatory activity	12	12823	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.78, 1.15]
6.2 Each of the combinations of the three categories versus placebo or standard of care	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]
6.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]

6.2. Analysis.

Comparison 6: Lung injury: acute lung injury/respiratory distress syndrome or tracheostomy as described by study authors, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 7. Brain injury: stroke as described by study authors.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Each of the three categories of intervention versus placebo or standard of care	84	23365	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.69, 1.01]
7.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	61	7878	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.53, 0.97]
7.1.2 Interventions that attenuate ischaemia reperfusion injury	3	172	Risk Ratio (M‐H, Random, 95% CI)	Not estimable
7.1.3 Interventions with nonspecific anti‐inflammatory activity	20	15315	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.72, 1.17]
7.2 Each of the combinations of the three categories versus placebo or standard of care	2	200	Risk Ratio (M‐H, Random, 95% CI)	2.38 [0.36, 15.86]
7.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Risk Ratio (M‐H, Random, 95% CI)	3.15 [0.14, 72.88]
7.2.2 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	161	Risk Ratio (M‐H, Random, 95% CI)	2.02 [0.19, 21.89]

7.2. Analysis.

Comparison 7: Brain injury: stroke as described by study authors, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 8. Infection: surgical site infection (safety outcome).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Each of the three categories of intervention versus placebo or standard of care	40	19368	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.83, 1.01]
8.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	23	4667	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.66, 1.06]
8.1.2 Interventions that attenuate ischaemia reperfusion injury	3	217	Risk Ratio (M‐H, Random, 95% CI)	0.42 [0.08, 2.28]
8.1.3 Interventions with nonspecific anti‐inflammatory activity	14	14484	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.85, 1.05]

Comparison 9. Duration of hospital length of stay.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Each of the three categories of intervention versus placebo or standard of care	146	15313	Mean Difference (IV, Random, 95% CI)	‐0.52 [‐0.70, ‐0.34]
9.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	94	6017	Mean Difference (IV, Random, 95% CI)	‐0.66 [‐0.94, ‐0.38]
9.1.2 Interventions that attenuate ischaemia reperfusion injury	17	1069	Mean Difference (IV, Random, 95% CI)	‐0.48 [‐0.78, ‐0.17]
9.1.3 Interventions with nonspecific anti‐inflammatory activity	36	8227	Mean Difference (IV, Random, 95% CI)	‐0.31 [‐0.58, ‐0.03]
9.2 Each of the combinations of the three categories versus placebo or standard of care	3	234	Mean Difference (IV, Random, 95% CI)	‐0.25 [‐0.83, 0.34]
9.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.90, 0.70]
9.2.2 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	195	Mean Difference (IV, Random, 95% CI)	‐0.42 [‐1.30, 0.46]

9.2. Analysis.

Comparison 9: Duration of hospital length of stay, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 10. C3a.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 Each of the three categories of intervention versus placebo or standard of care	28	1767	Std. Mean Difference (IV, Random, 95% CI)	‐1.08 [‐1.46, ‐0.69]
10.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	28	1767	Std. Mean Difference (IV, Random, 95% CI)	‐1.08 [‐1.46, ‐0.69]

Comparison 11. C5a.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Each of the three categories of intervention versus placebo or standard of care	3	160	Std. Mean Difference (IV, Random, 95% CI)	‐0.43 [‐0.75, ‐0.12]
11.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	3	160	Std. Mean Difference (IV, Random, 95% CI)	‐0.43 [‐0.75, ‐0.12]

Comparison 12. C5b9/terminal complement complex.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 Each of the three categories of intervention versus placebo or standard of care	15	1153	Std. Mean Difference (IV, Random, 95% CI)	‐0.53 [‐0.85, ‐0.20]
12.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	15	1130	Std. Mean Difference (IV, Random, 95% CI)	‐0.53 [‐0.87, ‐0.19]
12.1.2 Interventions with nonspecific anti‐inflammatory activity	1	23	Std. Mean Difference (IV, Random, 95% CI)	‐0.49 [‐1.44, 0.45]
12.2 Each of the combinations of the three categories versus placebo or standard of care	1	34	Std. Mean Difference (IV, Random, 95% CI)	‐0.64 [‐1.33, 0.05]
12.2.1 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	34	Std. Mean Difference (IV, Random, 95% CI)	‐0.64 [‐1.33, 0.05]

12.2. Analysis.

Comparison 12: C5b9/terminal complement complex, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 13. D‐dimer.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
13.1 Each of the three categories of intervention versus placebo or standard of care	20	956	Std. Mean Difference (IV, Random, 95% CI)	‐1.17 [‐1.61, ‐0.73]
13.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	20	956	Std. Mean Difference (IV, Random, 95% CI)	‐1.17 [‐1.61, ‐0.73]

Comparison 14. Neutrophil elastase.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
14.1 Each of the three categories of intervention versus placebo or standard of care	30	1413	Std. Mean Difference (IV, Random, 95% CI)	‐0.96 [‐1.46, ‐0.47]
14.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	29	1376	Std. Mean Difference (IV, Random, 95% CI)	‐0.95 [‐1.46, ‐0.44]
14.1.2 Interventions with nonspecific anti‐inflammatory activity	2	37	Std. Mean Difference (IV, Random, 95% CI)	‐1.36 [‐4.15, 1.44]
14.2 Each of the combinations of the three categories versus placebo or standard of care	2	47	Std. Mean Difference (IV, Random, 95% CI)	‐1.50 [‐3.72, 0.73]
14.2.1 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	47	Std. Mean Difference (IV, Random, 95% CI)	‐1.50 [‐3.72, 0.73]

14.2. Analysis.

Comparison 14: Neutrophil elastase, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 15. ICAM.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
15.1 Each of the three categories of intervention versus placebo or standard of care	10	612	Std. Mean Difference (IV, Random, 95% CI)	‐0.80 [‐1.42, ‐0.19]
15.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	7	294	Std. Mean Difference (IV, Random, 95% CI)	‐1.14 [‐2.10, ‐0.17]
15.1.2 Interventions that attenuate ischaemia reperfusion injury	1	241	Std. Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.35, 0.16]
15.1.3 Interventions with nonspecific anti‐inflammatory activity	2	77	Std. Mean Difference (IV, Random, 95% CI)	‐0.14 [‐1.32, 1.04]

Comparison 16. E‐Selectin.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
16.1 Each of the three categories of intervention versus placebo or standard of care	4	155	Std. Mean Difference (IV, Random, 95% CI)	‐0.41 [‐1.17, 0.34]
16.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	3	111	Std. Mean Difference (IV, Random, 95% CI)	‐0.81 [‐1.21, ‐0.41]
16.1.2 Interventions with nonspecific anti‐inflammatory activity	1	44	Std. Mean Difference (IV, Random, 95% CI)	0.53 [‐0.07, 1.13]

Comparison 17. CRP.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
17.1 Each of the three categories of intervention versus placebo or standard of care	43	2500	Std. Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.51, 0.04]
17.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	26	1566	Std. Mean Difference (IV, Random, 95% CI)	0.10 [‐0.18, 0.39]
17.1.2 Interventions that attenuate ischaemia reperfusion injury	4	306	Std. Mean Difference (IV, Random, 95% CI)	‐0.23 [‐0.95, 0.48]
17.1.3 Interventions with nonspecific anti‐inflammatory activity	13	628	Std. Mean Difference (IV, Random, 95% CI)	‐0.99 [‐1.74, ‐0.25]
17.2 Each of the combinations of the three categories versus placebo or standard of care	2	200	Std. Mean Difference (IV, Random, 95% CI)	‐1.03 [‐1.32, ‐0.73]
17.2.1 Interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Std. Mean Difference (IV, Random, 95% CI)	‐1.04 [‐1.72, ‐0.37]
17.2.2 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	161	Std. Mean Difference (IV, Random, 95% CI)	‐1.03 [‐1.35, ‐0.70]

17.2. Analysis.

Comparison 17: CRP, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 18. TNFα.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
18.1 Each of the three categories of intervention versus placebo or standard of care	61	2609	Std. Mean Difference (IV, Random, 95% CI)	‐0.95 [‐1.29, ‐0.60]
18.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	37	1570	Std. Mean Difference (IV, Random, 95% CI)	‐0.28 [‐0.64, 0.08]
18.1.2 Interventions that attenuate ischaemia reperfusion injury	6	311	Std. Mean Difference (IV, Random, 95% CI)	‐1.12 [‐2.30, 0.05]
18.1.3 Interventions with nonspecific anti‐inflammatory activity	18	728	Std. Mean Difference (IV, Random, 95% CI)	‐2.34 [‐3.12, ‐1.56]
18.2 Each of the combinations of the three categories versus placebo or standard of care	2	70	Std. Mean Difference (IV, Random, 95% CI)	‐0.65 [‐2.96, 1.65]
18.2.1 Interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Std. Mean Difference (IV, Random, 95% CI)	‐1.85 [‐2.73, ‐0.98]
18.2.2 Interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	40	Std. Mean Difference (IV, Random, 95% CI)	0.50 [‐0.13, 1.13]

18.2. Analysis.

Comparison 18: TNFα, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 19. IL‐10.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
19.1 Each of the three categories of intervention versus placebo or standard of care	38	1808	Std. Mean Difference (IV, Random, 95% CI)	0.19 [‐0.28, 0.65]
19.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	18	933	Std. Mean Difference (IV, Random, 95% CI)	‐0.64 [‐1.18, ‐0.11]
19.1.2 Interventions that attenuate ischaemia reperfusion injury	7	365	Std. Mean Difference (IV, Random, 95% CI)	1.89 [0.07, 3.71]
19.1.3 Interventions with nonspecific anti‐inflammatory activity	13	510	Std. Mean Difference (IV, Random, 95% CI)	0.95 [0.16, 1.74]
19.2 Each of the combinations of the three categories versus placebo or standard of care	1	30	Std. Mean Difference (IV, Random, 95% CI)	0.39 [‐0.34, 1.11]
19.2.1 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Std. Mean Difference (IV, Random, 95% CI)	0.39 [‐0.34, 1.11]

19.2. Analysis.

Comparison 19: IL‐10, Outcome 2: Each of the combinations of the three categories versus placebo or standard of care

Comparison 20. Malondialdehyde.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
20.1 Each of the three categories of intervention versus placebo or standard of care	10	620	Std. Mean Difference (IV, Random, 95% CI)	‐2.52 [‐3.43, ‐1.60]
20.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	2	89	Std. Mean Difference (IV, Random, 95% CI)	‐5.62 [‐15.47, 4.23]
20.1.2 Interventions that attenuate ischaemia reperfusion injury	7	491	Std. Mean Difference (IV, Random, 95% CI)	‐2.06 [‐2.98, ‐1.15]
20.1.3 Interventions with nonspecific anti‐inflammatory activity	1	40	Std. Mean Difference (IV, Random, 95% CI)	‐1.98 [‐2.75, ‐1.21]

Comparison 21. Myeloperoxidase.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
21.1 Each of the three categories of intervention versus placebo or standard of care	9	316	Std. Mean Difference (IV, Random, 95% CI)	0.33 [‐0.02, 0.68]
21.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	8	281	Std. Mean Difference (IV, Random, 95% CI)	0.39 [0.01, 0.77]
21.1.2 Interventions with nonspecific anti‐inflammatory activity	1	35	Std. Mean Difference (IV, Random, 95% CI)	‐0.08 [‐0.75, 0.59]

Comparison 22. IL‐6 (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
22.1 Each category and combination of intervention versus placebo or standard of care	152	6485	Std. Mean Difference (IV, Random, 95% CI)	‐0.77 [‐0.97, ‐0.58]
22.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	89	3830	Std. Mean Difference (IV, Random, 95% CI)	‐0.46 [‐0.68, ‐0.23]
22.1.2 Interventions that attenuate ischaemia reperfusion injury	18	862	Std. Mean Difference (IV, Random, 95% CI)	‐0.29 [‐0.87, 0.28]
22.1.3 Interventions with nonspecific anti‐inflammatory activity	43	1620	Std. Mean Difference (IV, Random, 95% CI)	‐1.82 [‐2.32, ‐1.33]
22.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Std. Mean Difference (IV, Random, 95% CI)	0.24 [‐0.39, 0.87]
22.1.5 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	74	Std. Mean Difference (IV, Random, 95% CI)	‐0.60 [‐2.18, 0.99]
22.1.6 interventions attenuating ischaemia reperfusion + interventions with nonspecific anti‐inflammatory activity	1	60	Std. Mean Difference (IV, Random, 95% CI)	‐1.45 [‐2.02, ‐0.88]

Comparison 23. IL‐8 (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
23.1 Each category and combination of intervention versus placebo or standard of care	72	2963	Std. Mean Difference (IV, Random, 95% CI)	‐0.92 [‐1.20, ‐0.65]
23.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	47	1902	Std. Mean Difference (IV, Random, 95% CI)	‐0.72 [‐0.99, ‐0.46]
23.1.2 Interventions that attenuate ischaemia reperfusion injury	9	373	Std. Mean Difference (IV, Random, 95% CI)	‐0.14 [‐1.07, 0.78]
23.1.3 Interventions with nonspecific anti‐inflammatory activity	17	624	Std. Mean Difference (IV, Random, 95% CI)	‐2.04 [‐2.95, ‐1.12]
23.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Std. Mean Difference (IV, Random, 95% CI)	‐2.32 [‐3.28, ‐1.37]
23.1.5 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	34	Std. Mean Difference (IV, Random, 95% CI)	‐0.76 [‐1.46, ‐0.06]

Comparison 24. In‐hospital or 30‐day mortality (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
24.1 Each category and combination of intervention versus placebo or standard of care	190	24817	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.68, 0.91]
24.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	132	11904	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.54, 0.85]
24.1.2 Interventions that attenuate ischaemia reperfusion injury	17	1016	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.30, 2.98]
24.1.3 Interventions with nonspecific anti‐inflammatory activity	40	11672	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.71, 1.06]
24.1.4 Interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Risk Ratio (M‐H, Random, 95% CI)	Not estimable
24.1.5 Interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	195	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.23, 4.90]

Comparison 25. Myocardial injury: low cardiac output as defined by study authors (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
25.1 Each category and combination of intervention versus placebo or standard of care	86	23118	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.95, 1.14]
25.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	51	7007	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.74, 1.00]
25.1.2 Interventions that attenuate ischaemia reperfusion injury	8	467	Risk Ratio (M‐H, Random, 95% CI)	1.39 [0.86, 2.25]
25.1.3 Interventions with nonspecific anti‐inflammatory activity	27	15614	Risk Ratio (M‐H, Random, 95% CI)	1.14 [1.02, 1.27]
25.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]

Comparison 26. Kidney injury: stage 3 AKI or haemofiltration as defined by study authors (safety outcome) (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
26.1 Each category or combination of intervention versus placebo or standard of care	79	23103	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.96, 1.16]
26.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	49	7156	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.71, 1.24]
26.1.2 Interventions that attenuate ischaemia reperfusion injury	7	539	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.91, 1.27]
26.1.3 Interventions with nonspecific anti‐inflammatory activity	22	15369	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.77, 1.31]
26.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Risk Ratio (M‐H, Random, 95% CI)	0.21 [0.01, 4.11]

Comparison 27. Infection: surgical site infection (safety outcome) (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
27.1 Each of categories or combinations of intervention versus placebo or standard of care	40	19368	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.83, 1.01]
27.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	23	4667	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.66, 1.06]
27.1.2 Interventions that attenuate ischaemia reperfusion injury	3	217	Risk Ratio (M‐H, Random, 95% CI)	0.42 [0.08, 2.28]
27.1.3 Interventions with nonspecific anti‐inflammatory activity	14	14484	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.85, 1.05]

Comparison 28. Duration of hospital length of stay (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
28.1 Each of the categories or combinations of interventions versus placebo or standard of care	148	15547	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐0.69, ‐0.34]
28.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	94	6017	Mean Difference (IV, Random, 95% CI)	‐0.66 [‐0.94, ‐0.38]
28.1.2 Interventions that attenuate ischaemia reperfusion injury	17	1069	Mean Difference (IV, Random, 95% CI)	‐0.48 [‐0.78, ‐0.17]
28.1.3 Interventions with nonspecific anti‐inflammatory activity	36	8227	Mean Difference (IV, Random, 95% CI)	‐0.31 [‐0.58, ‐0.03]
28.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.90, 0.70]
28.1.5 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	2	195	Mean Difference (IV, Random, 95% CI)	‐0.42 [‐1.30, 0.46]

Comparison 29. Brain injury: stroke as described by study authors (pooled).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
29.1 Each category and combination of intervention versus placebo or standard of care	86	23565	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.70, 1.02]
29.1.1 Interventions that attenuate haematological activation by the CPB circuit or surgical field	61	7878	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.53, 0.97]
29.1.2 Interventions that attenuate ischaemia reperfusion injury	3	172	Risk Ratio (M‐H, Random, 95% CI)	Not estimable
29.1.3 Interventions with nonspecific anti‐inflammatory activity	20	15315	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.72, 1.17]
29.1.4 interventions attenuating haematological activations + interventions attenuating ischaemia reperfusion	1	39	Risk Ratio (M‐H, Random, 95% CI)	3.15 [0.14, 72.88]
29.1.5 interventions attenuating haematological activations + interventions with nonspecific anti‐inflammatory activity	1	161	Risk Ratio (M‐H, Random, 95% CI)	2.02 [0.19, 21.89]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abbas 2007.

Study characteristics
Methods	Single‐centre, prospective, randomised control trial Location: Pakistan Number of centres: 1 Run‐in period: February 2006 to April 2006
Participants	Ninety‐six patients undergoing CABG or valve replacement. Those under 14 years or with haemodynamic instability at the end of CPB were excluded.
Interventions	Intervention: MUF for 15 minutes after CPB Control: standard CPB without post‐CPB MUF
Outcomes	Post‐bypass Hb, postoperative bleeding, reopening, transfusion requirements, need for inotropic support, duration of ventilation, ICU stay, and overall morbidity
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information about randomisation; states 96 participants were 'randomly selected and divided into 2 groups', suggesting this could have been exposed to selection bias
Allocation concealment (selection bias)	Unclear risk	No description about allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding discussed; suspect they were not blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information provided on blinding of outcome assessors or who conducted outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant data provided as expected; no participants lost
Selective reporting (reporting bias)	Low risk	All expected data items reported
Other bias	Unclear risk	None noted

Abd El‐Hakeem 2003.

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind placebo‐controlled study Registration dates: not specified Number of study centres: 1 Location: Egypt
Participants	Twenty people undergoing elective mitral valve replacement or AVR with CPB People with severely impaired LV function (EF < 50%), ischaemic heart disease, low cardiac output syndrome, uncontrolled rapid AF, or tight mitral stenosis and those with severe systemic noncardiac disease, infectious disease, or poorly controlled diabetes mellitus, as well as those receiving corticosteroids or other immunosuppressive treatment, were excluded from the study.
Interventions	Intervention group: 100 mg dexamethasone before induction Control group: placebo control
Outcomes	Perioperative TNF‐α, IL‐6, IL‐8, IL‐10, and postoperative cardiac index, PCWP and pulmonary function
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was done by using a sealed‐envelope method, by third party.
Allocation concealment (selection bias)	Low risk	Randomisation performed by an anaesthesiologist not involved in the participants' care
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Intervention given by third party prior to commencement of procedure, and placebo given to controls
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information was given on blinding specifically, but randomisation was performed by a third party, and we have judged the overall risk as low because intervention and control placebo were given preprocedurally and a separate team delivered care.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All expected outcomes reported for 20 participants as expected from methodology
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported in line with methodology, and outcomes that would be expected from this study design were reported.
Other bias	Low risk	None noted

Abdel‐Rahman 2005.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: August 2003 and May 2003 Number of study centres: 1 Location: Germany
Participants	Two hundred four people listed for elective CABG One hundred one to intervention group (mean age 67.7 years, M 74) and 103 to standard treatment (mean age 66.6 years, M 86); none withdrawn Exclusion criteria: re‐operations or combined procedures, emergencies. Standard surgical risk.
Interventions	Pre‐op randomly assigned either to CorX system (n = 101, group A) or a standard CPB with cardiotomy reservoir (n = 103, group B) (CorX system is a closed extracorporeal circuit)
Outcomes	Intraoperative blood loss, retransfused cell‐saver blood, and chest drainage loss during the first 12 hours. The amount of transfused packed RBCs and FFP was evaluated in the first 24 hours. Respiratory function was assessed by routine lung function test (inspiratory vital capacity, FEV1) preoperatively and on the fifth POD. In addition, the oxygenation index (PaO2 in mmHg/FiO2 in percent) was analysed 1 hour and 3 hours after admission to the ICU. Renal function and myocardial protection were assessed by serum values of Cr and CK‐MB preoperatively as well as 1 hour, 6 hours, and 24 hours postoperatively. Polymorphonuclear elastase (PMNE, ELISA; Milenia Biotec, Bad Nauheim, Germany) and terminal complement complex (TCC, ELISA; Gambro, Hechingen, Germany) served to assess inflammatory response; blood samples for determining PMNE and TCC were taken preoperatively as well as after aortic declamping and 1 hour post‐CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information about randomisation; study states 204 participants randomly assigned to group A (intervention) or group B, but does not give further detail on how this was performed.
Allocation concealment (selection bias)	Unclear risk	Method of concealment is not described with sufficient details. Again, study reports randomisation but does not mention blinding or allocation concealment. It is unlikely to have been implemented but should not have significantly impacted outcomes.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information provided by study. Blinding is not discussed. It is unlikely to have been possible to blind the surgical teams, as the CPB circuits were of different types for intervention or controls, and that would be visible intraoperatively. However, participants are likely to have been blinded in this regard. The lack of personnel blinding is unlikely to have significantly impacted outcomes.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information. No information provided on who performed postoperative outcome assessment. Parameters assessed are unlikely to be impacted by lack of blinding, however, as they are largely objective (i.e. intraoperative blood loss volume).
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study, with no withdrawals.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported in line with methodology described, and expected outcomes for this study type were also reported.
Other bias	Low risk	No funding was disclosed, although the impact of this is not thought to be significant.

Al‐Ruzzeh 2003.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Number of centres: 1 Location: UK
Participants	Twenty patients undergoing elective isolated CABG
Interventions	Intervention: with CPB (on‐pump) Control: without CPB (off‐pump)
Outcomes	WCC, CD11, and IL‐8 levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomisation was carried out with a computer program that uses minimization and block randomization on the basis of patient age, sex, diabetes occurence and left ventrincular function"
Allocation concealment (selection bias)	High risk	Insufficient information to assess the process of allocation concealment (unlikely to be implemented owing to the nature of the intervention)
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding discussed and unlikely to be possible to blind personnel. Participants could have been blinded, although this is not specified.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No information is provided regarding who performed outcome assessment, although even if personnel were aware of the treatment allocations, this is unlikely to have impacted the objective outcome measures.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing participants or exclusions
Selective reporting (reporting bias)	Low risk	All expected outcomes reported for methodology and study type
Other bias	Low risk	None noted

Alam 2015.

Study characteristics
Methods	Single‐centre, prospective, randomised, placebo‐controlled, parallel‐group clinical trial Run‐in period: June 2011 and September 2013 Number of study centres: 1 Location: UK
Participants	Eighty‐seven consecutive patients undergoing elective CABG surgery, aged > 17 years and needing two or more grafts
Interventions	IV 200 mg elafin after induction of anaesthesia, prior to sternotomy; placebo, normal saline
Outcomes	Plasma elastase, serum elafin, MPO, IL‐6, IL‐8, troponin, and CRP
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Study reports that it is a double‐blinded, placebo‐controlled, parallel‐group clinical trial. It does not elaborate on how randomisation was performed but states it was performed by the Edinburgh Clinical Trials Unit. Therefore, it is reasonable to assume it is methodologically sound, in line with trial unit SOPs.
Allocation concealment (selection bias)	Low risk	Blinding of study drug preparation was performed, and although the study does not provide detail, it states that allocation concealment was ensured.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of staff involved with care of participants (double‐blinded)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of who undertook outcome assessment but does note that participants and staff were blinded to intervention
Incomplete outcome data (attrition bias) All outcomes	High risk	Four participants were excluded after randomisation (of total 87 randomised), and reasons for this are not given. It is unclear why these participants were excluded.
Selective reporting (reporting bias)	Unclear risk	CRP results not provided in detail like other outcome measures; otherwise, all other outcome measures reported as expected per the methodology and for a study of this nature
Other bias	Low risk	Proteo Biotech AG supported the study by supplying the elafin required free of charge. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Therefore, risk of bias in this domain is low.

Albacker 2008.

Study characteristics
Methods	Single‐centre, prospective, randomised trial Run‐in period: June 2005 to July 2006 Number of centres: 1 Location: Canada
Participants	Fifty‐two patients who presented for elective CAB were randomised to receive intraoperative IV insulin infusion, titrated to maintain blood glucose concentrations less than 180 mg/dL (group I standard, n = 25 control group), or receive intraoperative fixed high dose of IV insulin infusion (5 mU/kg/min) with dextrose 20% infused separately to maintain a blood glucose level between 70 and 110 mg/dL (group II intervention, n = 27 intervention group). Blood samples were collected at different time points to determine outcomes.
Interventions	Intraoperative fixed high dose of IV insulin infusion (5 mU/kg/min) with dextrose 20% infused separately to maintain a blood glucose level between 70 and 110 mg/dL (group II the high‐dose insulin therapy group/clamp group/intervention group, n = 27)
Outcomes	Inflammatory markers were studied, including IL‐6 and IL‐8, TNF‐α, and complement factor 3 and 4 (C3 and C4). These markers were collected preoperatively (on arrival at the operating room) and 0 hours, 6 hours, 24 hours, and 48 hours postoperatively. Additionally, the levels of IL‐6, IL‐8 were measured intraoperatively at the following time points: prior to pericardiotomy (prior to heparin administration), before starting CPB, and immediately after separation from CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed with use of computerised randomisation table.
Allocation concealment (selection bias)	Low risk	Blinded envelopes were opened sequentially by study personnel after participants signed the patient consent form.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Study personnel opened the envelopes dictating treatment intervention or control after participant consent and were therefore not blinded to the treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Medical personnel not blinded to intervention. There is insufficient information given as to who collected postoperative samples, so it is unclear whether these individuals were aware of treatment allocation. However, the collection of blood samples, even if personnel were aware of intervention received, should not significantly impact objective outcome measurements such as blood cell counts; therefore, it is unlikely to result in significant bias.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Of the 55 patients enrolled in the study, 27 were randomised to control and 28 to intervention of high‐dose insulin. Two participants were excluded from control owing to surgical approaches not in line with surgical protocol (for technical reasons – i.e. calcification of ascending aorta), and one participant was excluded after the development of mesenteric ischaemia and its known association with inflammation. Missing outcome data were balanced in numbers across intervention groups, with similar reasons for missing data across groups.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported as per the protocol and the expected outcomes to be reported in a study of this type.
Other bias	Low risk	None noted

Aldea 2002.

Study characteristics
Methods	Single‐centre, prospective, randomised trial Run‐in period: not specified Number of centres: 1 Location: USA
Participants	Thirty‐six patients undergoing first‐time, nonemergency CABG with CPB were randomly assigned to one of three treatment groups: group I, non‐heparin‐bonded circuits with the use of cardiotomy suction (n = 12); group II, Duraflo II (BCR‐3500; Jostra Bentley Corp, Irvine, Calif) heparin‐bonded circuits with cardiotomy suction (n = 12); and group III, Duraflo II heparin‐bonded circuits without cardiotomy suction (n = 12). Exclusion criteria: LVEF < 25%; allergy to heparin or heparin‐induced thrombocytopaenia; steroid treatment; inability to perform the neuropsychologic test battery; other vascular surgery planned with the CABG (such as aortic arch, carotid endarterectomy, or ventricular aneurysmectomy); dementia; alcohol or drug abuse; hereditary complement deficiency; any serious medical condition that would bias the study or endanger the patient with respect to taking part in the study; and inability to fully understand any of the aspects of the protocol or provide informed, written consent.
Interventions	Control group received non‐heparin‐bonded circuit with cardiology suction and two intervention groups received Duraflo heparin‐bonded circuit with or without cardiology suction (two separate intervention groups).
Outcomes	Laboratory markers for inflammation (PMN‐Elastase, IL‐6, and IL‐8), coagulation (PF‐1.2), Plt activation TG, PF‐4, soluble P‐selectin, terminal complement activation (C5b‐9), neuronal injury (NSE, S100), and fibrinolysis (tPA, PAI‐1) were assayed in all study participants. Blood samples were drawn at baseline (after anaesthesia induction, before bypass), at 5 minutes on bypass, after bypass (just after protamine), 4 hours after bypass, and on the first POD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A computer‐generated randomisation scheme was used to randomise the participants.
Allocation concealment (selection bias)	Low risk	Participants were assigned to one of three study groups by the perfusionist on the basis of a randomisation scheme that was produced preoperatively by the investigational drug pharmacist.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Postoperative but not preoperative teams were blinded to treatment allocation. This is unlikely to have had a significant impact on postoperative outcomes, which include objective blood level measurements, and all postoperative outcome‐dependent blood samples were drawn on the same postoperative day regardless of allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The postoperative team was blinded to treatment group assignment, and assessment was performed using blood drawn on POD 2.
Incomplete outcome data (attrition bias) All outcomes	High risk	Total exclusions 16.6%. Six participants were excluded before or during the operation because autologous blood from the cell saver was transfused during the operation, the time frame of surgery changed, the participant withdrew before the operation, the heparinisation protocol was not followed, or a ventricular aneurysm was discovered during the operation. These withdrawals could have had significant postoperative outcomes or change in operative course owing to the intervention used.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported per methodology described and in line with outcomes expected from this study type.
Other bias	Low risk	None noted

Allah 2019.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Number of study centres: 1 Location: Egypt
Participants	Sixty‐four adult patients with rheumatic heart disease undergoing cardiac valve replacement surgery
Interventions	Intervention: 80 mg atorvastatin 12 hours and 2 hours preoperatively, then on second, third, fourth, and fifth PODs Control: placebo at the same time points
Outcomes	The effect of statin on postop AF incidence. The secondary outcome was to inspect the anti‐inflammatory and cardioprotective effects as reflected in the WBC count, CRP, IL‐6, cardiac troponin, and postoperative inotropic score.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were equally and randomly allocated through computer generated tables into 2 groups"
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment or discussion of whether medical personnel treating the patients were aware of allocations
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Trial description states 'double‐blinded' but no further information given. It is likely that personnel and participants were aware of allocations, particularly as a placebo was used.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	States outcomes were measured by authors but mentions earlier (without detail) that the study was double‐blinded, so it is likely that treatment allocations were not known to those assessing outcomes.
Incomplete outcome data (attrition bias) All outcomes	High risk	Three participants discontinued intervention owing to medical complications; therefore, outcomes were not analysed. However, these complications could have been due to treatment allocation and should not have been excluded.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported per study type
Other bias	Low risk	None noted

Allen 2007.

Study characteristics
Methods	Single‐centre, prospective, randomised trial Run‐in period: not specified Number of centres: 1 Location: Ireland
Participants	Thirty‐seven OPCAB surgery patients were randomly allocated to control (retransfusion of unwashed shed blood) and treatment (retransfusion of washed shed blood or discarding of unwashed blood) groups.
Interventions	Intervention: treatment group n = 18 (retransfusion of washed shed blood or discarding of unwashed blood) and control group n = 19 (retransfusion of unwashed shed blood) Control: in the control group, all cardiotomy suction blood was collected in a cardiotomy suction reservoir and returned to the patient after completion of the last graft. In the treatment group, this shed blood was either discarded (if < 500 mL total volume) or processed in a Dideco (Mirandola, Italy) compact cell saver and then returned to the patient.
Outcomes	Circulating blood and urinary samples were obtained. In addition, in the treatment group, blood was withdrawn from the shed mediastinal blood before and after it was processed in the cell saver. Plasma samples were assayed for TNF‐α, IL‐8, and IL‐6, as well as the anti‐inflammatory cytokines, IL‐10, IL‐1ra, and TNF soluble receptor‐2 (TNFsr‐2). Samples were also assayed for serum creatinine. Shed mediastinal blood was assayed for TNF‐α, IL‐6, IL‐8, TNFsr‐2, fHB, and full blood count. Urinary samples were assayed for IL‐1ra, TNFsr‐2, creatinine,α‐1‐microglobulin, albumin, and β‐NAG. All cytokine kits were from RD Systems (Minneapolis, MN) and were sandwich enzyme‐linked immunosorbent assay kits.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly allocated to control (retransfusion of unwashed shed blood) or treatment (retransfusion of washed shed blood or discarding of unwashed blood) using an a priori random number generation software (GraphPad StatMate; San Diego, CA).
Allocation concealment (selection bias)	Unclear risk	No description about allocation concealment, although it states that blood was processed in line with treatment allocation, suggesting that personnel were not blinded to the allocations. This is unlikely to have had a significant impact on the outcomes reported if they were aware of allocations, though.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information about blinding is provided. Information provided suggested no blinding occurred.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information. No information about blinding is provided. Information provided suggested no blinding occurred.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing data are identified, and outcomes for all participants were reported. No participants were withdrawn from analysis.
Selective reporting (reporting bias)	Unclear risk	Expected outcomes from methodology were reported, as were outcomes one would expect from this study design. The study does note that it was not powered to detect clinical differences but that the treatment group had an increased requirement for Plt transfusion. The need for this and the transfusion in itself could introduce significant variation in the outcomes.
Other bias	Low risk	None noted

Almansob 2012.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: September 2010 and June 2011 Number of centres: 1 Location: China
Participants	Patients referred for elective noncoronary artery cardiac surgery between September 2010 and June 2011 were recruited, and 151 patients were eligible and were randomly assigned to either the statin group (n = 77) or control group (n = 74). Simvastatin (20 mg) was administered every day for 5 days to 7 days preoperatively, but not on the day of surgery in the statin group. Then simvastatin was re‐administered on the second day postoperatively. The control group was administered all the same routine medications as the statin group, such as digoxin and furosemide, but without statin therapy.
Interventions	Statin vs standard medical treatment excluding statin
Outcomes	Plasma TnT and the isoenzyme of creatine kinase (CK‐MB) levels were measured preoperatively and at 6 hours, 12 hours, 24 hours, and 72 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation performed using a random number produced by a computer
Allocation concealment (selection bias)	Low risk	Randomly assigned and all participants blinded, including the participants themselves; therefore, low risk
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The participants, surgeons, anaesthetists, perfusionists, ultrasound physicians, and individuals collecting the samples were all blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	People performing data analysis were blinded from intervention or control allocation.
Incomplete outcome data (attrition bias) All outcomes	High risk	Total loss: 12.6% of participants. From 151 patients, only 132 patients (68 in the statin group and 64 in the control group) completed the study.
Selective reporting (reporting bias)	Low risk	The outcomes expected from the methodology and from this study type were reported.
Other bias	Low risk	This study was supported by Sun Yat‐sen University Clinical Research 5010 Program; the National Natural Science Foundation of China (30971261 and 81170271); Ministry of Education of China (20100171110057); The Science and Technology Research for the Returned Overseas Chinese Scholars from Ministry of Human Resources and Social Security of China (2011); and Bureau of Science and Information Technology of Guangzhou, China (2010GN‐E00221). All these organisations appear to be governmental rather than industry; therefore, we have judged the risk of potential financial bias as low.

Ammar 2017.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: June 2012 and February 2014 Location: Egypt Number of centres: 1
Participants	Fifty ASA Class II or III patients scheduled for cardiac surgery with CPB
Interventions	The intervention group received controlled infusion of dexmedetomidine initiated 5 minutes before CPB until 6 hours after surgery, and the control group received placebo with an equivalent volume of normal saline infusion.
Outcomes	Static pulmonary thoracic and dynamic compliance and pulmonary oxygenation and blood samples were taken for IL‐8 or MDA and PMN counts.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation done using random number table in Microsoft Excel and central randomisation performed by an independent statistician
Allocation concealment (selection bias)	Low risk	An independent statistician performed randomisation and concealed this from participants and investigators throughout to ensure 'proper allocation concealment'.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Personnels and participants were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Fully blinded – one investigator was blinded regarding the anaesthetic study drug during conduct of the study by covering the infusion pump and lines and by numeric codes during the whole process of data evaluation. In addition, physicians who were in charge of postoperative care of the participants and their discharge from the ICU and hospital were effectively blinded to the study protocol.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Outcomes for all participants reported
Selective reporting (reporting bias)	Low risk	All outcome measures reported as outlined in methodology and as expected from study type
Other bias	Low risk	Not noted

An 2008.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Location: China Number of centres: 1
Participants	Thirty low‐risk patients (ASA status II‐III) scheduled for cardiac valve replacement
Interventions	Intervention: in patients in the propofol group, a target‐controlled infusion of propofol (Diprivan, Astra Zeneca, UK) was commenced and maintained to achieve a steady‐state target concentration of 8 μg/mL until 5 minutes after unclamping, at which time it was decreased to 4 μg/mL until the end of surgery. Control: approximately 15 minutes before aortic unclasping, patients in the control group received an infusion of normal saline.
Outcomes	Pulmonary compliance and oxygenation, as well as blood levels of IL‐8, MDA, PMN count
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were assigned in equal groups based on computer‐generated randomisation.
Allocation concealment (selection bias)	Unclear risk	No information was provided, although methodology suggests that allocation concealment was not performed because participants were receiving propofol or normal saline, which have grossly differing appearances.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information about blinding was provided, but methodology suggests that personnel would have been aware of allocations. This is unlikely to have severe impact but could have resulted in small changes in participants' treatment.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information about blinding was provided, but methodology suggests that personnel would have been aware of allocations. This is unlikely to have severe impact but could have resulted in small changes in participants' treatment or in the interpretation of subjective outcomes in terms of pulmonary compliance measurements.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing participant data
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None noted

Andreasen 2004.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Number of centres: 1 Location: Denmark
Participants	Forty‐six consecutive patients scheduled for elective, primary isolated CABG using CPB
Interventions	One group received TA (Cyklokapronw, Phizer Consumer Healthcare, Denmark) 1.5 g as an IV bolus via central venous catheter beginning at the induction of the anaesthesia, followed by a constant infusion of 200 mg/h until an additional 1.5 g was given. The other group received placebo (0.9% sodium chloride).
Outcomes	Blood samples were taken the day before the operation as baseline measurements, and these included Hb 312 and serum Cr concentrations, Hct, platelet number, activity of antithrombin III (AT3), APTT, INR, ACT, and plasma concentrations of b‐TG, PF4, fibrinogen, and D‐dimer. Blood samples were also drawn 6 and 12 hours after the operation. Plasma samples for analyses of b‐TG and PF4 were prepared and stored at ‐80 °C for later analyses. CK‐MB isoenzyme was measured 6 hours postoperatively and the next morning, with further measurements if the value exceeded 50 U/L. Primary outcome measures were postoperative blood loss and the proportion of participants requiring allogeneic transfusion. Bleeding was monitored as chest tube output at hourly intervals in the ICU until the next morning. The chest tube output from 0 hours to 6 hours and 0 hours to 12 hours postoperatively were registered. The amount of autotransfused shed mediastinal blood was recorded at 6 hours. The type of allogeneic transfusions (packed RBCs, platelets and FFP) during and after the operation was registered. Secondary outcome measures were the following: development of perioperative MI (peak CK‐MB 50 U/L, development of new Q waves, or both), acute renal insufficiency (Cr value twice the baseline or need for dialysis), transient ischaemic attacks or stroke, early mortality (30 days hospital mortality), and mediastinal infection within 30 days. The need for surgical re‐exploration within 24 hours and findings were recorded. Participants requiring re‐exploration were excluded from analyses regarding postoperative bleeding, need for allogeneic transfusions, and analyses of postoperative biochemical data.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Two groups in a double‐blinded, prospective, placebo‐controlled study by a random number sequence
Allocation concealment (selection bias)	Low risk	The randomisation schedule was provided in sealed envelopes, and preparation of the drug or placebo was carried out just prior to anaesthesia by a staff member not involved in the treatment of the participant.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study, so personnel and participants not aware of allocations
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcome assessors do not appear to be blinded; however, this is unlikely to have influenced the assessment of laboratory outcome.
Incomplete outcome data (attrition bias) All outcomes	High risk	Two participants in the TA group were excluded from the study after randomisation, as these participants had received clopidogrel or indometacin within 1 week prior to surgery. One unexpected death
Selective reporting (reporting bias)	High risk	Clinical outcome not properly addressed. Participants requiring re‐exploration were excluded from analyses regarding postoperative bleeding, need of allogeneic transfusions, and analyses of postoperative biochemical data.
Other bias	Low risk	None noted

Anic 2004.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: 1 March to 1 May 1996 Number of centres: 1 Location: Croatia
Participants	Twenty patients undergoing elective isolated coronary revascularisation. All participants prior to inclusion were carefully evaluated for signs of infection.
Interventions	Participants in the intervention group received a single dose of MP (30 mg/kg) prior to induction of anaesthesia, whereas the patients in the control group did not receive any corticosteroid either before or after the surgical procedure.
Outcomes	IL‐1 beta, IL‐6, IL‐8, TNF‐α, CRP
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors used a list of computer‐randomised odd and even numbers to allocate participants into two equally sized groups.
Allocation concealment (selection bias)	Unclear risk	No concealment was discussed, and methodology of using the computer‐generated list to randomise suggests that authors were probably aware of allocations. The bias this introduces is unclear – not likely to be very high but could be significant.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding information provided; suspect nonblinded from methodology description. This is unlikely to have had a major impact on the outcomes being assessed though.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No blinding information provided; suspect nonblinded from methodology description. This is unlikely to have had a major impact on the outcomes being assessed though, as participants were anaesthetised at the time of administration of intervention and are unlikely to have been aware of the intervention they themselves received.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant information adequately reported
Selective reporting (reporting bias)	Low risk	Two data‐based outcomes, TNF‐α and IL‐1 beta reported simply as consistently below detectable value and raw data not provided. This is likely to be a logistical issue with laboratory reporting outcomes and unlikely to represent underreported data.
Other bias	Low risk	None noted

Antunes 2008.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Number of centres: 1 Location: Brazil
Participants	Forty patients undergoing CABG
Interventions	Intervention group received ultrafiltration. Control group did not receive ultrafiltration during ECC. All patients received 30 mg/kg MP in each group.
Outcomes	SOFA score, GCS, liver function, renal function, hypotension, and TNF‐α, IL‐1α, IL‐6, IL‐8, C3a, and C4a
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote "Participants [sic] were assigned to one of two groups according to alternating designation".
Allocation concealment (selection bias)	Unclear risk	No concealment discussed, although interventions would be hard to conceal intraoperatively, so it is likely that personnel at least were aware of allocations, and method of randomisation could be subject to introduction of bias if the order in which participants were processed was chosen while knowing which allocation would be received next
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding information is provided. It is likely that personnel were aware intraoperatively of allocations, as intervention is hard to conceal.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No discussion of who assessed outcomes and no blinding information given. However, this is unlikely to have impacted outcomes significantly.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants reported upon
Selective reporting (reporting bias)	Low risk	All expected outcomes reported and no expected data missing
Other bias	Low risk	None noted

Aouifi 1999.

Study characteristics
Methods	Single‐centre, prospective controlled trial Run‐in period: not specified Number of centres: 1 Location: France
Participants	Thirty‐six patients undergoing elective cardiac surgery with or without CPB, in whom an uneventful postoperative course was anticipated
Interventions	Intervention group received CABG off‐pump (without CPB). Control groups underwent valve replacement or CABG on CPB.
Outcomes	Serum PCT and CRP levels, as well as SIRS assessment (clinical assessment, including body temperature, microbiological and radiological examinations, were performed daily until hospital discharge). SIRS was defined according to the classification of the American College of Chest Physicians/Society of Critical Care Medicine 13 when two or more of the following signs were present: body temperature > 38 °C or <36 °C; persistent tachycardia (heart rate >90 beats/min); tachypnoea (>20 respirations per minute); leucocytosis (leucocytes> 12,000 g/L) or leucopenia (leucocytes <4000 g/L).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	No randomisation information given, and states participants were 'allocated', suggesting no randomisation, and no information is given on how participants were initially identified to be included within the study.
Allocation concealment (selection bias)	High risk	No concealment discussed, and discussion of allocation suggests authors (and possibly participants) were aware of allocations
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding information provided, but suggestion from methodology that study was not blinded and personnel and participants could have been aware of allocations
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information provided regarding how outcome assessment was performed, and no suggestion of blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for in results reporting
Selective reporting (reporting bias)	Low risk	All expected data outcomes reported
Other bias	Low risk	Supported in part by Brahms, France, who supply testing equipment, but the same equipment was used for all participants and is unlikely to be able to influence results; therefore, deemed low risk of bias

Ascione 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised trial Number of centres: 1 Run‐in period: not specified Location: UK
Participants	Sixty patients (48 males, mean age 60.9 +/‐ 9.1 years) undergoing first‐time CABG
Interventions	Group A (on‐pump) underwent conventional myocardial revascularisation with CPB and cardioplegic arrest of the heart, whereas group B (off‐pump) underwent beating heart surgery without CPB.
Outcomes	Blood samples for inflammatory markers assay were collected at the anaesthetic induction, 30 min post‐cessation of CPB in the on‐pump group or 30 min after completion of all anastomoses in the off‐pump group, and thereafter at 4 hours, 12 hours, and 24 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by card allocation.
Allocation concealment (selection bias)	High risk	Insufficient information to assess the process of allocation concealment (unlikely to be implemented owing to the nature of the intervention)
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study states that some outcome measures (those for infection) were assessed by a blinded responsible physician. Otherwise, no blinding was reported. Per above, blinding personnel was not feasible. This is unlikely to have caused a significant bias on the objective blood level outcomes assessed.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	For some parameters, blinding reported as discussed above
Incomplete outcome data (attrition bias) All outcomes	Low risk	All expected participant outcomes reported
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly and in line with what would be expected for this study design.
Other bias	Low risk	This work was supported by the British Heart Foundation and the Garfield Weston Trust. These are independent organisations, and the risk of bias is low.

Ashraf 1997.

Study characteristics
Methods	Single‐centre, prospective randomised controlled trial Run‐in period: not specified Location: UK Number of centres: 1
Participants	Thirty‐eight patients undergoing CABG with effort‐induced angina pectoris refractory to maximal anti‐anginal therapy and multivessel disease (> 70% occlusion) but EF > 0.4
Interventions	Intervention: low‐dose aprotonin Control: placebo
Outcomes	TNF‐α, IL‐1b, IL‐6, and IL‐8, TCC, plasma neutrophil elastase, FBC, blood loss and transfusion, cardiac output and index, and PVR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation information is provided, although the study states that participants were "randomised"; therefore, the risk of major bias is likely to be low.
Allocation concealment (selection bias)	High risk	Insufficient information to assess the process of allocation concealment (unlikely to be implemented owing to the nature of the intervention)
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information is provided to suggest that this study was blinded. No information is given regarding who administered the intervention, and so it is not clear whether, for example, the surgical team performing surgery would be aware of allocations.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified. No detail provided regarding who performed blood draws for outcome assessment and whether they were present intraoperatively and so would know allocations. This is unlikely to have a significant impact on the objectively measured outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant data available
Selective reporting (reporting bias)	Low risk	All expected outcomes reported in line with methodology and as would be expected for this study design
Other bias	Low risk	None noted

Asimakopoulos 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind controlled trial Run‐in period: not specified Location: UK Number of centres: 1
Participants	Eighteen patients, undergoing primary elective CABG
Interventions	Intervention group: full‐dose aprotonin Control group: placebo
Outcomes	Leucocyte B1 and B2 integrins
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients [sic] were randomised into two groups in a double‐blind fashion". No further information regarding randomisation is provided.
Allocation concealment (selection bias)	Low risk	No allocation information is provided other than to state that participants were randomised in a double‐blind fashion. No information is provided regarding when the aprotinin was administered or by whom. On the basis of it being reported as double‐blind, we can assume the risk of bias here is low.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	States double‐blinded, although no additional information given regarding method of blinding or who was blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	States double‐blinded, although no additional information given regarding method of blinding or who was blinded. No information was provided regarding who performed outcome assessment and whether they were aware of allocations. Even if they were aware, this is likely to have minimal impact on the observed outcome parameters; therefore, we have deemed the risk to be low.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing participant data; results for all included participants reported.
Selective reporting (reporting bias)	Low risk	No missing data outcomes, and all expected outcomes were available
Other bias	Low risk	No information

Aydin 2015.

Study characteristics
Methods	Single‐centre, prospective randomised controlled study Location: Turkey Number of centres: 1 Run‐in period: not specified
Participants	Sixty consecutive patients who underwent elective isolated CABG from January to December 2012
Interventions	Intervention: each extubated patient was given 40 mg of atorvastatin per day, starting from an average of 6 hours after the operation, to the end of the first month. Control: placebo
Outcomes	CRP levels were assessed in all patients before and 1, 7, and 14 days after CABG; AF rates post‐op.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed in 2 blocks of 30 participants, without taking any of the participants' demographic characteristics into account. Personnel were not blinded, which introduces a risk of bias, although this did not impact selection because it was noted that participant characteristics were not considered during randomisation.
Allocation concealment (selection bias)	Low risk	The assigned therapy was single‐blinded; the individual participants did not know whether they were included in the study or control group.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	"The assigned therapy was single‐blinded; the individual subjects did not know whether they were included in the study or control group." Healthcare personnel were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Although personnel were not blinded, this is unlikely to introduce significant bias in terms of objective outcome assessments (binary diagnoses and measurements).
Incomplete outcome data (attrition bias) All outcomes	Low risk	No excluded participants; all participants reported as enroled
Selective reporting (reporting bias)	Low risk	No missing data points; expected data points for this study design were available.
Other bias	Low risk	None noted

Bakhtiary 2008.

Study characteristics
Methods	Single‐centre, prospective randomised trial Location: Germany Number of centres: 1 Run‐in period: not specified
Participants	Fifty patients undergoing elective CABG with a preoperative high risk
Interventions	Intervention: leucocyte depletion (leucocyte filter [Pall LeukoGuard6, LG6, Pall Biomedical, NY, USA] in the arterial line of the CPB) Control: no leucocyte depletion filter
Outcomes	Polymorphonuclear elastase, MPO, S‐100B and NSE, mortality, CK‐MB, TnT
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Study states that participants were enroled 'randomly' into intervention groups with no additional details. The impact of this is uncertain.
Allocation concealment (selection bias)	Unclear risk	No allocation information is provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding information is provided. It is anticipated that it would be hard to blind personnel to intervention; therefore, they likely were not blinded, but the impact of this is unclear.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information provided regarding who performed outcome assessment and whether they were blinded. This is unlikely to have significant impact on the outcome findings.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants missing/excluded
Selective reporting (reporting bias)	Unclear risk	Not all data of the outcomes prespecified in Methods section were reported. The study stated that the authors evaluated Hct, WBC, amount of donated blood products used, CPB duration, cross‐clamp time, mechanical ventilation time, ICU stay, and hospital stay. The results mention that Hct and leucocyte counts were similar in both groups and that there were no significant differences between groups for other clinical parameters. Some of these (i.e. WBC) are critical for assessing this specific intervention, thus numerical values should have been reported.
Other bias	Low risk	None noted

Baki 2013.

Study characteristics
Methods	Two‐centre, prospective randomised trial Location: Turkey Number of centres: 2 Run‐in period: not specified
Participants	Total of 40 adult patients who had CAD and were scheduled for coronary bypass surgery were included in the study.
Interventions	Control: desflurane (n = 20) anaesthesia with desflurane 1 MAC to maintain the state entropy value between 40 and 60 Intervention: propofol anaesthesia with continuous infusion of propofol 5 mg/kg/h
Outcomes	Blood samples for IL‐6, IL‐8, TNF‐α, and S100β: just prior to the operation before the induction of anaesthesia; before CPB; after CPB; and 4 hours postoperatively at the ICU.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised by envelope; no further description given
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information provided
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding not discussed or specified; suspect nonblinded study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding not discussed or specified; suspect nonblinded study
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants lost; outcomes for all participants as expected reported
Selective reporting (reporting bias)	Low risk	Study protocol of the project was registered, and outcomes reported were in line with this.
Other bias	Low risk	None noted

Baksaas 1998.

Study characteristics
Methods	Single‐centre, prospective randomised trial Run‐in period: not specified Number of centres: 1 Location: Norway
Participants	Forty patients with two‐vessel or three‐vessel disease, undergoing elective CABG at The National Hospital in Oslo, Norway, with EF > 40%
Interventions	Intervention: ECC using a leucocyte‐depleting filter (group L, n = 20) Control: ECC with no leucocyte filter (group C, n = 20)
Outcomes	Leucocytes, platelets, Hb concentration, complement activation (C3bc and TCC), Hct, granulocyte degradation (plasma MPO), IL‐6, and TNF‐α
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States randomised, but no randomisation information given and methodology not described
Allocation concealment (selection bias)	Unclear risk	No allocation information provided and no mention of blinding
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unlikely that personnel were blinded, owing to the difficulty of blinding their intervention. Participants were probably blinded, although this is not specified. Even if they were not blinded, this is unlikely to introduce significant bias to objective result measurements.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified; unlikely that personnel were blinded, owing to the difficulty of blinding their intervention. Participants were probably blinded, although this is not specified. Even if they were not blinded, this is unlikely to introduce significant bias to objective result measurements.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None specified

Balaguer 2013.

Study characteristics
Methods	Single‐centre, prospective randomised study Number of centres: 1 Location: USA Run‐in period: June 2007 to June 2012
Participants	Patients (N = 115) undergoing primary elective on‐pump cardiac surgery including CABG or valvular surgery
Interventions	Intervention 1: HOE 140 (also known as icatibant; bradykinin B2 receptor antagonist; Clinalfa, Switzerland, and AnaSpec, Freemont, CA) (IV bolus of 22 μg/kg over 30 minutes followed by an infusion of 18 μg/kg/h) Intervention 2: EACA or HOE 140, a bradykinin B2 receptor antagonist Control: placebo
Outcomes	D‐dimer, blood loss, blood product transfusion, t‐PA antigen, PAI‐1, inflammatory marker, and clinical outcome: re‐exploration for bleeding, prolonged ventilation, new‐onset postoperative AF, placement of a permanent pacemaker, AKI, length of hospital stay, or rate of readmission
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation schedule was generated using a statistical analysis software to generate blocks of three to six for three treatment arms.
Allocation concealment (selection bias)	Low risk	Allocation by the investigational pharmacy. Study drugs were all prepared and labelled in an identical fashion.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Trial was double‐blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Trial was double‐blinded, but no detail was provided regarding who performed outcome assessment. Study states that unblinding only occurred after trial protocol was completed; therefore, risk of bias is low.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no protocol violations for study drug administration. All participants completed the study according to ITT.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Baran 2012.

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind, placebo‐controlled trial Run‐in period: February 2010 to May 2010 Number of centres: 1 Location: Turkey
Participants	Sixty consecutive patients who underwent isolated, first‐time CABG on CPB (on‐pump)
Interventions	Intervention: 14‐day atorvastatin (40 mg/day) preoperatively Control: placebo preoperatively
Outcomes	Endothelial progenitor cells, troponin, CRP, cholesterol
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	An imperial statistician created a computer‐generated randomised list.
Allocation concealment (selection bias)	Low risk	Concealment was ensured through the use of sequentially numbered, sealed envelopes.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded; allocation concealment was ensured.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Treatment allocation was blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Total exclusions: 6%. Of 64 included in randomisation, four were excluded for more complex surgery or off‐pump technique but excluded equally from both groups prior to analysis; therefore, bias is minimal from this. One control group participant was excluded from analysis, but reasons for this were not explained. This may have introduced bias.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported properly
Other bias	Low risk	None noted

Barkhordari 2011.

Study characteristics
Methods	Single‐centre, double‐blind randomised clinical trial Run‐in period: January to June 2010 Location: Iran Number of centres: 1
Participants	Twenty‐eight consecutive patients undergoing elective CABG
Interventions	Intervention: PTX (a nonspecific phosphodiesterase inhibitor) 5 mg/kg IV bolus injection, followed by 1.5 mg/kg/h continuous IV infusion until 3 hours after cessation of CPB Control: placebo
Outcomes	UNGAL, serum creatinine, and CRP
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation information provided other than stating participants were randomised into their treatment or control groups
Allocation concealment (selection bias)	Unclear risk	States that study was double‐blinded, but no blinding or allocation information given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	States that study was double‐blinded, but no blinding or allocation information given. It is likely that participants and personnel were blinded. This is unlikely to contribute significant bias to objective result testing and therefore is deemed low risk.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	States that study was double‐blinded, but no blinding or allocation information given. It is likely that participants and personnel were blinded. This is unlikely to contribute significant bias to objective result testing and therefore is deemed low risk.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Bastani 2021.

Study characteristics
Methods	Single‐centre, prospective randomised clinical trial Run‐in period: not specified. Published 2021 Number of centres: 1 Location: Iran
Participants	One hundred patients who were undergoing on‐pump CABG in Nemazee hospital in Shiraz
Interventions	Intervention group: high‑dose (80 mg) atorvastatin daily for 3 days before surgery and low dose of 20 mg atorvastatin daily during the ICU stay, after surgery, and after discharge. The treatment continued at home with a single dose of 20 mg atorvastatin daily. Control group: low‑dose (20 mg) atorvastatin up to surgery. During the 1‐month follow‑up period, they continued their medication with 20 mg atorvastatin daily.
Outcomes	The primary outcomes of this study were TnI, CK‑MB, and hs‑CRP levels; the incidence of postoperative arrhythmia and ventricular fibrillation; and mechanical ventilation duration in the ICU. Secondary outcomes included DC shock frequency, ICU and hospital stay length, EF level, ICU blood intake, need for inotrope at the off‐pump time or in the ICU, blood glucose profile, liver status GFR, and urine output. Postoperative complications such as delirium rate, postsurgery bleeding, low cardiac output, haemodynamic instability (any perturbation in heart rate, blood pressure, or central vein pressure), major advanced cardiac events including death, nonfatal MI, repeated revascularisation incidence due to stroke, and cardiac arrest either during ICU stay or at 1 month of follow‐up were also recorded.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were selected by randomised block design and divided into two groups of A and B.
Allocation concealment (selection bias)	Low risk	The atorvastatin pills were prepared in separated packages (each package contains 3 pills of 80 mg or 3 pills of 20 mg atorvastatin). The participant, tester, and monitoring committee were unaware of which package was 80 mg or 20 mg atorvastatin, so the experiment was performed in a triple‐blind way through groups A and B.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Triple‐blind randomised clinical trial: participants and investigators were all blind to group allocation until the final statistical analysis was completed.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Triple‐blind randomised clinical trial: participants and investigators were all blind to group allocation until the final statistical analysis was completed.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported (CONSORT flowchart). Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Baseline data were not reported properly. Inflammatory outcomes without clear SD in the figures
Other bias	Low risk	There are no conflicts of interest.

Bauer 2010.

Study characteristics
Methods	Prospective, randomised controlled trial Number of centres: 1 Location: USA Run‐in period: not specified
Participants	Forty patients undergoing elective isolated CABG for double‐vessel, triple‐vessel, or four‐vessel disease; patients aged between 18 years and 80 years, weighing between 60 kg and 100 kg, and having an EF > 40%
Interventions	Intervention: MiECC Control: conventional CPB circuit
Outcomes	Regional tissue perfusion. Secondary endpoints were transfusion rate and Hct.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation list
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information provided; very difficult to conceal this intervention, so unlikely that personnel were blinded to allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information about blinding provided; given nature of the intervention, likely that personnel were not blinded. Participants may have been blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified; blinding not discussed. Personnel not present intraoperatively may still have been blinded to intervention/control; therefore, bias level is uncertain.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Adequate data reporting for all participants after randomisation
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in this paper.

Bauer 2018.

Study characteristics
Methods	Single‐centre, prospective randomised controlled study Run‐in period: October 2014 to June 2015 Location: Germany Number of centres: 1
Participants	Total N = 66 Prospective patients planned for elective CABG using MiECC
Interventions	Intervention: shed‐blood separation and cell‐saver Control: direct re‐transfusion of shed blood
Outcomes	TNF‐α, CRP, procalcitonin, and clinical outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation information provided beyond stating that participants were randomly allocated to intervention groups
Allocation concealment (selection bias)	Unclear risk	No specification regarding allocation concealment or description of any blinding, so reasonable to assume allocations were known
Blinding of participants and personnel (performance bias) All outcomes	High risk	No specification regarding allocation concealment or description of any blinding, so reasonable to assume allocations were known by personnel. Participants may have been blinded, but this was not specified.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No specification regarding allocation concealment or description of any blinding, so reasonable to assume allocations were known. Knowing the allocation is, however, unlikely to have significantly influenced objective outcome measures.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Ten participants were excluded perioperatively: five had a change of indication and additional surgery and fulfilled the exclusion criteria after written informed consent and randomisation were obtained, and one participant abstained from written informed consent. Four participants were excluded owing to matching perioperative exclusion criteria. None were excluded from analysis after intervention, however; therefore, there is uncertain risk of bias, as patients technically were excluded because of exclusion criteria.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly. Study protocol of the project was registered.
Other bias	Low risk	None noted

Baufreton 2005.

Study characteristics
Methods	Single‐centre, prospective randomised study Run‐in period: not specified Number of centres: 1 Location: France
Participants	Thirty patients undergoing elective primary CABG
Interventions	Intervention: heparin‐coated extracorporeal circuits Control: standard noncoated extracorporeal circuits
Outcomes	C5b‐9 and protein s100, cognitive functions
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by statistical analysis software (StatMate 1.01; GraphPad, San Diego, CA)
Allocation concealment (selection bias)	Unclear risk	"Investigators not in the operating room were blinded to the randomisation" suggests that healthcare personnel present in the operating theatre were aware of allocations. This introduces a small risk of bias.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	"Investigators not in the operating room were blinded to the randomisation" suggests those investigators present in the operating theatre were of allocations. This introduces a small risk of bias.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators who were involved in the biochemical and the neuropsychological assessment were blinded to the randomisation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Sample size is not clear: one participant in the NC group was removed from the study after completion of the protocol; two died (one in each group).
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Baumbach 2016.

Study characteristics
Methods	Single‐centre, prospective randomised study Run‐in period: March 2011 to April 2013 Location: Germany Number of centres: 1
Participants	Two hundred patients undergoing isolated minimally invasive valve procedures
Interventions	Intervention: minimally invasive group (MECC; n = 101) Control: CECC (n = 99)
Outcomes	Procalcitonin, IL‐6, IL‐8, IL‐10, TNF‐α, IFN‐g, and clinical outcomes, haptoglobin, cell‐free Hb
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States that participants were randomised without any additional detail
Allocation concealment (selection bias)	Unclear risk	No specification of allocation concealment or blinding
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No specification of allocation concealment or blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified who performed outcome assessment (although these are objective measurements, so even if the personnel performing this were unblinded, impact is likely minimal)
Incomplete outcome data (attrition bias) All outcomes	Low risk	Adequate data completion
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	Drs Baumbach, Ahad, and Franke disclose a financial relationship with Edwards Lifesciences. What impact this has on the findings is unclear. It is unlikely to introduce significant bias.

Belboul 2000.

Study characteristics
Methods	Single‐centre, prospective randomised, single‐blind, clinical trial Location: Sweden Number of centres: 1 Run‐in period: not specified
Participants	Thirty‐nine patients undergoing open‐heart surgery
Interventions	Intervention: CPB circuits with heparin (n = 20) Control: CPB or without heparin‐coating (n = 19)
Outcomes	Complement activation, neutrophil counts, MDA, and cardiac enzymes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stated that participants were randomised, but no detail provided
Allocation concealment (selection bias)	High risk	Allocations were not concealed from personnel.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Personnel were not blinded from allocations, although objective outcome measures such as blood levels are unlikely to be significantly altered by this awareness.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Adequate data completion
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	None noted

Bernardi 2016.

Study characteristics
Methods	Single‐centre, blinded, randomised controlled pilot study Location: Austria Number of centres: 1 Run‐in period: 10 September 2013 to 6 May 2015
Participants	Thirty‐seven patients undergoing elective CPB surgery
Interventions	Intervention: HA of cytokines used a new sorbent CytoSorb™ installed in the CPB circuit Control: standard CPB circuit
Outcomes	Differences in cytokine levels (IL‐1β, IL‐6, IL‐18, TNF‐α, and IL‐10), TNF‐α, 30‐day mortality, blood products, catecholamine support, and ICU stay
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed as block randomisation by a web‐based service (online Randomizer for Clinical Trials 1.7.0).
Allocation concealment (selection bias)	Unclear risk	Only participants were blinded.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Only single‐blind study. Personnel were aware of allocations.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified who performed outcome assessment sampling. Awareness of allocations is unlikely to have had significant impact on blood outcomes as objective.
Incomplete outcome data (attrition bias) All outcomes	High risk	Nine participants missing after randomisation; one participant missing from each group owing to loss of follow‐up. Other participants were excluded because of scheduling, and one for clinical deterioration. No ITT analysis is given.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	High risk	Travel funding from CytoSorbents Europe GmbH (i.e. the company that manufactures the product), which may explain the analysis plan excluding missing participants

Bert 2008.

Study characteristics
Methods	Single‐blind, prospective randomised clinical trial Location: Belgium Run‐in period: not specified Number of centres: 1
Participants	Fifty patients undergoing elective OPCAB surgery
Interventions	Intervention: aprotinin was given as a loading dose (2 × 106 KIU) followed by a continuous infusion at 5 × 105 KIU/h until skin closure. Control: standard treatment; no placebo given
Outcomes	cTnI; IL‐6, IL‐8, and IL‐10; TNF; CRP; serum Cr; and PMN elastase
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation list
Allocation concealment (selection bias)	Unclear risk	No allocation information provided and blinding not specified
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding specified or suggested
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Persons involved in blood sample analysis were not aware of allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Adequate data completion
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Bical 2006.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Number of centres: 1 Location: France Run‐in period: December 2002 to December 2003
Participants	Forty patients who underwent singular AVR
Interventions	Intervention: miniaturised CPB MECC1 system (Jostra) with heparin‐coated tubing (Bioline) Control: standard CPB circuit with copolymer surface modified SMARTxT tubing
Outcomes	Demographic and medical history and postoperative course, including DPB time, aortic cross‐clamp, and routine postoperative tests already collected, plus venous blood for leucocytes, neutrophils, CRP, IL‐1beta, IL‐6, IL‐10, soluble VCAM‐1, TNF‐α, and neutrophil elastase
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stated participants were randomly assigned with no further information or detail provided regarding how
Allocation concealment (selection bias)	High risk	No discussion of blinding or allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No discussion of blinding or allocation concealment, although it is likely that participants were blinded because intervention was intraoperative
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No discussion of blinding or allocation concealment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Bicer 2014.

Study characteristics
Methods	Single‐centre, prospective randomised study Location: Turkey Run‐in period: not specified Number of centres: 1
Participants	Fifty patients undergoing CABG surgery in the Department of Cardiovascular Surgery of the university's Faculty of Medicine. Patients with multivessel coronary disease were included if they were deemed, by means of consensus, equally treatable with on‐pump and off‐pump CABG.
Interventions	Control: CABG utilising CPB (on‐pump) Intervention: CABG without CPB (off‐pump)
Outcomes	Serum MDA, hs‐CRP, M30, M65, and clinical outcomes (mortality and morbidity)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States prospective randomisation but no further detail given
Allocation concealment (selection bias)	High risk	Unlikely to have been implemented, owing to the nature of the intervention
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to assess the process of allocation concealment (unlikely to be implemented, owing to the nature of the intervention)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified who performed the outcome assessment, including who took the samples
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing data
Selective reporting (reporting bias)	Low risk	All expected data were reported properly.
Other bias	Low risk	None noted

Boeken 2002 (A).

Study characteristics
Methods	Single‐centre, prospective, double‐blind, multi‐arm, randomised controlled trial Number of centres: 1 Run‐in period: not specified Location: Germany
Participants	Forty patients undergoing CABG. Reoperation, emergency intervention, an EF > 35%, and severe accompanying diseases were exclusion criteria. Ten into control group, 10 into each of 3 intervention arms
Interventions	Intervention: high‐dose aprotinin (Hammersmith protocol) Control: standard care, no placebo
Outcomes	Leucocyte count, plasma elastase, procalcitonin, and C‐esterase inhibitor
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States participants were randomised but no additional randomisation information provided
Allocation concealment (selection bias)	Unclear risk	No information provided, but states double‐blinded, so allocation is likely to have been concealed from participants and some personnel
Blinding of participants and personnel (performance bias) All outcomes	Low risk	States double‐blinded without further detail given
Blinding of outcome assessment (detection bias) All outcomes	Low risk	States double‐blinded but, again, no information provided regarding who performed outcome assessment and whether they were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant data reported as expected
Selective reporting (reporting bias)	High risk	Some key information required to make conclusions from the paper are lacking (i.e. basic demographic data regarding participant demographics within each group).
Other bias	Low risk	None noted

Boeken 2002 (B).

Study characteristics
Methods	Single‐centre, prospective, double‐blind, multi‐arm, randomised controlled trial Number of centres: 1 Run‐in period: not specified Location: Germany
Participants	Forty patients undergoing CABG. Reoperation, emergency intervention, an EF of > 35%, and severe accompanying diseases were exclusion criteria. Ten into control group, 10 into each of 3 intervention arms
Interventions	Intervention: prostacyclin, 10 ng/kg/min initially when starting operation, 20 ng/kg/min with beginning of ECC, and infusion finishing at the end of the operation Control: standard care, no placebo
Outcomes	Leucocyte count, plasma elastase, procalcitonin, and C‐esterase inhibitor
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States participants were randomised but no additional randomisation information
Allocation concealment (selection bias)	Unclear risk	No information provided but states double‐blinded, so allocation is likely to have been concealed from participants and some personnel
Blinding of participants and personnel (performance bias) All outcomes	Low risk	States double‐blinded without further detail given
Blinding of outcome assessment (detection bias) All outcomes	Low risk	States double‐blinded but, again, no information provided regarding who performed outcome assessment and whether they were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant data reported
Selective reporting (reporting bias)	High risk	Some key information required to make conclusions from the paper are lacking (i.e. basic demographic data regarding participant demographics within each group).
Other bias	Low risk	None noted

Boeken 2002 (C).

Study characteristics
Methods	Single‐centre, prospective, double‐blind, multi‐arm, randomised controlled trial Number of centres: 1 Run‐in period: not specified Location: Germany
Participants	Forty patients undergoing CABG. Reoperation, emergency intervention, an EF of > 35%, and severe accompanying diseases were exclusion criteria. Ten into control group, 10 into each of 3 intervention arms
Interventions	Intervention: combination of prostacyclin and aprotonin Control: standard care, no placebo
Outcomes	Leucocyte count, plasma elastase, procalcitonin, and C‐esterase inhibitor
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States participants were randomised but no additional randomisation information
Allocation concealment (selection bias)	Unclear risk	No information provided but states double‐blinded, so allocation is likely to have been concealed from participants and some personnel
Blinding of participants and personnel (performance bias) All outcomes	Low risk	States double‐blinded without further detail given
Blinding of outcome assessment (detection bias) All outcomes	Low risk	States double‐blinded but, again, no information provided regarding who performed outcome assessment and whether they were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participant data reported
Selective reporting (reporting bias)	High risk	Some key information required to make conclusions from the paper are lacking (i.e. basic demographic data regarding participant demographics within each group).
Other bias	Low risk	None noted

Boga 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised study Location: Turkey Run‐in period: not specified Number of centres: 1
Participants	Forty patients undergoing CABG
Interventions	Intervention: modified haemofiltration that was performed via ultrafilters, filtering particles of 15 kDa to 50 kDa in molecular weight and membrane materials made from polyacrylonitrile (AN 69) after CPB but before heparin neutralisation, during the rewarming period and continued for 10 minutes to 15 minutes Control: standard treatment without ultrafiltration
Outcomes	Blood concentrations of IL‐6, IL‐8, and neopterin and mortality
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States prospectively randomised without further information provided
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information provided
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; no suggestion of blinding. Intervention would be challenging to fully conceal intraoperatively. It is likely participants were blinded to the allocations.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No suggestion of blinding of outcome assessors, although the impact on objective blood test measures is likely to be minimal. Intervention would be challenging to fully conceal intraoperatively.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	None noted

Bourbon 2004 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; authors from Groupe Hospitalier Pitie‐Salpetriere, France
Participants	Thirty‐six adult patients undergoing cardiac surgery Mean age: 62 years vs 60 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes Exclusion criteria: redo operation, age < 50 years or > 80 years, weight < 60 kg or > 90 kg, renal or hepatic dysfunction, haematologic or coagulation disorders, infection during the week preceding surgery, preoperative use of antibiotics or corticosteroids, and WBC count > 11,000/mm3
Interventions	Group A: no treatment Group B: MP, 5 mg/kg body weight
Outcomes	Blood gas, WBC, and Plt counts; Hct, Hb, electrolyte (NaC, KC, Ca2C), BUN, Cr, serum glucose, and lactate levels; IL‐6; TNF‐α; and mortality Blood samples were taken from the arterial line at the following time points: after anaesthesia induction; before aortic cannulation; 10 minutes after aortic cross‐clamp release; 5 minutes after protamine sulphate administration; 4 hours after protamine sulphate administration; and 24 hours after protamine sulphate administration.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process is not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed or reported.

Bourbon 2004 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; authors from Groupe Hospitalier Pitie‐Salpetriere, France
Participants	Thirty‐six adult patients undergoing cardiac surgery were included. Mean age: 62 years vs 60 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes Exclusion criteria: redo operation, age < 50 years or > 80 years, weight < 60 kg or > 90 kg, renal or hepatic dysfunction, haematologic or coagulation disorders, infection during the week preceding surgery, preoperative use of antibiotics or corticosteroids, and WBC count > 11,000/mm3
Interventions	Group A: no treatment Group C: MP, 10 mg/kg body weight
Outcomes	Outcomes: WBC and Plt counts; Hct, Hb, electrolytes (NaC, KC, Ca2C), BUN, Cr, serum glucose, and lactate levels; IL‐6; TNF‐α; and mortality Blood samples were taken from the arterial line at the following time points: after anaesthesia induction; before aortic cannulation; 10 minutes after aortic cross‐clamp release; 5 minutes after protamine sulphate administration; 4 hours after protamine sulphate administration; and 24 hours after protamine sulphate administration.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process is not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed or reported.

Brettner 2019.

Study characteristics
Methods	Single‐centre, prospective, randomised, interventional, double‐blind pilot trial Run‐in period: June 2007 to August 2008 Registration date: 25 June 2007 Number of study centres and locations: single centre; University Hospital of Munich, Munich, Germany
Participants	Thirty patients undergoing cardiac surgery with CPB were included. Mean age: 67 years vs 64 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes Exclusion criteria: age < 18 years, pregnancy, preoperative IL‐6 levels > 10 pg/mL, hepatic insufficiency (bilirubin > 3 mg/dL), renal insufficiency (Cr > 2 mg/dL), a positive serologic test result for HIV, manifest insulin‐dependent diabetes mellitus, adipositas per magna (BMI > 30), use of steroidal or nonsteroidal ant‐inflammatory drugs during the last 7 days (apart from 100 mg acetylsalicylic acid per day), any septic focus, or chronic or acute inflammatory diseases
Interventions	Intervention: hydrocortisone (100 mg over 10 minutes) before surgery Control: placebo (saline control) before surgery
Outcomes	Outcomes: plasma concentrations of glycocalyx constituents (syndecan‐1, heparan sulphate), inflammatory markers (CRP and IL‐6), and various clinical parameters (mortality, respiratory and renal function, use of vasopressors, length of stay at the ICU) Plasma concentrations of constituent parts of the endothelial glycocalyx were determined preoperatively (T0), after induction of general anaesthesia (T1), 30 min after onset of surgery (T2), 30 min after end of CPB (T3), 1 h after termination of CPB (T4), and 4 h after termination of CPB (T5). Inflammatory parameters were measured preoperatively (CRP and leucocytes) and at PODs 1, 2, and 3 (IL‐6, CRP, and leucocytes), cytokine and CRP being determined in serum. Blood gas was analysed at every time point.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by block randomisation.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind study, but process not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	The present study was supported by the Else‐Kroner‐Fresenius foundation (charitable organisation).

Breuer 2015.

Study characteristics
Methods	Single‐centre, single‐blinded, prospective, randomised controlled trial Run‐in period: not specified Registration date: 17 January 2011 Number of study centres and locations: single centre; authors from University Hospital of the RWTH, Aachen, Germany
Participants	Thirty participants Mean age: 66 years vs 68 years Sex (female/male ratio): 18/12 Low‐risk patients (elective cardiac surgery) Inclusion criteria: adult patients undergoing elective cardiac surgery with the use of CPB were included, both sexes Exclusion criteria: emergency operations, known or suspected pregnancy, patients’ age < 18 years, and failure to obtain informed consent
Interventions	Intervention: intraoperative balanced anaesthesia using xenon (end‐expiratory concentrations of 45 % vol to 50 % vol) combined with continuous infusion of sufentanil (0.5 μg/kg/h to 1.5 μg/kg/h) Control: intraoperative sevoflurane (end‐expiratory concentrations of 1.0 % vol to 1.4 % vol) combined with continuous infusion of sufentanil (0.5 μg/kg/h to 1.5 μg/kg/h)
Outcomes	Primary outcomes: occurrence of adverse events Secondary outcomes: feasibility criteria (bispectral index, perioperative haemodynamic, and respiratory profile) and safety parameters (dosage of study treatments, renal function, intraoperative blood loss, need for inotropic support, and regional cerebral tissue oxygenation). Furthermore, at predefined time points, systemic and pulmonary haemodynamics were assessed by the use of a pulmonary artery catheter. Other outcomes: IL‐6 as well as IL‐8/CXCL8 and their anti‐inflammatory counterpart IL‐10, followed by the analysis of ORP Time points: preoperative, intraoperative, postoperative (24 hours) Serum samples from participants were drawn directly before induction of anaesthesia (preoperatively), immediately before termination of surgery (intraoperatively), and 24 hours postoperatively. We measured cytokine concentrations, determined serum ORP as an indicator of oxidative stress, and investigated serum influence on the migration of PBMCs.
Notes	Clinical study – Stoppe
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but the process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Single‐blinded to research staff performing mental score assessment and laboratory analyses; no blinding described for personnel delivering intervention or participants
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Single‐blinded to research staff performing mental score assessment and laboratory analyses
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up werenreported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	FG grants CO 799/3‐1 funding was disclosed or reported.

Brinkman 2015.

Study characteristics
Methods	Single‐centre, prospective, unblinded, randomised controlled trial Run‐in period: February 2010 to October 2011 Registration date: not specified Number of study centres and locations: single centre; authors from Texas, United States of America
Participants	One hundred two participants Mean age: 62.8 years vs 63.7 years vs 66.4 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective isolated CABG, both sexes Exclusion criteria: patients with contraindications for the use of CPB or the use of heparin
Interventions	Group A: intraoperative OPCAB Group B: Intraoperative MECC Group C: Intraoperative conventional ONCAB
Outcomes	Outcomes: preoperative, perioperative, and postoperative data (including RBC use) were recorded, and serial blood samples were collected at specified time points; serum IL‐6 and CRP values were measured per standard laboratory protocols. Time points: preoperative (baseline), skin closure, and 24 hours postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants quasi‐randomised and assigned by surgeon preference
Allocation concealment (selection bias)	High risk	Unblinded study
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded study
Blinding of outcome assessment (detection bias) All outcomes	High risk	No blinding; surgeon‐driven selection
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	Supported in part by charitable organisation

Bulow 2016.

Study characteristics
Methods	Prospective, randomised study Design: single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; authors from Universidade Federal de Santa Maria, Santa Maria, Brazil
Participants	Twenty‐three participants Mean age: 65 years vs 60 years Sex (female/male ratio): 8/15 Low‐risk patients (elective CABG) Inclusion criteria: 30 clinical patients ASA II to III who underwent CABG with mini‐CPB, both sexes Exclusion criteria: severe ventricular dysfunction (LVEF < 40%), re‐intervention surgery, the requirement for blood products from the start of CPB, preoperative history of liver or kidney dysfunction, immunological disease, preoperative intake of corticosteroids or anti‐inflammatory drugs (except for acetylsalicylic acid), and a history of recent MI (in the past 2 weeks)
Interventions	Group A: total IV anaesthesia with propofol and sufentanil Group B: total IV anaesthesia with propofol and sufentanil plus dexmedetomidine
Outcomes	Cytokines, CRP, and cardiac enzymes Outcomes: cytokines (IL‐1, IL‐6, IL‐10, TNF‐α, and INF‐γ), TBARS and ALA‐D activity in erythrocytes, and CRP levels Arterial blood was sampled at radial catheterisation before induction of anaesthesia (baseline), 90 minutes after starting mini‐CPB (during surgery), 5 hours after starting mini‐CPB (within 2 hours to 3 hours after the end of surgery), and 24 hours after the end of surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but the process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	High risk	Thirty participants recruited but only 23 reported and 7 withdrawn/not analysed
Selective reporting (reporting bias)	High risk	Outcomes for seven participants missing, but all outcomes reported
Other bias	Unclear risk	This study was supported by Fundação de Amparo a Pesquisa do Estado do Rio Grande do Sul (FAPERGS), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Rede Instituto Brasileiro de Neurociência ((FINEP (IBN‐Net) #01.06. 0842‐00), FAPERGS‐PRONEX‐ NPQ, Instituto Nacional de Ciência e Tecnologia em Excitotoxicidade e Neuroproteção (INCT‐EN), and FAPERGS‐PQ‐ AUCHO‐2014‐2015 (2251/14‐7).

Butler 2002.

Study characteristics
Methods	Single‐centre, double‐blinded, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; authors from Western Infirmary, Glasgow, United Kingdom
Participants	Forty participants Mean age: 61 years vs 60.5 years Sex (female/male ratio): 6/34 Low‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective, isolated CABG, both sexes Exclusion criteria: administration of corticosteroids or nonsteroidal anti‐inflammatory agents or IV heparin infusion
Interventions	Intervention: intraoperative Duroflo II heparin‐bonded bypass circuit Control: intraoperative conventional circuit
Outcomes	Outcomes: IL‐6, IL‐8, IL‐10, and PMN‐Elastase Blood samples were withdrawn from indwelling cannulas at induction of anaesthesia, skin closure, and 2 hours, 4 hours, 6 hours, 24 hours, and 48 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation described
Allocation concealment (selection bias)	Low risk	Double‐blinding without other details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of all study personnel; no details given on blinding of participants
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Cagli 2005.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; authors from Turkiye Yuksek Ihtisas Hospital, Ankara, Turkey
Participants	Sixty participants Mean age: 49 years vs 52 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective CABG, both sexes Exclusion criteria: recent MI, ventricle performance score ≥ 9, a history of systemic inflammatory disease, collagen tissue disease, the use of immunosuppressive agents, steroids, diltiazem or sodium nitroprusside, diabetes mellitus, renal disease, hepatic disease, haemorrhagic diathesis and coagulopathy, or a history of allergic reactions induced by PTX
Interventions	Group A: control – normal saline (500 mL 0.9% sodium chloride) Group B: intraoperative PTX (200 mg intravenously and 100 mg in cardioplegia)
Outcomes	Outcomes: CRP, IL‐6, and TNF‐α Blood samples were taken before PTX infusion (T0), after induction of anaesthesia (T1), 30 minutes after weaning off CPB (T2), 6 hours postoperatively (T3), and 24 hours postoperatively (T4).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but the process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No obvious trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Carrascal 2016.

Study characteristics
Methods	Single‐centre, interventional, open‐label, randomised, prospective, and parallel‐assignment trial Run‐in period: February 2011 to October 2013 Registration date: not specified Number of study centres and locations: single centre; University Hospital of Valladolid, Spain
Participants	Ninety participants Mean age: 67 years vs 65 years Sex (female/male ratio): not specified Low‐risk patients (elective heart valve surgery) Inclusion criteria: men and women > 18 years of age with sinus rhythm and primary diagnosis of heart valve disease (isolated or associated with CAD) satisfying the requirements for heart surgery under CPB; women of childbearing potential were asked to use effective contraception and commit to maintaining it throughout the study. Exclusion criteria: urgent surgery; endocarditis; patients with previous episodes of AF; beta‐blocker therapy at randomisation; severe LV dysfunction (LVEF < 30%); chronic use of nonsteroidal anti‐inflammatory drugs, corticosteroids, or both; uncontrolled thyroid disease; previous statin therapy; active liver disease, history of previous chronic liver disease, or both; alcoholism; predisposing factors to statin side effects such as increased transaminase levels at baseline (3 times normal value), renal failure (Cr levels > 42 mg/dL), previous diagnosis of myopathy of any aetiology, and known hypersensitivity to calcium atorvastatin or lactose monohydrate or both; women with positive pregnancy test result on the day of inclusion in the study; unsigned informed consent form; and inability to understand study objectives. The exclusion criteria after study initiation were consent withdrawal and changes in liver function laboratory parameters (transaminases > 3 times the normal value) or CPK level suggesting adverse effects of statins.
Interventions	Intervention: 40 mg/day of atorvastatin 7 days before and after the surgery Control: no treatment
Outcomes	Primary outcomes: efficacy of atorvastatin in preventing POAF after valve surgery Secondary outcomes: correlation between variation of the inflammatory markers during CPB and POAF incidence, indicating a possible atorvastatin modulator effect; modifications in echocardiographic parameters (owing to atorvastatin) conventionally related to POAF; frequency, duration, and clinical characteristics of AF after heart valve surgery; clinical and haemodynamic consequences of POAF; and economic repercussion, depending on hospital stay prolongation Time points: 0 hours, 1 h, 6 h, 24 h, 48 h, and 72 h
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation performed (at the time of participant inclusion) by an independent statistician; no further information
Allocation concealment (selection bias)	Low risk	Randomiser and data collector were both blinded.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unclear whether participants were blinded to the treatment allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Data assessors were blinded to the treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	The research was supported by a grant from the 'Gerencia de Salud, Consejería de Sanidad, Junta de Castilla y Leon' [GRS 308/A/ 08] and ‘Caja Burgos Foundation’. The funding organisation had no involvement in the study delivery and analysis.

Casati 2004 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: 3 January 2001 to 30 June 2001 Registration date: not specified Number of study centres and locations: single centre; San Raffaele Hospital, Milan, Italy
Participants	Fifty‐one participants Mean age: 61 years vs 64 years Sex (female/male ratio): 10/41 Low‐risk patients (elective CABG – OPCAB and ONCAB) Inclusion criteria: consecutive patients undergoing OPCAB and consecutive patients undergoing ONCAB only if three or fewer distal anastomoses were to be performed, single critical stenosis of the anterior descending coronary artery, without the possibility of a catheter‐based intervention, restenosis after angioplasty or stent placement, or multivessel coronary disease in patients for whom CPB was contraindicated because of associated comorbid conditions (severe COPD, cerebrovascular diseases, severe peripheral vascular disease), both sexes Exclusion criteria: history of haematologic disease, chronic renal insufficiency (Cr level > 2 mg/dL), and liver disease (active chronic hepatitis or cirrhosis). Preoperative treatment with aspirin or subcutaneous low–molecular‐weight heparin was not a contraindication to inclusion in the study.
Interventions	Intervention: TXA and OPCAB (IV) Control: placebo (equivalent volume of saline solution) and OPCAB
Outcomes	Primary outcomes: bleeding in the first 24 postoperative hours Secondary outcomes: requirement for allogeneic transfusions, thrombotic complications, outcomes, and monitoring of coagulation, fibrinolysis, and inflammation Blood samples were taken at the following times: after the induction of anaesthesia (time 1), 10 minutes after patient arrival in the ICU (time 2), and 24 hours after the operation (time 3).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated algorithms described briefly
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The staff in the operating room and the intensive care unit (ICU) were blinded regarding treatment: the correct procedure was ensured by means of coded syringes, prepared by a fellow in anesthesia not directly involved in the perioperative patient care and in the treatment of clinical data". It is unlikely that participants were not blinded from description.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	Supported in part by a grant of the Italian Ministry of Health (contract ICS 160.3/ RF99.88)

Casati 2004 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: 3 January 2001 to 30 June 2001 Registration date: not specified Number of study centres and locations: single centre; San Raffaele Hospital, Milan, Italy
Participants	Fifty‐one participants Mean age: 60 years vs 64 years Sex (female/male ratio): 6/45 Low‐risk patients (elective CABG – OPCAB and ONCAB) Inclusion criteria: consecutive patients undergoing OPCAB and consecutive patients undergoing ONCAB only if three or fewer distal anastomoses were to be performed, single critical stenosis of the anterior descending coronary artery, without the possibility of a catheter‐based intervention, restenosis after angioplasty or stent placement, or multivessel coronary disease in patients for whom CPB was contraindicated because of associated comorbid conditions (severe COPD, cerebrovascular diseases, severe peripheral vascular disease), both sexes Exclusion criteria: history of haematologic disease, chronic renal insufficiency (Cr level > 2 mg/dL), and liver disease (active chronic hepatitis or cirrhosis). Preoperative treatment with aspirin or subcutaneous low–molecular‐weight heparin was not a contraindication to inclusion in the study.
Interventions	Intervention: TXA and ONCAB (IV) Control: placebo (equivalent volume of saline solution) and ONCAB
Outcomes	Primary outcomes: bleeding in the first 24 postoperative hours Secondary outcomes: requirement for allogeneic transfusions, thrombotic complications, outcomes, and monitoring of coagulation, fibrinolysis, and inflammation Blood samples were taken at the following times: after the induction of anaesthesia (time 1), 10 minutes after patient arrival in the ICU (time 2), and 24 hours after the operation (time 3).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated algorithms described briefly
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The staff in the operating room and the intensive care unit (ICU) were blinded regarding treatment: the correct procedure was ensured by means of coded syringes, prepared by a fellow in anesthesia not directly involved in the perioperative patient care and in the treatment of clinical data". It is unlikely that participants were not blinded from description.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	Supported in part by a grant of the Italian Ministry of Health (contract ICS 160.3/ RF99.88)

Castano 2015 (A).

Study characteristics
Methods	Single‐centre, prospective, double‐blind, placebo study, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Salamanca University Hospital, Salamanca, Spain
Participants	Thirty participants divided into three groups Mean age: 64.6 years vs 66.9 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients with dyslipidaemia under chronic (> 15 days) statin treatment scheduled for non‐emergent on‐pump CABG under aortic cross‐clamp; unclear whether both sexes were included Exclusion criteria: haemodynamic instability (cardiogenic shock, acute pulmonary oedema, inotropic or IABP support), emergent/salvage surgery, rest angina in the previous 48 hours or AMI 4 weeks before the procedure, preoperative renal or liver insufficiency, pravastatin allergy, concomitant medications that could increase pravastatin levels and adverse effects (cyclosporine, erythromycin, clarithromycin, nicotinic acid, fibrates), immunosuppressive therapy or disease, concomitant inflammatory processes, alcohol abuse, or pregnancy
Interventions	Group A: pravastatin 80 mg administered 2 hours before anaesthetic induction Group C: placebo
Outcomes	Outcomes: drug‐related adverse effects; myocardial damage markers (Tn‐I and CPK‐MB); liver, renal, and musculoskeletal adverse effects (CPK, AST, ALT, and Cr), superoxide anion, TNF‐α, IFNg, ICAM‐1, P‐selectin, L‐selectin, CD18, IL‐10, NF‐kB, NO, and iNOs Blood samples were collected at pre‐CPB, 30 minutes after cross‐clamp release, 24 hours, and day 5.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation, but no further details
Allocation concealment (selection bias)	Low risk	Sealed envelope used by staff pharmacologist and no further access to patient management of data
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All investigators and participants were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details given, but likely blinding was maintained
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Castano 2015 (B).

Study characteristics
Methods	Single‐centre, prospective, double‐blind, placebo study, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Salamanca University Hospital, Salamanca, Spain
Participants	Thirty participants divided into three groups Mean age: 64.6 years vs 66.9 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients with dyslipidaemia under chronic (> 15 days) statin treatment scheduled for non‐emergent on‐pump CABG under aortic cross‐clamp; unclear whether both sexes were included Exclusion criteria: haemodynamic instability (cardiogenic shock, acute pulmonary oedema, inotropic or IABP support), emergent/salvage surgery, rest angina in the previous 48 hours or AMI 4 weeks before the procedure, preoperative renal or liver insufficiency, pravastatin allergy, concomitant medications that could increase pravastatin levels and adverse effects (cyclosporine, erythromycin, clarithromycin, nicotinic acid, fibrates), immunosuppressive therapy or disease, concomitant inflammatory processes, alcohol abuse, or pregnancy
Interventions	Group B: pravastatin 40 mg administered 2 hours before anaesthetic induction Group C: placebo
Outcomes	Outcomes: drug‐related adverse effects; myocardial damage markers (Tn‐I and CPK‐MB); liver, renal, and musculoskeletal adverse effects (CPK, AST, ALT, and Cr); superoxide anion, TNF‐α, IFNg, ICAM‐1, P‐selectin, L‐selectin, CD18, IL‐10, NFkB, NO, and iNOs Blood samples were collected at pre‐CPB, 30 minutes after cross‐clamp release, 24 hours, and day 5.
Notes	Castano 2015 split into A and B – 2 doses pravastatin vs placebo
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation, but no further details
Allocation concealment (selection bias)	Low risk	Sealed envelope used by staff pharmacologist and no further access to patient management of data
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All investigators and participants were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details given, but likely blinding was maintained
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Cavalca 2006.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; University of Milan, Milan, Italy
Participants	Fifty participants Mean age: 63.4 years vs 68.4 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: first‐time isolated low‐risk (EuroSCORE 6) coronary bypass surgery patients for whom an on‐pump and an off‐pump procedure were deemed technically feasible and not contraindicated (i.e. porcelain aorta for CABG), both sexes Exclusion criteria: Q‐wave MI in the past 6 weeks, unstable angina, or poor LV function (EF < 30%), with no age restrictions
Interventions	Group A: intraoperative on‐pump CABG Group B: intraoperative OPCABG
Outcomes	Outcomes: urinary isoprostane iPF2‐III levels, plasma levels of free MDA, and total antioxidant status were measured before, during, and up to 24 hours after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but the process was not described.
Allocation concealment (selection bias)	High risk	No details given around concealment or blinding of allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding described; possible for participants but unlikely to have personnel blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear who was blinded in this study
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Data were missing or not included in the analysis for three participants.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Cavalca 2008.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; University of Milan, Milan, Italy
Participants	Forty‐three participants Mean age: 67.7 years vs 65.2 years Sex (female/male ratio): not specified Low‐risk patients (elective cardiac surgery) Inclusion criteria: patients scheduled to undergo elective cardiac surgery with CPB, stable angina, LVEF > 40%, and aged 60 years to 80 years, both sexes Exclusion criteria: aortic valve stenosis, angina on arrival in the operating room, and AMI during the past 7 days
Interventions	Intervention: propofol 2.5 g/mL to 4 g/mL (infusion protocol: 2.5–4 g/mL propofol, 10–12 ng/mL remifentanil). Anaesthesia was maintained with propofol (1.5 g/mL) and remifentanil (8–12 ng/mL), decreasing to 1.2 and 4 ng/mL, respectively, during CPB). Control: inhalation of sevoflurane (1% to 2%) (in an oxygen–air mixture and remifentanil (8–12 ng/ml, target‐controlled infusion) for the entire procedure, except for CPB, during which midazolam (50–100 g/kg/h) was administered intravenously and remifentanil was reduced (4 ng/mL))
Outcomes	Outcomes: effect of propofol on gamma‐T levels Secondary outcomes: plasma levels of alpha‐T, individual antioxidant capacity, MDA, and IL‐10 were measured before, during, and after anaesthesia. In addition, levels of the proinflammatory prostaglandin E2 as a marker of cyclooxygenase‐2 activity and those of IL‐10 were measured in whole blood cultured with bacterial lipopolysaccharide. Blood was collected before induction of anaesthesia (T0, baseline), 30 minutes after the beginning of CPB (T1), after protamine administration and on‐pump weaning (T2), at arrival in the intensive care unit (T3), and 24 hours after anaesthesia induction (T4).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	High risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes; one withdrawal after randomisation; no losses to follow‐up reported in the included participants, but no ITT analysis
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Cavarocchi 1986.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Mayo Clinic and Foundation, Rochester, MN, United States of America
Participants	Ninety‐one participants Mean age: not specified Sex (female/male ratio): not specified Low‐risk patients Inclusion criteria: patients undergoing CPB, both sexes Exclusion criteria: not specified
Interventions	MP‐coated bubble oxygenator vs bubble oxygenator Group I: bubble oxygenator (30 patients) Group II: bubble oxygenator and MPSS (30 mg/kg) (31 patients) Group III: membrane oxygenator (30 patients)
Outcomes	Outcomes: Complement 3a, Complement 5a Blood samples were taken during CPB and 25 minutes after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Celik 2004.

Study characteristics
Methods	Single‐centre, prospective, double‐blinded, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Meram Medical School of Selcuk University, Konya, Turkey
Participants	Sixty participants Mean age: 60 years vs 62 years Sex (female/male ratio): not specified Low‐risk patients (CABG) Inclusion criteria: patients with CAD undergoing CABG Exclusion criteria: patients with severely impaired LV function (EF < 40%), pulmonary disease (e.g. previous lung surgery, asthma, and chronic obstructive lung disease on therapy and supplemental oxygen or mechanical ventilatory requirement), severe systemic noncardiac disease (renal or liver impairment and cerebral stroke), insulin‐dependent diabetes, recent MI (6 weeks), infectious disease just before operation, and preoperative use of steroids and contraindications to steroid administration
Interventions	Intervention: MP IV 30 mg/kg, 6 times perioperatively (10 minutes before CPB, immediately after CBP, and then every 6 hours for the next 24 hours in an ICU) Control: placebo IV normal saline
Outcomes	Outcomes: TNF‐α, IL‐6, IL‐8, and IL‐10, CK, and CK‐MB Blood samples were taken for TNF‐α, IL‐6, IL‐8, and IL‐10 before induction of anaesthesia (T0, control value), after induction (T1), before starting CPB (T2), after aortic declamping (T3), at the end of CPB (T4), and 6 h (T5), 12 h (T6), and 24 h (T7) after skin closure. CK and CK‐MB were evaluated at the following intervals: T0, T5, T6, and T7.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	High risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All investigators blinded: "An anaesthesia research nurse performed the randomization and prepared the two syringes of solution that were administered by the anaesthesiologist managing the case"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators blinded to outcome data based on description
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chee 2017.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: March 2013 to May 2014 Registration date: not specified Number of study centres and locations: single centre; Mater Misericordiae University Hospital, Dublin, Ireland
Participants	Thirty participants Mean age: 67 years vs 64 years Sex (female/male ratio): 5/25 Low‐risk patients (underwent CPB – majority CABG – 21/30) Inclusion criteria: patients who underwent CPB‐related surgery, both sexes Exclusion criteria: patients on rosuvastatin because of the differences in its potency compared to that of atorvastatin and simvastatin; patients who had undergone emergency surgery, revision surgery, or use of an IABP preoperatively; patients with underlying infection disease or other significant disease such as chronic inflammatory disease, chronic liver disease, chronic kidney disease, and malignancy; patients who were being treated with corticosteroids or anti‐inflammatory or immunosuppressants
Interventions	Intervention: 80 mg atorvastatin at least 2 weeks prior to surgery Control: patients who continued initial statin therapy of ≤ 40 mg
Outcomes	Outcomes: white cell count, neutrophils and serum Cr, serum IL‐8 and MMP‐9 levels Time points: white cell count and neutrophils at baseline and after surgery; IL‐8 and MMP Blood samples were taken at patient baseline, 5 minutes after cross‐clamp removal (reflecting CPB activation and reperfusion injury), and 4 h after cross‐clamp removal (reflecting CPB activation and ischaemic‐reperfusion injury); serum high‐sensitivity cardiac troponin (hs‐cTnI) was determined in arterial blood samples at anaesthesia induction and 4 hours after cross‐clamp removal. NGAL levels were determined in urine samples 4 hours after cross‐clamp removal.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random allocation cards and sealed envelopes were used.
Allocation concealment (selection bias)	Low risk	Independent person responsible for randomisation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All investigators were blinded; unlikely breach of blinding to participants
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators were likely blinded to outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	Funding from pharmaceutical company but disclosure of no involvement of funder

Chello 2006.

Study characteristics
Methods	Single‐centre, double‐blinded, placebo‐controlled, randomised study Run‐in period: May 2003 to September 2004 Registration date: not specified Number of study centres and locations: single centre; University Campus Bio‐Medico, Rome, Italy
Participants	Forty participants Mean age: 65.7 years vs 63.7 years Sex (female/male ratio): 9/31 Low‐risk patients (elective CABG) Inclusion criteria: patients scheduled to undergo elective CABG surgery, both sexes Exclusion criteria: patients with diabetes, renal or hepatic impairment, congestive heart failure, active inflammatory or immunomodulatory diseases, or a history of MI 6 months previously and pregnant women
Interventions	Intervention: atorvastatin 20 mg/day for 3 weeks Control: placebo for 3 weeks; no further details
Outcomes	Outcomes: serum cytokine levels, apoptosis, and indices of activation Time points: before anaesthesia; at the end of surgery; and at 4 hours, 24 hours, 48 hours, and 72 hours postoperatively; neutrophil apoptosis and indices of activation at 12 hours and 24 hours
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process is not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and physicians were both blinded to the drug assignment group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details were given, but all the investigators were likely blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chello 2007.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: January 2005 to February 2006 Registration date: not specified Number of study centres and locations: single centre; University Campus Bio‐Medico, Rome, Italy
Participants	Thirty participants Mean age: 67.7 years vs 66.3 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients scheduled to undergo elective CABG surgery, both sexes Exclusion criteria: patients with diabetes, renal or hepatic impairment, congestive heart failure, active inflammatory or immunomodulatory diseases, or a history of MI within 6 months
Interventions	Intervention: simvastatin 40 mg/day for 3 weeks before surgery Control: placebo; no further details provided A third group of 15 patients undergoing OPCABG served as the control group.
Outcomes	Outcomes: serum cytokine levels, apoptosis, and indices of activation Blood samples were taken before anaesthesia; at the end of surgery; and at 4 hours, 24 hours, 48 hours, and 72 hours postoperatively. Neutrophil apoptosis and indices of activation at 12 hours and 24 hours
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and physicians were both blinded to the drug assignment group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details were given, but all investigators were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chen 2004.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Kaohsiung Medical University Hospital, Kaohsiung, Taiwan
Participants	Thirty‐two participants Mean age: 60.3 years vs 61.2 years Sex (female/male ratio): 6/26 Low‐risk patients (elective CABG or heart valve surgery) Inclusion criteria: adult patients undergoing CPB for CABG or heart valve operation Exclusion criteria: prior cardiac operation, infection, emergency operation, congestive heart failure, AMI within the previous 1 month, corticosteroid therapy, and severe asthma or chronic obstructive lung disease
Interventions	Intervention: intraoperative leucocyte depletion Control: standard arterial filter
Outcomes	Outcomes: P‐selectin, ICAM‐1, IL‐8, and PECAM‐1; plasma MDA concentration Blood samples were taken before sternotomy, at 30 minutes and at 60 minutes of CPB, at 5 minutes after coronary reperfusion, at the end of CPB, and at 2 hours and 24 hours after the cessation of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	"This study was supported by grants from the National Science Council of the Republic of China (NSC 89–2314‐B037–095)."

Chen 2005 (A).

Study characteristics
Methods	Multi‐centre, randomised, double‐blinded, placebo‐controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: 65; United States of America
Participants	Nine hundred fourteen participants allocated to three groups Mean age: 67 years vs 68 years Sex (female/male ratio): 162/452 Low‐risk patients (elective CABG or valve surgery) Inclusion criteria: patients undergoing non‐emergent CABG surgery ± valve surgery, both sexes Exclusion criteria: not specified
Interventions	Intervention: monoclonal antibody (PEX Bolus (2.0 mg/kg) + placebo infusion) at time of surgery and 24 hours after Control: placebo bolus + placebo infusion at time of surgery and 24 h after
Outcomes	Outcomes: laboratory testing including haematology, coagulation, chemistry, and urinalysis was performed preoperatively (screening); on PODs 1, 3, and 7 (or on the day of hospital discharge); and at the termination interview on POD 30. ECGs were taken preoperatively and on PODs 1, 4, 7 (or discharge, if < 7 days), and 30 and were screened for the presence of unreported MI by a blinded core ECG laboratory. Perioperative haemodynamics, concomitant medications, adverse events, chest tube drainage, and transfusion requirements were recorded. Time points: 0 hours, 1 hour, 4 hours, 8 hours, 12 hours, 24 hours, 48 hours, 168 hours
Notes	Chen 2005 divided into two studies because there were two intervention groups and one control
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible but unlikely for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	"Supported in part by grants from Alexion Pharmaceuticals, Cheshire, CT and Procter and Gamble Pharmaceuticals, Mason, OH"

Chen 2005 (B).

Study characteristics
Methods	Multi‐centre, randomised, double‐blinded, placebo‐controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: 65; United States of America
Participants	Nine hundred fourteen participants were allocated to three groups. Mean age: 68 years vs 68 years Sex (female/male ratio): 142/464 Low‐risk patients (elective CABG or valve surgery) Inclusion criteria: patients undergoing non‐emergent CABG surgery ± valve surgery, both sexes Exclusion criteria: not specified
Interventions	Intervention: monoclonal antibody (PEX Bolus (2.0 mg/kg) + PEX infusion (0.05 mg/kg/hour for 24 h) Control: placebo bolus + placebo infusion at time of surgery and 24 hours after
Outcomes	Outcomes: laboratory testing including haematology, coagulation, chemistry, and urinalysis was performed preoperatively (screening); on POD 1, 3, and 7 (or on the day of hospital discharge); and at the termination interview on POD 30. ECGs were taken preoperatively and on PODs 1, 4, 7 (or discharge, if < 7 days), and 30 and were screened for the presence of unreported MI by a blinded core ECG laboratory. Perioperative haemodynamics, concomitant medications, adverse events, chest tube drainage, and transfusion requirements were recorded. Time points: 0 hours, 1 hour, 4 hours, 8 hours, 12 hours, 24 hours, 48 hours, 168 hours
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible but unlikely for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	"Supported in part by grants from Alexion Pharmaceuticals, Cheshire, CT and Procter and Gamble Pharmaceuticals, Mason, OH"

Chen 2013 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Chinese PLA General Hospital, Beijing, China
Participants	One hundred twenty participants, four groups: the blank control (Group C), TXA group (Group T), UST group (Group U), and TXA–UST combination group (Group D) Mean age: 50 years vs 50 years Sex (female/male ratio): 36/24 Low‐risk patients (open‐heart valve replacement surgery) Inclusion criteria: patients with rheumatic heart disease and New York Heart Association (NYHA) cardiac function of Grade II and III undergoing open‐heart valve replacement surgery, both sexes Exclusion criteria: age > 80 years; reoperation; use of hormone and antibiotics 1 week prior to the surgery; preoperative examinations that revealed severe coagulation abnormalities such as significant prolongation of PT and significant reduction in thrombocytes; severe liver and renal failure; detection of pericardial adhesions during surgery; and receipt of treatment with recombinant human coagulation factor VII during and after surgery
Interventions	Intervention: TXA 15 mg/kg body weight prior to the surgery and then maintained by IV infusion of TXA at a dose of 15 mg/kg body weight per hour until the termination of the surgery Control: placebo (saline)
Outcomes	Outcomes: Hct, D‐dimer, TNF‐α, IL‐6, neutrophil elastase, tissue plasminogen activator, α2‐antiplasmin, postsurgical pericardial mediastinal drainage Blood samples were taken after the induction of anaesthesia (T1), at the time of unclamping the ascending aorta (T2), 1 hour after termination of CPB (T3), and 24 hours after termination of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number method (computer randomisation) was used.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were likely included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chen 2013 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Chinese PLA General Hospital, Beijing, China
Participants	One hundred twenty participants, four groups: the blank control (Group C), TXA group (Group T), UST group (Group U), and TXA–UST combination group (Group D) Mean age: 49 years vs 50 years Sex (female/male ratio): 34/26 Low‐risk patients (open‐heart valve replacement surgery) Inclusion criteria: patients with rheumatic heart disease and NYHA cardiac function of Grade II and III undergoing open‐heart valve replacement surgery, both sexes Exclusion criteria: age > 80 years; reoperation; use of hormone and antibiotics 1 week prior to the surgery; preoperative examinations that revealed severe coagulation abnormalities such as significant prolongation of PT and significant reduction in thrombocytes; severe liver and renal failure; detection of pericardial adhesions during surgery; and receipt of treatment with recombinant human coagulation factor VII during and after surgery
Interventions	Timing: intraoperative Type: UST Combination: N/A Comparison: placebo (saline) Intervention: UST 1.2 × 104 IU/kg body weight via the internal jugular vein prior to the surgery; another half was added into the CPB priming solution Control: placebo (saline)
Outcomes	Outcomes: Hct, D‐dimer, TNF‐α, IL‐6, neutrophil elastase, tissue plasminogen activator, α2‐antiplasmin, postsurgical pericardial mediastinal drainage Blood samples were taken after the induction of anaesthesia (T1), at the time of unclamping the ascending aorta (T2), 1 hour after termination of CPB (T3), and 24 hours after termination of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number method (computer randomisation) was used.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were likely included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chen 2013 (C).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Chinese PLA General Hospital, Beijing, China
Participants	One hundred twenty participants, four groups: the blank control (Group C), TXA group (Group T), UST group (Group U), and TXA–UST combination group (Group D) Mean age: 49 years vs 50 years Sex (female/male ratio): 32/28 Low‐risk patients (open‐heart valve replacement surgery) Inclusion criteria: patients with rheumatic heart disease and NYHA cardiac function of Grade II and III undergoing open‐heart valve replacement surgery, both sexes Exclusion criteria: age > 80 years; reoperation; use of hormone and antibiotics 1 week prior to the surgery; preoperative examinations that revealed severe coagulation abnormalities such as significant prolongation of PT and significant reduction in thrombocytes; severe liver and renal failure; detection of pericardial adhesions during surgery; and receipt of treatment with recombinant human coagulation factor VII during and after surgery
Interventions	Timing: intraoperative Comparison: placebo (saline) Intervention: TXA 15 mg/kg body weight prior to the surgery and then maintained by IV infusion of TXA at a dose of 15 mg/kg body weight per hour until the termination of the surgery + UST 1.2 × 104 IU/kg body weight via the internal jugular vein prior to the surgery; another half was added into the CPB priming solution Control: placebo (saline)
Outcomes	Outcomes: Hct, D‐dimer, TNF‐α, IL‐6, neutrophil elastase, tissue plasminogen activator, α2‐antiplasmin, postsurgical pericardial mediastinal drainage Blood samples were taken after the induction of anaesthesia (T1), at the time of unclamping the ascending aorta (T2), 1 hour after termination of CPB (T3), and 24 hours after termination of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number method (computer randomisation) was used.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were likely included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Chen 2015.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: March 2013 to March 2014 Registration date: not specified Number of study centres and locations: single centre; The Second Affiliated Hospital, Nanchang University, Jiangxi, China
Participants	Sixty‐five participants Mean age: 62 years vs 62 years Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: patients who had elective on pump CABG surgery, both sexes Exclusion criteria: patients who had systemic inflammatory diseases, received immunosuppressive or anti‐inflammatory medication within a month before administration, had previous cardiac surgery, had MI within 6 weeks before administration, and had severely impaired LV function (EF < 40%). Patients were also excluded if an emergency operation for unstable angina was performed.
Interventions	Intervention: dexmedetomidine (Dex pump infusion at a loading dose of 0.5 µg/kg for 10 minutes, followed by a continuous pump infusion 0.5 µg/kg per hour until the completion of the surgery) Control: normal saline for the same period
Outcomes	Outcomes: cTnI and CK‐MB were chosen to evaluate myocardial protection, and pro‐inflammatory cytokines TNF‐α, IL‐6, IL‐8, and anti‐inflammatory cytokine IL‐10. Central venous blood samples were taken at T1 (baseline), T2 (15 minutes after declamping the aorta), T3 (2 hours after declamping the aorta), and T4 (24 hours).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No funding was disclosed or reported.

Chen 2016.

Study characteristics
Methods	Single‐centre, prospective, randomised, controlled study with a 1:1 allocation ratio Run‐in period: July 2014 to March 2015 Registration date: not specified Number of study centres and locations: single centre; State Key Laboratory of Cardiovascular Disease and the Chinese Academy of Medical Science & Peking Union Medical College, Beijing, China
Participants	Sixty participants Mean age: 47 years vs 47 years Sex (female/male ratio): 18/42 High‐risk patients (total arch replacement) Inclusion criteria: patients receiving aortic arch replacement with DHCA, both sexes Exclusion criteria: reoperation, anti‐Plt drug treatment within 7 days before the operation, infectious diseases, and chronic renal or hepatic dysfunction
Interventions	Intervention: circuit coating (coated with human albumin by circulating the prime for 5 minutes at 4 L/min before CPB); study group (circulating the prime with 40 g albumin for 5 minutes before CPB) Control: noncoating circuit; control group (receiving 40 g human albumin 5 minutes after the initiation of CPB)
Outcomes	Time points: baseline, after anaesthesia induction and before CPB (T1), and 10 min after heparin reversal before any blood product transfusion (T2) Outcomes: MA at 10 min after heparin reversal before any blood product transfusion; other r‐TEG parameters; complete blood count containing HB, Plt, and WBC; coagulation test involving PT, APTT, and fibrinogen; inflammatory cytokines, comprising IL‐1, IL‐6, IL‐10, TNF‐α, and PAF; the amount of bleeding and transfusion data Blood samples were taken from the arterial line at the following time points: after anaesthesia induction; before aortic cannulation; 10 minutes after aortic cross‐clamp release; 5 minutes after protamine sulphate administration; 4 hours after protamine sulphate administration; and 24 hours after protamine sulphate administration.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random number list
Allocation concealment (selection bias)	High risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The attending medical team was blinded to the randomisation. Perfusionist was not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No funding was disclosed or reported.

Chew 2016.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: 2009 to 2012 Registration date: not specified Number of study centres and locations: single centre; Singapore General Hospital, Singapore
Participants	Sixty‐eight participants Mean age: 58 years vs 60 years Sex (female/male ratio): 18/49 Low‐risk patients (elective CABG) Inclusion criteria: patients presenting for elective CABG; Asian patients aged between 21 years and 85 years with no previous cardiac surgery done; both sexes Exclusion criteria: patients with poor LVEF (< 30%), immunologic disease or malignancies, acute inflammatory disease, coagulopathy, steroid treatment, significant carotid disease, preoperative renal impairment, or dialysis
Interventions	Intervention: minimised CPB circuit Control: CCPB circuit
Outcomes	Outcomes: AKI, Cr, cystatin‐C, any outcomes Time points: preoperative, intraoperative, during CPB, peak within 48 hours in ICU
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Low risk	Sealed envelope was used.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Single‐blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Single‐blinded, with plausible detection bias for research personnel
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes; one withdrawal, no losses to follow‐up were reported, but no ITT analysis. Data were missing on one participant.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	Noncommercial funding (NMRC IRG07nov119)

Cho 2009.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Catholic University of Korea, Medical College, Incheon, Korea
Participants	Thirty participants Mean age: 57 years vs 48 years Sex (female/male ratio): 24/6 Low‐risk patients (valve replacement surgery) Inclusion criteria: patients undergoing CPB, both sexes Exclusion criteria: patients with serious systemic inflammation or low cardiac output and patients using vasoactive agents were excluded from the study.
Interventions	Intervention: sevoflurane (2% sevoflurane was delivered through the oxygenator) Control: sufentanil (20 g of sufentanil was added to the priming solution)
Outcomes	Outcomes: postoperative data included acute renal failure (Cr 2.0 mg/dL), postoperative arrhythmias, pneumonia or pulmonary oedema, and postoperative inotropes, as well as the duration of mechanical ventilation, the length of stay in the ICU and the hospital, IL‐6, IL‐8, IL‐10, and TNF‐α. Time points: A, baseline; B, before CPB; C, after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel, but unlikely for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Choi 2011.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: October 2008 to November 2009 Registration date: not specified Number of study centres and locations: single centre; Yonsei University College of Medicine, Seoul, Republic of Korea
Participants	Seventy‐six participants Mean age: 57 years vs 60 years Sex (female/male ratio): 46/30 High‐risk patients (includes complex cardiac surgery procedures) Inclusion criteria: adult patients scheduled for elective, complex valvular heart surgery, which was defined as double‐valve surgery; combined valve and CABG procedures; Bentall operation; combined mitral valve surgery and tricuspid annuloplasty or reoperation; both sexes Exclusion criteria: patients older than 80 years and those with left main CAD > 50%, hepatic or pulmonary disease, active infective endocarditis, LVEF < 30%, MI within 3 weeks, pre‐existing renal dysfunction (serum Cr level > 1.6 mg/dL for men and > 1.4 mg/dL for women), or peripheral vascular disease affecting the lower limbs. In addition, patients taking the antidiabetic sulphonylurea glyburide (INN glibenclamide) or receiving nicorandil drug therapy were excluded because these agents have been shown to abolish preconditioning. Patients who required surgery for tricuspid valve replacement or hypothermic circulatory arrest were excluded.
Interventions	Intervention: RIPC (10 minutes before CPB) Control: deflated cuff
Outcomes	Outcomes: serum biomarkers of renal injury and the incidence of AKI, comparison of serum CK‐MB levels at 12 hours and 24 hours after surgery Time points: for primary outcome, baseline, ICU arrival, day 1, day 2, and day 3
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated table of randomisation
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Single‐blinding of personnel, but no details
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Single‐blinding, but no further details
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No funding was disclosed or reported.

Collart 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; University of Marseille, Marseille, France
Participants	Eighty participants Mean age: 65 years vs 66 years Sex (female/male ratio): 14/66 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, aged 18 years to 80 years, and signed consent to participate; both sexes Exclusion criteria: emergency cardiac operation, previous cardiac operation, nonisolated CABG, LVEF < 40%, preoperative liver disease or coagulopathy, renal disease requiring dialysis, severe pulmonary disorders, neurologic disorders, diabetes, and inability to obtain signed consent
Interventions	Intervention: heparin‐coated circuit group Control: control group (noncoated)
Outcomes	Outcomes: mortality rate, intubation time, ICU stay, blood loss, transfusion, inotropic support, myocardial ischaemia, arrhythmias, respiratory function, neurologic events, healing process, pericardial effusion, extensive laboratory tests Time points: days 0, 1, 2, and 3 and at discharge
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Corbeau 1995 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Angers, France
Participants	Sixty‐three participants Mean age: 66 years vs 63 years Sex (female/male ratio): 14/49 Low‐risk patients (elective CABG or AVR) Inclusion criteria: adults undergoing either CABG or AVR, both sexes Exclusion criteria: minors, previous cardiac surgery, anti‐Plt agent treatment within 10 days prior to operation, coagulopathy
Interventions	Intervention: aprotinin (2 × 106 KIU (280 mg) after induction, followed by an infusion of 0.5 × 106 KIU/h (70 mg/h) until chest closure, with a supplement to the oxygenator prime of 2 × 106 KIU) Control: routine care
Outcomes	Outcomes: amount of blood collected from the chest tube drainage, Plt counts, Hct, coagulation times, ACT, fibrinogen concentration, Cr, creatine phosphokinase, ECG criteria for MI Time points: admission to ICU as well as 4 hours, 8 hours, and 18 hours for chest tube drainage; postoperative for other outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Corbeau 1995 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Angers, France
Participants	Sixty‐one participants Mean age: 62 years vs 63 years Sex (female/male ratio): 19/42 Low‐risk patients (elective CABG or AVR) Inclusion criteria: adults undergoing either CABG or AVR, both sexes Exclusion criteria: minors, previous cardiac surgery, anti‐Plt agent treatment within 10 days prior to operation, coagulopathy
Interventions	Intervention: TXA, 15 mg/kg between the injection of heparin (400 IU/kg) and the beginning of ECC, 15 mg/kg after protamine injection (1.3 mg/100 lU of heparin) Control: routine care
Outcomes	Outcomes: amount of blood collected from the chest tube drainage, Plt counts, Hct, coagulation times, ACT, fibrinogen concentration, Cr, creatine phosphokinase, ECG criteria for MI Time points: admission to ICU as well as 4 hours, 8 hours, and 18 hours for chest tube drainage; postoperative for other outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Corcoran 2004.

Study characteristics
Methods	Single‐centre, prospective, double‐blinded, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Cork University/Mercy Hospitals and University College Cork, Cork City, Ireland
Participants	Twenty‐one participants Mean age: 58 years vs 62.5 years Sex (female/male ratio): 2/19 Low‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective CABG Exclusion criteria: impaired LV function (determined by ventriculography, EF < 50%), clinically significant valvular dysfunction, MI within the previous 6 weeks, diabetes mellitus, renal dysfunction, autoimmune disease, concurrent medication with anti‐inflammatory or immunosuppressant agents, or a history of allergy to propofol
Interventions	Intervention: propofol infusion (target‐controlled infusion of propofol (Diprifusor) was commenced to achieve a steady‐state target site concentration of 6 mg/mL to 8 mg/mL at the time of cross‐clamp release) Control: placebo (saline) (15 minutes before release of cross‐clamp infusion of normal saline 0.9%)
Outcomes	Outcomes: IL‐4, IL‐6, IL‐8, IL‐10, leucocyte 2 integrin expression, TnI Time points: IL‐6/IL‐8 T2 (1 minute to 20 minutes), Trc (30 minutes), T4 (4 hours), and T5 (12 hours); IL‐4/IL‐10 T2 (1 minute to 20 minutes), T3 (1 hour), T5 (12 hours), and T7 (36 hours)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described; blinding of participants likely
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators were likely blinded to outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Corcoran 2006.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; Cork University/Mercy Hospitals and University College Cork, Cork City, Ireland
Participants	Twenty‐seven participants Mean age: not specified Sex (female/male ratio): 7/20 Low‐risk patients (elective CABG) Inclusion criteria: patients with an LVEF < 40% undergoing elective CABG, both sexes Exclusion criteria: clinically significant valvular dysfunction, MI within the previous 6 weeks, diabetes mellitus, renal dysfunction, autoimmune disease, concurrent medication with anti‐inflammatory or immunosuppressant agents, or a history of allergy to propofol
Interventions	Intervention: propofol infusion (target‐controlled infusion of propofol (Diprifusor) was commenced to achieve a steady‐state target site concentration of 6 mg/mL to 8 mg/mL at the time of cross‐clamp release) Control: placebo (saline) (15 minutes before release of cross‐clamp infusion of normal saline 0.9%)
Outcomes	Outcomes: IL‐6, IL‐8, IL‐10, leucocyte respiratory burst or phagocytic activity, cTnI Time points: IL‐6/IL‐8 during CPB, 20 minutes after cross‐clamp removal, and at 1 hour, 4 hours, and 12 hours; IL‐10 during CPB, at 4 hours, and at 24 hours; cTnI during pre‐induction, 24 hours after release of cross‐clamp
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described; blinding of participants likely
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators were likely blinded to outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Cui 2020.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: July 2019 and September 2019 Registration date: 6 March 2021 Number of study centres and locations: single centre; Guangdong Provincial People’s Hospital, Guangzhou, China
Participants	Fifty‐seven participants Mean age: 48 years vs 49 years Sex (female/male ratio): 23/34 Low‐risk patients (thoracoscopic mitral valve surgery) Inclusion criteria: patients aged between 18 years and 65 years; ASA score II or III and NYHA class II or III; an estimated ECC ≤ 180 min; no active infective endocarditis; and a RhF level < 20 IU/mL. Exclusion criteria: elderly patients ≥ 70 years old; patients without sick sinus syndrome and atrioventricular block (bradycardia, heart rate < 60 beats per minute, atrioventricular block above the first degree, or indoor conduction block); patients with coronary heart disease or cerebrocardiac syndrome; patients with Gold grade < 2; patients with severe pulmonary hypertension, with a pulmonary artery systolic pressure > 60 mmHg; patients with immune system diseases; patients with abnormal renal function with a blood Cr level > 110 μmol/L; and patients in whom CPB or the prevention of all‐cause mortality from reoperation was difficult.
Interventions	Intervention: dexmedetomidine (infusion of 0.5 μg/kg DEX within 10 minutes using a microinfusion pump. DEX was then continuously infused at 0.5 μg/kg/h until the operation was completed) Control: placebo (normal saline instead of DEX)
Outcomes	Outcomes: the oxygenation (PaO2/FiO2) and respiratory indexes (P(A‐a)O/PaO2); the serum cytokines IL‐6, TNF‐α, and ICAM‐1; procalcitonin; the postoperative extubation time; length of ICU stay; and pulmonary infection rate Data were calculated 5 minutes after tracheal intubation (T1), 2 hours after operation (T2), 6 hours after operation (T3), and 24 hours after operation (T4).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly divided using the random number table method.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	High risk	In total, 76 patients were enroled, of whom 16 dropped out of the study after randomisation (shorter CPB than planned in six patients, surgical complications of thoracotomy and secondary chest closure in five, irregular randomisation in one, heavily calcified large blood vessels that made cannulation impossible for CPB in one, and loss of follow‐up in three).
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	The authors report no conflict of interest.

Czerny 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified Registration date: not specified Number of study centres and locations: single centre; University of Vienna Medical School, Vienna, Austria
Participants	Thirty participants Mean age: 63 years vs 65 years Sex (female/male ratio): 7/23 Low‐risk patients (CABG) Inclusion criteria: patients undergoing CABG, both sexes Exclusion criteria: not described
Interventions	Intervention: off‐pump coronary revascularisation Control: on‐pump coronary revascularisation
Outcomes	Outcomes: parameters of the inflammatory response (IL‐6, IL‐10, ICAM‐1, P‐selectin) and of myocardial injury (myoglobin, CK‐MB, TnI) (intraoperatively, 4 hours, 8 hours, 16 hours, 24 hours, and 48 hours after surgery), clinical outcomes Time points: preoperative, postoperative, 4 hours, and 6 hours for IL‐6, IL‐10, P‐selectin, and ICAM‐1. Myoglobin area – 8 hours, CK‐MB mass area – 24 hours, and cTnI area – 48 hours. Secondary outcomes – post‐surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that patients were randomised, but process was not described.
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described for participants or personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Damgaard 2010.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: January 2003 to February 2004 Registration date: 12 September 2005 Number of study centres and locations: single centre; Copenhagen University Hospital, Copenhagen, Denmark
Participants	Twenty‐nine participants Mean age: 66 years vs 68 years Sex (female/male ratio): 6/23 Low‐risk patients (elective CABG) Inclusion criteria: age older than 18 years, need for CABG, and informed consent, both sexes Exclusion criteria: off‐pump, redo or valve operations, current infection or antibiotic treatment, serum Cr concentration > 200 μmol/L, liver disease, immune disease, and anti‐inflammatory or immune‐modulating treatment, except for nonsteroidal anti‐inflammatory drugs and aspirin
Interventions	Intervention: pericardial blood washing and autotransfusion Control: direct retransfusion of the suction blood and the CPB circuit blood
Outcomes	Outcomes: patient plasma concentrations of IL‐6; plasma concentrations of the inflammatory markers IL‐1, IL‐8, IL‐10, IL‐12, TNF‐α, soluble tumour necrosis factor receptors I and II, and procalcitonin; bleeding; allogenic transfusions; cell‐saver effectiveness regarding inflammatory marker reduction; and complications Time points: 6 hours, 24 hours, and 72 hours postoperatively, except for bleeding, allogenic transfusions, cell‐saver effectiveness regarding inflammatory marker reduction, and complications measured at a single time point
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described clearly.
Allocation concealment (selection bias)	Low risk	Sealed, opaque, and numbered envelopes were used.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; plausible for personnel and participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, one withdrawal, and no losses to follow‐up were reported, but no ITT analysis. Data from all the included participants (29) were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Unclear risk	No funding was disclosed or reported.

Danielson 2018.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: January 2013 to January 2017 Registration date: 24 December 2012 Number of study centres and locations: single centre; University of Gothenburg, Gothenburg, Sweden
Participants	Thirty participants Mean age: 70.7 years vs 68.9 years Sex (female/male ratio): 7/23 Low‐risk patients (elective CABG, SAVR) Inclusion criteria: elective open ARVR surgery (SAVR) with a biological valve prosthesis due to aortic stenosis with or without concurrent CABG; normal preoperative coagulation tests (i.e. partial thromboplastin time < 45 seconds and PT [INR] < 1.5 and a Plt count > 80,000); absence of recent (< 1 week) treatment with thrombolytic or potent anti‐Plt drugs; and preoperative LVEF ≥ 50% Exclusion criteria: stroke with sequelae and abnormal coagulation tests (see above) or thromboelastograms in the morning on the day after surgery
Interventions	Timing: after induction of anaesthesia Intervention: MP (15 mg/kg) administered after induction of anaesthesia Control: placebo (saline) (0.9% sodium chloride) administered after induction of anaesthesia
Outcomes	Outcomes: assessment of systemic and brain inflammation (IL‐6, IL‐8, TNF‐α), axonal injury (total‐tau, neurofilament light chain protein), neuronal injury (NSE), astroglial injury (S‐100B, glial fibrillary acidic protein), and the BBB integrity (CSF/serum albumin ratio) Time points: the day before and 24 hours after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state that patients were randomised, but process was not described.
Allocation concealment (selection bias)	Low risk	A 1:1 ratio, using concealed, sequentially numbered opaque envelopes; randomiser not involved in study or care
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described adequately
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	"The study was supported by Swedish State Support for Clinical Research (ALFGBG‐721141, ALFGBG‐144341, and ALFGBG‐139671) and grants from the Gothenburg University, Sahlgrenska Academy, the Gothenburg Medical Society, the Swedish Research Council (K2010‐61X‐14002, K2010‐63P‐21562‐01‐4, and K2011‐61X‐20401‐05‐6), Frimurarestiftelsen and Hjärnfonden"

De Amorim 2014.

Study characteristics
Methods	Single centre, prospective randomised study Run‐in period: not specified. Study registered 2007. Study completed 2010. Published 2014 Number of study centres and location: single centre. Instituto do Coração, Faculdade de Medicina da Universidade de São Paulo, São Paulo, SP, Brazil Setting: secondary care Date of study: 2013
Participants	N randomised: 24 N lost: 4 N analysed: 20 CABG patients undergoing CPB leucocyte filtration (n = 9) or standard CPB (n = 11) Inclusion: individuals undergoing CABG who had their physical state classified as PII or PIII, according to the ASA. Exclusion: individuals older than 70 years or those presenting with BMI > 35, with CHF class > III (NYHA) or with LVEF < 40%, who had recently undergone other surgery, had Cr levels ≥ 1.4 mg/dL, or were using oral anticoagulants
Interventions	Timing: intraoperative Type: CPB leucocyte filtration Comparison: standard CPB
Outcomes	Primary outcomes: N/A Secondary outcomes: N/A Any outcomes: haemodynamic data and the PaO2/FiO2 index value were obtained preoperatively; after induction of anaesthesia (after induction); 5 minutes before CPB (beginning of CPB); 5 minutes after end of CPB (end of CPB); during skin suturing (end of surgery); and at 6 hours, 12 hours, and 24 hours after surgery. Intrapulmonary shunt fraction (Qs/Qt), TNF‐α, IL‐1β, IL‐6, IL‐8, IL‐10, IL‐1rA, elastase, and MPO were also evaluated at the following times: after induction; at beginning of CPB; at end of CPB; at end of surgery; and at 6 hours, 12 hours, and 24 hours postoperatively. Cell counts were performed at the following times: after induction, after 5 minutes of CPB, after 25 minutes of CPB, after 50 minutes of CPB, during suturing, and at 12 hours and 24 hours postoperatively. The presence and type of any infection were recorded postoperatively. Time points: as above
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process: "individuals were allocated randomly into the control or filtered group according to simple randomisation occurring immediately before surgery"
Allocation concealment (selection bias)	Unclear risk	No description of allocation: "individuals were allocated randomly into the control or filtered group. … In the filtered group an additional leucocyte filter was placed alongside the standard filter"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding described: "individuals were allocated randomly into the control or filtered group according to simple randomisation occurring immediately before surgery. In the filtered group an additional leucocyte filter was placed alongside the standard filter"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No blinding described: "individuals were allocated randomly into the control or filtered group according to simple randomisation occurring immediately before surgery. In the filtered group an additional leucocyte filter was placed alongside the standard filter"
Incomplete outcome data (attrition bias) All outcomes	Low risk	"A total of 26 individuals were initially assessed for eligibity [sic] of participation, but only 20 completed the study: 11 in the control group and 9 in the filtered group. One did not meet the inclusion criteria during the interview. Another refused to participate after the interview. Twenty‐four individuals were randomised: 12 to control group and 12 to filtered group. Two individuals in the filtered group did not received [sic] the allocated intervention. Another was excluded to the final analysis. In the control group, 1 individual was excluded…" No trial group changes, ITT analysis: reported explanation of participants' data losses after randomisation (off‐pump CABG, haemodynamic instability before surgery, postoperative pneumothorax)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

De Backer 1996.

Study characteristics
Methods	Randomised double‐blind study Run‐in period: not specified. Received: 2 May 1995; accepted: 26 April 1996 Number of study centres and location: single centre. Location not specified. Authors affiliated with University of Antwerp (UIA), Universiteitsplein 1, B‐2610 Antwerpen‐Wilrijk, Belgium Setting: secondary care Date of study: 1996
Participants	N randomised: 18 N lost: 0 N analysed: 18; intervention group (n = 10), control group (n = 8) Inclusion: 'low‐risk' patients undergoing saphenous‐vein elective coronary artery bypass surgery (none had a history of recent peptic ulcer disease, asthma, or obstructive or other pulmonary disease; no steroid use prior or during surgery; no ARDS; no history of left ventricular failure)
Interventions	Timing: 24 hours to 12 hours before surgery Type: high dose of N‐acetylcysteine (72 mg/kg IV as a bolus, later 72 mg/kg over 12 hours) Comparison: placebo (saline in a continuous IV infusion over 13 h before standard CPB surgery)
Outcomes	Primary outcomes: neutrophil activation products and IL‐8 in blood samples and in BAL Secondary outcome: lung injury Time points: before surgery, start of ECC, end of ECC, end of surgery, 4 hours after surgery, 24 hours after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded: preoperative care and medications, general anaesthesia, priming solution, type of extracorporeal support system, and postoperative care were the same in all patients.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified: only reported "double‐blind study"
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This study was supported by a research grant of Inpharzam Belgium (a division of Zambon).

De Seigneux 2012.

Study characteristics
Methods	Double‐blinded, randomised controlled study Duration: 1 year Number of study centres and location: single centre. Geneva University Hospitals, Geneva, Switzerland Setting: secondary care Date of study: 2012
Participants	N randomised: 80 N lost: 0 N analysed in part A: 60 Inclusion criteria: all adults (> 18 years old) admitted to the ICU post‐cardiac surgery and at risk for AKI (mechanical ventilation, sepsis, postoperative state, haemodynamic impairment, previous chronic kidney disease) were screened for inclusion. Patients able to read, ask questions before inclusion, and sign a consent form were eligible for inclusion. Exclusion criteria: malignant hypertension or systolic blood pressure ≥ 150 mmHg at enrolment, Hb levels ≥ 120 g/L at randomisation, acute coronary syndrome, pregnancy, urinary output < 600 mL/12 h, patients unable to understand the information for any reason or in an emergency situation that prevented prior consent
Interventions	Timing: after cardiac surgery Type: α‐Epoetin 40,000 IU Comparison: control (IV isotonic sodium chloride)
Outcomes	Primary outcomes: change in urinary NGAL concentration from baseline and 48 hours after r‐HuEPO injection Secondary outcomes: changes in traditional renal function markers (serum Cr and cystatin C) and in cytokine levels Time points: 24 h before the surgery, in the morning 48 h and 96 h after the injection
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation code was generated by a computer.
Allocation concealment (selection bias)	Low risk	"Envelopes with allocation were prepared by the quality of care unit. A nurse from the Nephrology Unit opened the envelopes and prepared the syringes for injection."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described: all treatments were administered intravenously as a single dose. Investigators and patients were blinded to the treatment.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up data missing only on one participant
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	Funding from pharmaceutical company but disclosure of no involvement of funder

De Vroege 2004.

Study characteristics
Methods	Double‐blinded, randomised controlled trial Duration: N/A Number of study centres and location: not specified. Authors affiliated with different institutions in the Netherlands Setting: secondary care Date of study: 2004
Participants	N randomised: 51 N lost: 0 N analysed: 51 Inclusion criteria: patients scheduled for elective first‐time CABG, aged 45 years to 70 years, and with an LVEF > 40% Exclusion criteria: patients with known previous neurologic deficit, preoperative immunosuppressive therapy, preoperative use of NSAID, intra‐aortic balloon support requiring aneurysmectomy, insulin‐dependent diabetes mellitus, or a plasma Cr level > 150 μmol/L
Interventions	Timing: intraoperative Type: heparin‐coated circuit Comparison: uncoated circuit
Outcomes	Primary outcomes: inflammatory markers of complement activation (C3b/c), elastase‐ 1‐antitrypsin complex, and sPLA2 during and after CPB, partial oxygen pressure (PaO2 and PvO2), partial carbon dioxide pressure (PaCO2 and PvCO2), saturation (SaO2 and Sv2), aHb, vHb, aHCT, and vHCT. Each time arterial and venous blood samples were obtained, the FiO2, gas flow, nasopharyngeal temperature, and blood flow were also taken. Time points: samples – before CPB (before the induction of anaesthesia), on arrival in ICU, 2 h in the ICU, 8 h in the ICU, and POD 1
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described: physicians involved were blinded for randomisation, as were the physicians involved in postoperative care.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up data missing on only one participant
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	Supported in part by Jostra Medizintechnik GmbH, Hirrlingen, Germany

De Vroege 2005.

Study characteristics
Methods	Double‐blinded, randomised controlled trial Duration: N/A Number of study centres and location: N/A Setting: secondary care Date of study: 2005
Participants	N randomised: 51 N lost: 0 N analysed: 51 Inclusion criteria: patients scheduled for elective, first‐time CABG; age 45 years to 70 years, and with an LVEF > 40% Exclusion criteria: patients with known previous neurologic deficit, preoperative immunosuppressive therapy, preoperative use of NSAID, intra‐aortic balloon support requiring aneurysmectomy, insulin‐dependent diabetes mellitus, or a plasma Cr level > 150 μmol/L
Interventions	Timing: intraoperative Type: heparin‐coated circuit Comparison: uncoated circuit
Outcomes	Primary outcomes: partial oxygen pressure (PaO2 and PvO2), partial carbon dioxide pressure (PaCO2 and PvCO2), saturation (SaO2 and SvO2), Hb (aHb and vHb), and Hct (aHCT and vHCT, the FiO2, gas flow, nasopharyngeal temperature, cardiac output or blood flow, cardiac index or pump flow index, arterial blood pressure, pulmonary artery pressure, central venous pressure, wedge pressure, S100b, N‐acetyl‐beta‐D‐glucosaminidase, nitrotyrosine, C4b/c activation products, CRP Time points: samples – after the induction of anaesthesia and before CPB (pre‐CPB), 15 min after starting CPB, at termination of CPB, 30 min after the start of protamine infusion (post‐CPB), at arrival in the ICU, after 2 h in the ICU, and after 8 h in the ICU
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The extracorporeal tubing packs with oxygenator, coated or uncoated, were randomly picked by perfusionist.
Allocation concealment (selection bias)	Unclear risk	The extracorporeal tubing packs with oxygenator, coated or uncoated, were randomly picked by perfusionist.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described, also physicians involved in postoperative care
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up data missing only on one participant
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

DECS 2012.

Study characteristics
Methods	Multi‐centre, randomised, double‐blind, placebo‐controlled trial Duration: May 2006 to November 2011. Published 2017 Number of study centres and location: eight centres in the Netherlands Setting: secondary care Date of study: 2012
Participants	N randomised: 4494 N lost: 12 N analysed: 4482 Inclusion criteria: patients aged 18 years or older undergoing cardiac surgery with CPB Exclusion criteria: an emergent or off‐pump procedure and a life expectancy of < 6 months
Interventions	Timing: after induction of anaesthesia but before initiation of CPB Type: dexamethasone (1 mg/kg) Comparison: placebo (saline)
Outcomes	Primary outcomes: a composite of death, MI, stroke, renal failure, or respiratory failure within 30 days of randomisation Secondary outcomes: each separate component of the primary end point (i.e. death, MI, stroke, renal failure, or respiratory failure within the first 30 days); postoperative infections; postoperative AF; highest serum glucose concentration in the ICU; highest body temperature in the ICU; postoperative delirium (defined as the postoperative indication for treatment with neuroleptic drugs); time to weaning from postoperative mechanical ventilation; and time to discharge from the ICU and from the hospital Timing: within 30 days of randomisation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated 1:1 randomisation scheme was used.
Allocation concealment (selection bias)	Low risk	"The study drug was supplied in packaged ampoules, each assigned to a unique study number. Packages and ampoules of dexamethasone and placebo were identical and contained an equal volume dexamethasone solution or normal saline, respectively"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described: participants and caregivers were unaware of study group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators blinded: researchers were unaware of study group assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up data missing only on 12 participants: different proper reasons were reported into enrolment flowchart.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Defraigne 2000 (A).

Study characteristics
Methods	Randomised control trial Run‐in period: not specified. Published 2000 Number of study centres and location: single centre. University Hospital of Liege, Liege, Belgium
Participants	One hundred participants were assessed in part A of this study. N randomised: 100 N lost: 0 N analysed: 100 Inclusion criteria: patients undergoing CPB for elective cardiac surgery Exclusion: age > 75 years; emergency surgery; EF < 30%; left end‐diastolic pressure > 25 mmHg; heparin treatment at the time of surgery; coagulopathy; severe pulmonary, renal, hepatic, and cerebrovascular diseases; and neoplasia
Interventions	Timing: intraoperative Type: heparin‐coated circuits with aprotinin administration (a high‐dose regimen) Comparison: uncoated circuit without aprotinin
Outcomes	Primary outcomes: cytokine release and neutrophil activation (TNF‐a, IL‐6, IL‐8, MPO, and elastase levels) Secondary outcomes: postoperative chest drainage volume and amounts of transfused blood products Time points: after the induction of anaesthesia (T1), 5 minutes after aorta unclamping (T2), after protamine administration (T3), and 24 hours after protamine administration (T4)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation was accomplished using a random number table.
Allocation concealment (selection bias)	Low risk	"The perfusionist performed the assignment immediately preoperatively by opening a sealed numbered envelope"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The cannulae tubing, and oxygenator used in these two circuits were identical in appearance, so that all members of the surgical and anesthesia teams, excluding the perfusionist were blinded to the patients [sic] assignment"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

Defraigne 2000 (B).

Study characteristics
Methods	Randomised control trial Run‐in period: not specified. Published 2000 Number of study centres and location: single centre. University Hospital of Liege, Liege, Belgium
Participants	One hundred participants were assessed in part B of this study. N randomised: 100 N lost: 0 N analysed: 100 Inclusion criteria: patients undergoing CPB for elective cardiac surgery Exclusion criteria: age > 75 years; emergency surgery; EF < 30%; left end‐diastolic pressure > 25 mmHg; heparin treatment at the time of surgery; coagulopathy; severe pulmonary, renal, hepatic, and cerebrovascular diseases; and neoplasia
Interventions	Timing: intraoperative Type: uncoated circuits with aprotinin administration Comparison: uncoated circuit without aprotinin
Outcomes	Primary outcomes: cytokine levels (TNF‐a, IL‐6, IL‐8), MPO, and elastase Secondary outcomes: postoperative chest drainage volume and amounts of transfused blood products Time points: after the induction of anaesthesia (T1), 5 minutes after aorta unclamping (T2), after protamine administration (T3), and 24 hours after protamine administration (T4)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation was accomplished using a random number table.
Allocation concealment (selection bias)	Low risk	"The perfusionist performed the assignment immediately preoperatively by opening a sealed numbered envelope"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The cannulae tubing, and oxygenator used in these two circuits were identical in appearence [sic], so that all members of the surgical and anesthesia teams, excluding the perfusionist were blinded to the patients [sic] assignment"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

Dekker 2020.

Study characteristics
Methods	Prospective, randomised, blinded study Run‐in period: not specified. Published 2019 Number of study centres and location: single centre. Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam Cardiovascular Sciences, Amsterdam, the Netherlands
Participants	A total of 26 adults undergoing elective CABG surgery with CPB were divided. Exclusion criteria: previous cardiac surgery; emergency surgery; type 1 diabetes mellitus; BMI > 35; and haematologic, hepatic, or renal disease (estimated GFR < 50 mL/min/1.73 m2)
Interventions	Type: heparin‐coated circuit Control: PC‐coated circuit Timing: intraoperative
Outcomes	Hct, capillary RBC concentration, perfused boundary region, syndecan 1, heparan sulphate Timing: before, during, and after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Low risk	"Before first inclusion, envelopes with PC and HC coating allocation were prepared, sealed, and randomly numbered. The day before surgery, once informed consent was obtained, the sealed envelope with coating allocation was opened by the clinical perfusionist on duty and corresponding CPB coating was prepared."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Triple‐blinded: "Before first inclusion, envelopes with PC and HC coating allocation were prepared, sealed, and randomly numbered. The day before surgery, once informed consent was obtained, the sealed envelope with coating allocation was opened by the clinical perfusionist on duty and corresponding CPB coating was prepared. All study measurements were performed by a researcher who was blinded for group allocation. Group allocation was revealed when data acquisition and analysis were completed."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"All study measurements were performed by a researcher who was blinded for group allocation. Group allocation was revealed when data acquisition and analysis were completed."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	"Three patients received different coating strategies than randomized for (rescheduling of surgery) and were excluded from the final analysis (Fig 1). A system crash (n = 9) and an inaccessible sublingual area owing to postoperative use of an oxygen mask (n = 1) resulted in the exclusion of 10 patients, leaving a total of 26 patients that were included in the final analysis (CONSORT flow diagram of participant allocation, follow‐up, and analysis)."
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The authors report no conflict of interest.

Demir 2009.

Study characteristics
Methods	Randomised controlled trial Duration: not specified Number of study centres and location: single centre. Istanbul Medical Faculty, Istanbul University, Capa, Istanbul, Turkey Setting: secondary care Date of study: 2009
Participants	N randomised: 30 N lost: 0 N analysed: 30 Inclusion criteria: adults scheduled for CABG with CPB Exclusion criteria: history of a previous cardiac surgery; need for valve replacement or repair of ventricular aneurysm in addition to coronary bypass surgery; anti‐inflammatory drug intake except acetylsalicylic acid; and presence of a cause precluding the use of corticosteroids, insulin‐dependent diabetes mellitus, or active inflammatory disease
Interventions	Timing: before CPB Type: MP (1 g) Comparison: control (no additional medication)
Outcomes	Primary outcomes: NSE levels Secondary outcomes: IL‐6, IL‐10 Time points: (T1) prior to skin incision, (T2) 10 min after CPB, (T3) 4 h after CPB, and (T4) 24 h after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

Diego 1997 (A).

Study characteristics
Methods	Prospective, randomised, blinded study Run‐in period: not specified. Published 1997 Number of study centres and location: single centre. University of Nebraska Medical Center, Omaha, Nebraska, United States.
Participants	Fourteen participants were assessed in part A of this study. Forty adult male human patients scheduled for myocardial revascularisation
Interventions	Methylprednisolone 1 g intravenously before CPB vs control Forty adult male human patients scheduled for myocardial revascularisation were divided into four groups (n = 10): control; Methylprednisolone 1 g intravenously before CPB; aprotinin low‐dose protocol; and aprotinin full‐dose protocol.
Outcomes	IL‐6 Timing: measured at baseline and 1 h and 24 h after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by a computer‐generated sequence.
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No statement on funding sources

Diego 1997 (B).

Study characteristics
Methods	Prospective, randomised, blinded study Run‐in period: not specified. Published 1997 Number of study centres and location: single centre. University of Nebraska Medical Center, Omaha, Nebraska, United States.
Participants	Thirteen participants were assessed in part B of this study. Forty adult male human patients scheduled for myocardial revascularisation
Interventions	Aprotinin full‐dose protocol vs control Forty adult male human patients scheduled for myocardial revascularisation were divided into four groups (n = 10): control; MPS, 1 g intravenously before CPB; aprotinin low‐dose protocol; and aprotinin full‐dose protocol.
Outcomes	IL‐6
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how patients were randomised
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No funding reported

Draxler 2019.

Study characteristics
Methods	Single‐centre cohort of multi‐centre randomised controlled trial Duration: 9 years and 7 months (March 2006 to October 2015) Number of study centres and location: multi‐centre international trial. Exact number of centres not specified. Centres in Australia, UK, Canada, Hong‐Kong, New Zealand, Singapore. Setting: secondary care Date of study: 2019
Participants	N randomised: 617 N lost: 4 N analysed: 613 Inclusion criteria: males and females, age ≥ 18 years; written, informed consent; elective coronary artery surgery (on‐pump or off‐pump); patient is at increased risk of major complications (defined by any of the following: age ≥ 70 years, LV impairment [fractional area change < 20%, EF < 40%, or at least moderate impairment on ventriculography], concomitant valvular or aortic surgery, LV aneurysmectomy, repeat cardiac surgery ['re‐do'], COPD, renal impairment [serum Cr > 150 µmol/L or Cr clearance < 45 mL/min], obesity [BMI > 25], pulmonary hypertension [mean pulmonary artery pressure > 25 mmHg], peripheral vascular disease). Exclusion criteria: poor (English) language comprehension; clinician preference for antifibrinolytic therapy; urgent surgery for unstable coronary syndromes in which, for clinical reasons, anti‐Plt medication cannot be discontinued; active peptic ulceration; allergy or contraindication to aspirin or TXA; aspirin therapy within 4 days of surgery (until January 2013); warfarin or clopidogrel therapy within 7 days of surgery or GIIb/IIIa antagonists within 24 hours of surgery; thrombocytopenia or any other known history of bleeding disorder; severe renal impairment (serum Cr > 250 µmol/L or estimated Cr clearance < 25 mL/min); recent haematuria; thromboembolic disease relating to history of postoperative or spontaneous pulmonary embolism, spontaneous arterial thrombosis, or familial hypercoagulability (e.g. lupus anticoagulant, protein C deficiency); pregnancy.
Interventions	Timing: after induction of anaesthesia but before CPB Type: TXA (at a dose of 100 mg/kg over 30 minutes for the first 183 patients at the lead site; this dose was reduced to 50 mg/kg for all subsequent patients because of a concern for seizure risk) Comparison: control
Outcomes	Primary outcomes: the rate of healthcare‐associated infection in the first 30 days postoperatively Secondary outcomes: specific sites of infection, surgical blood loss, blood transfusions, and duration of intensive care and hospital stay Any outcomes: N/A Time points: 30 days
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated code that was accessed by means of an automated telephone voice‐recognition or web‐based service
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of personnel described
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No trial group changes were noted, but data from four participants were not analysed.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

El Azab 2002 (A).

Study characteristics
Methods	Randomised controlled trial Duration: not specified Number of study centres and location: single centre. Thoraxcentre Ignatius Breda, PO Box 90108, 4800 RA Breda, The Netherlands Setting: secondary care Date of study: 2002
Participants	A total of 15 participants were assessed in part A of this study. Group 1 (10), control group (5) Inclusion criteria: patients with angina pectoris NYHA class II–IV and EF > 40% undergoing elective CABG surgery Exclusion criteria: patients with unstable angina pectoris, severely impaired LV function (EF < 40%), severe systemic noncardiac disease, renal or liver impairment, insulin‐dependent diabetes, recent MI (< 6 weeks), or immunosuppressive treatment
Interventions	Timing: induction/intraoperative Type: sevoflurane Control: a moderate dose of sufentanil and midazolam
Outcomes	Primary outcomes: IL‐6, IL‐8, TNF‐α Time points: Time 0, before induction of anaesthesia; Time 1, immediately before skin incision; Time 2, immediately before CPB; Time 3, after aortic declamping; Time 4, end of CPB; Time 5, 2 h after skin closure; Time 6: 24 h after skin closure
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

El Azab 2002 (B).

Study characteristics
Methods	Randomised controlled trial Duration: not specified Number of study centres and location: single centre. Thoraxcentre Ignatius Breda, PO Box 90108, 4800 RA Breda, The Netherlands Setting: secondary care Date of study: 2002
Participants	Intervention group: 10; control group: 5 Inclusion criteria: patients with angina pectoris NYHA class II–IV and EF > 40%, undergoing elective CABG surgery Exclusion criteria: patients with unstable angina pectoris, severely impaired LV function (EF < 40%), severe systemic noncardiac disease, renal or liver impairment, insulin‐dependent diabetes, recent MI (< 6 weeks), or immunosuppressive treatment
Interventions	Timing: induction/intraoperative Type: propofol (2 mg/kg) and sufentanil (0.7 μg/kg) Comparison: control (sufentanil and midazolam)
Outcomes	Primary outcomes: IL‐6, IL‐8, TNF‐α Time points: Time 0, before induction of anaesthesia; Time 1, immediately before skin incision; Time 2, immediately before CBP; Time 3, after aortic declamping; Time 4, end of CBP; Time 5, 2 h after skin closure; Time 6, 24 h after skin closure
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

El‐Hamamsy 2007.

Study characteristics
Methods	Randomised, double‐blind, placebo‐controlled clinical trial Duration: 1 year and 2 months (January 2003 to March 2004). Received for publication June 8, 2005; revisions received April 25, 2006; accepted for publication May 24, 2006. Number of study centres and location: single centre. Research Center and Department of Surgery, Montreal Heart Institute and Université de Montréal, Montreal, Quebec, Canada. Setting: secondary care Date of study: 2007
Participants	N randomised: 100 N lost: 0 N analysed: 100 Inclusion criteria: adults undergoing primary CABG with CPB Exclusion criteria: emergency operations, AMI within < 3 weeks, prior cardiac surgery, age > 80 years, EF < 20%, and concomitant procedures.
Interventions	Timing: day before surgery Type: N‐acetylcysteine (600 mg orally the day before and the morning of the operation, a bolus of 150 mg/kg of IV N‐acetylcysteine before skin incision, followed by perfusion at 12.5 mg/kg over 24 h) Comparison: placebo
Outcomes	Primary outcomes: mean postoperative release of cTnT levels between the two groups Secondary outcomes: the rate of MI (as defined by CK‐MB level 50, new Q wave on ECG in a given territory, or both), renal function (Cr), bleeding, low cardiac output syndromes, arrhythmias, and mean levels of CK‐MB Time points: 1 h, 2 h, 4 h, 8 h, 12 h, and 24 h postoperatively, then 2 days, 3 days, and 4 days postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: oral and IV preparations were similar for both groups.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed.

Ellam 2020.

Study characteristics
Methods	Single‐centre, prospective, randomised, open‐label clinical trial Run‐in period: February 2010 to August 2014. Published online 2021 Number of study centres and location: single centre. Kuopio University Hospital, Kuopio, Finland.
Participants	Two hundred forty adults undergoing CABG for the first time Exclusion criteria: history of episodes of AF or atrial flutter, sick sinus syndrome, II‐ or III‐degree atrioventricular block, heart failure, corticosteroid or immunosuppressive medication, thyroid disease, redo or emergency surgery, enrolled to another study.
Interventions	Intervention: MECC Control group: CECC
Outcomes	Primary outcomes: incidence of postoperative AF Secondary outcomes: incidence of postoperative MI, first postoperative CK‐MB, stroke, and re‐sternotomy caused by bleeding
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by a computer‐generated randomisation list.
Allocation concealment (selection bias)	High risk	Open‐label trial
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label trial and staff aware of allocation of intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Trial is open‐label, but authors do not specify whether outcome assessors were blinded (it could have been done, as the intervention was intraoperative and some of the outcomes are laboratory‐based).
Incomplete outcome data (attrition bias) All outcomes	Low risk	Three participants excluded from follow‐up due to change to intraoperative management: small number compared to overall cohort
Selective reporting (reporting bias)	Unclear risk	Registered protocol only specifies primary outcome. All the other secondary outcomes described in the methodology are reported in the results, but not clear if it was prespecified.
Other bias	Low risk	The author(s) declared no potential conflicts of interest with respect to the research, authorship, or publication of this article. This study was supported by the research grant from the Government Research Foundation of Finland.

Enc 2006.

Study characteristics
Methods	Double‐blind, single‐centre, randomised, prospective study Run‐in period: not specified. Received: July 5, 2005; accepted: November 4, 2005 Number of centres and study period: single centre. Thoracic and Cardiovascular Surgery Center, Istanbul, Turkey
Participants	Forty nondiabetic male patients with three‐vessel disease undergoing first‐time bypass surgery; mean age 58.35 years
Interventions	Intervention group: methylprednisolone 25 mg/kg Control group: saline Timing: 1 h before CPB
Outcomes	Primary outcomes: cTnI levels and clinical outcomes Timing: before surgery; at the second hour after CPB; and postoperatively 6 h, 24 h, and day 5
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind study but no further description
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not documented
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported. Data from all participants were included in the final analysis. No ITT analysis is included.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Engelman 1995.

Study characteristics
Methods	Randomised controlled trial Duration: not specified. Ethical approval granted 1990. Published 1995 Number of study centres and location:single centre. Division of Cardiac Surgery, Baystate Medical Center, Springfield, Massachusetts, USA Setting: secondary care Date of study: 1995
Participants	N randomised: 19 N lost: 0 N analysed: 19 Inclusion criteria: patients undergoing elective coronary revascularisation Exclusion criteria: N/A
Interventions	Intervention: steroids (1 g of MP intravenously before bypass and 4 mg of dexamethasone every 6 h for 4 doses during the first 24 h of recovery) Comparison: conventional management.
Outcomes	Primary outcomes: complement (C3a, C5a) and cytokine (IL‐1, IL‐3, IL‐8) levels Time points: before operation; immediately after completion of CPB; and at 24 h, 48 h, and 72 h postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed: supported by National Institutes of Health grants HL 22559‐14 and HL 34360‐07.

Fabbri 2001.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: November 1999 to March 2000. Published 2001 Number of study centres and location: single centre. Department of Cardiovascular Surgery, Vicenza General Hospital, Italy
Participants	Forty participants were randomly allocated to the leucocyte‐depletion group (group F, 20 participants) or to the control group (group C, 20 participants). Mean age: 70 years Gender (female/male ratio): 30% No high‐risk population Inclusion criteria: underwent isolated CABG Exclusion criteria: abnormal preoperative lung function, allergy or infection, the presence of haematological diseases (such as leukaemia or lymphoma), CPB < 60 minutes
Interventions	Intervention group: leucocyte‐depleting arterial blood filter (LeukoGard 6; Pall Biomedical Products, East Hills, NY, USA). The filter was incorporated in the arterial circuit before the standard arterial filter.
Outcomes	Primary outcomes: blood cell count, elastase, IL‐8, TNF‐α, pulmonary function (evaluated by arterial blood gas monitoring levels + pulmonary respiratory index), intubation time, and LOS (ICU LOS and H‐LOS) Secondary outcomes: 12 h to 24 h of bleeding by chest drainage; averaged transfusion requirement in the perioperative period included units of packed RBCs, platelet concentrates, and FFP. Timing: preoperatively (T1), at aortic declamping (T2), at 60 min, after CPB (T3), after arriving at the ICU (T4), and 24 h after the operation (T5)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	There was no significant difference between the two groups. No detailed description of randomisation method is given.
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No ITT analysis; no clear number of participants reported in the results
Selective reporting (reporting bias)	Unclear risk	IL‐8 and TNF‐α parameters were too dishomogeneous to be compared either to each other or to each of the other variables. For this reported reason, there is a lack of these data.
Other bias	Low risk	No funding was disclosed.

Farag 2017 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Received January 2016; revised February 2016. Number of study centres: not specified. Authors affiliated with institutions in Germany and the UK
Participants	Sixty consecutive participants undergoing CABG Mean age: 63.65 years Sex (female/male ratio): 13.33% No high‐risk population Inclusion criteria: not specified Exclusion criteria: not specified
Interventions	Intervention group: patients in group A were operated with the MECC (Maquet, Rastatt, Germany). The MECC system comprised a closed circuit with a Quadrox membrane oxygenator and a centrifugal pump (Maquet). Control group: CECC (Maquet); for CECC (group C), a membrane oxygenator with cardiotomy reservoir and a roller pump (Maquet)
Outcomes	Myocardial protection (TnT and CK‐MB); IL‐1b, IL‐6, IL‐10, TNF‐α, MIF, RAGE, and CD40L; need for blood transfusion; mechanical ventilation time; ischaemic time; total duration of surgery; or the number of grafts and clinical outcomes (major adverse cardiac events) Arterial blood samples were collected at intervals before CPB (T0), 30 min after initiation (T1), after termination of CPB (T2), 6 h after CPB (T3), and 24 h after CPB (T4).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Consecutive patients with comparable preoperative characteristics but no further randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors declared no conflict of interest.

Farag 2017 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Received January 2016; revised February 2016 Number of study centres: not specified. Authors affiliated with institutions in Germany and the UK
Participants	60 consecutive adults undergoing CABG Mean age: 63.65 years Sex (female/male ratio): 13.33% No high‐risk population Inclusion criteria: not specified Exclusion criteria: not specified
Interventions	Intervention group: operated on with ECC.O (Sorin, Italy). The ECC.O system comprised a small membrane oxygenator and centrifugal pump (Stockert, Sorin, Germany). Control group: CECC (Maquet); for CECC (group C), a membrane oxygenator with cardiotomy reservoir and a roller pump (Maquet)
Outcomes	Myocardial protection (TnT and CK‐MB); IL‐1b, IL‐6, IL‐10, TNF‐a, MIF, RAGE, and CD40L; need for blood transfusion; mechanical ventilation time; duration of ischaemic time; total duration of surgery; or the number of grafts and clinical outcomes (major adverse cardiac events) Arterial blood samples were collected at intervals before CPB (T0), 30 min after initiation (T1), after termination of CPB (T2), 6 h after CPB (T3), and 24 h after CPB (T4).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Consecutive patients with comparable preoperative characteristics but no further randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up reported, but no ITT analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors declared no conflict of interest.

Farneti 2008.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study date: 2008 Number of study centres and locations: single centre; Cardiac Surgery Department "G. Pasquinucci" Hospital, Massa (MS), Italy
Participants	Twenty participants were randomly assigned to standard CPB (n = 10) or Synergy (n = 10). Mean age: 62.3 years Gender (female/male ratio): not specified No high‐risk patients Inclusion criteria: patients diagnosed with ischaemic heart disease Exclusion criteria: not specified
Interventions	Intervention group: MECC system (mini CPB Synergy Sorin Group), a closed‐miniaturised apparatus (maximum blood flow rate of 8 L/min) associating a centrifugal pump (Revolution, Cobe Cardiovascular, Arvada, CO, USA), a microporous hollow‐fibre membrane oxygenator (surface area of 2.0 m2), and a heat exchanger (surface area of 0.14 m2) Control group: standard CPB, a membrane oxygenator with a cardiotomy reservoir (Monolyth‐Pro®, Sorin Group)
Outcomes	Plt expression of PAC‐1; monocyte/granulocyte‐platelet conjugates, leucocyte‐platelet adhesion index, IL‐6, TNF‐α, TATc, PF‐1.2, β‐TG, and sP selectin (sCD62P); and clinical outcomes Blood samples were drawn at the time of anaesthesia (T1), at the end of CPB (T2), and at 4 hours (T3) and 24 hours (T4) after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	All participants had comparable preoperative characteristics, but no further randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Plasma levels of β‐TG, TNF‐α, and IL‐6 did not differ significantly in the two groups, so data were not shown.
Other bias	High risk	This study was supported by the Sorin Group.

Farsak 2012 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: over a 4‐year period Study dates: 19 November 2010; accepted 22 March 2011; published online 10 November 2011 Number of study centres and locations: single centre, Department of Cardiovascular Surgery, Bayindir Hospital, Sogutozu, 06530 Ankara, Turkey
Participants	Eight hundred seventy‐five participants were prospectively randomised to one of four perfusion protocols: Group 1, PMEA‐coated circuits + leucocyte filters (n = 214); Group 2, uncoated ECC + full Hammersmith aprotinin (n = 212); Group 3, PMEA‐coated ECC + leucocyte filters + full Hammersmith aprotinin (n = 199); and Group 4, control, no treatment (n = 250). Mean age: 65.1 years Sex (female/male ratio): 51.34% High‐risk patients Inclusion criteria: CABG (EuroSCORE 6+) Exclusion criteria: known coagulopathy, endocarditis, and the inability to obtain informed consent
Interventions	Intervention group: PMEA‐coated circuits (Capiox SX 18; Terumo, Ann Arbor, MI, USA) + leucocyte filters (LG6B and BC2; Pall, East Hills, NY, USA). Leucocyte filters were deployed strategically approximately 30 minutes before cross‐clamp release. Control group: no treatment
Outcomes	IL‐2, C3a, CK‐MB, and lactate levels (CS blood); neutrophil CD11b/CD18; TATc; complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts); PT; activated partial thromboplastin time; fibrinogen levels; and standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions Perioperative clinical outcomes: haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, Cr clearance, duration of ICU and hospital stay, and perioperative mortality Platelet function was evaluated by TEG (ROTEG, Pentapharm, Munich, Germany) during the operation. Blood samples were collected at times T1 (after the induction of anaesthesia); T2 (after heparin administration and before the initiation of CPB); T3 (15 minutes after CPB); T4 (before cessation of CPB); T5 (15 minutes after protamine reversal); and T6 (in the ICU on the first POD at 8:00 a.m.).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No authors of this study have a conflict of interest.

Farsak 2012 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: over a 4‐year period Study dates: 19 November 2010; accepted 22 March 2011; published online 10 November 2011 Number of study centres and locations: single centre, Department of Cardiovascular Surgery, Bayindir Hospital, Sogutozu, 06530 Ankara, Turkey
Participants	Eight hundred seventy‐five participants were prospectively randomised to one of four perfusion protocols: Group 1, PMEA‐coated circuits + leucocyte filters (n = 214); Group 2, uncoated ECC + full Hammersmith aprotinin (n = 212); Group 3, PMEA‐coated ECC + leucocyte filters + full Hammersmith aprotinin (n = 199); and Group 4, control, no treatment (n = 250). Mean age: 69.19 years Gender (female/male ratio): 53.22% High‐risk patients Inclusion criteria: CABG (EuroSCORE 6+) Exclusion criteria: known coagulopathy, endocarditis, and the inability to obtain informed consent
Interventions	Intervention group: uncoated ECC (Capiox SX 18) and full Hammersmith aprotinin (10,000 KIU IV test dose, 2 million KIU aprotinin through a central line before sternotomy, and 500,000 KIU/h until the end of the operation) Control group: no treatment
Outcomes	IL‐2, C3a, CK‐MB, and lactate levels (CS blood); neutrophil CD11b/CD18; TATc; complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts); PT; activated partial thromboplastin time; fibrinogen levels; and standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions Perioperative clinical outcomes: haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, Cr clearance, duration of ICU and hospital stay, and perioperative mortality Plt function was evaluated by TEG (ROTEG, Pentapharm, Munich, Germany) during the operation. Blood samples were collected at times T1 (after the induction of anaesthesia); T2 (after heparin administration and before the initiation of CPB); T3 (15 minutes after CPB); T4 (before cessation of CPB); T5 (15 minutes after protamine reversal); and T6 (in the ICU on the first POD at 8:00 a.m.).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No authors of this study have a conflict of interest.

Farsak 2012 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: over a 4‐year period Study dates: 19 November 2010; accepted 22 March 2011; published online 10 November 2011 Number of study centres and locations: single centre, Department of Cardiovascular Surgery, Bayindir Hospital, Sogutozu, 06530 Ankara, Turkey
Participants	Eight hundred seventy‐five participants were prospectively randomised to one of four perfusion protocols: Group 1, PMEA‐coated circuits + leucocyte filters (n = 214); Group 2, uncoated ECC + full Hammersmith aprotinin (n = 212); Group 3, PMEA‐coated ECC + leucocyte filters + full Hammersmith aprotinin (n = 199); and Group 4, control, no treatment (n = 250). Mean age: 62.25 years Gender (female/male ratio): 45.39% High‐risk patients Inclusion criteria: CABG (EuroSCORE 6+) Exclusion criteria: known coagulopathy, endocarditis, and the inability to obtain informed consent
Interventions	Intervention group: PMEA‐coated ECC + leucocyte filters + full Hammersmith aprotinin (10,000 KIU IV test dose, 2 million KIU aprotinin through a central line before sternotomy, and 500,000 KIU/h until the end of the operation) Control group: no treatment
Outcomes	IL‐2, C3a, CK‐MB, and lactate levels (CS blood); neutrophil CD11b/CD18; TaTc; complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts); PT; activated partial thromboplastin time and fibrinogen levels; and standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions Perioperative clinical outcomes: haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, Cr clearance, duration of intensive care unit and hospital stay, and perioperative mortality Plt function was evaluated by TEG (ROTEG, Pentapharm, Munich, Germany) during the operation. Blood samples were collected at times T1 (after the induction of anaesthesia); T2 (after heparin administration and before the initiation of CPB); T3 (15 minutes after CPB); T4 (before cessation of CPB); T5 (15 minutes after protamine reversal); and T6 (in the ICU first POD at 8:00 a.m.).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No authors of this study have a conflict of interest.

Fattouch 2007.

Study characteristics
Methods	Single‐centre prospective, randomised, placebo‐controlled, double‐blind study Run‐in period: between February 2003 and April 2006. Number of study centres and location: single centre. Unit of Cardiac Surgery, University of Palermo, Palermo, Italy.
Participants	Eighty patients affected with STEMI who underwent emergent CABG. Participants were assigned to two groups: C1‐INH group (38 participants) and placebo group (42 participants). Mean age: 59.9 years Gender (female/male ratio): 32.5% High‐risk population Inclusion criteria: STEMI and coronary artery reperfusion with CABG within 12 hours after the onset of symptoms or the ischaemic event, recurrent myocardial ischaemia refractory to medical therapy, no candidates for fibrinolytic or primary PTCA, primary PTCA failure with persistent symptoms or haemodynamic instability, life‐threatening ventricular arrhythmias in patients with left main stenosis or three‐vessels disease, and patients with multivessel or left main disease and haemodynamic instability Exclusion criteria: mechanical complications of MI; patients requiring cardiopulmonary resuscitation; onset of cardiac shock > 6 hours; concomitant cardiac surgical procedures; associated renal, hepatic, or pulmonary diseases; patients with known complement deficiency or immune deficiency syndrome, known autoimmune disease, or evidence of infection; and coagulation disorder
Interventions	Participants were assigned in two groups: C1‐INH group (received 500 IU as IV bolus 10 min before reperfusion (aortic unclamping) followed by an IV infusion of 500 UI for 3 hours after surgery) and placebo group (received a saline solution).
Outcomes	Primary outcomes: mortality and MI size or the extent of irreversible myocardial cell injury Secondary outcomes: the incidence of LCOS; IABP and CPB support; inotropic support; time of intubation; ICU LOS and hospital LOS; major bleeding, infection, and renal failure requiring temporary haemodialysis; mechanical ventilation support; complement inhibition; haemodynamic data (mean arterial blood pressure, cardiac index, stroke volume); C1‐INH, C3a, C4a complement activation fragments, and cTnI serum levels Data were measured before, during, and after surgery at six different times: T1 (before induction of anaesthesia); T2 (after heparin dose); T3 (10 min after drug administration); T4 (2 h post‐CPB); T5 (12 h post‐CPB); and T6 (24 h after surgery).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The two groups did not differ in baseline characteristics, but there was no description of randomisation method.
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind study, no further description
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This work is supported by the Italian National Council of Research (CNR). No conflicts of interest were disclosed.

Fillinger 2002.

Study characteristics
Methods	Prospective, double‐blind, placebo‐controlled, randomised study Run‐in period: not specified. Published 2002 Number of centres and location: single centre. Dartmouth–Hitchcock Medical Center, Lebanon, New Hampshire, USA
Participants	Thirty adults undergoing elective CABG Patients were included if they were the first case of the day. Exclusion criteria: history of immune dysfunction, prior cardiac surgery, bacteria infection in past 30 days, or taking drugs, including glucocorticoids known to affect immune function
Interventions	Intervention group: 125 mg/kg intravenously MP 1 h before surgery and 0.3 mg/kg intravenously every 6 h × 4 doses Control: placebo (etomidate)
Outcomes	Primary outcomes: IL‐6 and IL‐10 Timing: at 60 min after CPB, on the morning of the first (P < 0.01) and third (P < 0.05) PODs
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States randomised, but no information on randomisation given
Allocation concealment (selection bias)	Low risk	Randomisation was performed by the hospital pharmacy with prepared syringes of solution administered by the anaesthesiologist intraoperatively and by the CTICU nurses postoperatively; no further description is given.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Study patients, investigators, and other physicians and nurses caring for the patients perioperatively were blinded to the treatment group.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Samples taken by ITU/anaesthetic team, not investigators
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No disclosures by the study authors

Florens 2001.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study date: 2001 Number of study centres and location: single centre; Department of Cardiovascular Surgery, Groupe Hospitalier Bichat‐Claude Bernard, 46 Rue Henri Huchard, 75018 Paris
Participants	Twenty patients undergoing various cardiac operations who required CPB were randomised. Mean age: 65 years Sex (female/male ratio): 30% No high‐risk population Inclusion criteria: normal preoperative cholesterol and triglyceride levels and not receiving lipid‐lowering therapy (fibrates or statins) for at least 3 months prior to the date of surgery Exclusion criteria: chronic steroid medication
Interventions	Intervention: oral dose of atorvastatin (40 mg the evening before and 40 mg the morning of surgery at 7:00 a.m.) Control: no statin
Outcomes	Primary outcomes: measurements of several inflammatory mediators (P‐selectin, IL‐6, IL‐8, sICAM‐1, lactoferrin), neutrophil CD11b surface adhesion molecule, oxidative burst, and the expression of the transcription NF‐kB (atrial biopsies were taken before and at the end of CPB) Secondary outcomes: clinical outcomes (time to extubation, temperature, and leucocytosis at 24 hours after the end of bypass; blood losses and weight gain over the first 24 postoperative hours; and serum Cr levels at POD 1) and indices of myocardial injury (new onset MI, peak postoperative levels of TnI, and requirements for inotropic or vasoactive drugs) Timing: P‐selectin at the end of bypass immediately before administration of protamine, IL‐6 4 hours after the end of bypass, IL‐8 at the end of bypass and 4 hours thereafter, and sICAM‐1 24 hours after the end of bypass
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The preoperative and intraoperative participant characteristics were not significantly different between the two groups; no detailed description of randomisation method is given (just that participants were randomised prospectively).
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All investigators involved in clinical care and biochemical assays were blinded as to the patient group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All investigators involved in clinical care and biochemical assays were blinded as to the patient group.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); there were no trial group changes, no withdrawals, and no losses to follow‐up reported but also no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	The post‐bypass neutrophil oxidative burst data in response to stimulation by PMA and zymosan were not shown because the response was markedly greater than during the pre‐bypass period, but the difference did not reach the level of statistical significance.
Other bias	Low risk	No funding was disclosed.

Formica 2009.

Study characteristics
Methods	Prospective, randomised study Run‐in period: April 2007 to October 2007 Number of study centres and locations: Single centre. San Gerardo Hospital, Monza, Italy
Participants	Sixty consecutive patients were randomised to MECC (n = 30) or off‐pump coronary revascularisation (OPCABG, n = 30). Mean age: 65.6 years Sex (female/male ratio): 31.67% No high‐risk population Inclusion criteria: primary and isolated CABG operation, at least two‐vessel disease, EF ≥ 40%, 30 years to 85 years of age, serum Cr < 1.8 mg/100 mL, and absence of acute or chronic inflammatory syndrome Exclusion criteria: small, calcified, and intramyocardial coronaries; recent or current steroid treatments; urgent or salvage operation; recent MI (< 10 days); unstable angina with IV medications; and conversion of OPCABG to standard CPB during the operation
Interventions	Intervention group: MECC (the MECC is a Bioline fully heparin‐coated system that includes a centrifugal pump (Rotaflow RF‐32, Maquet‐Jostra AG, Hirrlingen, Germany) and a continuous diffusion polymethylpentene membrane oxygenator Quadrox D with an integrated heat exchanger (Maquet‐Jostra AG, Hirrlingen, Germany) Control group: off‐pump coronary revascularisation (in the OPCABG group, heparin was given at a dose of 1.5 mg/kg with an activated clotting time target of 250 seconds. The anastomosis between the left ITA and the left anterior descending artery was always performed first, and then the proximal anastomosis in the aortic root was performed with the HeartString II Proximal Seal System device [Maquet AG, Hirrlingen, Germany])
Outcomes	Clinical outcomes (postoperative bleeding, requirement of blood transfusions, and incidence of AF), inflammatory response (CRP, leucocyte and neutrophil, monocyte cell level, IL‐6, and TNF‐α at the end of operation, post 24 hours and 48 hours), blood lactate, haemodilution (Hct and Hb levels), coagulative disorder (fibrinogen and Plt count), and markers of myocardial damage (cTnT and CK‐MB and CK mass, myoglobin) Blood samples for myocardial damage, haemodilution, and coagulative disorder were serially collected the day before surgery (T0), at the end of operation (T1), 24 hours after operation (T2), and 48 hours after operation (T3). Blood lactate and cytokine levels were analysed before the anaesthesia induction (T0), before CPB initiation in the MECC group or before the first distal anastomosis in the OPCABG group (T1), 5 minutes after aortic clamp removal in the MECC group or 5 minutes after the last distal anastomosis in the OPCABG group (T2), at the end of operation (T3), 12 hours after operation (T4), and 24 hours after operation (T5).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A randomisation list was generated by a computer algorithm. Preoperative characteristics did not differ between groups.
Allocation concealment (selection bias)	Unclear risk	All operations were performed by the same senior surgeon, but there was no information about how patients were placed in groups.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participant was withdrawn from the study. One participant in the OPCABG group died of acute severe respiratory disease on the fifth POD. There was no ITT analysis, but there were accurate descriptions of group size and cause of the observed death.
Selective reporting (reporting bias)	Unclear risk	One of the outcomes (coagulative disorder) was not properly assessed as the others.
Other bias	Low risk	No funding was disclosed.

Formica 2013 (A).

Study characteristics
Methods	Prospective, parallel‐group, randomised clinical trial Run‐in period: between June and December 2011 Number of study centres and locations: single centre; Cardiac Surgery Clinic, Department of Surgical Science and Interdisciplinary Medicine, University of Milano‐Bicocca, San Gerardo Hospital, Monza, Italy
Participants	Sixty‐one participants undergoing isolated CABG were prospectively randomised to MECC (n = 19), SECC (n = 20), or OPCABG (n = 22). Mean age: 69.35 years Sex (female/male ratio): 21.11% Low‐risk patients Inclusion criteria: first and isolated CABG operation, at least two‐vessel disease, EF ≥ 40%, age between 18 years and 85 years, serum Cr levels < 1.8 mg/100 mL, and absence of inflammatory syndromes and haematological disorders Exclusion criteria: calcified and intramyocardial coronary arteries, recent or current steroid treatments, emergency or urgency operation, recent MI (< 10 days), unstable angina with IV medications, and preoperative IABP
Interventions	Intervention group: MECC. The MECC system (Maquet‐Jostra AG, Hirrlingen, Germany) was a closed miniaturised circuit with no blood–air contact and no open venous reservoir. The system components included a centrifugal Rotaflow pump, a polimethylpentene membrane Quadrox D oxygenator, a heat exchanger, a venous bubble trap VBT160 located between the venous line and the centrifugal pump, an arterial filter, and a 1000‐mL closed bag used to prime and substitute volume during CPB. Control group: standard ECC system. The standard extracorporeal system consisted of a polyvinylchloride heparin‐coated circuit (Maquet‐Jostra, Hirrlingen, Germany), a hollow‐fiber polypropylene oxygenator (Quadrox Maquet‐Jostra, Hirrlingen, Germany), an open reservoir, a roller pump (Stockert, Munchen, Germany), a system of blood suction from the surgical field, a heat exchanger, and an arterial filter (Maquet‐Jostra, Hirrlingen, Germany).
Outcomes	Primary outcomes: perioperative mortality and morbidity (AF, new MI) Secondary outcomes: blood lactate; haemodilution; markers for inflammation and endothelial activation such as TNF‐α, IL‐6, monocyte chemotactic protein‐1, and E‐selectin; clinical outcomes; TnT; and CK‐MB Timing: during and after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Preoperative characteristics did not differ amongst the three groups; no randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	All the operations were performed by the same senior surgeon; no allocation information is given.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Two participants in the OPCABG group were withdrawn from the trial because of intraoperative conversion of OPCABG to standard CPB after haemodynamic instability; no ITT analysis.
Selective reporting (reporting bias)	Unclear risk	Inflammatory outcomes were not reported properly: TNF‐α and E‐selectin without SD in the figures and D‐dimer data were not shown.
Other bias	Low risk	This work was funded by the University fund for scientific research (Fondo di Ateneo per la Ricerca), University of Milano‐Bicocca.

Formica 2013 (B).

Study characteristics
Methods	Prospective, parallel‐group, randomised clinical trial Run‐in period: between June and December 2011; submitted for consideration January 2013; accepted for publication in revised form July 2013 Number of study centres and locations: single centre; Cardiac Surgery Clinic, Department of Surgical Science and Interdisciplinary Medicine, University of Milano‐Bicocca, San Gerardo Hospital, Monza, Italy
Participants	Sixty‐one participants undergoing isolated CABG were prospectively randomised to MECC (n = 19), SECC (n = 20), or OPCABG (n = 22). Mean age: 69.35 years Sex (female/male ratio): 21.11% Low‐risk patients Inclusion criteria: first and isolated CABG operation, at least two‐vessel disease, EF ≥ 40%, age between 18 years and 85 years, serum Cr levels < 1.8 mg/100 mL, and absence of inflammatory syndromes and haematological disorders Exclusion criteria: calcified and intramyocardial coronary arteries, recent or current steroid treatments, emergency or urgency operation, recent MI (< 10 days), unstable angina with IV medications, and preoperative IABP
Interventions	Intervention group: OPCABG. The anastomosis between the left ITA and the LAD artery was always performed first. The proximal anastomosis in the aortic root was done with the HeartString II Proximal Seal System device (Maquet AG, Hirrlingen, Germany). The distal anastomosis on the marginal branches and the right coronary was performed at the end. The heart was stabilised with the Octopus device (Medtronic, Minneapolis, MN). An Axius intracoronary shunt (Maquet, Hirrlingen, Germany) was used. Control group: standard ECC system. The standard extracorporeal system consisted of a polyvinylchloride heparin‐coated circuit (Maquet‐Jostra, Hirrlingen, Germany), a hollow‐fiber polypropylene oxygenator (Quadrox Maquet‐Jostra, Hirrlingen, Germany), an open reservoir, a roller pump (Stockert, Munchen, Germany), a system of blood suction from the surgical field, a heat exchanger, and an arterial filter (Maquet‐Jostra, Hirrlingen, Germany).
Outcomes	Primary outcomes: perioperative mortality and morbidity (AF, new MI) Secondary outcomes: blood lactate; haemodilution; markers for inflammation and endothelial activation such as TNF‐α, IL‐6, monocyte chemotactic protein‐1, and E‐selectin; clinical outcomes; TnT and CK‐MB Timing: during and after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Preoperative characteristics did not differ amongst the three groups; no randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	All the operations were performed by the same senior surgeon; no allocation information is given.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Two participants in the OPCABG group were withdrawn from the trial because of intraoperative conversion of OPCABG to standard CPB after haemodynamic instability; no ITT analysis.
Selective reporting (reporting bias)	Unclear risk	Inflammatory outcomes were not reported properly: TNF‐α and E‐selectin withur SD in the figures and Ddimer data were not shown.
Other bias	Low risk	This work was funded by the University fund for scientific research (Fondo di Ateneo per la Ricerca), University of Milano‐Bicocca.

Freyholdt 2003.

Study characteristics
Methods	Prospective, double‐blinded study Run‐in period: not specified; received 23 September 2002; accepted 24 April 2003 Number of study centres and locations: Department of Physiology, University of Munich; Department of Cardiothoracic Surgery, University of Essen; and Departments of Anesthesiology and Cardiovascular Surgery, German Heart Center, Munich, Germany
Participants	Participants undergoing CABGs (n = 30) with severely compromised LVEF (< 40%) were treated with either SNP (0.5 µg/kg/min) or placebo. Only participants with complete data sets (n = 24) were considered for analysis of cytokines. Mean age: 65 years Sex (female/male ratio): 28% High‐risk population Inclusion criteria: LVEF < 40%, age 80 years, and elective CABG surgery Exclusion criteria: known hypersensitivity against SNP or any other nitro substances, inflammatory diseases, intake of immunosuppressive drugs, and MI within the 14 days before operation
Interventions	Intervention group: sodium nitroprusside (0.5 µg/kg/min (Nipruss; Schwarz Pharma, Monheim, Germany) Control group: 5% glucose solution (the solvent for Nipruss) as placebo Timing: both infused via the central venous line of the Swan‐Ganz catheter for the first 60 minutes of reperfusion after cardiac arrest
Outcomes	Relationship between cytokines and cardiac index (IL‐6, IL‐8, and TNF‐α), haemodynamic measurements (cardiac index and systemic and pulmonary vascular resistance, cardiac output, heart rate, MAP, mean pulmonary artery pressure, central venous pressure, and pulmonary capillary wedge pressure), and clinical outcomes (mortality, inotropic support, duration of operation) Timing – Samples were drawn from the radial artery and the CS at the following times: before CPB; 5 minutes after release of the aortic cross‐clamp; after termination of CPB (approximately 35 minutes of reperfusion); and at the end of surgery (approximately 75 minutes of reperfusion).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind study, but no further description
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	High risk	Only participants with complete data sets (n = 24) were considered for analysis of cytokines. There was no accurate description as to why the absence of a complete dataset for these participants resulted in a different size for both groups.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The work was supported by the Deutsche Forschungsgemeinschaft.

Fromes 2002.

Study characteristics
Methods	Prospective study Run‐in period: not specified; received 18 September 2001; received in revised form 9 April 2002; accepted 14 June 2002 Number of study centres and locations: single centre; Department of Cardiac Surgery, Fondation Hospital Saint Joseph, 185 Rue Raymond Losserand, 75 674 Paris Cedex 14, France
Participants	Sixty consecutive patients were randomly assigned to either standard normothermic CPB (n = 30) or the MECC system(n = 30). Mean age: 61.6 years Sex (female/male ratio): not specified No high‐risk population Inclusion criteria: not specified Exclusion criteria: not specified
Interventions	Intervention group: MECC (a system with a reduced priming volume, no aortic venting, and no venous reservoir, excluding the blood–air interface. It is a closed system associating a Rotaflow centrifugal pump [Jostra, France] and a Quadrox membrane oxygenator [Jostra, France]) Control group: standard normothermic CPB with a membrane oxygenator with a cardiotomy reservoir with a roller pump (Stoeckert, France) at a nonpulsatile flow of 2.6 L/min/m2
Outcomes	Haematologic parameters and inflammatory cytokines (IL‐1b, IL‐6, TNF‐a, neutrophil elastase, b‐TG, Hct, monocyte count, S100 protein, polymorphonuclear phagocyte counts, Plt count), clinical outcomes (number of grafted vessel, mortality), and myocardial protection (cTnI) Timing: Serial blood samples were taken prior to the onset (T1), after initiation (T2), at the end (T3), and after weaning of the CPB (T4); further samples were collected 6 hours (T5) and 24 hours (T6) after the end of the CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Consecutive patients with similar preoperative and intraoperative characteristics were randomised, but no proper description of randomisation method was reported.
Allocation concealment (selection bias)	Unclear risk	The authors declared a blinded random allocation for this study but gave no further description.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants survived and completed the study, with no trial group changes, no withdrawals, and no losses to follow‐up reported, but with no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were properly assessed.
Other bias	Low risk	No funding was disclosed.

Fujii 2010.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified; received 8 October 2009; received in revised form 17 February 2010; accepted 3 March 2010 Number of study centres and locations: single centre; Division of Cardiovascular Surgery, Department of Surgery, Nippon Medical School Hospital (Tokyo, Japan)
Participants	Twelve participants were treated with either sivelestat 0.2 mg/kg/h (sivelestat group, n = 6) or 0.9% saline (control group, n = 6). Mean age: 70.75 years Sex (female/male ratio): 66.67% No high‐risk population Inclusion criteria: > 20 years of age and scheduled to undergo elective AVR surgery Exclusion criteria: presence of cardiogenic pulmonary oedema, emergency or urgent cases, pre‐existing asthma or COPD, and major chest wall abnormalities
Interventions	Intervention group: sivelestat sodium hydrate, a novel synthesised PMN‐Elastase inhibitor (dissolved in 50 mL of 5% glucose solution and administered intravenously at a rate of 0.2 mg/kg/h for 24 hours as a continuous infusion after tracheal intubation) Control group: 0.9% saline from the start of surgery
Outcomes	Clinical parameters (mortality, major complications, reoperation, time to extubation), pulmonary function (AaDO2, PaO2/FiO2 ratio), IL‐6 and IL‐8, and PMN‐Elastase Timing: after tracheal intubation, 1 hour after CPB introduction, and 3 hours after CPB termination. Data were drawn from plasma and trough bronchoscopic microsampling procedure.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	There were no differences in baseline characteristics. Patients were randomised into two groups using a computer‐generated randomisation table.
Allocation concealment (selection bias)	Unclear risk	The drug administration was blinded for the surgical and anaesthesiologic teams, but there was no further description of how allocation was made.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Surgical and anaesthesiologic teams were blinded. All determinations of cytokines in the alveolar biochemical constituents were carried out by a technician blinded to treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All determinations of cytokines in the alveolar biochemical constituents were carried out by a technician blinded to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants survived and completed the study, with no trial group changes, no withdrawals, and no losses to follow‐up reported but with no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Inflammatory marker data were detected only from the alveolar biochemical constituents. They are not clearly reported in the Results section of the paper.
Other bias	Low risk	No funding was disclosed.

Gabel 2013.

Study characteristics
Methods	Prospective, open, randomised study Run‐in period: not specified; received 2 November 2012; received in revised form 16 December 2012; accepted 27 December 2012 Number of study centres and locations: single centre; Sahlgrenska University Hospital, Gothenburg, Sweden
Participants	Twenty‐five CABG patients were randomised to either cell salvage of cardiotomy suction blood (n = 13) or no cell salvage before retransfusion (n = 12) Mean age: 68 years Sex (female/male ratio): 20% No high‐risk population Inclusion criteria: age 40 years to 80 years, two‐ or three‐vessel coronary disease with angina pectoris and appropriate coronary anatomy for CABG, LVEF > 40%, and no other significant disorders Exclusion criteria: preoperative use of steroids or nonsteroidal anti‐inflammatory drugs
Interventions	Intervention group: cell salvage of cardiotomy suction blood (cardiotomy suction blood was evacuated from the reservoir and processed in a cell‐saver device [autoLog™, Medtronic, Minneapolis, MN, USA] before retransfusion) Control group: no cell salvage before retransfusion (cardiotomy suction blood was collected and retransfused without further processing by opening the hatch between the suction reservoir and the venous reservoir. Ringer’s acetate was added to the processed blood to maintain a retransfusion volume and Hct value comparable with those of the unprocessed group)
Outcomes	Clinical course (stroke, number of transfusion), three anti‐inflammatory cytokines (IL‐1 receptor antagonist, IL‐4, and IL‐10) and two proinflammatory cytokines (TNF‐α and IL‐6), the IL‐6:IL‐10 ratio, and Hct Timing: at baseline (before surgery); during CPB immediately before retransfusion of cardiotomy suction blood; before weaning (after retransfusion); 10 minutes after CPB; 2 hours after CPB; and 24 hours after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There were no statistically significant differences between the groups with respect to baseline characteristics, but there was no further randomisation method description.
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants survived and completed the study, with no trial group changes, no withdrawals, and no losses to follow‐up reported but no ITT analysis. Data from all participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	All outcomes were properly assessed.
Other bias	Low risk	This work was supported by the Swedish Heart and Lung Foundation, Sahlgrenska University Hospital, and Gothenburg Medical Society. The authors declared no conflict of interest.

Garau 2019.

Study characteristics
Methods	Prospective, interventional, single‐centre, randomised controlled pilot trial Run‐in period: September 2013 to June 2015; article first published online 22 November 2018; manuscript accepted 5 November 2018; manuscript revised: 23 October 2018; manuscript received 21 March 2018 Number of study centres and locations: single centre; University Heart Center, Hamburg‐Eppendorf, Hamburg, Germany
Participants	Forty‐three participants were recruited and randomised into two groups: one with haemadsorption on CPB (study group, N = 21) and one without (control group, N = 22). One participant dropped out of the study group, and two participants dropped out of the control group. Finally, analysis of primary and secondary outcomes included 20 participants for each group. Mean age: 70.3 years Sex (female/male ratio): not specified No high‐risk population Inclusion criteria: patients scheduled for CABG, AVR, or a combined procedure, with an expected CPB time of > 120 minutes Exclusion criteria: an age of < 18 years, a BMI of < 18, pregnancy, AF, use of immunosuppressive medication, leukopenia, emergency and urgent surgery, re‐thoracotomy, serum Cr of > 2 mg/dL, transplant surgery, and refusal of written informed consent
Interventions	Intervention group: haemadsorption (Cytosorb®, CytoSorbents Europe GmbH, Berlin, Germany) during CPB. The system was installed into a parallel arm of the extracorporeal circuit, and a blood flow of 300 mL/min was maintained over the CPB period. Control group: no haemadsorption
Outcomes	Primary outcomes: cytokine measurements (IL‐8, IL‐6, and TNF‐α) and procalcitonin Secondary outcomes: cardiovascular function and haemodynamic assessment (cardiac index, extravascular lung water, global end diastolic volume, and SVR index), fluid and catecholamine requirements, duration of surgery, CPB time, time in ICU, and hospital discharge time Timing (from the arterial line): before (M1) and at the end (M2) of CPB and 6 hours (M3) and 24 hours after CPB (M4)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation was performed on the day of the surgery, but there was no further randomisation method description. Demographic data and surgical data did not differ between the two groups.
Allocation concealment (selection bias)	Low risk	A sealed envelope was drawn to determine which group the patient would be assigned to.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Twenty‐two participants were randomised into the control group without the haemadsorption device, and 21 participants were allocated to the study group, in which haemadsorption was performed during CPB. ITT analysis: two participants were excluded in the control group because of extension of surgery, and another participant was excluded in the study group because of immediate reoperation. Based on an anticipated difference in IL‐6 concentration of 25% between the study group and the control group, an α error of 0.05, and a power of 0.9, a group size of 18 patients would be necessary to detect a significant difference. For dropout compensation, a sample size of 40 patients was finally chosen and analysed for analysis. The protocol was approved by the local government ethics committee (Ethics Committee Hamburg Medical Board, protocol number PV4420, date of approval 8 July 2013).
Selective reporting (reporting bias)	Unclear risk	Linear mixed models were used to analyse the effect of haemadsorption in comparison to control on changes from baseline to the follow‐up time points. In the case of an insignificant interaction term, only the main effects group and follow‐up time points were included. This decision was met by using the likelihood ratio test for model comparison. Moreover, all models were adjusted for the baseline level of the outcome parameter, age, aortic cross‐clamp time, duration of ECC, and duration of surgery.
Other bias	High risk	The authors certify that there is no conflict of interest with any financial organisation regarding the material discussed in the manuscript, but this work was supported by an unrestricted grant from CytoSorbents Europe GmbH.

Garcia‐Camacho 2020.

Study characteristics
Methods	Single‐centre, randomised controlled trial Run‐in period: September 1st, 2017 to February 28th, 2018. Received: April 17, 2020; Accepted: October 30, 2020; Published: November 23, 2020 Number of study centres and locations: single centre; Puerta del Mar University Hospital, Andalusian Health Service, Cadiz, Andalusia, Spain
Participants	Fifty‐four adults undergoing ECC for normothermic surgical procedures and patients who had a minimum time before decannulation of > 60 minutes (myocardial reperfusion completed, unclamped aorta, and ECC completed) Exclusion criteria: previous renal or hepatic failure and procedures without ECC
Interventions	Intervention: continuous high‐volume haemofiltration with volume replacement by the use of a polyethersulfone membrane during ECC Control: no filter
Outcomes	Primary outcome: lactate levels before commencement of CBP Secondary outcomes: maximum lactate level intraoperatively; lactate levels at 1 minute after CBP initiation; potassium levels; EuroSCORE; Hct, ICU length of stay, time of intubation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Generated manually by the lead investigator in eight permuted blocks without stratification; no clear indication if he was blinded to participant identity
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Operating theatre staff were not blinded, but participants, ICU staff, and outcome assessors were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors blinded to allocation
Incomplete outcome data (attrition bias) All outcomes	High risk	Data not analysed according to ITT: some patients were excluded from the analysis even though they underwent surgery with or without the intervention and were all excluded after experiencing complications (rebleeding, death, mediastinitis), which are potentially severe and therefore should have been included. The authors' report on the PRISMA diagram is misleading when they claim that zero participants were lost to follow‐up, and they make it appear as if participants excluded from analysis never underwent the intervention.
Selective reporting (reporting bias)	High risk	There is a protocol attached to the paper, but it does not appear to have been registered anywhere; thus, it is impossible to prove it was written a priori. Secondly, the authors report multiple other outcomes not specified in the initial protocol.
Other bias	Low risk	The authors received no specific funding for this work, and the authors have declared that no competing interests exist.

Gasz 2004.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified; published 2004 Number of study centres and locations: single centre; Department of Cardiac Surgery, Zala Country Hospital, Hungary
Participants	Twenty patients undergoing CABG with (CPB group: 10 patients) or without (OPCABG group: 10 patients) pump Mean age: 63.25 years Sex (female/male ratio): 75% No high‐risk population Inclusion criteria: not specified Exclusion criteria: recent MI (< 3 months); acute operation; reoperation; infection; immunological disease; tumour; acute or chronic renal failure; respiratory impairment; previous stroke; and coagulopathy
Interventions	Intervention group: OPCAB Control group: on‐pump CPB
Outcomes	TNF‐α, adhesion molecules of leucocyte (CD11, CD18), WBC and neutrophil PMN count, myocardial injury (cTnI), and clinical data (mortality, pulmonary, and neurological complication; postoperative blood loss) Timing: three times during the operation and on PODs 1, 2, 3, and 7
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups), with no trial group changes, no withdrawals, and no losses to follow‐up reported but with no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Because marked haemodilution was present in the CPB group, a correction needed to be used to determine the final value of WBCs, neutrophil count, and TNF‐α.
Other bias	Unclear risk	This study was supported by OTKA T038035 and OTKA T34810.

Gianetti 2004.

Study characteristics
Methods	Prospective, randomised, nonblinded study The study took place between July 2000 and April 2002. Number of study centres and locations: single centre; Pasquinucci Hospital, Massa, Italy
Participants	Twenty‐nine consecutive participants undergoing aortic valve replacement combined with aortocoronary bypass were randomly allocated to either 20 ppm of inhaled nitric oxide (n = 14) or no additional inhalatory treatment (n = 15). Mean age: 69.48 years Sex (female/male ratio): 58.62% No high‐risk evaluation Inclusion criteria: nonemergency AVR combined with aortocoronary bypass Exclusion criteria: an active infection, EF < 30%, or malignancy; a history of haematologic, hepatic, or renal disorders; corticosteroid or a nonsteroidal anti‐inflammatory treatment within the previous 7 days or during surgery; or a postoperative treatment with nitrates or sodium nitroprussiate
Interventions	Intervention group: 20 ppm of inhaled NO administered for 8 hours during and after CPB. In particular, NO delivery was started immediately after patient intubation in the operating theatre via the ventilatory circuit, then continued throughout the CPB directly into the pump prime and in the ICU via the ventilator circuits until suspension of mechanical ventilation NO gas was mixed with nitrogen (500 ppm). The mixture of gases was continuously introduced into the patient’s ventilatory circuit and was connected to a time‐cycled anaesthesia machine ventilator (Servo 900D; Siemens‐Elema, Stockholm, Sweden). The cardioplegic solution circuit was designed to administer oxygenated hyperkalaemic blood enriched with NO. Control group: no additional inhalatory treatment
Outcomes	Myocardial injury (total CK, CK‐MB fraction, and TnI), left ventricular dysfunction (BNP), endothelial cell activation (P‐selectin) Timing: before surgery and at 4 hours, 12 hours, 24 hours, and 48 hours postsurgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Consecutive patients were randomised, but no further randomisation method description was reported. Demographic variables, chemistry laboratory values, and surgical parameters did not significantly differ between the two patient groups.
Allocation concealment (selection bias)	Unclear risk	No information around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups), with no trial group changes, no withdrawals, and no losses to follow‐up reported but with no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Giomarelli 2003.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled clinical study Study period: between January 2002 and April 2002; accepted for publication 24 January 2003 Number of study centres and locations: single‐centre, Institute of Thoracic and Cardiovascular Surgery, University of Siena, Siena, Italy
Participants	Twenty patients undergoing CABG received either preoperative steroids (n = 10, protocol group) or no steroids (n = 10, control group). Mean age: 64 years Sex (female/male ratio): 35% No high‐risk population Inclusion criteria: not specified Exclusion criteria: urgency/emergency surgery; previous heart surgery; valve or combined CABG and valve surgery; LVEF < 0.35; diabetics on insulin therapy; active gastropathic disorder; COPD on therapy; preoperative use of steroids and contraindications to steroid administration; Cleveland Clinic score of 4 or higher
Interventions	Intervention group: steroid (1 g IV MP preoperatively and 125 mg at the end of CPB; in the ICU, 4 additional 125 mg doses were given every 6 hours) Control group: no steroids (similar volumes of isotonic sodium chloride solution at the same time)
Outcomes	TNF‐α, IL‐6, IL‐8, IL‐10, haemodynamic and pulmonary measurements (AaDo2, RI, shunt [Qs/Qt], dead space [Vd/Vt], a‐vo2D, OER, and oxygen delivery), heart protection (MAP, SVR, pulmonary artery pressure, PVR, CVP, cardiac index, CK and CK‐MB), and clinical outcomes (blood glucose level, extubation time, ICU or hospital LOS, postoperative major complication as adverse events, infection) Samples were measured at the following intervals: T1, 15 minutes after intubation; T2, 5 minutes after aortic cross‐clamp release; T3, 10 minutes after CPB; T4, 3 hours after CPB; T5, 12 hours after CPB; T6, 24 hours after CPB; and T7, 4 days after operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised according to a computer‐generated sequence. Participants were similar with regard to preoperative and intraoperative data.
Allocation concealment (selection bias)	Low risk	An anaesthesia nurse performed the randomisation and prepared the syringes of blinded solution that were administered by the anaesthesiologist managing the case.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: all caregivers were blinded to treatment group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All physicians and nursing staff caring for the participants perioperatively were unaware of the treatment groups.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Clinical data were not extensively reported as the other laboratory outcomes.
Other bias	Low risk	No funding was disclosed, and no other source for risk of bias was identified.

Glumac 2017.

Study characteristics
Methods	Randomised, double‐blind, placebo‐controlled, parallel‐arm trial Study period: March 2015 to January 2016. Published online: 18 May 2017 Number of study centres and locations: single university teaching hospital, the University Hospital of Split, Croatia
Participants	A total of 169 patients scheduled for elective cardiac surgery were enrolled, and 161 participants were included in the analyses. Participants were randomised to receive dexamethasone (n = 80) or placebo (n = 81). Mean age: 63.95 years Sex (female/male ratio): 21.12% No high‐risk population Inclusion criteria: age between 41 years and 84 years, scheduled for elective CABG surgery, heart valve surgery, or combined surgery (CABG and valve surgery) with or without CPB Exclusion criteria: any cerebrovascular incident in the last 3 years; mental illness; visual, hearing, or motor impairment interfering with cognitive assessment; previous cardiac or carotid surgery; LVEF < 35%; adrenal gland disease requiring steroid treatment for longer than 7 days in the past year; alcohol (> 50 g/d or > 500 g/week) or controlled substance abuse; a preoperative MMSE score < 26 points; a preoperative BDI‐II score > 19 points; a preoperative CRP level > 5 mg/L; perioperative stroke; and additional corticosteroid treatment throughout the study period
Interventions	Intervention group: dexamethasone (a single IV bolus of 0.1 mg /kg) Control group: placebo (0.9% NaCl) Timing: 10 hours before the surgery
Outcomes	Primary outcome: POCD on the sixth day after surgery Secondary outcomes: SIRS, PCR, S100, and clinical outcomes (mortality, stroke) Measurements were collected at PODs 1, 2, and 3.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The pharmacy of the University Hospital of Split prepared the trial medication in computer‐randomised blocks of five indistinguishable, sequentially numbered vials" A random sequence generator was used to determine patient allocation.
Allocation concealment (selection bias)	Low risk	The injections were performed by the attending anaesthesiologist. The pharmacy of the University Hospital of Split prepared the trial medication in computer‐randomised blocks of five indistinguishable, sequentially numbered vials containing either a clear solution of dexamethasone or placebo.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The participants, their treating physicians, the biochemists, and the investigators were blind to the treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The participants, their treating physicians, the biochemists, and the investigators were blind to the treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	One hundred seventy‐one participants provided consent for this trial. Two participants were excluded because they were unable to complete baseline cognitive test. One hundred sixty‐nine participants completed baseline assessment, and 161 participants were included in the final analyses. Intention‐to‐treat analysis and causes were reported.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly. TRIAL REGISTRATION Clinicaltrials.gov identifier: NCT02767713
Other bias	Low risk	This study was supported by the Clinical Department of Anaesthesiology and Intensive Care, University Hospital of Split, Split, Croatia. The authors declared no conflicts of interest. No other source of bias was identified.

Gorjipour 2022.

Study characteristics
Methods	Single‐centre randomised control trial Study period: 2018. Published 2022 Number of centres and location: Imam Hossein educational hospital, Shahid Beheshti University of Medical Sciences (SBUMS) in Tehran, Iran
Participants	Forty‐three participants: RIPC group (21 participants), control group (22 participants) Inclusion criteria: adults undergoing first‐time CABG Exclusion criteria: EF < 45%, recent inflammatory disease, abnormal upper limb examination, diabetes mellitus, re‐do cardiac surgery, renal disease, and pre‐existing conditions known to affect the systemic inflammatory response to CPB
Interventions	Intervention group: RIPC (four 5‐minute cycles of upper limb ischaemia, induced by a blood pressure cuff inflated to 15 mmHg to 20 mmHg above systolic blood pressure, with an intervening 5 minutes of reperfusion by deflating the cuff) Control group: no RIPC
Outcomes	Levels of inflammatory cytokines (IFNγ, IL‐4, IL‐8, and IL‐10) and troponin after anaesthesia induction (pre‐skin incision), after CPB, and 24 hours after ICU admission
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"For randomization, each of patients randomly selected one group code; R as RIPC group and C as control group" Unclear how random sequence was generated; authors state that patient chose a code at random but unclear if the codes were hidden from patient before choice
Allocation concealment (selection bias)	Unclear risk	Not described; unlikely because of the nature of the intervention
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described; in view of the nature of the intervention, it would be difficult to blind participant or the personnel administering the preconditioning.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described, insufficient information
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No PRISMA diagram is available, and the authors do not disclose how many patients were screened/how many met inclusion criteria, etc.
Selective reporting (reporting bias)	High risk	No prespecified protocol is registered or published. In the methods, the only outcomes prespecified are the levels of inflammatory cytokines and troponin, but the authors also report additional outcomes.
Other bias	Low risk	No funding received; no other source of bias identified

Gorki 2016 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised trial Study date: 2016 Number of study centres and locations: single centre; Department of Cardiothoracic and Vascular Surgery, University of Ulm, Ulm, Germany
Participants	One‐hundred twenty CABG patients were prospectively randomised for three surgical techniques: MECC, OPCAB, and ONCAB. Forty participants were allocated for each investigated surgical technique. One hundred twelve participants were assessed for final analysis. Mean age: 68.33 years Sex (female/male ratio): 11.67% Low‐risk patients Inclusion criteria: first‐time operation of isolated CAD with at least three target vessels, LVEF > 40% and age between 18 years and 85 years, preoperative anti‐Plt medication Exclusion criteria: emergency patients and patients with coagulation disorders, increased bleeding risk, or findings precluding one of the three investigated surgical techniques (e.g. porcelain aorta, diffuse calcified target vessels)
Interventions	Intervention group: MECC. MECC used a closed system with a centrifugal pump (Rotaflow, Maquet), heparin‐coated tubing, a membrane oxygenator with albumin/heparin coating and about 1.8 m2 surface (Quadrox, Maquet), and a cell saver (BT844, Sorin) for pericardial fluids with vacuum‐assisted suction. Control group: ONCAB used an open system with a roller pump (Stöckert S5, Sorin, Munich, Germany) and vacuum‐assisted venous return if necessary, uncoated tubing, a membrane oxygenator with amphiphilic bio‐passive coating and about 2.5 m2 surface (Capoix FX25, Terumo, Eschborn, Germany), a reservoir with defoamer and filter, and recirculation of PS via an extra cardiotomy reservoir (Maquet, Rastatt, Germany).
Outcomes	Primary outcomes: initial and cumulative heparin, protamine, protamine/cumulative heparin ratio, and blood loss Secondary outcomes: clinical data (mortality, MI, stroke, repeat revascularisation), coagulation (coagulation factors I, II, V, VIII, X, antithrombin III, TAT, D‐dimers, tissue factor pathway inhibitor, leucocyte and Plt counts, P selectine, PF‐1.2, INR, aPTT), and inflammatory markers (TNF‐α, sICAM, CRP) Blood samples were obtained from a peripheral arterial access (preferred) or the central venous line at induction (time point 'A'), before administration of protamine ('B'), at arrival in the ICU ('C'), and at 6 hours ('D'), 12 hours ('E'), 24 hours ('F'), and 72 hours ('G') postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Few risk factors (hyperlipidaemia P = 0.039, hypertension P = 0.015) and medication (statins P = 0.035) differ significantly between groups. No randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No allocation information is given. All procedures were performed by the same surgeon. Concealment is unlikely because of the nature of the intervention.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unlikely because of the nature of the intervention, but might not have influenced the study procedures significantly
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified; insufficient information
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis: 266 patients were assessed for eligibility, 146 patients were excluded, and 120 patients were randomised in 3 different groups. Preoperative and postoperative data referred to 120 participants (intention to treat) and study‐relevant intraoperative and postoperative data referred to 112 participants (per protocol, reasons for exclusion of 8 participants were documented).
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly. The study is registered at the German Clinical Trial Registry: registration number DRKS00007580.
Other bias	High risk	The authors declared no potential conflicts of interest with respect to the research, authorship, or publication of this article. The study was, however, supported financially by the companies Medtronic and Maquet.

Gorki 2016 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised trial Study date: 2016 Number of study centres and locations: single centre; Department of Cardiothoracic and Vascular Surgery, University of Ulm, Ulm, Germany
Participants	One hundred twenty CABG patients were prospectively randomised for three surgical techniques: MECC, OPCAB, and ONCAB. Forty participants were allocated for each investigated surgical technique. One hundred twelve participants were assessed for final analysis. Mean age: 68.67 years Sex (female/male ratio): 15% Low‐risk patients Inclusion criteria: first‐time operation of isolated CAD with at least three target vessels, LVEF > 40% and age between 18 years and 85 years, preoperative anti‐Plt medication Exclusion criteria: emergency patients and patients with coagulation disorders, increased bleeding risk, or findings precluding one of the three investigated surgical techniques (e.g. porcelain aorta, diffuse calcified target vessels)
Interventions	Intervention group: OPCAB. OPCAB used deep pericardial stitches and, occasionally, an apical suction device (Urchin, Medtronic, Minneapolis, MN, USA) for exposure of the left lateral and posterior targets of the beating heart, vacuum suction for local stabilisation (Octopus IV, Medtronic), intraluminal shunts, and a cell saver. Control group: ONCAB used an open system with a roller pump (Stöckert S5, Sorin, Munich, Germany) and vacuum‐assisted venous return if necessary, uncoated tubing, a membrane oxygenator with amphiphilic bio‐passive coating and about 2.5 m2 surface (Capoix FX25, Terumo, Eschborn, Germany), a reservoir with defoamer and filter, and recirculation of PS via an extra cardiotomy reservoir (Maquet, Rastatt, Germany).
Outcomes	Primary outcomes: initial and cumulative heparin, protamine, protamine/cumulative heparin ratio, and blood loss Secondary outcomes: clinical data (mortality, MI, stroke, repeat revascularisation), coagulation (coagulation factors I, II, V, VIII, X, antithrombin III, TAT, D‐dimers, tissue factor pathway inhibitor, leucocyte and Plt counts, P selectin, PF‐1.2, INR, aPTT), and inflammatory markers (TNF‐α, sICAM, CRP) Blood samples were obtained from a peripheral arterial access (preferred) or the central venous line at induction (time point 'A'), before administration of protamine ('B'), at arrival in the ICU ('C'), and at 6 hours ('D'), 12 hours ('E'), 24 hours ('F'), and 72 hours ('G') postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given. Several baseline characteristics and risk factors (hyperlipidaemia P = 0.039, hypertension P = 0.015) and medication (statins P = 0.035) differ significantly between groups.
Allocation concealment (selection bias)	Unclear risk	No allocation information is available. All procedures were performed by the same surgeon.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; insufficient information; unlikely to have introduced bias
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Intention‐to‐treat analysis: 266 patients were assessed for eligibility, 146 patients were excluded, and 120 patients were randomised into 3 different groups. Preoperative and postoperative data referred to 120 participants (intention to treat), and study‐relevant intraoperative and postoperative data referred to 112 participants (per protocol, reasons for exclusion of 8 participants were documented).
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly. The study is registered at the German Clinical Trial Registry: registration number DRKS00007580.
Other bias	High risk	The authors declared no potential conflicts of interest with respect to the research, authorship, or publication of this article. The study was supported financially by the companies Medtronic and Maquet.

Gorki 2017.

Study characteristics
Methods	Prospective, randomised study Run‐in period: participants were enrolled from July 2015 to January 2016. Received: 31 July 2018 Accepted after revision: 21 December 2018 Number of study centres and locations: single centre, Department of Cardiothoracic and Vascular Surgery, University of Ulm, Ulm, Germany
Participants	Forty‐eight participants operated on with CCPB for myocardial revascularisation were randomised either for direct recirculation of PS fluids (merged group 1, PS n = 24) or for cell saving (merged group 2, n = 24). Forty‐seven participants were assessed in the final analysis but only 37 participants for the inflammatory marker assessment. Mean age: 67.35 years Sex (female/male ratio): 12.5% Low‐risk patients Inclusion criteria: first‐time operation of isolated CAD with at least three target vessels, LVEF > 40% and age between 18 years and 85 years, preoperative anti‐Plt medication Exclusion criteria: emergency patients and patients with coagulation disorders, increased bleeding risk, or findings precluding one of the three investigated surgical techniques (e.g. porcelain aorta, diffuse calcified target vessels)
Interventions	Intervention group: pericardial blood was recirculated via pericardiotomy suction. Control group: shed pericardial fluids were collected separately and, if exceeding 500 mL, retransfused after Cell Saver as autologous RBC concentrates.
Outcomes	Primary outcomes: TAT and D‐dimers, the applied doses of heparin, protamine, and the heparin/protamine ratio Secondary outcomes: inflammation markers (serum levels of leucocytes, TNF‐α, sICAM‐1, and soluble sP‐selectin), Hct, and leucocyte counts Blood samples were collected from a peripheral arterial catheter at induction of anaesthesia, before protamine administration, at ICU arrival, and 6 hours, 12 hours, and 24 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The software package R (R Foundation for Statistical Computing, Vienna, Austria) was used to generate random allocation sequences. The groups were comparable in most relevant data.
Allocation concealment (selection bias)	Low risk	Sealed and sequentially numbered envelopes containing the group assignments were provided by the Institute of Epidemiology and Medical Biometry, Ulm.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding was reported. Participants were operated on by the same surgeon.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Demographic parameters were evaluated in an intention‐to‐treat analysis: 48 participants were randomised but 47 were assessed in the final analysis. Although the coagulation markers and leucocytes were investigated in 47 participants, blood samples of only 37 participants (19 PS group and 18 CS group) were available for investigation of inflammation (TNF‐α, sICAM‐1, and sP‐selectin). No explanation as to why this reduction was given. The study was registered at the German Clinical Trials Register (registration number DRKS00008832).
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors have no disclosure or conflict of interest. No other source of bias is identified.

Goudeau 2007.

Study characteristics
Methods	Prospective, randomised study Received for publication: 13 July 2006 Accepted: 17 October 2006 Number of study centres and locations: single centre, Anaesthesia and Intensive Care Unit, Bocage Hospital, Dijon, France
Participants	Twenty‐seven participants were included in the study: control patients (n = 14) underwent conventional CPB, and treated patients (n = 13) underwent a CPB with Baxter Duraflo II heparin‐coated circuits, high doses of aprotinin, and pre‐CPB haemofiltration. Mean age: 76.2 years Sex (female/male ratio): 48.15% High‐risk patients Inclusion criteria: undergoing coronary artery revascularisation or valvular replacement (alone or combined) under CPB Exclusion criteria: surgical emergencies, heparin or aprotinin allergy, intercurrent inflammatory chronic pathologies, coagulation troubles (except heparin or aspirin treatment), and aneurysmectomy‐associated surgery
Interventions	Intervention group: treated patients underwent a CPB with Baxter Duraflo II heparin‐coated circuits (Baxter Healthcare Corp, Bentley Laboratories Division, Irvine, California), plus high doses of aprotinin (complete Royston dose) and pre‐CPB haemofiltration (preoperative low‐volume ultrafiltration [10 mL/kg/min]). Control group: underwent CCPB
Outcomes	Primary outcomes: inflammation (CRP, IL‐6, and C3a, C4, B factor) and oxidative stress status (free radical) Secondary outcomes: clinical criteria (inotropic drugs, blood transfusions, blood losses, postoperative complications such as death, arrhythmia, blood loss, infection, and organ failure and ICU LOS), haemodynamic criteria (heart rate, mean pulmonary artery pressure, pulmonary capillary wedge pressure, cardiac index, and SVR), and biological criteria (blood formula, Plt count, renal function, blood gases, pH, lactic acid, CK‐MB, CTnI) Markers were measured before, during, and to the second POD. Clinical complications were detected for the first postoperative month.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised to one of the two study groups, based on a table of random numbers. There were no significant initial differences between the two groups in preoperative participants' characteristics.
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No clear description of blinding. The same surgeon operated on all the participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This work was supported in part by CHU Dijon, the Region Bourgogne, Faculté de Médecine de Dijon, Fondation de France, and Association de Cardiologie de Bourgogne.

Greilich 2001.

Study characteristics
Methods	Randomised, double‐blind, placebo‐controlled study Run‐in period: not specified Study date: 2001 Number of study centres and locations: Number of centres not specified. Dallas Veterans Affairs Medical Center – the Department of Anesthesiology and Pain Management, the Department of Physiology, and the Department of Cardiovascular and Thoracic Surgery, University of Texas Southwestern Medical Center, Dallas
Participants	Seventy‐two participants undergoing CABG were randomly chosen to receive high‐dose aprotinin (n = 24), EACA (n = 23), or saline placebo (n = 25). Mean age: 62.98 years Sex (female/male ratio): not specified No high‐risk population Inclusion criteria: elective, primary CAB with ECC. Participants were not excluded from the study if they were receiving salicylates, nonsteroidal anti‐inflammatory drugs, or heparin before surgery. Exclusion criteria: corticosteroids, dipyridamole or anticoagulants, Plt or coagulation abnormalities, thrombolytic therapy within 5 days of surgery, Cr level > 2.0 mg/dL, EF < 30%, and a history of adverse reaction to aprotinin
Interventions	Intervention group: full‐dose aprotinin (2 × 106 KIU [load], 2 × 106 KIU [pump prime], 5 × 105 KIU/h [infusion]) Control group: saline (200 mL [load], 200 mL [pump prime], 50 mL/h [infusion])
Outcomes	IL‐10, IL‐6, and clinical outcomes (heart rate; mean arterial blood pressure; peripheral arterial saturation; CVP; mean pulmonary arterial pressure; pulmonary capillary wedge pressure; cardiac output; Hb, Hct, and Plt counts; fibrinogen concentration; WBC counts; mediastinal chest tube drainage; blood product administration; duration of postoperative mechanical ventilation; duration of inotropic support; and the length of the ICU stay) Blood samples were measured at five time points: baseline (before induction); 10 minutes of ECC; 30 minutes after ECC (after heparin reversal); 3 hours after ECC; and 18 hours after ECC.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There were no significant differences between the participants, but no further randomisation method description.
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Both clinical and inflammatory markers were not properly assessed (IL‐6 data not reported).
Other bias	Low risk	This study was funded by the Department of Veterans Affairs and by Research Starter Grants from the Society of Cardiovascular Anesthesiologists and Bayer Corporation.

Greilich 2003.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled study Received for publication: 17 January 2003 Revisions requested: 25 March 2003 Revisions received: 17 April 2003 Accepted for publication: 24 April 2003 Number of study centres and locations: Number of centres not specified. Cardiovascular and Thoracic Surgery, University of Texas Southwestern Medical Center, Dallas Veterans Affairs Medical Center, Dallas, TX
Participants	Sixty participants were randomised to receive EACA, aprotinin, or saline (placebo). Mean age: 63.5 years Sex (female/male ratio): 0% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: corticosteroid, dipyridamole, or anticoagulant therapy; documented Plt or coagulation abnormalities or treatment with thrombolytic therapy within 5 days of surgery; serum Cr > 2.0 mg/dL; EF<30%; or a history of adverse reaction to aprotinin
Interventions	Intervention group: aprotinin (broad‐spectrum serine protease inhibitor – 2 × 106 KIU [loading dose], 2 × 106 KIU [pump prime], and 5 × 105 KIU/h [infusion]) Control group: saline (200 mL [load], 200 mL [pump prime], and 50 mL/h [infusion]) Timing: in similar dosing regimens (loading dose, pump prime, and infusion)
Outcomes	Primary outcomes: IL‐6 and IL‐8, D‐dimer Secondary outcomes: duration of aortic clamping and CPB, the number and type of bypass grafts, mediastinal chest tube drainage, blood product administration, duration of postoperative mechanical ventilation, and inotropic support Arterial blood samples were collected at the following times: baseline (before anaesthetic induction); 10 minutes after aortic cross‐clamp removal; 15 minutes after protamine administration (heparin reversal); 3 hours after separation from CPB; and 18 to 24 hours after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There were no significant differences between the participants, but there was no randomisation method information: "A total of 60 patients were randomized to receive aprotinin, ϵ‐aminocaproic acid, or saline".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: "patients were randomized in a double‐blind fashion"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This study was supported in part by grants from the Department of Veterans Affairs and Society of Cardiovascular Anesthesiologists.

Greilich 2004.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled trial Accepted for publication: 11 July 2003 Number of study centres and location: single centre, University of Texas Southwestern Medical Center, Dallas, Texas
Participants	Thirty‐six patients scheduled to undergo cardiac surgery with CPB were randomised to receive EACA, aprotinin, or saline (placebo). Data were analysed from a total of 35 patients randomised to EACA (n = 11), aprotinin (n = 12), or saline (n = 12). Mean age: 63.5 years Sex (female/male ratio): 0% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: corticosteroids, anticoagulant therapy, IV tissue plasminogen activators or streptokinase (within 5 days of surgery), documented Plt or coagulation abnormalities and use of dipyridamole or any other anti‐Plt therapy other than aspirin (within 2 weeks of surgery), serum Cr concentration > 2.0 mg/dL, EF < 30%, or had a history of adverse reaction to aprotinin
Interventions	Intervention group: aprotinin (2 × 106 KIU loading dose, 5 × 105 KIU/h infusion rate, and 2 × 106 KIU in the pump prime) Control group: saline (placebo)
Outcomes	Markers of plasmin activity (D‐dimer concentrations), Plt activation (CD62P, P selectine), leucocyte activation (CD11b) and leucocyte–Plt adhesion (monocyte– and neutrophil–Plt conjugates), blood cell counts, duration of aortic clamping and CPB, number and type of bypass grafts, mediastinal chest tube drainage, and blood product administration Radial artery blood samples were collected at the following times: at baseline (before induction); at the end of hypothermic CPB (before warming); after the termination of CPB (after heparin reversal by protamine); and 24 hours after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Significant differences in participant demographic and surgical data between treatment groups were found for EF, transfusions, and mediastinal chest tube drainage. No randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: "prospective, randomized, double‐blind, placebo‐controlled trial"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	One participant randomised to the EACA group was excluded from the analysis. In a second participant, who was randomised to aprotinin, reinstitution of CPB was required because of the combination of a failed mitral valve repair and major surgical haemorrhage. This participant was subsequently reheparinised and underwent mitral valve replacement without incident. All post‐CPB data were collected after the second bypass run and included in the analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	Supported by a New Investigator Award from the Department of Veterans Affairs, Washington, D.C. (to Dr. Greilich) and grant no. HL47193 from the National Institutes of Health, Bethesda, Maryland (to Drs. C. Rinder and B. Smith)

Grunenfelder 2000 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: between July 1997 and May 1999 Received: 7 September 1999 Received in revised form: 19 November 1999 Accepted: 29 November 1999 Number of study centres and locations: single centre, University Hospital, Rämistrasse 100, 8091 Zürich, Switzerland
Participants	Ninety‐seven participants undergoing elective CABG. Fifty patients were operated on using normothermic and 47 patients using hypothermic CPB. The normothermic group was subdivided into a group with MUF (n = 30) and a group without MUF (n = 20). In the hypothermic group, 30 participants received MUF compared with 17 participants serving as controls. Mean age: 62.1 years Sex (female/male ratio): not specified No high‐risk patients Inclusion criteria: elective CABG Exclusion criteria: diabetes, preoperative ischaemia, or previous CAB surgery
Interventions	Intervention: normothermic CPB with MUF using a Gambro FH22 filter with an effective membrane area of 0.2 m2 and a pore size of 5 nm (Gambro Dialysatoren GmbH and Co. KG, Hechingen, Germany) Timing: MUF was instituted after CPB for 15 minutes through the arterial and venous bypass circuit lines. Control group: normothermic CPB without MUF
Outcomes	Cytokines (IL‐6, IL‐8, TNF‐α, IL‐2R) and adhesion molecules (sE‐selectin, sICAM‐1) and clinical parameters (intubation time [hours], blood loss in 24 hours [mL], given transfusion [mL], PAaO2, inotropic requirement, pulmonary infection requiring antibiotic treatment, AF, body temperature [max, min], ICU stay [days], leucocyte count and CRP for 6 days [min, max], length of hospital stay [days]) Blood samples were measured preoperatively, pre‐MUF, in the ultrafiltrate, and 24 hours, 48 hours, and 6 days after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "prospective randomized study"
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unlikely for personnel because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Grunenfelder 2000 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: between July 1997 and May 1999 Received: 7 September 1999 Received in revised form: 19 November 1999 Accepted: 29 November 1999 Number of study centres and locations: single centre, University Hospital, Rämistrasse 100, 8091 Zürich, Switzerland
Participants	Ninety‐seven participants undergoing elective CABG. Fifty patients were operated on using normothermic and 47 patients using hypothermic CPB. The normothermic group was subdivided into a group with MUF (n = 30) and a group without MUF (n = 20). In the hypothermic group, 30 participants received MUF compared with 17 participants serving as controls. Mean age: 58.7 years Sex (female/male ratio): not specified No high‐risk patients Inclusion criteria: elective CABG Exclusion criteria: diabetes, preoperative ischaemia, or previous CAB surgery
Interventions	Intervention group: hypothermic CPB with MUF using a Gambro FH22 filter with an effective membrane area of 0.2 m2 and a pore size of 5 nm (Gambro Dialysatoren GmbH and Co. KG, Hechingen, Germany) Timing: MUF was instituted after CPB for 15 minutes through the arterial and venous bypass circuit lines. Control group: hypothermic CPB without MUF
Outcomes	Cytokines (IL‐6, IL‐8, TNF‐α, IL‐2R) and adhesion molecules (sE‐selectin, sICAM‐1) and clinical parameters (intubation time [hours], blood loss in 24 hours [mL], given transfusion [mL], PAaO2, inotropic requirement, pulmonary infection requiring antibiotic treatment, AF, body temperature [max, min], ICU stay [days], leucocyte count and CRP for 6 days [min, max], length of hospital stay [days]) Blood samples were measured preoperatively, pre‐MUF, in the ultrafiltrate, and 24 hours, 48 hours, and 6 days after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "prospective randomized study"
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, but unlikely for personnel because of the nature of the intervention
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Gu 1993.

Study characteristics
Methods	Prospective, randomised study Accepted for publication: 20 July 1992 Number of study centres and locations: single centre; Thorax Centre, University Hospital, Groningen, the Netherlands
Participants	Thirty patients undergoing CABG were randomly perfused with a heparin‐coated circuit (Duraflo II, n = 15) or with a similar noncoated circuit (control, n = 15). Mean age: 52.5 years Sex (female/male ratio): not specified No high‐risk patients Inclusion criteria: not specified Exclusion criteria: preoperative heart failure, pulmonary diseases, bleeding diathesis, and an age > 70 years
Interventions	Intervention group: heparin‐coated circuit (all blood‐contacting components in the circuit (including oxygenator, venous reservoir, tubing, and cannulas) except the cardiotomy reservoir were coated with heparin (Duraflo II; Bentley Baxter Inc)) Control group: noncoated circuit
Outcomes	Elastase, TNF‐α, C3a, blood loss, clinical complications (death; number of anastomoses; bypass time; aortic cross‐clamp time; cardiac, pulmonary, or renal complications), cell counts (leucocyte and granulocyte counts), plasma Hb Blood samples were taken after the induction of anaesthesia, 5 minutes before CPB, 5 minutes and 30 minutes after the start of CPB, 5 minutes before and after release of the aortic cross‐clamp, at the end of CPB, and 30 minutes after protamine administration from the indwelling radial arterial catheter.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There was no significant difference between the two groups with regard to the preoperative variables. No randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified, might be unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	High risk	This study was in part supported by Bentley Laboratories Europe, Uden, the Netherlands.

Gu 1996.

Study characteristics
Methods	Prospective, randomised study Accepted for publication: 21 December 1995 Number of study centres and locations: single centre; Departments of Cardiothoracic Surgery University Hospital, Groningen, the Netherlands
Participants	Thirty patients undergoing elective heart operations were randomly allocated to a leucocyte‐depletion group (n = 20) or a control group (n = 10). Mean age: 60.67 years Sex (female/male ratio): 26.67% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: allergy or recurrent infection, reoperation, and emergency operation
Interventions	Intervention group: all residual blood was filtered by leucocyte removal filters and re‐infused after CPB. Control group: an identical amount of residual blood after CPB was re‐infused without filtration.
Outcomes	Leucocyte depletion (leucocyte and granulocyte counts, Plt levels), lung function (partial arterial oxygen tension, PAP, PCWP, PVR), other clinical parameters (duration of postoperative intubation, blood loss, ICU LOS and hospital LOS, infections), laboratory parameters, and inflammatory mediators (thromboxane B2, IL‐6, and IL‐2) Blood samples were taken at the baseline before operation, at the end of CPB before transfusion of the leucocyte‐depleted blood, at the end of operation during skin closure, 1 hour and 3 hours after the patient's arrival in the ICU, and at 6 a.m. the next day in the ICU.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There were no significant differences between the leucocyte‐depletion group and the control group. No randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No allocation information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unlikely for personnel because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	IL‐2 was not detectable in any of the samples (but this could be in agreement with other cardiac surgery studies). The other outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Gu 1998 (A).

Study characteristics
Methods	Prospective study Run‐in period: from June 1995 to June 1996 Accepted for publication: 5 August 1997 Number of study centres and locations: single centre; Department of Cardiothoracic Surgery, University Hospital Groningen, the Netherlands
Participants	Sixty‐two consecutive patients were assigned randomly to two groups: 31 participants underwent minimally invasive CABG, and 31 participants underwent conventional CABG with CPB. In a subgroup of 10 participants in each group, subclinical markers were measured to determine the level of the inflammatory response generated during the operation. Mean age: 60.5 years Sex (female/male ratio): 33.87% No high‐risk patients Inclusion criteria: patients with isolated stenosis of the left anterior descending coronary artery Exclusion criteria: the presence of any associated cardiac disease, such as a left ventricular aneurysm or valvular disease
Interventions	Intervention group: without CPB through a small anterolateral thoracotomy (MICABG, a skin incision approximately 8 cm to 10 cm long was made at the level of the fifth intercostal space, with the medial edge of the incision 3 cm to 4 cm lateral to the LIMA) Control group: CPB‐CABG (a median sternotomy)
Outcomes	Haematologic and biochemical measurements (leucocyte elastase, leucocyte and Plt count, b‐Tg, and complement C3a), clinical parameters (duration of operation, blood loss, ventilatory support, postoperative hospital stay, and postoperative morbidity and mortality) Blood sample was taken during wound closure in the MICABG group and after CPB but before protamine administration in the CPB‐CABG group.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Consecutive patients were enroled, but no further randomisation method description is given: "62 consecutive patients with isolated stenosis of the left anterior descending coronary artery were assigned randomly to two groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; very unlikely for personnel and participants because of the nature of the intervention, although unclear how lack of blinding could influence outcomes
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	In a subgroup of 10 participants in each group, subclinical markers were measured to determine the level of the inflammatory response generated during the operation. Criteria for selecting the participants in the subgroup are unclear.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Gu 1998 (B).

Study characteristics
Methods	Prospective, randomised study Accepted for publication: 26 December 1997 Number of study centres and locations: single centre; Department of Cardiothoracic Surgery, University Hospital Groningen, the Netherlands
Participants	Twenty participants undergoing CABG were randomly divided into two groups: the SMA group (n = 10) and the control group (n = 10) participants. The final analysis did not include two participants with postoperative complications in the intervention group. Mean age: 58.5 years Sex (female/male ratio): 10% No high‐risk patients Inclusion criteria: no evidence of severe heart failure or renal or hepatic dysfunction, had no history of bleeding diathesis, no Plt‐inhibiting drugs within 3 days before the operation Exclusion criteria: not specified
Interventions	Intervention group: SMA (all blood‐contacting surfaces in the CPB circuit coated, including oxygenator, venous reservoir, tubing, connectors, and cannula with SMA‐treated circuit [Cobe Cardiovascular Inc, Arvada, CO]) Control group: participants were perfused with an identical circuit without treatment.
Outcomes	Plt count and b‐Tg, prothrombin activation, Plt glycoprotein IIIa, C3a and terminal complex C5b‐9, C4a, leucocyte activation, elastase, IL‐8, chest tube drainage, blood transfusion, duration of ventilatory support, hospital LOS Blood samples were taken before the start of CPB, 5 minutes and 30 minutes after the start of CPB, 5 minutes after release of the aortic cross‐clamp, and at the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	There was no statistical difference between the two groups, but no further randomisation method description is given: "Twenty patients undergoing coronary artery bypass grafting were randomly divided into two groups".
Allocation concealment (selection bias)	Unclear risk	Untreated circuit was similar to treated ECC, but no relevant allocation description was reported.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No ITT is given, and data from the final analysis do not include two participants in the intervention group with postoperative complications.
Selective reporting (reporting bias)	Unclear risk	Overall acceptable selection of outcomes. Study protocol not available. Data do not include two participants with postoperative complications in the intervention group.
Other bias	Low risk	No funding was disclosed.

Gu 1999.

Study characteristics
Methods	Prospective, randomised study Accepted for publication: 13 July 1998 Number of study centres and location: single centre, Departments of Cardiothoracic Surgery and Anaesthesiology, University Hospital Groningen, Groningen, the Netherlands
Participants	Forty cardiac surgical patients undergoing CPB were allocated randomly to a leucocyte depletion group (n = 20) and a control group (n = 20). Mean age: 65.5 years Sex (female/male ratio): 32.5% No high‐risk patients Inclusion criteria: undergoing CABG, heart valve replacement, or a combined procedure Exclusion criteria: allergy, infection, or preoperative pulmonary dysfunction
Interventions	Intervention group: leucocyte depletion was achieved with 2 filter sets (Pall Duplex filter sets (J1647G; Pall Biomedical, Portsmouth, UK)) located between the venous drainage and the venous reservoir and was driven by a separate roller pump of the heart‐lung machine. Leucocyte filtration was commenced after the start of rewarming but before the release of the aortic cross‐clamp. Control group: no leucocyte depletion filter
Outcomes	Mean Plt and leucocyte removal rate, inflammatory markers (IL‐8, elastase, L‐selectine), and clinical observations (postoperative PaO2 or pulmonary haemodynamics, postoperative intubation time and blood loss, urine output, length of stay in the ICU and in the hospital) Blood samples from participants were taken from the radial arterial catheter before operation, at several points during CPB, at the end of CPB, at the end of operation during skin closure, 1 hour and 3 hours after transfer to the ICU, and at 6 a.m. the next morning in the ICU.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "The patients were allocated randomly to a leukocyte depletion group (n = 20) and a control group (n = 20)"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Levels of inflammatory mediators were determined mainly during the postoperative course in exclusively 10 participants in each group, but unclear criteria of selection for the subgroup.
Selective reporting (reporting bias)	Unclear risk	All expected outcomes were reported properly.
Other bias	High risk	Sponsorship provided by Pall Biomedical UK

Gulielmos 2000 (A).

Study characteristics
Methods	Prospective, randomised trial Run‐in period: not specified Received: 14 December 1999 Received in revised form: 14 July 2000 Accepted: 15 August 2000 Number of study centres and locations: single centre, Cardiovascular Institute, University Hospital Dresden, Fetscherstrasse 76, 01307 Dresden, Germany
Participants	Thirteen participants included in part A of this study Four surgical techniques were compared: group 1, median sternotomy with CPB in 10 patients; group 2, median sternotomy and off‐pump in 10 patients; group 3, mini‐thoracotomy with CPB in 10 patients; group 4, mini‐thoracotomy and off‐pump in 10 patients. Mean age: 63.73 years Sex (female/male ratio): 30% No high‐risk patients Inclusion criteria: single LIMA bypass to the LAD, because of CAD or patients suffering from coronary artery double‐vessel or multivessel disease, with only the LAD being amenable for surgery Exclusion criteria: impaired LVEF (< 30%), with impaired lung and renal function, unstable angina, major calcification of the ascending aorta, and obesity (BMI ≥30)
Interventions	Intervention group: median sternotomy and off‐pump. At surgery over a median sternotomy using stay sutures between the left half of the pericardium and the soft tissue, the heart was rotated to the right, resulting in medialisation of the LAD, thus providing better access to this coronary vessel. Using the 'Octopus' (Octopus; Medtronic, Inc.), the coronary artery was occluded using a snare (4.0 Prolene; Ethicon, Norderstedt, Germany). After preconditioning (20 minutes temporary occlusion of the coronary vessel, reperfusion 20 minutes, and again occlusion for the anastomosis), the anastomosis was performed with the aid of an air blower to keep the operative sites free from blood. Control group: median sternotomy with CPB
Outcomes	IL‐1, IL‐6, and myocardial enzyme (troponin T and CK‐MB; clinical data (major complications, operative time, postoperative ventilation, CPB time [if used], cross‐clamping time, and coronary occlusion time) Venous blood samples were drawn at the following times: at 5 minutes before induction of anaesthesia; on induction of CPB or occlusion of the LAD; at 15 minutes after induction of CPB or occlusion of the LAD; at 1 hour after induction of CPB or occlusion of the LAD; at 2 hours after induction of CPB or occlusion of the LAD; at 6 hours after induction of CPB or occlusion of the LAD; at 12 hours after induction of CPB or occlusion of the LAD; at 24 hours after induction of CPB or occlusion of the LAD; and at 48 hours after induction of CPB or occlusion of the LAD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Between the four groups, there were no differences in terms of age, BMI, or incidence of previous infarction. No randomisation method description. One out of four techniques was prospectively chosen for each patient at random, but no other description about how the allocation was made is given.
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unspecified and unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All procedures were completed as they were planned. No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Clinical data were not properly reported as the other outcomes.
Other bias	Low risk	No funding was disclosed.

Gulielmos 2000 (B).

Study characteristics
Methods	Prospective, randomised trial Run‐in period: not specified Received: 14 December 1999 Received in revised form: 14 July 2000 Accepted: 15 August 2000 Number of study centres and locations: single centre, Cardiovascular Institute, University Hospital Dresden, Fetscherstrasse 76, 01307 Dresden, Germany
Participants	Twelve participants in part B of this study Four surgical techniques were compared: group 1, median sternotomy with CPB in 10 patients; group 2, median sternotomy and off‐pump in 10 patients; group 3, minithoracotomy with CPB in 10 patients; group 4, minithoracotomy and off‐pump in 10 patients. Mean age: 60.8 years Sex (female/male ratio): 30% No high‐risk patients Inclusion criteria: single LIMA bypass to the LAD, because of CAD or patients suffering from coronary artery double‐vessel or multivessel disease, with only the LAD being amenable for surgery Exclusion criteria: impaired LVEF (< 30%), with impaired lung and renal function, unstable angina, major calcification of the ascending aorta, and obesity (BMI ≥30)
Interventions	Intervention group: minithoracotomy with CPB. The LIMA was harvested via a minithoracotomy. The right atrium was cannulated transfemoral percutaneously for venous drainage for CPB. The ascending aorta was cannulated and dissected from the pulmonary trunk on pump. A conventional aortic clamp was used for external cross‐clamping, and cardioplegia was applied via the ascending aorta. During cardioplegic arrest, the relaxed heart was rotated for exposure of the LAD. Anastomosis of the LIMA to the LAD was performed. Control group: median sternotomy with CPB
Outcomes	IL‐1, IL‐6, and myocardial enzyme (troponin T and CK‐MB); clinical data (major complications, operative time, postoperative ventilation, CPB time [if used], cross‐clamping time, and coronary occlusion time) Venous blood samples were drawn at the following times: at 5 minutes before induction of anaesthesia; on induction of CPB or occlusion of the LAD; at 15 minutes after induction of CPB or occlusion of the LAD; at 1 hour after induction of CPB or occlusion of the LAD; at 2 hours after induction of CPB or occlusion of the LAD; at 6 hours after induction of CPB or occlusion of the LAD; at 12 hours after induction of CPB or occlusion of the LAD; at 24 hours after induction of CPB or occlusion of the LAD; and at 48 hours after induction of CPB or occlusion of the LAD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Between the four groups, there were no differences in terms of age, BMI, or incidence of previous infarction. No randomisation method description. One out of four techniques was prospectively chosen for each patient at random, but no other description about how the allocation was made is given.
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unspecified and unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All procedures were completed as they were planned. No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Clinical data were not reported properly as the other outcomes.
Other bias	Low risk	No funding was disclosed.

Gulielmos 2000 (C).

Study characteristics
Methods	Prospective, randomised trial Run‐in period: not specified Received: 14 December 1999 Received in revised form: 14 July 2000 Accepted: 15 August 2000 Number of study centres and locations: single centre, Cardiovascular Institute, University Hospital Dresden, Fetscherstrasse 76, 01307 Dresden, Germany
Participants	A total of 12 participants were assessed in part C of this study. Four surgical techniques were compared: group 1, median sternotomy with CPB in 10 patients; group 2, median sternotomy and off‐pump in 10 patients; group 3, minithoracotomy with CPB in 10 patients; group 4, minithoracotomy and off‐pump in 10 patients. Mean age: 62.08 years Sex (female/male ratio): 30% No high‐risk patients Inclusion criteria: single LIMA bypass to the left LAD, because of CAD or patients suffering from coronary artery double‐vessel or multivessel disease, with only the LAD being amenable for surgery Exclusion criteria: impaired LVEF (< 30%), with impaired lung and renal function, unstable angina, major calcification of the ascending aorta, and obesity (BMI ≥ 30)
Interventions	Intervention group: minithoracotomy and OPCABG Control group: median sternotomy with CPB
Outcomes	IL‐1, IL‐6, and myocardial enzyme (troponin T and CK‐MB); clinical data (major complications, operative time, postoperative ventilation, CPB time [if used], cross‐clamping time, and coronary occlusion time) Venous blood samples were drawn at the following times: at 5 minutes before induction of anaesthesia; on induction of CPB or occlusion of the LAD; at 15 minutes after induction of CPB or occlusion of the LAD; at 1 hour after induction of CPB or occlusion of the LAD; at 2 hours after induction of CPB or occlusion of the LAD; at 6 hours after induction of CPB or occlusion of the LAD; at 12 hours after induction of CPB or occlusion of the LAD; at 24 hours after induction of CPB or occlusion of the LAD; and at 48 hours after induction of CPB or occlusion of the LAD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Between the four groups, there were no differences in terms of age, BMI, or incidence of previous infarction. No randomisation method description. One out of four techniques was prospectively chosen for each participant at random, but no other description about how the allocation was made is given.
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unspecified and unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	All procedures were completed as they were planned. No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Clinical data were not properly reported as the other outcomes.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2006 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants (45 in each group of the study, i.e. A, B, C D, E, and F) undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified High‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (A), we report the comparison between the PMEA‐coated circuits/leucocyte filters group and the control group in the high‐risk cohort.
Interventions	Intervention group: high‐risk patients, PMEA‐coated circuits (Capiox SX 18; Terumo Medical Corp., Somerset, NJ, USA)/leucocyte filters (LG6B and BC2; Pall Biomedical Products, East Hills, NY, USA) Timing: leucocyte filters were deployed strategically approximately 30 minutes before cross‐clamp release. Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF; CD11b/CD18 expression; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; and T6: first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2007 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2007 Number of study centres and locations: not specified
Participants	Forty participants were assessed in part A of this study, including the high‐risk group. One hundred twenty participants were allocated into two groups (n = 60): Group 1 – PMEA‐coated circuits + leucocyte filters (Terumo, USA); Group 2 – control, uncoated circuits (Terumo, USA). Each group was further divided into three subgroups (n = 20) with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Inclusion criteria: CABG Exclusion criteria: coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: high‐risk patients, PMEA‐coated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA) + leucocyte filters (LG6B and BC2, Pall Biomedical Products, East Hills, NY, USA) Control group: uncoated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA)
Outcomes	Heart injury, AF, IL‐2, C3a, CK‐MB, complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt counts), PT, activated partial thromboplastin time, lactate levels, neutrophil and monocyte CD11b/CD18, phagocytic capacity of fibres; haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality Blood samples were obtained from radial artery and CS blood at the following intervals: T1, baseline – after induction of anaesthesia (before administration of heparin); T2, after heparin administration and before the initiation of CPB (TEG control); T3, on CPB – 15 minutes after initiation of CPB; T4, off CPB – before cessation of CPB; T5, protamine – 15 minutes after reversal with protamine; T6, ICU – first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective randomized study, 120 patients undergoing coronary artery bypass grafting (CABG) were allocated into two groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants' blinding unspecified, personnel blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Not specified for outcome assessors, data collectors blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Unclear risk	No funding was disclosed.

Gunaydin 2007 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2007 Number of study centres and locations: not specified
Participants	Forty participants were assessed in part C of this study, including the low‐risk group. One hundred twenty participants were allocated into two groups (n = 60): Group 1 – PMEA‐coated circuits + leucocyte filters (Terumo, USA); Group 2 – control, uncoated circuits (Terumo, USA). Each group was further divided into three subgroups (n = 20) with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Inclusion criteria: CABG Exclusion criteria: coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: low‐risk patients, PMEA‐coated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA) + leucocyte filters (LG6B and BC2, Pall Biomedical Products, East Hills, NY, USA) Control group: uncoated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA)
Outcomes	Heart injury, AF, IL‐2, C3a, CK‐MB, complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt counts), PT, activated partial thromboplastin time, lactate levels, neutrophil and monocyte CD11b/CD18, phagocytic capacity of fibres; haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality Blood samples were obtained from radial artery and CS blood at the following intervals: T1, baseline – after induction of anaesthesia (before administration of heparin); T2, after heparin administration and before the initiation of CPB (TEG control); T3, on CPB – 15 minutes after initiation of CPB; T4, off CPB – before cessation of CPB; T5, protamine – 15 minutes after reversal with protamine; T6, ICU – first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective randomized study, 120 patients undergoing coronary artery bypass grafting (CABG) were allocated into two groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants' blinding unspecified, personnel blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Not specified for outcome assessors, data collectors blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Unclear risk	No funding was disclosed.

Gunaydin 2009 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: over a 10‐month period Study date: 2009 Number of study centres and locations: not specified
Participants	Forty patients undergoing CABG were prospectively randomised to one of two perfusion protocols (n = 20): Group 1, minimised extracorporeal circuits (Mini‐CPB); Group 2, CECCs. Mean age: 63.3 years Sex (female/male ratio): 10% High‐risk patients Inclusion criteria: EuroSCORE classification (6+) Exclusion criteria: coagulopathy, endocarditis, or inability to obtain informed consent
Interventions	Intervention group: minimised extracorporeal circuits (Mini‐CPB) (ROCsafe MPC, Terumo, Ann Arbor, MI, USA) (a minimised “closed‐loop” circulatory and respiratory support circuit) Control group: CECCs (Capiox SX18, Terumo, USA) with a roller pump (System 1, TCM Heat exchanger, Terumo, USA) via an open system (Capiox Cardiotomy Reservoir, Terumo, USA) with 40‐mm blood filter (AV6SV, Pall)
Outcomes	Primary outcomes: air handling (intraoperative microembolic signals (GME)), perioperative rSO2, need for conversion to CECC and any complication, mean number of distal anastomoses Secondary outcomes: inflammation (IL‐6 and C3a), WBC and Plt counts, serum CK‐MB, neutrophil CD11b/CD18 levels, free Hb haemodilution, lactate levels, haemolysis and clinical outcome (haemodynamic parameters), perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and FFP, incidence of arrhythmia, use of inotropic support, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality, and Hct, erythrocyte, serum albumin, and globulin fractions Blood samples were collected at T1, after induction of anaesthesia (before administration of heparin); T2, TEG control; T3, 15 minutes after commencement of CPB; T4, before cessation of CPB (before protamine infusion); T5, 15 minutes after protamine reversal; and T6, ICU first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Low risk	Closed envelope allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Investigators were blinded (anaesthesia technicians collected data) to the allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were blinded (anaesthesia technicians collected data) to the allocation; it is unclear whether assessment of outcomes was blinded, but plausible.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Haemolysis and infection parameters were not clearly assessed.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2009 (B1).

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2‐year period Received: 10 June 2008 Revised: 1 October 2008 Accepted: 21 October 2008 Number of study centres and locations: four large tertiary care hospitals. Locations not specified
Participants	Fifty participants were assessed in part B1 of this study, with 25 patients in the intervention group receiving full‐dose heparin. One hundred patients were allocated into two equal groups (n = 50): Group 1 was treated with hyaluronan‐based heparin‐bonded preconnected circuits (Vision HFOGBS, Gish, California, USA) and Group 2 with identical uncoated controls (Vision HFO, Gish, USA). In the study group, half of the patients (n = 25) received low‐systemic heparin (125 IU/kg, ACT > 250 seconds) and half received a full dose of heparin. Mean age: 57.5 years Sex (female/male ratio): 36% No high‐risk patients Inclusion criteria: reoperation Exclusion criteria: coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: hyaluronan‐based heparin‐bonded extracorporeal circuits (Vision HFOGBS, Gish, California, USA) plus full dose heparin Control group: identical uncoated controls circuit (Vision HFO, Gish, USA)
Outcomes	Plt count; IL‐2; C3a; troponin T (CS); TAT; haemodynamic parameters; perfusion and cross‐clamp duration; intubation period; postoperative haemorrhage; the use of blood and plasma postoperatively; incidence of arrhythmia; use of inotropic support; complications and infection; the duration of ICU stay and hospital stay; perioperative mortality; Hb, Hct, erythrocyte, and WBC count; serum albumin and globulin fractions; neutrophil and monocyte CD11b/CD18; free plasma Hb Blood samples were collected after induction of anaesthesia (T1), heparin administration before CPB (T2), 15 minutes after initiation of CPB (T3), before cessation of CPB (T4), 15 minutes after reversal with protamine (T5), and the first POD at 8:00 a.m. (T6).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified for participants, although ICU personnel were blinded: "ICU staff was blind to the type of the circuit and heparin dose adapted"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2009 (C1).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2009 Number of study centres and locations: not specified
Participants	Thirty‐eight participants were assessed in part C1 of this study; the 25 participants from the condensed, coated circuit group (Group 1) are presented in this comparison. Seventy‐five patients were randomly allocated into three groups (n = 25): Group 1, CondECC; Group 2, ECC; Group 3, CONT. Mean age: 60.28 years Sex (female/male ratio): 39.47% No high‐risk patients Inclusion criteria: reoperation for CAB surgery Exclusion criteria: coagulopathy or ongoing anticoagulation, steroid therapy, nonsteroidal anti‐inflammatory drugs or aspirin within 5 days preoperatively
Interventions	Intervention group: condensed dual‐function open/closed configuration circuit (CondECC) (US patent no: 6946099) with components tip‐to‐tip coated with PMEA (Capiox SX 18; Terumo Medical Corporation, Somerset, NJ, USA), shortened tubing and components, and a priming volume under 800 mL, including a centrifugal pump (Sarns Centrifugal System, CXSP45; Terumo) and a venous air removal device (US patent no.: 6852280) incorporated shunt which bypasses the reservoir for the closed configuration. CPB was instituted either on open configuration, with a hard‐shell venous reservoir (Capiox SX; Terumo) and cardiotomy (CXCRXA; Terumo), or closed configuration, with a flexible venous reservoir (CX‐VRA0401T2; Terumo). Control group: a roller pump (System 1; Terumo) and conventional open, uncoated circuits (Capiox SX 18; Terumo), with a hard‐shell venous reservoir (Capiox SX; Terumo) and cardiotomy (CXCRXA; Terumo)
Outcomes	Inflammatory mediators (IL‐2, C3a), CK‐MB, troponin T and flow cytometry (CD11b/CD18), complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, fibrinogen levels, haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality; standard blood and urine chemistry, especially serum albumin and globulin fractions; free plasma Hb Blood samples were collected at the following intervals: (T1) baseline – after induction of anaesthesia (before administration of heparin); (T2) TEG control; (T3) on CPB – 15 minutes after initiation of CPB; (T4) off CPB – before cessation of CPB; (T5) protamine – 15 minutes after reversal with protamine; (T6) ICU – first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Allocation concealment (selection bias)	Low risk	Closed envelope allocation: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The investigators were blinded to the allocation; it is unclear whether patients were blinded: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors declare they have no proprietary interest in the products or devices described in the article.

Gunaydin 2010 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2 months Received for publication: 28 April 2010; accepted: 16 November 2010 Number of study centres and locations: not specified
Participants	Sixteen participants were assessed in part A of this study; patients from group 1 are included in this comparison. Forty‐eight participants were prospectively randomised to one of four perfusion protocols – Group 1: closed and totally hyaluronan‐based heparin‐free coated ECC with a soft‐shell coated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 2: closed and totally uncoated identical ECC with soft‐shell uncoated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 3: open, totally hyaluronan‐based heparin‐free coated ECC (n = 12); and Group 4: control open, uncoated ECC (n = 12). Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have a history of coagulation factor insufficiency. Mean age: 54.72 years Sex (female/male ratio): 25% High‐risk patients Inclusion criteria: EuroSCOREs ≥6 undergoing coronary revascularisation Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: closed and totally hyaluronan‐based heparin‐free coated (Vision HFO‐GBS‐HF ™, Gish Biomedical, Rancho Santa Margarita, CA) ECC with a soft‐shell coated venous reservoir (SVR11S2‐HFC™, Gish Biomedical) and a hard‐shell cardiotomy (CAPVRF44, Gish Biomedical) Control group: control open, uncoated ECC (n = 12)
Outcomes	IL 6, CK‐MB (CS), neutrophil CD11b/CD18, TAT, free plasma Hb, clinical outcomes (requirement for inotropes, atrial fibrillation, haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood (as units) and FFP (blood product as units) during the hospital stay, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality) and complete blood count (Hb, Hct, erythrocyte, WBC, and Plt counts) Blood samples were collected at T1: baseline (after induction of anaesthesia (before administration of heparin); T2: 15 minutes after CPB initiation; T3: before cessation of CPB but before protamine infusion; T4: 15 minutes after protamine reversal; and T5: in the ICU first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Allocation concealment (selection bias)	Low risk	Closed envelope allocation. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether participants were blinded. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether outcome assessment was blinded, but likely.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have history of coagulation factor insufficiency. No ITT analysis and no reported sample size in figures and tables are given.
Selective reporting (reporting bias)	Unclear risk	Haemodynamic, biological, and clinical data were reported but not in a detailed manner (each outcome was assessed in a very synthetic way).
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in this paper. This study was supported by University of Kirikkale Research Fund.

Gunaydin 2006 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified Medium‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (B), we report the comparison between the PMEA‐coated circuits/leucocyte filters group and the control group in the medium‐risk cohort.
Interventions	Intervention group: medium‐risk patients, PMEA‐coated circuits (Capiox SX 18; Terumo Medical Corp., Somerset, NJ, USA)/leucocyte filters (LG6B and BC2; Pall Biomedical Products, East Hills, NY, USA) Timing: leucocyte filters were deployed strategically approximately 30 min prior to cross‐clamp release. Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF; CD11b/CD18 expression; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; and T6: first POD at 08:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2006 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified Low‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (C), we report the comparison between the PMEA‐coated circuits/leucocyte filters and the control group in the low‐risk cohort.
Interventions	Intervention group: low‐risk patients, PMEA‐coated circuits (Capiox SX 18; Terumo Medical Corp., Somerset, NJ, USA)/leucocyte filters (LG6B and BC2; Pall Biomedical Products, East Hills, NY, USA) Timing: leucocyte filters were deployed strategically approximately 30 minutes before cross‐clamp release. Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF; CD11b/CD18 expressions; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; and T6: first POD at 08:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2006 (D).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified High‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (D), we report the comparison between the PMEA‐coated heparin‐bonded circuits and the control group in the high‐risk cohort.
Interventions	Intervention group: high‐risk patients, polypeptide‐based heparin‐bonded circuits with reduced heparinisation (Bioline† Coating; Maquet, Hirrlingen, Germany) Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF; CD11b/CD18 expression; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; T6: first POD at 08:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2006 (E).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified Medium‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (E), we report the comparison between the PMEA‐coated heparin‐bonded circuits and the control group in the medium‐risk cohort.
Interventions	Intervention group: medium‐risk patients, polypeptide‐based heparin‐bonded circuits with reduced heparinisation (Bioline† Coating; Maquet, Hirrlingen, Germany) Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF; CD11b/CD18 expressions; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of intensive care unit and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; and T6: first POD at 08:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2006 (F).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2006 Number of study centres and locations: not specified
Participants	Two hundred seventy participants undergoing CABG were allocated into three groups (n = 90): Group 1, PMEA‐coated circuits/leucocyte filters; Group 2, polypeptide‐based heparin‐bonded circuits with reduced heparinisation; and Group 3, control, uncoated circuits. Each group was further divided into three subgroups (n = 30), with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Mean age: not specified Sex (female/male ratio): not specified Low‐risk patients Inclusion criteria: not specified Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent In Gunaydin 2006 (F), we report the comparison between the PMEA‐coated heparin‐bonded circuits and the control group in the low‐risk cohort.
Interventions	Intervention group: low‐risk patients, polypeptide‐based heparin‐bonded circuits with reduced heparinisation (Bioline† Coating; Maquet, Hirrlingen, Germany) Control group: uncoated circuits (Capiox SX 18; Terumo)
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, and fibrinogen; standard blood and urine biochemistry, especially total protein, albumin, and globulin fractions; IL‐2 and C3a; CK‐MB; lactate postoperative bleeding; respiratory support time and incidence of AF, CD11b/CD18 expression; TATc; perioperative follow‐up (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality) Blood samples were collected at T1: after induction of anaesthesia; T2: after heparin administration; T3: 15 minutes after CPB; T4: before cessation of CPB; T5: 15 minutes after protamine reversal; and T6: first POD at 08:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective, randomized study, 270 patients undergoing coronary artery bypass grafting (CABG) were allocated to three groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding, but unlikely amongst the personnel because of the nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2007 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2007 Number of study centres and locations: not specified
Participants	Forty participants were assessed in part B of this study, including the medium‐risk group. One hundred twenty participants were allocated into two groups (n = 60): Group 1 – PMEA‐coated circuits + leucocyte filters (Terumo, USA); Group 2: control, uncoated circuits (Terumo, USA). Each group was further divided into three subgroups (n = 20) with respect to low‐ (EuroSCORE 0 to 2), medium‐ (3 to 5), and high‐risk (6+) patients. Inclusion criteria: CABG Exclusion criteria: coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: medium‐risk patients, PMEA‐coated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA) + leucocyte filters (LG6B and BC2, Pall Biomedical Products, East Hills, NY, USA) Control group: uncoated circuits (Capiox SX 18®, Terumo Medical Corporation, Ann Arbor, MI, USA)
Outcomes	Heart injury, AF, IL‐2, C3a, CK‐MB, complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt counts), PT, activated partial thromboplastin time, lactate levels, neutrophil and monocyte CD11b/CD18, phagocytic capacity of fibres; haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications, duration of ICU and hospital stay, perioperative mortality Blood samples were obtained from radial artery and CS blood at the following intervals: T1, baseline – after induction of anaesthesia (before administration of heparin); T2, after heparin administration and before the initiation of CPB (TEG control); T3, on CPB – 15 minutes after initiation of CPB; T4, off CPB – before cessation of CPB; T5, protamine – 15 minutes after reversal with protamine; T6, ICU – first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "In a prospective randomized study, 120 patients undergoing coronary artery bypass grafting (CABG) were allocated into two groups"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants' blinding unspecified, personnel blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Not specified for outcome assessors, data collectors blinded: "The operating room and intensive care unit staff collecting data were blinded to the entire study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Baseline, clinical, and inflammatory outcomes were not reported properly for each group.
Other bias	Unclear risk	No funding was disclosed.

Gunaydin 2009 (B2).

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2‐year period Received: 10 June 2008 Revised: 1 October 2008 Accepted: 21 October 2008 Number of study centres and locations: four large tertiary care hospitals. Locations not specified
Participants	Fifty participants were assessed in part B2 of this study, with 25 patients in the intervention group receiving low‐dose heparin. One hundred patients were allocated into two equal groups (n = 50): Group 1 was treated with hyaluronan‐based heparin‐bonded preconnected circuits (Vision HFOGBS, Gish, California, USA) and Group 2 with identical uncoated controls (Vision HFO, Gish, USA). In the study group, half of the patients (n = 25) received low‐systemic heparin (125 IU/kg, ACT > 250 seconds) and half received the full dose of heparin. Mean age: 59 years Sex (female/male ratio): 40% High‐risk patients Inclusion criteria: reoperation Exclusion criteria: coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: hyaluronan‐based heparin‐bonded extracorporeal circuits (Vision HFOGBS, Gish, California, USA) plus low‐dose heparin Control group: identical uncoated controls circuit (Vision HFO, Gish, USA)
Outcomes	Plt count; IL‐2; C3a; troponin T (CS); TAT; haemodynamic parameters; perfusion and cross‐clamp duration; intubation period; postoperative haemorrhage; the use of blood and plasma postoperatively; incidence of arrhythmia; use of inotropic support; complications and infection; the duration of ICU stay and hospital stay; perioperative mortality; Hb, Hct, erythrocyte, and WBC counts; serum albumin and globulin fractions; neutrophil and monocyte CD11b/CD18; free plasma Hb Blood samples were collected after induction of anaesthesia (T1), heparin administration before CPB (T2), 15 minutes after initiation of CPB (T3), before cessation of CPB (T4), 15 minutes after reversal with protamine (T5), and the first POD at 8:00 a.m. (T6).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified for participants, although ICU personnel were blinded: "ICU staff was blind to the type of the circuit and heparin dose adapted"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Gunaydin 2009 (C2).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Study publication date: 2009 Number of study centres and locations: not specified
Participants	Thirty‐seven participants were assessed in part C2 of this study; the 25 participants from the conventional coated circuit group (Group 2) are presented in this comparison. Seventy‐five participants were randomly allocated into three groups (n = 25): Group 1, CondECC; Group 2, ECC; Group 3, CONT. Mean age: 60.28 years Sex (female/male ratio): 39.47% No high‐risk patients Inclusion criteria: reoperation for CAB surgery Exclusion criteria: coagulopathy or ongoing anticoagulation, steroid therapy, nonsteroidal anti‐inflammatory drugs or aspirin within 5 days preoperatively
Interventions	Intervention group: PMEA‐coated ECC Control group: a roller pump (System 1; Terumo) and conventional open, uncoated circuits (Capiox SX 18; Terumo), with a hard‐shell venous reservoir (Capiox SX; Terumo) and cardiotomy (CXCRXA; Terumo)
Outcomes	Inflammatory mediators (IL‐2, C3a), CK‐MB, troponin T and flow cytometry (CD11b/CD18), complete blood count (Hb, Hct, erythrocyte, leucocyte [WBC], and Plt counts), PT, activated partial thromboplastin time, fibrinogen levels, haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia, use of inotropic support, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality; standard blood and urine chemistry, especially serum albumin and globulin fractions; free plasma Hb Blood samples were collected at the following intervals: (T1) baseline – after induction of anaesthesia (before administration of heparin); (T2) TEG control; (T3) on CPB – 15 minutes after initiation of CPB; (T4) off CPB – before cessation of CPB; (T5) protamine – 15 minutes after reversal with protamine; (T6) ICU – first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Allocation concealment (selection bias)	Low risk	Closed envelope allocation: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The investigators were blinded to the allocation; it is unclear whether patients were blinded: "Patients were randomly allocated to one of the three groups (closed envelope allocation) with the investigators blinded to the allocation."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors declare they have no proprietary interest in the products or devices described in the article.

Gunaydin 2009 (D1).

Study characteristics
Methods	Prospective, randomised pilot study Run‐in period: 4‐month period from March until June 2008 Study publication date: 2009 Number of study centres and locations: not specified.
Participants	Fourteen participants were assessed in part D1 of this study; participants from group 1 were included in comparison D1. Forty participants (EuroSCORE 6+) undergoing coronary revascularisation were prospectively randomised to one of four perfusion protocols – Group 1 (n = 10): conventional circuits (ECC) + 2 leucocyte filters (LG6B, Pall, USA) with the method of 2‐phase (continuous + strategic) leucofiltration; Group 2 (n = 10): ECC + single leucocyte filter with the method of continuous leucofiltration; Group 3 (n = 10): ECC + single leucocyte filter with the method of strategic leucofiltration; Group 4 (n = 10) control: ECC without leucocyte filtration. Mean age: 68.72 years Sex (female/male ratio): 42.86% High‐risk patients Inclusion criteria: high risk – EuroSCORE 6+ CABG Exclusion criteria: not specified
Interventions	Intervention group: Group 1 (n = 10) – conventional circuits (ECC) (compactflo EVO circuits [Dideco, Mirandola, Italy]) + 2 leucocyte filters (LG6B; Pall Biomedical Products, East Hills, NY) with the method of two‐phase (continuous + strategic) leucofiltration Timing: continuous leucofiltration started as soon as the CPB was on using the first leucofilter. Approximately 30 minutes before cross‐clamp release, the first filter was clamped and the second leucofilter was deployed strategically. Control group: ECC (compactflo EVO circuits [Dideco, Mirandola, Italy]) without leucocyte filtration
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt count); standard blood and urine biochemistry and total protein, albumin, and globulin fractions; IL‐6, TNF‐α, and procalcitonin; CK‐MB (CS blood); BNP (serum level); MMP‐9, S100B, and D‐dimer (serum level); clinical outcomes (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia (AF)); use of inotropic support; complications; the duration of ICU and hospital stay; perioperative mortality Plt function was evaluated by TEG (ROTEG; Pentapharm, Munich, Germany) during the operation. CT, CFT, α‐angle, MCF, and A5 were measured in samples T1 to T4. Blood samples were collected at T1: baseline (after induction of anaesthesia and before administration of heparin); T2: on CPB, 5 minutes after initiation of CPB; T3: cross‐clamp (5 minutes after cross‐clamping of aorta);T4: off CPB, 5 minutes after cessation of CPB; T5: ICU24 (first POD at 8:00 am); and T6: ICU48 (second POD at 8:00 am).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "… were prospectively randomized to one of the four perfusion protocols"
Allocation concealment (selection bias)	Unclear risk	Investigators were blinded to the allocation but no further allocation method description: "… with the investigators blinded to the allocation".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	No adequate information about blinding of participants. Investigators were blinded to the allocation: "… with the investigators blinded to the allocation".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in the paper.

Gunaydin 2009 (D2).

Study characteristics
Methods	Prospective, randomised pilot study Run‐in period: 4‐month period from March until June 2008 Study publication date: 2009 Number of study centres and locations: not specified.
Participants	Thirteen participants were assessed in part D2 of this study; patients from group 2 were included in comparison D2. Forty participants (EuroSCORE 6+) undergoing coronary revascularisation were prospectively randomised to one of four perfusion protocols – Group 1 (n = 10): conventional circuits (ECC) + 2 leucocyte filters (LG6B, Pall, USA) with the method of 2‐phase (continuous + strategic) leucofiltration; Group 2 (n = 10): ECC + single leucocyte filter with the method of continuous leucofiltration; Group 3 (n = 10): ECC + single leucocyte filter with the method of strategic leucofiltration; Group 4 (n = 10) control: ECC without leucocyte filtration. Mean age: 65.87 years Sex (female/male ratio): 61.54% High‐risk patients Inclusion criteria: high risk – EuroSCORE 6+ CABG Exclusion criteria: not specified
Interventions	Intervention group: ECC (compactflo EVO circuits [Dideco, Mirandola, Italy]) + single leucocyte filter with the method of continuous leucofiltration Timing: continuous leucofiltration started as soon as the CPB was on until the end of the procedure. Control group: ECC (compactflo EVO circuits [Dideco, Mirandola, Italy]) without leucocyte filtration
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt count); standard blood and urine biochemistry and total protein, albumin, and globulin fractions; IL‐6, TNF‐α, and procalcitonin; CK‐MB (CS blood); BNP (serum level); MMP‐9, S100B, and D‐dimer (serum level); clinical outcomes (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia (AF)); use of inotropic support; complications; the duration of ICU and hospital stay; perioperative mortality Plt function was evaluated by TEG (ROTEG; Pentapharm, Munich, Germany) during the operation. CT, CFT, α‐angle, MCF, and A5 were measured in samples T1 to T4. Blood samples were collected at T1: baseline (after induction of anaesthesia and before administration of heparin); T2: on CPB, 5 minutes after initiation of CPB; T3: cross‐clamp (5 minutes after cross‐clamping of aorta); T4: off CPB, 5 minutes after cessation of CPB; T5: ICU24 (first POD at 8:00 am); and T6: ICU48 (second POD at 8:00 am).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "… were prospectively randomized to one of the four perfusion protocols"
Allocation concealment (selection bias)	Unclear risk	Investigators were blinded to the allocation but no further allocation method description: "… with the investigators blinded to the allocation".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	No adequate information about blinding of participants. Investigators were blinded to the allocation: "… with the investigators blinded to the allocation".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in the paper.

Gunaydin 2009 (D3).

Study characteristics
Methods	Prospective, randomised pilot study Run‐in period: 4‐month period from March until June 2008 Study publication date: 2009 Number of study centres and locations: not specified
Participants	Thirteen participants were assessed in part D3 of the study; patients from group 3 were included in comparison D3. Forty participants (EuroSCORE 6+) undergoing coronary revascularisation were prospectively randomised to one of four perfusion protocols – Group 1 (n = 10): conventional circuits (ECC) + 2 leucocyte filters (LG6B, Pall, USA) with the method of 2‐phase (continuous + strategic) leucofiltration; Group 2 (n = 10): ECC + single leucocyte filter with the method of continuous leucofiltration; Group 3 (n = 10): ECC + single leucocyte filter with the method of strategic leucofiltration; and Group 4 (n = 10) control: ECC without leucocyte filtration. Mean age: 67.82 years Sex (female/male ratio): 53.85% High‐risk patients Inclusion criteria: high risk – EuroSCORE 6+ CABG Exclusion criteria: not specified
Interventions	Intervention group: ECC (compactflo EVO circuits [Dideco, Mirandola, Italy]) + single leucocyte filter with the method of strategic leucofiltration Timing: the single filter was deployed strategically ~30 minutes before cross‐clamp release. Control group: ECC (compactflo EVO circuits [Dideco, Mirandola, Italy]) without leucocyte filtration
Outcomes	Complete blood count (Hb, Hct, erythrocyte, leucocyte (WBC), and Plt count); standard blood and urine biochemistry and total protein, albumin, and globulin fractions; IL‐6, TNF‐α, and procalcitonin; CK‐MB (CS blood); BNP (serum level); MMP‐9, S100B, and D‐dimer (serum level); clinical outcomes (haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood and plasma, incidence of arrhythmia [AF]); use of inotropic support; complications; the duration of ICU and hospital stay; perioperative mortality Plt function was evaluated by TEG (ROTEG; Pentapharm, Munich, Germany) during the operation. CT, CFT, α‐angle, MCF, and A5 were measured in samples T1 to T4. Blood samples were collected at T1: baseline (after induction of anaesthesia and before administration of heparin); T2: on CPB, 5 minutes after initiation of CPB; T3: cross‐clamp (5 minutes after cross‐clamping of aorta); T4: off CPB, 5 minutes after cessation of CPB; T5: ICU24 (first POD at 8:00 am); and T6: ICU48 (second POD at 8:00 am).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description: "… were prospectively randomized to one of the four perfusion protocols"
Allocation concealment (selection bias)	Unclear risk	Investigators were blinded to the allocation but no further allocation method description: "… with the investigators blinded to the allocation".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	No adequate information about blinding of participants. Investigators were blinded to the allocation: "… with the investigators blinded to the allocation".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in the paper.

Gunaydin 2010 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2 months Received for publication: 28 April 2010; accepted: 16 November 2010 Number of study centres and locations: not specified
Participants	Sixteen participants were assessed in part B of this study; participants from group 2 are included in this comparison. Forty‐eight participants were prospectively randomised to one of four perfusion protocols – Group 1: closed and totally hyaluronan‐based heparin‐free coated ECC with a soft‐shell coated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 2: closed and totally uncoated identical ECC with soft‐shell uncoated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 3: open, totally hyaluronan‐based heparin‐free coated ECC (n = 12); and Group 4: control open, uncoated ECC (n = 12). Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have a history of coagulation factor insufficiency. Mean age: 54.72 years Sex (female/male ratio): 25% High‐risk patients Inclusion criteria: EuroSCOREs ≥ 6 undergoing coronary revascularisation Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: closed and totally uncoated identical ECC with soft‐shell uncoated venous reservoir and a hard‐shell cardiotomy Control group: control open, uncoated ECC
Outcomes	IL 6, CK‐MB (CS), neutrophil CD11b/CD18, TAT, free plasma Hb, clinical outcomes (requirement for inotropes, atrial fibrillation, haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood (as units) and FFP (blood product as units) during the hospital stay, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality and complete blood count (Hb, Hct, erythrocyte, WBC, and Plt counts) Blood samples were collected at T1: baseline (after induction of anaesthesia (before administration of heparin); T2: 15 minutes after CPB initiation; T3: before cessation of CPB but before protamine infusion; T4: 15 minutes after protamine reversal; and T5: in the ICU first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Allocation concealment (selection bias)	Low risk	Closed envelope allocation. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether participants were blinded. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether outcome assessment was blinded, but likely.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have history of coagulation factor insufficiency. No ITT analysis and no reported sample size in figures and tables are given.
Selective reporting (reporting bias)	Unclear risk	Haemodynamic, biological, and clinical data were reported but not in a detailed manner (each outcome was assessed in a very synthetic way).
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in this paper. This study was supported by University of Kirikkale Research Fund.

Gunaydin 2010 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2 months Received for publication: 28 April 2010; accepted: 16 November 2010 Number of study centres and locations: not specified
Participants	Sixteen participants were assessed in part C of study; participants from group 3 are included in this comparison. Forty‐eight participants were prospectively randomised to one of four perfusion protocols – Group 1: closed and totally hyaluronan‐based heparin‐free coated ECC with a soft‐shell coated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 2: closed and totally uncoated identical ECC with soft‐shell uncoated venous reservoir and a hard‐shell cardiotomy (n = 12); Group 3: open, totally hyaluronan‐based heparin‐free coated ECC (n = 12); Group 4: control open, uncoated ECC (n = 12). Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have a history of coagulation factor insufficiency. Mean age: 54.72 years Sex (female/male ratio): 25% High‐risk patients Inclusion criteria: EuroSCORE ≥6 undergoing coronary revascularisation Exclusion criteria: known coagulopathy, endocarditis, and inability to obtain informed consent
Interventions	Intervention group: open, totally hyaluronan‐based heparin‐free coated ECC Control group: control open, uncoated ECC
Outcomes	IL‐6, CK‐MB (CS), neutrophil CD11b/CD18, TAT, free plasma Hb, clinical outcomes (requirement for inotropes, AF, haemodynamic parameters, perfusion and cross‐clamp duration, intubation period, postoperative haemorrhage, the use of blood (as units) and FFP (blood product – as units) during the hospital stay, complications and infection, the duration of ICU stay and hospital stay, perioperative mortality, and complete blood count (Hb, Hct, erythrocyte, WBC, and Plt counts) Blood samples were collected at T1: baseline (after induction of anaesthesia and before administration of heparin); T2: 15 minutes after CPB initiation; T3: before cessation of CPB but before protamine infusion; T4: 15 minutes after protamine reversal; and T5: in the ICU, first POD at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Allocation concealment (selection bias)	Low risk	Closed envelope allocation. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether participants were blinded. No accurate description of a randomisation method: "prospectively randomized (closed envelope allocation) to one of the four perfusion protocols with the investigators blinded to the allocation"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were blinded to the allocation; it is unclear whether outcome assessment was blinded, but likely.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Four participants were excluded from the study. Three of them refused to participate in this study, and one was reported to have history of coagulation factor insufficiency. No ITT analysis and no reported sample size in figures and tables are given.
Selective reporting (reporting bias)	Unclear risk	Haemodynamic, biological, and clinical data were reported but not in a detailed manner (each outcome was assessed in a very synthetic way).
Other bias	Low risk	The senior author has stated that authors have reported no material, financial, or other relationship with any healthcare‐related business or other entity whose products or services are discussed in this paper. This study was supported by University of Kirikkale Research Fund.

Gygax 2018.

Study characteristics
Methods	Randomised controlled trial Run‐in period: between June 2011 and November 2014 Study date: 2018 Number of study centres and locations: single centre, Department of Cardiovascular Surgery, University Hospital Bern, University of Bern, Bern, Switzerland
Participants	Fifty participants were enrolled and randomised into two groups: MiECC group (n = 24) and CECC group (n = 26). Because of a logistics error in patient inclusion (patient aged 2 months above 80 years), this patient had to be excluded from the MiECC group in retrospect. The next patient was randomly assigned to the CECC regimen. Mean age: 66.1 years Sex (female/male ratio): 44% No high‐risk patients Inclusion criteria: elective isolated SAVR Exclusion criteria: absence of written informed consent, emergency surgery, redo surgery, age below 18 years or above 80 years, pregnancy, pre‐existing neurological deficit or neurocognitive disorder, alcohol or drug abuse, carotid artery stenosis 70%, clinically or laboratory apparent infection, and treatment with oral anticoagulants other than acetylsalicylic acid.
Interventions	Intervention group: type II minimal‐invasive extracorporeal circuit. The type II MiECC system (Maquet Deutschland GmbH) was equipped with a centrifugal pump (Rotaflow, Maquet Deutschland GmbH). Identical to the CECC setting, microporous capillary membrane (polypropylene) oxygenators with or without integrated arterial line filter were used. Shed blood was reintroduced into the MiECC using an automated suction device (Smart Suction System, Cardiosmart, Muri, Switzerland). At the end of perfusion, the MiECC was completely emptied by retransfusion of the remaining blood through the arterial line to the participant. Control group: conventional extracorporeal circuits (CECC) (Maquet Deutschland GmbH, Rastatt, Germany) included a roller pump (HL‐20), oxygenator types with or without filter Quadrox (Maquet Deutschland GmbH), Capiox FX (Terumo Corp., Tokyo, Japan) or Affinity Fusion (Medtronic Inc., Minneapolis, MN, USA), and an external arterial line filter (40 lm; Pall AL6, Terumo Corp.). The CECC system consisted of a venous hard‐shell reservoir where the aspirated blood was collected. The CECC setup included a cell‐saving device (Autolog, Medtronic Inc.) to process aspirated blood. Remaining blood in the CECC circuit was cell saved and then retransfused to the participant.
Outcomes	Inflammatory responses (IL‐6, TNF‐α, sCD40‐ ligand), complement activation (sC5b‐9), selected clinical endpoints (included perfusion time, transfusion of allogeneic blood products, postoperative bleeding, sepsis, new onset of AF, stroke, and in‐hospital mortality), and activation of the coagulation system (D‐dimer, TAT) Blood samples were taken at three time points: before operation and at 2 hours and 24 hours after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The two groups were similar in baseline characteristics, but no randomisation method was described: "Fifty patients were prospectively randomized to MiECC or CECC perfusion regimen."
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis. Because of a logistic error in patient inclusion (participant aged 2 months above 80 years), one participant had to be excluded from the MiECC group in retrospect. The following participant was randomly assigned to the CECC regimen.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	High risk	The study was supported by Grants from Fumedica AG and Maquet, Muri, Switzerland. The authors declare no conflicts of interest.

Hamada 2001 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Received: December 2000; revised: March 2001 Number of study centres and locations: single centre, Department of Surgery II, Ehime University School of Medicine, Shigenobu, Ehime, Japan
Participants	Fifteen participants were assessed in part A of the study; this comparison included the participants randomised to the H group. Thirty patients were allocated randomly to equal groups with a conventional circuit and arterial line filter (C group, n = 10), a heparin‐coated circuit with a conventional filter (H group, n = 10), or a heparin‐coated circuit with a leucocyte‐depleting arterial line filter (HF group, n = 10). Mean age: 67 years Sex (female/male ratio): 26.66% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: not specified
Interventions	Intervention group: a heparin‐coated circuit with a conventional filter (H group); all surfaces in contact with blood were treated with heparin (Duraflo II; Baxter Healthcare Co., Irvine, CA, USA), and a conventional arterial line filter was used. Control group: conventional circuit and arterial line filter; a CPB circuit without heparin coating was used with a conventional arterial line filter.
Outcomes	Cytokines (IL‐6, IL‐8) and respiratory function (RI, myocardial complication, leucocyte counts) Blood samples were collected at four time points: after induction of anaesthesia but before CPB, immediately after termination of CPB, at the end of the operation, and at 4 hours and 12 hours after operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given; however, no significant differences were noted between the three groups: "All patients were allocated randomly to 1 of 3 groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Hamada 2001 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Received: December 2000; revised: March 2001 Number of study centres and locations: single centre, Department of Surgery II, Ehime University School of Medicine, Shigenobu, Ehime, Japan
Participants	Fifteen participants were assessed in part B of the study; this comparison included the participants randomised to the HF group. Thirty patients were allocated randomly to equal groups with a conventional circuit and arterial line filter (C group, n = 10), a heparin‐coated circuit with a conventional filter (H group, n = 10), or a heparin‐coated circuit with a leucocyte‐depleting arterial line filter (HF group, n = 10). Mean age: 68.33 years Sex (female/male ratio): 26.66% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: not specified
Interventions	Intervention group: heparin‐coated circuit with a leucocyte‐depleting arterial line filter (HF group), the CPB circuit combined heparin coating with a leucocyte‐depleting filter (LG6; Pall Biomedical Products, East Hills, NY, USA) Control group: conventional circuit and arterial line filter
Outcomes	Cytokines (IL‐6, IL‐8) and respiratory function (RI, myocardial complication, leucocyte counts) Blood samples were collected at four time points: after induction of anaesthesia but before CPB, immediately after termination of CPB, at the end of the operation, and at 4 hours and 12 hours after operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given; however, no significant differences were noted between the three groups: "All patients were allocated randomly to 1 of 3 groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Hamano 2001.

Study characteristics
Methods	Prospective, randomised study Run‐in period: from April 1997 to December 1998 Study date: 2001 Number of study centres and locations: not specified
Participants	Forty‐eight adult patients were included in this study, 27 of whom underwent conventional surgery and 21 MICS. Mean age: 62.25 years Sex (female/male ratio): 41.67% No high‐risk patients Inclusion criteria: patients who underwent surgery for single valve disease Exclusion criteria: not specified
Interventions	Intervention group: minimally invasive cardiac surgery (a small skin incision) Control group: conventional cardiac surgery
Outcomes	Duration and degree of SIRS, the level of CRP, the operating times, perfusion times or aorta clamp times, mean volume of blood transfusion, infection or organ failure Blood samples were collected 1 day, 3 days, and 6 days postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear randomisation method description is given; however, the two groups had similar baseline characteristics: "of the 48 patients, 21 were randomly selected to undergo MICS".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; unfeasible because of the type of intervention; unclear whether it influenced management in the postoperative period and outcomes
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Hao 2019.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registered: 14 February 2011 Number of study centres and locations: single centre at Beijing Anzhen Hospital, Capital Medical University
Participants	Thirty‐six participants were randomly assigned to an MP group (MG, N = 18) and a non‐MP group (NMG, N = 18). Mean age: 53.3 years Sex (female/male ratio): 58.33% No high‐risk patients Inclusion criteria: not specified Exclusion criteria: infections, tumours, pregnancy, or autoimmune disorders
Interventions	Intervention group: MP (500 mg in the CPB priming) Control group: non‐MP before, during or after the operation
Outcomes	Monocyte and CD4+CD25+ Treg subsets, HLA‐DR, CD86, CD64, and toll‐like receptor 4; incidences of AF; bypass time; cross‐clamp time; length of ICU stay; duration of mechanical ventilation; and length of hospital stay Blood samples were collected before the operation, immediately after the operation, and once a day for 7 days post‐operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description, overall adequate balance of baseline characteristics: "Thirty‐six patient [sic] who underwent heart surgery with CPB were randomly assigned"
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly. Trial registration – Clinicaltrials.gov: NCT01296074
Other bias	Low risk	The authors declared no potential conflicts of interest with respect to the research, authorship, or publication of this article. The work was supported by the National Natural Science Foundation of China.

Harig 1999 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Fourteen participants were included in part A of the analysis, including the participants from group A. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone preoperatively and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 62 years Sex (female/male ratio): 66% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal or hepatic insufficiency and known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and postoperatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6, IL‐8, and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given; however, no significant differences were noted between the four groups: "Cohorts of 10 patients were randomized independently".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Harig 1999 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Thirteen participants were included in part B of the analysis, including the participants from group B. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone preoperatively and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 62 years Sex (female/male ratio): 66% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal or hepatic insufficiency and known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and postoperatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6, IL‐8, and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given; however, no significant differences were noted between the four groups: "Cohorts of 10 patients were randomized independently".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Harig 1999 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Thirteen participants were included in part C of the analysis, including the participants from group C. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone preoperatively and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 62 years Sex (female/male ratio): 66% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal or hepatic insufficiency and known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and post‐operatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6, IL‐8, and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given; however, no significant differences were noted between the four groups: "Cohorts of 10 patients were randomized independently".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Harig 2001 (A).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available, published in 2001 Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Fourteen participants were included in part A of the analysis, including the participants from group A. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone pre‐ and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 63 years Sex (female/male ratio): 67% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal, pulmonary, or hepatic insufficiency and a known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and postoperatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6 and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given, and very limited wording was adopted; however, no significant differences were noted between the four groups: "… prospectively after randomizing them into four groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Harig 2001 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available, published in 2001 Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Thirteen participants were included in part B of the analysis, including the participants from group B. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone pre‐ and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 63 years Sex (female/male ratio): 67% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal, pulmonary, or hepatic insufficiency and a known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and postoperatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6 and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given, and very limited wording was adopted; however, no significant differences were noted between the four groups: "… prospectively after randomizing them into four groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Harig 2001 (C).

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available, published in 2001 Number of study centres and locations: unspecified, authors from Erlangen, Germany
Participants	Thirteen participants were included in part C of the analysis, including the participants from group C. Forty patients with CAD undergoing elective CABG were prospectively randomised to four groups of 10 – Group A: prednisolone pre‐ and postoperatively (2 × 250 mg); Group B: aprotinin perioperatively (6 million KIU); Group C: heparin‐coated circuits ('Bioline' by Jostra); Group D: no special measures were taken (controls). Mean age: 63 years Sex (female/male ratio): 67% No high‐risk patients (CABG) Inclusion criteria: normal LV function (i.e. left‐LVEDP < 14 mmHg and EF > 55); all underwent elective CABG. Exclusion criteria: renal, pulmonary, or hepatic insufficiency and a known anaphylaxis to aprotinin
Interventions	Group A: prednisolone 250 mg pre‐ and postoperatively Group B: aprotinin as follows – a loading dose of 2,000,000 KIU over a 15‐minute period at the start of the operation was followed by a continuous infusion of 500,000 KIU/h throughout the operation. Another 2,000,000 KIU were added to the prime volume of the heart‐lung machine. Group C: heparin‐coated circuits modified by the Bioline procedure Group D: the control group, only standard anaesthetic techniques and operative procedures
Outcomes	Cytokine release (IL‐6 and IL‐10) In each group, blood samples were taken from the radial artery at the following times: after induction of anaesthesia and before the infusion of aprotinin or prednisolone preoperatively and 30 minutes after the start of CPB, at the end of CPB, and 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 12 hours after the end of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description is given, and very limited wording was adopted; however, no significant differences were noted between the four groups: "… prospectively after randomizing them into four groups".
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; likely for participants, unlikely for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hatemi 2016.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available, published in 2016 Number of study centres and locations: unspecified, authors from Istanbul University, Istanbul, Turkey
Participants	A total of 32 participants were assessed. Mean age: 61 years Sex (female/male ratio): 23% No high‐risk patients (CABG) Inclusion criteria: patients with a diagnosis of ischaemic heart disease undergoing elective CABG Exclusion criteria: low EF (EF < 30%), emergency operation, LV aneurysm, reoperation, concomitant procedures (e.g. valve replacement, carotid endarterectomy, and LV aneurysm resection), perioperative MI, old age (> 70 years), renal insufficiency, and consent denial
Interventions	Intervention: PC‐coated ECC system Control: noncoated open ECC system
Outcomes	IL‐6, TNF‐α, IL‐β, procalcitonin, IL‐8, IL‐10 Blood samples for haematologic, biochemical, and immunologic parameters were collected preoperatively (0), before cross‐clamp removal (1), after CPB termination (2), at the sixth postoperative hour (3), and at the first postoperative week (4). Clinical signs of inflammation were also followed and recorded until the seventh POD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Adequate randomisation protocol: "randomly divided into two groups on the basis of odd or even number poll"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were blinded while study personnel were unblinded: "enrolled in the study group (single blinded)".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given; unlikely, although it might not have influenced values of laboratory outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed. "The authors declare that they have no competing interests."

Hatori 1994.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Registration date: not available, published in 1994 Number of study centres and locations: unspecified
Participants	A total of 14 participants were assessed. Mean age: 62 years Sex (female/male ratio): 27% No high‐risk patients (CABG) Inclusion criteria: absence of major noncardiac illness, normal liver function, no renal insufficiency, no severely impaired LV function, and no history of recent use of steroids and acetylsalicylic acid.
Interventions	Intervention: heparin‐coated membrane oxygenator Control: uncoated oxygenator
Outcomes	Plt activation, prostaglandin production, complement activation, and activated granulocyte released substance Measured prior to, during, and 6 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "11 males and 3 females, underwent coronary artery bypass surgery and were randomly divided into 2 groups of 7 patients each"
Allocation concealment (selection bias)	Unclear risk	No details given about allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given; plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given; unlikely, but unclear whether outcomes could be influenced
Incomplete outcome data (attrition bias) All outcomes	Low risk	No details given
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hayashi 2003.

Study characteristics
Methods	Randomised controlled trial Run‐in period: 1997 to 2000 Registration date: not available, published in 2003 Number of study centres and locations: unspecified, authors from Osaka, Japan
Participants	A total of 54 participants were assessed in this study. Mean age: 57 years Sex (female/male ratio): 59% No high‐risk patients (AVR) Inclusion criteria: adult patients undergoing elective AVR Exclusion criteria: reoperation, concomitant mitral valve disease to be surgically treated, CAD requiring revascularisation, and severe LV hypertrophy shown as LVMI over 150 g/m2 by echocardiography
Interventions	Intervention: leucocyte‐depleted terminal blood cardioplegia Control: standard CPB management
Outcomes	Clinical outcomes, depletion efficiency of leucocytes and neutrophils, aortic cross‐clamp time, plasma PMN‐Elastase, plasma Nox Plasma CK‐MB was measured every 6 hours after operation using immunoinhibition assay, and the maximum value during the first 24 postoperative hours was used as peak CK‐MB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "patients were randomly divided into 2 groups"
Allocation concealment (selection bias)	Unclear risk	No details given around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given; plausible for participants, unlikely for personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given; potentially not performed, unclear whether it could have influenced laboratory outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Heyer 2002.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Registration date: not available, published in 2002 Number of study centres and locations: unspecified, authors from Columbia–Presbyterian Medical Center, New York, NY
Participants	A total of 61 participants were assessed. Mean age: 63 years Sex (female/male ratio): 19% No high‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective CABG Exclusion criteria: patients with evidence of pre‐existing neurologic or psychiatric illness
Interventions	Intervention: heparin‐bonded CPB Control: uncoated circuits
Outcomes	Serum samples were analysed to evaluate markers of complement activation (C3a), proinflammatory cytokines (TNF‐α, IL‐1β, and IL‐6), and coagulation (TAT). Blood samples were obtained preoperatively, at 15 minutes and 60 minutes after the onset of CPB, and 6 hours and 24 hours after the completion of surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "Patients were randomly assigned to receive HB‐CPB circuits (Carmeda, Medtronic, Inc) or NH‐CPB circuits (Medtronic, Inc, Minneapolis, MN) for the procedure."
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Described as blinded in the abstract, although no details given: "Prospective, randomized, blinded clinical trial."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Study is described as blinded in the abstract, although no details were given: "Prospective, randomized, blinded clinical trial."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hill 1995.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Registration date: not available, published in 1995 Number of study centres and locations: unspecified, authors from University of Nebraska Medical Center, Omaha, Nebraska
Participants	Twenty participants were assessed in total. Mean age: 60 years Sex (female/male ratio): not recorded No high‐risk patients (CABG) Inclusion criteria: adult male patients who were scheduled for elective myocardial revascularisation Exclusion criteria: patients who were smoking, or who had quit smoking less than 6 months prior to surgery, and those with a history of asthma
Interventions	Intervention: MP Control: no MP
Outcomes	TNF‐α, IL‐6, airway NO, measured at 6 hours after arrival of the patient in the ICU
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "were randomly assigned to one of two groups"
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given around blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given around blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hill 1995 (A).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Registration date: not available, published in 1995 Number of study centres and locations: unspecified, authors from University of Nebraska Medical Center, Omaha, Nebraska
Participants	Twelve participants were assessed in part A of this comparison, which includes patients from group C in the original study. Twenty‐four male patients scheduled for elective aorta‐coronary bypass were randomised equally to one of three groups: a control group (group A); a group that received MP (1 gram intravenously) administered 5 minutes before CPB (group B); and a group that received aprotinin, 140 mg intravenously as a loading dose, 140 mg in the pump prime and 35 mg/h as an IV constant infusion until chest closure (group C). Inclusion criteria: male patients scheduled for elective CABG Exclusion criteria: none provided
Interventions	Group A: control group Group B: MP (1 gram intravenously) administered 5 minutes before CPB Group C: aprotinin, 140 mg intravenously as a loading dose, 140 mg in the pump prime and 35 mg/h as an IV constant infusion until chest closure
Outcomes	CD11b, TNF‐α, measured at baseline (after placement of the arterial and IV catheters but before anaesthetic drug administration), after 50 minutes of CPB, and 30 minutes after termination of CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "24 male patients scheduled for elective aorta‐coronary bypass were randomized equally to one of three groups"
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Only laboratory personnel were blinded: "Laboratory personnel were blinded as to which group of the study each patient was assigned".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hill 1995 (B).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Registration date: not available, published in 1995 Number of study centres and locations: unspecified, authors from University of Nebraska Medical Center, Omaha, Nebraska
Participants	Twelve participants were assessed in part B of this comparison, which includes patients from group B in the original study. Twenty‐four male patients scheduled for elective aorta‐coronary bypass were randomised equally to one of three groups: a control group (group A); a group that received MP (1 gram intravenously) administered 5 minutes before CPB (group B); and a group that received aprotinin, 140 mg intravenously as a loading dose, 140 mg in the pump prime and 35 mg/h as an IV constant infusion until chest closure (group C). Inclusion criteria: male patients scheduled for elective CABG Exclusion criteria: none provided
Interventions	Group A: control group Group B: MP (1 gram intravenously) administered 5 minutes before CPB Group C: aprotinin, 140 mg intravenously as a loading dose, 140 mg in the pump prime and 35 mg/h as an IV constant infusion until chest closure
Outcomes	CD11b, TNF‐α, measured at baseline (after placement of the arterial and IV catheters but before anaesthetic drug administration), after 50 minutes of CPB, and 30 minutes after termination of CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "24 male patients scheduled for elective aorta‐coronary bypass were randomized equally to one of three groups"
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Only laboratory personnel were blinded: "Laboratory personnel were blinded as to which group of the study each patient was assigned".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No funding was disclosed, although no supporting statement was provided.

Hoedemaekers 2005.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Registration date: not available, published in 2005 Number of study centres and locations: unspecified, authors from University Hospital Nijmegen, the Netherlands
Participants	Twenty participants were assessed in this study. Inclusion criteria: patients aged 18 years or older scheduled for elective CABG Exclusion criteria: history of diabetes, fasting blood glucose levels above 100 mg/dL on the day before surgery, MI within 4 weeks before surgery, cardiogenic shock or renal failure (serum Cr level above 1.7 mg/dL). Patients were also excluded if they had used any medication within 4 weeks before surgery known to modulate the inflammatory response (for example, nonsteroidal anti‐inflammatory drugs or steroids) or when there were clinical signs of infection or inflammatory disease. Patients undergoing off‐pump cardiac surgery were excluded.
Interventions	Intervention: IT Control: standard glycaemic management
Outcomes	Concentrations of the proinflammatory cytokines TNF‐α and IL‐ 6 and the anti‐inflammatory cytokine IL‐11 at 0 hours, 1 hour, 2 hours, 4 hours, 8 hours, 12 hours, 16 hours, and 24 hours after admission to the ICU
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given around the randomisation method: "We performed a randomized, controlled study in 20 nondiabetic patients undergoing elective coronary bypass surgery"
Allocation concealment (selection bias)	Low risk	Sealed envelope: "Assignments to the treatment groups were made with the use of sealed envelopes"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given for blinding, unfeasible for study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given, unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Did not report graph for TNF‐α
Other bias	Low risk	The authors did not declare any financial support. "The authors declare that they have no competing interests."

Hu 2016.

Study characteristics
Methods	Single‐centre, randomised, prospective, and double‐blinded study Run‐in period: between March 2012 and May 2014 Registration date: not available, published in 2016 Number of study centres and locations: single centre, authors from Central‐South University, Changsha, Hunan, China
Participants	A total of 201 participants were assessed. Mean age: 47 years Sex (female/male ratio): 164% No clear description of risk for patients included (AVR, MVR, AVR + MVR) Inclusion criteria: patients with rheumatic valvular diseases admitted for valve replacement surgery Exclusion criteria: CAD, previous heart surgery history, need for AF ablation, infective endocarditis, peripheral vascular diseases, hypertension, diabetes or abnormal hepatic, and renal or pulmonary function
Interventions	Intervention: RIPerc. Cuff was immediately inflated to 600 mmHg – in accordance with the orthopaedic standard for blocking the blood flow of the lower limb – for 5 minutes followed by a 5‐minute reperfusion interval, and it was repeated three times. Control: no RIPerc; the cuff was only slightly inflated.
Outcomes	Cardiac troponins, acute lung injury, postoperative drainage, biomarkers for renal injury, and systemic inflammation Biomarkers that indicate the extent of injury to lungs, liver, and kidneys were measured within 48 hours after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given: "performed a single‐center, randomized, prospective, and double‐blinded study"
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded study: "We performed a single‐center, randomized, prospective, and double‐blinded study." "Patients, surgeons, perfusionists, intensive care unit staff, and laboratory technicians were blinded to which group the patients were allocated."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Patients, surgeons, perfusionists, ICU staff, and laboratory technicians were blinded as to which group the participants were allocated.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed: "Disclosure: None"

Hua 2017.

Study characteristics
Methods	Randomised controlled trial Run‐in period: February 2013‐December 2014 Registration date: not available. Received 4 June 2017, revised 8 September 2017, accepted 26 September 2017, published 8 November 2017 Number of centres and location: single centre. Sun Yat‐sen Memorial Hospital, Sun Yat‐sen University, Guangzhou 510120, China
Participants	A total of 130 participants were assessed in the study. Inclusion criteria: patients with valvulopathy who underwent surgery with CPB in the department of cardiac surgery at our hospital from February 2013 to December 2014 Exclusion criteria: coronary heart disease; allergy to simvastatin; active hepatitis, severe renal impairment (serum Cr > 200 mol/L), or both; pregnancy or lactation; taking medication(s) that may increase the effective concentration of simvastatin (e.g. a tetralol‐derived calcium‐channel blocker such as mibefradil); and hyperlipidaemia.
Interventions	Preoperative simvastatin. Participants in the control group were given placebo.
Outcomes	Inflammatory indices (TNF‐α, IL‐6, and IL‐8) 6 hours pre‐ and postoperatively and cardiac muscle impairment indices (serum cTnT and CK‐MB) 6 hours, 12 hours, 24 hours, and 72 hours pre‐ and postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"by assigning random numbers from a random number table"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	High risk	Twenty participants failed to complete the study and were not included in the results.
Selective reporting (reporting bias)	High risk	Figures for inflammatory markers not reported
Other bias	Low risk	"The authors declare that they have no conflicts of interest."

Huang 2011 (A).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not stated Registration date: not stated. Published 2011 Number of centres and locations: not stated
Participants	One hundred and twenty ASA class II to III patients, aged 53–76 years, presenting for scheduled primary elective CABG were assigned according to a computer‐generated random code to one of four groups: a control group receiving midazolam and fentanyl (group C; n = 30); propofol alone (group P; n = 30); isoflurane alone (group I; n = 30); and a combination of isoflurane and propofol (group IP; n = 30). In this analysis we present the findings for group P. Age (mean): 62.17 Sex (female/male ratio): 23.9% High‐risk patients
Interventions	Induction of anaesthesia was standardised for all groups: etomidate at 0.3 mg/kg of body weight, fentanyl at 8 μg/kg of body weight and pancuronium bromide at 0.1 mg/kg of body weight intravenously. After induction, all patients received continuous infusions of fentanyl at 0.6μg/ kg of body weight/min and pancuronium bromide at 15 μg/kg of body weight/min during surgery. Group P: maintenance of anaesthesia with propofol at 100 μg /kg of body weight/ min before and during CPB followed by propofol at 60 μg/kg of body weight /min for 15 min after aortic declamping Group C (control): maintenance of anaesthesia with fentanyl and midazolam
Outcomes	Haemodynamic data: MAP, HR, CVP, PCWP perioperative plasma TNF‐α, IL‐6, SOD, and MDA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"computer‐generated random code"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Incomplete outcome data (attrition bias) All outcomes	Low risk	One participant's data were not included in analysis because of death.
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"Funding: This study is supported, in part, by the National Natural Science Foundation of China (NSFC) [grant number 30872447, 30672033] and by a Society of Cardiovascular Anesthesiologists (SCA) Foundation Starter grant (to Z.X.)."

Huang 2011 (B).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not stated Registration date: not stated. Published 2011 Number of centres and locations: not stated
Participants	One hundred and twenty ASA class II to III patients, aged 53–76 years, presenting for scheduled primary elective CABG were assigned according to a computer‐generated random code to one of four groups: a control group receiving midazolam and fentanyl (group C; n = 30); propofol alone (group P; n = 30); isoflurane alone (group I; n = 30); and a combination of isoflurane and propofol (group IP; n = 30). In this analysis we present the findings for group I. Age (mean): 62.17 Sex (female/male ratio): 23.9% High‐risk patients
Interventions	Induction of anaesthesia was standardised for all groups: etomidate at 0.3 mg/kg of body weight, fentanyl at 8 μg/kg of body weight and pancuronium bromide at 0.1 mg/kg of body weight intravenously. After induction, all patients received continuous infusions of fentanyl at 0.6μg/ kg of body weight/min and pancuronium bromide at 15 μg/kg of body weight/min during surgery. Group I: maintenance of anaesthesia with isoflurane 1–1.5 MAC end‐tidal throughout the surgery Group C (control): maintenance of anaesthesia with fentanyl and midazolam
Outcomes	Haemodynamic data: MAP, HR, CVP, PCWP perioperative plasma TNF‐α, IL‐6, SOD, and MDA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"computer‐generated random code"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Two participants' data were not included in analysis because of death.
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"Funding: This study is supported, in part, by the National Natural Science Foundation of China (NSFC) [grant number 30872447, 30672033] and by a Society of Cardiovascular Anesthesiologists (SCA) Foundation Starter grant (to Z.X.)."

Huang 2011 (C).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not stated Registration date: not stated. Published 2011 Number of centres and locations: not stated
Participants	One hundred and twenty ASA class II to III patients, aged 53–76 years, presenting for scheduled primary elective CABG were assigned according to a computer‐generated random code to one of four groups: a control group receiving midazolam and fentanyl (group C; n = 30); propofol alone (group P; n = 30); isoflurane alone (group I; n = 30); and a combination of isoflurane and propofol (group IP; n = 30). In this analysis we present the findings for group IP. Age (mean): 62.17 Sex (female/male ratio): 23.9% High‐risk patients
Interventions	Induction of anaesthesia was standardised for all groups: etomidate at 0.3 mg/kg of body weight, fentanyl at 8 μg/kg of body weight and pancuronium bromide at 0.1 mg/kg of body weight intravenously. After induction, all patients received continuous infusions of fentanyl at 0.6μg/ kg of body weight/min and pancuronium bromide at 15 μg/kg of body weight/min during surgery. Group IP: maintenance of anaesthesia with isoflurane 1–1.5 MAC end‐tidal before CPB and switched to propofol at 100μg/kg of body weight/min during CPB followed by propofol 60 μg /kg of body weight /min 15 min after aortic declamping Group C (control): maintenance of anaesthesia with fentanyl and midazolam
Outcomes	Haemodynamic data: MAP, HR, CVP, PCWP perioperative plasma TNF‐α, IL‐6, SOD, and MDA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"computer‐generated random code"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The surgeons, research assistants and medical and nursing staff in the operation room were blinded to the group assignments, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Three participants' data were not included in analysis because of death.
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"Funding: This study is supported, in part, by the National Natural Science Foundation of China (NSFC) [grant number 30872447, 30672033] and by a Society of Cardiovascular Anesthesiologists (SCA) Foundation Starter grant (to Z.X.)."

Hurst 1997.

Study characteristics
Methods	Randomised controlled trial Run‐in period: June 1993‐June 1994. Published 1997 Number of centres and location: not specified
Participants	In total, 24 participants were assessed. Inclusion criteria: patients scheduled for elective open‐heart valve surgery between July 1993 and June 1994 Exclusion criteria: patients were excluded if we could not obtain informed consent or if they were undergoing emergency surgery.
Interventions	Intervention: Neutrophil depletion filter Control: Standard filter
Outcomes	PAO2, cardiac index, EF, haemodynamic variables, use of inotropes, spirometry (FEV1, FVC), and hospitalisation duration
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	"The study was supported in part by a grant from the Saskatchewan Lung Association."

Huybregts 2007.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Published 2007 Number of centres and location: not specified
Participants	A total of 49 participants were assessed. Inclusion criteria: elective first‐time CABG, aged 45 years to 75 years, with an LVEF > 40% Exclusion criteria: patients with preoperative immunosuppressive therapy, preoperative use of nonsteroidal anti‐inflammatory drug, requiring intra‐aortic balloon support or aneurysmectomy, with insulin‐dependent diabetes mellitus, or with a plasma Cr level > 150 mmol/L
Interventions	Intervention: Mini‐CPB system Control: Conventional CPB
Outcomes	Urinary IL‐6, urine N‐acetyl‐glucosaminidase, urinary intestinal fatty acid binding protein, urinary thromboxane B2
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	"Pseudo double‐blind (masking of the laboratory personnel)"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	Not clearly stated

Ikuta 2004 (A).

Study characteristics
Methods	Randomised controlled trial Run‐in period: April 2001‐February 2002 Published 2004 Number of centres and location: not specified
Participants	Twenty‐three participants were assessed in part A of this study. Inclusion criteria: first‐time elective CABG Exclusion criteria: previous cardiac surgery, renal or liver dysfunction, and preoperative coagulopathy
Interventions	Intervention: Heparin‐coated CPB circuit Control: Uncoated circuit.
Outcomes	Plt preservation by number of Plts C activation by C3a and C4a levels Inflammatory response by IL‐6 and IL‐8 levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Forty‐five patients undergoing CABG were randomly assigned to PMEA‐coated (group P, n = 15), heparin‐coated (group H, n = 15), and noncoated (group N, n = 15) circuit groups.
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	Not clearly stated

Ikuta 2004 (B).

Study characteristics
Methods	Randomised controlled trial Run‐in period: April 2001‐February 2002 Number of centres and location: unspecified
Participants	Twenty‐two participants were assessed in part B of this study. Inclusion criteria: first‐time elective CABG Exclusion criteria: previous cardiac surgery, renal or liver dysfunction, and preoperative coagulopathy
Interventions	Intervention: PMEA‐coated CPB circuit Control: uncoated circuit
Outcomes	Plt preservation by number of Plts C activation by C3a and C4a levels Inflammatory response by IL‐6 and IL‐8 levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	Not clearly stated

Ilmakunnas 2005.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received for publication Feb 13, 2004; revisions received June 29, 2004; accepted for publication July 14, 2004
Participants	A total of 50 participants were assessed in this study. Inclusion criteria: patients undergoing elective CABG Exclusion criteria: patients with significant valvular dysfunction or any renal, liver, or pancreatic disease were excluded.
Interventions	Intervention: Leucocyte‐depleting filter (Pall LeukoGuard 6 leukocyte‐depleting filter [LG6; Pall Biomedical, Portsmouth, United Kingdom] during CPB) Control: no arterial line filter
Outcomes	Monocyte CD11b expression, neutrophil hydrogen peroxide production, and lactoferrin plasma levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants undergoing elective CABG were randomly allocated into two groups: the filter group (n = 25) with a pall LeukoGuard 6 leucocyte‐depleting filter (LG6; Pall Biomedical, Portsmouth, UK) used during CPB and the control group (n = 25) with no arterial line filter.
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	Supported by grants from the Clinical Research Institute of the University of Helsinki Central Hospital, Helsinki, Finland

Immer 2007.

Study characteristics
Methods	Randomised substudy – cohort of larger patients Run‐in period: not specified. Accepted for publication May 29, 2007 Number of centres and location: University Hospital, Berne, Switzerland
Participants	A total of 1053 patients were assessed in this study; 60 participants were randomised to MiECC or CCPB circuit. Inclusion criteria: patients undergoing CABG Exclusion criteria: re‐do CABG and combined CABG operations (with valve surgery, aortic surgery, or both)
Interventions	Intervention: MiECC Control: CCPB
Outcomes	cTnI, CK‐MB, IL‐6, SC5b‐9, human lactoferrin (HL)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Inui 1999.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received November 1998; revised February 1999. Published 1999 Number of study centres and location: not specified
Participants	Twenty‐four participants were assessed in this study. Inclusion criteria: adult patients undergoing cardiac surgery
Interventions	Intervention: Duraflo II heparin‐coated CPB circuits Control: uncoated circuit
Outcomes	TNF‐α, IL‐6, IL‐8, MPO, TAT, PIC, TM
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Twenty‐four patients were randomly assigned to be investigated. Nine participants underwent CABG, and 15 underwent valve replacement or valve repair surgery.
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Iskesen 2009.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received December 23, 2008; accepted February 13, 2009. Published 2009 Number of centres and location: not specified
Participants	A total of 40 participants were assessed in this study. Inclusion criteria: patients undergoing elective CABG surgery Exclusion criteria: severe LV dysfunction (defined as an LVEF of < 30% or an end‐diastolic pressure > 16 mmHg); emergency or re‐do operation; recent MI in the last 4 weeks; pulmonary disease; renal or hepatic dysfunction; insulin‐dependent diabetes; age > 80 years; WBC count > 10,000/mL; infection during the week preceding surgery; preoperative use of antibiotics, corticosteroids, nonsteroidal anti‐inflammatory drugs, and aspirin; or smoking in the last month
Interventions	Intervention: oral PTX 400 mg twice a day (800 mg/day, orally) for the last 5 days preoperatively Control: no PTX
Outcomes	TNF‐α, IL‐6, and IL‐8, CK‐MB and CK Time points: before induction of anaesthesia (T0), 10 minutes after placement of aortic cross‐clamping (during the cross‐clamp period) (T1), 10 minutes after the end of CPB (T2), 4 hours after skin closure (T3), and at the 24th hour postoperatively (T4)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	In a prospective, randomised study, 40 participants undergoing CABG surgery received either PTX (study group, n = 21) or not (control group, n = 19).
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"double‐blinded study in patients undergoing coronary artery" "The surgical team and biochemical analysts were blinded with respect to the study medication."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	This study was supported by a grant from the "Celal Bayar University Scientific Investigation Fund".

Jansen 1991.

Study characteristics
Methods	Placebo‐controlled, double‐blind, prospective, randomised study Run‐in period: not specified Number of centres and location: not specified
Participants	Twenty‐five patients undergoing elective CABG procedures were studied in a randomised double‐blind trial. Patients with insulin‐dependent diabetes mellitus, preoperative use of corticosteroids, or chronic obstructive lung disease were not included. Exclusion criteria were a CPB time longer than 180 minutes, use of an IABP, and reoperation.
Interventions	Intervention: Dexamethasone 1 mg/kg after induction of anaesthesia Control: placebo (0.9% sodium chloride)
Outcomes	C3a, elastase, TNF, LTB4, tPA, leucocytes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how participants were randomised
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment but states double blinded
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The study was performed in a double‐blind, placebo‐controlled manner, but no description of blinding process is given.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The study was performed in a double‐blind, placebo‐controlled manner, but no description of blinding process is given.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	No financial support declared

Jansen 1995.

Study characteristics
Methods	Randomised controlled trial Run‐in period: Received for publication July 25, 1994. Accepted for publication Dec. 23, 1994 Number of centres and location: not specified
Participants	A total of 30 participants were included in this study. Inclusion criteria: adult patients scheduled for cardiac surgery with CPB Exclusion criteria: patients with endocarditis and those admitted with cardiogenic shock or with intra‐aortic counterpulsation balloons
Interventions	Intervention: heparin‐coated CPB circuits Control: uncoated CPB circuits
Outcomes	Complement activation and PMN‐Elastase
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"Physicians involved in patient care and biochemists were blinded for randomization."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	"Supported in part by Bentley Laboratories Europe BV, Uden, The Netherlands."

Jimenez 2007.

Study characteristics
Methods	Randomised controlled trial Run‐in period: February to May 2004. Received: 17 July 2006. Revisions received: 25 May 2007 Accepted: 7 November 2007 Published: 7 November 2007 Number of centres and location: single centre. Hospital Universitario de Canarias, Ofra s/n La Cuesta, La Laguna, 38320, Spain
Participants	A total of 50 participants were assessed in this study. Inclusion criteria: patients undergoing elective first‐time CABG; LVEDP less than 25 mmHg and acetylsalicylic acid stopped for > 3 days Exclusion criteria: emergency interventions, patients with a history of chronic coagulopathy (PT < 50% or INR > 2 and Plts < 50,000/mm3 or aggregation dysfunction), renal failure (Cr> 2 mg/dL), chronic hepatopathy (Child B or higher degree), use of immunosuppressant drugs, endocarditis, sepsis in the first 24 hours after intervention, or unwillingness to enrol
Interventions	Intervention: TXA 2 g pre‐CPB and post‐CPB after protamine administration Control: placebo (0.9% sodium chloride)
Outcomes	Soluble tumour necrosis factor receptor (STNFR)‐1, IL‐6, quantitative PAI‐1 antigen and tPA antigen levels Core body temperature, biochemical determinations (haematology, inflammation, coagulation, and fibrinolysis), and haemodynamic parameters were recorded before intervention (baseline), on admission to the ICU after surgery (0 hours), and at 4 hours and 24 hours after intervention, blood loss measured by tube chest drainage and the amount of hemoderivatives used
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomly assigned by independent pharmacists using a list of pseudorandomized numbers"
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The code was revealed once recruitment, data collection, and laboratory analyses were completed."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The code was revealed once recruitment, data collection, and laboratory analyses were completed."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	This study was supported by FUNCIS (Fundación Canaria de Investigación y Salud) 2202.

Jin 2019.

Study characteristics
Methods	Randomised controlled trial Run‐in period: July 2012 to July 2015. Received February 5, 2018; accepted December 28, 2018, published 2019 Number of centres and location: single centre. First Affiliated Hospital of Wenzhou Medical University (Wenzhou, China)
Participants	A total of 241 participants were assessed in the study. Inclusion criteria: patients who received valve surgery at the First Affiliated Hospital of Wenzhou Medical University; mitral valve disease, aortic valve disease, or combined valvular disease; and patients with stable haemodynamic blood Exclusion criteria: infection, chronic lung disease, medications that may interfere with RIPC, pregnancy, renal disease, cardiac arrest during hospital admission and peripheral arterial disease affecting the limbs, complicated CAD, complicated hypertension, congenital heart valve disease, preoperative stroke, simultaneous radiofrequency ablation of AF, and reoperation
Interventions	Interventions: RIPC: two 5‐min cycles of simultaneous upper arm and thigh cuff inflation and deflation (simultaneous inflation to 200 mmHg, left inflation for 5 min and then deflation to 0 mmHg and left deflated for 5 min) Control: no preconditioning
Outcomes	Postoperative MI detected using hs‐cTnT, sICAM‐1, ET‐1, MDA, NO, blood gas indexes, ALIand length of ICU stay, length of hospital stay and major adverse cardiovascular events at 90 days (death, myocardial infarction or stroke) Blood samples were collected preoperatively (T1) and at 5 min (T2), 2 h (T3), 6 h (T4) and 24 h (T5) after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	"The investigators who analyzed the data were blinded to the treatment allocation."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given other than "The investigators who analyzed the data were blinded to the treatment allocation"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	The authors declare that they have no competing interests.

Johansson‐Synnergren 2004.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received October 4, 2003; accepted November 30, 2003 Number of centres and location: not specified. Authors from the Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Gőteborg, Sweden
Participants	A total of 52 participants were assessed in this study. Inclusion criteria: patients scheduled for CABG between January 1999 and January 2001 Exclusion criteria: stenoses in the distal circumflex area, LVEF < 30%, preoperative stroke, known peripheral vascular disease, infection, or preoperative treatment with anti‐inflammatory drugs
Interventions	Intervention: OPCAB Control: ONCAB
Outcomes	Plasma levels of complement (C3a), cytokines (IL‐8, TNF‐α), ET‐1, and neopterin; blood flow in the forearm
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis.
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Johnell 2002.

Study characteristics
Methods	Randomised controlled trial Run‐in period: Not specified. Received August 14, 2002; accepted September 4, 2002. Published 2022 Number of study centres and location: not specified. Authors from University Hospital, Uppsala, Sweden
Participants	A total of 60 participants were assessed in this study. Inclusion criteria: patients under the age of 75 scheduled for elective CABG Exclusion criteria: patients with inflammatory or renal disease or severely low EFs
Interventions	Intervention: Corline heparin surface (CHS). The CHS was applied to a complete set of an extracorporeal circuit (tubing, cannula, oxygenator and reservoir). The CHS utilises a unique macromolecular conjugate in which approximately 70 heparin molecules are covalently linked to a polymer carrier with a molecular weight of 50 kD. The CHS coating comprises a conditioning layer of a polymeric cationic amine on to which the macromolecular conjugate is irreversibly attached by multiple ionic interactions. Control: Uncoated circuit
Outcomes	Cell count, TF and CD11b expression on monocytes, and monocytic TFmRNA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	"The postoperative course and mobilization was uneventful in all but one patient, included in group 1, who died the day after surgery due to myocardial failure. This patient was excluded from the study."
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"This study was supported by the Swedish Foundation for Strategic Research and the Swedish Medical Research Council."

Jongman 2013.

Study characteristics
Methods	Prospective, randomised trial Run‐in period: June‐ December 2011 Number of centres and location: single centre. University Medical Center Groningen, University of Groningen
Participants	A total of 59 participants were assessed in this study. Inclusion criteria: patients older than 18 years with CAD who were suitable for both on‐pump CABG and OPCABG surgery Exclusion criteria: patients with a history of head trauma or stroke, history of neurosurgery, severe or symptomatic carotid artery disease, requirement for valve surgery in addition to CABG, pre‐existing acute or chronic renal dysfunction, urgent or emergency surgery, or difficulty with testing for cognitive dysfunction
Interventions	Intervention: OPCAB. In the OPCAB group, cardiac stabilisation and displacement were obtained using Acrobat and XPOSE 4 devices (Maquet Netherlands BV, Hilversum, The Netherlands). Control: ONCAB. In this group, the ascending aorta and right atrium were cannulated and CPB with non‐pulsatile flow was started. To induce and maintain cardiac arrest, myocardial protection was obtained using antegrade administration of cold blood cardioplegia through the aortic root and/or retrograde administration via the coronary sinus according to the patient’s anatomy.
Outcomes	TNF‐α, IL‐6, IL‐10, MPO, soluble E‐selectin, soluble VCAM‐1, and sICAM‐1, Ang1, Ang2, and soluble Tie2, VEGF, soluble VEGFR1 and VEGFR2. Blood was drawn from the arterial catheter at four time points, after the arrival of the patient in the operation room (start), after sternum closure (end), 6 h after closure, and 24 h postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Low risk	Patients were randomised to either on‐pump or off‐pump surgery using a sealed envelope technique.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	"One patient who was randomized for on‐pump CABG surgery was excluded from the study."
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"Conflict of interest: None declared" "Competing interests: The authors declare that they have no competing interests."

Joyeux‐Faure 2011.

Study characteristics
Methods	Randomised, placebo‐controlled trial Run‐in period: Received 7 April 2011; revised 20 June 2011; accepted 30 August 2011 Number of centres and locations: single centre. Grenoble University Hospital, Grenoble, France
Participants	A total of 50 participants were assessed in this study. Inclusion criteria: patients scheduled for elective nonemergency CABG surgery at the department of cardiac surgery (Grenoble University Hospital, France), aged > 18 years and < 80 years, with LVEF > 40% and who gave a written informed consent Exclusion criteria: nonisolated coronary surgery such as valvular surgery or associated carotid graft surgery or surgery without CPB or with beating heart, history of prior stroke or infarction (of < 1 month), history of cardiac surgery, renal failure (Cr level > 200 μM), uncontrolled hypertension, unstable angina, risk of deep venous thrombosis, cancer, allergy to EPO, phenylketonuria, pregnant or lactating women, and history of planned blood donation
Interventions	Intervention: erythropoietin. 800 IU/kg epoietin beta (NeoRecormon; Roche Laboratories, Neuilly‐sur‐Seine, France), slowly administered intravenously (in 60 mL, in 30 min) 1–3 h before CPB beginning, at the anaesthesia induction Control: placebo (60 mL saline) administered at the same timing
Outcomes	TnT, NT‐proBNP, CK‐MB, S100b protein, circulating EPO, TNF‐α, IL‐6, and IL‐10 at T6. The following clinical safety parameters were also collected: mortality during hospitalisation, the volume of postoperative bleeding, the need for surgical re‐intervention in the 15 days following CABG, and the need for positive inotropic drug administration in the 48 h following the surgery. Blood samples were collected at the induction of anaesthesia (Ti for baseline), just after the end of CPB (T0) and 6 h (T6), 12 h (T12), 24 h (T24), and 48 h (T48) later.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The randomisation key was prepared by a statistician independent of the study and held by an independent pharmacist who prepared the administered treatment.
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"Investigators, patients, and the medical team were blinded to treatment allocation."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"Investigators, patients, and the medical team were blinded to treatment allocation."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in analysis. Two participants were excluded from analysis after randomisation.
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	This study was supported by a grant from the regional 'Programme Hospitalier de Recherche Clinique' in Grenoble, France. We would like to thank Roche Laboratories who kindly provided the epoietin beta (NeoRecormon) used in this study and its subsidiary company, Roche Diagnostics Laboratories, which kindly provided all kit reagents used for the measurement of cardiac and cerebral ischaemic and inflammatory markers.

Kaminishi 2004.

Study characteristics
Methods	Prospective, randomised, blinded study Run‐in period: not specified Number of study centres and location: single centre. Hitachi General Hospital, Ibaraki, Japan.
Participants	Eighteen patients undergoing coronary surgery
Interventions	Aprotinin (heparin plus aprotinin, 2 × 104 KIU/kg; n = 6) Three groups: the control group (heparin, 4 mg/kg; n = 6); the nafamostat mesilate group (heparin plus nafamostat, 0.2 mg/kg bolus followed by 2 mg/kg/h during CPB; n = 6); and the aprotinin group (heparin plus aprotinin, 2 × 104 KIU/kg; n = 6). Nafamostat group excluded as not within protocol
Outcomes	Plt count, Plt aggregation, thromboglobulin, PF‐1.2, TAT, plasminogen activator inhibitor‐1, 2‐plasmin inhibitor‐plasmin complex, D‐dimer, neutrophil elastase, and IL‐6 were measured before, during, and after bypass. Bleeding times and blood loss were recorded. Blood samples were obtained for analysis before heparin administration (baseline), at 60 minutes of CPB, at 30 minutes after protamine administration (protamine), on admission to the ICU, and at 12 hours after surgery (12 hours).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how patients were randomised
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	Nineteen consecutive participants included, outcomes reported for 18 participants only
Selective reporting (reporting bias)	High risk	Nineteen consecutive participants included, outcomes reported for 18 participants only
Other bias	Low risk	This work was supported by The University of Tsukuba Research Project in 2001.

Karuppasamy 2011.

Study characteristics
Methods	Randomised controlled trial Run‐in period: October 2008 to September 2009. Received: 10 December 2010; revised: 28 March 2011; accepted: 20 April 2011; published online: 5 May 2011 Number of centres and location: not specified. Authors from King’s College Hospital NHS Foundation Trust, Denmark Hill, London SE5 9RS, United Kingdom
Participants	A total of 54 participants were assessed in this study. Inclusion criteria: adult patients referred for elective CABG surgery without the concomitant valve or aortic surgery Exclusion criteria: participants > 85 years, with unstable angina, significant hepatic, pulmonary or renal disease, and those with peripheral vascular disease in the upper limbs were excluded. We also excluded patients taking sulfonylureas or other drugs that may activate or inhibit preconditioning.
Interventions	Intervention: remote intermittent ischaemia: three 5‐min cycles of left upper arm ischaemia, which was induced by a blood pressure cuff inflated to 200 mmHg. Control: placebo. A deflated cuff was placed on the upper left arm for 30 min.
Outcomes	cTnI, BNP, CK‐MB,IGF‐1, CK‐MB, myostatin, INF‐γ, IL‐2, IL‐4, IL‐6, IL‐8, IL‐10, IL‐1α, IL‐1β, TNF‐α and MCP‐1. Blood samples were collected preoperatively and after 6 hours, 12 hours, 24 hours, and 48 hours.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Low risk	"Patients were randomised by sealed envelopes."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and surgeons were blinded to the treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	"All completed the study with no patients lost to follow up."
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	This study was supported by a grant from the Department of Research and Development, King’s College Hospital Foundation Trust, London. The authors declare that they have no conflict of interest.

Kazemi 2013.

Study characteristics
Methods	Randomised controlled trial Run‐in period: August 2010‐September 2011. Received January 14, 2013, revised March 23, 2013, accepted April 18, 2013 Number of study centres and location: not specified. Authors from Tabriz University of Medical Sciences, Tabriz, Iran
Participants	A total of 240 participants were assessed in this study. Inclusion criteria: patients scheduled for their first open‐heart surgery Exclusion criteria: age < 18 years (n = 2), thyroid dysfunction (n = 1), left atrial size > 55 mm in M‐mode, parasternal long axis view (n = 9), previous open‐heart surgery (n = 14), emergent surgery (n = 18), documented history of AF or atrial flutter before the operation (n = 19), and incomplete monitoring (n = 20). We also excluded patients with LVEF ≤ 25% (n = 54) or NYHA Functional Class ≥ III (n = 21) because of an increased tendency to develop AF.
Interventions	Intervention: NAC 1,200 mg orally from 48 hours before and up to 72 hours after open‐heart surgery Control: placebo at the same time
Outcomes	The primary endpoint was the incidence of POAF on NAC or placebo treatment. The secondary endpoints were the postoperative morbidity, total mortality, ICU, and hospital stay.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	No details given
Allocation concealment (selection bias)	Low risk	Randomisation was performed by a nurse in the operating room with closed and opaque envelopes containing the study assignment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Two experienced electrophysiologists (B.K. and F.A.), blinded to the participant's randomisation group, analysed all arrhythmias, their duration, time of occurrence, and treatments and collected study data. This was a prospective, randomised, double‐blind, placebo‐controlled trial. The participants or physicians carrying out the study or collecting data were completely blinded to treatment group allocations.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Two experienced electrophysiologists (B.K. and F.A.), blinded to the participant's randomisation group, analysed all arrhythmias, their duration, time of occurrence, and treatments and collected study data. This was a prospective, randomised, double‐blind, placebo‐controlled trial. The participants or physicians carrying out the study or collecting data were completely blinded to treatment group allocations.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Low risk	"Conflict of Interest: None"

Kiaii 2012.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Accepted for publication February 2012 Number of centres and location: not specified. Authors from London Health Sciences Centre, University Hospital, 339 Windermere Road, London, ON, Canada
Participants	A total of 60 participants were assessed in this study. Inclusion criteria included age > 18 years, EF > 30%, ≥ 2 CADs, and suitability for on‐pump coronary artery revascularisation. Exclusion criteria were an inability to give informed consent, emergency operations, use of immunosuppressive drugs including steroids within 4 weeks before surgery, use of nonsteroidal anti‐inflammatory drugs 1 week before surgery, serum Cr level > 200 μmol/L, previous cardiac operation, and concomitant valve operations.
Interventions	Intervention: mini CPB Control: conventional CPB
Outcomes	Changes in levels of plasma proteins, including inflammatory cytokines (IL‐6, IL‐10, and TNF‐α), chemokines (IL‐8, MIG, RANTES, MCP‐1, and IP‐10) and acute phase proteins (CRP and C3)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	Surgeon could not have been blinded to the treatment group.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kiessling 2018.

Study characteristics
Methods	Randomised controlled trial: Run‐in period: March‐June 2016 Number of study centres and location: single centre. Johann Wolfgang Goethe University Hospital, Frankfurt, Germany
Participants	A total of 72 participants were assessed in this study. Inclusion criteria: planned CABG procedures with ECC support and age > 64 years Exclusion criteria: CABG procedures in combination with operations on the carotid arteries, aorta, or valves. Patients who underwent a reoperation, had an AMI (ST‐elevation MI or non‐ST‐elevation MI < 7 days ago), had elevated serum Cr levels (> 1.8 mg/dL), had elevated liver enzymes (AST, ALT > 2× above norm), or had decreased Hb (Hb < 11 mg/dL) were also excluded.
Interventions	Intervention: MiECC system Control: standard CPB
Outcomes	Renal function, inflammatory response, ischaemia, coagulation, and haemolysis and clinical data
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kilger 2003.

Study characteristics
Methods	Randomised controlled trial Run‐in period: Number of centres and location: not specified. Authors from Grosshadern University Hospital, LM‐University, Munich, Germany
Participants	A total of 91 participants were assessed in this study. Inclusion criteria: planned CABG with ≥ 4 or more grafts; planned valve surgery combined with CABG; or preoperative EF of < 40%. Exclusion criteria: EF < 35% was an exclusion criterion in this study. Participants were also excluded if pregnancy, IL‐6 concentrations > 10 pg/mL preoperatively, hepatic insufficiency (bilirubin > 3 mg/dL), renal insufficiency (Cr > 2 mg/dL), a positive serologic test for HIV, manifest insulin‐dependent diabetes mellitus, adipositas permagna (BMI > 30), use of steroidal or nonsteroidal antiphlogistics during the last 7 days (except 100 mg of acetylsalicylic acid per day), an extracardial septic focus, or chronic or acute inflammatory diseases.
Interventions	Intervention: hydrocortisone. Loading dose (100 mg intravenously for 10 minutes) before induction of anaesthesia, followed by a continuous infusion of 10 mg/h for 24 hours (POD 1), which was reduced to 5 mg/h on POD 2, then tapered to 3 × 20 mg on POD 3 and 3 × 10 mg on POD 4 Control: placebo (0.9% sodium chloride)
Outcomes	IL‐6 before induction of anaesthesia and 6 hrs after termination of CPB (pg/mL), maximum dose of norepinephrine and epinephrine (g/kg) during the first 24 h duration of vasopressor or inotropic support, PaO2/FIO2 ratio after admission at the ICU, duration of mechanical ventilation, need for transfusion, lowest serum concentration of ATIII during the first 24 h, postoperative fluid requirement, postoperative blood loss, peak serum concentration of lactic acid during the first 24 h lowest pH value in the blood during the first 24 h, Simplified Acute Physiology Score II during the first 24 h in the ICU, Therapeutic Intervention Scoring System (TISS) daily during the stay in the ICU, length of stay (LOS) in the ICU and in hospital, mortality rate until discharge from the hospital
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not in a double‐blind manner
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kilger 2011.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received on September 3, 2010, accepted for publication on November 19, 2010 Number of centres and location: not specified. Authors from Grosshadern University Hospital, LM‐University, Munich, Germany
Participants	A total of 305 participants were assessed in this study. Inclusion criteria: patients undergoing CABG, able to give informed consent The exclusion criteria were the following: age < 18 years; pregnancy; preoperative IL‐6 levels > 10 pg/mL; hepatic insufficiency (bilirubin > 3 mg/ dL); renal insufficiency (Cr > 2 mg/dL); a positive serologic test for HIV; manifest insulin‐dependent diabetes mellitus; adipositas permagna (BMI > 30); use of steroidal or nonsteroidal antiphlogistics during the last 7 days (except 100 mg acetylsalicylic acid per day); chronic or acute inflammatory diseases.
Interventions	Intervention: hydrocortisone. Loading dose (100 mg intravenously for 10 minutes) before induction of anaesthesia, followed by a continuous infusion of 10 mg/h for 24 hours (POD 1), which was reduced to 5 mg/h on POD 2, then tapered to 3.2 mg on POD 3 and 3.10 mg on POD 4 Control: placebo (0.9% sodium chloride)
Outcomes	Amount of vasopressor or inotropic therapy and maximal dose of norepinephrine, epinephrine and milrinoen administered during the first 48 hours; gas exchange; duration of mechanical ventilation; highest serum level of bilirubin and CK‐MB during the first 24 hours, length of stay at the ICU, the incidence of postoperative atrial fibrillation, re‐thoracotomy, re‐intubation, use of CPAP therapy and/or IABP; mortality during the first 30 days after surgery; any infections (wound infection, pneumonia, urinary tract infection, catheter‐associated infection, and bacteraemia) occurring from enrolment until discharge from the hospital
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	High risk	The study was performed in a prospective, randomised controlled trial without double blinding.
Blinding of outcome assessment (detection bias) All outcomes	High risk	The study was performed in a prospective, randomised controlled trial without double blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kilic 2009 (A).

Study characteristics
Methods	Randomised controlled trial Run‐in period: January to March 2007. Received 26 December 2007; returned for revision 12 July 2008; received from final revision 12 November 2008; accepted 16 November 2008. Published online first on 6 March 2009 Number of study centres and location: not specified. Authors from University of Kirikkale, School of Medicine, Kirikkale, Turkey
Participants	A total of 20 participants were assessed in part A of this study. Inclusion criteria: high‐risk patients (EuroSCORE 6+) undergoing CABG surgery Exclusion criteria: patients with preoperative immunosuppressive or nonsteroidal therapy; EF < 20%; heparin treatment during surgery; disseminated intravascular coagulation; preoperative treatment with steroids; severe pulmonary, renal, hepatic, or cerebrovascular disease; neoplasia; or infectious disease
Interventions	Intervention: Polyethyleneoxide (PEO) based heparin bonded ECC without leukofiltration. Control: Uncoated ECC
Outcomes	Hb, haematocrit, erythrocyte, WBC and Plt, standard blood and urine biochemistry; IL‐10, TNF‐α, procalcitonin, MMP 9, D‐dimer, CK‐MB Blood samples were obtained via radial artery catheter in potassium‐EDTA tubes at the following intervals : T1: Baseline: Following induction of anaesthesia (before administration of heparin) T2 : On CPB : 5 min. following initiation of CPB T3. X‐Clamp: 5 min. following cross‐clamping of the aorta T4 : Off CPB : 5 min. following cessation of CPB T5 : ICU24 : First postoperative day in intensive care unit at 8 :00 a.m. T6. ICU48 : Second postoperative day in intensive care unit at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Only participants blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The operating room and intensive care unit staff collecting data were blinded to the entire study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Unclear risk	IL‐10 data not provided
Other bias	Unclear risk	No details given

Kilic 2009 (B).

Study characteristics
Methods	Randomised controlled trial Run‐in period: January to March 2007. Received 26 December 2007; returned for revision 12 July 2008; received from final revision 12 November 2008; accepted 16 November 2008. Published online first on 6 March 2009 Number of study centres and location: not specified. Authors from University of Kirikkale, School of Medicine, Kirikkale, Turkey
Participants	A total of 20 participants were assessed in part B of this study. Inclusion criteria: high‐risk patients (EuroSCORE 6+) undergoing CABG surgery Exclusion criteria: patients with preoperative immunosuppressive or nonsteroidal therapy; EF < 20%; heparin treatment during surgery; disseminated intravascular coagulation; preoperative treatment with steroids; severe pulmonary, renal, hepatic, or cerebrovascular disease; neoplasia; or infectious disease
Interventions	Intervention: uncoated ECC + leukofiltration Control: uncoated ECC
Outcomes	Hb, haematocrit, erythrocyte, WBC and Plt, standard blood and urine biochemistry; IL‐10, TNF‐α, procalcitonin, MMP 9, D‐dimer, CK‐MB. Blood samples were obtained via radial artery catheter in potassium‐EDTA tubes at the following intervals : T1: Baseline: Following induction of anaesthesia (before administration of heparin) T2 : On CPB : 5 min. following initiation of CPB T3. X‐Clamp: 5 min. following cross clamping of the aorta T4 : Off CPB : 5 min. following cessation of CPB T5 : ICU24 : First postoperative day in intensive care unit at 8 :00 a.m. T6. ICU48 : Second postoperative day in intensive care unit at 8:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Only participants blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The operating room and intensive care unit staff collecting data were blinded to the entire study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Unclear risk	IL‐10 data not provided
Other bias	Unclear risk	No details given

Kilickan 2008 (A).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Number of centres and location: not specified
Participants	A total of 39 participants were included in part A of this study. Inclusion criteria: patients undergoing elective CABG with CPB Exclusion criteria: LVEF < 40%, diabetes, active gastropathic disorder; preoperative use of steroids and contraindication to steroid administration, contraindication to epidurals, systemic or local infection, preoperative signs of infection, chronic inflammatory disease, treatment with COX inhibitors, drugs inhibiting thrombocyte function within 7 days of operation
Interventions	Intervention: 6‐methylprednisolone 15 mg/kg intravenously 60 min before induction Control: conventional management
Outcomes	IL‐10, glucose, clinical outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kilickan 2008 (B).

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified Number of centres and location: not specified
Participants	Inclusion criteria: patients undergoing elective CABG with CPB Exclusion criteria: LVEF < 0.40, diabetes, active gastropathic disorder; preoperative use of steroids and contraindication to steroid administration, contraindication to epidurals, systemic or local infection, preoperative signs of infection, chronic inflammatory disease, treatment with COX inhibitors, drugs inhibiting thrombocyte function within 7 days of operation
Interventions	Intervention: 20 mg bupivacaine ibolus 60 min before induction followed by 20 mg/h bupivacaine infusion intraoperatively and postoperatively via epidural catheter Control: conventional management
Outcomes	IL‐10, glucose, clinical outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kipfer 2003.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Received 11 December 2002; accepted 10 February 2003 Number of study centres and locations: single centre. University Clinic for Cardiovascular Surgery, Inselspital, CH‐3010 Bern, Switzerland
Participants	A total of 30 participants were assessed in this study. Inclusion criteria: patients scheduled for CABG Exclusion criteria: previous cardiac surgery, LVEF < 40%, unstable angina, acetylsalicylic acid within the last 7 days, IV heparin within the last 3 days and therapy with oral anticoagulants, age > 75 years, recent CVA, severe pulmonary disorders, impairment of renal or liver function, active inflammatory diseases, or known coagulopathies
Interventions	Intervention: aprotinin 2Mio KIU added to the prime volume of CPB Control: no aprotinin
Outcomes	Morbidity, mortality, TAT, PAP, PAI, IL‐6, sTNF‐IIR, or SC5b‐9 Blood samples were collected at seven different time points: baseline(after induction of anaesthesia); 5 min after first delivery of cardioplegia; 5 min after aortic declamping; 10 min after administration of protamine sulfate; 2, 6 and 24h postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details given
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data included in analysis
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Kofidis 2008.

Study characteristics
Methods	Randomised controlled trial Run‐in period: not specified. Published 2008 Number of study centres and location: not specified
Participants	A total of 80 participants were assessed in this study. Inclusion criteria: patients destined for a CABG procedure because of advanced CAD Exclusion criteria: not given
Interventions	Intervention: Minimised ECC system Control: standard CPB
Outcomes	Intraoperative parameters: operation time, bypass time, cross‐clamp time, mean number of grafts. Postoperative parameters: ventilation time, ICU and hospital stay, chest‐tube drainage in the first 24 hours postoperatively, cell saver volume returned, requirements for packed erythrocytes, fresh frozen plasma and thrombocyte concentrates. Adverse post‐operative events: myocardial ischaemia, atrial fibrillation, re‐thoracotomy, respiratory failure, re‐intubation, stroke and mortality. Laboratory values were measured and compared between the two groups: CK (24 h), CK‐MB (24 h), troponin‐T at 8 h postoperatively, IL‐6 and IL‐8 at 24 h postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comparable patients destined for a CABG procedure because of advanced CAD were included in the pool, from which they were picked out randomly for either minimised ECC system or CCPB.
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No details given
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details given
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	"20 patient charts where [sic] not worked up at the time of the data processing."
Selective reporting (reporting bias)	Low risk	No reporting bias apparent
Other bias	Unclear risk	No details given

Koskenkari 2006.

Study characteristics
Methods	Prospective, randomised, and unblinded single‐centre clinical study Run‐in period: not specified. Accepted for publication 3 May 2006 Number of centres and location: single centre, Oulu University Hospital, Finland
Participants	Forty participants were included in this study. Inclusion criteria: patients with unstable angina pectoris scheduled for urgent coronary artery revascularisation Exclusion criteria: patients on insulin or diabetic medication, emergency operation and need for inotropic medication, significant end‐target organ failure, evident infection or markedly elevated CRP (> 80 mg/L), corticosteroid medication, elevated serum creatinine (> 150 mg/L) or other severe chronic disease Mean age 67.11
Interventions	Intervention: separate glucose and insulin infusion. After induction of anaesthesia, 30% glucose solution (1000 mL with 20 mmol KCl and 20 mmol Mg) at a rate of 1.5 mL/kg/h (0.45 g/kg/h) and a fast‐acting insulin infusion (5 IU/mL) in a dose of 1 IU/kg/h were started via a central venous line. Blood glucose level was targeted at 6.0 mmol/L to 8.0 mmol/L by adjusting the glucose infusion rate. In the high‐dose insulin group, blood glucose levels were checked at 15‐minute intervals during the first hour of surgery and after that every half hour until the end of the operation. Weaning from the insulin infusion was started gradually after admission to the ICU, and insulin was discontinued within 12 hours to 14 hours for all patients. The glucose infusion rate was decreased gradually in response to the blood glucose levels. Control: no insulin. The control group received saline (1000 mL with 20 mmol KCl and Mg 20 mmol) solution at a rate of 1.5 mL/kg/h, and blood glucose levels were checked once in an hour during the operation. The blood glucose level was targeted to 6.0 mmol/L to 8.0 mmol/L with the aid of short‐acting insulin 4 IU to 8 IU IV bolus doses or an insulin infusion (1 IU/mL) as necessary.
Outcomes	IL‐6, IL‐8, and CRP, cTnI, lactate, glucose. Clinical outcomes: Time to extubation, ICU and hospital stay, first POD weight gain and perioperative blood loss were recorded for all patients. Postoperative morbidity (cardiac, renal, pulmonary, neurological, gastroenterological)
Notes	No supplementary material online, median converted via RevMan calculator
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of random‐sequence generation
Allocation concealment (selection bias)	Unclear risk	No statement regarding allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unblinded outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Kosour 2016.

Study characteristics
Methods	Prospective, randomised, controlled trial. Run‐in period: not specified Number of centres and location: not specified. Authors from 2 institutions in Campinas, Brazil
Participants	Forty patients undergoing CABG Exclusion criteria: emergency cardiac surgery, recent MI, mechanical complications after MI, indication for mitral repair or other surgical procedure in addition to the planned CABG, unstable angina, age < 18, hyperglycaemia (4180 mL/dL), other inflammatory pathologies, pregnancy, renal insufficiency with a Cr > 2.0 mg/dL, liver insufficiency with a TB > 2.5 mg/dL and transaminase or alkaline phosphatase elevation > 2 x the higher limit of normal values, use of acetylsalicylic acid or oral anticoagulant in a period shorter than 7 days, GCS < 10, and recent bleeding of the digestive system
Interventions	Intervention: ultrafiltration during the entire CPB time. It was estimated that 1000 mL/h of ultrafiltrate was withdrawn. Control: standard care
Outcomes	Primary outcomes were evaluated to determine whether ultrafiltration had an influence on pulmonary dysfunction after surgery, mechanical ventilation time, and ICU hours. The secondary outcome was inflammatory response, which was assessed by serum levels of IL‐1β, IL‐6, IL‐8, and TNF‐α. Blood samples were analysed before induction of anaesthesia (T1); 5 minutes before the start of CPB (T2); 30 minutes after CPB (T3); and 6 hours (T4), 12 hours (T5), 24 hours (T6), 36 hours (T7), and 48 hours (T8) after the surgery.
Notes	No supplementary material online, median converted via RevMan calculator
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation technique
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded to surgeon and anaesthesiologist
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The physicians responsible for postoperative care were blinded in respect to the study group.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No evidence of funding bias

Koster 2004.

Study characteristics
Methods	Prospective, randomised, single‐centre study Run‐in period: not specified Number of study centres and location: single‐centre. Location not specified
Participants	Two hundred participants undergoing elective standard cardiac surgery Exclusion criteria: reoperations, combined procedures; only operations in normothermic CPB. No patient had warfarin or anti‐Plt therapy within 7 days before surgery.
Interventions	Intervention: heparin according to the results of the Hep‐HMS PlusTM plus aprotinin (bolus of 1 × 106 KIU immediately before initiation of CPB, 1 × 106 KIU in the priming solution of the CPB, and a continuous infusion of 250,000 KIU/h during CPB) Control: heparin was given according to the results of the Hep‐HMS PlusTM.
Outcomes	To evaluate the effects of aprotinin, administered according to a half‐dose regimen, on activation of the haemostatic/inflammation system during CPB when used with heparin‐level‐based heparin management, postoperative blood loss, haemostasis, and complement activation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Surgeons were blinded regarding the allocation of the participants to either of the groups.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Surgeons were blinded regarding the allocation of the participants to either of the groups.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Kudlova 2007.

Study characteristics
Methods	Prospective randomised control study Run‐in period: not specified. Received 10 July 2007; Accepted 4 September 2007 Number of centres and location: single centre. Charles University in Prague, School of Medicine and University Hospital in Hradec Kra'love, 500 05 Hradec Kralove, Czech Republic
Participants	Forty patients referred for first‐time CABG, elective patients Patients treated with anti‐inflammatory agents, either steroids or NSAID, were excluded from the study, as were patients with serum Cr ≥ 130 μmol/L or with hepatic disorders. No patients were known to suffer from concomitant malignancies. Patients with active infectious diseases are not admitted to elective CABG in our department.
Interventions	Intervention: off‐pump CABG Control: on‐pump CABG
Outcomes	Serum levels of LBP and sCD14 in cardiac surgical patients undergoing CABG either with the use of CPB (on pump) or without the use of CPB (off pump) To ascertain the sources of sCD14, levels of IL‐6, CRP, and long pentraxin (PTX3) were tested.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of randomisation apart from the abstract
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding (unlikely for the personnel)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	IL‐6 and PTX3 levels tested but not reported
Selective reporting (reporting bias)	High risk	IL‐6 and PTX3 not reported
Other bias	Low risk	No clear evidence of funding bias

Kumar 2020.

Study characteristics
Methods	Double‐blind, randomised, placebo‐controlled trial Run‐in period: September 2015‐September 2018. Trial registered October 2015 Number of study centres and location: single centre (tertiary cardiac surgery centre). Glenfield General Hospital, Leicester, UK
Participants	One hundred twenty‐nine randomised; five participants withdrew before surgery (two intervention, three placebo), five had major protocol deviations, and eight patients withdrew after surgery; after 3 months follow‐up, 117 participants Exclusion criteria included patients with pre‐existing AKI, sepsis, Stage 5 chronic kidney disease, severe hepatic impairment, allergy to PDE Type 5 inhibitors, or recent treatment with cytochrome P450 3A4 inhibitors or guanylate cyclase stimulators.
Interventions	Intervention: sildenafil citrate 12.5 mg in dextrose 5% (65 mL) intravenously over 150 minutes starting at the time of skin incision. Control: placebo. The same volume of 5% of dextrose solution was administered over 150 minutes.
Outcomes	The primary outcome of the trial was serum Cr concentration over time in both groups. Secondary outcomes include eGFR and multiple organ dysfunction scores from baseline to 96 hours postsurgery, with a final serum Cr sample for estimation of eGFR at 6 weeks; consensus clinical definitions of acute kidney, lung, liver, gut, brain, and myocardial injury; and sepsis, death, and a composite of these outcomes. Biomarkers of the inflammatory response (serum IL‐6, IL‐8, and IL‐ 10) and myocardial injury (serum TnI), and urine biomarkers of inflammation (NGAL) and injury (Timp2*IGF‐binding protein‐7 [Timp2*IGFBP7]) were measured in serial serum and urine samples.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Internet‐based randomisation system
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, researchers, and clinical staff were blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants, researchers, and clinical staff were blinded to group allocation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Five participants withdrew before surgery (two intervention, three placebo), five had major protocol deviations, and eight participants withdrew after surgery.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The British Heart Foundation funded the study.

Later 2009 (A).

Study characteristics
Methods	Double‐blind, randomised, placebo‐controlled trial. Run‐in period: June 204‐October 2006 Number of study centres and location: single centre. Leiden University Medical Centre
Participants	A total of 165 participants were assessed in part A of this study. Three hundred thirty‐three randomised, 114 placebo (103 received, 11 did not), 108 TXA (99 received, 9 did not), 111 aprotinin (96 received, 15 did not); first‐time, non‐complex (1 or 2 procedures) heart surgery with use of CPB Exclusion criteria: patients with previous sternotomy, known bleeding disorders, an abnormal preoperative coagulation profile for reasons other than anticoagulant therapy or treatment with anti‐Plt agents within 5 days before surgery, known or suspected allergy to aprotinin, plasma Cr concentration > 1.36 mg/dL (120 mmol/L), pregnancy, a history of thrombosis, or an emergency operation
Interventions	Intervention: TXA according to a full‐dose regimen (1 g loading dose, 500 mg added to the CPB system prime, and a continuous infusion of 400 mg/h) Control: placebo (0.9% saline solution) according to an identical regimen
Outcomes	Primary endpoints were total postoperative blood loss and transfusion requirements. Secondary endpoints included in‐hospital mortality, re‐exploration, perioperative MI, IABP use, mediastinitis, renal failure and use of CVVH, neurological complications, sepsis, postoperative intubation time, and length of ICU and hospital stay.
Notes	For demographic data, the whole population was considered (placebo, TXA, and aprotinin). For AKI, we considered renal failure; for neurological complications, we considered stroke; for myocardial damage, only MI; and for infection, only mediastinitis was considered. Later 2009 (A) considers the comparison between TXA and placebo, and Later 2009 (B) considers the comparison between placebo and aprotinin.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation list
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All caretakers were blinded to medication allocation. An independent anaesthesia assistant then prepared the trial medication overnight in identical‐appearing syringes labelled with the participants' trial number only. Codes were revealed once recruitment, data collection, and laboratory analyses were completed.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Codes were revealed once recruitment, data collection, and laboratory analyses were completed.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No commercial funding was used for this work.

Later 2009 (B).

Study characteristics
Methods	Double‐blind, randomised, placebo‐controlled trial; single‐centre study at Leiden University Medical Centre, enrolment period from June 2004 to October 2006
Participants	A total of 168 participants were assessed in part B of this study. Three hundred thirty‐three randomised, 114 placebo (103 received, 11 did not), 108 TXA (99 received, 9 did not), 111 aprotinin (96 received, 15 did not); first‐time, non‐complex (1 or 2 procedures) heart surgery with use of CPB Exclusion criteria: patients with previous sternotomy, known bleeding disorders, an abnormal preoperative coagulation profile for reasons other than anticoagulant therapy, or treatment with anti‐Plt agents within 5 days before surgery, known or suspected allergy to aprotinin, plasma Cr concentration > 1.36 mg/dL (120 mmol/L), pregnancy, a history of thrombosis, or an emergency operation
Interventions	Intervention: high‐dose aprotinin (Trasylol, Bayer AG Germany) according to the Hammersmith protocol (2 × 106 KIU aprotinin loading dose, 2 × 106 KIU added to the CPB system prime, and a continuous infusion of 5 × 105 KIU/h during CPB) Control: placebo (0.9% saline solution) according to an identical regimen
Outcomes	Primary endpoints were total postoperative blood loss and transfusion requirements. Secondary endpoints included in‐hospital mortality, re‐exploration, perioperative MI, IABP use, mediastinitis, renal failure and use of CVVH, neurological complications, sepsis, postoperative intubation time, and length of ICU and hospital stay.
Notes	For demographic data, the whole population was considered (placebo, TXA, and aprotinin). For AKI, we considered renal failure; for neurological complications, we considered stroke; for myocardial damage, only MI; and for infection, only mediastinitis was considered. Later 2009 (A) considers the comparison between TXA and placebo, and Later 2009 (B) considers the comparison between placebo and aprotinin.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation list
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All caretakers were blinded to medication allocation. An independent anaesthesia assistant then prepared the trial medication overnight in identical‐appearing syringes labelled with the participants' trial number only. Codes were revealed once recruitment, data collection, and laboratory analyses were completed.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Codes were revealed once recruitment, data collection, and laboratory analyses were completed.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No commercial funding was used for this work.

Lazar 2007.

Study characteristics
Methods	Prospective, multi‐centre, randomised, double‐blind, placebo‐controlled, phase 2 trial Run‐in period: May 2004 to October 2005 Number of centres and location: multi‐centre. United States
Participants	One hundred ninety‐seven high‐risk (urgent surgery, CABG, valve, reoperations, EF < 30%) female participants Inclusion criteria: female patients undergoing cardiac surgery on CPB requiring re‐do sternotomy, urgent or emergent CABG in the absence of an evolving MI, left main stenosis > 50%, three‐vessel coronary disease with NYHA Class III or IV angina, or the combination of CABG valve repair or replacement Exclusion criteria: presence of an MI within 72 hours before randomisation, the presence of an LBBB, the need for isolated valve or aneurysm surgery, the presence of hepatic and renal failure, sepsis, and patients who were pregnant and those with documented complement or immune deficiency syndromes or HIV infection
Interventions	Intervention: single IV infusion of TP10 (5 mg/kg) over 30 minutes before surgery Control: placebo (sodium chloride 0.9%) over 30 minutes before surgery All infusions were completed before CPB.
Outcomes	The primary efficacy endpoint was the composite incidence of death or MI within 28 days of surgery. The primary safety endpoint was the incidence of reported adverse events. Secondary endpoints included death, MI, changes in baseline levels of SC5b‐9 and CH50, hospital and ICU length of stay, and changes in ECG.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "After anesthetic induction, patients were assigned to a treatment group using a blocked randomization (SAS version 8.2)"
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No clear description of blinding for participants and healthcare personnel
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded MI assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "All efficacy analysis were performed on intent to treat (ITT) population, whereas safety analysis were done on the safety population (as treated)." Overall, a small number of participants did not complete the study (17 out of 297) for a number of reasons. Statistical analysis was adjusted accordingly.
Selective reporting (reporting bias)	Low risk	No trial registration number was reported; thus, we were unable to access prospective trial registration. However, the outcomes are clearly stated in the methodology section and are all reported.
Other bias	High risk	Comment: Avant Therapeutics which manufactures TP10 provided grants for study funding and for the PI of the study. Some of the company employees are authors on the paper and have worked on the trial. It is unclear how much influence Avant Therapeutics may have exerted on the trial design.

Lei 2003.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Published 2003 Number of centres and location: not specified
Participants	Thirty‐two patients (age = 55.5 years ± 10.5 years old; male n = 23 and female n = 9) with pulmonary arterial hypertension (pulmonary arterial pressure > 60 mmHg) undergoing mitral valve replacement or AVR, with heart function NYHA class III to IV
Interventions	Intervention groups: aprotinin 1.0 × 105 KIU/kg Control group: it was not specified what this group received.
Outcomes	Proinflammatory cytokines, including IL‐1β, IL‐8, and TNF‐α, and the anti‐inflammatory cytokine IL‐10
Notes	As per protocol, we considered the highest dosage for the intervention group. IL and TNF measures were expressed in pg/L, reported because they were compatible with other studies' units.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No clear description of blinding procedures for participants and healthcare personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Procedure for the blinding of outcome assessors not described in the paper
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants were lost to follow‐up or did not complete the study.
Selective reporting (reporting bias)	Low risk	It was not possible to ascertain whether the trial was prospectively registered (no trial registry number available). However, the outcomes were clearly stated in the methodology section and reported.
Other bias	Low risk	There was no evidence of funding bias.

Leyh 2001.

Study characteristics
Methods	Prospective, single‐centre, randomised study Run‐in period: not specified. Received 25 May 2000, received in revised form 3 November 2000, accepted 15 November 2000 Number of centres and location: single centre. University of LuÈbeck, LuÈbeck, Germany
Participants	Forty‐eight patients with stable angina and preserved LV function (EF > 45%) undergoing primary elective CABG Exclusion criteria were an age equal to or more than 75 years, administration of acetylsalicylic acid within the 7 days prior to surgery, a Plt count < 150,000/mL, and a documented or reported coagulopathy assessed by routine coagulation parameters and patient history.
Interventions	Intervention A (n = 16): MUF Intervention B (n = 16): conventional ultrafiltration Control group (n = 16): treatment not specified
Outcomes	PF 4, FBC, Plt, fibrinogen, ATIII activity, TT, aPTT, ACT, TAT, prothrombin fragments, fibrinomonomers, tPA, D‐dimer, PAP were assessed in 36 patients. In the remaining 12 patients (four patients in each group), only routine parameters were evaluated. Postoperative blood loss and transfusion requirements
Notes	No clinical outcomes reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were prospectively randomized into a control group (Con.; n = 16), a conventional ultrafiltration (CUF) group (n = 16) and a MUF group (n = 16)." No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment of participants after randomisation
Blinding of participants and personnel (performance bias) All outcomes	High risk	No clear description of processes to blind participants and healthcare personnel. It would be possible to blind study participants. However, it would not be feasible to blind healthcare personnel because of the visible difference between the interventions and the control.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of processes to blind outcome assessors to participant allocations
Incomplete outcome data (attrition bias) All outcomes	Low risk	All outcomes were reported for all participants.
Selective reporting (reporting bias)	Low risk	Although it was not possible to access the prospective trial registration, the outcomes of interest were clearly described in the methodology section and were all reported in the results section.
Other bias	Low risk	No evidence of funding bias

Liakopoulos 2007.

Study characteristics
Methods	Prospective, single‐centre, placebo‐controlled study Run‐in period: November 2003 to July 2004. Accepted for publication in 2007 Number of centres and locations: single centre. University of Göttingen, Göttingen, Germany
Participants	Eighty patients undergoing elective CABG with CPB Exclusion criteria were previous emergency or concomitant cardiac surgical procedures, age > 80 years, compromised LVEF (< 30%), AMI (within 4 weeks), acute and chronic infections, known neoplasm, renal or hepatic dysfunction, autoimmune disease, or preceding anti‐inflammatory treatment. Patients who routinely received intraoperative aprotinin because of surgeons' preferences were not enrolled for this study to exclude known confounding anti‐inflammatory actions.
Interventions	Intervention: single IV bolus of 15 mg/kg MP 30 minutes before the start of CPB Control: placebo (NaCl 0.9%) at the same time
Outcomes	Activation of myocardial transcriptional proteins of inflammation and systemic cytokine response (TNF‐α, IL‐6, IL‐8, and IL‐10) as well as variables of postoperative haemodynamic, myocardial, and pulmonary function
Notes	IL‐10 values not measurable at 24 hours
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Following a computer‐generated sequence, patients were randomly assigned to receive either a single intravenous bolus of 15 mg/kg methylprednisolone (Urbason, Sanofi‐Aventis, Frankfurt, Germany; MP group) or placebo (NaCl 0.9%; PLA group) 30 minutes before CPB was instituted." Use of computer‐generated sequence to randomise patients
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	No clear description of whether trial participants or healthcare personnel have been blinded in the paper. However, in the discussion below, the author state that the study was not blinded. Blinding would have been possible given the type of intervention.
Blinding of outcome assessment (detection bias) All outcomes	High risk	No clear description of whether outcome assessors were blinded. However, in the discussion below, the author state that the study was not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Authors did not specify whether intention‐to‐treat analysis was planned. Two participants did not complete the study and appear to have been excluded from the final analysis; however, as it is a small number in proportion to the total, the impact on results is likely to be insignificant.
Selective reporting (reporting bias)	Low risk	It was not possible to access the prospective trial registration (if any) as no trial registration number was reported in the paper. However, outcomes were clearly stated in the methodology section and were reported in the results.
Other bias	Low risk	No clear evidence of funding bias

Lin 2020.

Study characteristics
Methods	Prospective, randomised, placebo‐controlled double‐blinded clinical trial Run‐in period: not specified. Received: September 11, 2019, accepted January 31, 2020, available online February 21, 2020, published April 20, 2020 Number of centres and location: not specified. Authors from Second Affiliated Hospital of Harbin Medical University, Harbin, Heilongjiang, P.R. China
Participants	Patients undergoing elective cardiac surgery with CPB Exclusion criteria: participants with BMI > 35, polycythaemia, smoking cessation < 2 weeks, and severe lung dysfunction including FVC and FEV1 < 50% of the predicted values. Participants with a history of severe anaemia, immune dysfunction, systemic infection, hypoproteinaemia, serious lung infection, or pleural effusion were also excluded from this study. In addition, those participants with hepatic, renal, or coagulation disorder, COPD, or asthma were also excluded, as were participants who received a second cardiac surgery.
Interventions	Intervention: preoperative EPO 100 IU/kg, diluted into 50 mL saline, continuously infused over 3 days Control: 0.9% sodium chloride
Outcomes	The primary endpoint was to compare postoperative lowest PaO2/FiO2 after CPB. The ventilation function, including dynamic compliance, peak airway pressure, and plateau pressure, were recorded. The levels of TNF‐α, IL‐1b, and IL‐10 in serum and arterial blood gas were analysed. The mechanical ventilation time in the ICU, the length of time spent in the ICU, the time from operation to discharge, and the total time of hospitalisation were recorded. Adverse events in the ICU were monitored and recorded.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The first anesthesiologist randomized all the patients using the random number table generated by the computer and prepared the saline or EPO." Randomisation was performed using a random number table generated by the computer.
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "This study was a double‐blinded clinical trial. Three anesthesiologists were recruited to perform the randomization, anesthesia, and observation. The first anesthesiologist randomized all the patients using the random number table generated by the computer and prepared the saline or EPO. The second anesthesiologist only performed the general anesthesia for all the patients."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "This study was a double‐blinded clinical trial. Three anesthesiologists were recruited to perform the randomization, anesthesia, and observation. The last anesthesiologist only collected research sample and recorded the data."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the trial and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	The trial protocol was prospectively registered. All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Lindholm 2004.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Accepted for publication June 2, 2004 Number of study centres and location: not specified. Authors from Sahlgrenska University Hospital, Gothenburg, Sweden
Participants	Forty‐one elderly patients (mean age 73.1 years, 66% men) were included in the study. Inclusion criteria were age > 60 years, coronary disease with stable angina pectoris or aortic stenosis, LVEF > 20%, and an expected CPB time of > 90 minutes. Exclusion criteria were reoperation and preoperative use of steroids or nonsteroidal anti‐inflammatory drugs. Aspirin and clopidogrel were discontinued 5 days to 7 days before surgery.
Interventions	Intervention: biocompatible circuit group (n = 21). This consisted of a closed and totally heparinised extracorporeal system. All components, including cannulas, tubing connectors, and polyvinyl chloride tubing, were surface‐heparinised with Carmeda Bioactive Surface (CBAS; Medtronic Inc). Control: conventional management with an open, nonheparinised system was used in the conventional group (n = 20).
Outcomes	Markers of complement activation, cytokine release, granulocyte degradation, coagulation, and fibrinolytic activity
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Forty‐one patients were randomly allocated to either of two regimens by a computerized program for randomization with sequential allocation depending on sex, age, weight, and type of operation" At low risk for random sequence generation
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	It is not explicitly stated in the paper whether participants and personnel were blinded. While it would have been possible to blind participants, it is not possible to blind personnel to this intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not described whether outcome assessors were blinded and, if so, what procedures were in place to achieve blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no prospectively registered protocol is available, outcomes were clearly stated in the methodology section and were reported in the results.
Other bias	Low risk	Quote: "The study was supported by grants from The Swedish Heart and Lung Foundation and the Gothenburg Medical Association". Funding sources were clearly disclosed. Funding was received from nonindustry sources and unlikely to affect the outcome of the study.

Liu 2009.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Received 15 February 2009, received in revised form 17 May 2009, accepted 16 June 2009; available online 15 August 2009 Number of centres and locations: single‐centre. Beijing Anzhen Hospital, Capital Medical University, Beijing 100029, PR China
Participants	Thirty participants with rheumatic heart valve disease receiving valve replacement surgery Exclusion criteria: participants with comorbidities of coronary heart disease and with hepatic, renal, or pulmonary disease were excluded. Participants having additional cardiac diseases, severe noncardiac diseases, hypertension, diabetes mellitus, congenital heart disease, or undergoing a repeat operation were also excluded.
Interventions	Intervention: adenosine 100 mg/kg/min over 10 minutes before application of the aortic cross‐clamp and antegrade 1 mmol/L ADO high‐potassium cold (12 °C) blood cardioplegia after clamp‐on (n = 15) Control: regular institutional high‐potassium ([K+] = 20 mol L1) cold (12 °C) blood cardioplegia (n = 15)
Outcomes	Myocardial recovery and inflammatory response, cTnI, CK‐MB, and inflammatory factors (IL‐6 and IL‐8) (obtained from serial venous blood samples after induction, 5 minutes after cross‐clamp of aorta, 10 minutes after clamp‐off, 1 hour after return to the ICU, and postoperatively 24 hours and 48 hours)
Notes	No clinical outcomes reported, besides death at 30 days
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation process not clearly described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was not clearly described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "The cardiologists in the intensive care unit (ICU) were blinded to group allocation in this investigation." It is unclear whether participants were blinded to the intervention. One group of staff caring for participants in the postoperative period (the cardiologists in the ICU) were blinded to the intervention, but it is unclear whether theatre staff were. We cannot therefore make a firm judgement on the risk of bias for this domain.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors state that the cardiologists in the ICU were blinded to group allocation. However, it is not clear from the paper whether they were involved in assessing the outcomes, which are heavily laboratory‐based tests.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the trial and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	It was not possible to access a prospectively registered trial record. However, all the outcomes clearly specified in the methods were reported in the results.
Other bias	Low risk	No clear evidence of funding bias

Lomivorotov 2013.

Study characteristics
Methods	Prospective, double‐blinded, placebo‐controlled, randomised study Run‐in period: 2009 (months not specified). Received 2 February 2012; received in revised form 1 May 2012; accepted 1 August 2012; available online 28 August 2012 Number of centres and location: not specified. Authors from Rechkunovskaya Street 15, Novosibirsk 630055, Russia
Participants	Patients undergoing a CABG in 2009 Exclusion criteria were age > 70 years, LVEF < 40%, diabetes mellitus, COPD, and chronic renal disease.
Interventions	Intervention: MP (20 mg/kg intraoperatively immediately after anaesthesia induction) Control: placebo (20 mL of 0.9% NaCl)
Outcomes	Serum ET‐1 was the primary endpoint. The secondary endpoints included E‐selectin, pro‐ and anti‐inflammatory markers (IL‐6 and IL‐10), oxygenation index (PaO2/FiO2), and serum glucose. Microalbuminuria was measured to assess capillary leak syndrome. Parameters were measured at the following: immediately prior to anaesthesia induction; 10 minutes after CPB; 2 hours after CPB; 4 hours after CPB; and 24 hours after surgery. We also assessed hospital mortality, ventilation time after surgery, frequency of inotropic support, frequency of AF during the first 24 hours after surgery, infectious complications, ICU and hospital stay durations, and ICU readmission.
Notes	IL‐10 values not measurable at 24 hours
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomisation was performed using opaque sealed envelopes."
Allocation concealment (selection bias)	Low risk	Quote: "Randomisation was performed using opaque sealed envelopes."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The solution for injections was prepared in a non‐transparent syringe by an independent pharmacist." Authors also state that the trial is double blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessors is not clearly described.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Three participants were excluded from the study and the final analysis. The reasons for exclusions were varied (mitral valve surgery, off‐pump surgery, and severe bleeding). Although they are a small proportion of the cohort, the reasons for exclusion are significant event, and their impact is unclear.
Selective reporting (reporting bias)	Low risk	Although there was no trial registration available, all outcomes of interest were clearly explained on the methodology section and fully reported.
Other bias	Low risk	No clear evidence of funding bias

Lucchinetti 2012.

Study characteristics
Methods	Prospective, randomised, placebo‐controlled trial Run‐in period: September 2008 to July 2010 Number of centres and location: single centre. University of Alberta Hospital
Participants	Fifty‐five participants scheduled for elective ONCABG surgery were finally enrolled and assigned to RIPC treatment or placebo at the University of Alberta Hospital between September 2008 and July 2010. Inclusion criteria were being scheduled for elective ONCABG surgery and age of 50 years to 85 years. Exclusion criteria were emergency surgery, MI within 48 hours before surgery as defined by increased plasma concentrations for cardiac enzymes, diabetes mellitus, a BMI > 35, concomitant noncardiac surgery, or severe peripheral vascular disease. Fifty‐seven participants gave consent, and two participants withdrew because of logistical reasons.
Interventions	Intervention: RIPC (four 5‐minute cycles of lower‐limb ischaemia reperfusion induced by a tourniquet inflated to 300 mmHg) Control: placebo. No ischaemic preconditioning was performed, but cuff was placed on the leg and hidden from participants.
Outcomes	The primary outcome was hsTnT release. Secondary endpoints were plasma levels of NT‐proBNP, hs‐CRP, S100 protein, and short‐ and long‐term clinical outcomes. Gene expression profiles were obtained from atrial tissue using microarrays.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A 1:1 block randomization (block size 10) with no further stratification was generated by an independent person using a computer random number generator"
Allocation concealment (selection bias)	Low risk	Quote: "The results were stored in numbered, sealed, opaque envelopes"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All participants had the cuff placed on the lower limb but would have been able to discriminate whether it was inflated or not. The intervention was delivered by a technician who was not involved in any other phase of the study. Therefore, overall we rate the risk of bias low.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Collection and analyses of all clinical and laboratory data were performed by study personnel blinded for group assignment."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Two participants withdrew because of logistic reasons, but it is unclear from which group.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Luo 1998.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified. Published in 1998 Number of centres and location: not specified. Authors from the Institute for Cardiovascular Disease of Tongji Medical University
Participants	Twenty patients undergoing elective heart valve surgery (MVR, AVR, DVR) Exclusion criteria: salicylate up to 7 days before operation; allergo‐immunologic, hepatic, renal, or coagulation disorders
Interventions	Intervention: aprotinin 30 x 105KIU Control: no intervention
Outcomes	RC3bRR and RICR were done. Plasma C3, C4, CH50, and IgG levels were measured. Samples were collected preoperatively at the end of CPB and on the first, third, and seventh PODs.
Notes	No clinical outcomes reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description of the randomisation technique
Allocation concealment (selection bias)	Unclear risk	No description of the allocation concealment measures
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not describe whether the study was blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of outcome assessment blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The study was funded by a national institution. Therefore, we rate the risk of funding bias as low.

Luo 2004.

Study characteristics
Methods	Prospective, randomised, double‐blind study Run‐in period: November 2002‐ May 2003. Received 21 November 2003; received in revised form 18 February 2004; accepted 23 February 2004 Number of study centres and location: single centre. Xiang Ya Hospital, Hunan Medical University, Changsha, Hunan 410008, China
Participants	Thirty patients undergoing elective valve replacement were randomised to receive either aminophylline treatment (aminophylline, n = 15) or no aminophylline (control, n = 15). Exclusion criteria for the experiment were infective valve disease, valve disease with CAD, reoperation of valve, or COPD.
Interventions	Intervention: aminophylline (5 mg/kg) intravenously immediately after induction of anaesthesia over 5 minutes, followed by continuous IV administration of aminophylline at a rate of 0.5 mg/kg/h until the end of CPB Control group: same volume of 0.9% sodium chloride
Outcomes	Release of cytokines (IL‐8, IL‐10, and TNF‐α and neutrophil count) and pulmonary function (mean arterial pressure, heart rate, CVP, pulmonary artery pressure, cardiac index, and RI) Blood samples were collected from the peripheral arterial lines after anaesthesia induction but before administration of aminophylline, at 1 hour, 8 hours, and 24 hours after termination of CPB.
Notes	Lung damage considered as prolonged ventilation; myocardial injury, such as ventricular arrhythmias
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were randomized by drawing prearranged cards in an envelope before anesthesia"
Allocation concealment (selection bias)	Unclear risk	The allocation cards were in an envelope, but it was not specified whether it was opaque.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "A nurse of operating room performed the randomization and prepared the syringes of blinded solution. Only anesthesiologist and one nurse of operating room were aware of the treatment groups. All operations were performed by one surgeon" The participants and likely the outcome assessors were blinded. However, as intraoperative haemodynamic parameters were reported in the trial, the fact that the anaesthesiologist was aware of trial arm allocation may introduce some bias into the results.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Only anesthesiologist and one nurse of operating room were aware of the treatment groups"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Luo 2007.

Study characteristics
Methods	Prospective, randomised, blinded trial Run‐in period: not specified. Received 17 August 2006; received in revised form 26 November 2006; accepted 28 November 2006; available online 26 December 2006 Number of study centres and location: single centre. Xiang Ya Hospital, Central South University, Changsha, Hunan, Peoples Republic of China
Participants	Thirty adult patients with rheumatic heart valve disease undergoing elective isolated valve replacement Exclusion criteria for this study included infective valve disease, valve disease with CAD, previous valve repair, and severe LV enlargement (LVDED > 70 mm).
Interventions	Intervention: aminophylline bolus (5 mg/kg) intravenously immediately after induction of anaesthesia and insertion of radial arterial and central venous catheters but before skin incision over 5 minutes Control: bolus of 0.9% sodium chloride (5 mg/kg)
Outcomes	Outcomes assessed were cTnI after anaesthesia induction and at 30 minutes, 8 hours, and 24 hours after aortic declamping; myocardial cAMP and tissue MPO activity just before start of CBP and 30 minutes after aortic declamping; transcardiac neutrophil count before cardioplegic arrest and 30 minutes after aortic declamping; MAP, heart rate, CVP, pulmonary artery pressure, and average dose of inotropes used for first 24 hours postoperatively.
Notes	MPO from tissue sections, not comparable to other studies
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were randomized by randomized number table before anesthesia"
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "Only the anesthesiologist and one operating room nurse were aware of the treatment groups" The participants and likely the outcome assessors were blinded. However, as intraoperative haemodynamic parameters were reported in the trial, the fact that the anaesthesiologist was aware of trial arm allocation may introduce some bias into the results.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Only the anesthesiologist and one operating room nurse were aware of the treatment groups"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no prospectively registered protocol was available, all outcomes described in the methods were included in the paper.
Other bias	Low risk	No clear evidence of funding bias

Mansourian 2015.

Study characteristics
Methods	Single‐centre, prospective, placebo‐controlled, randomised clinical trial Run‐in period: June 21, 2011 to July 21, 2012. Accepted November 5, 2014, available online November 5, 2014 Number of centres and location: single centre. Tehran Heart Center, Tehran University of Medical Sciences, Tehran, Iran
Participants	One hundred seventy‐eight CABG candidates with EF ≤ 30%, divided into two equal groups (PTX and control) Exclusion criteria: previous RF, recent MI (< 4 weeks ago), uncontrolled diabetes, and use of anti‐inflammatory drugs
Interventions	Intervention: PTX 400 mg orally 3 times a day for 3 days before the operation in addition to usual medications Control: placebo at the same time in addition to usual medications
Outcomes	Levels of TNF‐α and IL‐6, cardiac troponin T before entry to the operating room and immediately after arrival in the ICU; preoperative and postoperative WBC, Hb, Hct, BUN, Cr, and Plt count Clinical outcomes: need for inotropes, transfusion, and IABP in the ICU; rates of postoperative CVA, RF, MI, ventilation dependency, ICU stay, and mortality; postoperative LV function
Notes	In spite of the randomisation and allocation of the participants between the two groups, some preoperative values between groups were significantly different.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Randomization was done simply by using a random number table by one of the authors who was not involved in the data gathering and clinical investigations". At the end of the paper, the authors state "In spite of the randomization and allocation of the patients between the two groups, some pre‐operative values between groups were significantly different, which made us consider it in the statistical analysis and compare the change in values between the two groups." However, Table 1 reporting patient characteristics at baseline does not show any significant differences between cohorts apart from the levels of IL‐6 and TNF‐a which could be attributed to the intervention. As such, we have rated the risk of bias for this domain unclear.
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment processes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The study was blinded for the patient groups and surgeons"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of outcome assessment blinding
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All participants completed the study and were included in the final analysis. The authors adjusted the analysis in view of differences in preoperative baseline levels of inflammatory markers.
Selective reporting (reporting bias)	High risk	There was no prospectively registered trial protocol. Some outcomes were prespecified in the methodology section. However, the authors then report additional outcomes such as BUN and CBP time. Furthermore, their statistical analysis plan appear to have been changed after realising some differences in the group characteristics.
Other bias	Low risk	No clear evidence of funding bias

Marberg 2010.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Run‐in period: September 2006 to May 2007 Number of study centres and location: single centre. Sahlgrenska University Hospital, Goteburg, Sweden
Participants	Eighty consecutive patients with stable angina pectoris admitted for elective CABG at Sahlgrenska University Hospital between September 2006 and May 2007 Exclusion criteria: known liver, kidney, or bleeding disorder; perioperative use of aprotinin; or clopidogrel treatment within 5 days before surgery
Interventions	Intervention: autotransfusion of all mediastinal shed blood during the first 12 postoperative hours Control: no autotransfusion of mediastinal shed blood (any mediastinal shed blood was discarded)
Outcomes	Primary endpoint was postoperative bleeding volume during the first 12 postoperative hours. Predefined secondary endpoints were transfusion requirements, Hb levels, thrombo‐elastometric variables, and plasma concentrations of IL‐6, TAT, and D‐dimer 24 hours after surgery.
Notes	IL‐6 report not included because measured after 24 hours
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random assignment was conducted using unmarked envelopes, each containing a card indicating autotransfusion or no‐autotransfusion".
Allocation concealment (selection bias)	Low risk	Quote: "Random assignment was conducted using unmarked envelopes, each containing a card indicating autotransfusion or no‐autotransfusion".
Blinding of participants and personnel (performance bias) All outcomes	High risk	The authors state that the trial was open label. Blinding of the theatre and ICU personnel would have been impossible because of the nature of the intervention; however, participants and the rest of the trial staff could have been blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors state that the trial was open label. It is unclear whether any blinding of outcome assessors has been implemented, especially for the laboratory variables.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "Three patients were excluded after inclusion, two due to perioperative use of aprotinin and one due to ongoing clopidogrel treatment, overseen at the inclusion" Although the number of participants excluded from the study and the analysis is small compared to the total, it leaves the study underpowered (power calculations by the authors showed that they needed 80 participants, which is the exact number recruited).
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The study was supported by a charity and internal hospital funding; thus, we rate this at low risk of funding bias.

Marcheix 2008 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: December 2003 to July 2005. Accepted for publication August 22, 2007 Number of centres and location: not specified. Authors from Montreal Heart Institute and University of Montreal, Montreal, Quebec, Canada
Participants	One hundred patients undergoing isolated ONCABG surgery. Participants were randomised in a 2 × 2 factorial design (thus resulting in 2 trials for the purpose of this review with 50 participants in each trial). Exclusion criteria: patients with valvular and aortic surgery
Interventions	Intervention: cell‐saving device alone Control: no cell‐saving device and no VACPB
Outcomes	Postoperative inflammatory response, with a specific focus on the implication of the complement pathway
Notes	In our review, we considered group C (as intervention, no VACPB with CS); the control was group A (no VACPB and no CS).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was based on random numbers in a sealed card opened by perfusionists just before surgery."
Allocation concealment (selection bias)	Low risk	Quote: "Randomization was based on random numbers in a sealed card opened by perfusionists just before surgery."
Blinding of participants and personnel (performance bias) All outcomes	High risk	This study was not blinded for participants and personnel.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not describe whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the prespecified analysis.
Selective reporting (reporting bias)	Unclear risk	No prospectively registered protocol. Outcomes are not clearly listed in the method sections; for instance, the authors describe some of the laboratory markers under investigation but not what clinical outcomes. Therefore, it is unclear whether any additional outcomes were added or whether any were not reported.
Other bias	Low risk	No evidence of funding bias

Marcheix 2008 (B).

Study characteristics
Methods	Prospective, randomised study Run‐in period: December 2003 to July 2005. Accepted for publication August 22, 2007 Number of centres and location: not specified. Authors from Montreal Heart Institute and University of Montreal, Montreal, Quebec, Canada
Participants	One hundred patients undergoing isolated ONCABG surgery. Participants were randomised in a 2 × 2 factorial design (thus resulting in 2 trials for the purpose of this review with 50 participants in each trial). Exclusion criteria: patients with valvular and aortic surgery
Interventions	Intervention: VACPB and cell‐saving Control: VACPB
Outcomes	Postoperative inflammatory response, with a specific focus on the implication of the complement pathways
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was based on random numbers in a sealed card opened by perfusionists just before surgery".
Allocation concealment (selection bias)	Low risk	Quote: "Sealed card opened by perfusionists just before surgery".
Blinding of participants and personnel (performance bias) All outcomes	High risk	The trial was open label.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis and completed the trial.
Selective reporting (reporting bias)	Unclear risk	No prospectively registered protocol. Outcomes are not clearly listed in the method sections; for instance, the authors describe some of the laboratory markers under investigation but not what clinical outcomes. Therefore, it is unclear whether any additional outcomes were added or if any were not reported.
Other bias	Low risk	No evidence of bias due to funding

Massoudy 1999.

Study characteristics
Methods	Prospective, randomised, double‐blind study Run‐in period: not specified. Approved by ethics committee August 1996. Accepted for publication October 14, 1998 Number of centres and location: not specified. Authors from the German Heart Center Munich, and the Department of Physiology, University of Munich, Munich, Germany
Participants	Twenty‐two patients undergoing elective CABG. Inclusion criteria were the following: LVEF ≤ 55%; age ≤ 75 years ; and no MI (as defined by significant elevation of serum CK levels or electrocardiographic signs for MI) within 14 days before the operation. Exclusion criteria were known hypersensitivity against SNP or any other nitro substances, inflammatory diseases, or the intake of immunosuppressive drugs.
Interventions	Intervention: sodium nitroprusside (Nipruss 0.5 μg/kg/min after release of the aortic cross‐clamp for 20 minutes through the central venous line of a Swan‐Ganz catheter) Control: 5% glucose solution Protamine (1 mg/kg) was given after CPB in both groups.
Outcomes	ILs and mortality
Notes	No clinical outcomes reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding procedures
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of blinding procedures for outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the trial and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The sodium nitroprusside was donated by a pharmaceutical company who was not otherwise involved in the study. Thus, we rate this at low risk of funding bias.

Massoudy 2000.

Study characteristics
Methods	Prospective, randomised, double‐blind study Run‐in period: not specified. Received for publication March 22, 1999; revisions requested July 26, 1999; revisions received Oct 25, 1999; accepted for publication Oct 26, 1999 Number of study centres and location: not specified. Authors from centres in Essen and Munich, Germany
Participants	Thirty patients undergoing elective CABG. Inclusion criteria were the following: an LVEF ≤ 40%; age ≤ 80 years; and no MI (as defined by significant elevation of serum CK levels or electrocardiographic signs for MI) within 14 days before the operation. Exclusion criteria: known hypersensitivity against sodium nitroprusside or any other nitro substances, inflammatory diseases, or the intake of immunosuppressive drugs. A further exclusion criterion was unstable angina.
Interventions	Intervention: sodium nitroprusside (Nipruss 0.5 μg/kg/min) after release of the aortic cross‐clamp through the central venous line of a Swan‐Ganz catheter over 60 minutes Control: 5% glucose solution as placebo through the central venous line of a Swan‐Ganz catheter over 60 minutes Protamine (1 mg/kg) was given after CPB to both groups.
Outcomes	The proinflammatory cytokines IL‐6 and IL‐8, leucocyte adhesion molecule CD11b, Plt adhesion molecules CD41 and CD62, and CD41 on leucocytes. Samples were taken at the following time points: after application of heparin; 1 minute, 5 minutes, and 10 minutes after release of the aortic cross‐clamp; after release of the Lambert‐Kay clamp (approximately 25 minutes after release of the aortic cross‐clamp); and after application of protamine (approximately 35 minutes after release of the aortic cross‐clamp).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Authors state that this was a double‐blind study.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not describe whether outcome assessors were blinded and, if so, how.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Matata 2000.

Study characteristics
Methods	Single‐centre, prospective study Run‐in period: 1987‐1998. Received for publication October 1, 1999; revisions requested December 16, 1999; revisions received January 27, 2000; accepted for publication February 24, 2000 Number of centres and location: Single centre. Authors from Glenfield Hospital, Leicester, United Kingdom
Participants	Twenty patients with single‐ or double‐vessel coronary disease undergoing elective CABG Exclusion criteria: patients with diabetes, diseases of the circumflex and the left main stem, valvular diseases, ventricular aneurysm, heart failure, and poor LV function
Interventions	Intervention: off‐pump CPB Control: standard CPB
Outcomes	Oxidative stress and inflammatory reaction
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	The authors do not describe whether participants were blinded. Operating theatre personnel could not be blinded because of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not describe whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	The authors describe funding received by the hospital, charitable bodies, and industry sources (Medtronics Inc). However, no commercial products from the company were employed in the study, and the company does not appear to have had direct involvement in the trial; thus, we rate this at low risk of bias.

Mayumi 1997.

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind, controlled trial Run‐in period: December 1993 to July 1994. Accepted for publication July 19, 1996 Number of centres and location: single centre. Kyushu University Hospital. Fukuoka, Japan
Participants	Twenty‐seven adult patients who underwent elective valve replacement. Only the patients with preserved autologous frozen blood and similar severity were selected by the senior author (H.M.). Exclusion criteria: patients with cardiac cachexia due to end‐stage valvular disease
Interventions	Intervention: MPSS, 20 mg/kg body weight before and after bypass Control: placebo before and after bypass
Outcomes	Blood cell count, CRP, lymphocyte surface markers (CD3, CD4, CD8, CD16, and CD20), phytohaemagglutinin response, IL‐2 production, and natural killer cell activity were examined on admission through day 7.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A chief anesthesiologist, who was not directly involved in the present study, was responsible for opening an envelope indicating the drug, and also for preparing the drug in a covered syringe."
Allocation concealment (selection bias)	Low risk	Quote: "A chief anesthesiologist, who was not directly involved in the present study, was responsible for opening an envelope indicating the drug, and also for preparing the drug in a covered syringe."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "A chief anesthesiologist, who was not directly involved in the present study, was responsible for opening an envelope indicating the drug, and also for preparing the drug in a covered syringe. The patient names, but not drug names, included in the two groups were told by the anesthesiologist to the senior author at the end of study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "A chief anesthesiologist, who was not directly involved in the present study, was responsible for opening an envelope indicating the drug, and also for preparing the drug in a covered syringe. The patient names, but not drug names, included in the two groups were told by the anesthesiologist to the senior author at the end of study. After the statistical analysis was completed, the used drug for each group and the used dose of the steroid in each patient were disclosed to the senior author."
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Patients who underwent intraaortic balloon pumping (n = 1) or allogeneic blood transfusion (n = 2) were excluded from this study at the halfway point." The overall number of participants excluded from the final analysis is small; they represent a large proportion because of the small sample size. Furthermore, the exclusion reasons were for clinical complications which could have affected the results.
Selective reporting (reporting bias)	High risk	Quote: "Only the patients with preserved autologous frozen blood and similar severity were selected by the senior author (H.M.)"
Other bias	Low risk	Quote: "This study was supported by Grants‐in‐Aid for Scientific Research from the Ministry of Education, Science and Culture, and the Ministry of Health and Welfare, Japan."

McBride 2004.

Study characteristics
Methods	Prospective, double‐blind, randomised study Run‐in period: not specified. Received 30 September 2003; received in revised form 17 March 2004; accepted 30 March 2004 Number of study centres and locations: not specified. Authors from different institutions in Belfast
Participants	Thirty‐six ASA grade III/IV patients undergoing elective CABG with CPB Exclusion criteria were unstable angina, previous steroid therapy, MI within the previous 3 months, diabetes, heart or liver failure, and patients with documented renal dysfunction (plasma Cr >125 mmol/L).
Interventions	Intervention: MPSS 30 mg/kg just prior to induction Control: placebo
Outcomes	Changes in urinary TNFsr and IL‐1ra. Blood samples were obtained at baseline (sample A), 10 minutes after aortic cross‐clamp release (sample B), and 2 hours and 24 hours after revascularisation (samples C and D). Urine samples were obtained at baseline (sample A), 10 minutes after aortic cross‐clamp release (sample B), and 2 hours, 24 hours, 48 hours, and 72 hours after revascularisation (samples C to F, respectively). Urine samples were assayed for IL‐1ra and TNFsr‐2 and TGFb1, and plasma samples were assayed for TNF‐α, IL‐8, IL‐10, IL‐1ra, and TNFsr‐2.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The study was double blind with respect to the patient and the laboratory investigators. Ethical considerations required that the peri‐operative physician was aware of the randomisation group." Overall, we rate the study at low risk of bias for this domain.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The study was double blind with respect to the patient and the laboratory investigators."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No missing outcome data, although one participant from the control group was excluded because of deterioration on anaesthetic induction.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	Quote: "We acknowledge a project grant from the Heart Fund Trust, Northern Ireland. Dr Allen was in receipt of a Clinical Research Fellowship from the Royal Group of Hospitals Trust, Belfast, Northern Ireland. Dr Gormley was in receipt of a Clinical Research Fellowship grant from the Department of Health and Social Services, Northern Ireland." Funding was received from charitable and hospital funds with no evidence of conflict of interest.

Mirow 2001.

Study characteristics
Methods	Prospective, randomised study, single‐centre study. Run‐in period: December 1997 to June 1999 Number of study centres and locations: single centre. Heart Center NRW, Ruhr University of Bochum, Bad Oeynhausen, Germany
Participants	Two hundred forty‐three patients undergoing elective CABG were included. Exclusion criteria: presumed high‐risk patients, age > 80 years, BMI > 30, EF < 30%, LVEDP > 20 mmHg, persistent angina or decompensation, moderate or severe valve disease, concomitant cardiomyopathy, relevant pulmonary disease, relevant carotid stenosis (> 70%), CVA events, syncope, fever or leucocytosis, blood coagulation disorders, redo operation
Interventions	Intervention: coated ECC with low–molecular‐weight heparin Control: standard uncoated ECC set In both groups, IV heparin was administered at an initial dose of 400 IE/kg body weight. During ECC, ACT was monitored at 30‐minute intervals and was kept at 3480 seconds.
Outcomes	Clinical feasibility, hazards, and advantages of heparin‐coated CPB combined with either full‐dose or low‐dose systemic heparinisation
Notes	Group C, which studied reduced doses of heparin, was not considered for our meta‐analysis. Age was not retrievable from data because it was reported as mean and range.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that the trial was randomised, but they do not describe how the randomisation was performed.
Allocation concealment (selection bias)	Unclear risk	The authors do not describe whether allocation was concealed and, if so, how.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not describe whether participants or personnel were blinded. It would have been difficult to blind personnel in view of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not describe whether the outcome assessors were blinded to group allocations.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	No prospectively registered protocol available. Furthermore, the study aims were not stated in the methodology.
Other bias	Low risk	No evidence of funding bias

Moen 1997.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: not specified. Accepted for publication July 16, 1996 Number of study centres and location: single centre. Ullevål Hospital, University of Oslo, Oslo, Norway
Participants	Twenty‐four low‐risk patients accepted for elective CABG Exclusion criteria: EF < 30%; known pulmonary, renal, or hepatic failure; insulin‐dependent diabetes mellitus; active inflammatory or infectious disease; use of anti‐inflammatory drugs in the previous 8 days
Interventions	Intervention: heparin‐coated CPB Control: uncoated CPB
Outcomes	Expression of the chosen cell surface markers and complement activation C3bc, degranulation of neutrophils, and changes in leucocyte and Plt counts Test samples were drawn from the arterial line at the start of CPB, after 10 minutes of circulation, at the end of CPB, and 10 minutes after protamine administration.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not specify whether the trial was blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not specify whether the trial was blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the trial and were included in the analysis.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	Quote: "This study was supported by the European Working Group on Heparin‐Coated Extracorporeal Circulation Circuits. Financial support was provided by The Norwegian Council on Cardiovascular Disease." Funding was provided by a national body.

Morariu 2005.

Study characteristics
Methods	Prospective, double‐blind, placebo‐controlled randomised trial Run‐in period: not specified. Manuscript received March 17, 2005; revision accepted April 13, 2005 Number of study centres and location: single centre. University Medical Center Groningen, Groningen, the Netherlands
Participants	Twenty patients scheduled for first‐time coronary artery revascularisation with CAD, normal renal function (as assessed by a serum Cr level <120 μmol/L and normal urinalysis results) and normal hepatic, cerebral, and cardiac function (EF > 45%) Exclusion criteria: diabetes, recent MI, unstable angina, or recent use of radio‐contrast agents or corticosteroids
Interventions	Intervention: dexamethasone, 1 mg/kg, at induction of anaesthesia and 0.5 mg/kg 8 hours later Control: placebo at the same time points
Outcomes	Effects of dexamethasone on cytokine release and perioperative myocardial, pulmonary, renal, intestinal, and hepatic damage, as assessed by specific and sensitive biomarkers: IL‐6, IL‐8, IL‐10, CRP, tryptase, H‐FABP, TnI, CK‐MB, NAG, I‐FABP, alpha GST
Notes	Age SD was calculated according to the Cochrane Handbook from 95% CI. IL‐8 was measured from a graph at sternum closure and IL‐6 from a graph at 6 hours in ICU. IL‐10 was not measurable from a graph at 24 hours. CRP levels were not widely reported.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description regarding randomisation technique
Allocation concealment (selection bias)	Unclear risk	No clear description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Authors state that the trial was planned in a double‐blind fashion but do not describe how this was achieved.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors state that the trial was planned in a double‐blind fashion but do not describe how this was achieved or whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "All 20 patients included completed the study and survived the hospital stay." All participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Not all prespecified variables were extensively reported.
Other bias	Low risk	No clear evidence of funding bias

Moscarelli 2019.

Study characteristics
Methods	Dual‐centre, randomised controlled trial Run‐in period: February 2013 to June 2015. Number of centres and location: 2 centres. National Heart and Lung Institute, Cardiothoracic Surgery Department, Imperial College London, Du Cane Road, W12 0NN London, UK; Bristol Heart Institute, University of Bristol, Bristol Royal Infirmary, Marlborough Street, BS2 8HW Bristol, UK
Participants	One hundred twenty‐four adult patients undergoing first‐time isolated CABG and isolated AVR with CPB Exclusion criteria: patients who have experienced cardiogenic shock or cardiac arrest, suffer from significant peripheral arterial disease affecting the upper limbs or have neither upper limb available for the intervention, have preoperative renal failure (with a GFR < 30 mL/min/1.73 m2), or have been taking glibenclamide or nicorandil within 24 hours of surgery
Interventions	Intervention: RIPC on the upper arm after induction anaesthesia but before sternotomy. RIPC comprised four 5‐minute cycles of upper limb ischaemia, induced by a blood pressure cuff inflated to 200 mmHg, with an intervening 5 minutes of reperfusion by deflating the cuff. For participants with systolic blood pressure > 185 mmHg, the cuff will be inflated at least 15 mmHg above the participant’s systolic blood pressure. Control group: sham. A research nurse will position the cuff for administering RIPC as for the intervention group, under the operating cover, and will squeeze the bulb with the air valve opened at the same intervals as in the intervention group. This will ensure that all operating staff other than the research nurse placing the cuff will perceive the cuff to be inflated for all participants.
Outcomes	The primary outcome was myocardial injury, assessed by measuring myocardial TnI in serum from blood samples. The secondary outcomes will be obtained from data collected from myocardial biopsies and blood samples. Seven blood samples will be collected from each participant: two samples will be collected before the sternotomy is performed and before CPB (one before RIPC and one after RIPC) is started, and 5 postoperative samples will be collected at 6 hours, 12 hours, 24 hours, 48 hours, and 72 hours after the end of ischaemic cardioplegic arrest. Plasma concentrations of IL‐6, IL‐8, IL‐10, and TNF‐α were also measured.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random allocations, stratified by surgical stratum and centre, will be generated by computer".
Allocation concealment (selection bias)	Low risk	Accessed using a secure, internet‐based randomisation system to guarantee concealment until a participant’s identity and eligibility is confirmed and securely documented
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "A research nurse will apply the cuff for administering RIPC under the operating cover, placing the cuff around the participant’s upper arm, after induction of anaesthesia and before sternotomy, CPB and cardioplegic arrest. RIPC will be induced as described by others… A research nurse will position the cuff for administering RIPC as for the intervention group, under the operating cover and will squeeze the bulb with the air valve opened, at the same intervals as in the intervention group. This will ensure that all operating staff other than the research nurse placing the cuff will perceive the cuff to be inflated for all participants." Participants were blinded because the intervention is delivered after induction of anaesthesia, and there will be no visible signs of having had RIPC. The operating staff and data collectors were blinded as per the quote above.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The operating staff and data collectors will be blinded because a research nurse, who will have no further role in the trial, will deliver RIPC or sham RIPC." Data collectors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Prospectively registered study with published protocol. All prespecified variables were reported.
Other bias	Low risk	The study was funded by a grant from the British Heart Foundation.

Murphy 2011.

Study characteristics
Methods	Two centres, randomised, double‐blind, placebo‐controlled investigation Run‐in period: not specified. Published in 2011 Number of study centres and location: 2 centres. NorthShore University HealthSystem and Northwestern University Feinberg School of Medicine, Chicago, USA
Participants	One hundred seventeen patients undergoing elective CABG surgery with CPB or single valvular repair/replacement surgery and anticipated early tracheal extubation Exclusion criteria: combined (CABG/valve) procedures, EF < 30%, preoperative use of inotropic agents or an IABP, preoperative use of steroids or antiemetic agents, acute or chronic renal failure, pulmonary disease necessitating oxygen therapy or any patient assessed as potentially requiring prolonged postoperative mechanical lung ventilation, poorly controlled diabetes, poor English comprehension or psychiatric/central nervous system disturbances precluding completion of the QoR‐40 questionnaire
Interventions	Intervention: dexamethasone (8 mg in 2 mL total volume) at the time of anaesthetic induction (approximately 45 minutes before surgical incision) and on initiation of CPB Intervention: 0.9% sodium chloride (2 mL total volume) at the same time points
Outcomes	The primary aim of this investigation was to determine the effect of dexamethasone on postoperative QoR‐40 scores (a 40‐item scoring system specifically developed to measure patients’ health status).
Notes	No relevant inflammatory outcomes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomized to receive either dexamethasone (dexamethasone group) or saline placebo (control group) using a computer‐generated randomization code".
Allocation concealment (selection bias)	Low risk	Quote: "The randomization code for the 117 subjects was provided to the operating room pharmacy before the start of the study. Study medications were prepared by the operating room pharmacy in a 3‐mL syringes labeled with the patient’s name".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Study medications were prepared by the operating room pharmacy in a 3‐mL syringes labeled with the patient’s name". All clinicians and researchers were blinded to group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	As the operating room pharmacy was responsible for preparing the syringes, all other clinicians and researchers were blinded to group assignment.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	One hundred seventeen participants were randomised. Seven participants were excluded in the dexamethasone group and one in the control group. Data analysis included 109 participants (60 in dexamethasone group, 20 CABG/40 AVR, and 49 in control group, 22 CABG/27 AVR). At follow‐up, three participants were excluded in the dexamethasone group and two participants in the control group. It is unclear of the overall impact on the study conclusion.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Nakanishi 2006.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled clinical study Run‐in period: N/R Number of study centres and location: N/R
Participants	Thirty participants were randomised into an intervention group (n = 15) or a control group (n = 15). Exclusion criteria: emergency, previous heart surgery, valve or combined CABG and valve surgery, age > 75 years, LVEF < 0.45, diabetes treated with insulin, active gastropathic disorder, treatment for COPD, and preoperative use of steroids
Interventions	Intervention: UST (5000 units/kg IV immediately before aortic cannulation) Control: 0.9% sodium chloride at the same time
Outcomes	Neutrophil elastase, TNF‐α, IL‐6, IL‐8, Hct, CK‐MB, troponin T, myosin light chain I, haemodynamic (stroke volume index), and pulmonary data (respiratory index and intrapulmonary shunt)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by computer‐generated sequence.
Allocation concealment (selection bias)	Low risk	An anaesthesia research nurse prepared the syringes of blinded solution.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: all physicians and nursing staff were unaware of the treatment groups.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	All physicians and nursing staff were unaware of the treatment groups. No mention of blinding of outcome assessment was made.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Two participants were excluded from the study because of MI.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	The authors have no financial interests to disclose.

Narayan 2011.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: January 2003 to October 2004. Received 28 April 2010; received in revised form 10 August 2010; accepted 16 August 2010; Available online 29 September 2010 Number of study centres and location: single‐centre. Bristol Heart Institute, University of Bristol, Bristol, UK
Participants	Eighty‐two patients undergoing elective, multivessel, isolated, primary CABG (40 on pump without cardioplegic arrest and 41 on pump with cardioplegic arrest) Exclusion criteria: an MI within the last 7 days; severe LV impairment (EF < 30%); symptomatic cerebrovascular disease; renal insufficiency (serum Cr>130 μmol/L); presence of bleeding diathesis, bullous emphysema on CXR, or significant chronic obstructive airway disease (FEV1/FVC < 40%); history of pericarditis, median sternotomy, thoracotomy, chest irradiation, or pleurodesis; smoking 1 month prior to surgery; significant peripheral vascular disease; and use of aspirin within 4 days prior to surgery
Interventions	Intervention: on‐pump CPB with coronary grafting performed on the beating heart Control: on‐pump CPB with cardiac arrest
Outcomes	To evaluate the surgical technique of on‐pump without CA versus conventional surgery by assessing its impact on inflammation (IL‐6, IL‐8, and IL‐10 preoperatively and at 1 hour, 4 hours, 12 hours, and 24 hours after surgery); myocardial (TnI preoperatively and at 1 hour, 4 hours, 12 hours, 24 hours, and 48 hours, postoperatively), cerebral (serum S100 protein preoperatively and 1 hour, 4 hours, 12 hours, and 24 hours postoperatively), and renal injury (NAG and urinary Cr were measured from urine samples collected immediately before and 1 hour, 24 hours, and 48 hours after CPB); early health outcome; and 5‐year event‐free survival in elective patients undergoing CABG
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random treatment allocations, stratified by consultant team (two strata), were generated by computer in blocks of unequal length."
Allocation concealment (selection bias)	Low risk	Quote: "Allocation details were concealed in sequentially numbered, opaque sealed envelopes."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unclear whether patients were blinded to the interventions. They could have been blinded. Theatre personnel were not blinded because of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not clearly describe whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis. One participant had to be switched from noncardiac arrest group to cardiac arrest because of technical difficulties. All participants completed the study.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Nederlof 2017.

Study characteristics
Methods	Single‐centre, randomised controlled clinical trial Run‐in period: not specified. Published 2017 Number of study centres and location: single‐centre. Academic Medical Center, Amsterdam, The Netherlands
Participants	Male participants (> 18 years) undergoing elective first‐time isolated ONCABG surgery Exclusion criteria were diabetes mellitus, unstable angina pectoris, increased baseline troponin levels, concomitant procedures, severe COPD, EF < 40%, MI within 2 weeks before surgery, peripheral vascular disease affecting the upper limbs, female, and nicorandil use.
Interventions	Intervention: RIPC (3 × 5 minutes with BP cuff inflated to 200 mmHg) before the start of surgery Control: no treatment; unclear whether sham procedure performed
Outcomes	Primary outcome in this study was cTnT. Secondary outcome measures were mitochondrial hexokinase binding and activity after RIPC; Akt and AMPK phosphorylation after RIPC; CRP levels before and 6 hours, 12 hours, 24 hours, and 48 hours after surgery; and MIF, IL‐6, IL‐10, and TNF‐α levels before and after RIPC. Blood was collected for cTnT and CRP determination before and 6 hours, 12 hours, 24 hours, and 48 hours after surgery.
Notes	Continuous infusion of sufentanil 0.3 μg/kg/h and sevoflurane were used for maintenance of hypnosis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "In the operating room, patients were randomized in blocks of 6 to control or RIPC treatment using a computer program (ALEA) by the researcher."
Allocation concealment (selection bias)	Low risk	Quote: "In the operating room, patients were randomized in blocks of 6 to control or RIPC treatment using a computer program (ALEA) by the researcher."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Patients, anesthetists, surgeons and Intensive Care Unit (ICU) staff were blinded to treatment allocation."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Samples were in a blinded manner."
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "6 patients were excluded because surgery was performed off‐pump (2), propofol was used at induction of anesthesia (1), increased pre‐operative levels of cTnT (1), concomitant procedures were performed (1) or it was unable to perform the protocol during surgery (1)." Six participants out of 36 did not complete the study for a variety of reasons. This represents around 15% of the sample size and further reduces the power of a study which is already underpowered.
Selective reporting (reporting bias)	Low risk	Study was registered at the Netherlands Trial Register (NTR2915; www.trialregister.nl/trialreg/admin/rctview.asp?TC=2915). All prespecified outcomes were reported; however, mitochondria samples were not available for five participants.
Other bias	Low risk	This study was supported by the Dutch Heart Foundation (CJZ; NHS2010B011).

Nesher 2006.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: January to June 2004 Number of study centres and location: single centre. Tel‐Aviv Sourasky Medical Center and Sackler Faculty of Medicine, Tel‐Aviv University, Tel‐Aviv, Israel
Participants	One hundred twenty patients for non‐emergent CABG, aged 40 years to 80 years with LVEF > 25% (as assessed by echocardiography or by angiographic contrast left ventriculography) Exclusion criteria: concomitant debilitating noncardiac disease, severe peripheral vascular disease (defined by a history of intermittent claudication within walking distance of < 100 m), uncontrolled insulin‐dependent diabetes mellitus (preoperative fasting glucose levels > 250 g/dL), fever or infection within 1 week prior to surgery, recent steroidal or antifibrinolytic therapy (within the last 6 weeks), or clinically significant laboratory abnormalities (Cr ≥ 2.0 mg/dL, TB ≥ 1.5 mg/dL, Hb ≤ 10.0 g/dL, Plt count ≤ 100,000 cells/mL, clotting abnormality, or WBC count < 3000 or > 14,000 cells/mL3)
Interventions	Intervention: OPCABG Control: CCABG
Outcomes	Haemodynamic and respiratory data, serum CK‐MB mass fraction, TnI, and IL‐6, IL‐8, and IL‐10 levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients then gave written informed consent and were randomized by means of a computer‐generated random number table to either CCAB or OPCAB on the day of surgery".
Allocation concealment (selection bias)	Unclear risk	Allocation concealment procedures were not clearly described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The study design was blinded to the surgeons who assigned the patients to surgery, to the operating room (OR) team, and to the postoperative ward nursing staff".
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "All perioperative data were recorded in real time by an independent technician, starting at patient’s placement on the OR table until 24 hours after surgery. Invasive arterial blood pressure, heart rate, pulmonary artery pressure, and cardiac index were continuously recorded".
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Intraoperative conversion from OPCAB to CCAB could take place if hemodynamic stability could not be achieved during stabilization of the heart, despite aggressive fluid or catecholamine administration, or if the coronary vessels turned out to be small and atheromatotic, calcified, or deeply intramural, thus unsuitable for OPCAB. Intraoperative conversion from CCAB to OPCAB could be carried out if a porcelain aorta was found. Epi‐aortic echo was performed only on those with high index of suspicious calcification seen on angiogram, or chest X‐rays, or during palpation upon the exposure of the aorta. On both occasions, patients were dropped from the study to eliminate bias and were replaced by others using the same randomization method". Although the switch from one group to the other was specified in the protocol, this risk introducing bias is high as participants were excluded from the trial and replaced.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes were reported.
Other bias	Low risk	No evidence of funding bias

Ng 2015.

Study characteristics
Methods	Prospective, single‐blinded, randomised controlled trial Run‐in period: February 2009 and December 2012 Number of centres and location: single centre. Authors affiliated to different institutions in Singapore
Participants	Seventy‐eight patients undergoing elective isolated CABG, Asian patients, aged between 21 and 85 years, no previous cardiac surgery Exclusion criteria: poor LVEF (< 30%), immunologic disease or malignancies, acute inflammatory disease, coagulopathy, steroid treatment, significant carotid disease, and on dialysis
Interventions	Intervention group: miniaturised CPB. This is a closed loop including a venous bubble trap, centrifugal pump, heat exchanger, and oxygenator (all integrated into a single device). The system includes a hardware platform based on the Stöckert heart‐lung machine. The system is provided in a 100% PC‐coated circuit. A cell‐saver device was used to suction blood from the field. Control group: conventional circuit consisted of a nonpulsatile roller pump (Stöckert Instruments, Munich, Germany) and PC‐coated tubing circuit. A cell‐saver device was used to suction blood from the field.
Outcomes	TNF‐α, IL‐6, CRP, LDH, and postoperative clinical outcomes (AKI, pulmonary oedema, length of stay, and mortality), coagulation profile (blood loss, Plt count, FFP and Plt transfusion) Data were collected at baseline, upon arrival at the ICU, and at 6 hours, 12 hours, 24 hours, and 48 hours in the ICU.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "These 78 patients were then randomly assigned to the MCPB or the CCPB group, using the sealed envelope technique."
Allocation concealment (selection bias)	Low risk	Quote: "These 78 patients were then randomly assigned to the MCPB or the CCPB group, using the sealed envelope technique."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Authors state it was a single‐blinded study.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Authors state that this was a single‐blinded study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	Quote: "The authors declare that there is no conflict of interest. This research was supported by a grant from the National Medical Research Council, Singapore." Grant funding was received from a national organisation; therefore, this was considered to be at low risk of bias.

Nguyen 2016.

Study characteristics
Methods	Parallel‐group, randomised controlled trial Run‐in period: not specified Accepted for publication: 8 September 2015 Number of study centres and locations: single centre, United Kingdom
Participants	From 50 potentially eligible participants: 24 participants were excluded, and 26 participants were randomised to receive either cCPB (n = 13) or mCPB (n = 13). Mean age: 67.35 years Sex (female/male ratio): 30.77% No high‐risk patients Inclusion criteria: primary isolated CABG performed by a single surgeon Exclusion criteria: < 18 years of age, emergency operations, EF < 30%, recent CVA, carotid artery stenosis > 75%, renal impairment (serum Cr level > 177 mmol/L), pre‐existing coagulopathy, pre‐existing liver dysfunction or recent (within 5 days) use of anti‐Plt agents (aspirin/clopidogrel)
Interventions	Intervention group: miniaturised CPB Control group: CCPB
Outcomes	Primary outcome: ROS in granulocytes Secondary outcomes: ROS in monocytes and lymphocytes, NF‐kB, p38 MAPK within leucocytes, leucocyte accumulation in cantharidin‐induced blisters, WCCs, serum CRP and Cr levels Blood samples were collected preoperatively and five times post‐CPB within 5 hours.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were randomised by the Imperial College Trial Unit staff not otherwise involved in the study. Random allocations were blocked and generated by computer. Randomisation was carried out by 1 of the investigators after the patient’s eligibility had been checked and written informed consent obtained."
Allocation concealment (selection bias)	Low risk	Quote: "Random allocations were blocked and generated by computer. The random allocations were concealed at the time of recruitment."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "Participants were blinded to the allocation." It is unclear whether healthcare personnel were blinded to trial arm allocation, although highly likely. Therefore, we have rated the risk of bias for this domain as unclear.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is unclear whether outcome assessors were blinded to trial arm allocation, although highly likely. Therefore, we have rated the risk of bias for this domain as unclear.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "All data analyses were performed on an intention‐to‐treat basis irrespective of the final actual treatment modality received" and "All patients allocated to cCPB had operative treatment performed as allocated. One participant allocated to mCPB underwent surgical intervention with cCPB because the perfusionist was not fully trained to use mCPB."
Selective reporting (reporting bias)	Low risk	Prospectively registered protocol was available to access. All expected outcomes were reported properly.
Other bias	Low risk	Quote: "The study was funded by Heart Research UK, the British Heart Foundation and the National Institute for Health Research Bristol Biomedical Research Unit in Cardiovascular Medicine. Patient recruitment was made possible, in part, by funding through the National Institute of Health Research Comprehensive Biomedical Research Centre at Imperial College Healthcare NHS Trust." As funding was received from a national funding body, we rated the risk of funding bias as low.

Ninomiya 2003.

Study characteristics
Methods	Prospective, randomised clinical study Run‐in period: November 2000 to January 2001. Accepted for publication September 2002 Number of study centres and location: single centre. Faculty of Medicine, University of Tokyo, Tokyo, Japan
Participants	Twenty‐two participants were randomised to PMEA‐coated (group P; Capiox RX25; n = 11) or uncoated (group U; Capiox SX10; n = 11) circuit group. Mean age: 62.2 years Sex (female/male ratio): 40.9% No high‐risk patients Inclusion criteria: CABG or valve operations, adult, first‐time elective surgery, absence of major noncardiac illness, normal liver function, and no history of recent use of steroids, acetylsalicylic acid, or anti‐inflammatory drugs Exclusion criteria: major noncardiac illness, abnormal liver function, and history of recent use of steroids, acetylsalicylic acid or anti‐inflammatory drugs
Interventions	Intervention: PMEA‐coated bypass circuits and oxygenators (Capiox RX25, Terumo Corp, Tokyo, Japan) Control: uncoated circuit group (Capiox SX10, Terumo Corp)
Outcomes	Clinical outcomes (major postoperative complications, the timing of extubation, and the duration of ICU stay), C3a (including C3a‐desArg), polymorphonuclear elastase concentrations, IL‐6 concentrations, number of Plts Blood samples were taken at five time points: immediately before CPB, 1 hour and 2 hours after start of CPB, and 1 hour and 24 hours after the start of protamine infusion.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The randomisation was conducted using a coin toss method by one of the investigators. No significant differences were observed between groups.
Allocation concealment (selection bias)	Unclear risk	The randomisation was conducted using a coin toss method by one of the investigators, but no further allocation method description is given.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The randomisation was conducted using a coin toss method by one of the investigators, and the other medical staff members were not informed of the randomisation throughout the study.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The randomisation was conducted using a coin toss method by one of the investigators, and the other medical staff members were not informed of the randomisation throughout the study. It is unclear whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Nollert 2005.

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2003 to 2004. Accepted for publication May 17, 2005 Number of study centres and location: single centre. University of Munich, Munich, Germany
Participants	Thirty participants were prospectively randomly assigned to undergo CABG with CCPB or MCPB. Mean age: 67.9 years Sex (female/male ratio): 23% No high‐risk patients Inclusion criteria: elective, isolated CAB surgery Exclusion criteria: combined surgical procedures, age < 18 years, weight > 80 kg, severe neurological or psychiatric disorder, acute infection, use of steroids, emergencies, MI or use of clopidogrel within the preoperative week, coagulation disorders with INR > 2, anaemia with Hb values <8 g/dL, terminal renal insufficiency with dialysis, oncologic diseases, reoperation, preoperative IABP, and LVEF < 30%
Interventions	Intervention: miniaturised closed CPB without cardiotomy suction Control group: CCPB
Outcomes	Inflammation (IL‐2 receptor, IL‐6, IL‐10, TNF receptor 55 and 75, CRP, leucocyte differentiation), coagulation and use of blood products (d‐dimers, fibrinogen, and thrombocytes, major complication, INR, partial thromboplastin time, antithrombin III), clinical outcome (arterial blood pressure, use of inotropes, or ventilation, ICU stay, hospital stay), routine parameters (Hb, CK, CK‐MB, TnI, blood cell count), and intraoperative handling Data were collected at six time points: preoperatively; 30 minutes after commencement of CPB; 15 minutes after declamping the ascending aorta; at the end of surgery; 6 hours postoperatively; and on the postoperative morning.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation method was not described.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment methods were not described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not specify whether the trial was blinded and, if so, how blinding was achieved. It would have been difficult to blind operating theatre personnel because of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not specify whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "The study was initially planned as a larger trial including 30 patients in each study arm. However, safety concerns and no perceivable benefit for the patients in the MCPB group led to premature discontinuance of the study." Although it was appropriate to terminate the study early on safety grounds, it is unclear whether this would have significantly impacted the trial findings.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	High risk	Quote: "Medtronic, Inc, funded the laboratory investigations and documentation of the study. All cardiopulmonary bypass circuits, oxygenators and further equipment were purchased. The authors had full control of the design of the study, methods used, outcome parameters, analysis of data and production of the written report." The study received funding from the manufacturers of the miniaturised CPB system. This could have introduced bias.

Ohata 2007.

Study characteristics
Methods	Prospective, randomised trial Run‐in period: June 2005 to August 2006. Received: December 12, 2006; accepted: January 23, 2007 Number of study centres and location: single centre. Hyogo College of Medicine, 1‐1 Mukogawa‐cho, Nishinomiya, Hyogo 663‐8501, Japan
Participants	Thirty adult participants undergoing CABG Mean age: 68.5 years Sex (female/male ratio): not specified No high‐risk patients Exclusion criteria: not specified
Interventions	Intervention: closed mini‐CPB system Control: CCPB
Outcomes	IL‐6, IL‐8, neutrophil elastase levels, neutrophil count and CRP, intraoperative blood loss, and transfusion volume Serial blood samples were taken preoperatively, after surgery, and on PODs 1 and 2.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	All consecutive participants had similar pre‐ and intraoperative characteristics, but no randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment processes
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not specify whether the study was blinded in any way.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not specify whether the study was blinded in any way.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Ohata 2008.

Study characteristics
Methods	Prospective, randomised study Run‐in period: 2002 to 2006. Submitted for consideration July 2007; accepted for publication in revised form November 2007 Number of study centres and location: single centre. Hyogo College of Medicine, Nishinomiya, Hyogo 663‐8501, Japan
Participants	Ninety‐eight participants were randomly assigned to mini‐CPB (n = 34) or CCPB (n = 64). Mean age: 67.73 years Sex (female/male ratio): not specified No high‐risk patients Inclusion criteria: ischaemic heart disease, CABG Exclusion criteria: OPCABG, emergent CABG, and blood priming cases such as haemodialysis
Interventions	Intervention: closed mini‐CPB system Control group: CCPB
Outcomes	IL‐8 and neutrophil elastase; blood loss and haemodilution ratio; other clinical outcomes such as Type I or II neurological injury, renal failure, wound complication, or Hct Serial blood samples for this purpose were taken preoperatively, immediately after the operation, and on PODs 1 and 2.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	All consecutive participants had similar pre‐ and intraoperative characteristics, but no further randomisation method description is given.
Allocation concealment (selection bias)	Unclear risk	The authors do not describe how allocation to different treatment arms was concealed.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not describe whether the trial was blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not describe whether the trial was blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Sample size of control group is twice the size of intervention group.
Other bias	Low risk	No funding was disclosed.

Olivencia‐Yurvati 2003.

Study characteristics
Methods	Randomised, prospective controlled study Run‐in period: not specified. Published 2003 Number of study centres and location: single centre. University of North Texas Health Science Center, Fort Worth, TX, USA
Participants	Two hundred twenty‐five participants were randomised to into a control group (n = 110) or a study group (n = 115). Mean age: 63.49 years Sex (female/male ratio): 45.78% No high‐risk patients Inclusion criteria: isolated CABG, adult, use of blood cardioplegia as a technique for myocardial preservation Exclusion criteria: none specified
Interventions	Intervention group: leucocyte depletion. The leucoreduction study group (n = 115) received only leucoreduced transfused (allogeneic) blood (packed cells, platelets, and fresh frozen plasma) and leucoreduced re‐infused salvaged (autologous) blood (RSI; Pall Biomedical Products, East Hills, NY, USA). In this group, total leucoreduction and strategically implemented filtration were accomplished by passing blood through a leucoreducing arterial line filter (LGB; Pall Biomedical Products, East Hills, NY, USA). A leucodepleting cardioplegia line filter was also used. In addition, the study group received standard Hammersmith dosing of aprotinin. Control group: moderately hypothermic CBP alone (no leucocyte depletion) and full‐dose aprotinin Timing: 30 minutes before pulmonary vascular reperfusion and before cross‐clamp release
Outcomes	PMVP, shunt fractions, pulmonary artery pressure, incidence of pulmonary dysfunction, postoperative total ventilator time, hospital stay, and total inpatient charges These variables were measured prior to bypass and at 1 hour, 3 hours, 6 hours, 12 hours, 18 hours, and 24 hours post‐bypass.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The randomisation method was not described.
Allocation concealment (selection bias)	Unclear risk	The allocation concealment method was not described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Authors NT and RTM were blinded as to which group was strategically leukodepleted".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "Authors NT and RTM were blinded as to which group was strategically leukodepleted". Two of the study authors were blinded to intervention; however, it is unclear who collected the data and whether they were blinded to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants were excluded from data analysis, and there were no operative deaths.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Onorati 2010.

Study characteristics
Methods	Prospective randomised control trial Run‐in period: not specified. received 16 July 2009; received in revised form 3 November 2009; accepted 6 November 2009; available online 16 December 2009 Number of study centres and location: single centre. Magna Graecia University of Catanzaro, Catanzaro, Italy
Participants	Sixty adult patients undergoing isolated elective primary CABG Mean age: 71.5 years Sex (female/male ratio): 22.5% No high‐risk patients Exclusion criteria: chronic renal insufficiency or failure, liver failure, COPD, abdominal aortic aneurysm with abdominal arteriopathy, autoimmune disease, severe anaemia (Hb < 8 g/dL), known coagulation disorder, or unstable angina
Interventions	Intervention: OPCABG Control group: standard linear CPB
Outcomes	Primary outcomes: endothelial activation (VEGF, MCP‐1) and IL‐10 Secondary outcomes: inflammatory response (IL‐2, IL‐6, and IL‐8), intubation times, ICU stay, and hospital stay Blood was collected from the peripheral arterial line preoperatively (time 0), after construction of proximal anastomoses in pulsatile and linear groups or completion of the last distal anastomosis in OPCABG group (time 1), at the end of surgery (time 2), and at 12 hours (time 3) and 24 hours (time 4) postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Eligible patients were randomised to receive PCPB, LCPB or OPCABG by lottery, drawing from sealed envelopes containing the group assignment."
Allocation concealment (selection bias)	Low risk	Quote: "Eligible patients were randomised to receive PCPB, LCPB or OPCABG by lottery, drawing from sealed envelopes containing the group assignment."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors do not specify whether the study was blinded. It would have been difficult to blind operating theatre personnel to treatment group allocation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors do not specify whether the study was blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study and were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes were reported.
Other bias	Low risk	No funding sources were disclosed.

Onorati 2011.

Study characteristics
Methods	Prospective, randomised study Run‐in period: December 2009 to April 2010 Number of study centres and location: two centres. University of Verona Medical School, Verona; Varese University Hospital, University of Insubria, Varese, Italy
Participants	Sixty adult participants admitted to two university institutions because of degenerative calcified severe aortic stenosis and scheduled for elective primary AVR Mean age: 76.35 years Sex (female/male ratio): 25% High‐risk patients Inclusion criteria: from mild to moderate COPD (FEV1 40% to 80%), degenerative calcified severe aortic stenosis, elective primary AVR Exclusion criteria: low risk for post‐pump disease (Global Initiative for COPD stage 0) or at very high risk for post‐pump lung disease (severe COPD, i.e. Global Initiative for COPD stage IIb or III); age < 65 years or > 85 years; emergent, urgent, or salvage procedures; LVEF < 40%; associated cardiac or vascular surgical procedures; re‐do surgery; recent (< 8 weeks) AMI; preoperative IABP assistance; obesity (BMI > 30); current smoker; renal disease; previous irradiation; previous thoracic surgery; previous transfusion (within 6 months); recent (< 30 days) infections; liver dysfunction; ongoing steroid therapy; alcohol or drug abuse; neurologic diseases or recent (< 3 months) stroke; acquired/congenital deficits of the immune system; autoimmune diseases; or cancer)
Interventions	Intervention group: a LeukoGuard LG arterial filter connected to the arterial line and a second LeukoGuard BC2 cardioplegia filter connected to the cardioplegic line throughout the duration of CPB Control group: standard CPB with standard arterial filters
Outcomes	Primary outcome: AaDO2 Secondary outcomes: inflammatory markers (IL‐6, IL‐8, TNF‐α, IL‐10, perioperative serum neutrophils, and Plt count), other determinants of respiratory function (PaO2/FiO2 lung compliance, rLIS, fluid balance, cardiac index (CI) and ISVR) and clinical data (length of intubation, ICU stay, hospital stay, need for non‐invasive positive‐pressure ventilation, acute lung injury, pneumonia, haemodynamic status, major postoperative morbidity (transient low output stay, paroxysmal AF, type I and type II neurologic outcomes, acute renal insufficiency, re‐exploration for bleeding, hospital mortality, postoperative chest drainage, total and red packed cell transfusions)) Data were collected from peripheral arterial blood at ICU arrival and at 24 hours, 48 hours, and 72 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The day of surgery, patients were randomly allocated by lottery, drawing previously prepared sealed envelopes containing the group assignment."
Allocation concealment (selection bias)	Low risk	Quote: "The day of surgery, patients were randomly allocated by lottery, drawing previously prepared sealed envelopes containing the group assignment."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "With the exception of the main investigator and of the CPB technician, all the other physicians (surgeons, anesthesiologists, biochemists, and so forth) dealing with perioperative care, biological sampling and collection and analysis of data were blinded toward the group assignment until the end of the study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "With the exception of the main investigator and of the CPB technician, all the other physicians (surgeons, anesthesiologists, biochemists, and so forth) dealing with perioperative care, biological sampling and collection and analysis of data were blinded toward the group assignment until the end of the study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No sources of funding were disclosed.

Onorati 2013.

Study characteristics
Methods	Prospective, randomised study Run‐in period: April 2010 to January 2011. Received 20 October 2011; received in revised form 3 January 2012; accepted 5 January 2012 Number of study centres and location: single centre. University of Verona Medical School, Verona, Italy
Participants	Thirty participants were randomised to leucocyte filtration of blood cardioplegia (n = 15) or to standard cold blood cardioplegia (S‐Group, n = 15). Mean age: 75.65 years Sex (female/male ratio): 26.67% High‐risk patients Inclusion criteria: degenerative calcified severe aortic valve stenosis, elective primary isolated AVR Exclusion criteria: coronary disease, age < 60 years or > 85 years, emergent/urgent/salvage procedures, LVEF < 35%, associated cardiac or vascular surgical procedures, redo surgery, recent (< 8 weeks) AMI, preoperative IABP assistance, severe COPD (≥ stage IIIa of GOLD classification), obesity (BMI > 30), renal disease (KDOQI class ≥ 2), previous irradiation, previous thoracic surgery, previous transfusion (within 6 months), recent (< 30 days) infections, liver dysfunction, ongoing steroids or statin therapy, drug abuse, neurological diseases or recent (< 3 months) stroke, acquired/congenital deficits of the immune system, autoimmune diseases, or cancer
Interventions	Intervention group: leucocyte filtration during CPB Control group: standard CPB without leucocyte filtration
Outcomes	Primary outcome: myocardial damage (TnI and lactate) Secondary outcomes: cytokines (IL‐6, IL‐8, TNF‐α, IL‐10), perioperative haemodynamic indices (cardiac index, ISVR, CCE, and CVP), and clinical outcome (the need for intraoperative ventricular defibrillation after aortic declamping, the length of intubation, perioperative AMI, transient low output syndrome, paroxysmal AF, acute renal insufficiency, major morbidity, inotropic support (dose and duration in hours), ICU stay, hospital stay, hospital mortality, and the leading cause of death) Blood samples were taken from CS and peripheral vessel at admission and in the ICU at 6 hours, 8 hours, 12 hours, 18 hours, 24 hours, 36 hours, and 60 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomized by lottery on the day of surgery (drawing pre‐prepared sealed envelopes containing the group assignment) to either cLkF during CPB (15 patients, cLkF‐Group) or standard CPB without cLkF (15 patients, S‐Group)."
Allocation concealment (selection bias)	Low risk	Quote: "Patients were randomized by lottery on the day of surgery (drawing pre‐prepared sealed envelopes containing the group assignment) to either cLkF during CPB (15 patients, cLkF‐Group) or standard CPB without cLkF (15 patients, S‐Group)."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "With the exception of the main investigator and the CPB technician, all other physicians (surgeons, anaesthesiologists, intensivists, biochemists, etc.) dealing with perioperative patient care, biological sampling and/or collection and analysis of clinical data were blinded to the group assignment until the end of the study."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "With the exception of the main investigator and the CPB technician, all other physicians (surgeons, anaesthesiologists, intensivists, biochemists, etc.) dealing with perioperative patient care, biological sampling and/or collection and analysis of clinical data were blinded to the group assignment until the end of the study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No conflict of interest

Otani 2008.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Received: 5 July 2007 Accepted: 22 October 2007 Number of study centres and locations: not specified
Participants	In total, there were 30 participants split into 2 groups, A and B, with 15 in each group. Thirty adult participants were randomly separated into two experimental groups and one control group of 10 participants each. Mean age: 63.77 years Sex (female/male ratio): 20% No high‐risk patients Inclusion criteria: first‐time CABG Exclusion criteria: recent MI, unstable angina, acute infection, known immunological disease, insulin‐dependent diabetes, acute or chronic renal failure, re‐do surgery, respiratory impairment, and coagulopathy
Interventions	Intervention group: PTX, a methylxanthine derivative (oral 900 mg/day) Control group: no PTX Timing: administration was started on preoperative day 5 and continued for 5 days.
Outcomes	Serum levels of PTX, respiratory index, IL‐6, CRP, and fibrinogen Blood samples were collected for serological assay just before the start of CPB (pre‐CPB) and at 4 hours after the completion of CPB (post‐CPB).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description was reported. However, there were no differences in baseline parameters between the two groups.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; no mention of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified; no mention of blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Ozguler 2015.

Study characteristics
Methods	Randomised controlled study Run‐in period: not specified Manuscript submitted 23 January 2015; manuscript accepted 30 May 2015 Number of study centres and locations: single centre. Firat University Hospital, Department of Cardiovascular Surgery, Elazig, Turkey
Participants	Fifty participants were randomly divided into two groups: control group (n = 25) and rosuvastatin‐treated group (n = 25). Mean age: 62.05 years Sex (female/male ratio): 26% No high‐risk patients Inclusion criteria: CABG because of CAD Exclusion criteria: corticosteroids, salicylates, dipyridamole or anticoagulant, receiving thrombolytic therapy within 5 days or who had platelet or coagulation disorders, hypoxaemia, hypercapnia, congestive heart failure, diabetes, chronic renal failure, COPD and hepatic insufficiency, heart rate < 60/min and EF < 30% or active infections or with a history of previous heart surgery, MI in the last month, thyroid disease or surgery, chronic antiarrhythmic drugs, emergency operations, additional surgical procedures
Interventions	Intervention group: 20 mg/d of rosuvastatin tablets Control group: placebo Timing: starting at preoperative day 7 through POD 28
Outcomes	IL‐6, IL‐10, IL‐18, and hs‐CRP and LDL, HDL, VLDL, TC, and TG Blood samples were taken at 6 time points: before induction of anaesthesia (T1), during CPB (T2), 5 minutes after removal of cross‐clamp (T3), after protamine infusion (T4), POD 3 (T5), and POD 28 (T6).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method description was reported. However, there were no differences in baseline parameters between the two groups.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method description
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified; no mention of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified; no mention of blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All expected outcomes were reported properly.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This study was supported by FUBAP (Firat University Scientific Research and Design Center). Project Number: TF: 1110 – 2012 Firat University Elazig, Turkey.

Pahari 2013.

Study characteristics
Methods	Randomised, prospective pilot study Run‐in period: between July and December 2010 Received: February 2012; revised April 2012 Number of study centres and locations: single centre. Department of Cardiothoracic and Vascular Surgery, Klinikum Braunschweig, Germany
Participants	Twenty‐eight participants were randomly divided into an MPC group (n = 14) and a CCPB group (n = 14). Mean age: 67 years Sex (female/male ratio): 28.57% No high‐risk patients Inclusion criteria: CABG, elective patients Exclusion criteria: concomitant procedure; EF < 50%; NYHA III to IV; acute coronary syndrome within the last 30 days; any pathological valve; previous cardiac surgery; intracardiac thrombus; pacemaker or AF; previous CVA; stenosis of an internal carotid artery > 70%; Cr > 1 mg/dL or dialysis; insulin‐dependent diabetes mellitus; peripheral arterial vascular disease; intake of nicotine; history of any malignancy, immunological disease, or preoperative immunosuppressive therapy; hepatitis; preoperative use of steroid and nonsteroidal anti‐inflammatory drugs; intake of warfarin, heparin infusion, or known coagulopathy; thrombocyte count < 105
Interventions	Intervention group: minimised perfusion circuit (ROCsafeRX, Terumo Corporation, Tokyo, Japan). It was a totally closed perfusion circuit and consisted of a centrifugal pump (Terumo Cardiovascular Systems, Ann Arbor, MI, USA), a hollow‐fiber oxygenator (Terumo RX 15, Terumo Corporation, Tokyo, Japan), and a 40‐mm arterial blood filter (AL8x, Pall, East Hills, NY, USA) in the arterial line. In the MPC group, pericardial blood was aspirated into a cell‐saver device, reprocessed, and retransfused when at least 300 mL blood could be collected. Control group: conventional circulatory perfusion bypass (Biomedicus, BP80, Medtronic, Minneapolis, MN, USA). It was an open perfusion circuit with cardiotomy reservoir (Hardshell reservoir, Maquet Cardiopulmonary AG, Rastatt, Germany) composed of a centrifugal pump (Biomedicus, BP80), membrane oxygenator (A. L. One, Eurosets, Medolla, Italy), arterial filter (Affinity 3838, Medtronic), and an arterial‐venous line (HMT GmbH, Maisach, Germany).
Outcomes	Brain biomarker carnosinase, BFABP, IL‐6, CRP, ET‐1, leucocyte count; Hb and Hct were measured systemically for all time points. Blood samples were taken at different time points: 5 minutes postanaesthetic application (T1), 5 minutes after sternum closing (T2), after 2 hours in ICU (T3), first POD (T4), second POD (T5), and fifth POD (T6).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patient randomised according to a computer generated algorithm"
Allocation concealment (selection bias)	Unclear risk	Unclear allocation concealment method: "Patients blindly randomised"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified blinding of surgical team. Participants likely to be blinded to treatment: "Patients blindly randomised". No mention regarding personnel blinding. However, surgical team was likely to be aware of the circuit used.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Pang 2016.

Study characteristics
Methods	Prospective, randomised study Run‐in period: between January 2012 and June 2014 Study date: 2016 Number of study centres and locations: single centre. Department of Cardiac Surgery, Qilu Hospital of Shandong University, Shandong, China
Participants	The study included 60 participants undergoing CPB who were randomly divided into a UTL group (n = 30) and a control group (n = 30). Mean age: 51.85 years Sex (female/male ratio): 40% No high‐risk patients Inclusion criteria: participants underwent CPB for CABG surgery, valve replacement, or combined operation. Exclusion criteria: previous cardiac surgery; severely impaired liver or kidney function; allergies; and pregnant and breastfeeding women.
Interventions	Intervention group: UST, 5000 U/kg UTL (Guangdong Tianpu Biochemical MEDICINE Co, Ltd, 12,000 U/teams, batch number: 20040505) intravenously in 50 mL normal saline for 20 minutes after the induction of anaesthesia, heparin, and protamine, respectively Control group: similar volumes of normal saline and similar timing
Outcomes	Perioperative inflammatory response (WBC count, IL‐2, IL‐8, IL‐6, TNF‐α, neutrophil elastin) and pulmonary function (PA‐aDO2, RI, PaO2, FiO2, oxygenation index, tidal volume, peak airway pressure, plateau airway pressure, PEEP, pulmonary Cd, and pulmonary Cs), PBMC apoptosis, TLR4 and HSP70 expressions, and clinical data (operation time, heart resuscitation, duration of postoperative mechanical ventilation, length of ICU stay, time of drainage tube usage, and postoperative hospitalisation length of stay) Blood samples were drawn through the internal jugular vein after anaesthetic induction (T1), immediately after aortic valve opening (T2), and 4 hours (T3) and 24 hours (T4) after weaning from CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method described; however, no significant differences in baseline characteristics between groups
Allocation concealment (selection bias)	Unclear risk	No allocation concealment information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors had no conflicts of interest to declare.

Papadopoulos 2013.

Study characteristics
Methods	Prospective, randomised trial Run‐in period: from April 2009 to January 2010 Study date: 2013 Number of study centres and locations: single centre, Department of Thoracic and Cardiovascular Surgery, J.‐W. Goethe University Hospital, Frankfurt, Germany
Participants	Fifty patients were assigned to a control group (n = 25) or a normovolaemic MUF N‐MUF group (n = 25). Mean age: 74.5 years Sex (female/male ratio): 42% High‐risk patients Inclusion criteria: elective, complex cardiac surgery (defined as combined procedures, double‐valve surgery, aortic surgery, and redo procedures), high‐risk patients, age > 65 years, logistic EuroSCORE II > 5%, and expected duration of ECC > 120 minutes Exclusion criteria: patients with an age < 65 years, malignancy, systemic inflammatory disease, and failure to obtain consent
Interventions	Intervention group: N‐MUF performed after protamine infusion. Ultrafiltration was normovolaemic and with no concentrating effect with a BC 140 plus haemofilter. Simultaneous infusion of 150 mL/min of crystalloid solution was performed, using a second roller pump to maintain normovolaemia. N‐MUF was stopped after a filtration volume of 3000 mL. Control group: standard CPB, with no MUF
Outcomes	Primary outcomes: plasma levels of LBP, TCC (C5b9), and cytokines (IL‐6, IL‐10, IL‐1β, TNF‐α) Secondary outcomes: clinical outcome (postoperative chest tube blood loss, length of ICU and in‐hospital‐stay, intubation time, pulmonary and wound infections, neurological complications, AF, inotropic requirements, blood loss in 24 hours, transfusions, and need for surgical revisions due to postoperative bleeding, haemodynamic monitoring, and lactate). Conventional laboratory analyses including leucocytes, CRP, Cr, and urea were also recorded. Blood samples were taken from a central venous line after the induction of anaesthesia, before CPB, before CPB weaning, 30 minutes after CPB, and at 6 hours, 24 hours, and 48 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Consecutive participants were assigned preoperatively by blockwise randomisation. There were no significant differences between the control and N‐MUF group regarding demographic characteristics.
Allocation concealment (selection bias)	Unclear risk	No details regarding allocation concealment method reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants or personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors declare that there are no conflicts of interest. This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors.

Paparella 2012.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: from July 2008 to December 2010 Number of study centres and locations: single‐centre. University hospital, University of Bari, Bari, Italy
Participants	Eighty‐six participants were randomly assigned either to the Physio group (n = 43) (closed venous reservoir, PC coating, and no cardiotomy suction) or to the Standard group (n = 43). Three participants assigned to the Physio group withdrew their consent and were excluded. Mean age: 67.48 years Sex (female/male ratio): 50.6% No high‐risk patients Inclusion criteria: CABG surgery Exclusion criteria: haemolytic or haemostatic disorders, anticoagulant treatment, anti‐Plt treatment within 5 days before surgery (not including aspirin), anti‐inflammatory treatment within 30 days before surgery, chronic inflammatory illness, class B to C Child‐Pugh liver disease, and re‐do surgery
Interventions	Intervention group: closed venous reservoir, PC coating, and no cardiotomy suction Control group: open, noncoated, and cardiotomy suction used
Outcomes	Primary outcome: IL‐6 (a marker of inflammation), PF‐1.2 (a marker of thrombin generation), PAP (a marker of fibrinolysis), and PF4 (a marker of Plt activation) Secondary outcome: NF‐kB, cTnI Blood samples were obtained from the CVC or the peripheral vein at 6 different time points: T0, before surgery; T1, 30 minutes after the beginning of CPB; T2, 15 minutes after aorta cross‐clamp release; T3, 2 hours after the end of CPB; T4, 24 hours after the end of surgery; and T5, 5 days after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was computer generated and took place the day before surgery. For baseline characteristics, no differences were found between the two groups.
Allocation concealment (selection bias)	Unclear risk	No details of allocation concealment method used
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not blinded. However, authors state that clinical decisions in operating room and ICU were taken following standard algorithms, and this should mitigate the risk of bias.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention regarding blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Amongst the 86 eligible participants, 43 were assigned to the Physio group and 43 to the Standard group. Three participants assigned to the Physio group withdrew their consent and were excluded.
Selective reporting (reporting bias)	Low risk	Overall adequate reporting of clinical outcomes. For extraction of NF‐kB, MNCs were obtained from the peripheral blood of 10 participants per group before surgery, at the end of surgery, and 24 hours after the end of surgery.
Other bias	Low risk	Supported by a Regione Puglia Government Research grant (Progetto Esplorativo 127)

Parolari 1999.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Paper accepted: 4 February 1998 Number of study centres and locations: not specified
Participants	Thirty participants were randomly allocated to be treated either with a circuit completely coated with surface‐bound heparin (n = 15) or with an uncoated circuit (n = 15). One participant from the heparin‐coated circuits group underwent surgical re‐exploration for excessive bleeding and was excluded from the study. Mean age: 57 years Sex (female/male ratio): 16.67% Low‐risk patients Inclusion criteria: patients undergoing primary elective CABG Exclusion criteria: age > 75 years; renal or liver disease; Hct <30%; coagulopathy; intake of drugs that affect Plt function, coagulation, or fibrinolysis within 7 days prior to surgery; and the presence of active inflammatory disease
Interventions	Intervention group: 'high‐dose' aprotinin (Trasylol, Bayer, Leverkusen, Germany) following the 'Hammersmith' protocol – the patients received a bolus (30 min) of 2 × 106 KIU dissolved in 200 mL saline after anaesthesia induction and after the introduction of the Swan‐Ganz catheter. Aprotinin was continuously infused at a dosage of 5 × 105 KIU in 50 mL saline/h until the end of operation. An additional 2 × 106 KIV was added to the priming solution of the oxygenator) plus heparin‐coated circuit (Duraflo II, Baxter). In addition, for this group of patients, the aortic cannula (Sherwood Argile THI Aortic Perfusion Cannula, Sherwood Medical, Tullamore, Ireland), the right atrium cannula (William Harvey Venous Cannulae, Bard Inc., Tewkesbury, MA), the antegrade cardioplegia and aortic vent catheter (MT 90301, Medical Technology s.r.l., Novedrate, CO, Italy), the retroplegia catheter (RC‐014, Research Medical, Inc., Midvale, UT), and connectors and taps (RVB2VG, Medical Technology) were precoated with heparin following the Duraflo II technique. Control group: 'high‐dose' aprotinin (as intervention group) plus an uncoated circuit
Outcomes	Clinical course (bleeding and transfusional requirements, the time spent on a ventilator or in duration of stay in the ICU, number of distal anastomoses, number of arterial distal anastomoses, heparin dose [IU], protamine dose [mg], intervention time [minutes], CPB time [minutes], haemostasis time [minutes], re‐exploration for bleeding, perioperative neurological deficit, perioperative MI, maximal body temperature reached during ICU stay [OC]) and haematological and inflammatory factors (blood cell count [Hct, blood Plt, total white cell, neutrophil, lymphocyte, and monocyte counts], IL‐6, and CRP) Data were collected from a peripheral vein and from radial artery catheter at the following times: the day before surgery; after sternotomy; at 30 minutes and 60 minutes after the beginning of CPB; at 10 minutes after completion of protamine; and at 4 hours, 24 hours, 48 hours, 3 days, and 6 days after the end of surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method is described. However, no significant differences existed between the two groups with regard to preoperative clinical features.
Allocation concealment (selection bias)	Unclear risk	Participants had identical circuits. All the participants were managed by the same cardiac surgical and anaesthesiology team; management of participants during the operation was similar, but no details regarding allocation concealment method are given.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The physicians involved in patient care and the biochemists, were blind to patient randomisation."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The physicians involved in patient care, and the biochemists, were blind to patient randomisation."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	One participant from the heparin‐coated circuits group underwent surgical re‐exploration for excessive bleeding and was excluded from the study. No ITT analysis was mentioned.
Selective reporting (reporting bias)	Unclear risk	One participant from the heparin‐coated circuits group underwent surgical re‐exploration for excessive bleeding and was excluded from the study.
Other bias	Low risk	No funding was disclosed.

Parolari 2007.

Study characteristics
Methods	Prospective, randomised study Run‐in period: from January 2004 to June 2005 Accepted for publication: 16 April 2007 Number of study centres and locations: not specified. Department of Cardiac Surgery, Centro Cardiologico Monzino I.R.C.C.S, and Department of Pharmacological Sciences, University of Milan, Italy
Participants	Thirty participants, candidates for coronary surgery, were randomised to undergo CABG (n = 16) or OPCAB (n = 14). One participant of the OPCAB group was excluded as he refused to participate in the study after surgery. Mean age: 66.52 years Sex (female/male ratio): 20.69% Low‐risk patients Inclusion criteria: elective primary surgical myocardial revascularisation (OPCAB or CABG) Exclusion criteria: age > 80 years, renal or liver disease, intake of drugs affecting Plt function or coagulation or fibrinolysis within 10 days prior to surgery, postoperative bleeding or re‐exploration for bleeding, perioperative MI, stroke, or renal failure requiring dialysis
Interventions	Intervention group: off‐pump operation on the beating heart (all OPCABs were performed through a midline sternotomy; mechanical stability of the coronary arteriotomy area was achieved with a suction stabiliser, and a soft plastic coronary flow‐shunt was always introduced into the coronary arteriotomy to maintain some degree of distal flow, to reduce myocardial ischaemia, and to improve visualisation of the anastomosis area) Control group: a conventional on‐pump operation (a non‐pulsatile roller pump, hollow‐fiber oxygenator with integrated heat exchanger, arterial filter, open cardiotomy reservoir, and polyvinyl tubing system were used)
Outcomes	Cytokines and neutrophil activation (plasma TNF‐α and IL‐6, neutrophil elastase levels), acute phase proteins (hs‐CRP and fibrinogen), and relationship between postoperative changes in cytokines and acute phase protein Blood samples were collected before the intervention, 5 minutes after protamine administration, and at 4 days, 8 days, and 30 days after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random sequence generation method is not described. However, the authors report "The randomisation codes were concealed in numbered, sealed, opaque envelopes", and no significant differences in clinical characteristics were found between the two groups.
Allocation concealment (selection bias)	Low risk	The randomisation codes were concealed in numbered, sealed, opaque envelopes. The treatment allocation for a participant was determined by opening the next envelope the evening before the operation.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel. However, because of the nature of the intervention which consists of different operating techniques, surely the cardiac surgical team and anaesthetic team were not blinded to the procedure. Unclear for the postoperative care.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Twenty‐nine participants of the 30 completed the study; one participant of the OPCAB group was excluded as he refused to participate in the study after surgery. There were no changes in group allocation after participant randomisation.
Selective reporting (reporting bias)	Unclear risk	Participants undergoing postoperative bleeding or re‐exploration for bleeding, perioperative MI, stroke, or renal failure requiring dialysis were excluded from analysis.
Other bias	Low risk	No funding was disclosed.

Paulitsch 2009.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Received: 17 October 2008 Revised: 28 November 2008 Accepted: 10 November 2008 Number of study centres and locations: not specified
Participants	A total of 84 participants were assessed in this study. Participants were assigned prospectively and randomly to either on‐pump (n = 41) or off‐pump (n = 51) coronary bypass surgery. Eight participants were excluded from the analysis. Mean age: 60.46 years Sex (female/male ratio): 35% Low‐risk patients Inclusion criteria: elective myocardial revascularisation with either method (CPB or OPCAB) (documented proximal multivessel coronary stenosis > 70%, stable angina, preserved ventricular function) Exclusion criteria: emergency surgery, concomitant major surgery, unstable angina requiring emergency revascularisation, valvular disease, LV aneurysm requiring repair, heart failure (EF < 40%), renal failure, or antecedent stroke
Interventions	Intervention group: OPCAB surgery with an Octopus stabiliser (Medtronic Inc., Minneapolis, Minnesota, USA) placed on the beating heart on both sides of the target coronary artery Control group: on‐pump surgery with CPB in combination with cold crystalloid cardioplegia for myocardial protection
Outcomes	Biochemical measurements (CRP, fibrinogen, D‐dimer, and plasminogen activator inhibitor type‐1), perioperative characteristics (duration of CPB, duration of surgery, number of grafts, number of territories grafted, ICU LOS and hospital LOS, blood loss), clinical events (chest infection, MI, CK‐MB, acute renal failure, stroke, systolic heart failure, and death) Blood samples were collected before and after (1 hour and 24 hours) surgery; clinical data were also assessed after 1 year.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details regarding randomisation method used. However, baseline characteristics of the two groups were similar.
Allocation concealment (selection bias)	Unclear risk	No mention regarding the allocation concealment method used. However, similar surgical (full median sternotomy) and anaesthetic procedures were used for both groups.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention regarding blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention regarding blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Eight participants were excluded from analysis because of an insufficient sample of blood (blood haemolysed or clotted). Analyses were performed on blood from 92 participants and included samples from 41 participants assigned to on‐pump and 51 participants assigned to off‐pump surgery.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This study is supported by Zerbini Foundation. The authors have no conflicts of interest to declare.

Poulsen 2009.

Study characteristics
Methods	Prospective, double‐blind, placebo‐controlled and randomised study Run‐in period: not specified Study date: 2009 Number of study centres and location: single tertiary care at the Department of Cardiac Surgery, Copenhagen University Hospital (Rigshospitalet)
Participants	A total of 45 participants were assessed in this study. Patients were randomised to receive EPO (n = 22) or placebo (n = 21). Forty‐five participants were included. One participant in the EPO group and one participant in the placebo group were excluded after the administration of the first dose of test medicine because of postponing of the operation. Mean age: 64.05 years Sex (female/male ratio): 25.58% Low‐risk patients Inclusion criteria: patients scheduled for CABG surgery with CPB Exclusion criteria: pregnancy; concomitant malignant disease; history of polycythemia vera or Hb concentration > 13 mmol/L; untreated or uncontrolled arterial hypertension; previous treatment with recombinant EPO; major trauma or major surgery < 3 months before inclusion; treatment with steroids < 3 months before inclusion; or previous thromboembolic complications from surgery.
Interventions	Intervention group: recombinant human EPO (each dose intravenously as bolus injection of 500 IU/kg, EPREX, epoetin alpha; Janssen‐Cilag, Birkeroed, Denmark) was diluted in isotonic saline to a total volume of 20 mL. Control group: placebo = 20 mL of isotonic saline Timing: 2 preoperative doses (the first dose was given 12 to 18 hours preoperatively, and the second dose was given the next day in the operating room just after the induction of anaesthesia and before surgery)
Outcomes	TNF‐α, IL‐1, IL‐1ra, IL‐6, IL‐10, and NT‐proBNP; CRP and duration of surgery; administration of vasoactive surgery drugs or crystalloids; Macrodex; blood products; use of aprotinin; aortic clamp/CPB time; perfusion pressure; FiO2; postoperative extubation; postoperative bleeding; body temperature; glucose, LDH, and lactate levels; p‐Cr; ICU LOS; AF; and in‐hospital death Secondary outcomes: adverse effects of treatment (allergic reactions and epileptic seizures) and arterial blood pressure and plasma concentrations of Hb, reticulocytes, and thrombocytes. Samples were collected at nine time points in the perioperative period: the day before CABG, immediately before the administration of the first dose of EPO or placebo; just after the induction of anaesthesia and before administration of the second dose of EPO or placebo; at the end of CPB; 2 hours after arrival in the postoperative care unit; the next morning; the second day after CABG surgery; the third day after CABG surgery; the fourth day after CABG surgery; and the fifth day after CABG surgery. All samples from the first POD onward were obtained between 7:00 and 11:00 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	An independent investigator generated a restricted randomisation list with blocks of 10 using a computer program.
Allocation concealment (selection bias)	Low risk	After patient inclusion, the independent investigator opened a sealed opaque envelope containing the treatment code and prepared the medication. He did not participate in patient inclusion, patient management, study measurements, or data analysis. The two types of medication were indistinguishable.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and investigators were blinded (except who made allocation).
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Study measurements of data analysis was made by blind investigators.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Forty‐five participants were included. One participant in the EPO group and one participant in the placebo group were excluded after the administration of the first dose of test medicine because of postponing of the operation. One participant in the EPO group was excluded at day 3 because of problems with venous access.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	Supported by grants from the Heart Centre and from the scientific board of Copenhagen University Hospital (Rigshospitalet)

Prasongsukan 2005.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Period: 1 August 2000 to 28 February 2001 Number of study centres and locations: single centre; St Paul’s Hospital, University of British Columbia, Vancouver, British Columbia, Canada
Participants	A total of 88 participants were assessed in this study. Inclusion criteria: elective first‐time CABG and normal sinus rhythm. Exclusion criteria included a history of heart block; a permanent pacemaker; any documented or suspected supraventricular or ventricular arrhythmias, including isolated atrial or ventricular premature depolarisation noted on preoperative surface electrocardiography; requirement for additional procedures, such as valvular operation or LV aneurysmectomy; refusal to participate in this study; use of a radial artery for grafting; steroid dependency; steroid allergy; and participation in another investigational protocol.
Interventions	Participants were randomly assigned to a placebo group receiving maintenance fluids (5% dextrose water with potassium chloride 20 mEq/L) or to a steroid group receiving 1 g of IV MPSS (Solu‐Medrol; Upjohn, Kalamazoo, MI) before CPB and 4 mg of IV dexamethasone (Decadron; Merck Sharp & Dohme, West Point, PA) every 6 hours for a total of 4 doses in the first 24 hours after surgery.
Outcomes	Primary: AF Secondary: secondary endpoints were the length of hospital stay and the adverse effects (complications) of steroids. Supplementary study: TNF‐α, IL‐6, IL‐8, and IL‐10 at 4 hours and 24 hours after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details about the method of random sequence generation used. Authors state "All vials … were prepared and randomized by the hospital pharmacy."
Allocation concealment (selection bias)	Low risk	All vials of the steroid and placebo medications were prepared and randomised by the hospital pharmacy. The steroid and placebo solutions were visually indistinguishable.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All vials of the steroid and placebo medications were prepared and randomised by the hospital pharmacy. The steroid and placebo solutions were visually indistinguishable.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Authors state: "double‐blind protocol in which the surgical staff, principal investigators, and patients were blinded to the assigned therapy. Clinical data were collected and recorded in the database by independent blinded investigators".
Incomplete outcome data (attrition bias) All outcomes	High risk	It is unclear which patients were selected for the supplementary study on cytokines. Authors report "22 patients in the steroid group and in 86 similarly treated control patients".
Selective reporting (reporting bias)	High risk	Only IL‐6 reported
Other bias	Unclear risk	No funding information given

Prieto 2013.

Study characteristics
Methods	Prospective, randomised study Run‐in period: between March and June 2009 Study date: 2013 Number of study centres and location: single centre.Hospital Universitario La Princesa, Madrid, Spain
Participants	Fifty‐seven patients were prospectively randomised to undergo cardiac surgery without (n = 28) or with a cell saver (n = 29). Mean age: 64.7 years Sex (female/male ratio): 31.6% Low‐risk patients Inclusion criteria: first‐time, nonemergency cardiac surgery with the use of CPB Exclusion criteria: age over 80 years, redo or emergency surgery, aortic or pericardial disease, endocarditis, need for triple‐valve surgery, refusal of blood products, logistic EuroSCORE > 10%, and a high risk of bleeding (Cr > 2.2 mg/mL, liver insufficiency (Child‐Pugh B or C), severe lung disease, body surface area < 1.6 m2, preoperative Hb < 13 g/dL in males or < 12 g/dL in females, Plt count < 50,000/mL or any Plt disorder, coagulation disorder, intake of aspirin 3 days before surgery or clopidogrel 7 days before surgery)
Interventions	Intervention group: all remaining blood inside the circuit was recovered and concentrated by the cell saver and transfused to the participants using a 200‐m filter (Venisystem; Hospira, USA). Cardiotomy suction was applied when the participant was heparinised. This blood was re‐infused into the participant continuously during ECC. Control group: all blood in the surgical field was aspirated using only cardiotomy suction. All blood aspirated before starting heparin administration and after protamine administration was lost.
Outcomes	Complete blood count, proinflammatory cytokines (IL‐8, p40 subunit of IL‐12, IL‐6, IL‐1β, INF‐γ, IL‐23), and clinical outcomes Blood samples for inflammatory marker assays were collected from the arterial line on induction of anaesthesia, at the end of CPB, 1 hour after surgery, and 24 hours after surgery. Clinical data were also recorded on the day of discharge. All participants were interviewed in person or by telephone 30 days after the procedure.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised using a statistical program the day before surgery. There were no significant differences between the two groups for baseline characteristics.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcomes assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors. No conflict of interest

Rasmussen 2007.

Study characteristics
Methods	Prospective, randomised study Run‐in period: not specified Accepted for publication: 19 May 2007 Number of study centres and locations: single centre. Aalborg Hospital and Aarhus University Hospital, Denmark
Participants	Thirty‐five participants were randomly assigned to one of two groups (CABG, n = 17, or OPCAB, n = 18). Two participants were excluded: one participant was converted from OPCAB to CABG and one for re‐exploration. Mean age: 67.5 years Sex (female/male ratio): 18.18% No high‐risk population Inclusion criteria: low‐risk patients scheduled for elective coronary revascularisation Exclusion criteria: LVEF < 0.40, acute coronary syndrome, previous cardiac surgery, concomitant valve disease, AF or atrial flutter, pulmonary dysfunction and chronic dialysis, re‐exploration, conversion to CABG
Interventions	Intervention group: off‐pump CABG Control group: ONCAB surgery
Outcomes	Oxygenation (shunt [per cent] and ventilation‐perfusions mismatch, described as DPO2 [kPa]) and systemic inflammatory response (IL‐6, IL‐8, IL‐10, CRP, and neutrophils), haemodynamic profiles (cardiac index, SvO2, PCWP, lactate), pulmonary cytokine production, and elimination and myocardial markers (CK‐MB, TnT). Data were recorded for up to 5 days postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly allocated to OPCAB or CABG based on computer‐generated codes sealed in sequentially numbered, opaque envelopes. The randomisation resulted in 2 demographically comparable groups; perioperative and postoperative data were equal.
Allocation concealment (selection bias)	Low risk	Patients were randomly allocated to OPCAB or CABG based on computer‐generated codes sealed in sequentially numbered, opaque envelopes.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unclear whether patients were aware of the group they were assigned. Unclear whether the postoperative care team was aware of the treatment. Because of the nature of the intervention, surgical and anaesthetic team were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Two participants were excluded: one participant was converted from OPCAB to CABG and one for re‐exploration.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The Research Foundation of the Northern County of Jutland, Denmark and The Research Foundation of Hertha B. Christensen, Aalborg, Denmark provided support.

Rastan 2005.

Study characteristics
Methods	Randomised prospective study Run‐in period: not specified Received: 29 November 2004 Received in revised form: 27 February 2005 Accepted: 9 March 2005 Available online: 8 April 2005 Number of study centres and locations: single centre. Department of Cardiac Surgery, University of Leipzig, Heart Center Leipzig, Struempellstr. 39, 04289 Leipzig, Germany
Participants	Forty participants were assigned to OPCAB or OnP‐BH surgery. One OPCAB participant crossed over to OnP‐BH. Mean age: 64.15 years Sex (female/male ratio): 20% No high‐risk population Inclusion criteria: elective, patients with normal EF and three‐vessel CAD (stenosis > 70%) Exclusion criteria: patients > 75 years of age, combined cardiac procedures, cardiac reoperation, emergency surgery, left main stenosis > 70%, and EF < 50%
Interventions	Intervention group: OPCAB surgery Control group: on‐pump beating heart surgery
Outcomes	Intraoperative myocardial ischaemia (pH, lactate, pO2) and oxidative stress (MDA) (CS blood) and myocardial necrosis (CK‐MB, cTnI), myocardial dysfunction (NT‐proBNP), and inflammation (CRP) (arterial blood) and clinical data Arterial blood was analysed 4 hours, 12 hours, and 24 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear method of randomisation used. Authors state "Randomisation was performed by chance the day prior to surgery using randomisation envelops [sic]". Preoperative patient characteristics were similar in both groups.
Allocation concealment (selection bias)	Unclear risk	No clear method of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and postoperative care. Because of the nature of the intervention, surgical team could not be blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	One OPCAB patient crossed over to OnP‐BH, and no further or unexpected events or conversions occurred. Two patients with evidence of perioperative MI required surgical re‐intervention and were excluded from the laboratory data analyses.
Selective reporting (reporting bias)	Unclear risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Remadi 2006.

Study characteristics
Methods	Prospective randomised study Run‐in period: from 1998 to December 2003 Submitted: 15 January 2004 Accepted: 26 March 2005 Number of study centres and locations: single centre. South Hospital, Amiens Cedex 1, France
Participants	Four hundred patients underwent elective CABG using a standard CPB (200 patients) or a Jostra MECC System (200 patients) Mean age: 64.7 years Sex (female/male ratio): 12% No high‐risk population Inclusion criteria: elective CABG Exclusion criteria: renal dysfunction (Cr > 1.2 mg/dL) and re‐do procedures
Interventions	Intervention group: MiECC, a fully heparinised closed‐loop CPB system (Bioline‐Jostra, Gretz, France) Control group: a standard CPB
Outcomes	Mortality rate, low‐cardiac‐output syndrome, inflammatory response, CRP, Hct and Hb rate, intraoperative transfusion rate, postoperative blood Cr and urea, haemodynamic data
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of random sequence generation. However, there was no statistically significant difference in preoperative comorbidity between groups.
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment method used
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel. However, because of the nature of the intervention, operating room and surgical team were aware of the circuit used for CPB.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention regarding blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up reported, but no intention‐to‐treat analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Measures of systemic inflammatory response were not performed because the number of participants was too high.
Other bias	Low risk	The authors had no financial or conflict of interest relative to this work.

Rimpilainen 2011.

Study characteristics
Methods	Prospective randomised study Run‐in period: between February 2007 and May 2009 Study date: 2011 Number of study centres and locations: single centre. Oulu University Hospital, Oulu, Finland
Participants	40 participants were randomised to mCPB (n = 20) or CCPB (n = 20). Mean age: not specified Sex (female/male ratio): not specified No high‐risk population Inclusion criteria: AVR with or without CABG Exclusion criteria: antithrombotic (excluding aspirin) and immunosuppressive medication, unstable CAD, significant LV dysfunction (EF < 35%), and severe renal or liver failure
Interventions	Intervention group: mCPB consisted of biocompatible material (PMEA)‐coated lines, a membrane oxygenator, a manual venous line bubble trap, an arterial line filter, a collapsable reservoir bag (ROCsafeTM, Terumo Europe NV, Leuven, Belgium), and a centrifugal pump. Control group: CCPB consisted of biocompatible material (PC)‐coated lines; a membrane oxygenator; an arterial line filter; a hard‐shell, dual‐chamber, venous reservoir (DidecoTM S.p.A., Mirandola, Italy); and a roller pump.
Outcomes	Markers of inflammation (leucocyte PMN‐Elastase, MPO, TNF‐α, IL‐6, and IL‐8), endothelial activation (vWF release, TM), thrombin‐dependent Plt activity and sGPV, coagulation, Plt activation, and Hct Blood samples were collected at eight time points: before the induction of anaesthesia; 45 minutes after the start of CPB; after protamine infusion; at 2 hours after the termination of CPB; at 4 hours after the termination of CPB; on the morning of the first POD; on the morning of the second POD; and on the morning of the third POD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Forty patients were randomised the day before surgery by using sealed envelopes, numbered according to a computer‐generated list, using a 1:1 randomisation to conventional CPB or minimized CPB." Participant demographics did not differ significantly between the two groups.
Allocation concealment (selection bias)	Low risk	"Forty patients were randomised the day before surgery by using sealed envelopes, numbered according to a computer‐generated list, using a 1:1 randomisation to conventional CPB or minimized CPB."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel, only mention of ophthalmologist blinding assessing retinal emboli, which is not an outcome of interest for our study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Inflammatory marker data were not reported in the paper. Authors state "Other measured markers of inflammation (MPO, IL‐6, IL‐8, TNF‐α) were without significant intergroup differences".
Other bias	Low risk	This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors. No conflict of interest was reported.

Rubens 1999.

Study characteristics
Methods	Prospective, double‐blind trial Run‐in period: during a 12‐month interval Received: 29 September 1998 Received in revised form: 21 December 1998 Accepted: 8 January 1999 Number of study centres and locations: single centre. University of Ottawa Heart Institute, Ottawa, Ontario, Canada
Participants	Thirty‐four patients randomised to surgery, with either a standard circuit or a circuit treated ('tip to tip') with the surface‐modifying additive Mean age: 57.95 years Sex (female/male ratio): 5.88% No high‐risk population Inclusion criteria: patients scheduled for elective primary CAB operations necessitating CPB Exclusion criteria: major systemic illnesses (e.g. insulin‐dependant diabetes with organ failure, renal failure), preoperative Plt or coagulation abnormalities, Hct during CPB < 21%, or anticoagulant or anti‐Plt medications
Interventions	Intervention group: surgery with the test circuit (COBE SMARxTTM Biocompatible Extracorporeal Circuit, COBE Cardiovascular, Arvada, CO) Control group: a standard untreated circuit
Outcomes	Perioperative haemodynamics (MAP, WBC count), volume replacement, α‐agonist, inotrope use, and TCC (SC5b‐9) The sampling was carried out at the following time points: insertion of the central line; 5 minutes after heparin administration; 10 minutes, 20 minutes, 40 minutes, and 60 minutes on CPB; 20 minutes after administration of protamine; and 24 hours after the administration of protamine.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation was accomplished using a random number table. There were no differences in demographic or operative variables between the two cohorts, although there was a trend towards an increased number of class III ventricles in the control group.
Allocation concealment (selection bias)	Low risk	"The perfusionist assigned the treatment immediately preoperatively, by opening a sealed, numbered envelope prior to set‐up of the extracorporeal circuit."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The perfusionist assigned the treatment immediately preoperatively, by opening a sealed, numbered envelope prior to set‐up of the extracorporeal circuit. The cannulae and tubing used in these two circuits were identical in appearance, so that all members of the surgical and anaesthesia teams, excluding the perfusionist, were blinded to the patient designation."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	There were two exclusions from the study (both in the control group): in the first case, there were technical problems with a gastroepiploic graft, and in the second case, the venting through the standard aortic root angiocatheter was inadequate, and the aortic root was opened directly for aspiration. In neither case were the complications related to the type of bypass circuit utilised. Amongst the remaining 34 participants, there was complete follow‐up throughout the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This study was supported from a grant from COBE Cardiovascular, Arvada, CO, USA.

Rubens 2005.

Study characteristics
Methods	Prospective clinical trial Run‐in period: not specified Accepted for publication: 19 July 2004 Number of study centres and locations: single centre. University of Ottawa, Ottawa, Ontario, Canada
Participants	In total, there were 68 participants randomised into 3 groups, A (n = 22), B (n = 23), and C (n = 23). Participants were randomly assigned to CPB utilising control circuits or a circuit prepared with an SMA (SMA‐CPB), with or without MP (1 g IV) (4 groups, n = 17 per group). Mean age: 54 years Sex (female/male ratio): 6% No high‐risk population Inclusion criteria: patients undergoing CABG Exclusion criteria: steroids or Coumadin, emergency, reoperative surgery or other cardiac procedures in addition to CABG, coagulopathy, bleeding diathesis, thrombocytopenia, severe COPD (FEV1 < 1.5 L), recent peptic ulcer disease (< 6 months), chronic renal failure (Cr > 120 µmol/L), or steroid dependency
Interventions	Intervention group: surgery with the modified CPB circuit (SMA‐CPB) and MP (1 g IV) given before CPB Control group: surgery with standard circuit and placebo
Outcomes	Primary outcomes: ACE levels, tissue plasminogen activation and BK generation, complement activation (TCC), cytokine release (IL‐6, IL‐8), elastase Secondary outcomes: clinical outcomes (blood loss, transfusion, arterial pressure response and hypotensive episodes, postoperative cardiac and pulmonary functions (PVR, pulmonary compliance, AaDO2, DO2I, avDO2, OER, and VO2; PaO2/FiO2; CK and troponin T; atrial fibrillation; cardiac index; and SVR), insulin need and inotropic use, occurrence of infections (positive blood cultures, clinical sepsis, wound infection, pneumonia, UTI) Sampling was done at the following time points: at insertion of the central line; before heparin; after heparin administration but before initiation of CPB (post‐sternotomy); at 5 minutes, 10 minutes, 15 minutes, 20 minutes, 40 minutes, and 60 minutes after initiation of CPB; and 1 hour after separation from CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was in blocks of 4, generated using SAS Version 8.2 (SAS, Cary, NC).
Allocation concealment (selection bias)	Low risk	The perfusionist performed the treatment assignment immediately preoperatively by opening a sealed, numbered envelope. Syringes containing the MP or the placebo (saline) were prepared in the hospital pharmacy and labelled with a code. The cannulas and tubing used in all of the CPB circuits were identical in appearance, so that all members of the surgical and anaesthesia teams, except the perfusionist, were blinded to the circuit assignment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All members of the surgical and anaesthetic teams were blinded (except the perfusionist).
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Three participants were excluded after randomisation but before intervention because of surgeon request, and no sampling was done. All other participants (n = 17 per group) completed the study with full follow‐up.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	This trial was funded by operating grant UOP 44198 from the Canadian Institute of Health Research University‐Industry Program.

Rubens 2007.

Study characteristics
Methods	Randomised, double‐blind study Run‐in period: between September 2001 and February 2006 Study date: 2007 Number of study centres and locations: single centre. Divisions of Cardiac Surgery (F.D.R., M.B., T.M.) and Cardiac Anesthesia (D.W., H.J.N.) and the Department of Epidemiology and Community Medicine (G.W.), University of Ottawa Heart Institute, Ottawa, Ontario, Canada
Participants	A total of 266 participants were assessed in this study. Participants were randomised to receive unprocessed blood (control, n = 134) or cardiotomy blood that had been processed by centrifugal washing and lipid filtration (treatment, n = 132). Preoperative neuropsychological testing was completed in 247 (93%) participants (control (n = 125), treatment (n = 122)). Mean age: 58.9 years Sex (female/male ratio): 10.53% No high‐risk population Inclusion criteria: patients undergoing isolated, non‐emergent coronary or aortic valve surgery using CPB Exclusion criteria: co‐procedure, OPCAB/MiDCAB, central neurological deficits, renal and hepatic impairment, comorbidity, unstable, language, distance, other study, no time
Interventions	Intervention group: cardiotomy blood had been processed by centrifugal washing and lipid filtration (Pall LeukoGuard RS Filter; Pall Biomedical Products Company, East Hills, NY). Control group: unprocessed blood
Outcomes	Primary outcomes: need for one or more RBC transfusions and incidence of postoperative cognitive dysfunction Secondary outcomes: blood loss, total non‐RBC transfusion requirements (factors, Plts), quality of life, and the incidence of cognitive deficit at 3 months and coagulation parameters (coagulation profiles, including Hb, Plts, partial thromboplastin time, thrombin time, INR, and fibrinogen) Data were collected during anaesthetic induction, 15 minutes after protamine administration, and 2 hours, 4 hours, and 12 hours after separation from CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation, supervised by the data manager, was computer generated in blocks of 8 (SAS version 8.2; SAS Institute, Cary, NC) stratified for age (75 years), and the assignment was concealed until the interventions were assigned. Baseline characteristics were similar between treatment and control groups.
Allocation concealment (selection bias)	Low risk	A sealed opaque envelope containing the treatment allocation was opened by the research coordinator just before the participant receiving heparin and the initiation of CPB. The perfusionist was informed of the treatment assignment. Intraoperative blinding of all members of the surgical team (except for the perfusionist) was accomplished by the positioning of an opaque drape over the CPB circuit and the cell saver. All intraoperative decisions to transfuse during CPB were made by the anaesthetist, who was unaware of treatment assignment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: participants and treating physicians were blinded to treatment assignment. The perfusionist was informed of the treatment assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The participants, psychometrists, and all medical staff were blinded to treatment assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data were collected from all participants on an ITTt basis, and no participants were lost to follow‐up for the bleeding and transfusion data. Preoperative neuropsychological testing was completed in 247 (93%) participants (control (n = 125), treatment (n = 122)).
Selective reporting (reporting bias)	Low risk	Participants undergoing AVR were excluded from the neuropsychological testing.
Other bias	High risk	This trial was funded by Clinical Trials Grants MCT‐44149 and MCT‐70887 from the Canadian Institute of Health Research. The LeukoGuard RS filters were provided by Pall Biomedical (East Hills, NY).

Rubino 2012.

Study characteristics
Methods	Prospective, randomised study Run‐in period: between December 2009 and June 2010 Accepted: 21 February 2012 Number of study centres and locations: not specified
Participants	Seventy participants were randomised to receive leucocyte depletion filters on both arterial and cardioplegia lines (filter group, n = 35) or standard arterial filters (control group, n = 35) during CPB. Mean age: 63.3 years Sex (female/male ratio): 25.7% No high‐risk patients Inclusion criteria: isolated primary elective CABG surgery Exclusion criteria: not specified
Interventions	Intervention group: both arterial and cardioplegia lines (filters). A Pall LeukoGuard® LG6B arterial filter (Pall Biomedical, Portsmouth, UK) was connected to the arterial line, and a second Pall LeukoGuard® BC2B cardioplegia filter (Pall Biomedical, Portsmouth, UK) was connected to the cardioplegia line. Control group: standard arterial filter, Pall AL8 filter (Pall Biomedical, Portsmouth, UK), was adopted, and no filters were connected to the cardioplegia line. Timing: systemic leucofiltration started at the onset of CPB and lasted throughout the entire procedure.
Outcomes	Primary outcomes: PaO2/FiO2 for pulmonary function Secondary outcomes: intubation time, need for post‐extubation NIV with CPAP hood, respiratory failure and postoperative pneumonia for pulmonary function and inflammatory response (IL‐6, IL‐8, IL‐1β, IL‐10, IL‐2, TNF‐α), hospital outcomes (mortality, intensive care, and in‐hospital lengths of stay), and haematological results (neutrophil and Plt counts, need for postoperative transfusion). Samples were collected from peripheral arterial line before induction of anaesthesia, at ICU admission, and at 12 hours, 24 hours, and 48 hours postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Consecutive participants were randomised by lottery. The two groups proved homogeneous in preoperative baseline characteristics.
Allocation concealment (selection bias)	Low risk	Participants were allocated by preprepared sealed envelopes containing the group assignment. All patients underwent surgery at 8:00 a.m., and CPB and surgical techniques were standardised and did not change during the study period. In all patients, CABG was performed through a median sternotomy.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Anaesthesiologists and cardiac surgeons caring for the patients during the postoperative course were blinded to the intraoperative group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Inflammatory data could not be extracted from figure provided in the paper.
Other bias	Low risk	No conflict of interest

Rumalla 2001.

Study characteristics
Methods	Randomised controlled study Run‐in period: 6 months Study date: 2001 Setting: single centre. Robert Wood Johnson Medical School (New Brunswick, New Jersey) and affiliated tertiary care hospital
Participants	Thirteen patients were randomly assigned to an intervention group (n = 6) or a control group (n = 7). Mean age: 61.92 years Sex (female/male ratio): not specified No high‐risk patients Inclusion criteria: CABG Exclusion criteria: immunodeficency, short‐ or long‐term steroid therapy
Interventions	Intervention group: 1 g of MP on anaesthetic induction Control group: not specified
Outcomes	PMN cell‐associated receptors (CD95, TNF receptor), cytokines (TNF‐α, IL‐6, IL‐8, and IL‐10), soluble receptors (sTNFR1 and sTNFR2), and clinical outcomes Blood samples were drawn before induction, 20 minutes after sternotomy, and 20 minutes after bypass, postoperatively in the recovery room, and on POD 1.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation was performed by the investigators, but no randomisation was method described.
Allocation concealment (selection bias)	Unclear risk	No mention of the allocation concealment method used
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	IL‐8 data not shown
Other bias	Low risk	This study was supported by grant GM‐34695 from the US Public Health Service, Bethesda, MD.

Ryugo 2006.

Study characteristics
Methods	Clinical, double‐blind, randomised study Run‐in period: not specified Received: 2 May 2005 Accepted: 13 September 2005 Number of study centres and locations: single centre Osaka University Hospital, Japan
Participants	Fourteen patients were divided into a SiV group (n = 7) and a control group (n = 7). Mean age: 61.7 years Sex (female/male ratio): 28.57% No high‐risk patients Inclusion criteria: cardiovascular surgery with > 2 hours CPB Exclusion criteria: not specified
Interventions	Intervention group (SiV): PMN‐Elastase inhibitor SiV (novel synthesised drug, sodium N‐[2‐[4 (2,2dimethylpropionyloxy) phenylsulfonylamino] benzoyl] aminoacetate tetrahydrate, Ono Pharma, Osaka, Japan). Participants received a continuous IV infusion of 0.2 mg/kg/h prepared as 300 mg of SiV dissolved in 500 mL of saline, from their admission to the ICU until the morning of POD 4. Control group: saline. The control group received an infusion of the same volume for 4 days from the time of arrival in the ICU.
Outcomes	Inflammatory reactions and respiratory data (WBC count, serum PMN‐Elastase and IL‐8, respiratory index, PaO2/FiO2, CD18, intubation time) and clinical complications Arterial blood was taken from the radial arterial catheter just before, then 6 hours and 24 hours after the start of the SiV or saline infusion.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No randomisation method is described. However, the two groups had the same baseline characteristics.
Allocation concealment (selection bias)	Unclear risk	No method of allocation concealment is described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding was disclosed.

Sahlman 2001.

Study characteristics
Methods	Open, randomised controlled trial Run‐in period: not specified. Published 2001 Number of study centres and locations: single centre in a large tertiary centre in Finland
Participants	A total of 60 participants were included in this study. Patients undergoing elective CABG Exclusion criteria were LVEF < 30%; significant valvular dysfunction; any renal, liver, or pancreatic disease; transient ischaemic attack or cerebral infarction during the 6‐month period prior to the operation; or any cerebral disease limiting independent activity.
Interventions	Intervention: leucocyte‐depleting arterial filter Control: conventional management
Outcomes	Authors did not specify primary and secondary outcomes. Outcomes recorded were Hb, Hct, and neutrophil count, preoperative and postoperatively; values for BUN, serum Cr, serum sodium and potassium, serum osmolality, urine creatine, urine sodium and potassium, and urine osmolality at baseline, at 6 hours and 24 hours after CPB, and on the fifth POD; C3 and C4 before surgery and 24 hours after CPB; the MBfraction of the creatine kinase (CK‐MB) on the morning of the first POD; and amount of crystalloid administered, urine output, chest drain output, day of discharge from ITU, and hospital length of stay.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors do not state how the participants were randomised.
Allocation concealment (selection bias)	High risk	No mention of allocation concealment method used; however, open‐label trial
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label trial
Blinding of outcome assessment (detection bias) All outcomes	High risk	No mention of blinding of outcome assessment and open study
Incomplete outcome data (attrition bias) All outcomes	Low risk	It would appear that all participant data were collected and analysed. The only incomplete data collection was related to the length of hospital stay for the participants who were transferred to other institutions for continuation of postoperative care. Overall, however, the risk of incomplete outcome data bias is low.
Selective reporting (reporting bias)	Unclear risk	No protocol available, and the outcomes are not prespecified in the methodology.
Other bias	High risk	Study was supported by the manufacturer of the filter, but authors do not detail how this support was given and whether it was of financial nature. The authors do not state whether there was any conflict of interest.

Sano 2006.

Study characteristics
Methods	Randomised controlled trial, single centre Run‐in period: not specified. Published 2006 Number of study centres and locations: single centre. Kyushu University, 3‐1‐1 Maidashi Higashi‐ku Fukuoka 812‐8582, Japan
Participants	A total of 60 participants were assessed in this study. Patients undergoing elective cardiac surgery
Interventions	Intervention group: 31 participants received hydrocortisone (50 mg/kg) before and after CPB (steroid group). Control group: the other 29 participants received saline as controls (nonsteroid group).
Outcomes	Risk of infection after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors stated that participants were prospectively randomised into two groups by the operation‐registry staff, who were not involved in the study. However, no method of randomisation is described.
Allocation concealment (selection bias)	Unclear risk	Authors do not state whether or how the allocation was concealed.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data were analysed; no participants were missing or excluded from the analysis.
Selective reporting (reporting bias)	Unclear risk	No protocol available, and all the outcomes were not prespecified in the methodology.
Other bias	Unclear risk	No statement about funding or conflict of interest

Sato 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised double‐blind trial Run‐in period: unspecified Registration date: unspecified, study published in 2000 Number of study centres and locations: single centre; authors from Gunma University School of Medicine, Maebashi, Japan
Participants	Fifteen adult patients undergoing elective cardiac surgery for ischaemic heart or valvular heart disease Mean age: 65 years Sex (female/male ratio): 0.36 Low‐risk patients (elective CABG or valvular) Exclusion criteria: history of disease associated with cytokine elevation
Interventions	Group U (n = 7): UST (Mochida Pharmaceutical Corp., Ltd., Tokyo) was administered during and after surgery. 600,000 units of UST were introduced into a priming solution of the CPB circuit, and 300,000 units of UST were introduced into the CPB circuit just before the removal of aortic cross‐clamping. An additional 150,000 units of UST were administered intravenously twice a day for 5 days after surgery. Group C (n = 8) served as controls, receiving no PI, steroids, or other agents.
Outcomes	Alpha 1 antitripsin, IL‐6, IL‐8, PMN‐Elastase Arterial blood samples were obtained on an event‐based schedule: shortly after anaesthesia was induced; immediately after CPB; on POD 1; on POD 2; and on POD 5.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of randomisation process
Allocation concealment (selection bias)	Unclear risk	No allocation concealment description
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description. Authors state "Double blinded study".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected data were reported properly.
Other bias	Low risk	No funding was disclosed.

Schmartz 2003.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial, single centre Received for publication: 16 April 2001 Number of study centres and locations: Departments of Anesthesiology, Intensive Care Medicine, and Anesthesiology, Erasme University Hospital, and the Department of Immunology, Brugmann University Hospital, Brussels, Belgium
Participants	A total of 60 participants were assessed in this study. Inclusion criteria: adults, male, undergoing elective CABG Exclusion criteria: severe LV dysfunction, autoimmune disease, presumed or documented infection, preoperative administration of corticosteroids or nonsteroidal inflammatory medication during the 3 days before the operation, ingestion of drugs with a known anti‐Plt effect during the 3 days before the operation, emergency procedure, previous cardiac operation, preoperative Hct level of less than 25%, pre‐existing liver dysfunction, pre‐existing renal dysfunction, genital or acquired bleeding disorders, known hypersensitivity to aprotinin, previous exposure to aprotinin, allergic diathesis or atopy, or alcohol or drug abuse
Interventions	The 60 participants were divided into 3 groups by means of computerised randomisation: a control group that received a placebo solution; a low‐dose group that received a priming solution of 1,000,000 KIU of aprotinin (Trasylol; Bayer, Leverkusen, Germany), followed by a continuous infusion of 250,000 KIU/h and 1,000,000 KIU added to the pump prime; and a high‐dose group that received a priming solution of 2,000,000 KIU of aprotinin (Trasylol, Bayer), followed by a continuous infusion of 500,000 KIU/h and 2,000,000 KIU added to the pump prime. For this analysis, we considered groups 1 and 3 as per protocol.
Outcomes	Inflammatory response to CPB: IL‐6, IL‐8, IL‐10, TNF‐α, C3a, C5a, C1q, C3, C4, prekallikrein, PGT2
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors stated that patient randomisation was carried out using a computer.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	In the introduction, the study is described as a randomised controlled trial with double blinding, but no further details are given.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	In the introduction, the study is described as a randomised control trial with double blinding, but no further details are given.
Incomplete outcome data (attrition bias) All outcomes	Low risk	One participant had to be withdrawn from the study, and one further participant was recruited instead. All participants were accounted for in the final analysis.
Selective reporting (reporting bias)	Low risk	Although there is not a protocol available, the authors clearly define the tests that they are running in the methodology section, and these are all reported in the results.
Other bias	High risk	This work was supported by a grant from Bayer, Leverkusen, Germany. Aprotinin used in the trial is Trasylol produced by Bayer, Leverkusen, Germany.

Schoenebeck 2010 (A).

Study characteristics
Methods	Prospective, randomised controlled trial Enrolment period: March 2004 and December 2006 Number of study centres and locations: single centre. Department of Cardiovascular Surgery, University Heart Center Hamburg, Hamburg, Germany
Participants	A total of 120 patients underwent elective CABG and were randomised into three groups: mini‐CPB using a Bioline‐coated (group A, n = 40) circuit; mini‐CPB uncoated (group B, n = 40) circuit; CCPB (group C, n = 40). In this analysis part A, participants from group A (mini‐CPB coated circuit, 40 participants) versus half of group C (CCPB, 20 participants) were assessed. Inclusion criteria: adults undergoing elective CABG Exclusion criteria were reoperation, concomitant valve procedures, relevant aortic valve insufficiency, recent MI within the last 2 weeks before operation, impaired LV function (EF 35%), previous transient ischaemic attack or CVA, significant carotid artery stenosis (70%), renal failure (Cr 2 mg/dL), evidence of neurologic disease (epilepsy, Alzheimer disease, and dementia), and urgent surgery.
Interventions	Miniaturised heparin‐coated CPB circuit CCPB
Outcomes	Blood levels of myocardial (CK, CK‐MB, troponin T) and inflammatory markers (IL‐6, IL‐8, and TNF‐α) and Hct and Plt counts were serially measured at different time points (preoperatively and 0 hours, 1 hour, 6 hours, 12 hours, 24 hours, and 48 hours after CPB). Operation time (OR time, minutes), ECC and cross‐clamp time (minutes), number of distal anastomoses and grafts used, ventilation time (minutes), ICU stay (hours), hospital stay (days), chest‐tube drainage (mL/d), transfusion rates of packed RBCs and FFP, incidence of myocardial ischaemia, bypass stenosis, respiratory failure (acute respiratory distress syndrome), low cardiac output syndrome, AF, re‐thoracotomy, deep sternal wound infection, CVA, early mortality, and early clinical outcome.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but random sequence generation method is not described.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel. However, because of the nature of the intervention, surgical team was likely to be aware of the circuit used.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No blinding of outcome assessment method described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants are accounted for in the analysis.
Selective reporting (reporting bias)	High risk	Although the authors clearly define the tests that they are running in the method section and these are all reported in the results, there are significant concerns over the graph depicting levels of IL‐6, IL‐8, and IL‐10. The authors do not specify whether there is a 95% CI represented or a standard error. Furthermore, each time point has only got a single error bar which is unclear to which value it refers.
Other bias	High risk	Study funded by two of the manufacturers of the circuits used

Schoenebeck 2010 (B).

Study characteristics
Methods	Prospective, randomised controlled trial Enrolment period: March 2004 and December 2006 Number of study centres and locations: single centre. Department of Cardiovascular Surgery, University Heart Center Hamburg, Hamburg, Germany
Participants	A total of 120 participants underwent elective CABG and were randomised into three groups: mini‐CPB using a Bioline‐coated (group A, n = 40) circuit; mini‐CPB uncoated (group B, n = 40) circuit; and CCPB (group C, n = 40). In this analysis part B, participants from group B (mini‐CPB uncoated circuit, 40 participants) versus half group C (CCPB, 20 participants) were assessed. Inclusion criteria: adults undergoing elective CABG Exclusion criteria were reoperation, concomitant valve procedures, relevant aortic valve insufficiency, recent MI within the last 2 weeks before operation, impaired LV function (EF 35%), previous transient ischaemic attack or CVA, significant carotid artery stenosis (70%), renal failure (Cr 2 mg/dL), evidence of neurologic disease (epilepsy, Alzheimer disease, and dementia), and urgent surgery.
Interventions	Miniaturised uncoated CPB circuit CCPB
Outcomes	Blood levels of myocardial (CK, CK‐MB, troponin T) and inflammatory markers (IL‐6, IL‐8, and TNF‐α) and Hct and Plt counts were serially measured at different time points (preoperatively and 0 hours, 1 hour, 6 hours, 12 hours, 24 hours, and 48 hours after CPB). Operation time (OR time, minutes), ECC and cross‐clamp time (minutes), number of distal anastomoses and grafts used, ventilation time (minutes), ICU stay (hours), hospital stay (days), chest‐tube drainage (mL/d), transfusion rates of packed RBCs and FFP, incidence of myocardial ischaemia, bypass stenosis, respiratory failure (acute respiratory distress syndrome), low cardiac output syndrome, AF, re‐thoracotomy, deep sternal wound infection, CVA, early mortality, and early clinical outcomes.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that patients were randomised, but random sequence generation method is not described.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel. However, because of the nature of the intervention, the surgical team was likely to be aware of the circuit used.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No blinding of outcome assessment method described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants are accounted for in the analysis.
Selective reporting (reporting bias)	High risk	Although the authors clearly define the tests that they are running in the method section and these are all reported in the results, there are significant concerns over the graph depicting levels of IL‐6, IL‐8, and IL‐10. The authors do not specify whether there is a 95% CI represented or a standard error. Furthermore, each time point has only got a single error bar which is unclear to which value it pertains. This makes extracting data from the graph impossible.
Other bias	High risk	Study funded by two of the manufacturers of the circuits used

Scholz 2002.

Study characteristics
Methods	Randomised controlled trial, single centre Run‐in period: N/R Number of study centres and locations: single centre. Johann Wolfgang Goethe University, Frankfurt am Main, Germany
Participants	A total of 80 participants were assessed in this study as adults undergoing CABG. Elective CABG patients (n = 80) older than 50 years with EF of > 55% and Parsonnet score < 20 from our clinic were allocated to four groups (n = 20 each) without (group I) or with leucocyte filtration beginning with onset of CPB (group II), beginning 5 minutes before aortic declamping (group III), and with aortic declamping (group IV). In our analysis, we considered patients with leucocyte filters at all timings as one intervention group (II, III, and IV) and without leucocyte‐depleting arterial filters as a control.
Interventions	Intervention: leucocyte‐depleting arterial filter at onset of CPB Control: conventional management
Outcomes	Number of leuckocytes; leukocyte kinetics to assess cells with phagocytic activity; PMN‐Elastase and MPO
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors do not describe their randomisation protocol.
Allocation concealment (selection bias)	Unclear risk	Authors do not state the allocation concealment method employed.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Authors do not state whether participants or personnel were blinded to the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors do not state whether assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	Unclear how many participants were included in analysis. Note below bar charts states that data from 31 patients were used in the first time point, but it is unclear why only 31 and whether any participants were withdrawn from analysis.
Selective reporting (reporting bias)	Low risk	Although there is not a protocol available, the authors clearly define the tests that they are running in the methodology section, and these are all reported in the results.
Other bias	Unclear risk	No statement about funding sources

Schulz 2002.

Study characteristics
Methods	Prospective, randomised study Run‐in period: August 2000 and April 2001 Number of study centres and locations: single centre. University of Munich, Germany
Participants	Thirty‐six patients with CAD of various degrees undergoing CABG surgery using ECC were prospectively enroled in this study. Exclusion criteria were impaired renal function with serum Cr levels > 1.6 mg/dL, liver function test results indicating impaired liver function, a history of diabetes mellitus, acute infections, and pre‐existing chronic inflammatory disease.
Interventions	Group 1 (n = 18): CPB using extracorporeal circuits coated with PC (Stöckert Physio, Munich, Germany). Coated components included arterial and venous lines, the cardioplegia line, cannulas, filters, and membrane oxygenators. The identical circuit was used in group 2 (control; n = 18) without the use of coatings.
Outcomes	TNF‐α, IL‐10, P‐selectin, and NO levels Blood samples were collected before cannulation and at 2 hours, 24 hours, and 48 hours postoperation. Clinical data measured perioperatively included haemodynamics, aortic clamp time, duration of bypass, time to extubation, catecholamine requirement, length of ICU stay, postoperative blood loss, and amount of blood transfused.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment method not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected data were reported properly.
Other bias	Low risk	No funding was disclosed.

Schulze 2009.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2009 Number of study centres and locations: single centre, University of Alberta, Edmonton, Canada
Participants	A total of 30 consecutive participants undergoing first‐time elective CABG Mean age: 64.5 years Sex (female/male ratio): 0.11 Low‐risk patients (elective CABG) Exclusion criteria were previous cardiac surgery or combined procedure, renal or liver dysfunction, and preoperative coagulation disorder with or without impaired Plt function or decreased Plt count (< 100 × 109/L).
Interventions	PC group: n = 15, PC‐coated circuit C‐coated Control group: n = 15, noncoated circuit
Outcomes	Proinflammatory cytokines (IL‐6 and TNF‐α), markers for Plt activation (CD41a, CD42b, and CD62p), and total Plt count Whole blood samples from the arterial line were obtained at the following times: prior to induction of anaesthesia (Pre); 15 minutes after the initiation of CPB (CPB); 30 minutes after termination of CPB (Post). Clinical data were recorded, and intraoperative variables including the duration of aortic cross‐clamping, the duration of CPB, number of grafts, and the amount of postoperative (24 hours) total blood loss were analysed.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were prospectively randomised, but there is no explanation of how this was done.
Allocation concealment (selection bias)	Unclear risk	No mention of allocation concealment method used
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding of outcome assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	High risk	This study was supported by Dideco s.r.l. (Mirandola, Italy), which is the producer of the PC‐coated components for the CPB circuit.

Schurr 2001.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: August 1999 and November 2000 Registration date: study published in 2001 Number of study centres and locations: single centre, authors from University Hospital Zurich, Zurich, Switzerland
Participants	A total of 50 adults undergoing primary isolated CABG Mean age: 64 years Sex (female/male ratio): 0.16 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes Exclusion criteria: participants with insulin‐dependent diabetes mellitus, peptic ulcer history, malignant tumours, immunologic deficiencies, renal or hepatic insufficiency, and COPD
Interventions	Group A (n = 24): 10 mg/kg MPSS (Solu‐Medrol; Pharmacia & Upjohn AG, Duebendorf, Switzerland) intravenously 4 hours before the operation Group B (n = 26) served as a control.
Outcomes	Preoperative blood samples (10 mL) were taken before the administration of 10 mg/kg body weight MPSS, delivered 4 hours before induction of anaesthesia. Postoperatively, blood samples were collected after 24 hours, 48 hours, and on the sixth POD. The following measurements were taken: IL‐2R, IL‐6, IL‐8, TNF‐α, and the soluble adhesion molecules sE‐selectin and sICAM‐1. Medical history, demographic data, and the clinical course were analysed for each participant. Postoperatively, fluid balance; haemodynamic measurements; time on respirator; blood loss; occurrence of AF; renal, hepatic, and coagulation disorders; and pulmonary infection requiring antibiotic treatment were registered, as well as the duration of ICU stay and hospitalisation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described: "…Fifty patients undergoing elective coronary operations under normothermic CPB were randomized into two groups".
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No blinding of participants and personnel described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	Some prespecified outcomes are not reported, i.e. IL‐8, IL‐2, and sICAM‐1 authors state "…were also reduced, but again the difference did not reach statistical significance", but no values are reported.
Other bias	Unclear risk	No statement provided with regard to possible conflict of interest

Scrascia 2012.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: September 2009 to March 2010 Registration date: study published in 2012 Number of study centres and locations: single centre, authors from University of Bari "Aldo Moro", Bari, Italy
Participants	A total of 34 first‐time, elective, isolated CABG patients were enrolled. Mean age: 68.5 years Sex (female/male ratio): 0.62 Low‐risk patients (elective CABG) Exclusion criteria: age > 80 years old; preoperative Hb < 12 g/dL; BSA < 1.7 m2; re‐do or emergency surgery; valvular, thoracic aorta, or combined procedures; liver insufficiency (Child‐Pugh B or C class); Plt count below 50,000 or anti‐Plt treatment taken within 5 days before surgery; pre‐existing haemolytic or haemostatic disorders; anticoagulant treatment; inflammatory disorders or steroids treatment
Interventions	Cell salvage group (n = 17): processed CPB circuit residual blood infusion Control group (n = 17): conventional management (circuit blood discarded)
Outcomes	The primary endpoint was to evaluate the influence of CPB circuit residual blood salvage infusion after cell‐saving treatment on inflammatory, coagulative, and fibrinolytic system activation, measuring specific parameters (PF 1.2, PAP, PAI‐1, IL‐6, Hb, and heparin levels). Blood samples were collected from a peripheral arterial line after the induction of anaesthesia (preoperatively), 2 hours after the end of CPB (T0) (at this time, the collecting bag was transfused to each patient), and 24 hours after surgery (T1). Blood samples were also taken from the collecting bag after the washing and concentration procedure and prior to infusion into the patient (Tbag). The secondary endpoint was to evaluate the influence of pump blood salvage on postoperative Hb levels and transfusion rate.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomization was computer generated and took place the day before surgery."
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	No funding received, and authors declared conflicts of interests (none to declare).

Serrano 2010.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified. Author stated 31 months. Registration date: unspecified, study published in 2010 Number of study centres and locations: single centre, authors from University of São Paulo, Brazil
Participants	A total of 81 participants were enrolled in this study. Mean age: 62.8 years Sex (female/male ratio): 0.52 Low‐risk patients (elective CABG) Inclusion criteria: adults with low‐risk stable CAD and preserved LVEF Exclusion criteria: age above 75 years, combined cardiac procedures, cardiac reoperations, left main stenosis of greater than 70%, LVEF lower than 0.50, inflammatory and consumptive diseases, acute infection, colagenoses, renal failure (Cr N 1.5 mg/dL), respiratory distress, emergency surgeries, acute coronary syndromes, use of vasoactive drugs 7 days before surgery date, coagulopathy, stroke, and ICU patients
Interventions	Off‐pump CPB
Outcomes	Complications of cardiac surgery. The aim of this study was to compare specific aspects of inflammation and myocardial injury between on‐pump and off‐pump coronary surgery. Plasma concentrations of CRP, IL‐6 and IL‐8, sP‐selectin, sICAM‐1, CK‐MB, and cTnI were determined for this purpose. In addition, associations between perioperative inflammation and myocardial damage and 1‐year all‐cause mortality were sought. Peripheral venous blood samples were obtained from patients preoperatively and 1 hour, 24 hours, and 5 days after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomization was performed by chance the day before surgery using a computerized randomization system."
Allocation concealment (selection bias)	Unclear risk	Authors do not describe how this was achieved.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The authors state that "All incisions and closure techniques were the same for groups, limiting variability and maintaining blinding of group assignment for patients, family, and referring cardiologists", which would suggest that the participants were blinded. The blinding was not present for the operating theatre personnel, and it is unclear whether the person making the decisions to alter the bypass protocol was involved in the study analysis.
Blinding of outcome assessment (detection bias) All outcomes	High risk	As discussed previously, although the participants were allocated to an intervention group randomly, there was the option to change the allocation during surgery if the conditions did not allow for either on‐ or off‐pump CPB. Although this was appropriate to maintain patient safety, it is not clear whether the operating surgeon then had any involvement in data collection and analysis. Involvement in both the action of breaking the randomisation protocol and the analysis could introduce significant bias and could have been controlled by excluding the operating surgeons from the analysis. The authors do not clarify this point in the method sections.
Incomplete outcome data (attrition bias) All outcomes	High risk	Six participants were withdrawn from the intervention arm and switched to the control arm; while this was done for clinical reasons, this could have significantly impacted the results. Secondly, because of the study duration, only 80 participants completed the full follow‐up, which could have underpowered the study.
Selective reporting (reporting bias)	Low risk	Although no protocol was reported, the authors clearly state which biomarkers they wish to measure in the introduction and method sections, and these are reported in the results section.
Other bias	Low risk	Funding source declared as a nonprofit organisation. No apparent conflict of interests

Shimamoto 2000.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: March 1995 and March 1997 Registration date: study published in 2000 Number of study centres and locations: single centre, authors from Mie University School of Medicine, Tsu, Japan
Participants	A total of 32 adults who were scheduled for elective CABG surgery using CPB were included. Mean age: 62.1 years Sex (female/male ratio): 0.42 Low‐risk patients (elective CABG) Inclusion criteria: adult patients undergoing elective CABG Exclusion criteria: not specified
Interventions	Heparin‐coated circuit
Outcomes	"We measured the changes in inflammatory cytokine (tumor necrosis factor‐a [TNF‐a], interleukin [IL]‐1b, IL‐6, and IL‐8), terminal complement complex (C5b‐9), and polymorphonuclear elastase (PMN‐E) concentrations following CPB to analyze the biocompatibility of the siliconecoated oxygenator with respect to its effect on the systemic inflammatory response." Arterial blood was withdrawn at the following times: before CPB; 5 minutes and 1 hour after the start of CPB; immediately after CPB; 1 hour and 3 hours after the termination of CPB; and on the first POD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that the participants were prospectively randomised, but they do not state how this was achieved.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment method described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No mention of blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No mention of blinding outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Although no protocol was reported, the authors clearly state which measures they would like to measure and analyse.
Selective reporting (reporting bias)	Low risk	Although no protocol was reported, the authors clearly state which measures they would like to measure and analyse.
Other bias	Low risk	No concerns

SIRS 2015.

Study characteristics
Methods	Double‐blind, randomised controlled trial substudy Run‐in period: not specified Study date: published 2015 Number of study centres and locations: 80 hospital‐based cardiac surgery practices in 18 countries by dedicated local research teams
Participants	7507 participants who were randomised to receive either IV MPSS (n = 3755) or placebo (n = 3752). Eighty‐one participants were enrolled in the substudy (37 placebo versus 44 in treatment group). Mean age: 71.09 years Sex (female/male ratio): 43.21% High‐risk patients Inclusion criteria: any ethnicity, greater than 18 years of age, CPB for any cardiac surgical procedure (such as CABG, valve, aorta, or combined procedures); patients aged 18 years or older were eligible if they had a EuroSCORE of at least 6 and provided written informed consent. Exclusion criteria: if patients were taking or expected to receive systemic steroids in the immediate postoperative period or had a history of bacterial or fungal infection in the preceding 30 days
Interventions	Intervention group: 500 mg of MPSS (250 mg at anaesthetic induction and 250 mg at initiation of CPB) Control group: placebo
Outcomes	PF1+2 (thrombin generation), PAP (fibrinolysis), PF4 Plt activation, and inflammatory and coagulation biomarkers (fibrinogen, CRP, WBC count, antithrombin, bilirubin, aspartate transaminase, and alanine transaminase) and clinical outcome (postoperative bleeding and need for blood product transfusion, Hct and Hb, death, stroke, renal and lung injury, infection) Blood samples were drawn from a peripheral vein or arterial catheter at the following time intervals: preoperatively (within 7 days before surgery) (T0), 30 minutes post‐CPB initiation (T1), 15 minutes after aortic cross‐clamp release (T2), 2 hours postoperatively (T3), day 1 postoperatively (T4), and day 5 postoperatively or upon discharge (T5).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Research personnel randomised participants via a 24‐hour interactive web randomisation system maintained by the coordinating centre at the Population Health Research Institution, which is part of the Hamilton Health Sciences and McMaster University, Hamilton, Ontario, Canada. The randomisation process for the main study uses block randomisation stratified by centre. Participants in this substudy were consecutively selected from participating centres. Baseline clinical and operative characteristics of participants were similar between the study groups.
Allocation concealment (selection bias)	Low risk	"Methylprednisolone was obtained from the centre's local pharmacy, and the study drug was prepared and masked by the local pharmacy following procedures described in a provided study manual."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, healthcare providers, data collectors, and outcome adjudicators were blinded to treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants, healthcare providers, data collectors, and outcome adjudicators were blinded to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	7507 participants were included in the trial; 3755 were allocated to the treatment group and 3752 to the placebo group. All participants were included in the 30‐day intention‐to‐treat analysis. At 6‐month follow‐up intention‐to‐treat analysis, three lost at follow‐up were reported for the treatment group and five lost at follow‐up for the placebo group. Quote: "The 30‐day follow‐up was complete in 7507 patients (100%) and the 6‐month follow‐up was complete in 7499 patients (99·9%)."
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	The authors have no conflict of interest to declare. The study received funding from the Canadian Institutes of Health Research and the University of Bari Aldo Moro.

Sirvinskas 2007.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: 2005 to 2006 Registration date: 2007 Number of study centres and locations: single centre, authors from Kaunas University of Medicine, Lithuania
Participants	A total of 90 patients undergoing valve procedure, CABG, or both were included in this study. Mean age: 62.9 years Sex (female/male ratio): 0.5 Low‐risk patients (elective CABG or valve procedure or both) Inclusion criteria: elective cardiac surgery, valvular, CABG, or both Exclusion criteria: previous cardiac surgery, emergency operations, preoperative use of steroids, nonsteroidal anti‐inflammatory or anti‐Plt drugs, patients with inherent and acquired coagulation disorders, perioperative MI, postoperative low output heart failure, and reoperation for bleeding
Interventions	Shedding of mediastinal blood
Outcomes	The aim of this study was to evaluate the efficacy of reinfusion of autologous shed mediastinal blood on a patient’s postoperative course. "Haemoglobin (Hb), haematocrit (Hct), C‐reactive protein values, and leucocyte count were compared before surgery, at 4 h and 20 h after surgery, and on the fifth postoperative day. We have measured serum procalcitonin (PCT) concentration at 4 h and 20 h after CPB. We assessed drained blood loss within 20 postoperative hours".
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Authors do not state whether the participants were randomly assigned to either intervention arm.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	It is unclear how many participants were included in the final analysis. Furthermore, the exclusion criteria include "perioperative myocardial infarction, postoperative low output heart failure, and reoperation for bleeding", which would increase the concern that not all participants were included in the final analysis and could lead to biased data.
Selective reporting (reporting bias)	High risk	No clear protocol and no clear statement of the outcome of the study
Other bias	Unclear risk	No statement with regard to funding or any conflict of interest

Skrabal 2006 (A).

Study characteristics
Methods	Single‐centre, prospective randomised trial Run‐in period: unspecified Registration date: unspecified, study published in 2006 Number of study centres and locations: single centre, authors from University of Rostock, Germany
Participants	A total of 39 adults who were scheduled for elective CABG surgery using CPB were included. Mean age: 63.5 years Sex (female/male ratio): 0.22 Low‐risk patients (elective CABG) Exclusion criteria: age older than 80 years, urgency/emergency operation, valve surgery, redo or combined cardiac surgery, significant stenosis of the carotid arteries, severe renal dysfunction with/without requiring dialysis, hepatic disorder, clinically evident cognitive impairment preoperatively, and history of previous neurological or neurocognitive disorders
Interventions	Group A (n = 19): PMEA‐coated lines and oxygenators (group A, CAIPOX RX 25, Terumo Corp., Tokyo, Japan) Group B (n = 20): noncoated circuits (group B, CAIPOX SX, Terumo Corp., Tokyo, Japan)
Outcomes	At each sampling point, leucocyte counts, plasma concentration of CRP, IL‐6, PMN‐Elastase, anaphylatoxin C3a, and b‐TG were measured. Blood samples were taken before CPB (T0) and 20 minutes (T1) and 90 minutes (T2) after CPB initiation. Neurological assessment was performed to investigate the participants' neurocognitive outcome after cardiac surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state the participants were randomised but do not give details of how this was done.
Allocation concealment (selection bias)	Unclear risk	No details given about concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not described: unclear whether participants or theatre staff were blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors state that the investigator carrying out the neurocognitive testing was blinded. It is unclear whether the assessors of the biological variables were blinded; however, these were less likely to be affected by lack of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	High risk	Sponsored by the manufacturer of the coated circuit

Skrabal 2006 (B).

Study characteristics
Methods	Single‐centre, prospective randomised trial Run‐in period: unspecified Registration date: unspecified, study published in 2006 Number of study centres and locations: single centre, authors from University of Rostock, Germany
Participants	A total of 35 adults undergoing CABG were assessed in this study. Mean age: not reported Sex (female/male ratio): not reported Low‐risk patients (elective CABG) Exclusion criteria: not specified
Interventions	Group A (n = 15, retransfusion group): the pericardial suction blood collected during CPB was retransfused. Group B (n = 20, no‐retransfusion group): pericardial suction blood was separated using an innovative hard shell reservoir with separate chambers for venous return and cardiotomy suction blood (D 903 Avant, Dideco s.r.l, Mirandola, Modena, Italy); in this group, the separated suction blood was retransfused in the ICU only if necessary. Otherwise, it was discarded.
Outcomes	At each sampling point, leucocyte counts; levels for CRP, IL‐6, and PMN‐Elastase; Plt counts; b‐TG; soluble P‐selectin; fHb; LDH; and anaphylatoxin C3a were measured. Blood samples were collected in citrate and EDTA tubes before (T0) and after CPB (T1), but latest after 90 minutes on CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Thirty‐five patients subjected to elective coronary artery bypass grafting were prospectively randomized into two groups. Randomization plan with five sub‐blocks was generated by an open‐access web‐based tool."
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, and it would be unlikely that trial personnel would have been blinded to the interventions in view of the different devices used. As the trial personnel could decide whether to retransfuse the discarded blood or not, and it is not clear from the methodology whether clinical personnel making the decision to transfuse were involved in the trial and whether there were set criteria for retransfusion of shed mediastinal blood, this bias was rated as high.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No declaration in regard to sponsorship

Sobieski 2008.

Study characteristics
Methods	Single‐centre, placebo‐controlled, prospective, randomised trial Run‐in period: unspecified Registration date: unspecified, study published in 2008 Number of study centres and locations: single centre, authors from Division of Cardiac Surgery, Advocate Christ Medical Center, Oak Lawn, Illinois
Participants	A total of 28 adults who were scheduled for elective CABG surgery using CPB were included. Mean age: 63.3 years Sex (female/male ratio): 0.18 Low‐risk patients (elective CABG) Inclusion criteria: patients undergoing elective coronary artery revascularisation, less than 80 years old Exclusion criteria: clinically significant end‐organ dysfunction, AMI, IABP; patients actively receiving corticosteroids, anti‐inflammatory (aspirin, dipyridamole, ibuprofen) or anticoagulation therapies (Coumadin, Lovenox) for other medical conditions
Interventions	Study group (n = 13) received 100 mg of dexamethasone. Control group (n = 15) received sterile saline.
Outcomes	The goal of this study was to evaluate the effectiveness of a single dose of dexamethasone given after induction of mechanical ventilation and general anaesthesia but before initiation of CPB to determine whether it could successfully blunt the effects of CPB‐induced SIRS. Secondary goals included its effect on clinical outcomes, including the possible adverse event of surgical site infections from the use of steroids. Blood samples were obtained at baseline after intubation, 30 minutes after initiation of CPB, and at 24 hours and 72 hours after termination of CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"On the day of surgery, patients enrolled in the study were randomized by the pharmacy."
Allocation concealment (selection bias)	Low risk	"The anesthesiologist was given 10 ml syringes labeled only with the patients’ names. Patients in the control group received 10 ml of sterile saline whereas the patients in the study group received 100 mg of dexamethasone (diluted to 10 ml of solution). At the completion of the study, the pharmacy unblinded all the patient treatment groups."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The anesthesiologist was given 10 ml syringes labeled only with the patients’ names. Patients in the control group received 10 ml of sterile saline whereas the patients in the study group received 100 mg of dexamethasone (diluted to 10 ml of solution). At the completion of the study, the pharmacy unblinded all the patient treatment groups."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described – the blinding is lifted at the end of the process, but it is unclear whether the assessors were blinded or not.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the analysis.
Selective reporting (reporting bias)	High risk	No protocol and no outcomes clearly specified – however, in the introduction, the authors specify that their goal was also to investigate secondary outcomes such as the impact of dexamethasone on possible wound complications, which is not addressed in the paper.
Other bias	Unclear risk	No declaration with regards to sponsorship

Sohn 2009 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: from August 2007 to April 2008 Registration date: not available, published in 2009 Number of study centres and locations: single centre, authors from University of Saskatchewan, Saskatoon, Canada
Participants	Patients from part A were included in this analysis (see Interventions). A total of 78 participants undergoing elective CABG were assessed. The study was divided into four parts, i.e. A, B, C, and D. The number of participants was as follows – A: 20, B: 20, C: 19, and D:1 9. There were 16 participants in each group, except for the control group (n = 14). Mean age: 65 years Sex (female/male ratio): 0.54 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: subjects who recently attended emergency or who suffered from any chronic inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, Crohn’s disease, and COPD
Interventions	Part A: trillium‐coated circuits Part B: Bioline‐coated circuits Part C: PC‐coated circuits Part D: PMEA‐coated circuits Control: an uncoated circuit
Outcomes	TNF‐α, IL‐6, and IL‐10 Samples collected before CPB, 6 hours after CPB, and 72 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that patients were randomised, but process was not described: "…were randomly assigned to five groups with different biocompatible coated circuits".
Allocation concealment (selection bias)	Unclear risk	No details regarding allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	One of the authors received a non‐industry‐funded scholarship.

Sohn 2009 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: from August 2007 to April 2008 Registration date: not available, published in 2009 Number of study centres and locations: single centre, authors from University of Saskatchewan, Saskatoon, Canada
Participants	Patients from part B were included in this analysis (see Interventions). A total of 78 participants undergoing elective CABG were assessed. The study was divided into four parts, i.e. A, B, C, and D. The number of participants was as follows – A: 20, B: 20, C: 19, and D: 19. There were 16 participants in each group, except for the control group (n = 14). Mean age: 65 years Sex (female/male ratio): 0.54 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: subjects who recently attended emergency or who suffered from any chronic inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, Crohn’s disease, and COPD
Interventions	Part A: trillium‐coated circuits Part B: Bioline‐coated circuits Part C: PC‐coated circuits Part D: PMEA‐coated circuits Control: an uncoated circuit
Outcomes	TNF‐α, IL‐6, and IL‐10 Samples collected before CPB, 6 hours after CPB, and 72 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that patients were randomised, but process was not described: "…were randomly assigned to five groups with different biocompatible coated circuits".
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	One of the authors received a non‐industry‐funded scholarship.

Sohn 2009 (C).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: from August 2007 to April 2008 Registration date: not available, published in 2009 Number of study centres and location: single centre, authors from University of Saskatchewan, Saskatoon, Canada
Participants	Patients from part C were included in this analysis (see Interventions). A total of 78 participants undergoing elective CABG were assessed. The study was divided into four parts, i.e. A, B, C, and D. The number of participants was as follows – A: 20, B: 20, C: 19, and D: 19. There were 16 participants in each group, except for the control group (n = 14). Mean age: 65 years Sex (female/male ratio): 0.54 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: subjects who recently attended emergency or who suffered from any chronic inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, Crohn’s disease, and COPD
Interventions	Part A: trillium‐coated circuits Part B: Bioline‐coated circuits Part C: PC‐coated circuits Part D: PMEA‐coated circuits Control: an uncoated circuit
Outcomes	TNF‐α, IL‐6, and IL‐10 Samples collected before CPB, 6 hours after CPB, and 72 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that patients were randomised, but process was not described: "…were randomly assigned to five groups with different biocompatible coated circuits".
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	One of the authors received a non‐industry‐funded scholarship.

Sohn 2009 (D).

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: from August 2007 to April 2008 Registration date: not available, published in 2009 Number of study centres and locations: single centre, authors from University of Saskatchewan, Saskatoon, Canada
Participants	Participants from part D were included in this analysis (see Interventions). A total of 78 participants undergoing elective CABG were assessed. The study was divided into four parts, i.e. A, B, C, and D. The number of participants was as follows – A: 20, B: 20, C: 19, and D: 19. There were 16 participants in each group, except for the control group (n = 14). Mean age: 65 years Sex (female/male ratio): 0.54 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: subjects who recently attended emergency or who suffered from any chronic inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, Crohn’s disease, and COPD
Interventions	Part A: trillium‐coated circuits Part B: Bioline‐coated circuits Part C: PC‐coated circuits Part D: PMEA‐coated circuits Control: an uncoated circuit
Outcomes	TNF‐α, IL‐6, and IL‐10 Samples collected before CPB, 6 hours after CPB, and 72 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors state that participants were randomised, but process was not described: "…were randomly assigned to five groups with different biocompatible coated circuits".
Allocation concealment (selection bias)	Unclear risk	No details given around concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described, plausible for participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No clear details given
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Low risk	One of the authors received a non‐industry‐funded scholarship.

Soleimani 2018.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: between August 2015 and April 2016 Registration date: 2014, Iranian Registry of Clinical Trials Database (IRCT2015040921669N1) Number of study centres and locations: single centre; authors from University of Mazandaran University of Medical Sciences, Sari, Iran
Participants	A total of 150 adults who were scheduled for elective CABG surgery using CPB were included. Mean age: 62 years Sex (female/male ratio): 0.83 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes (age, 35 years to 75 years) Exclusion criteria: history of previous cardiac surgery, prior treatment with antiarrhythmic medications (except beta‐blockers), a history of heart failure with EF < 30%, left atrium size > 55 mm, obstructive sleep apnoea, a need for more than 4 grafts, renal failure (serum Cr > 1.5 mg/dL on 2 consecutive tests), liver dysfunction (hepatic enzymes elevated over 1.5 times the normal value), a history of COPD, hyperthyroidism, grade II to III atrioventricular block, serum pH < 7.25 or > 7.55, or experienced MI after surgery
Interventions	Group A: after induction of anaesthesia, participants received IV NAC infusion at a dose of 50 mg/kg, diluted to a total volume of 50 cc of normal saline, over a period of 30 minutes. Furthermore, these participants received 2 IV doses of 50 mg/kg NAC (with each dose occurring over a period of 30 minutes) on days 1 and 2 after surgery. Group B: normal saline (as a placebo) with the same volume at the same time interval
Outcomes	Postoperative variations of hs‐CRP levels, the length of ICU and hospital stay, and the incidence of MACCEs, which was defined as the occurrence of any of the following complications during the perioperative period: AMI, nonfatal cardiac arrest, unstable angina pectoris, heart failure, postoperative stroke, and death Prior to and on the third day after the surgery, blood samples were obtained for evaluation of the hs‐CRP level.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbering system: "Patients who fulfilled the inclusion criteria were randomly divided into groups A and B (75 patients in each group) using a sealed envelope technique with a computer‐generated random numbering system with the help of a nurse anaesthetist who was blinded to the study groups."
Allocation concealment (selection bias)	Low risk	Use of opaque sealed envelopes: "using a sealed envelope technique"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Research staff was likely blinded to the intervention: "All pharmaceutical preparations were done by an anaesthesiology resident who was not involved in the conduct of the study."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Not clearly described but plausible
Incomplete outcome data (attrition bias) All outcomes	High risk	Participants were excluded from the analysis if they developed cardiovascular instability, needed reoperation, or needed a pacemaker; as two of these factors could occur if participants developed postoperative AF, this could have introduced bias.
Selective reporting (reporting bias)	Low risk	Protocol available: all prespecified outcomes were reported.
Other bias	Low risk	Declaration of conflicts of interest present, and authors declare no conflicts of interest.

Song 2013.

Study characteristics
Methods	Single‐centre, prospective, placebo‐controlled, double‐blinded, randomised controlled trial Run‐in period: between March 2008 and December 2008 Registration date: not specified Number of study centres and locations: single centre, authors from Konkuk University Medical Center, Konkuk University School of Medicine, Seoul, Korea
Participants	Twenty‐six participants undergoing elective cardiac surgery requiring CPB Mean age: 58 years Sex (female/male ratio): 1.2 Preoperative exclusion criteria were urgent/emergency surgery, previous heart surgery, diabetes, ischaemic heart disease, combined surgery with a CABG procedure, age > 75 years, LVEF < 45%, diabetes, active gastropathic disorder, COPD, preoperative administration of furosemide, and renal failure requiring renal replacement therapy. The intraoperative exclusion criteria included intraoperative use of an IABP, intraoperative administration of steroids or TXA, and intraoperative transfusion of FFP or Plt concentrates during CPB. The postoperative exclusion criteria included reoperation for surgical correction of an intractable postoperative bleeding within 2 hours after the end of surgery and transfusion of any banked blood product.
Interventions	Intervention: UST (5000 U/kg) administered intravenously 30 minutes to 40 minutes before initiating CPB Control: placebo. Same volume of 0.9% sodium chloride as UST
Outcomes	PaO2/FiO2, CK‐MB, TnI, GFR, and Cr IL‐6 and TNF‐α after anaesthesia induction, 1 hour after weaning from CPB, and at the end of surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Adequate randomisation process: "Twenty‐six patients undergoing elective cardiac surgery requiring CPB were assigned to a prospective, double‐blinded, randomized fashion to two groups: control (group C, n = 13) and ulinastatin pretreatment (group U, n = 13) during our study period. Patients were allocated to the randomization process using a patient identification number, which was stratified in a 1: 1 ratio, to receive either ulinastatin (Ulistin; Hallyim Inc., Markham, ON, Canada) or the same volume of normal saline."
Allocation concealment (selection bias)	Low risk	Use of patient identification number, unclear details around the process, but plausible allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The investigators and physicians who managed the participants were blinded to the participants' group allocation. All physicians and nursing staff who cared for the participants were unaware of the group allocations.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The investigators and physicians who managed the participants were blinded to the participants' group allocation. All physicians and nursing staff who cared for the participants were unaware of the group allocations. It is unclear whether outcome assessors were blinded, although laboratory outcomes were probably not affected by detection bias.
Incomplete outcome data (attrition bias) All outcomes	High risk	Some participants were removed from the final analysis as they met some of the intraoperative exclusion criteria; it is understandable why some of these criteria may have been instituted as these interventions could affect the way UST acts, but they are very artificial and would possibly lead to findings which could not be valid for actual clinical practice.
Selective reporting (reporting bias)	Low risk	No protocol is available, but the authors clearly detailed which parameters they were measuring in the methodology.
Other bias	Low risk	Authors declared sources of funding. At low risk of bias

Stammers 1997.

Study characteristics
Methods	Randomised, prospective, double‐blind, placebo‐controlled study Run‐in period: unspecified Registration date: unspecified, published in 1997 Number of study centres and locations: single centre, authors from University of Nebraska Medical Center, Omaha, Nebraska
Participants	A total of 20 adults undergoing elective CABG were assessed in this study. Mean age: 65 years Sex (female/male ratio): 0.25 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, both sexes (age, 19 years to 80 years) Exclusion criteria: semilunar valvular stenosis with aortic valve area ≤ 0.6 cm2, atrioventricular or semilunar valve regurgitation, preoperative coagulation disorders as assessed by a PT > 1.5 times normal, preoperative Hct < 33%, or patient weight < 50 kg
Interventions	Intervention: after a test dose of 1 mL, a loading dose of 2 × 106 (280 mg) KIU was administered intravenously immediately after the induction of anaesthesia. 2 × 106 KIU were placed in the priming volume of the extracorporeal circuit, and patients received a constant infusion of 5 × 105 KIU (70 mg) intravenously hourly until chest closure. Control: equal volume of saline administered in the same manner
Outcomes	"The purpose of the present study was to examine the anti‐inflammatory effects (total complement, fibrinogen split products (FSP), D‐dimers, protein C and protein S, factor XII and PMN elastase) of aprotinin in patients undergoing coronary artery bypass grafting (CABG) with CPB, preoperatively (post‐induction and prior to surgical incision), postoperatively (at chest closure), and on the first postoperative day (24 hours post‐chest closure)."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state that the study was a randomised double‐blind trial: "The study design included a randomized, prospective, double‐blind, placebo controlled study of aprotinin administration".
Allocation concealment (selection bias)	Unclear risk	Not described in sufficient detail: "All drugs were drawn up immediately prior to use by a pharmacist (DC) and placed in a 500 ml evacuated glass bottle which was labeled with the patient's name, registration number and date. No other clinician knew of the treatment received by the patient".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All medications drawn up by trial pharmacist and labelled with participant's name and study number and date. No other member of staff knew what the participant received.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blinded study. All medications drawn up by trial pharmacist and labelled with participant's name and study number and date. No other member of staff knew what the participant received. Unspecified whether outcome assessment was blinded as well, but unlikely source of detection bias because of outcome selection
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	No protocol was given, but all the tests were clearly detailed in the methodology.
Other bias	High risk	Industry funding: "The authors wish to thank Bayer Pharmaceuticals, West Haven, CT, for providing us with funding for this study".

Stammers 2001.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, published in 2001 Number of study centres and locations: single centre, authors from Mayo Clinic and Mayo Foundation, Rochester, MN
Participants	Thirty adult patients undergoing elective CABG surgery Mean age: 66 years Sex (female/male ratio): 0.3 Low‐risk patients (elective CABG) Inclusion criteria: adult patients undergoing elective CABG at authors' institution Exclusion criteria: none reported
Interventions	Intervention: trillium‐coated circuit Control: uncoated circuit
Outcomes	Hb, Hct, leucocyte count, Plt count, plasma TCC (SC5b‐9), complement activation products (C3a), MPO, b‐TG, PF1.2, PAP, heparin concentration, kaolin‐based ACT, Plt glass bead retention, and fibrinogen concentration; perioperative allogeneic and autologous (cell salvage) blood product usage; duration of endotracheal intubation; length of hospitalisation; and perioperative morbidity were recorded.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state that the participants were randomised in a double‐blinded fashion: "30 adult patients scheduled for elective primary coronary artery bypass graft surgery were randomized, using a double‐blinded procedure".
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Authors state that the trial was double‐blinded. However, they do not describe how this was achieved given the use of two different circuits.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors state that the trial was double‐blinded. No mention regarding blinding of outcome assessment specifically
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unclear whether all participants were included in the final analysis
Selective reporting (reporting bias)	Low risk	No protocol was given, but all the tests were clearly detailed in the methodology.
Other bias	Unclear risk	No statement about funding sources

Stefanou 2001.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre; authors from Imperial College School of Medicine, Hammersmith Hospital Campus, London
Participants	Twenty adults undergoing CABG Sex (female/male ratio): 0.1 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG
Interventions	Intervention: leucocyte‐depleting filter Control: standard CPB circuit
Outcomes	Our aim was to establish whether the use of leucocyte depletion filters is associated with any measurable clinical benefits. Clinical parameters were measured during inpatient recovery. Expression of CD11 b on neutrophils and production of MPO and lactoferrin were measured in arterial samples between induction and 3 hours post‐bypass.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear whether participants were randomised – methods state the participants were divided between two groups, but they were allocated by gender: "Twenty patients undergoing elective coronary revascularization were randomly divided into two groups".
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All members of staff blinded: "Throughout inpatient care, all members of staff were blind regarding which patients were part of the study and to which groups they belonged."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unclear whether all participants were included in the final analysis
Selective reporting (reporting bias)	Unclear risk	No protocol described
Other bias	Unclear risk	No statement around funding resources

Steinhoff 1996.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Run‐in period: from May through December 1995 Registration date: not specified, published in 1996 Number of study centres and locations: single centre, authors from Department of Heart and Vascular Surgery, University Kiel, Germany
Participants	A total of 28 adults undergoing CAB were assessed in this study. Mean age: 57 years Sex (female/male ratio): 0.08 Low‐risk patients (elective CABG) Inclusion criteria: age 40 years to 70 years; coronary three‐vessel disease (NYHA 2); absence of unstable angina; EF > 50%; primary open‐heart surgery: no calcification of the aorta; sinus rhythm until the start of CPB; no impairment of renal function; no MI in the 3 months before operation Exclusion criteria: malignancies, diabetes type I, and organ transplantation; early or severe ECG changes; extrasystoles or loss of sinus rhythm during the ischaemic preconditioning procedure
Interventions	Normothermic preconditioning
Outcomes	Parameters of endothelial activation and myocardial injury were analysed. ICU stay, postoperative ventilation, CVP, LAD‐pressure, mean systolic blood pressure, hourly diuresis, catecholamine dose, blood loss, reoperation, 30‐day outcome, hospital stay IL‐6, IL‐8, TNF‐α, 3 hours to 12 hours after reperfusion
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state the trial was randomised: "28 patients were included in a randomized study".
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described, unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	High risk	Although all participants were included in the analysis, the authors use different names for the groups in the tables, and they do not clearly label all diagrams, thus making it difficult to interpret the data and possibly introducing bias.
Selective reporting (reporting bias)	High risk	No protocol. Some of the outcomes that the authors stated they wanted to measure are not reported in the text.
Other bias	Unclear risk	No statement on funding sources

Sucu 2004.

Study characteristics
Methods	Single‐centre, prospective, placebo‐controlled, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: single centre, authors from Mersin University, Medical School, Mersin, Turkey
Participants	A total of 40 adults undergoing CABG were assessed in this study. Mean age: 65 years Sex (female/male ratio): 0.38 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: poor ventricular function (EF < 30%) or diabetes mellitus; remarkably abnormal pulmonary, endocrine, metabolic, or neurologic pathology; and those requiring emergency surgery
Interventions	Intervention: NAC 50 mg/kg intravenously per day, given as 25 mg/kg twice in 24 hours Control: saline 0.5 mL/kg twice in 24 hours for 3 days
Outcomes	Neutrophil activation, lipid peroxidation, and the acute‐phase reactants of patients undergoing CPB, by measuring MPO activity, IL‐6, MDA, CRP, and AAGP levels during surgery and postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state that the study participants were randomly allocated: "Forty patients undergoing coronary artery bypass grafting (CABG) were randomly divided".
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described clearly
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants included in the analysis
Selective reporting (reporting bias)	Low risk	No protocol was given, but all the variables investigated were clearly detailed in the introductions and reported on.
Other bias	Unclear risk	No statement with regard to funding or any conflict of interest

Svenmarker 2003.

Study characteristics
Methods	Single‐centre, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, trial published in 2003 Number of study centres and locations: single centre, authors from Department of Surgical and Perioperative Science, Cardiothoracic Division, Umea University Hospital, Sweden
Participants	A total of 33 adult males undergoing elective CABG were assessed in this study. Mean age: 58 years Sex (female/male ratio): 0 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: unspecified
Interventions	Intervention: washing of pericardial shed blood Control: retransfusion directly from the cardiotomy reservoir without processing
Outcomes	Plasma fHb, C3a, IL‐6, IL‐8, TNF‐α, eosinophils, and basophils Blood samples were collected at the following time points: before CPB; on release of the aorta clamp; during rewarming from the cardiotomy reservoir; during rewarming from the system circulation; at termination of CPB; 5 hours after CPB; and on the first POD.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Authors state that participants were randomly assigned to interventions: "Patients were randomized to receive the accumulated PSB volume using two strategies".
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described, and it would be difficult to blind staff to the different practice.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	It would appear that all the participants were included in the analysis.
Selective reporting (reporting bias)	Low risk	Although there was no protocol, all the outcomes were detailed in the methods and introduction.
Other bias	Low risk	Funding source declared as a national body. No conflicts of interest detected

Svitek 2009.

Study characteristics
Methods	Randomised controlled trial Run‐in period: between December 2006 and December 2007 Registration date: unspecified, study published in 2009 Number of study centres and locations: single centre, authors from University Hospital in Hradec Kralove, Czech Republic
Participants	A total of 54 adults undergoing CABG were assessed in this study. Mean age: 67 years Sex (female/male ratio): 0.17 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: urgent operations, reoperation, combined operations, operative risk higher than 5% (logistic EuroSCORE), preoperative level of serum Cr >130 mmol/L, hepatic disease, and malignancies
Interventions	Intervention: MECC system Control: standard CPB
Outcomes	IL‐6, PMN‐Elastase, and MCP‐1, collected before surgery (after induction to anaesthesia), at the initiation (clamp placement) and end (declamping) of ECC, the end of surgery, and on the first, third, and seventh PODs
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation with envelope system: "Randomization occured [sic] during admittance to the hospital using the envelope method"
Allocation concealment (selection bias)	Low risk	Participants were randomised using sealed opaque envelopes.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not stated; unlikely that personnel were blinded because of the presence of a different bypass circuit
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	No protocol and not stated clearly by the authors
Other bias	Low risk	Funding received by the Czech government; low risk of conflict of interests

Tabaei 2018.

Study characteristics
Methods	Double‐blinded, single‐centre, randomised controlled study Run‐in period: during the year 2017 Registration date: registered at Iranian Registry of Clinical Trials (IRCT) with the registration code of IRCT2017042127617N3, study published in 2018 Number of study centres and locations: single centre, authors from Iran University of Medical Sciences, Tehran, Iran
Participants	A total of 56 adults undergoing CABG were assessed in this study. Mean age: 59 years Sex (female/male ratio): 0.24 Low‐risk patients (elective CABG) Inclusion criteria: CPB time between 50 minutes and 120 minutes and the number of grafts up to 4 Exclusion criteria: patients with active cardiac disorders affecting post‐CABG recovery period; history of a sternotomy, cardiac arrest, or using a defibrillator in CABG operation; preoperative infection or coagulation disorders; and LVEF < 35%
Interventions	Intervention: during the CABG operation, participants received CUF and then were continued on MUF. Control: during the CABG operation, participants received CUF.
Outcomes	Inflammatory cytokine response, haemodilution, and rotational thromboelastometry outcomes in adults' CABG operation First sample, one minute after unclamping of the aorta clamp; secondary sample, after MUF (end of bypass); and the third sample 24 hours after ICU entrance
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised using a simple number table: "Simple randomization was used to allocate the patients to two groups, as previously described".
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Authors describe the trial as double‐blinded but do not describe how this was achieved. The risk is unclear because of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors describe the trial as double‐blinded but do not describe how this was achieved.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants included in the analysis
Selective reporting (reporting bias)	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Other bias	Low risk	Funding received by Iranian government

Taleska 2020.

Study characteristics
Methods	Double‐blinded, single‐centre, prospective, randomised controlled trial Run‐in period: between February 2016 and December 2017 Registration date: registered at ClinicalTrials.gov (NCT02666703) in 2015 Number of study centres and locations: Clinical Department of Cardiovascular Surgery at the University Medical Centre in Ljubljana, Slovenia
Participants	This study was split into parts A and B, with 30 participants in part A and 30 in part B; they were adults undergoing cardiac surgery. Inclusion criteria: patients > 18 years old who were admitted for elective complex cardiac surgery with an expected CPB duration of > 90 minutes. The surgery thus included combined valve and coronary bypass grafting surgery, concomitant surgery of 2 or more valves, surgery of ascending aorta and aortic arch, and reoperations of the same types. Exclusion criteria: refusal to participate in the study; age < 18 years; pregnant women; emergency procedures; heart transplantation; implantation of the LV assist device, right ventricular assist device, or total artificial heart; treatment with chemo/immunosuppressive therapy; treatment with antileucocyte drugs or TNF‐α blockers; immunocompromised patients (e.g. with AIDS); leucopenia (< 4.0 × 109 cells/L); clinical or laboratory signs of infection (i.e. CRP > 2 mg/dL (20 mg/L); procalcitonin > 0.5 μg/L; leucocytes > 10.0 × 109 cells/L); serum Cr > 2 mg/dL (176 μmol/L); bilirubin > 2 mg/dL (34.2 μmol/L); history of stroke; malnourished patients; BMI < 18 kg/(m2)
Interventions	Intervention: 1 g of MP added in the CPB priming solution Control: usual care
Outcomes	Cytokine levels (TNF‐α, IL‐1β, IL‐6, IL‐8, and IL‐10) and complement C5a, as well as expression of CD64 and CD163 markers on monocytes, granulocytes, and lymphocytes Changes in serum hs‐CRP and procalcitonin levels, leucocyte count, albumin, fibrinogen, and haemodynamic measurements (i.e. cardiac index, SVR index, central venous oxygen saturation, and MAP); duration of postoperative mechanical ventilation, length of ICU stay, use of inotropic/vasoactive drugs, use of fluid/blood products and insulin, length of in‐hospital stay, and 30‐day mortality
Notes	The only intervention arm considered in this review is the one with MP, as cytokine filters are not included in the protocol.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was carried out by one of the members of the study team a day before surgery and achieved by using identical sealed envelopes, whereby each participant selected an envelope that assigned him/her to 1 of 3 treatment groups. Randomisation allocation numbers were generated by the Research Randomizer (https://www.randomizer.org/).
Allocation concealment (selection bias)	Low risk	Randomisation was carried out by one of the members of the study team a day before surgery and achieved by using identical sealed envelopes, whereby each participant selected an envelope that assigned him/her to 1 of 3 treatment groups. Randomisation allocation numbers were generated by the Research Randomizer (https://www.randomizer.org/).
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, ICU and ward personnel, and laboratory staff who participated in the trial were "blinded" for assigned treatment throughout the duration of the study. Exception from being blinded was for personnel in the operating theatre, who, on the other hand, were not included in data collection and analysis.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants, ICU and ward personnel, and laboratory staff who participated in the trial were "blinded" for assigned treatment throughout the duration of the study. Exception from being blinded was for personnel in the operating theatre, who, on the other hand, were not included in data collection and analysis.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Appropriate power calculations were performed, and all participants were included in the analysis.
Selective reporting (reporting bias)	Low risk	Outcomes clearly prespecified in the protocol; all outcomes reported
Other bias	Low risk	No concerns

Tallman 2002.

Study characteristics
Methods	Single‐centre, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2002 Number of study centres and location: single centre, the Ohio State University College of Medicine, Columbus, Ohio, USA
Participants	Thirty adults undergoing cardiac surgery Mean age: 63 years Sex (female/male ratio): 0.36 Low‐risk: elective CABG or single valve replacement Inclusion criteria: adult cardiac patients less than 80 years old undergoing either CABG or single valve replacement surgery; ASA classes II to IV and with Hb > 10 g/dL, Hct > 30%, Cr < 1.8 mg/dL, WBC 5‐10 × 103/mL, and platelet count 150‐400 × 103/mL Exclusion criteria: other severe medical conditions such as renal, pulmonary, or liver disease, EF less than 35%, concomitant use of aprotinin
Interventions	Intervention: ultrafiltration (ZBUF) Control: CCPB management
Outcomes	IL‐1, IL‐6, TNF‐α, C3a, and C5a, up to 24 hours postoperatively It is the purpose of this study to determine whether ZBUF is capable of removing significant amounts of these inflammatory mediators in adults undergoing open‐heart surgery and, in addition, whether removing these mediators has any effect on the patient’s blood plasma concentrations and whether other signs of clinical morbidity can be recognised as benefiting from ZBUF.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear randomisation process: "This study was designed as a prospective, randomized, blinded study."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Although ICU staff were blinded to the allocation of participants, the theatre staff were aware of the group allocation, and this could have introduced bias. "Also, the ICU staff was [sic] blinded to the patient’s group."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Adequate blinding process: "the laboratory technician who performed the diagnostic work had no knowledge of the patient’s group"
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported properly.
Other bias	Unclear risk	No statement about funding

Tamayo 2009.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2009 Number of study centres and locations: single centre, authors from University of Valladolid, Valladolid, Spain
Participants	In total, 44 adults undergoing cardiac surgery with bypass were assessed in this study. Mean age: 68 years Sex (female/male ratio): 0.27 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG Exclusion criteria: renal or hepatic impairment, congestive heart failure, severely impaired LVEF, active inflammatory disease, history of MI < 6 months, preoperative use of steroids
Interventions	Intervention: Simvastatin (20 mg/day) Control: unspecified
Outcomes	Levels of inflammatory cytokines (IL‐6, IL‐8, TNF‐α) measured at surgery and 48 hours after bypass
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomised, but the method is unspecified.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All staff blinded with the exception of perfusionist
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All staff blinded with the exception of perfusionist
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants included in the analysis
Selective reporting (reporting bias)	Low risk	Although no protocol was published, all the outcomes stated in the introduction and methods have been reported on.
Other bias	Unclear risk	No statement on funding sources

Tang 2002.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2002 Number of study centres and locations: single centre, authors from Southampton General Hospital, Southampton, United Kingdom
Participants	In total, 40 adults undergoing elective CABG were assessed in this study. Mean age: 64 years Sex (female/male ratio): 0.15 Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, normotensive, stable angina, no usage of nephrotoxic agents Exclusion criteria: age above 80 years, pre‐existing renal disease, serum Cr above 135 μmol/L, LVEF below 40%, diabetes mellitus
Interventions	Intervention: leucodepletion Control: CPB alone
Outcomes	This study aimed to investigate the influence of leucodepletion on differential renal injury in low‐risk patients undergoing coronary revascularisation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were allocated randomly to either group; no mention of the technique used.
Allocation concealment (selection bias)	Unclear risk	Not described, unlikely
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not stated; unlikely that personnel were blinded because of the presence of a different filter in the circuit
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	High risk	Participants were withdrawn from the analysis because of the use of inotropes intraoperatively; this could confound the results as these participants were likely to be more unwell than the rest of the cohort.
Selective reporting (reporting bias)	Low risk	Although no protocol was given, the authors clearly stated their outcomes in the introduction and methods.
Other bias	Low risk	Funding obtained from the RCS – low risk of bias

Tassani 1999.

Study characteristics
Methods	Single‐centre, prospective randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 1999 Number of study centres and locations: single centre, German Heart Center Munich at the Technical University, Munchen, Germany
Participants	A total of 43 adults undergoing CABG Mean age: not specified Sex (female/male ratio): not specified Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, NYHA groups II and III, Higgins score of 1.4 ± 0.15 Exclusion criteria: aged > 75 years, body weight >/< 30% ideal body weight, LVEF < 40%, haemodynamic instability or emergency surgery, additional valvular diseases, complete bundle‐branch block, third‐degree atrioventricular block, renal (Cr level > 1.2 mg/dL) or hepatic failure, Hct < 30%
Interventions	Intervention: zero‐balanced ultrafiltration Control: standard CPB management
Outcomes	IL‐6, IL‐8, IL‐10, and IL‐1ra Blood samples were obtained and haemodynamic measurements were performed at the following time points: before induction of anaesthesia; after induction of anaesthesia, just before skin incision; during ECC, before the start of zero‐balanced ultrafiltration; end of ECC; after MUF, 20 minutes after ECC; 2 hours after ECC; 4 hours after ECC; and 24 hours after ECC.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomised with an unspecified method.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The physicians responsible for postoperative care were blinded in respect to the study group. Because of the nature of the intervention, blinding of theatre staff is unfeasible.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	Power analysis done, no participants lost to follow‐up
Selective reporting (reporting bias)	Unclear risk	No protocol, unclear which prespecified outcomes are being assessed
Other bias	Unclear risk	Not clearly stated, no supporting statement

Tassani 1999 (A).

Study characteristics
Methods	Single‐centre, prospective randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 1999 Number of study centres and location: single centre, German Heart Center Munich at the Technical University, Munchen, Germany
Participants	A total of 52 adults undergoing CABG Mean age: unspecified Sex (female/male ratio): unspecified Low‐risk patients (elective CABG) Inclusion criteria: elective CABG, first operation of the day Exclusion criteria: age > 75 years, body weight >/< 30% ideal body weight, LVEF ≤ 40%, haemodynamic instability or emergency operations, additional valvular diseases, complete bundle‐branch block, third degree atrioventricular block, renal (Cr level > 1.2 mg/dL) or hepatic failure, and Hct < 30%
Interventions	Intervention: 1 g of MP 30 minutes before CPB Control: placebo
Outcomes	Proinflammatory (IL‐6 and IL‐8) and anti‐inflammatory (IL‐10 and IL‐1ra) mediators Blood samples were obtained and haemodynamic measurements were performed at the following time points: before induction of anaesthesia; after induction of anaesthesia, just before skin incision; during ECC before rewarming; end of ECC; 30 minutes after ECC; 2 hours after ECC; 4 hours after ECC; and 24 hours after ECC.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomised with an unclear method.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The physicians responsible for postoperative care were blinded in respect to the study group.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	Power analysis done, no participants lost to follow‐up
Selective reporting (reporting bias)	Unclear risk	No protocol, unclear which prespecified outcomes are being assessed
Other bias	Unclear risk	Not clearly stated

Tassani 2000.

Study characteristics
Methods	Single‐centre, prospective, double‐blinded, placebo‐controlled, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2000 Number of study centres and locations: single centre, Department of Cardiac Surgery, and the Institute of Laboratory Analysis, Deutsches Herzzentrum, München
Participants	Twenty patients (NYHA II to III) with stable angina and otherwise healthy scheduled for elective CABG surgery Exclusion criteria: unspecified
Interventions	Intervention: aprotinin (2 × 106 KIU pre‐CPB, 2 × 106 KIU in prime, 500,000 KIU/h during CPB) Control: placebo
Outcomes	Proinflammatory (IL‐6) and anti‐inflammatory (IL‐1‐receptor antagonist) mediators Blood samples were obtained and haemodynamic measurements were performed at the following time points: before induction of anaesthesia; after induction of anaesthesia, just before skin incision; during CPB before rewarming; end of CPB; 30 minutes after CPB; 2 hours after CPB; 4 hours after CPB; and 24 hours after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomised with an unclear method.
Allocation concealment (selection bias)	Unclear risk	Not described clearly
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study was performed in a double‐blind, placebo‐controlled manner.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors state that the study was double‐blinded, but they do not specify who was blinded and to what.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	No clear protocol and no clear statement of the outcome of the study
Other bias	Unclear risk	Not clearly stated

Thielmann 2006.

Study characteristics
Methods	Single‐centre, prospective, double‐blinded, placebo‐controlled, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2006 Number of study centres and locations: single centre, West‐German Heart Center Essen, University Hospital of Essen, Hufelandstraße 55, 45122 Essen
Participants	A total of 57 adults undergoing CABG after acute STEMI were assessed in this study (admitted with an acute ST‐elevation MI and underwent emergency CABG within 24 hours after the onset of symptoms or the ischaemic event). Mean age: 64.5 years Sex (female/male ratio): 0.3 High‐risk patients (recent STEMI) Exclusion criteria: recent (4 weeks) MI; new onset of LBBB; concomitant mechanical MI complications; reoperations; concomitant heart surgery procedures besides CABG; known or suspected hereditary complement deficiency; known or suspected autoimmune disease such as systemic lupus erythematosus; or evidence of serious infection.
Interventions	Intervention: IV bolus of C1‐INH (Berinert®P, ZLB Behring, Marburg, Germany; 40 IU/kg) starting 5 minutes before reperfusion (at aortic unclamping) followed by an IV infusion of C1‐INH (20 IU/kg) 6 hours after surgery (group 1, CABG + STEMI + C1‐INH) Control: placebo
Outcomes	"The present study, therefore, sought to analyse the effects of adjunctive C1‐INH treatment during reperfusion on complement activation, perioperative myocardial cell injury, and clinical outcome in patients undergoing emergency CABG due to acute STEMI."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomly assigned before surgery to receive either an IV bolus of C1‐INH (Berinert®, ZLB Behring, Marburg, Germany; 40 IU kg−1) starting 5 minutes before reperfusion (at aortic unclamping) followed by an IV infusion of C1‐INH (20 IU kg−1) 6 hours after surgery (group 1, CABG + STEMI + C1‐INH) or receiving a placebo (NaCl) bolus followed by placebo infusion (group 2, CABG + STEMI). Insufficient information regarding the randomisation technique
Allocation concealment (selection bias)	Unclear risk	Not clearly stated
Blinding of participants and personnel (performance bias) All outcomes	High risk	Trial was open label.
Blinding of outcome assessment (detection bias) All outcomes	High risk	The trial was not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	The authors clearly stated their outcomes in the introduction and methods; however, not all the variables are then reported (i.e. CRP and D‐dimer).
Other bias	Unclear risk	Not clearly stated

Toft 1997.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: not specified. Submitted January 9, 1997. Accepted for publication May 12, 1997 Number of study centres and location: single centre. University Hospital of Arhus, Skejby Sygehus, Denmark
Participants	Sixteen participants who underwent 'open‐heart surgery'. Cardiovascular‐stable patients, scheduled for open‐heart surgery and in whom no surgical problems were expected, were included in the study. Patients with endocrine disorders were excluded. Mean age: 63.7 years Sex (female/male ratio): 0.14 Low‐risk patients (no surgical problems were expected)
Interventions	Intervention: MP (30 mg/kg intravenously) at the induction of anaesthesia Control: unspecified
Outcomes	L‐selectin (CD62L), H‐CAM (CD44), gp180 LFA‐1 (CDlla), gp165 Mac‐1 (CDllb), and gp95 integrin‐p2 (CD18). Samples were collected 5 minutes after aortic clamping and 5 minutes, 2 hours, and 3 hours after aortic declamping. The leucocyte and differential counts were measured at the same time intervals.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information provided
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information provided
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	No funding reported

Toikkanen 2018.

Study characteristics
Methods	Single‐centre, prospective, double‐blinded, placebo‐controlled, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2018 Number of study centres and locations: single centre, Department of Cardiothoracic Surgery SDSKIR Heart Center Tampere University Hospital, Finland
Participants	Forty elective primary CABG patients Mean age: unspecified Sex (female/male ratio): unspecified Low‐risk patients (elective primary CABG patients) The exclusion criteria were acute coronary syndrome or MI during the previous month, pre‐existent pulmonary disease, pulmonary hypertension with systolic PAP > 40 mmHg, renal insufficiency, previous ischaemic cerebral event, smoking, use of corticosteroids or nonsteroidal anti‐inflammatory drugs or COX‐2 inhibitors, use of angiotensin‐converting‐enzyme inhibitors in diabetics or known hypersensitivity to aprotinin.
Interventions	Intervention: aprotinin (280 mg (2 × 106 KIU of aprotinin immediately after induction of anaesthesia), followed by 280 mg in pump prime, and a constant infusion of 70 mg/h (5 × 105 KIU) until chest closure) Control: constant infusion of 500 mg of TXA just before chest closure
Outcomes	"In the present study, we investigated whether aprotinin administered during CABG impacts lung passage of some key inflammatory factors. With the aid of the PA/RA equation, IL‐6, IL‐8, IL‐10, MPO, and 8‐isoprostane were investigated in patients with CABG treated with a high aprotinin dose and compared with patients with a lower aprotinin regimen." Timing of samples: (T1), 1 min after releasing aortic cross‐clamp; (T2), 15 min after releasing aortic cross‐clamp; (T3), 1 hour after releasing aortic cross‐clamp; (T4), 20 min after releasing aortic cross‐clamp; (T5) 20 h after releasing aortic cross‐clamp
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was made using sealed blinded envelopes as previously described according to the aprotinin dose.
Allocation concealment (selection bias)	Low risk	Sealed envelopes for randomisation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not specified, unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Authors do not describe whether any participants were lost to follow‐up.
Selective reporting (reporting bias)	Low risk	Although no protocol is present, the variables analysed are clearly detailed in the introduction and methods and reported.
Other bias	Low risk	Funding received by government and academic institutions – low risk of bias

Torina 2010.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2010 Number of study centres and locations: Heart Surgery Division – Faculdade de Ciências Médicas Universidade Estadual de Campinas, Campinas, Brazil
Participants	A total of 37 adults undergoing CABG were assessed in this study. Mean age: 55.4 years Sex (female/male ratio): 0.12 Low‐risk patients (elective primary CABG patients) Inclusion criteria: CABG surgery, EF higher than 40% which was estimated by cardiac catheterism in anterior right oblique projection, elective surgery, diabetics or nondiabetics Exclusion criteria: presence of any kind of neoplasia, depressed renal function, which was defined as Cr clearance lower than 45 mL/min for males and lower than 40 mL/min for females
Interventions	Intervention: ultrafiltration (ultrafiltration by aspirating the blood from the ascending aorta by cardioplegic cannula with flow of 300 mL/min during 15 minutes. The blood originating from the cardioplegic cannula ran through the haemoconcentrator H500 (Braile Biomédica, São José do Rio Preto, Brazil), where it was heated at 38o C, then haemoconcentrated and finally returned by the venous line to the right atrium) Control: no ultrafiltration
Outcomes	To analyse the effects of ultrafiltration on cardiorespiratory function, requirement for blood transfusion, and bleeding risk of adult patients undergoing elective CABG
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	ITU personnel who were making decisions with regard to transfusion (one of the primary outcomes) were blinded to the group allocation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unclear how many participants were included in the trial and subsequently in the analysis
Selective reporting (reporting bias)	High risk	Some of the outcomes in the investigation plan such as rates of MI were not included in the results section.
Other bias	Low risk	Funding received from nonprofit academic organisation

Torina 2012.

Study characteristics
Methods	Single‐centre, double‐blinded, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2012 Number of study centres and locations: Heart Surgery Division – Faculdade de Ciências Médicas Universidade Estadual de Campinas, Campinas, Brazil
Participants	A total of 37 adults undergoing CABG were assessed in this study. Mean age: 55 years Sex (female/male ratio): 0.15 Low‐risk patients (elective primary CABG patients) Inclusion criteria were age between 30 years and 70 years, LVEF > 39%, and GFR (calculated by the Cockcroft‐Gault equation) > 45 mL/min/1.73 m2 in men or > 40 mL/min/1.73 m2 in women. Exclusion criteria were indications for mitral valve repair or other surgical procedures in addition to the planned CABG.
Interventions	Intervention: MUF "was performed in heparinized patients between the arterial and venous tubing of the CPB circuit. An H‐500 filter (polyethersulfone) was used, which had an effective membrane area of 0.5 m2, pore size of 5 nm, prime volume of 34 mL, maximum transmembrane pressure of 400 mm Hg, internal fiber diameter of 200 μm, and fiber wall thickness of 30 μm (Braile Biomedica, São José do Rio Preto, Brazil). The blood flow rate through the filter (300 mL/min) was maintained by a roller pump on the inlet part of the filter. Suction was applied to the filtrate port to achieve a negative pressure of 100 mm Hg. The process was carried out for 15 minutes in all patients who underwent MUF." Control: patients assigned to the control group were instead observed for 15 minutes.
Outcomes	"In the present study, using a randomised controlled trial design, we assessed the safety of MUF and its associated inflammatory modulation through tissue markers of inflammatory activity in adult patients undergoing elective coronary artery bypass grafting (CABG) with CPB support." Timing: first 48 hours after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned to either group.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Protocol states there was double‐blinding, implying participants were blinded. In the methods, it states that the theatre staff were not blinded which is due to the type of intervention.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"A blinded nurse intensivist independently rated the chest tube blood loss within the first 48 hours after surgery"
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the protocol were reported on.
Other bias	Low risk	Funding received from academic state‐funded organisation

Tsang 1996.

Study characteristics
Methods	Single‐centre, double‐blinded, prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 1996 Number of study centres and locations: Cardiothoracic Surgical Unit, University Hospital, Birmingham, United Kingdom
Participants	A total of 20 adults undergoing elective CABG were assessed in this study. Mean age: unspecified Sex (female/male ratio): 0.21 Low‐risk patients (elective primary CABG patients) Exclusion criteria: unspecified
Interventions	Intervention: PTX 400 mg orally, 1 week before the operation Control: placebo, for the same period
Outcomes	Release of inflammatory mediators, endothelial injury, and permeability Time points 2 hours, 6 hours, and 24 hours after termination of CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how participants were randomised
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and staff were blinded to interventions (had identical packaging for placebo/tablet).
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	Two participants were withdrawn from the analysis for the intervention arm (one because of intolerance to the medication and one because of postoperative death). Whilst the first was unavoidable, the latter should have been included in the analysis as it is a potentially severe adverse outcome.
Selective reporting (reporting bias)	Low risk	Although there was no protocol, all the outcomes were detailed in the methods and introduction and reported as prespecified.
Other bias	Unclear risk	No statement of funding resources

Turkoz 2001 (A).

Study characteristics
Methods	Prospective, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Departments of Anesthesiology, Biochemistry, and Cardiovascular Surgery, İnönü University Hospital, Malatya, Turkey
Participants	Thirty patients undergoing elective CABG surgery Mean age: 61.1 years Sex (female/male ratio): 0.17 Low‐risk patients (elective CABG patients) Exclusion criteria: patients undergoing a reoperation, who had an MI within 1 month, were suffering from an uncontrolled systemic disease (diabetes mellitus, hypertension, or renal failure), or were receiving long‐term glucocorticoids were excluded.
Interventions	Intervention: MP 30 mg/kg before CPB Control: no treatment
Outcomes	Plasma levels of proinflammatory cytokines (TNF‐α, IL‐1, IL‐6, and IL‐8) were measured in peripheral arterial blood immediately before the induction of anaesthesia, 5 minutes before CPB, 3 minutes after the start of CPB, 2 minutes after the release of the aortic cross‐clamp, 1 hour after CPB, 6 hours after CPB, and 24 hours after CPB and in CS blood immediately before CPB and 2 minutes after the release of the aortic cross‐clamp.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how patients were randomised
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Other bias	Low risk	No statement on funding sources

Turkoz 2001 (B).

Study characteristics
Methods	Prospective, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Departments of Anesthesiology, Biochemistry, and Cardiovascular Surgery, İnönü University Hospital, Malatya, Turkey
Participants	Thirty patients undergoing elective CABG surgery Mean age: 62 years Sex (female/male ratio): 0.11 Low‐risk patients (elective CABG patients) Exclusion criteria: patients undergoing a reoperation, who had an MI within 1 month, were suffering from an uncontrolled systemic disease (diabetes mellitus, hypertension, or renal failure), or were receiving long‐term glucocorticoids were excluded.
Interventions	Intervention: high‐dose aprotinin Control: no treatment
Outcomes	Plasma levels of proinflammatory cytokines (TNF‐α, IL‐1, IL‐6, and IL‐8) were measured in peripheral arterial blood immediately before the induction of anaesthesia, 5 minutes before CPB, 3 minutes after the start of CPB, 2 minutes after the release of the aortic cross‐clamp, 1 hour after CPB, 6 hours after CPB, and 24 hours after CPB and in CS blood immediately before CPB and 2 minutes after the release of the aortic cross‐clamp.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how patients were randomised
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No statement on funding sources

Türktan 2017.

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: approved by local ethics committee 20 June 2013. Patient recruitment December 2014‐ August 2015. Received: 27 December 2016; accepted/published online: 10 May 2017 Number of centres and study location: single centre. Çukurova University, Adana, Turkey
Participants	Participants were between the age of 18 years and 70 years, scheduled for elective CABG surgery, and had physical status of II–III according to the ASA. The exclusion criteria were emergency surgery, prior cardiac surgery (reoperation), EF < 50%, moderate or severe chronic restrictive or obstructive pulmonary disease, hypoxia (SpO2 < 90% in room air), preoperative mechanical ventilation, IABP, inotropic agent requirement, chronic liver disease, or renal failure (Cr > 1.6 mg/dL), bradycardia (heart rate < 50 beats/min), chronic atrial fibrillation, or elevated levels of cardiac enzyme within 48 hours before surgery.
Interventions	Intervention group: anaesthesia was maintained with 0.5%‍ to 1.5% sevoflurane in a 50%‍/50% oxygen/air mixture and 0.3 µg/kg/min to 0.9 µg/kg/min dexmedetomidine infusion after a 0.5 µg/kg/10‐min loading dose of dexmedetomidine. Control group: anaesthesia was maintained with 0.5%‍ to 1.5% sevoflurane in a 50%‍/50% oxygen/air mixture and 0.125‍ µg/kg/min to 0.25 µg/kg/min remifentanil infusion after a 1 µg/kg/10‐min loading dose of remifentanil.
Outcomes	Aortic cross‐clamping time, CPB time, extubation time, and length of stay in the ICU and hospital
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation
Allocation concealment (selection bias)	Unclear risk	Insufficient information
Blinding of participants and personnel (performance bias) All outcomes	Low risk	An independent researcher administered the infusions.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Clinical trial (NCT02405689) registration
Other bias	Low risk	No funding was disclosed.

Ueki 2014.

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2014 Number of study centres and locations: Department of Anaesthesiology, University of Occupational and Environmental Health, Kitakyushu, Japan
Participants	Thirty‐seven patients, > 18 years of age, undergoing elective cardiac surgery with CPB Mean age: 70 years Sex (female/male ratio): 1.31 Exclusion criteria were previous heart surgery, ongoing infection, and chronic administration of anti‐inflammatory drugs.
Interventions	Intervention: dexmedetomidine 1 μg/kg for 10 minutes after aortic cross‐clamping, and 0.5 μg/kg/h intraoperatively Control group received the same volume of saline.
Outcomes	Death, MI, stroke, IL‐6 Blood samples for AST, WCC, CRP, ALT, LDH, blood urine nitrogen, Cr, CK, and CK‐MB were taken before operation, immediately after operation, and on POD 1, POD 2, and POD 3.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sealed envelope randomisation
Allocation concealment (selection bias)	Unclear risk	Attending anaesthetists were not blinded (but did not participate in any study procedures other than preparation/administration of study medication).
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Surgeons, recovery staff, ward nurses, and participants were blinded to group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators responsible for assessing participant data and postoperative outcomes were fully blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Four participants in the intervention group did not complete the study protocol versus one in the control group.
Selective reporting (reporting bias)	Low risk	Adequate choice and reporting of outcomes
Other bias	Low risk	Supported by Grant‐in‐Aid for Scientific Research (C‐22591755 to TK) from the Japan Society for the Promotion of Science

Ueyama 2004 (A).

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: Department of Cardiovascular Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan
Participants	Thirty patients aged 42 years to 74 years and undergoing elective cardiac surgery with CPB were included. Mean age: 60.5 years Sex (female/male ratio): 0.52 Exclusion criteria: age > 75 years; emergency surgery; EF < 30%; heparin treatment at the time of surgery; disseminated intravascular coagulation; severe pulmonary, renal, hepatic, or cerebrovascular disease or neoplasia; preoperative treatment with steroids; and infectious disease
Interventions	Intervention group: PMEA‐coated oxygenator and circuit (n = 10) Control group: non–heparin‐coated oxygenator and circuit (n = 5)
Outcomes	Measurement of cytokine‐6 (IL‐6), BK, and C3a at the end of CPB and 1 hour, 3 hours, and 24 hours after the termination of CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unspecified method of randomisation. Patients were randomly divided into three groups: group N, non–heparin‐coated oxygenator and circuit (n = 10); group H, heparin‐coated oxygenator and circuit (n = 10); and group X, PMEA‐coated oxygenator and circuit (n = 10).
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no published protocol, the authors clearly state their outcomes and report on those in the results.
Other bias	Unclear risk	No clear statement on funding

Ueyama 2004 (B).

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: Department of Cardiovascular Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan
Participants	Thirty patients aged 42 years to 74 years and undergoing elective cardiac surgery with CPB were included. Mean age: 60.5 years Sex (female/male ratio): 0.52 Exclusion criteria: age > 75 years; emergency surgery; EF < 30%; heparin treatment at the time of surgery; disseminated intravascular coagulation; severe pulmonary, renal, hepatic, or cerebrovascular disease or neoplasia; preoperative treatment with steroids; and infectious disease
Interventions	Intervention group: PMEA‐coated oxygenator and circuit (n = 10) Control group: non–heparin‐coated oxygenator and circuit (n = 5)
Outcomes	Measurement of cytokine‐6 (IL‐6), BK, and C3a at the end of CPB and 1 hour, 3 hours, and 24 hours after the termination of CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unspecified method of randomisation. Participants were randomly divided into three groups: group N, non–heparin‐coated oxygenator and circuit (n = 10); group H, heparin‐coated oxygenator and circuit (n = 10); and group X, PMEA‐coated oxygenator and circuit (n = 10).
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the analysis.
Selective reporting (reporting bias)	Low risk	Although no published protocol, the authors clearly state their outcomes and report on those in the results.
Other bias	Unclear risk	No clear statement on funding

Van Den Goor 2004 (A).

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: Department of Cardio‐thoracic Surgery, Academic Medical Center of the University of Amsterdam, Amsterdam, the Netherlands
Participants	Thirty‐seven patients who underwent elective CABG assisted by CBP Mean age: 61 years Sex (female/male ratio): 0.04 Inclusion criteria: age 21 years to 75 years, elective CAB surgery, EF < 30%, body surface area < 1.66 m2, and preoperative Hb levels < 7.5 mmol/L Exclusion criteria: combined valve surgery or aneurysmectomy, re‐do operations, insulin‐dependent diabetes mellitus, Cr l < 300 mmol/L, preoperative intra‐aortic balloon pumping, preoperative use of nonsteroidal anti‐inflammatory drugs, preoperative use of warfarin, preoperative immunosuppressive therapy > 24 h, allergic reactions, and COPD
Interventions	Intervention: oxygenator coated with bio‐passive surface (Trillium; Medtronic: BPS; n = 25) Control: oxygenator was uncoated (n = 12).
Outcomes	TCCs, elastase, sPLA2, CRP, and clinical outcomes Blood samples were collected before bypass after induction of anaesthesia, 15 minutes after start of CPB, at the end of CPB, 30 minutes after administration of protamine sulphate, and on the first and second PODs.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of random sequence generation process
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment measures
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were accounted for in the final analysis.
Selective reporting (reporting bias)	High risk	No protocol with prespecified outcomes, and all the outcomes were not prespecified in the methodology section.
Other bias	High risk	Funded by Medtronic, Minneapolis, MN

Van Den Goor 2004 (B).

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: Department of Cardio‐thoracic Surgery, Academic Medical Center of the University of Amsterdam, Amsterdam, the Netherlands
Participants	Thirty‐four patients who underwent elective CABG assisted by CBP Mean age: 62 years Sex (female/male ratio): 0.02 Inclusion criteria: age 21 years to 75 years, elective CAB surgery, EF < 30%, body surface area < 1.66 m2, and preoperative Hb levels < 7.5 mmol/L Exclusion criteria: combined valve surgery or aneurysmectomy, re‐do operations, insulin‐dependent diabetes mellitus, Cr plasma level < 300 mmol/L, preoperative intra‐aortic balloon pumping, preoperative use of nonsteroidal anti‐inflammatory drugs, preoperative use of warfarin, preoperative immunosuppressive therapy > 24 h, allergic reactions, and COPD
Interventions	Intervention: Carmeda coated oxygenator (n = 21) Control: oxygenator was uncoated (n = 13).
Outcomes	TCCs, elastase, sPLA2, CRP, and clinical outcomes Blood samples were collected before bypass after induction of anaesthesia, 15 minutes after start of CPB, at the end of CPB, 30 minutes after administration of protamine sulphate, and on the first and second PODs.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation process was not clearly described.
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were accounted for in the final analysis.
Selective reporting (reporting bias)	High risk	No protocol with prespecified outcomes, and all the outcomes were not prespecified in the methodology section.
Other bias	High risk	Funded by Medtronic, Minneapolis, MN

Vanden Eynden 2008.

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: between July 2001 and October 2003 Registration date: unspecified, study published in 2004 Number of study centres and locations: Montreal Heart Institute, Montreal, Quebec, Canada
Participants	A total of 200 adults undergoing valve surgery or valve surgery and CABG Mean age: 65 years Sex (female/male ratio): 0.45 Inclusion criteria: patient planned for the valve surgery including AVR, mitral valve replacement, aortic or mitral valve repair, double or triple valve surgery, and valve surgery with concomitant CABG Exclusion criteria: pre‐existing coagulopathies, reoperation, and emergency operations where informed consent was unobtainable
Interventions	Intervention group (n = 99) was exposed to the Avecor oxygenator with TrilliumTM coating. Control group (n = 101) Monolyth (Sorin, Irvine, CA, USA) oxygenator without coating
Outcomes	Blood loss, transfusion of blood products, intubation time, ICU stay and hospital stay, MI, stroke, and death "Partial thromboplastin time (PTT), prothrombin time (PT), and thrombin time (TT), fibrinogen, haemoglobin, and haematocrit were measured before surgery, at five and 24 hours post‐surgery. Troponin‐T and creatinine kinase MB (CK‐MB) were measured at 12 and 24 hours post‐surgery. The platelet count was measured before surgery, at five hours, 12 hours, and 24 hours post‐surgery. The inflammatory factors interleukin (IL)‐ 8, C3a, plasminogen activator inhibitor 1 (PAI‐1), thrombin antithrombin (TAT), prothrombin fragment (PF1‐2), D‐dimer, the von Willebrand factor (vWF), and plasma haemoglobin were measured preoperatively and five hours postoperatively."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation was done with random numbers given to the perfusionist immediately before the surgery.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described, but unlikely that operating theatre staff could have been blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no protocol was given, all the outcomes are clearly stated in the introduction and methods and were reported.
Other bias	High risk	Funding received from manufacturer

Vento 2003.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2003 Number of study centres and locations: Department of Thoracic and Cardiovascular Surgery, Helsinki University Central Hospital, Helsinki, Finland
Participants	Thirty‐five male patients undergoing CABG Mean age: 61 years Sex (female/male ratio): all male patients Inclusion criteria: CABG Exclusion criteria: unspecified
Interventions	Intervention: NAC in a 0.04 mol/L solution Control: standard cardioplegia
Outcomes	Haemodynamic parameters monitored with invasive monitoring up to 24 hours post‐CBP Biochemical parameters: MPO, superoxide dismutase, ubiquinone, LDL oxidation products, conjugated dienes (LDL‐BDC), LDL‐TRAP
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation process not clearly described
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	No clear protocol. The methods appear to report prespecified outcomes for this study; however, in the results there are many more parameters reported than prespecified in the methods. For instance, CRP and CK‐MB are reported for time points up to 24 hours postprocedure which was not stipulated in the methods.
Other bias	Low risk	Funded by a grant from Helsinki Central Hospital – at low risk of having introduced bias

Verrier 2004.

Study characteristics
Methods	International, multi‐centre, randomised, double‐blind, parallel‐group, placebo‐controlled controlled trial 205 hospitals in North America and Western Europe from January 2002 to February 2003
Participants	3099 patients (over 18 years) undergoing CABG surgery Mean age: 64 years Sex (female/male ratio): 0.25 Inclusion criteria: "at least 18 years and had to have met at least 1 of the following baseline risk factors: require urgent intervention defined according to the American College of Cardiology–American Heart Association (ACC/AHA) guidelines as being patients who are required to be staying the hospital due to medical factors but may be scheduled and operated on within a normal scheduling routine; have been diagnosed as having diabetes mellitus; be a woman; have undergone prior CABG procedure; have a history of a neurological event (cerebrovascular accident, transient ischaemic attack, or carotid endarterectomy), congestive heart failure (New York Heart Association class III or IV), at least 2 MIs (excluding patients who have had an MI within 48 hours of undergoing CABG surgery) or experiencing an MI in not less than 48 hours but no more than 4 weeks before CABG surgery" Exclusion criteria: "scheduled to undergo planned aortic dissection repair and/or aortic root reconstruction; required salvage intervention; had current cardiogenic shock; had acute left ventricular, septal, or acute papillary muscle rupture; had uncontrolled diabetes (plasma blood glucose value > 400 mg/dL) within 3 days before surgery; had a history of renal failure and a serum creatinine value greater than 3.0 mg/dL (265.2 µmol/L), of chronic hepatic failure and/or hepatic cirrhosis, and of malignancy, excepting basal cell carcinoma and malignancies in remission; had known or suspected hereditary complement deficiency, any active infection that was clinically significant in the opinion of the investigator, participated in another investigational drug study, or was exposed to another investigational agent within 30 days; and had a known or suspected pregnancy, was breastfeeding, or intended to become pregnant during the study"
Interventions	Intervention: pexelizumab (C5a inhibitor); 2.0 mg/kg bolus followed by 0.05 mg/kg/h infusion of pexelizumab for 24 hours Control: placebo
Outcomes	Incidence of a composite of death or MI within 30 days of randomisation, a peak CK‐MB ≥ 100 ng/mL by day 4 (defined as non‐Q wave if no new evidence of Q wave existed); Q wave evidence of MI, along with CK‐MB ≥ 70 ng/mL by day 4; new Q wave evidence of MI by day 30 that was not present by day 4; or MI (Q wave or non‐Q wave) as identified by the investigator and confirmed by the Clinical Events Committee by day 30. CK‐MB measurements were collected at 4 hours, 8 hours, 12 hours, 16 hours, 24 hours, 72 hours, and 96 hours postoperatively and were analysed at a central core laboratory.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned in a double‐blind fashion by a central telephone‐based interactive voice randomization system to receive either intravenous pexelizumab (2.0 mg/kg bolus followed by 0.05 mg/kg per hour infusion of pexelizumab for 24 hours) or placebo (placebo bolus followed by 24‐hour placebo infusion). Stratification occurred within each site and was based on whether valve surgery was planned, the type of valve surgery (whether mitral or other valve), and whether they had previously undergone CABG surgery."
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Although the protocol states the study was double‐blinded, they do not describe how this was achieved.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Although the protocol states the study was double‐blinded, they do not describe how this was achieved.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Around 10 patients out of 1300 participants were lost to follow up in each group – this is a small amount and would not have significantly affected the results. Although the authors state that the analysis was conducted with intention of including patients who had CABG + valve surgery, the analysis for the prespecified primary endpoint is then reported only for patients undergoing CABG. In view of the fact that this is compliant with the initial protocol and within the parameters of the power calculations, this study is at low risk of bias.
Selective reporting (reporting bias)	Low risk	Prespecified outcomes reported
Other bias	High risk	Funded by two pharmaceutical companies

Visser 2005.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2005 Number of study centres and locations: Cardiothoracic Surgery, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands
Participants	34 patients with normal LVEF scheduled for CABG Mean age: 64 years Sex (female/male ratio): 0.25 Low‐risk patients (elective CABG) Exclusion criteria: diabetes mellitus, EF < 45%, unstable angina pectoris or atrioventricular conduction defects, corticosteroids or nonsteroidal anti‐inflammatory drugs, or undergoing additional surgical procedures
Interventions	Intervention: hyperinsulinaemic normoglycaemic clamping was started and continued throughout the period of CPB until 24 hours after release of the aortic cross‐clamp, buffered with concomitant glucose infusion. Control: no additional infusion
Outcomes	Markers of systemic inflammation Plasma concentrations of IL‐6, IL‐8, and IL‐10, CRP, and SAA were measured at baseline and at time points 4 to 10. Cytokines were measured at baseline and at 2 hours after reperfusion, on the DOS between 6 and 8 p.m., and on POD 1 between 6 and 8 a.m.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"A locally designed randomization programme was used to allocate patients to a control group or a GIK group." – unclear how this programme achieved randomisation
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	The authors report that 15 participants were withdrawn from the final analysis: 13 because of the presence of diabetes, unstable angina, insufficient insulin therapy during CBP, or chronic corticosteroid use and 2 because of acute complications (ischaemic pulmonary oedema resulting in death and MI). As the initial sample size was 34, this is a significant proportion of it (although no power calculations were actually made to assess the number of participants needed to detect an effect). Furthermore, although withdrawing 2 participants because of adverse events could be seen as acceptable, the 13 who were withdrawn because of the presence of exclusion criteria after randomisation highlight poor conduct of the study.
Selective reporting (reporting bias)	Low risk	Most of the parameters were reported properly. However, for the parameters of SIRS, the authors only mention neutrophils without detailing the other parameters.
Other bias	Low risk	The study was funded by a government grant with no association with the study.

Volk 2001.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Clinic for Cardiac Surgery (WK), University Hospital Charité, Campus Mitte, Humboldt‐University, Berlin, Germany
Participants	Thirty‐nine patients scheduled for conventional coronary surgery with three‐vessel disease Mean age: 62 years Sex (female/male ratio): 0.11 Low‐risk patients (elective CABG patients) Inclusion criteria: patients with three‐vessel disease older than 18 years with stable angina pectoris, LVEF ≥ 0.4, LVEDP < 17 mm Hg, absence of pre‐existing pulmonary diseases (determined by clinical examination, chest radiography, lung function tests, and blood gas analyses), absence of insulin‐dependent diabetes mellitus, and clinically relevant renal, hepatic, or cerebrovascular disease Exclusion criteria: preoperative signs of infection (such as WBC count > 12,000/μL, body temperature > 38 °C, CRP > 5 mg/dL), chronic inflammatory diseases, treatment with either cyclooxygenase inhibitors, steroids, or lazaroids within 7 days before the operation, emergency surgical intervention
Interventions	Intervention: preoperative (within 1.5 hours before ECC) application of MP (15 mg/kg) Control: placebo
Outcomes	IL‐6, IL‐8, MCP‐1, CRP, TNF‐α, IFN‐γ, IL‐12, ICU stay, and LOS Time points: after induction of general anaesthesia (time point 1), 1 hour (time point 2) and 4 hours after cross‐clamp release (time point 3), and on the first postoperative morning (time point 4)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The authors state that the study is double‐blind; however, they do not state how this was achieved.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors state that the study is double‐blind; however, they do not state how this was achieved.
Incomplete outcome data (attrition bias) All outcomes	High risk	One participant in the MP group accidentally was treated with both corticosteroids and aminosteroids and therefore had to be excluded from data analysis; as this represents almost 10% of the sample size, it could have affected the results.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the outcomes were clearly detailed in the methods and reported.
Other bias	Low risk	No funding disclosed

Vukovic 2011 (A).

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: between February 2008 and May 2009 Registration date: unspecified, study published in 2011 Number of study centres and locations: single centre. Department of Anesthesia, Dedinje Cardiovascular Institute, School of Medicine, University of Belgrade, Serbia
Participants	Ninety patients undergoing elective coronary surgery (43 in part A). All patients were found to have significantly impaired LV function (EF ≤ 30%). Mean age: 61 years Sex (female/male ratio): 0.16 Low‐risk patients (elective CABG patients) Exclusion criteria: patients with AMI (< 4 weeks), acute and chronic infections, autoimmune disease or preceding anti‐inflammatory therapy, severe renal dysfunction demanding preoperative dialysis, hepatic dysfunction and previous or concomitant cardiac surgical procedures
Interventions	Intervention: MP group (MP), a single IV bolus of MP (10 mg/kg) administered after induction of anaesthesia Control: standard treatment
Outcomes	IL‐6, postoperative cardiac index, LV stroke work index, postoperative AF rate, ICU stay, inotropic support, tracheal intubation time Blood samples were collected at the end of the surgery and at 4 hours and 24 hours after surgery, as well as 72 hours after the operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was achieved by the use of computer‐generated random numbers.
Allocation concealment (selection bias)	High risk	Participants were blinded to the drug assignment group. Caregivers were aware of the drug used, but were not involved in data collection and interpretation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were blinded to the drug assignment group, but medical personnel were aware of the drug used.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors state that medical personnel were aware of allocation but not involved in data collection. However, they do not state clearly whether the outcome assessors were blinded to the intervention.
Incomplete outcome data (attrition bias) All outcomes	High risk	The authors do not state anything about participants lost to follow‐up, but in the tables there are 29, 28, and 28 participants in each group, respectively, whereas in the methods the authors state that they randomised 90 participants in equal groups. While losing 5 participants out of 90 to follow up is not hugely concerning, the fact that this is not acknowledged is concerning.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors.

Vukovic 2011 (B).

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: between February 2008 and May 2009 Registration date: unspecified, study published in 2011 Number of study centres and locations: single centre. Department of Anesthesia, Dedinje Cardiovascular Institute, School of Medicine, University of Belgrade, Serbia
Participants	Ninety patients undergoing elective coronary surgery (43 in part B). All patients were found to have significantly impaired LV function (EF ≤ 30%). Mean age: 61 years Sex (female/male ratio): 0.16 Low‐risk patients (elective CABG patients) Exclusion criteria: patients with AMI (< 4 weeks), acute and chronic infections, autoimmune disease or preceding anti‐inflammatory therapy, severe renal dysfunction demanding preoperative dialysis, hepatic dysfunction, and previous or concomitant cardiac surgical procedures
Interventions	Intervention: Atorvastatin 20 mg once a day for 3 weeks prior to the operation Control: standard treatment
Outcomes	IL‐6, postoperative cardiac index, LV stroke work index, postoperative atrial fibrillation rate, ICU stay, inotropic support, tracheal intubation time Blood samples were collected at the end of the surgery and at 4 hours and 24 hours after surgery, as well as 72 hours after the operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was achieved by the use of computer‐generated random numbers.
Allocation concealment (selection bias)	High risk	Participants were blinded to the drug assignment group. Caregivers were aware of the drug used, but were not involved in data collection and interpretation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were blinded to the drug assignment group. Caregivers were aware of the drug used, but were not involved in data collection and interpretation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The authors state that medical personnel were not blinded to the intervention but that they were not involved in the data collection and analysis. It is unclear whether the outcome assessors were blinded or not.
Incomplete outcome data (attrition bias) All outcomes	High risk	The authors do not state anything about participants lost to follow‐up, but in the tables there are 29, 28, and 28 participants in each group, respectively, whereas in the methods the authors state that they randomised 90 participants in equal groups. While losing 5 participants out of 90 to follow‐up is not hugely concerning, the fact that this is not acknowledged is concerning.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors.

Wan 1999.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: 2‐month period, unclear dates Registration date: unspecified, study published in 1999 Number of study centres and locations: not specified. Authors from Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong, People's Republic of China and from University Hospital Erasme, Free University of Brussels, Brussels, Belgium.
Participants	Forty‐four participants assessed in total. Twenty adult patients undergoing CABG Mean age: 64 years Sex (female/male ratio): 0.3 Low‐risk patients (elective CABG patients) Exclusion criteria: patients undergoing a potentially short duration of CPB such as for CABG of fewer than 3 grafts, those undergoing a redo or an emergency procedure, those who had infectious disease before the operation, and those who had used steroids in the preoperative period
Interventions	Intervention: 10 mg/kg dose of MP (Solu‐Medrol; Pharmacia & Upjohn, Kalamazoo, MI) during induction of anaesthesia Control: placebo
Outcomes	Postoperative serum CK‐MB and lactate values were measured every 12 hours within 24 hours. Postoperative blood losses, duration of mechanical ventilation, and the length of ICU stay were also recorded.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The patients were allocated to each group by the opening of a sealed preprepared envelope.
Allocation concealment (selection bias)	Unclear risk	Not described in sufficient detail
Blinding of participants and personnel (performance bias) All outcomes	Low risk	It is unclear whether participants and theatre staff were blinded to the allocation, but ITU staff were.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	The authors do not show all the data for lactate and CK‐MB which were serially measured. Only a peak value is reported, and it is unclear when that was measured.
Other bias	Low risk	This study was supported by the Fondation pour la Chirurgie Cardiaque, Belgium; at low risk of bias introduced by funding.

Wan 2004.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: 2‐month period, unclear dates Registration date: unspecified, study published in 1999 Number of study centres and locations: single centre. Division of Cardiothoracic Surgery, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong, People's Republic of China
Participants	Thirty‐seven adult patients undergoing CABG Mean age: 65 years Sex (female/male ratio): 0.3 Low‐risk patients (elective CABG patients) Exclusion criteria: recent MI, unstable angina, on steroid treatment, age > 75, serum Cr > 200 μmol/L, single‐vessel disease, previous open cardiac procedures, and poor LV function (EF < 30%)
Interventions	Intervention: off‐pump beating CPB Control: on‐pump CPB, performed by the same surgeon
Outcomes	Inflammatory marker (IL‐6, IL‐8, and IL‐10, TNF‐α, vascular cell adhesion molecule‐1) and clinical outcomes (postoperative mortality or major complications). Samples collected on skin closure and 4 hours, 12 hours, and 24 hours after the operations
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A randomisation table was generated with a computer.
Allocation concealment (selection bias)	Low risk	Allocation was performed by sealed opaque envelopes. Both the participants and the surgeons were unaware of the result of randomisation until the patient was put under general anaesthesia and the sealed envelope was opened.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants blinded; no description of blinding personnel, but unlikely because of nature of intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	No funding received

Wang 2019.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2019 Number of study centres and locations: single centre. Department of Anesthesiology, Zhongshan Hospital, Fudan University, Shanghai, China
Participants	Sixty‐five adult patients undergoing OPCAB Mean age: 65 years Sex (female/male ratio): 0.3 Low‐risk patients (elective CABG patients) Exclusion criteria: age > 80 years; major combined surgery (such as valve surgery); MI in the last 28 days; severe infection in the last 7 days; severe hepatic, renal, pulmonary or haematological disease; use of an inotropic agent or a mechanical assist device; LVEF < 40%; or peripheral vascular disease affecting upper limbs
Interventions	Intervention: RIPC, induced with repeated 5‐minute ischaemia and 5‐minute reperfusion on the upper limb for four times; by using a blood pressure cuff inflation, patients in the RIPC group were exposed to a pressure 40 mmHg higher than the systolic arterial pressure. RIPC procedure was performed right after the end of anaesthetic induction. Control: sham placement of the pressure cuff without inflation
Outcomes	IL‐6, IL‐8, IL‐10, TNF‐α, cTnT, HFABP, HIF1‐a, IMA, and MDA. Blood samples were collected from the right internal jugular vein preoperatively and after the surgery (0 hours, 6 hours, 18 hours, 24 hours, 48 hours, 72 hours, and 120 hours).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned to control or RIPC group using a computer‐generated random list after they entered the operating room. Both surgeons and anaesthesiologists were blinded to the assignments.
Allocation concealment (selection bias)	Low risk	Participants were randomly assigned to control or RIPC group using a computer‐generated random list after they entered the operating room. Both surgeons and anaesthesiologists were blinded to the assignments.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Both surgeons and anaesthesiologists were blinded to the assignments.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	The registered protocol only registers ICU length of stay as their primary outcome. However, the methods section does include clearly other variables which are all reported clearly.
Other bias	Low risk	No funding disclosed

Wang 2020.

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: from September 2017 to September 2018 Registration date: unspecified, study published in 2020 Number of study centres and locations: single centre. Department of Anesthesiology, The First Affiliated Hospital of Hebei North University, Hebei, China
Participants	Sixty patients with elective operation of CPB heart surgery Mean age: 44 years Sex (female/male ratio): 0.7 Low‐risk patients (elective CABG patients) Inclusion criteria: first‐time cardiac surgery Exclusion criteria were patients with recent surgical treatment, patients with severe respiratory diseases, patients with liver and kidney dysfunction, and patients with coagulation dysfunction and other surgical contraindications.
Interventions	Intervention: dexmedetomidine 1 μg/kg was added into the membrane lung priming solution in the Dex group and was continuously infused at the rate of 0.5 μg/kg/h until the end of the operation. Control: normal saline
Outcomes	cTnI, MDA, IL‐6, and TNF‐α, cognitive function, occurrence of arrhythmia; time points before anaesthesia induction, before CPB, during CPB, immediately after CPB, and 12 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation
Allocation concealment (selection bias)	High risk	Insufficient information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	No prespecified protocol was registered.
Other bias	Low risk	No funding was disclosed.

Wang 2022.

Study characteristics
Methods	Single‐centre, randomised controlled study Run‐in period: from March 2020 to February 2021 Registration date: unspecified, study published in 2022 Number of study centres and locations: single centre. Department of Cardiovascular Surgery, Yijishan Hospital, The First Affiliated Hospital of Wannan Medical College, Wuhu City, China
Participants	Adults undergoing valve replacement surgery Mean age: 54 years Sex (female/male ratio): 0.67 Inclusion criteria: "patients diagnosed with congenital heart valve disease, degenerative disease, or rheumatic valvular disease; patients who underwent cardiac mitral valve replacement or cardiac tricuspid valve replacement under valve cardiopulmonary bypass; patients with New York Heart Association (NYHA) functional classification grade III‐IV" Exclusion criteria: "patients with drug allergy, patients with a history of other cardiac surgery, patients with multiple cardiovascular or cerebrovascular diseases, and patients with organ or immune dysfunction"
Interventions	Intervention: UST Control: normal saline
Outcomes	CPB status; quality of life; cardiac function (CK‐MB and troponin levels); inflammation (TNF‐α and IL‐10 levels); oxidative stress (MDA and SOD levels); haemodynamics Venous blood was sampled from each patient before surgery, at 8 hours after surgery, and at 24 hours after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described. In view of the nature of the intervention, it would be difficult to blind participants or the personnel administering the preconditioning.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No PRISMA diagram available, and the authors do not disclose how many patients were screened/how many met inclusion criteria, etc.
Selective reporting (reporting bias)	Unclear risk	No prespecified protocol was registered.
Other bias	Low risk	No conflict of interests declared

Weerwind 1995.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: from March 2020 to February 2021 Registration date: unspecified, study published in 2022 Number of study centres and locations: single study. Maastricht, the Netherlands (institution. not specified)
Participants	A total of 40 participants assessed; 14 adult patients undergoing elective CABG Mean age: 59 years Sex (female/male ratio): 0.17 Exclusion criteria: "previous cardiac surgery, congestive cardiac failure, neurologic disorders (e.g. cerebrovascular accident), severe pulmonary disorders (e.g. chronic obstructive pulmonary disease, emphysema), diabetes, renal diseases (e.g. renal failure), liver diseases, and preoperative coagulopathies"
Interventions	Intervention: Duraflo II heparin‐treated circuit Control: uncoated circuit
Outcomes	IL‐6, IL‐8, TNF‐Rs, sE‐Selectin, sICAM; after release of the aortic cross‐clamp and 6 hours and 12 hours after aortic cross‐clamp release
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation technique not clearly described
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described; unlikely because of the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigation were detailed in the method sections and reported.
Other bias	Low risk	No funding was disclosed.

Wehlin 2004.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: between March 2000 and June 2001 Registration date: unspecified, study published in 2004 Number of study centres and locations: single centre. Department of Clinical Immunology, Institution of Medicine, Karolinska Institutet, Stockholm, Sweden
Participants	Thirty‐eight adult patients admitted for elective CABG Mean age: 65 years Sex (female/male ratio): 0.23 Low‐risk (elective CABG) Exclusion criteria: "age < 50 or > 80 years, ejection fraction < 30%, serum creatinine > 150 μmol/L, tight main stem stenosis, re‐do operation, diffuse distal coronary artery disease, unstable angina, and history of cerebrovascular disease"
Interventions	Intervention: OPCAB Control: ONCAB
Outcomes	C5a, TCC, C59‐b, IL‐6, IL‐8; CD11b, CD35, and CD62L Blood samples were taken during and up to 24 hours after the operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation technique was not described.
Allocation concealment (selection bias)	Unclear risk	Insufficient details around allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	High risk	One participant in the OPCAB group was converted to on‐pump surgery because of bleeding from a septal branch during the anastomosis of the left internal mammary artery to the left anterior descending artery and was therefore excluded from the study. Furthermore, the authors report that serum was collected for 32 participants, but it is unclear how many were actually included in the analysis.
Selective reporting (reporting bias)	High risk	The outcomes were not clearly defined in the method section and there is no published protocol; thus, there is a high risk of selective reporting bias.
Other bias	Low risk	No funding was disclosed.

Wei 2001.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Tampere University Hospital, Tampere, Finland
Participants	Twenty‐one patients admitted for first‐time elective CABG Mean age: 64 years Sex (female/male ratio): 0 Low‐risk (elective CABG) Exclusion criteria: patients with unstable angina, poor LV function (EF < 30%) or valve disease, and those on corticosteroid medication
Interventions	Intervention: 2 × 106 KIU (280 mg) aprotinin (Trasylol; Bayer AG, Leverkusen, Germany) was added to the priming solution in the ECC. Control: no addition to the priming solution
Outcomes	IL‐6, IL‐8, IL‐10, CK‐MB, leucocyte count Blood samples were collected from the radial artery before induction and 5 minutes, 1 hour, 4 hours, 8 hours, and 20 hours after reperfusion.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation technique was not described.
Allocation concealment (selection bias)	Unclear risk	No sufficient description
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigation were detailed in the method sections and reported.
Other bias	Low risk	Funding received by a government organisation; thus, at low risk of funding bias

Wei 2001 (A).

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Division of Cardiovascular Surgery, Tampere University Hospital, Tampere, Finland
Participants	Twenty‐two male patients with multivessel coronary disease and stable angina Mean age: 64 years Sex (female/male ratio): 0 Low‐risk (elective CABG) Exclusion criteria: patients with unstable angina, poor LV function (EF < 30%) or valve disease, and those on corticosteroid medication
Interventions	Intervention: ischaemic preconditioning ("after the start of cardiopulmonary bypass, patients randomised to the IP group received two two‐minute periods of ischaemia by cross‐clamping the aorta, each separated by three three‐minute periods of reperfusion") Control: 10 minutes of normothermic CPB before cross‐clamping
Outcomes	TNF‐α, IL‐6, IL‐8 and IL‐10. Blood samples were collected from the radial artery before induction, 1 minute before CPB, and 5 minutes, 1 hour, 4 hours, and 20 hours after reperfusion. Another 2 blood samples from the CS were collected 1 minute before CPB and 5 minutes after reperfusion.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation process was not clearly described.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described (unlikely because of the nature of the intervention)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	High risk	Two participants, one from the ischaemic preconditioning group and one from the control, were excluded because of severe postoperative bleeding requiring re‐exploration which prohibited blood chemistry measurements.
Selective reporting (reporting bias)	High risk	TNF‐α values not shown
Other bias	Low risk	The present study was supported by a grant from the Medical Research Fund of Tampere University Hospital and the CIMO Foundation of Finland.

Wei 2001 (B).

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: between February 2000 and July 2000 Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Division of Cardiovascular Surgery, Tampere University Hospital, Tampere, Finland
Participants	Thirty male patients with multiple‐vessel CAD and stable angina admitted to the hospital for the first time for elective CAB surgery Mean age: 65 years Sex (female/male ratio): 0 Low‐risk (elective CABG) Exclusion criteria: patients with unstable angina, poor LV function (EF < 30%) or valve disease, and those on corticosteroid medication
Interventions	Intervention: adenosine infusion prior to the initiation of CPB ("infusion of Adenoscan (Sanofi; Winthrop, France) prior to initiation of CPB (after complete cannulation of appropriate vessels and before administration of cardioplegic solution) through a Swan‐Ganz catheter to the superior vena cava using a computer‐controlled pump infusion system. The initial infusion rate was a 50‐μg/kg increment to the dosage of 100 μg/kg/min at the second minute; after this, the infusion lasted for 6 min or until the patient developed a systolic arterial pressure < 70 mm Hg.") Control: unspecified
Outcomes	TNF‐α, IL‐6, IL‐8, and IL‐10, white cell count, haemodynamic monitoring; before induction of anaesthesia (baseline) and 5 minutes, 1 hour, 4 hours, 8 hours, and 20 hours after reperfusion to the myocardium
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of the randomisation process
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unclear whether participants and personnel were blinded to the intervention, but it is unlikely because of the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Although laboratory staff were blinded, it does not state whether the rest of the trial staff were blinded when assessing outcomes for these participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	Although all randomised patients were included in the analysis, the exclusion criteria of the study stipulate that if a patient suffered complications needing reoperation or cross‐clamp time of > 120 minutes, they would be excluded from the analysis. This has the potential to introduce significant bias in the results by excluding patients suffering from complications; the results of the trial may be overly positive.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigation were detailed in the method sections and reported.
Other bias	Low risk	The present study was supported by a grant from the Medical Research Fund of Tampere University Hospital, the CIMO Foundation of Finland and the Tampere Tuberculosis Foundation – low risk of funding bias.

Wei 2002.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2002 Number of study centres and locations: single centre. Division of Cardiovascular Surgery, Tampere University Hospital, Tampere, Finland
Participants	Sixty‐four adult patients admitted for first‐time CABG Mean age: 65 years Sex (female/male ratio): all male patients. Low‐risk (elective CABG) Exclusion criteria: patients with unstable angina, poor LV function (EF < 30%) or valve disease, and those on corticosteroid medication
Interventions	Intervention: 280 mg of aprotinin in the pump prime Control: unspecified
Outcomes	CK‐MB release, leucocyte counts, and haemodynamics data were recorded. Perioperative MPO, IL‐6, IL‐8, and IL‐10 levels were measured in a subgroup of participants (15 participants in each group). Blood samples were collected from the radial artery before induction of anaesthesia (baseline) and 5 minutes, 1 hour, 4 hours, and 20 hours after myocardium reperfusion.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation technique was not described.
Allocation concealment (selection bias)	Unclear risk	No description of concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Although laboratory staff was blinded, it does not state whether the rest of the trial staff were blinded when assessing outcomes for these participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	Two participants amongst the controls and one in the aprotinin group needed re‐exploration on account of severe surgical bleeding. These participants were excluded from the statistical analysis of all parameters reported. One participant of control group was excluded from statistical analysis because of CK‐MB levels. Perioperative MPO, IL‐6, IL‐8, and IL‐10 levels were measured in a subgroup of participants (15 participants in each group).
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	Funding received from a government organisation – low risk of bias

Weis 2009.

Study characteristics
Methods	Prospective, randomised, double‐blinded, placebo‐controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2009 Number of study centres and locations: single centre. Departments of Anesthesiology and Cardiac Surgery, University of Munich, Klinikum Grosshadern, Munich, Germany
Participants	Thirty‐six high‐risk patients scheduled for cardiac operations using CPB. High risk was defined as an LVEF < 39% or an expected duration of the CPB > 97 minutes (combined procedures or CABG with > 3 grafts planned). Mean age: 68 years Sex (female/male ratio): 0.71 Exclusion criteria: "age < 18 years, pregnancy, preoperative IL‐6 levels> 10 pg/mL, hepatic insufficiency (bilirubin> 3 mg/dL), renal insufficiency (creatinine >2 mg/dL), a positive serologic test for HIV, manifest insulin‐dependent diabetes mellitus, adipositas permagna (BMI 30), use of steroidal or nonsteroidal antiphlogistics during the last 7 days (except 100 mg acetylsalicyl acid per day), an extracardial septic focus, or chronic or acute inflammatory diseases"
Interventions	Intervention: stress doses of hydrocortisone (loading dose 100 mg intravenously during 10 minutes) before induction of anaesthesia and followed by a continuous infusion of 10 mg/hr for 24 hours (POD 1), which was reduced to 5 mg/hr on POD 2 and then tapered to 3 × 20 mg intravenously on POD 3 and 3 × 10 mg intravenously on POD 4. Group 2 served as a control group and received placebo (0.9% saline solution). Control: placebo (no detail of specific dose)
Outcomes	IL‐6 to IL‐10 ratio and other markers of systemic inflammation at predefined time points (1 hour, 4 hours, 24 hours, and 48 hours after termination of CPB); short‐term clinical outcome
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly allocated to two groups by block randomisation.
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No description of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No description of assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no ITT analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	No funding reported

Westerberg 2004.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2004 Number of study centres and locations: single centre. Cardiothoracic Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden
Participants	Twenty‐nine adult patients undergoing coronary surgery Mean age: 66 years Sex (female/male ratio): 0.14 Exclusion criteria: preoperative use of steroids or nonsteroidal anti‐inflammatory drugs, perioperative MI, postoperative low output heart failure, and reoperation for bleeding
Interventions	Intervention: discarding mediastinal‐shed blood Control: mediastinal blood reinfusion
Outcomes	Plasma concentrations of the cytokines (TNF‐α, IL‐6, and complement factor C3a), CRP, erythrocyte sedimentation rate, troponin T, and Hb levels; measured preoperatively and 10 minutes after CPB discontinuation (before protamine administration), and 2 hours and 24 hours after CPB
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description of randomisation process
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	High risk	Six participants were excluded from the study after surgery, four because of cardiotomy suction blood volume or postoperative bleeding more than 500 mL and 2 because of electrocardiographic signs of perioperative MI.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	The study was supported by the Gothenburg Medical Association.

Whitaker 2006.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: unspecified Registration date: unspecified, study published in 2006 Number of study centres and locations: single centre. Department of Cardiothoracic Surgery at the Middlesex Hospital (University College London Hospitals NHS Trust), London, UK
Participants	Sixty patients scheduled for elective first‐time CABG Mean age: 65 years Sex (female/male ratio): 0.14 Low‐risk elective CABG patients Exclusion criteria: age > 80 years, having re‐do procedures, history of transient ischaemic attacks or strokes, cardiac arrest in the preceding month, carotid bruits, insulin‐dependent diabetes mellitus, preoperative renal failure (defined as Cr > 150 mmol/L) or a preoperative leucocyte count outside the normal range
Interventions	Intervention: leucocyte‐depleting filters Control: standard CPB circuit
Outcomes	Troponin T, neutrophil counts, serum elastase, and ECGs; preoperatively, pre‐bypass, on bypass, off bypass, and 6 h, 24 h, 48 h, and 72 h after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was achieved with sealed envelopes given to the perfusion department, whose staff then set up the bypass circuit using the appropriate filter.
Allocation concealment (selection bias)	Low risk	Allocation was performed by sealed envelopes given to the perfusion department. All investigators were blinded to the participant's assignment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study – only the perfusionist was aware of participant allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The perfusionists kept a record of the filter used for each participant, and all other investigators were blinded to the participant's assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	High risk	The company manufacturing the filters provided the filters at no cost.

Whitlock 2006.

Study characteristics
Methods	Single‐centre, double‐blind, prospective, randomised controlled trial Run‐in period: from 1 April 2004 to 28 February 2005 Registration date: unspecified, study published in 2006 Number of study centres and locations: single centre. Department of Surgery, Division of Cardiovascular Surgery, McMaster University, Hamilton, Canada
Participants	60 patients undergoing CPB Mean age: 67 years Sex (female/male ratio): 0.27 Exclusion criteria: use of systemic steroids, history of bacterial or fungal infection in the previous 30 days, or steroid intolerance
Interventions	Intervention: low‐dose MP (250 mg intravenously twice) Control: placebo
Outcomes	Haemodynamic data, ventilator requirement, arrhythmia, metabolic data, IL‐6, ICU stay, LOS; nine time points (just before skin incision, just before CPB initiation, 10 min on CPB, upon cessation of CPB, ICU admission, 4 h postoperative, 8 h postoperative, 12 h postoperative, and 24 h postoperative)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation by a computer‐generated sequence
Allocation concealment (selection bias)	Low risk	Allocation was performed by pharmacy.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinded trial: all participants, clinicians, and statisticians were blinded until the completion of data analysis by group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All participants, clinicians, and statisticians were blinded until the completion of data analysis by group.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study.
Selective reporting (reporting bias)	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Other bias	Low risk	No funding was disclosed.

Wildhirt 2001.

Study characteristics
Methods	Prospective, randomised controlled study Run‐in period: unspecified Registration date: unspecified, study published in 2001 Number of study centres and locations: single centre. Department of Cardiac Surgery, Ludwig‐Maximilians University, Munich, Germany
Participants	Twenty‐six patients with single or multivessel CAD Mean age: 65 years Sex (female/male ratio): 0.19 Exclusion criteria: impaired renal function or liver function, signs of acute infection, reoperation
Interventions	Intervention: OPCAB Control: conventional ONCAB
Outcomes	TNF‐α, P‐selectin (24 h and 72 h after surgery), and ICAM‐1 (0.5 h, 6 h, and 24 h after surgery)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No clear description of random sequence generation
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding procedures not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	No clear description of prespecified outcomes
Other bias	Low risk	Funding was obtained from nonprofit institutions: "This study was in part supported by Dr Wandler Foundation (Munich, Germany), by the Friedrich Baur Foundation, and by the Walter Schulz Foundation (Munich, Germany). SMW was supported by the German Israeli Foundation (GIF), the Deutcsche Forschungsgemeinschaft, and the Research Support Foundation of the Ludwig‐ Maximilian University (Muncih [sic], Germany)."

Williams 2014.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: between May 2010 and June 2011 Registration date: registered on the Australian New Zealand Clinical Trials Registry (ACTRN 12609000965202), study published in 2001 Number of study centres and locations: single centre. Victoria University of Wellington, Wellington, New Zealand
Participants	Ninety‐six adult high‐risk cardiac surgery patients (defined as double, triple, or mitral valve replacement, CABG with EF < 50%, CABG + valve(s), or any re‐do cardiac operation) Mean age: 65 years Sex (female/male ratio): 0.19 High‐risk Exclusion criteria: patients with peripheral vascular disease affecting the upper limbs or requiring deep hypothermic circulatory arrest or radial artery conduit harvesting
Interventions	Intervention: RIPC (applied beginning with the first surgical incision by inflating the cuff to 200 mmHg for 5 minutes, followed by 5 minutes of deflation. This process was repeated three times) Control: unspecified
Outcomes	Primary outcomes: total inotrope use, worse level of AKI, troponin Secondary outcomes: time to extubation; proportion of patients extubated 12 h postweaning of CPB; IL‐6, IL‐8, IL‐10, and TNF‐α levels (measured preoperatively and 1 h, 2 h, 3 h, 6 h, and 12 h after cross‐clamp removal); volume of fluid infused; duration of vasopressor support
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were permuted‐block randomised in groups of eight by a third party using an online randomisation sequence generator with an allocation ratio of 1:1 to either RIPC or control.
Allocation concealment (selection bias)	Low risk	Treatment group allocation was concealed in sequentially numbered opaque envelopes until an anaesthetic technician applied the intervention.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	For the control group, the same intervention was applied to the tourniquet wrapped around the towel. Participants, theatre staff (with the exception of the anaesthetic technician), ICU staff, and investigators were masked as to treatment allocation until data collection was completed.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants, theatre staff (with the exception of the anaesthetic technician), ICU staff, and investigators were masked as to treatment allocation until data collection was completed.
Incomplete outcome data (attrition bias) All outcomes	Low risk	One participant was lost to follow‐up, although in view of the sample size, this was unlikely to have altered the outcomes or have introduced bias.
Selective reporting (reporting bias)	High risk	Not all the outcomes prespecified in the protocol are reported.
Other bias	Low risk	Funding from the New Zealand Lotteries Commission and the National Heart Foundation of New Zealand – low risk of bias

Wippermann 2005.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: between May 2010 and June 2011 Registration date: registered on the Australian New Zealand Clinical Trials Registry (ACTRN 12609000965202), study published in 2001 Number of study centres and locations: single centre. Department of Cardiothoracic and Vascular Surgery, University Hospital Jena, Jena‐Lobeda, Germany
Participants	In total, 30 participants were assessed in this study. Twenty adult patients were undergoing CABG. Mean age: 64 years Sex (female/male ratio): unspecified Low‐risk CABG Exclusion criteria: unspecified
Interventions	Intervention: CCECC Control: CCPB circuit
Outcomes	IL‐6, fHb, vWf activity, TATc, PF1.2, and PAPc were assessed preoperatively, perioperatively, and 24 h postoperatively.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of the randomisation technique
Allocation concealment (selection bias)	Unclear risk	Not described (unlikely because of the nature of the intervention)
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described (unlikely because of the nature of the intervention)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigations were detailed in the method sections and reported.
Other bias	Low risk	No funding was disclosed.

Xia 2006.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: not specified. Accepted for publication May 24, 2006 Number of study centres and location: not specified. Authors from Renmin Hospital, Wuhan University, Wuhan, China, Shenzhen Sun Yat‐sen Cardiovascular Hospital, China and the University of British Columbia, Vancouver, Canada
Participants	Fifty‐four ASA class II to III patients, aged 55 years to 73 years, presenting for scheduled CABG surgery Exclusion criteria: a preoperative history of liver or kidney dysfunction; recent intake of vitamin C or E; age < 18 years or > 80 years; and a history of an AMI or evolving MI.
Interventions	Intervention: high‐dose propofol Control: isoflurane
Outcomes	cTnI, cTnT, cardiac‐specific creatine kinase (CK‐MB), and plasma levels of MDA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "54 ASA class II to III patients, aged 55–73 yr, presenting for scheduled CABG surgery were assigned (according to randomization envelopes) to the small‐dose propofol (Group P; n = 18), large‐dose propofol (Group HiP; n = 18), or isoflurane (Group I; n = 18) groups". The randomisation scheme provided an equal number of participants from each study group for two surgical teams.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment procedures were not clearly described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Surgeons working in the operation room and intensive care unit (ICU) were blinded to treatment protocols, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "Surgeons working in the operation room and intensive care unit (ICU) were blinded to treatment protocols, facilitated by covering the drug infusion pump and lines and shielding the isoflurane vaporizer from view". This does not clearly detail whether outcome assessors were blinded or not.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	The protocol was not published prospectively; there are some outcomes detailed in the methodology section, but there are a lot more outcomes reported in the results.
Other bias	Low risk	Quote: "Supported, in part, by the National Natural Science Foundation of China (NSFC) (No. 30471659, to Z. Xia)". Funding from nonprofit organisation

Xu 2013.

Study characteristics
Methods	Prospective, double‐blinded, randomised controlled study Run‐in period: January 2007 to December 2009 Number of study centre and locations: single centre. Provincial Hospital Affiliated to Shandong University, Jinan, China
Participants	Thirty‐six patients with acute type‐A aortic dissection undergoing cardiac surgery using CPB under deep hypothermic circulatory arrest
Interventions	Intervention: high‐dose UST Control: 0.9% sodium chloride
Outcomes	TNF‐α, IL‐6, IL‐8, and PMN‐Elastase were measured after anaesthetic induction (T0), 30 min (T1) after aortic cross‐clamp, 3 (T2), and 6 h (T3) and 9 h (T4) after weaning from CPB. Pulmonary data were also recorded.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "According to a computer‐generated sequence, patients were prospectively randomized into 2 groups: the ulinastatin group (U group, n = 18) and the control group (C group, n=18)."
Allocation concealment (selection bias)	Unclear risk	Allocation concealment procedures were not described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Comment: "Blinded solution syringes were prepared by a well‐trained surgical nurse, and administered by the anesthesiologist. None of the nursing staff and physicians caring for these patients were aware of the study protocol. The surgeries on all patients were carried out by the same surgical team, which included the anesthesiologist and perfusionist."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clearly described whether outcome assessors were blinded to intervention.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	Quote: "This work is supported by the Shandong Provincial Key Technologies R&D Program of China (Project number 2008GG30002042)." Funding disclosed from national organisation; at low risk of bias

Xu 2017.

Study characteristics
Methods	Prospective, double‐blind, randomised controlled trial Run‐in period: February 2015 to February 2016 Number of study centres and location: single centre. Affiliated People’s Hospital, Nanjing Medical University, Wuxi, China
Participants	40 adult participants with rheumatic heart disease undergoing mechanical valve replacement with CPB Exclusion criteria: patients had valve replacement operation before admission; and patients had autoimmune diseases, chronic diseases of the digestive system, and systemic infectious diseases.
Interventions	Intervention: urinary trypsin inhibitor 5000 U/kg before CPB during the operation, and continuing with IV injection of 5000 U/kg for 3 days after the operation Control: 0.9% sodium chloride 5000 U/kg before CPB during the operation, and continuing with IV injection of 5000 U/kg for 3 days after the operation
Outcomes	TNF‐α, IL‐1β, IL‐6, and IL‐8; preoperative and postoperative liver and renal function index, in which the liver index included ALT, AST, TB, DB, and renal index including Cr; also, ventilator‐assisted time and ICU hospital stay of the 2 groups of patients
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation was not described.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment processes were not clearly described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It was not described whether participants and personnel were blinded to group allocation.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It was not described whether outcome assessors were blinded to group allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	No funding sources were disclosed.

Yamada 1996.

Study characteristics
Methods	Prospective, randomised controlled study Run‐in period: not specified. Published 1996 Number of study centres and location: single centre. Yokohama City University School of Medicine, Japan
Participants	Eighteen patients scheduled for elective CABGs Exclusion criteria not explicitly stated
Interventions	Intervention: heparin‐coated CPB circuit Control: uncoated CPB circuit (otherwise identical to the heparin‐coated circuit)
Outcomes	IL‐8, TNF‐α, neutrophil elastase, and neutrophils
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation process not described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment processes not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding of participants and personnel was not described.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear whether outcome assessors were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	No funding sources were disclosed.

Yared 2000.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled trial Run‐in period: January 15 1997 to September 25 1997. Published 2000 Number of study centres and location: single centre. The Cleveland Clinic Foundation, Cleveland, Ohio, USA
Participants	236 participants aged ≥ 20 years and having elective coronary or valvular heart surgery with CPB Exclusion criteria: patients with a history of hypersensitivity to dexamethasone or of diabetes mellitus treated with insulin or oral medications; and patients receiving corticosteroids already.
Interventions	Intervention: dexamethasone (0.6 mg/kg, after induction of anaesthesia) Control: equal volume of 0.9% sodium chloride
Outcomes	The primary endpoint of the randomised study was the comparison of the effect of dexamethasone versus placebo on the incidence of shivering. Secondary outcomes: the time elapsed from ICU admission to tracheal extubation, ICU and hospital length of stay, and mortality, as well as the incidence of major neurologic, renal, cardiac, infectious, and pulmonary morbidities, obtained from the Cardiothoracic Anesthesia Database. Additional data were collected about new onset atrial fibrillation in the first 3 PODs and administration of continuous infusions of insulin, vasopressors (norepinephrine), and inotropic drugs (epinephrine, dobutamine, or milrinone) in the ICU on the day of surgery. Early postoperative fever was defined as peak T °PA of 38 °C or more in the first 6 hours after ICU admission.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation procedures were not described.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment procedures were not described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It was not clearly described whether participants and personnel were blinded to the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It was not clearly described whether outcome assessors were blinded to the intervention.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Patients who were enrolled in the study but returned to the operating room because of surgical bleeding, or who received additional corticosteroids or other drugs such as aprotinin that interfere with the inflammatory response, were excluded from the analysis". Although the indication for this can be understood, it can also introduce bias, especially as the trial is looking at clinical outcomes.
Selective reporting (reporting bias)	Low risk	Although no protocol was published, the outcomes were clearly stated in the methodology and reported on.
Other bias	Low risk	No conflict of interest detected

Yared 2007.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled trial Run‐in period: October 2000 to January 2001. Published 2007 Number of study centres and location: single centre. Cleveland Clinic Foundation, Cleveland, Ohio, USA
Participants	Seventy‐eight adult patients undergoing combined valve and CABG surgery Exclusion criteria: not explicitly stated
Interventions	Intervention: dexamethasone (0.6 mg/kg, after induction of anaesthesia) Control: equal volume of 0.9% sodium chloride
Outcomes	IL‐6, IL‐8, and IL‐10; TNF‐α; and ET‐1, C‐4 and CRP, exhaled NO, and incidence of atrial fibrillation postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A computer‐generated random table was used."
Allocation concealment (selection bias)	Unclear risk	Allocation concealment processes were not described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It was unclear whether participants and personnel were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It was unclear whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	High risk	CRP, C‐4, and NO were only reported for selected group of participants.
Other bias	Low risk	No funding sources were disclosed.

Yeager 2005.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled trial Run‐in period: recruitment over a period for 16 months (dates not specified). Published 2005 Number of study centres and location: not specified. Authors from Dartmouth Medical School, Hanover and Dartmouth‐Hitchcock Medical Center, Lebanon, New Hampshire, USA.
Participants	Sixty elective cardiac surgical patients scheduled for CABG, cardiac valve replacement, or both Exclusion criteria: history of ongoing systemic inflammatory disease of any kind; treatment with systemic glucocorticoids for any reason within 3 months before admission.
Interventions	Intervention: etomidate and hydrocortisone (8 μg/kg over 6 h) Control: thiopentone and saline placebo
Outcomes	IL‐6, IL‐10, cortisol, MIF
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation procedures were not clearly described.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Authors state that the trial was double‐blinded, but they do not clearly describe how this was achieved.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	Quote: "Supported by grant NIAID AI051547 from the National Institutes of Health (PMG)" Funding source at low risk of bias

Yilmaz 1999.

Study characteristics
Methods	Double‐blind, prospective, randomised controlled trial Run‐in period: ethical approval granted December 1994. All other timelines unspecified Number of study centres and location: single centre. Uludag University Medical Faculty, Bursa
Participants	Twenty adult patients with isolated CAD and scheduled to undergo a primary elective CABG operation Exclusion criteria: concurrent use of corticosteroids, salicylates, dipyridamol, or anticoagulants; concurrent COPD, chronic renal failure, insulin‐dependent diabetes, congestive cardiac failure, peptic ulcer history; prior cardiac operations; recent (30 days) MI; steroid dependency.
Interventions	Intervention: low‐dose MPSS (1 mg/kg into the pump prime solution) Control: placebo
Outcomes	CK‐MB, WBC, IL‐6 and IL‐8 levels, and clinical outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation method not clearly described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment methods not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It is not clearly stated how participants and personnel were blinded even though the study is described as a double‐blinded study.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clearly stated how outcome assessors were blinded even though the study is described as a double‐blinded study.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Of the 25 enrolled patients, 20 patients completed the study. Three patients who required blood transfusion during or after surgery and two patients who had impaired glucose tolerance were excluded from further analyses." The reasons for excluding five participants indicate that these happened during the perioperative period. This uncertainty around numbers of randomised participants and the reasons for excluding five of the recruited participants is the reason for the high risk of bias judgement.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigations were detailed in the method sections and reported.
Other bias	Low risk	No funding sources were disclosed.

Youn 2011.

Study characteristics
Methods	Single‐centre, prospective, randomised controlled study Run‐in period: December 2007 to October 2008 Number of study centres and location: single centre. Yonsei Cardiovascular Hospital, Yonsei University College of Medicine, Yonsei University Health System, Seoul, South Korea
Participants	142 patients with stable angina who were scheduled to undergo surgical myocardial revascularisation Exclusion criteria: evidence of recent MI and unstable angina; emergent or urgent operation; history of prior cardiac surgery; any increase in liver enzymes; renal failure with serum Cr > 3 mg/dL; LVEF < 35%; and history of liver or muscle disease.
Interventions	Intervention: rosuvastatin 40 mg 12 h before surgery and 20 mg 2 h before surgery Control: no intervention
Outcomes	The primary endpoint was the occurrence of MACEs, which included the following: death from all causes, nonfatal MI, and repeat revascularisation by percutaneous intervention or bypass surgery within 30 days after the operation. The secondary endpoint was the difference in the degree of myocardial ischaemia between the 2 groups after OPCAB. The degree of myocardial ischaemia was measured by the highest level of cardiac enzymes after OPCAB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were assigned in a 1: 1 ratio using a computer‐generated randomization sequence."
Allocation concealment (selection bias)	Unclear risk	Allocation concealment methods were not clearly described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding methdology was not clearly described.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding methodology was not clearly described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Low risk	The primary and secondary outcomes are clearly stated in the methodology and are reported.
Other bias	Low risk	Authors do not declare any conflicts of interest.

Yuhe 2020.

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: February 2009 to December 2012. Submitted: October 8, 2018; accepted: January 9, 2019; published July 17, 2020 Number of study centres and location: single centre. Tertiary cardiac centre in Singapore
Participants	Seventy‐eight patients undergoing elective and isolated CABG surgery at a tertiary centre in Singapore Exclusion criteria: poor LVEF (< 30%); immunologic disease or malignancies; acute inflammatory disease; coagulopathy; steroid treatment; significant carotid disease; on dialysis.
Interventions	Intervention: miniature CPB circuit Control: CCPB circuit
Outcomes	Postoperative cognitive dysfunction
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "They were randomly assigned to the MCPB or CCPB group using the sealed envelope technique." No additional information was provided on how randomisation was achieved apart from mentioning the sealed envelope technique.
Allocation concealment (selection bias)	Low risk	Quote: "They were randomly assigned to the MCPB or CCPB group using the sealed envelope technique."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study was described as single‐blinded. It is unclear who was blinded to study group allocation and the impact on the results.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The study was described as single‐blinded. It is unclear who was blinded to study group allocation and the impact on the results. The authors report that a different team member from the surgeon assessed cognitive function; however, it is not clear whether they were blinded to group allocations.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Out of 78 patients, four patients were lost to follow‐up and three patients died before follow‐up. Therefore, data from 71 patients were analyzed." No intention‐to‐treat analysis
Selective reporting (reporting bias)	Low risk	All outcomes were reported properly.
Other bias	Low risk	Quote: "This work was supported by the National Medical Research Council, Singapore (grant number: NMRC IRG07nov119)."

Zahoor 2007.

Study characteristics
Methods	Single‐centre, prospective, randomised study Run‐in period: not specified. Published 2007 Number of study centres and location: single centre. Armed Forces Institute of Cardiology/ National Institute of Heart Diseases (AFIC/NIHD), Rawalpindi, Pakistan
Participants	100 patients aged over 14 years old Exclusion criteria: cardiac trauma; vascular emergencies; ASA Class IV or higher; and use of aprotinin at any point during hospital stay.
Interventions	Intervention: MUF after bypass Control: no intervention
Outcomes	Ultrafiltration impact on haemoconcentration and postoperative bleeding
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization (lottery method) was done using single blind technique to allocate the patients randomly to both groups of 50 each, and were allocated to receive modified ultrafiltration (MUF group) or not (control group)".
Allocation concealment (selection bias)	Unclear risk	Allocation concealment processes not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Single blinding
Blinding of outcome assessment (detection bias) All outcomes	High risk	The authors state the trial was single blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Four participants from the control group were excluded because of incomplete data.
Selective reporting (reporting bias)	Unclear risk	Unclear criteria to select outcomes
Other bias	Low risk	Authors do not declare any conflicts of interest.

Zhang 2010.

Study characteristics
Methods	Prospective, randomised controlled study Run‐in period: not specified. Published 2010 Number of study centres and location: Guangdong General Hospital, Guangdong Academy of Medical Science, Guangzhou, China.
Participants	52 patients with rheumatic valve lesions due to undergo elective mitral valve or mitral valve replacement and AVR surgery were enrolled randomly into 2 prospective groups.
Interventions	Intervention: leucocyte‐depleting filter (Pall LeukoGuard6, LG6; Pall Biomedical, NY, USA) Control: standard CPB circuit
Outcomes	CK, TnI, IL‐6, IL‐8, TNF‐α, inotropic support requirements, duration on mechanical ventilation, the duration of intensive care, and overall hospital stay
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not clearly described
Allocation concealment (selection bias)	Unclear risk	Not clearly described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clearly described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clearly described
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomised participants completed the study.
Selective reporting (reporting bias)	Low risk	Although no protocol was available, the biomarkers under investigations were detailed in the method sections and reported.
Other bias	Low risk	No funding was disclosed.

Zhang 2020 (A).

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind, controlled trial Run‐in period: April 2008 to December 2008. Retrospective registration February 2010. Received February 20, 2020; accepted May 4, 2020 Number of study centres and location: single centre. Cardiovascular Institute and Fuwai Hospital, Beijing, China
Participants	426 consecutive patients receiving open‐heart surgery with CPB were randomly assigned into 3 groups to receive UST (group U, n = 142), TXA (group T, n = 143), or normal saline (group C, n = 141). Inclusion criteria: between 18 years and 79 years old undergoing elective heart surgery with CPB, including CABG, valvular repair or replacement, or repair of congenital heart deformities Exclusion criteria: previous cardiac surgery, Hct level < 33%, Plt count < 100,000 × 103/L, allergy to TXA, and being recruited in other studies 486 assessed for eligibility 426 randomised UST: 142 allocated, 142 analysed, 22 lost to follow‐up, analysed at 10 years 119 Allocated to TXA: 143, analysed 143, lost to follow‐up 21, analysed at 10 years 122 Allocated to placebo: 141, analysed 141, lost to follow‐up 20, analysed at 10 years 120 Zhang 2020 (A) considered participants treated with TXA and control group participants treated with saline.
Interventions	Intervention: 30 mg/kg body weight of TXA (Jie Ning®; Changchun Tiancheng Pharmaceutical Co., Changchun, China) Control: equal volume of 0.9% sodium chloride There were 2 syringes prepared for each participant, labelled '1' and '2'. In the TXA group, both the syringes contained 15 mg/kg TXA to fulfil a total dosage of 30 mg/kg. In the control group, both the syringes contained saline. Syringes 1 and 2 were administered intravenously after anaesthetic induction and after the administration of protamine, respectively.
Outcomes	The primary outcome of this study was the total volume of postoperative blood loss. The secondary outcomes included stroke, postoperative MI, renal failure, respiratory failure, in‐hospital adverse outcomes, and long‐term morbidities and mortalities.
Notes	Zhang 2020 (A) considered participants treated with TXA and control group participants treated with saline.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence was generated by computer in permuted blocks by a 1:1:1 ratio and was masked in sealed, sequentially numbered, and opaque envelopes.
Allocation concealment (selection bias)	Low risk	Sequentially numbered and opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The participants, medical staff, and investigators were unaware of the treatment allocation until the end of the study.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The participants, medical staff, and investigators were unaware of the treatment allocation until the end of the study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Early outcomes (in hospital and 30 days) are reported for the complete population. Lost to follow‐up at 10 years are for UST group 22/142, 21/143 for TXA, and 20 for control group.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Zhang 2020 (B).

Study characteristics
Methods	Single‐centre, prospective, randomised, double‐blind, controlled trial Run‐in period: April 2008 to December 2008. Retrospective registration February 2010. Received February 20, 2020; accepted May 4, 2020 Number of study centres and location: single centre. Cardiovascular Institute and Fuwai Hospital, Beijing, China
Participants	426 consecutive patients receiving open‐heart surgery with CPB were randomly assigned into 3 groups to receive UST (group U, n = 142), TXA (group T, n = 143), or normal saline (group C, n = 141). Inclusion criteria: between 18 years and 79 years old undergoing elective heart surgery with CPB, including CABG, valvular repair or replacement, or repair of congenital heart deformities Exclusion criteria: previous cardiac surgery, Hct level < 33%, Plt count < 100,000 × 103/L, allergy to TXA, and being recruited in other studies 486 assessed for eligibility 426 randomised UST: 142 allocated, 142 analysed, 22 lost to follow‐up, analysed at 10 years 119 Allocated to TXA: 143, analysed 143, lost to follow‐up 21, analysed at 10 years 122 Allocated to placebo: 141, analysed 141, lost to follow‐up 20, analysed at 10 years 120 Zhang 2020 (B) intervention group: UST
Interventions	Intervention: 1,000,000 U UST Control: equal volume 0.9% sodium chloride. In the UST group, syringe 1 contained 1,000,000 U UST, and syringe 2 contained saline. In the control group, both the syringes contained saline. Syringes 1 and 2 were administered intravenously after anaesthetic induction and after the administration of protamine, respectively.
Outcomes	The primary outcome of this study was the total volume of postoperative blood loss. The secondary outcomes included stroke, postoperative MI, renal failure, respiratory failure, in‐hospital adverse outcomes, and long‐term morbidities and mortalities.
Notes	Zhang 2020 (B) intervention group: UST
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence was generated by computer in permuted blocks by a 1:1:1 ratio and was masked in sealed, sequentially numbered, and opaque envelopes.
Allocation concealment (selection bias)	Low risk	Sequentially numbered and opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The participants, medical staff, and investigators were unaware of the treatment allocation until the end of the study.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were unaware of the treatment allocation until the end of the study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Early outcomes (in hospital and 30 days) are reported for the complete population.
Selective reporting (reporting bias)	Low risk	All prespecified variables were reported.
Other bias	Low risk	No clear evidence of funding bias

Zheng 2010.

Study characteristics
Methods	Prospective, randomised controlled study. Run‐in period: from September 2005 to April 2007 Registration date: unspecified, study published in 2010 Number of study centres and location: single centre. Fourth Military Medical University, Xi’an, China
Participants	100 nondiabetic inpatients undergoing valve replacement Mean age: 44 years Sex (female/male ratio): 0.89 Exclusion criteria: preoperative liver or kidney disease or dysfunction, preoperative coagulation disorder, palliative operation, or a second operation
Interventions	Intervention: intensive IT Control: standard perioperative management
Outcomes	TNF‐α, IL‐6, IL‐1β, cTnI, cardiac index measurements, time on ventilator, time in ICU, length of hospital stay, nosocomial infection rate, mortality and rate of hypoglycaemic events; seven time points: after anaesthesia induction (T1), at the initiation of CPB (T2), after the termination of CPB (T3), at 6 hours after CPB (T4), at 12 hours after CPB (T5), at 24 hours after CPB (T6), and at 48 hours after CPB (T7)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The 100 patients were randomised with computerised randomisation tables.
Allocation concealment (selection bias)	Low risk	Study personnel opened the blinded envelopes sequentially after the participants had signed the patient consent form, and patients were assigned to receive either routine therapy (control group, n = 50) or IT (n = 50).
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	No protocol was available; however, the outcomes under investigation were clearly detailed in the methodology and reported.
Other bias	Low risk	No funding was disclosed.

Zheng 2016.

Study characteristics
Methods	Randomised, placebo‐controlled trial Run‐in period: from September 2011 through October 2013 Registration date: ClinicalTrials.gov number NCT01573143, study published in 2016 Number of study centres and location: single centre. Department of Cardiovascular Surgery, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, China
Participants	1922 patients who were scheduled for elective cardiac surgery to receive perioperative rosuvastatin (at a dose of 20 mg daily) (n = 960) or placebo (n = 962) Mean age: 59 years Sex (female/male ratio): 0.26 Exclusion criteria: moderate or severe mitral‐valve disease or known renal dysfunction (Cr level > 2.3 mg/dL (200 μmol/L)) or had contraindications to statin therapy
Interventions	Intervention: rosuvastatin at a dose of 20 mg once daily for up to 8 days before surgery and for 5 days thereafter Control: placebo tablets at a dose of 20 mg once daily for up to 8 days before surgery and for 5 days thereafter
Outcomes	Postoperative AF, myocardial injury within 120 hours after surgery (cTnI), major in‐hospital adverse events and death, duration of stay in the hospital and ICU, LV and renal function, and blood biomarkers (LDL cholesterol and CRP) Blood samples were obtained 6 hours, 24 hours, 48 hours, and 120 hours after surgery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation schedule by sequential number generated by Clinical Trial Service Unit at Oxford
Allocation concealment (selection bias)	Low risk	“Mixed” randomisation approach
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Adequate blinding ("evaluated the prespecified outcomes systematically in a blinded manner")
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Adequate blinding ("evaluated the prespecified outcomes systematically in a blinded manner")
Incomplete outcome data (attrition bias) All outcomes	Low risk	No significant number of trial group changes, withdrawals, or losses to follow‐up were reported. Data from all the patients were included in the final analysis.
Selective reporting (reporting bias)	Low risk	Prespecified outcomes reported
Other bias	Low risk	Study drug was bought.

Zhou 2019 (A).

Study characteristics
Methods	Prospective, randomised controlled trial Run‐in period: from August 2017 to December 2018 Registration date and number: ChiCTR‐IPR‐17012544, study published in 2019 Number of study centres and location: single centre. Department of Anesthesiology, Intensive Care Medicine and Pain Medicine, First Affiliated Hospital of Soochow University, Suzhou, China
Participants	180 patients with heart valve replacement surgery undergoing CPB, aged 60 years to 80 years, ASA grades I to III, education level above elementary school, and either gender Mean age: 70 years Sex (female/male ratio): 1.2 Exclusion criteria: MMSE score ≤ 23, acute or chronic infectious diseases, taking anti‐inflammatory drugs or immunosuppressants, a stroke in the prior 6 months or any other central nervous system diseases, peptic ulcer disease, stay in intensive care unit ≥ 3 days, BMI > 35, severe deafness or vision problems, illiteracy or communication difficulties related to pronunciation or dialect, postoperative delirium, and refusal or unexpected discharge
Interventions	Intervention: UST (20,000 IU/kg immediately after induction and the first day after surgery) Control: standard management
Outcomes	IL‐6, MMP‐9, TNF‐α (before anaesthesia [T0], immediately after surgery [T1], and immediately after extubation [T2]), cognitive testing, neuroimaging
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence without stratification was generated by a computer and sealed with consecutively numbered envelopes.
Allocation concealment (selection bias)	Low risk	The randomisation sequence without stratification was generated by a computer and sealed with consecutively numbered envelopes.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and investigators were all blind to group allocation until the final statistical analysis was completed.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants and investigators were all blind to group allocation until the final statistical analysis was completed.
Incomplete outcome data (attrition bias) All outcomes	High risk	No power calculation. Two patients were withdrawn from the study because of death. Although the number is small and unlikely to have affected the results, the reason for withdrawal (death) could have led to larger bias had the mortality rate been higher.
Selective reporting (reporting bias)	Low risk	Protocol published: prespecified outcomes are reported on.
Other bias	Low risk	This study was funded by the National Natural Science Foundation of China (81870857) – at low risk of funding bias.

Zhou 2019 (B).

Study characteristics
Methods	Prospective, single‐centre, randomised clinical trial Run‐in period: from 20 August to 30 October 2018 Registration date and number: JXEY‐20180718H01, study published in 2019 Number of study centres and location: single centre. Department of Anesthesiology, the Second Affiliated Hospital of Jiaxing University, Jiaxing City, China
Participants	Thirty‐five patients scheduled for elective valve replacement surgery under CPB, from 20 August to 30 October 2018 Mean age: 58 years Sex (female/male ratio): 1.3 Exclusion criteria: emergency surgery; age <18 years; NYHA cardiac function classification > grade IV; severe COPD; coronary heart disease, MI, or a high cTnI concentration (cTnI ≥ 0.2 ng/mL); renal insufficiency (Cr>41.6 mg/dL); hypotension (MAP < 60 mmHg); and bradycardia (HR < 50 beats/min) in the past 3 days
Interventions	Intervention: dexmedetomidine 0.5 µg/kg loading dose for 10 minutes and a 0.5‐μg/kg/h maintenance dose Control: placebo
Outcomes	cTnI concentration, MDA, superoxide dismutase, TNF‐α, IL‐6, and IL‐8. Samples were collected immediately after induction and 1 h, 6 h, 12 h, and 24 h after CPB.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation
Allocation concealment (selection bias)	Low risk	Double‐blinding, unlabelled syringes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Experimental solutions of dexmedetomidine or saline were unlabelled and diluted by pharmacists in the central pharmacy and were not known to the investigators.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information
Incomplete outcome data (attrition bias) All outcomes	Low risk	One participant per group lost to follow‐up
Selective reporting (reporting bias)	Low risk	Adequate choice and reporting of outcomes
Other bias	Low risk	Funding from the Zhejiang Science and Technology Project for Medical and Health (NO: 2015KYB388) and the Jiaxing Municipal Science and Technology Project (NO: 2016BY28019)

Ziegeler 2009.

Study characteristics
Methods	Prospective, randomised clinical trial Run‐in period: from 20 August to 30 October 2018 Registration date and number: JXEY‐20180718H01, study published in 2019 Number of study centres and location: single centre. Department of Thoracic and Cardiovascular Surgery, University of Saarland, Homburg ⁄Saar, Germany
Participants	Forty‐five male patients aged 35 to 80 (ASA physical status III) scheduled for CABG surgery Mean age: 64 years Sex (female/male ratio): 0 Exclusion criteria: previous cardiac surgery, coexisting congestive (EF < 40%) or valvular heart disease, MI during the last 3 months, or evidence of concomitant malignant or immunologic diseases
Interventions	Intervention: haemofiltration during CPB (15 mL ⁄ kg after rewarming) Control: no intervention, standard CPB management
Outcomes	Expression density of the monocytic surface receptor CD14, HLA‐DR expression and cytokine release (TNF‐α and IL‐10); blood samples were taken after induction of anaesthesia (T1), 20 minutes after CPB (T2), and 24 h postoperatively (T3)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Forty‐five male patients aged 35 to 80 (ASA‐physical status III) scheduled for CABG surgery were enroled and randomised to one of three groups according to a computer‐generated random list.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described (unlikely because of the nature of the intervention)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	The control group was smaller than the interventions group. One patient from the control group was withdrawn from the analysis because of death (not a prespecified exclusion criterion). Furthermore, the sample size is much smaller than planned.
Selective reporting (reporting bias)	High risk	Very vague prespecified outcomes in the published protocol
Other bias	Low risk	This study was supported by a grant (P03 ⁄ 05 ⁄ ⁄ A64 ⁄ 04) from the Else Kroener‐Fresenius‐Stiftung, Bad Homburg v.d.H., Germany. None of the authors has any financial or other potential conflict of interest. Funding was from a charitable organisation.

AaDO₂: alveolar‐arterial oxygen difference

AAGP: α1‐acid glycoprotein

ACE: angiotensin‐converting enzyme

ACT: activated clotting time

ADO: administration of adenosine

AF: atrial fibrillation

aHb: haemoglobin (from arterial sample)

aHCT: haematocrit (from arterial sample)

AKI: acute kidney injury

Akt: protein kinase B

ALI: acute lung injury

ALT: alanine aminotransferase

AMI: acute myocardial infarction

AMPK: adenosine monophosphate‐activated protein kinase

Ang1: angiopoietin 1

Ang2: angiopoietin 2

Ang3: angiopoietin 3

aPTT: activated partial thromboplastin time

ARDS: acute respiratory distress syndrome

ASA: American Society of Anesthesiologists

AST: aspartate aminotransferase

avDO₂: arteriovenous oxygen content difference

a‐vo2D: arterial‐venous difference in oxygen

AVR: aortic valve replacement

BAL: bronchoalveolar lavage

BBB: blood brain barrier

BDI‐II: Beck Depression Inventory – Second Edition

BFABP: brain type fatty binding protein

BK: bradykinin

BMI: body mass index

BNP: brain natriuretic peptide

BP: blood pressure

BPS: bronchopulmonary sequestration

BSA: body surface area

b‐TG: beta‐thromboglobulin

BUN: blood urea nitrogen

C: complement

C1‐INH

C3a: complement 3 activation

C3bc: complement 3 activation product

CA: cardiac arrest

CAB: coronary artery bypass

CABG: coronary artery bypass graft(ing)

CAD: coronary artery disease

cAMP: cyclic AMP

CCE: cardiac cycle efficiency

CCECC: closed‐circuit extracorporeal circulation

C‐coated

cCPB: conventional cardiopulmonary bypass

CCPB: conventional cardiopulmonary bypass

CD: cluster of differentiation

Cd: dynamic compliance

CD40L: cluster of differentiation 40 ligand

CECC: conventional extracorporeal circuit

CFT: clot formation time

CH50: complement total activity

CHF: congestive heart failure

CI: confidence interval

CK: creatine kinase

CK‐MB: creatinine kinase–MB

CondECC: condensed dual‐function open/closed configuration circuit

COPD: chronic obstructive pulmonary disease

COX: cyclooxygenase

CPAP: continuous positive airway pressure

CPB: cardiopulmonary bypass

CPK: creatine‐phosphokinase

Cr: creatinine

CRP: C‐reactive protein

CS: coronary sinus

Cs: static compliance

CSF: cerebrospinal fluid

CT: coagulation time

CTICU: cardiothoracic intensive care unit

cTnI: cardiac troponin I

cTnT: cardiac troponin T

CUF: conventional ultrafiltration

CVA: cerebrovascular accident

CVC: central venous catheter

CVP: central venous pressure

CVVH: continuous venovenous haemofiltration

CXR: chest x‐ray

DB: direct bilirubin

desArg: without arginine

DEX: dexamethasone

DHA: docosahexaenoic acid

DHCA: deep hypothermic circulatory arrest

DO₂I: oxygen delivery index

DPO_2: difference in partial pressure of oxygen

DVR: double valve replacement

EACA: ε‐aminocaproic acid

ECC.O: extracorporeal circulation circuit optimised

ECC: extracorporeal circulation

ECG: electrocardiogram

EDTA: ethylenediaminetetraacetic acid

EF: ejection fraction

eGFR: estimated glomerular filtration rate

ELISA: enzyme‐linked immunosorbent assay

EPO: erythropoietin

ET‐1: endothelin‐1

EuroSCORE: European System for Cardiac Operative Risk Evaluation

FBC: full blood count

FEV1: forced expiratory volume in 1 second

FFP: fresh frozen plasma

fHB: free haemoglobin

FiO₂: inspired oxygen fraction

FVC: forced vital capacity

GCS: Glasgow coma scale

GFR: glomerular filtration rate

GIK: glucose–insulin–potassium

GST: Glutathione S‐transferase

HA: haemadsorption

Hb: haemoglobin

H‐CAM: homing cellular adhesion molecule

Hct: haematocrit

HDL: high‐density lipoprotein

HFABP: ischaemic biomarker heart‐type fatty acid binding protein

H‐FABP: ischaemic biomarker heart‐type fatty acid binding protein

HLA‐DR: human leukocyte antigen – DR isotype

H‐LOS: hospital length of stay

HOE: D‐Arg‐[Hyp3, Thi5, D‐Tic7, Oic8]bradykinin

HR: heart rate

hs‐CRP: high‐sensitivity C‐reactive protein

hs‐cTnI: high‐sensitivity cardiac troponin I

hs‐cTnT: high‐sensitivity cardiac troponin T

HSP70: heat shock protein 70

IABP: intra‐aortic balloon pump

ICAM‐1: intercellular adhesion molecule‐1

ICU: intensive care unit

I‐FABP: intestinal fatty acid binding protein

IFN‐γ‐ interferon γ

IGF: insulin‐like growth factor

IgG: immunoglobulin G

IL: interleukin

IL‐1ra: interleukin‐1 receptor antagonist

INN: international nonproprietary name

iNO: inducible nitric oxide

INR: international normalised ratio

iPF2‐III: F2‐isoprostane III

IP10: interferon‐inducible protein‐10

ISVR: indexed systemic vascular resistance

IT: insulin therapy

ITA: internal thoracic artery

ITT: intention‐to‐treat

ITU: intensive therapy unit

IU: international units

IV: intravenous

KDOQI: Kidney Disease Outcome Quality Initiative

KIU: kallikrein inhibiting unit

KIV: Kallikrein Inhibitor Vnits

LAD: left anterior descending artery

LBBB: left bundle‐branch block

LBP: lipopolysaccharide‐binding protein

LCOS: low cardiac output syndrome

LDH: lactate dehydrogenase

LDL: low‐density lipoprotein

LDL‐BDC: low‐density lipoprotein baseline diene conjugation

LDL‐TRAP: LDL‐total peroxyl radical trapping antioxidant potential

LIMA: left internal mammary artery

LOS: length of stay

LTB4: leukotriene B4

LV: left ventricular

LVDED: left ventricular end‐diastolic diameter

LVEDP: left ventricular end‐diastolic pressure

LVEF: left ventricular ejection fraction

LVMI: left ventricular mass index

MA: maximal amplitude

MAC: monitored anaesthesia care

MACCE: major adverse cardiac and clinical event

MACE: major adverse cardiac event

MAP: mean arterial pressure

MAPK: mitogen‐activated protein kinase

MCF: mean clot firmness

MCP‐1: monocyte chemoattractant protein 1

mCPB: minimised cardiopulmonary bypass

MDA: malondialdehyde

MECC: minimised extracorporeal circulation

MG: methylprednisolone group

MI: myocardial infarction

MICABG: minimally invasive coronary artery bypass grafting

MICS: minimally invasive cardiac surgery

MiDCAB: minimally invasive direct coronary artery bypass

MiECC: minimised extracorporeal circulation

MIF: migration inhibitory factor

MIG: monokine induced by interferon‐γ

MMP: matrix metalloproteinase

MMP‐9: matrix metallopeptidase 9

MMSE: Mini‐Mental State Examination

MNC: mononuclear cell

MP: methylprednisolone

MPC: minimised perfusion circuit

MPO: myeloperoxidase

MPSS: methylprednisolone sodium succinate

mRNA: messenger ribonucleic acid

MUF: modified ultrafiltration

MVR: mitral valve replacement

N/A: not applicable

NAC: N‐acetylcysteine

NAG: N‐acetylglutamate

NC: nasal cannula

NF‐kB: nuclear factor kappa B

NGAL: neutrophil gelatinase‐associated lipocalin

NIV: non‐invasive mechanical ventilation

NMF: non‐methylprednisolone group

N‐MUF: normovolemic modified ultrafiltration

NO: nitric oxide

N/R: not reported

NSE: neuron‐specific enolase

NT‐proBNP: N‐terminal probrain natriuretic peptide

NYHA: New York Heart Association

OER: oxygen extraction rate

ONCAB: on‐pump coronary artery bypass

OnP‐BH: on‐pump beating‐heart

OPCAB: off‐pump coronary artery bypass

OPCABG: off‐pump coronary artery bypass graft

ORP: oxidation‐reduction potential

PA‐aDO2: pulmonary alveolar‐arterial oxygen gradient

PAaO₂: alveolar‐arterial oxygen gradient

PAC‐1: first procaspase activating compound

PAF: paroxysmal atrial fibrillation

PAI‐1: plasminogen activator inhibitor

PaO₂: partial pressure of oxygen

PAP: plasmin–antiplasmin

PAPc: plasmin–antiplasmin complex

PBMC: peripheral blood mononuclear cells

PC: phosphorylcholine

PCPB: percutaneous cardiopulmonary bypass

PCR: polymerase chain reaction

PCT: procalcitonin

PCWP: pulmonary capillary wedge pressure

PDE: phosphodiesterase

PECAM‐1: platelet‐endothelial cell adhesion molecule‐1

PEEP: peak end‐expiratory pressure

PF‐1.2: prothrombin fragment 1.2

PF4: platelet factor 4

PF‐4: platelet factor 4

PGT2: prostaglandin T2

PI: perfusion index

PIC: peripherally inserted catheter

Plt: platelets

PMEA: poly(2‐methoxyethyl acrylate)

PMN‐Elastase: polymorphonuclear neutrophils elastase

PMVP: pulmonary microvascular pressure

pO₂: partial pressure of oxygen

POAF: post‐operative atrial fibrillation

POCD: postoperative cognitive dysfunction

POD: postoperative day

PS: pericardial suction

PT: prothrombin time

PTCA: percutaneous transluminal coronary angioplasty

PTX: pentoxifylline

PTX3: long pentraxin

PVR: pulmonary vascular resistance

QoR‐40: quality of recovery score

Qs: Quantum sufficiat

RAGE: receptor for advanced glycation end‐product

RANTES: regulated on activation normal T cell expressed and secreted

RBC: red blood cell

RC3bRR: red blood cell C3b receptor ratio

RF: renal failure

RhF: rheumatic factor

r‐HuEPO: recombinant human erythropoietins

RI: respiratory index

RICR: red blood cell immune‐component ratio

RIPC: remote ischaemic preconditioning

RIPerc: remote ischaemic perconditioning

rLIS: radiographic lung injury score

ROS: reactive oxygen species

rSO₂: regional cerebral oxygen saturation

r‐TEG: rapid thromboelastography

SAA: serum amyloid A

SAVR: aortic valve replacement surgery

SC5b‐9: soluble complement 5b‐9

sCD14: soluble cluster of differentiation 14

SD: standard deviation

SECC: standard extra‐corporeal circulation

sE‐selectin: soluble endothelial selectin

sGPV: soluble plasma glycoprotein

sICAM: soluble intercellular adhesion molecule

SIRS: systemic inflammatory response syndrome

SiV: Sivelestat sodium

SMA: surface‐modifying active copolymer

SNP: single nucleotide polymorphism

SOD: superoxide dismutase

SOFA: sequential organ failure assessment

SOP: standard operating procedure

sPLA2: secretory phospholipase A2

sP‐selectin: soluble platelet selectin

STEMI: ST elevation myocardial infarction

sTNF‐IIR: soluble tumour necrosis factor receptor 2

sTNFR: soluble tumour necrosis factor receptor

SvO₂: venous oxygen saturation

SVR: systemic vascular resistance

TAT: thrombin‐antithrombin complex

TATc: thrombin–antithrombin‐III complex

TB: total bilirubin

TBARS: thiobarbituric acid reactive substances

TC: total cholesterol

TCC: terminal complement complex

TEG: thromboelastography

TF: tissue factor

TG: thromboglobulin

TG: triglyceride

TGFb1: transforming growth factor beta 1

Timp2: tissue inhibitor metalloproteinase‐2

TLR4: toll‐like receptor 4

TM: thrombomodulin

TNF‐R: tumour necrosis factor receptor

TNFsr: soluble tumour necrosis factor receptor

TNF‐α: tumour necrosis factor‐α

TnI: troponin I

TnT: troponin T

TP10: soluble complement receptor 1

tPA: tissue‐type plasminogen activator

Treg: regulatory T cells

TXA: tranexamic acid

UF: ultrafiltration

UNGAL: urine neutrophil gelatinase‐associated lipocalin

UST: ulinastatin

UTI: urinary tract infection

VACPB: vacuum‐assisted cardiopulmonary bypass

VCAM‐1: vascular cell adhesion molecule‐1.

Vd: volume of distribution

VEGF: vascular endothelial growth factor

VEGFR: vascular endothelial growth factor receptor

vHb: haemoglobin (venous)

vHCT: haematocrit (venous)

VLDL: very low‐density lipoprotein

VO₂: oxygen consumption index

Vt: ventricular tachycardia

vWF: von Willebrand factor

WBC: white blood cell

WCC: white cell count

ZBUF: zero‐balance ultrafiltration

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Adabag 2008	Established organ damage
Akowuah 2017	Protocol only
Alexiou 2004	Did not assess IL‐6 nor IL‐8 (or any other included inflammatory markers
Anastasiadis 2013	Systematic review and meta‐analysis
Baker 2009	Wrong publication type (systematic review)
Beaver 2018	Inadequate intervention and presentation of results
Bingol 2005	Outcomes did not meet protocol criteria ‐ outcomes FEV1/FVC
Brie 2022	Wrong population (percutaneous coronary intervention)
Brown 2009	Meta‐analysis
Bunenkov 2020	Non‐randomised
Burns 2005	Not eligible due to characteristics of the patient population
Casula 2022	Ineligible publication type
Chen 2022	Ineligible publication type
Diab 2022	Ineligible intervention type and population (cytokine filter)
Diegeler 2000	Not randomised
Dieleman 2011	Meta‐analysis
Dvirnik 2018	Meta‐analysis
Elgebaly 2020	None of the recorded outcomes was eligible as per our protocol.
Eren 2003	Outcomes did not meet protocol criteria.
Gott 1998	No comparison group
Gursu 2013	No control group
Haase 2007	No relevant outcomes
He 2018	Meta‐analysis of randomised controlled trials
Kim 2012	Inadequate intervention type (postoperative ischaemic conditioning)
Krivoy 2008	No control group
Landis 2014	Inadequate publication type (systematic review and meta‐analysis)
Later 2013	Uneligible publication type
Loubser 1997	No clear evidence of randomisation
Luciani 2009	Non‐relevant intervention
Morisaki 2013	Non‐randomised study (authors emailed to clarify, no response obtained)
Nader 2004	Sevoflurane as part of cardioplegia
NCT00484575	No control group
NCT00484575 2007	No control group. Registration for Steurer 2012, "Organ Protection With Sevoflurane Postconditioning After Cardiac Surgery With Cardiopulmonary Bypass".
NCT02118025 2006	No control group
Ng 2020	Systematic review of randomised controlled trials
Orhan 2006	Outcomes assessed during intraoperative period
Ozaydin 2008	Ineligible outcomes
Permanyer 2020	No control group
Ristikankare 2006	No relevant outcomes, established organ damage
Sano 2003	No relevant outcomes
Santarpino 2009	No control group
Schonebeck 2007	Congress abstract
Senay 2009	Outcomes measured intraoperatively
Sisillo 2008	Established organ damage
Tosun 2013	None of the recorded outcomes was eligible as per our study protocol.
Volk 2003	Duplicate reference for Volk 2001
Von Spiegel 2001	No relevant outcomes
Wang 2009	Not a relevant intervention (bradykinin preconditioning)
Wijeysundera 2007	Established organ damage
Yao 2020	Systematic review
Zakkar 2015	Best evidence topic
Zhang 2014	Systematic review and meta‐analysis
Zhen‐Han 2017	Meta‐analysis

IL: interleukin

FEV1: forced expiratory volume

FVC: forced vital capacity

Characteristics of studies awaiting classification [ordered by study ID]

Kawamura 1999.

Methods	Prospective, randomised, nonblinded study
Participants	Twenty‐one patients on cardiopulmonary bypass undergoing aortocoronary bypass surgery (10 control, 11 MPS).
Interventions	The methylprednisolone group [MPS] [11 patients] received an intravenous administration of 30 [micro sign]g/kg methylprednisolone before CPB surgery and before declamping of the aorta.
Outcomes	identifying the effects of methylprednisolone on anti‐inflammatory cytokines that may contribute to the low incidence of inflammatory clinical sequelae
Notes	Unable to retrieve full text

Miura 1998.

Methods	Unable to assess as full text not retrieved
Participants	Patients undergoing CPB
Interventions	Ulinastatin
Outcomes	Unable to assess as full text not retrieved
Notes	Unable to assess as full text not retrieved

Zhou 2010.

Methods	Unable to assess as full text not retrieved
Participants	Unable to assess as full text not retrieved
Interventions	Ulinastatin
Outcomes	Unable to assess as full text not retrieved
Notes	NA

CPB: cardiopulmonary bypass
MPS: methylprednisolone
NA: not applicable

Characteristics of ongoing studies [ordered by study ID]

Cardoso 2021.

Study name	Effect of sevoflurane on the inflammatory response during cardiopulmonary bypass in cardiac surgery: the study protocol for a randomized controlled trial
Methods	Randomised controlled trial
Participants	One hundred patients undergoing elective coronary artery bypass graft with cardiopulmonary bypass who will be randomised to receive either volatile anaesthetics during cardiopulmonary bypass or total intravenous anaesthesia
Interventions	Volatile anaesthetics during cardiopulmonary bypass or total intravenous anaesthesia
Outcomes	Inflammatory response during cardiopulmonary bypass.
Starting date
Contact information
Notes

ChiCTR2000038585.

Study name	Multimodal analgesia versus traditional opiate based analgesia
Methods	Randomised controlled trial
Participants	Elective on pump cardiac surgery through median sternotomy
Interventions	Multimodal analgesia
Outcomes	Analgesic effect; additional opioid consumption; postoperative delirium; incidence of acute kidney injury; dysglycemia; mechanical ventilation time; ICU stay; hospital stay; systemic immune inflammation index
Starting date	2020
Contact information	http://www.who.int/trialsearch/Trial2.aspx?TrialID=ChiCTR2000038585
Notes

ChiCTR2000041099.

Study name	Effects of dexmedetomidine on renal function in patients undergoing cardiac valve surgery: a randomized placebo‐controlled study
Methods	Randomised controlled trial
Participants	Elective general anaesthesia surgery under extracorporeal circulation in valve surgery
Interventions	Dexmedetomidine intravenous injection Placebo group: intravenous injection
Outcomes	Serum creatinine and urea; serum inflammatory markers; outcome of arterial blood gas analysis; postoperative recovery; haemodynamics; cardiac function and liver function
Starting date	2020
Contact information	https://trialsearch.who.int/Trial2.aspx?TrialID=ChiCTR2000041099
Notes

ChiCTR2100043950.

Study name
Methods	Randomised controlled trial
Participants	Adult cardiac surgery patients
Interventions	Ulinastatin treatment
Outcomes	PRIMARY OUTCOME: Postoperative cognitive dysfunction SECONDARY OUTCOMES: biochemical blood measurements; inflammatory cytokines; brain MRI
Starting date
Contact information
Notes

Fiorentino 2015.

Study name	Cardioprotection in CABG and AVR patients with RIPC
Methods	Two‐centre randomised controlled trial
Participants	Adults undergoing elective CABG or valve replacement surgery on bypass
Interventions	Remote ischaemic preconditioning
Outcomes	Primary outcome: Myocardial injury; time point(s): 2 preoperatively and 6, 12, 24, 48 and 72 hours after end of cardioplegic arrest Secondary outcomes: 1. Clinical endpoints measured from admission up until 3 months postoperatively 2. Inflammatory and oxidative stress measured two times preoperatively and 5 times postoperatively at 6, 12, 24, 48 & 72 hours after end of cardioplegic arrest
Starting date	06/02/2013
Contact information	f.fiorentino@imperial.ac.uk
Notes	‐

ISRCTN15255199 2019.

Study name	proMPT‐2
Methods	Randomised controlled trial
Participants	Anyone undergoing elective or urgent isolated CABG
Interventions	Placebo: Blood cardioplegia with sham supplementation (normal saline 0.9% weight/volume sodium chloride), Propofol ‐ LD: Blood cardioplegia with low dose (6 mcg/mL) propofol supplementation Propofol ‐ HD: Blood cardioplegia with high dose (12 mcg/mL) propofol supplementation
Outcomes	Primary outcome: Myocardial injury, assessed by serial measurements of cTnT in serum from blood samples collected preoperatively and during the first 48 hours post‐chest closure Secondary outcomes: 1. Systemic metabolic stress as measured by blood lactate at preop, 10 mins post‐cross clamp, 1, 6, 12, 24, 48 hours post‐chest closure 2. Renal function, as measured by creatinine in serum at preop, 1, 6, 12, 24, 48 hours post‐chest closure 3. Markers of Inflammation and oxidative stress as measured by tumour necrosis factor (TNF)‐alpha, interleukin (IL)‐10, IL‐8, IL‐6 and myeloperoxidase (MPO) in serum (Bristol cohort only) at preop, 1, 6, 12, 24, 48 hours post‐chest closure 4. Blood pH at preop, 1, 6, 12, 24, 48 hours post‐chest closure 5. Association between cTnT and circulating level of cardiac‐released microRNA‐1 and exosomal microRNA‐1 content at preop, 1, 6, 12, 24, 48 hours post‐chest closure 6. Association between cTnT and microRNA and exosomal microRNA‐1 content differs between groups (i.e. differs with the propofol supplementation received) at preop, 1, 6, 12, 24, 48 hours post‐chest closure 7. Length of intensive care unit (ICU) stay 8. Length of postoperative hospital stay 9. Clinical outcomes and serious adverse events, i.e. serious postoperative complications (e.g. MI, permanent stroke, acute kidney injury) and death from any cause 10. QoL measured using the Coronary Revascularisation Outcome Questionnaire (CROQ) and the EQ‐5D‐5L questionnaire at baseline, 3 months and 12 months
Starting date	01/12/2017
Contact information	Prompt2‐trial@bristol.ac.uk
Notes	NA

NCT02518087 2015.

Study name	Increased adsorption membranes during cardiopulmonary bypass
Methods	Single‐blind randomised controlled trial
Participants	Adults undergoing cardiac surgery on bypass with expected bypass time > 90 minutes (non‐emergency): double valve replacement or valve replacement and CABG
Interventions	oXiris® membrane set is composed of a 1.5 m2 copolymer of acrylonitrile and sodium methylsulfonate (AN 69) with polyethylenimine treated surface and adhered heparin during sets fabrication [oXiris® (Baxter Gambro)]
Outcomes	Primary outcome: Incidence of cardiac surgery associated AKI Secondary outcomes: 90‐day survival after cardiac surgery [time frame: 90 days] ICU length of stay after cardiac surgery [time frame: Participants will be followed for the duration of ICU stay, an expected average of 2 days] Cytokine circulatory levels during CPB and up to 24 hours after surgery. Cytokine removal: IL‐1, IL‐6, IL‐8, IL‐10, IL‐18, TNFα, MCP‐1, HMBP‐1, RANTES, and GROα levels will be determined in plasma. Determinations will be measured baseline (0 hours), and at T1 (CPB end), T2 (ICU admission), and T3 (24 hours). Urinary and blood levels of NAD will be at measured baseline (0 hours), and at T1 (CPB end), T2 (ICU admission), and T3 (24 hours). Renal function in terms of creatinine and urine output will be registered every 24 h after cardiac surgery up to the first week. KDIGO AKI score will be calculated according to these two items. Hospital length of stay after cardiac surgery [time frame: Participants will be followed for the duration of hospital stay, an expected average of 1 week] Lipopolysaccharide (LPS) removal: LPS levels will be determined in plasma. Determinations will be measured baseline (0 hours), and at T1 (CPB end), T2 (ICU admission), and T3 (24 hours).
Starting date	August 7th 2015
Contact information	JOSE LUIS PEREZ FERNANDEZ, Hospital Universitari de Bellvitge
Notes

NCT02984111 2016.

Study name	Erythropoietin effect on ischemic reperfusion injury in coronary artery bypass graft surgery
Methods	Parallel‐assignment randomised controlled trial
Participants	Adults undergoing elective first time (not re‐do) on pump CABG
Interventions	Arm 1: EPO during aortic cross‐clamp for 45‐60 minutes Arm 2: EPO after induction of anaesthesia and before bypass for 45‐60 minutes Arm 3: Placebo: 50 mL of saline during aortic cross‐clamp for 45‐60 minutes
Outcomes	Primary outcome: 1. Change in interleukin‐6(IL‐6) [time frame: change from baseline until 48 hours after surgery] Serum IL‐6 (an inflammatory marker) level is measured before operation (base), just after operation, 24 hours after operation, 48 hours after operation 2. Change in YKL‐40 [time frame: change from baseline until 48 hours after surgery] Chitinase‐3‐like protein 1 (CHI3L1), also known as YKL‐40 is an inflammatory marker. Serum YKL‐40 level is measured before operation (base), just after operation, 24 hours after operation, 48 hours after operation 3. Change in pro b‐type natriuretic peptide (pro BNP) [time frame: change from baseline until 48 hours after surgery] Serum pro BNP level is measured before operation (base), just after operation, 24 hours after operation, 48 hours after operation Secondary outcomes: 1. Volume of bleeding [time frame: during operation until 24 hours after surgery] 2. Need for blood transfusion [time frame: during and after operation, through study completion, an average of one week] Number of packed cell and fresh frozen plasma that were transfused in the period of time that the patient was hospitalised 3. Ventilation time [time frame: period of time after surgery that patient has been intubated in ICU, an average of 10 hours]
Starting date	December 2016
Contact information	Mahnoosh Foroughi, Clinical Professor, Shahid Beheshti University of Medical Sciences
Notes	NA

NCT03657225 2018.

Study name	Mini versus conventional cardiopulmonary bypass In CABG in Asian patients (Mini‐CPB)
Methods	Randomised controlled trial
Participants	Asian adults (aged 21‐85) undergoing first time on pump CABG revascularisation
Interventions	Mini‐CPB
Outcomes	Primary outcomes: 1. AKI meeting AKIN criteria Secondary outcomes: Neurocognitive decline, inflammation, atrial fibrillation
Starting date	November 2008
Contact information	Lian K Ti, MBBS, MMed National University Health System, Singapore
Notes	NA

NCT04632095 2020.

Study name	Parasternal vs. sternotomy approach for conventional aortic valve replacement
Methods	Randomised controlled trial
Participants	Adult cardiac surgery patients
Interventions	Minimally invasive (parasternal) cardiac surgery
Outcomes	Clinical and inflammatory outcomes
Starting date	2020
Contact information	https://clinicaltrials.gov/show/NCT04632095
Notes

NCT04648540 2020.

Study name	Multimodal opioid‐free anesthesia versus opioid‐based anesthesia for patients undergoing cardiac valve surgeries: RCT
Methods	Randomised controlled trial
Participants	Elective valve surgery patients
Interventions	Multimodal opioid free anaesthesia
Outcomes	Laboratory and clinical outcomes
Starting date	2020
Contact information	https://clinicaltrials.gov/show/NCT04648540
Notes

NCT05162742 2021.

Study name	Colchicine and inflammation in aortic stenosis
Methods	Randomised controlled trial
Participants	Cardiac surgery (aortic stenosis) patients
Interventions	Colchicine
Outcomes	Inflammatory and clinical outcomes
Starting date	2021
Contact information	https://clinicaltrials.gov/show/NCT05162742
Notes

Zhang 2022.

Study name	Evaluation of low‐dose colchicine in patients with cardiopulmonary bypass
Methods	Randomised controlled trial
Participants	132 patients undergoing coronary surgery
Interventions	Low‐dose colchicine
Outcomes	Postoperative myocardial injury and inflammatory response markers
Starting date	2020
Contact information	Nanjing University Medical College
Notes

AKI: acute kidney injury

CPB: cardiopulmonary bypass

CROQ: Coronary Revascularisation Outcome Questionnaire

EPO: erythropoietin

ICU: intensive care unit

IL: interleukin

MPO: myeloperoxidase

MRI: magnetic resonance imaging

NA: not available

TNF: tumor necrosis factor

Differences between protocol and review

We chose to report dichotomous outcomes as risk ratios (instead of odds ratios as planned) to improve the general readability of our work.

All authors were involved in transferring the data into RevMan Web (instead of Review Manager 5.4), and another author (different from the one involved in the upload process) spot‐checked the results against the data collection form.

We renamed our outcomes category (measures of treatment effect on inflammation, measures of treatment effect on clinical outcomes, biomarkers of the innate immune response) to improve readability and standardise our definitions with contemporary scientific literature.

Although not initially planned in the published protocol, a series of post hoc analyses (categorising the trials by individual intervention, by risk of detection bias, and by surgical urgency) was conducted to help investigate heterogeneity in the reporting of outcomes, and subgroup analyses were conducted for all outcomes (despite being initially planned exclusively for primary outcomes). Sensitivity analyses were restricted to trials at low risk of bias for allocation concealment, to allow for a meaningful number of reports to be pooled in a meta‐analysis.

Contributions of authors

RGA: formulated the research question, performed previous work that was the foundation of the current review, co‐ordinated the review, collected data, managed data entry, analysed the data, interpreted the data, and wrote the protocol and the review manuscript.

ST: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

MR: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

AR: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

SP: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

JR: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

GL: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

CL: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

AB: selected studies for inclusion, collected the data, entered data into RevMan, and reviewed and approved the protocol and the final manuscript.

FL: helped in designing the study, provided methodological expert opinion and reviewed and approved the protocol and the final manuscript.

TK: helped in designing the study, provided methodological expert opinion and reviewed and approved the protocol and the final manuscript.

MW: helped in designing the study, provided methodological expert opinion and reviewed and approved the protocol and the final manuscript.

GJM: supervised the study, formulated the research question, performed previous work that was the foundation of the current review, co‐ordinated the review, interpreted the data, and reviewed and approved the protocol and the final manuscript.

Sources of support

Internal sources

• Not applicable, Other

  Not applicable

External sources

• This project was supported by the National Institute for Health Research, via Cochrane Infrastructure funding to the Heart Group. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health., UK

  Cochrane Infrastructure

Declarations of interest

Riccardo Giuseppe Abbasciano: none to declare. RGA is a Cochrane Editor for the Cochrane Heart Group, but was not involved in the editorial process for the present study.

Sara Tomassini: none to declare

Marius Roman: none to declare

Angelica Rizzello: none to declare

Suraj Pathak: none to declare

Joussi Ramzi: none to declare

Carla Lucarelli: none to declare

Georgia Layton: none to declare

Ayesha Butt: none to declare

Florence Lai: none to declare

Tracy Kumar: none to declare

Marcin J Wozniak: My institution received funding from BHF, Hart Link, Van Geest and Zimmer Biomet.

Gavin Murphy: receiving research funding from Zimmer Biomet and support for educational activities from Vascutek. This author was also the research lead and senior author for one of the included studies (Kumar 2020), and therefore was not involved in making study eligibility decisions about, extracting data from, carrying out the risk of bias assessment for, or performing GRADE assessments for this specific trial.

New