Antioxidants for adults with chronic kidney disease

Abstract

Background

Chronic kidney disease (CKD) is a significant risk factor for cardiovascular disease (CVD) and death. Increased oxidative stress in people with CKD has been implicated as a potential causative factor. Antioxidant therapy decreases oxidative stress and may consequently reduce cardiovascular morbidity and death in people with CKD. This is an update of a Cochrane review first published in 2012.

Objectives

To examine the benefits and harms of antioxidant therapy on death and cardiovascular and kidney endpoints in adults with CKD stages 3 to 5, patients undergoing dialysis, and kidney transplant recipients.

Search methods

We searched the Cochrane Kidney and Transplant Register of Studies until 15 November 2022 using search terms relevant to this review. Studies in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE, conference proceedings, the International Clinical Trials Registry Platform (ICTRP) Search Portal, and ClinicalTrials.gov.

Selection criteria

We included all randomised controlled trials investigating the use of antioxidants, compared with placebo, usual or standard care, no treatment, or other antioxidants, for adults with CKD on cardiovascular and kidney endpoints.

Data collection and analysis

Titles and abstracts were screened independently by two authors who also performed data extraction using standardised forms. Results were pooled using random effects models and expressed as risk ratios (RR) or mean difference (MD) with 95% confidence intervals (CI). Confidence in the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

Main results

We included 95 studies (10,468 randomised patients) that evaluated antioxidant therapy in adults with non‐dialysis‐dependent CKD (31 studies, 5342 patients), dialysis‐dependent CKD (41 studies, 3444 patients) and kidney transplant recipients (21 studies, 1529 patients). Two studies enrolled dialysis and non‐dialysis patients (153 patients). Twenty‐one studies assessed the effects of vitamin antioxidants, and 74 assessed the effects of non‐vitamin antioxidants. Overall, the quality of included studies was moderate to low or very low due to unclear or high risk of bias for randomisation, allocation concealment, blinding, and loss to follow‐up.

Compared with placebo, usual care, or no treatment, antioxidant therapy may have little or no effect on cardiovascular death (8 studies, 3813 patients: RR 0.94, 95% CI 0.64 to 1.40; I² = 33%; low certainty of evidence) and probably has little to no effect on death (any cause) (45 studies, 7530 patients: RR 0.95, 95% CI 0.82 to 1.11; I² = 0%; moderate certainty of evidence), CVD (16 studies, 4768 patients: RR 0.79, 95% CI 0.63 to 0.99; I² = 23%; moderate certainty of evidence), or loss of kidney transplant (graft loss) (11 studies, 1053 patients: RR 0.88, 95% CI 0.67 to 1.17; I² = 0%; moderate certainty of evidence).

Compared with placebo, usual care, or no treatment, antioxidants had little to no effect on the slope of urinary albumin/creatinine ratio (change in UACR) (7 studies, 1286 patients: MD ‐0.04 mg/mmol, 95% CI ‐0.55 to 0.47; I² = 37%; very low certainty of evidence) but the evidence is very uncertain. Antioxidants probably reduced the progression to kidney failure (10 studies, 3201 patients: RR 0.65, 95% CI 0.41 to 1.02; I² = 41%; moderate certainty of evidence), may improve the slope of estimated glomerular filtration rate (change in eGFR) (28 studies, 4128 patients: MD 3.65 mL/min/1.73 m², 95% CI 2.81 to 4.50; I² = 99%; low certainty of evidence), but had uncertain effects on the slope of serum creatinine (change in SCr) (16 studies, 3180 patients: MD ‐13.35 µmol/L, 95% CI ‐23.49 to ‐3.23; I² = 98%; very low certainty of evidence).

Possible safety concerns are an observed increase in the risk of infection (14 studies, 3697 patients: RR 1.30, 95% CI 1.14 to 1.50; I² = 3%; moderate certainty of evidence) and heart failure (6 studies, 3733 patients: RR 1.40, 95% CI 1.11 to 1.75; I² = 0; moderate certainty of evidence) among antioxidant users. Results of studies with a low risk of bias or longer follow‐ups generally were comparable to the main analyses.

Authors' conclusions

We found no evidence that antioxidants reduced death or improved kidney transplant outcomes or proteinuria in patients with CKD. Antioxidants likely reduce cardiovascular events and progression to kidney failure and may improve kidney function. Possible concerns are an increased risk of infections and heart failure among antioxidant users. However, most studies were of suboptimal quality and had limited follow‐up, and few included people undergoing dialysis or kidney transplant recipients. Furthermore, the large heterogeneity in interventions hampers drawing conclusions on the efficacy and safety of individual agents.

Plain language summary

Antioxidants for adults with chronic kidney disease

People with chronic kidney disease have a high risk of early death, cardiovascular disease (heart disease and stroke), or kidney failure (dialysis or kidney transplantation). Antioxidants, like vitamin supplements, may be an easily available intervention to reduce these high risks.

What did we do?

We searched the literature up until November 2022 and assessed the effects of antioxidants on death, cardiovascular disease, kidney disease, and loss of kidney transplants. We determined the quality of studies and combined their results to estimate the effects of antioxidant supplements.

What did we find?

We included 95 studies with 10,468 adult patients, which tested 49 different antioxidants. Antioxidants did not reduce the risk of death or the loss of a kidney transplant. Antioxidants probably reduced the risk of heart disease and stroke and the risk of kidney failure (which would require dialysis). Antioxidants may also improve kidney function. However, we also observed an increase in the risk of heart failure and infections from antioxidants. Most studies were of poor quality. Therefore, better studies are needed to confirm the possible harms and benefits of antioxidants.

Conclusions

In adults with chronic kidney disease, antioxidants did not reduce the risk of death but probably reduced the risk of cardiovascular disease and kidney failure and improved kidney function. However, antioxidants might increase the risk of heart failure and infections.

Summary of findings

Summary of findings 1. Antioxidants versus control for adults with chronic kidney disease.

Antioxidants versus control for adults with chronic kidney disease
Patient or population: adults with CKD Settings: all settings Intervention: antioxidants Comparison: control (placebo, usual care, no treatment)
Outcomes	Anticipated absolute effects (95% CI)		Relative effect (95% CI)	No. of participants (RCTs)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	Control	Antioxidants
Cardiovascular death Follow‐up: 12 weeks to 2.25 years (median 9.8 months)	All adults (CKD 3‐5, CKD 5D, CKD 5‐transplant)		RR 0.94 (0.64 to 1.40)	3813 (8)	Low5 ⊕⊕⊝⊝
	59 per 1000	56 per 1000 (38 to 83)
	CKD 3‐5		RR 0.89 (0.46 to 1.74)	3269 (3)	Very low1,2,3 ⊕⊝⊝⊝
	53 per 1000	47 per 1000 (24 to 91)
	CKD 5D		RR 0.84 (0.49 to 1.44)	444 (4)	Moderate1 ⊕⊕⊕⊝
	119 per 1000	100 per 1000 (59 to 172)
	CKD 5‐transplant		RR 5.00 (0.25 to 101.58)	100 (1)	Very low1,7 ⊕⊝⊝⊝
	0 per 1000	0 per 1000 (0 to 0)
Death (any cause) Follow‐up: 1 month to 5 years (median 6 months)	All adults (CKD 3‐5, CKD 5D, CKD 5‐transplant)		RR 0.95 (0.82 to 1.11)	7530 (45)	Moderate1 ⊕⊕⊕⊝
	79 per 1000	75 per 1000 (65 to 87)
	CKD 3‐5		RR 0.1.00 (0.82 to 1.20)	4229 (13)	Low5 ⊕⊕⊝⊝
	77 per 1000	77 per 1000 (63 to 93)
	CKD 5D		RR 0.92 (0.69 to 1.22)	2261 (21)	Moderate1 ⊕⊕⊕⊝
	81 per 1000	75 per 1000 (59 to 99)
	CKD 5‐transplant		RR 0.82 (0.53 to 1.28)	1040 (11)	Very low3,5 ⊕⊝⊝⊝
	80 per 1000	65 per 1000 (42 to 102)
Cardiovascular disease Follow‐up: 12 weeks to 2 years (median 11 months)	All adults (CKD 3‐5, CKD 5D, CKD 5‐transplant)		RR 0.79 (0.63 to 0.99)	4768 (16)	Moderate1 ⊕⊕⊕⊝
	120 per 1000	95 per 1000 (76 to 119)
	CKD 3‐5		RR 0.76 (0.59 to 0.99)	4050 (8)	Moderate1 ⊕⊕⊕⊝
	126 per 1000	96 per 1000 (75 to 125)
	CKD 5D		RR 0.87 (0.47 to 1.62)	557 (6)	Low1,3 ⊕⊕⊝⊝
	104 per 1000	91 per 1000 (49 to 169)
	CKD 5‐transplant		RR 5.00 (0.25 to 101.58)	161 (2)	Very low5,7 ⊕⊝⊝⊝
	0 per 1000	0 per 1000 (0 to 0)
Heart failure Follow‐up: 16 weeks to 2 years (median 12 months)	All adults (CKD 3‐5, CKD 5D, CKD 5‐transplant)		RR 1.40 (1.11 to 1.76)	3733 (6)	Moderate1 ⊕⊕⊕⊝
	62 per 1000	86 per 1000 (68 to 108)
	CKD 3‐5		RR 1.40 (1.11 to 1.76)	3733 (6)	Moderate1 ⊕⊕⊕⊝
	62 per 1000	86 per 1000 (68 to 108)
	CKD 5D		‐	‐	No evidence
	‐	‐
	CKD 5‐transplant		‐	‐	No evidence
	‐	‐
Kidney failure Follow‐up: 3 months to 5 years (median 12 months)	All adults (CKD 3‐5, CKD 5‐transplant)		RR 0.65 (0.41 to 1.02)	3201 (10)	Moderate1 ⊕⊕⊕⊝
	77 per 1000	50 per 1000 (31 to 78)
	CKD 3‐5		RR 0.68 (0.40 to 1.17)	3024 (9)	Moderate1 ⊕⊕⊕⊝
	68 per 1000	46 per 1000 (27 to 80)
	CKD 5‐transplant		RR 0.53 (0.26 to 1.11)	177 (1)	Low1,3 ⊕⊕⊝⊝
	208 per 1000	110 per 1000 (54 to 231)
Change in eGFR (mL/min/1.73 m²) Follow‐up: 2 days to 2 years (median 6 months)	All adults (CKD 3‐5, CKD 5‐transplant)		‐	4128 (28)	Very low1,4,6 ⊕⊝⊝⊝
	The mean change in eGFR was 3.65 higher with antioxidants than control (2.81 higher to 4.50 higher)
	CKD 3‐5		‐	4057 (26)	Very low1,4,6 ⊕⊝⊝⊝
	The mean change in eGFR was 3.72 higher with antioxidants than control (2.85 higher to 4.58 higher)
	CKD 5‐transplant		‐	71 (2)	Very low1,2,3,4 ⊕⊝⊝⊝
	The mean change in eGFR was 3.97 higher with antioxidants than control (4.18 lower to 12.13 higher)
Graft loss Follow‐up: 1 month to 5 years (median 12 months)	CKD 5‐transplant		RR 0.88 (0.67 to 1.17)	1053 (11)	Moderate1 ⊕⊕⊕⊝
	155 per 1000	136 per 1000
Infection Follow‐up: 7 days to 3 years (median 12 months)	All adults (CKD 3‐5, CKD 5‐transplant)		RR 1.30 (1.14 to 1.50)	3697 (14)	Moderate1 ⊕⊕⊕⊝
	189 per 1000	246 per 1000 (216 to 284)
	CKD 3‐5		RR 1.46 (1.28 to 1.66)	2802 (6)	Moderate1 ⊕⊕⊕⊝
	203 per 1000	297 per 1000 (260 to 337)
	CKD 5D		RR 1.39 (0.73 to 2.65)	616 (5)	Very low1,7 ⊕⊝⊝⊝
	51 per 1000	72 per 1000 (253 to 418)
	CKD 5‐transplant		RR 1.00 (0.78 to 1.29)	279 (3)	Moderate1 ⊕⊕⊕⊝
	324 per 1000	324 per 1000 (253 to 418)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CKD: chronic kidney disease; CI: confidence interval; RR: risk ratio; eGFR: estimated glomerular filtration rate
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1 Evidence certainty was downgraded by one level due to risk of bias

2 Evidence certainty was downgraded by one level due to inconsistency (due to substantial heterogeneity (I² ranged from 50% to 90%)

3 Evidence certainty was downgraded by one level due to imprecision in CIs

4 Evidence certainty was downgraded by one level due to possible publication bias

5 Evidence certainty was downgraded by two levels due to risk of bias and difference in estimates in analysis with all studies and analysis with only studies with low risk of bias

6 Evidence certainty was downgraded by two levels for inconsistency due to considerable heterogeneity (I² ≥ 90%)

7 Evidence certainty was downgraded by two levels due to very large imprecision in CIs

Background

Description of the condition

Chronic kidney disease (CKD) is a major public health problem with an estimated prevalence of 700 million patients, causing 1.2 million deaths every year (GBD 2020). People with CKD are at high risk of premature cardiovascular disease (CVD) and death (Gansevoort 2013). This elevated risk already applies to people with a modest decline in estimated glomerular filtration rate (eGFR) or an increase in albuminuria, and the risk magnitude increases as people progress to more advanced stages of CKD (Gansevoort 2013). The absolute CVD risk is markedly higher among people with kidney failure compared to the general population, particularly among younger people who have a several hundred‐fold increase in CVD risk (Foley 1998).

The aetiology of CVD in CKD is complex and remains relatively poorly understood. Increased CVD risk for people with CKD may result from both an increased prevalence of traditional cardiovascular risk factors like hypertension, smoking, and hyperglycaemia, as well as risk factors that are unique to CKD such as vascular calcification, oxidative stress, inflammation, and increased proteinuria (Jankowski 2021).

Description of the intervention

Antioxidants are a group of substances that reduce oxidative stress by scavenging reactive oxygen and nitrogen species (Hunyadi 2019). They encompass a wide range of compounds such as beta carotene, selenium, flavonoids, and vitamins A, C, and E that are present in dietary sources, mainly fruit and vegetables. Alternatively, antioxidants can be taken as dietary supplements, which usually are readily available as over‐the‐counter medication at local drugstores or pharmacies. The current guidelines of the Kidney Disease: Improving Global Outcomes (KDIGO), American Society for Cardiology, and European Society of Cardiology for the prevention of CVD do not make any recommendations about the use of antioxidants in people with CKD (Arnett 2019; KDIGO 2013; Visseren 2021)

How the intervention might work

Increased oxidative stress, defined by Jones 2006 as “a disruption of redox signalling and control”, has long been implicated as a factor in the pathophysiology of cardiovascular and kidney disease (Daenen 2019; Giugliano 1995; Singal 1981). Increased oxidative stress in cardiac and vascular myocytes may be a consequence of an increase in the formation of reactive oxygen species and a decrease in the antioxidant reserve, or both (Woo 2021). Reactive oxygen species cause endothelial dysfunction, reduce nitric oxide availability, augment cardiac remodelling, and disrupt the handling of calcium in cardiomyocytes. These disturbances could lead to, among others, atherosclerosis, cardiomyopathy, cardiac arrhythmias, and heart failure (Pizzino 2017; Senomer 2019).

A range of naturally occurring vitamins (such as vitamins A, C, and E) and other substances (N‐acetylcysteine) have demonstrated antioxidant activity in vitro and in vivo (Brunet 1995; Frei 1989; Islam 2000; Meydani 1994). Studies assessing the effects of antioxidant therapy for the primary or secondary prevention of CVD in the general population have not identified cardiovascular benefits (Leopold 2015, Bjelakovic 2012). These studies did not specifically assess the effects of antioxidants in people with CKD. It is possible that people with CKD, contrary to the general population, benefit from antioxidant therapy since they have a much higher cardiovascular risk and because increased oxidative stress is highly prevalent, particularly for people with kidney failure (Daenen 2019).

Why it is important to do this review

Cardiovascular risk management traditionally involves the treatment of risk factors such as hypertension, dyslipidaemia, and hyperglycaemia, as well as antiplatelet therapy or anticoagulants (Arnett 2019; Visseren 2021). For patients with CKD, cardiovascular risk management can be expanded with therapies to normalise anaemia and phosphate‐lowering therapies (KDIGO 2013a). Despite this wealth of options, new strategies to reduce the cardiovascular risk for people with CKD are urgently required as their CVD risk is persistently elevated, and the preferred strategy to reduce this risk remains unclear.

A strong scientific rationale exists for the efficacy of antioxidants in the CKD population, and the fact that this therapy is easily available makes it an attractive proposition. The first version of this review (Jun 2012) did not find conclusive evidence for the use of antioxidants in people with CKD but could only base their results on a limited number of studies (10 studies, 1979 patients) of suboptimal quality. In the meantime, new studies investigating antioxidants in people with CKD have been published. This updated review looked at a broader set of antioxidant therapies and investigated the effect of antioxidants on markers of kidney damage and kidney transplant outcomes.

Objectives

The aim of this review was to investigate the benefits and harms of antioxidant therapy on death and cardiovascular and kidney endpoints in patients with CKD stages 3 to 5, patients undergoing dialysis, and kidney transplant recipients.

Methods

Criteria for considering studies for this review

Types of studies

All randomised controlled trials (RCTs) and quasi‐RCTs (RCTs in which the randomisation sequence involves a non‐random component) assessing the use of antioxidants in adults with CKD were considered. Studies with sequential designs were excluded.

Types of participants

We included adults (18 years or older) with CKD stages 3 to 5 as defined by the KDIGO guidelines (KDIGO 2013a) or by the author, who were undergoing dialysis (CKD 5D) or who received a kidney transplant (CKD 5‐transplant).

Types of interventions

We included studies that actively compared antioxidant supplements with placebo, usual or standard care, or no treatment regardless of dosing schedule or route of administration. Studies of agents in which a major mechanism of action was not thought to be antioxidants, such as statins, or which studied the effect of dietary antioxidants were excluded. Comparisons investigated were:

• Vitamin A versus placebo, usual or standard care, or no treatment

• Vitamin C versus placebo, usual or standard care, or no treatment

• Vitamin E versus placebo, usual or standard care, or no treatment

• N‐acetylcysteine versus placebo, usual or standard care, or no treatment

• Beta carotene versus placebo, usual or standard care, or no treatment

• Flavonoids versus placebo, usual or standard care, or no treatment

• Pentoxifylline compared with placebo, usual or standard care, or no treatment

• Bardoxolone methyl compared with placebo, usual or standard care, or no treatment

• Selenium compared with placebo, usual or standard care, or no treatment

• Polyphenols compared with placebo, usual or standard care, or no treatment

• Coenzyme Q10 compared with placebo, usual or standard care, or no treatment

• Astaxanthin compared with placebo, usual or standard care, or no treatment

• Sylimarin compared with placebo, usual or standard care, or no treatment

• Curcumin or turmeric compared with placebo, usual or standard care, or no treatment

• Sodium nitrite compared with placebo, usual or standard care, or no treatment

• Reservatrol compared with placebo, usual or standard care, or no treatment

• Pomegranate juice compared with placebo, usual or standard care, or no treatment

• Zinc compared with placebo, usual or standard care, or no treatment

• Grape seed extract compared with placebo, usual or standard care, or no treatment

• Ginsenoside Rb compared with placebo, usual or standard care, or no treatment

• Any antioxidant versus placebo, usual or standard care, or no treatment

• Any combination of antioxidants versus placebo, usual or standard care, or no treatment.

Compared to the 2012 version of this review (Jun 2012), the antioxidants which were investigated were expanded to include pentoxifylline, bardoxolone methyl, selenium, polyphenols, coenzyme Q10, astaxanthin, sylimarin, curcumin or turmeric, sodium nitrite, reservatrol, pomegranate juice, zinc, grape seed extract, and ginsenoside Rb.

Types of outcome measures

We focussed on death, cardiovascular and kidney function endpoints. We also considered dialysis access endpoints for patients receiving dialysis and transplant endpoints for kidney transplant recipients. We included the outcome measured at the latest time point in the follow‐up in our meta‐analysis. We did not assess the effect of antioxidants on eGFR or serum creatinine (SCr) outcomes for patients receiving dialysis treatment since these outcomes are heavily affected by patients’ dialysis treatment (Shafi 2018).

Primary outcomes

• Cardiovascular death

• Death (any cause)

Secondary outcomes

• Cardiovascular outcomes

  • CVD defined as a composite of the following outcomes: fatal and non‐fatal myocardial infarction (MI), fatal and non‐fatal stroke and cardiovascular death, or as defined by study authors

  • Coronary heart disease: fatal and non‐fatal MI as well as coronary revascularisation, or as defined by authors

  • Cerebrovascular disease, including stroke (overall and by subtype), transient ischaemic attacks (TIA), and cerebrovascular revascularisation

  • Heart failure: hospitalisation for heart failure, development of heart failure (New York Heart Association (NYHA) class III or IV), or as defined by authors

  • Peripheral vascular disease: lower limb revascularisation or amputation

• Kidney outcomes

  • Kidney failure: requirement for kidney replacement therapy (KRT) or death due to kidney disease

  • eGFR at the end of the study

  • Mean change in eGFR

  • SCr at the end of the study

  • Change in SCr

  • Urinary albumin/creatinine ratio (UACR) at the end of the study

  • Change in UACR

  • Proteinuria at the end of the study

  • Change in proteinuria

  • Albuminuria at the end of the study

  • Change in albuminuria

• Dialysis outcomes

  • Thrombosis of vascular access

• Transplant outcomes

  • Graft loss

  • Delayed graft function (DGF)

  • Graft failure

• Adverse events

  • Malignancy (including haematological malignancies)

  • Major bleeding

  • Infection

Search methods for identification of studies

Electronic searches

We searched the Cochrane Kidney and Transplant Register of Studies up to 15 November 2022 through contact with the Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources.

1. Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)

2. Weekly searches of MEDLINE OVID SP

3. Searches of kidney and transplant journals and the proceedings and abstracts from major kidney and transplant conferences

4. Searching of the current year of EMBASE OVID SP

5. Weekly current awareness alerts for selected kidney and transplant journals

6. Searches of the International Clinical Trials Registry Platform (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of search strategies, as well as a list of handsearched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website under CKT Register of Studies.

See Appendix 1 for search terms used in strategies for this review.

Searching other resources

1. Reference lists of nephrology textbooks, review articles and relevant studies.

2. Letters seeking information about unpublished or incomplete studies to investigators known to be involved in previous studies.

Data collection and analysis

Selection of studies

The search strategy was used to obtain titles and abstracts of studies relevant to this review. Two authors (JC and RV) independently screened titles and abstracts and discarded studies that were not eligible based on study design, patient population, or interventions. Subsequently, two authors (JC and RV) screened the studies on full text to determine if the studies satisfied the inclusion criteria. Disagreements between authors were resolved by discussion. We planned to consult a third author if the two authors could not agree on the selection of studies.

Data extraction and management

Two authors (JC and RV) used standardised extraction forms to extract data from included studies. One author (JC) performed the data extraction, which was verified by a second author (RV). Publication in a non‐English language was not an exclusion criterion. Results from studies with multiple publications were combined. Disagreements were resolved by discussion and assessing the original publication.

Assessment of risk of bias in included studies

The following items were assessed using the risk of bias assessment tool (Higgins 2022) (see Appendix 2).

• Was there adequate sequence generation (selection bias)?

• Was allocation adequately concealed (selection bias)?

• Was knowledge of the allocated interventions adequately prevented during the study (detection bias)?

  • Participants and personnel

  • Outcome assessors

• Were incomplete outcome data adequately addressed (attrition bias)?

• Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?

• Was the study apparently free of other problems that could put it at risk of bias?

We assumed a low risk of bias for detection bias for all studies because all outcomes of interest were considered objective and unlikely to be affected by the blinding of the outcome assessor. We attempted to retrieve protocols for all included studies to determine reporting bias. Studies were only rated as low risk of bias if all extracted outcomes were reported in the protocol. The risk of reporting bias was rated as unclear if no protocol was available.

Measures of treatment effect

Dichotomous outcomes were expressed as risk ratios (RR) with 95% confidence intervals (CI) and continuous scales as mean difference (MD) with 95% CIs.

Unit of analysis issues

Analyses were performed using the intention‐to‐treat principle wherever possible. SCr was converted to µmol/L, UACR to mg/mmol, proteinuria to g/day, and albuminuria to mg/day where necessary. This review included studies with non‐standard study designs, such as cross‐over RCTs, factorial RCTs, and RCTs with more than two study arms.

Cross‐over RCTs

For cross‐over study designs, we solely included outcomes from the first treatment period to avoid carry‐over effects to subsequent periods.

Factorial RCT

For factorial study designs, we pooled the study arms which received antioxidant interventions as the intervention arm and the arms which did not receive antioxidants as the control group for the meta‐analysis.

Multi‐arm RCTs

For multi‐arm study designs, we extracted data from all study arms that met the inclusion criteria. If two or more arms in the same study received antioxidants, we pooled the results of the study arms that received antioxidants as the intervention arm for the meta‐analysis, if possible. If one study arm received a placebo and another study arm received usual care, we only included the arm receiving placebo in the meta‐analysis.

Studies with mixed populations

Studies which included both adults with and without CKD were included if they reported stratified results for adults with CKD.

Dealing with missing data

We attempted to contact the authors when data were missing from the publication, when outcomes were reported in insufficient detail to enable meta‐analyses, or when only a conference abstract was available. If the authors did not respond to our enquiry, we narratively described the outcomes in our review. We excluded studies if we were unable to extract relevant information without additional information from the authors and the authors did not respond to our correspondence.

We extracted data that was only presented in graphs to the best of our abilities. For continuous variables for which only the median and range or median and interquartile range were reported, we estimated the mean and standard deviation (SD) as described by Wan 2014. We calculated the MD for continuous variables, which were reported values at baseline and the end of the study. Missing SD were imputed as described by Follmann 1992 and Abrams 2005. We calculated weighted correlation coefficients if more than one study was described in sufficient detail to calculate correlation coefficients. If studies only reported the SD at baseline or the end of the study, we assumed the SD was the same at both times. If no study was described in sufficient detail to calculate a correlation coefficient, we conservatively assumed a correlation coefficient of 0.80.

Assessment of heterogeneity

Heterogeneity was analysed using a Chi² test on n‐1 degrees of freedom, with an alpha of 0.05 used for statistical significance and with the I² test (Higgins 2003). I² were interpreted as follows:

• 0% to 40%: might not be important

• 30% to 60%: may represent moderate heterogeneity

• 50% to 90%: may represent substantial heterogeneity

• 75% to 100%: considerable heterogeneity.

Assessment of reporting biases

We assessed the risk of small study bias using funnel plots. Small study bias was only assessed if we managed to include a minimum of 10 studies per outcome. We assessed publication bias for the individual subgroups if we included a minimum of 10 studies per CKD subgroup (CKD 3‐5, CKD 5D, and CKD 5‐transplant).

Data synthesis

Data were pooled in meta‐analyses using random‐effects models. All analyses were stratified by the stage of kidney disease (CKD 3‐5, CKD 5D, and CKD 5‐transplant).

Subgroup analysis and investigation of heterogeneity

Heterogeneity was analysed using the Cochran Q test on N‐1 degrees of freedom, with P < 0.05 used to denote statistical significance, and the I² test (with uncertainty intervals). We performed subgroup analyses to assess plausible variations in treatment effects based on the type of antioxidants, risk of bias and duration of follow‐up.

Subgroup analyses were conducted according to the following characteristics.

• Interventions with vitamins

• Interventions without vitamins

• Follow up ≥ six months

Plausible explanations for variations in treatment effect were explored using subgroup analyses based on study quality and length of follow‐up.

Sensitivity analysis

Sensitivity analyses were undertaken to assess the impact of individual studies on the overall results if significant evidence of heterogeneity was observed. Additionally, we performed a sensitivity analysis to explore the effects of imputing SD for change on our results and the effect of antioxidants in studies with a low risk of bias. Studies were classified as low risk of bias if they had used adequate methods for randomisation and blinding of participants and study personnel and were unlikely to have attrition bias.

Summary of findings and assessment of the certainty of the evidence

The main results of this review are presented in 'Summary of findings' tables. These tables include key information on the quality of the evidence, the magnitude of the effects of the interventions examined, and the sum of the available data for the main outcomes (Schünemann 2022a). The 'Summary of findings' table also includes an overall grading of the evidence for each of the main outcomes based on the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (GRADE 2008; GRADE 2011). The GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. The quality of a body of evidence involves consideration of the within‐trial risk of bias (methodological quality), directness of evidence, heterogeneity, the precision of effect estimates, and the risk of publication bias (Schünemann 2022b).

The following outcomes are presented in the summary of findings table.

• Death (any cause)

• Cardiovascular death

• CVD

• Heart failure

• Kidney failure

• Change in eGFR

• Graft loss

• Infection

Results

Description of studies

The following section contains broad descriptions of the studies considered in this review. For further details on each individual study, please see the characteristics of studies tables.

Results of the search

2012 review

In the 2012 review (Jun 2012), 1080 reports were identified, and after duplicates were removed, 753 reports were screened. Ten studies (12 reports, 1979 patients) were included, 12 studies (13 reports) were excluded, and one ongoing study was identified. Reasons for exclusion were wrong study design, wrong population, wrong intervention, and no outcomes of interest.

2023 update

For this 2023 update, we searched the Cochrane Kidney and Transplant Register of Studies and identified 557 reports. After title and abstract screening, 404 reports were potentially eligible. Seventy‐nine new studies (142 reports) were included, 124 studies (185 reports) were excluded, and we identified 15 ongoing studies (17 reports). We also identified 59 reports of previously included and excluded studies (182 reports) (see Figure 1).

1.

Flow chart: 2023 review update study selection

We reassessed all previous studies. We retained nine included studies (BEAM 2011; HOPE 1996; Schneeberger 1990; Pollak 1993; Rabl 1993; Shoskes 2005; Singh 2000; SPACE 2000; Tepel 2003). Wijnen 2002 was excluded after closer examination because the study population did not include patients with CKD but patients undergoing infrarenal aneurysm surgery who had a mean creatinine clearance (CrCl) > 85 mL/min/1.73 m². (Figure 1). We reclassified reports of six previously excluded studies as included studies (ATIC 2005; Blackhall 2005; Noel 1997; Perkins 2009; Schramm 2002; Wlodarczyk 2000) because they reported outcomes of interest, and one ongoing study has now been included (BEACON 2013). Two non‐randomised studies have been deleted.

We contacted the authors of 24 studies via email (AIONID 2013; Aminorroaya 2005; Argani 2014; Bahmani 2016a; Biniaz 2014; Borges 2016; Fallahzadeh 2012; Friedman 2003; Gholnari 2018; Lin 2008; Lu 2007; Orban 2015; PREDIAN 2011; Rabizadeh 2018; Roozbeh 2009; Sattarinezhad 2019; Sengupta 2022; Shahidi 2015; Shema‐Didi 2012; Singer 2011; Soliman 2014; Voroneanu 2017; Zachara 2009; Zahed 2016) and received unpublished data from seven (AIONID 2013; Borges 2016; Friedman 2003; Lu 2007; Orban 2015; PREDIAN 2011; Rabizadeh 2018).

In total, we have included 95 studies (222 reports) and excluded 147 studies (189 reports). There are 15 ongoing studies, which will be assessed in a future update of this review.

Included studies

We included 95 studies (222 reports, 10,468 randomised patients).

There were two cross‐over RCTs (Blackhall 2005; Rivara 2015) and four factorial RCTs (AIONID 2013; Asemi 2016; Mori 2009; Shojaei 2011). The remaining studies were parallel RCTs. Eight studies were 3‐arm RCTs (Biniaz 2014; CoQ10 Biomarker 2016; Pergola 2009a; Sahraei 2015; Shoskes 2005; Song 2006a; Tonelli 2015; Yan 2017a), and two were 4‐arm RCTs (Bansal 2017; BEAM 2011). The remaining studies had two arms. Four studies were published only as conference abstracts (Bansal 2017; Martinez 2020; Pergola 2009a; Soliman 2014). Fourteen studies were not included in the meta‐analyses (Argani 2014; Biniaz 2014; Jern 2000; Guo 2013; Firuzi 2016; Friedman 2003; Gonzalez‐Espinoza 2012; Paniagua 1995; Roozbeh 2009; Sengupta 2022; Shojaei 2011; Singh 2000; Thaha 2009; Zachara 2009) because they measured eGFR or SCr in patients undergoing dialysis, only described their outcomes narratively, or both.

Twenty‐five studies were conducted in North America (AIONID 2013; Attallah 2006; Bansal 2017; BEAM 2011; Jern 2000; CoQ10 Biomarker 2016; DeVault 1994; Friedman 2003; Gonzalez‐Espinoza 2012; Himmelfarb 2007; Jimenez‐Osorio 2016; Lu 2007; Martinez 2020; Moist 2010; Moreillon 2013; NICE 2020; Paniagua 1995; PATH 2014; Pergola 2009a; Perkins 2009; Pollak 1993; Rivara 2015; Shoskes 2005; Tonelli 2015; Vincenti 1996), 20 in Europe (ATIC 2005; Demir 2006a; Giliberti 2022; Goicoechea 2012; Konigsrainer 1995; Navarro 1999; Navarro 1999a; Noel 1997; Norio 2003; Oko 1999; Orban 2015; PREDIAN 2011; Rabl 1993; Rassaf 2016; Schneeberger 1990; Schramm 2002; Tepel 2003; Voroneanu 2017; Wlodarczyk 2000; Zachara 2009), 22 in the Middle East (Atapour 2022; Ahmadi 2017; Argani 2014; Asemi 2016; Biniaz 2014; Firuzi 2016; Gholipour Baradari 2011; Haddadian‐Khouzani 2022; Hajian 2022; Hosseini 2021; Khabbazi 2012; Modarresi 2018; Modarresi 2017; Omar 2022; Rabizadeh 2018; Roozbeh 2009; Sahraei 2015; Salehi 2013; Shema‐Didi 2012; Shojaei 2011; SPACE 2000; Vafadar Afshar 2020), 13 in Asia (ACTIVE 2014; Guo 2013; Lin 2008; Okamoto 2020; Sengupta 2022; Singh 2000; Song 2006a; Tan 2019; Thaha 2009; TSUBAKI 2020; Xu 2017; Yan 2017a; Yang 2019a), five in Oceania (Blackhall 2005; HERO 2008; Mori 2009; Singer 2011; XANTHIN 2008), four in South America (Antunes 2014; Borges 2016; Danilovic 2011; Silveira 2019), four in Africa (Bejaoui 2022; Hajji 2021; Soliman 2014; Turki 2016), and two were multi‐continental (BEACON 2013; HOPE 1996).

Participants

Of included studies, 31 included patients with CKD 3‐5 (ATIC 2005; Bansal 2017; BEACON 2013; BEAM 2011; Bejaoui 2022; Borges 2016; Giliberti 2022; Goicoechea 2012; HOPE 1996; Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Moist 2010; Mori 2009; Navarro 1999; Navarro 1999a; NICE 2020; Pergola 2009a; Perkins 2009; PREDIAN 2011; Rabizadeh 2018; Sengupta 2022; Silveira 2019; Song 2006a; Tan 2019; Thaha 2009; TSUBAKI 2020; Turki 2016; Voroneanu 2017; Xu 2017; Yan 2017a) (5342 randomised patients), 41 included patients undergoing dialysis (CKD 5D) (ACTIVE 2014; Ahmadi 2017; AIONID 2013; Antunes 2014; Argani 2014; Asemi 2016; Atapour 2022; Attallah 2006; Biniaz 2014; CoQ10 Biomarker 2016; Firuzi 2016; Friedman 2003; Gholipour Baradari 2011; Gonzalez‐Espinoza 2012; Guo 2013; Haddadian‐Khouzani 2022; Hajian 2022; Hajji 2021; Himmelfarb 2007; Hosseini 2021; Jern 2000; Khabbazi 2012; Lu 2007; Moreillon 2013; Okamoto 2020; Omar 2022; Paniagua 1995; PATH 2014; Rassaf 2016; Rivara 2015; Roozbeh 2009; Salehi 2013; Shema‐Didi 2012; Shojaei 2011; Singh 2000; SPACE 2000; Tepel 2003; Tonelli 2015; Vafadar Afshar 2020; Yang 2019a; Zachara 2009) (3444 randomised patients), and 21 included kidney transplant recipients (Blackhall 2005; Danilovic 2011; Demir 2006a; DeVault 1994; Konigsrainer 1995; Modarresi 2017; Modarresi 2018; Noel 1997; Norio 2003; Oko 1999; Orban 2015; Pollak 1993; Rabl 1993; Sahraei 2015; Schneeberger 1990; Schramm 2002; Shoskes 2005; Soliman 2014; Vincenti 1996; Wlodarczyk 2000; XANTHIN 2008) (1529 randomised patients). Two studies included both patients with CKD 3‐5 and patients undergoing dialysis (HERO 2008; Singer 2011) (153 randomised patients; CKD 3‐5: 25, CKD 5D: 128).

The number of randomised patients ranged from 10 to 2185 (median 62) and 30 studies randomised fewer than 50 patients (Attallah 2006; Bejaoui 2022; Blackhall 2005; Borges 2016; Jern 2000; Demir 2006a; DeVault 1994; Friedman 2003; Gonzalez‐Espinoza 2012; Hajian 2022; Hosseini 2021; Lu 2007; Modarresi 2018; Moreillon 2013; Navarro 1999; Navarro 1999a; Oko 1999; Paniagua 1995; Perkins 2009; Rabizadeh 2018; Rabl 1993; Rivara 2015; Shoskes 2005; Silveira 2019; Singh 2000; Soliman 2014; Thaha 2009; Turki 2016; Vincenti 1996; Wlodarczyk 2000).

The mean age ranged from 25 to 75 years (median 58 years). In six studies, mean age was below 40 years (Demir 2006a; Gonzalez‐Espinoza 2012; Noel 1997; Paniagua 1995; Sahraei 2015; Vincenti 1996), and in 30 studies the mean age was over 60 years (Bansal 2017; BEACON 2013; BEAM 2011; Biniaz 2014; Borges 2016; Friedman 2003; Gholipour Baradari 2011; Goicoechea 2012; HERO 2008; HOPE 1996; Moist 2010; Navarro 1999; Navarro 1999a; NICE 2020; Okamoto 2020; Pergola 2009a; Perkins 2009; PREDIAN 2011; Rassaf 2016; Shema‐Didi 2012; Silveira 2019; Singer 2011; SPACE 2000; Tan 2019; Tepel 2003; TSUBAKI 2020; Turki 2016; Voroneanu 2017; Yang 2019a; Zachara 2009).

Overall, the population of the included studies comprised more males than females (59% of all patients were male). In 10 studies, more than 70% of patients were male (Bansal 2017; HOPE 1996; Moist 2010; Modarresi 2018; Rabl 1993; Rassaf 2016; Schneeberger 1990; Singh 2000; Tonelli 2015; XANTHIN 2008). Baseline eGFR in patients with CKD 3‐5 ranged from 19 to 64 mL/min/1.73 m² (median 40 mL/min/1.73 m² ). The mean baseline eGFR was below 30 mL/min/1.73 m² in four studies (BEACON 2013; Navarro 1999; Navarro 1999a; Yan 2017a).

Interventions

Forty‐nine different interventions were studied. The duration of the studies ranged from two days to five years (median follow‐up four months), and 37 studies had a follow‐up of six months or more (ACTIVE 2014; ATIC 2005; Attallah 2006; BEACON 2013; BEAM 2011; Bejaoui 2022; Blackhall 2005; Danilovic 2011; Giliberti 2022; Goicoechea 2012; HOPE 1996; Lin 2008; Schramm 2002; Lu 2007; Navarro 1999; Navarro 1999a; Noel 1997; Norio 2003; Okamoto 2020; Orban 2015; PATH 2014; Perkins 2009; PREDIAN 2011; Schneeberger 1990; Sengupta 2022; Shema‐Didi 2012; Silveira 2019; Song 2006a; SPACE 2000; Tan 2019; Tepel 2003; Tonelli 2015; Turki 2016; Vincenti 1996; Voroneanu 2017; XANTHIN 2008; Xu 2017).

Vitamin interventions

Twenty‐one studies compared antioxidants with vitamins with placebo, usual care, or no treatment (ACTIVE 2014; Asemi 2016; ATIC 2005; Attallah 2006; Biniaz 2014; Blackhall 2005; Gholipour Baradari 2011; Giliberti 2022; Hajian 2022; Himmelfarb 2007; HOPE 1996; Lu 2007; Omar 2022; Norio 2003; Singer 2011; SPACE 2000; Tan 2019; Tonelli 2015; Yang 2019a; PATH 2014; Sahraei 2015).

• Vitamin C

  • Vitamin C (8 studies) (Attallah 2006; Biniaz 2014; Gholipour Baradari 2011; Hajian 2022; Norio 2003; Singer 2011; Tan 2019; Yang 2019a)

  • Vitamin C plus iron and erythropoietin (1 study) (ACTIVE 2014)

  • Vitamin C plus n‐acetylcysteine (1 study) (Sahraei 2015).

  • Vitamin C plus rutin (1 study) (Omar 2022)

  • Vitamin C plus ferric sodium EDTA, folic acid, copper gluconate, zinc gluconate, and selenomethionine (1 study) (Giliberti 2022)

• Vitamin E

  • Vitamin E (4 studies) (Asemi 2016; HOPE 1996; Lu 2007; SPACE 2000)

  • Vitamin E plus docosahexaenoic acid (1 study) (Himmelfarb 2007)

  • Vitamin E plus pravastatin and vitamin B6 (1 study) (ATIC 2005)

  • Vitamin E plus alpha‐lipoic acid (1 study) (PATH 2014)

  • Vitamin E plus zinc, selenium and standard kidney vitamins (1 study) (Tonelli 2015)

• Vitamin C plus vitamin E

  • Vitamin C plus vitamin E and beta‐carotene (1 study) (Blackhall 2005)

Non‐vitamin interventions

Seventy‐four studies compared antioxidants without vitamins with placebo, no treatment, or usual care (Ahmadi 2017; AIONID 2013; Antunes 2014; Argani 2014; Atapour 2022; Bansal 2017; BEACON 2013; BEAM 2011; Bejaoui 2022; Borges 2016; Jern 2000; CoQ10 Biomarker 2016; Danilovic 2011; Demir 2006a; DeVault 1994; Firuzi 2016; Friedman 2003; Goicoechea 2012; Gonzalez‐Espinoza 2012; Guo 2013; Haddadian‐Khouzani 2022; Hajji 2021; HERO 2008; Hosseini 2021; Jimenez‐Osorio 2016; Khabbazi 2012; Konigsrainer 1995; Lin 2008; Schramm 2002; Martinez 2020; Modarresi 2018; Modarresi 2017; Moist 2010; Moreillon 2013; Mori 2009; Navarro 1999; Navarro 1999a; NICE 2020; Noel 1997; Okamoto 2020; Oko 1999; Orban 2015; Paniagua 1995; Pergola 2009a; Perkins 2009; Pollak 1993; PREDIAN 2011; Rabizadeh 2018; Rabl 1993; Rassaf 2016; Rivara 2015; Roozbeh 2009; Salehi 2013; Schneeberger 1990; Sengupta 2022; Shema‐Didi 2012; Shojaei 2011; Shoskes 2005; Silveira 2019; Singh 2000; Soliman 2014; Song 2006a; Tepel 2003; Thaha 2009; TSUBAKI 2020; Turki 2016; Vafadar Afshar 2020; Vincenti 1996; Voroneanu 2017; Wlodarczyk 2000; XANTHIN 2008; Xu 2017; Yan 2017a; Zachara 2009).

• Pentoxifylline

  • Pentoxifylline (16 studies) (Antunes 2014; Demir 2006a; DeVault 1994; Goicoechea 2012; Gonzalez‐Espinoza 2012; HERO 2008; Konigsrainer 1995; Lin 2008; Navarro 1999; Navarro 1999a; Noel 1997; Oko 1999; Perkins 2009; PREDIAN 2011; Vincenti 1996; Wlodarczyk 2000)

  • Pentoxifylline plus losartan (1 study) (AIONID 2013)

  • Pentoxifylline plus oral nutritional supplements (1 study) (Rabizadeh 2018)

• N‐acetylcysteine

  • N‐acetylcysteine (11 studies: Ahmadi 2017; Danilovic 2011; Friedman 2003; Modarresi 2018; Modarresi 2017; Moist 2010; Orban 2015; Sahraei 2015; Soliman 2014; Tepel 2003; Yan 2017a)

  • N‐acetylcysteine plus angiotensin‐converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARB) (1 study) (Thaha 2009)

  • N‐acetylcysteine plus silybin (1 study) (Bansal 2017)

  • N‐acetylcysteine plus taurine (1 study) (Sengupta 2022)

  • N‐acetylcysteine plus pyridoxamine dihydrochloride (1 study) (Sengupta 2022)

• Zinc

  • Zinc sulphate (6 studies) (Argani 2014; Hajji 2021; Hosseini 2021; Jern 2000; Paniagua 1995; Roozbeh 2009)

  • o Zinc gluconate (1 study) (Haddadian‐Khouzani 2022)

  • Zinc (1 study) (Guo 2013)

• Bardoxolone methyl (4 studies) (BEACON 2013; BEAM 2011; Pergola 2009a; TSUBAKI 2020)

• Coenzyme Q10

  • Coenzyme Q10 (2 studies) (CoQ10 Biomarker 2016; Singh 2000)

  • Coenzyme Q10 plus carnitine (1 study) (Shojaei 2011)

  • Coenzyme Q10 plus omega‐3 fatty acids (1 study) (Mori 2009)

• Curcumin

  • Curcumin (2 studies) (Jimenez‐Osorio 2016; Martinez 2020)

  • Curcumin plus Boswellia serrata (1 study) (Moreillon 2013)

  • Curcumin plus quercetin (1 study) (Shoskes 2005)

  • Nanocurcumin (1 study) (Vafadar Afshar 2020)

• Grape seed

  • Grape seed extract (1 study) (Turki 2016)

  • Grape seed flour (1 study) (Bejaoui 2022)

• Alpha‐lipoic acid (2 studies) (Khabbazi 2012; Song 2006a)

• Human recombinant superoxide dismutase (2 studies) (Pollak 1993; Schneeberger 1990)

• Selenium (2 studies) (Salehi 2013; Zachara 2009)

• Silymarin (2 studies) (Firuzi 2016; Voroneanu 2017)

• Cocoa‐flavonol (1 study) (Rassaf 2016)

• Ginsenoside Rb (1 study) (Xu 2017)

• Green tea polyphenols (1 study) (Borges 2016)

• L‐arginine (1 study) (Schramm 2002)

• Omnibiota (1 study) (Rabl 1993)

• Pomegranate juice (1 study) (Shema‐Didi 2012)

• Propolis (1 study) (Silveira 2019)

• Sodium nitrite plus isoquercetin (1 study) (NICE 2020)

We included two studies that compared antioxidant therapies head to head.

• Pomegranate juice versus pomegranate extract (1 study) (Rivara 2015)

• Zinc acetate versus polaprezinc (1 study) (Okamoto 2020)

Outcomes

Cardiovascular death

Nine studies reported cardiovascular death (BEACON 2013; Goicoechea 2012; HERO 2008; HOPE 1996; Konigsrainer 1995; Lu 2007; Salehi 2013; SPACE 2000; Tepel 2003).

Death (any cause)

Fifty‐one studies reported death (any cause) (ACTIVE 2014; Ahmadi 2017; AIONID 2013; Argani 2014; Asemi 2016; Atapour 2022; BEACON 2013; BEAM 2011; Biniaz 2014; CoQ10 Biomarker 2016; Danilovic 2011; DeVault 1994; Gholipour Baradari 2011; Goicoechea 2012; Haddadian‐Khouzani 2022; Hajji 2021; HERO 2008; Himmelfarb 2007; HOPE 1996; Khabbazi 2012; Konigsrainer 1995; Lu 2007; Modarresi 2017; Moreillon 2013; Mori 2009; NICE 2020; Noel 1997; Norio 2003; Okamoto 2020; Omar 2022; Orban 2015; PATH 2014; PREDIAN 2011; Rivara 2015; Roozbeh 2009; Salehi 2013; Schneeberger 1990; Schramm 2002; Shema‐Didi 2012; Silveira 2019; SPACE 2000; Tan 2019; Tepel 2003; Tonelli 2015; TSUBAKI 2020; Vafadar Afshar 2020; Vincenti 1996; Voroneanu 2017; XANTHIN 2008; Yan 2017a; Yang 2019a).

Four studies could not be included in the meta‐analysis because the authors only provided the total number of events, without stratification per study arm, and did not respond to our enquiries for further information (Argani 2014; Biniaz 2014; Moreillon 2013; Roozbeh 2009).

Cardiovascular disease

Seventeen studies reported CVD (ACTIVE 2014; ATIC 2005; Attallah 2006; BEACON 2013; BEAM 2011; HERO 2008HOPE 1996; Konigsrainer 1995; Lu 2007; PREDIAN 2011; Salehi 2013; Shema‐Didi 2012; Shojaei 2011; SPACE 2000; Tepel 2003; XANTHIN 2008; Yang 2019a). We observed heterogeneity across studies in their definition of CVD. Eleven studies did not specify their definition of CVD (ACTIVE 2014; ATIC 2005; Attallah 2006; BEAM 2011; HERO 2008; Konigsrainer 1995; Lu 2007; Salehi 2013; Shojaei 2011; XANTHIN 2008; Yang 2019a). Studies that did provide a definition for CVD include, among others, cardiovascular death, (non‐fatal) MI, (non‐fatal) stroke, hospitalisation for heart failure, peripheral vascular disease (revascularisation or amputation), unstable angina, and coronary angioplasty or bypass surgery. One study could not be included in the meta‐analyses because the authors only provided the total number of events, without stratification per study arm, and did not respond to our enquiries for further information (Shojaei 2011).

Coronary heart disease

Fourteen studies reported coronary heart disease (ATIC 2005; Attallah 2006; BEACON 2013; BEAM 2011; Jern 2000; HOPE 1996; Konigsrainer 1995; Lu 2007; PREDIAN 2011; Salehi 2013; Singer 2011; SPACE 2000; Tepel 2003; TSUBAKI 2020). There were no studies reporting coronary heart disease in kidney transplant recipients. One study could not be included in the meta‐analyses because the authors only provided the total number of events, without stratification per study arm, and did not respond to our enquiries for further information (Jern 2000).

Heart failure

Six studies reported heart failure (BEACON 2013; BEAM 2011; HOPE 1996; Perkins 2009; PREDIAN 2011; TSUBAKI 2020). There were no studies reporting results for people treated with dialysis or kidney transplant recipients. Most studies defined heart failure as hospitalisation for heart failure. HOPE 1996 required clinical and radiologic signs of congestion, and BEACON 2013 also included death due to heart failure in its definition.

Cerebrovascular disease

Eight studies reported cerebrovascular disease (ACTIVE 2014; ATIC 2005; HOPE 1996; Konigsrainer 1995; Salehi 2013; SPACE 2000; Tan 2019; Tepel 2003).

Peripheral arterial disease

Four studies reported peripheral arterial disease (HOPE 1996; PREDIAN 2011; SPACE 2000; Tepel 2003). There were no studies reporting peripheral arterial disease in kidney transplant recipients.

Kidney failure

Eleven studies reported progression to kidney failure (ATIC 2005; BEACON 2013; BEAM 2011; Goicoechea 2012; Perkins 2009; PREDIAN 2011; Schneeberger 1990; Singer 2011; Turki 2016; Voroneanu 2017; Yan 2017a). Most studies defined kidney failure as the initiation of dialysis or receiving a kidney transplant. Schneeberger 1990 defined kidney failure as progressive transplant dysfunction requiring dialysis, and Turki 2016 and Yan 2017a as an eGFR < 15 mL/min/1.73 m².

Glomerular filtration rate

Twenty‐eight studies reported eGFR at the end of the study (Ahmadi 2017; ATIC 2005; BEAM 2011; Bejaoui 2022; Borges 2016; Danilovic 2011; Giliberti 2022; Goicoechea 2012; Guo 2013; Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Modarresi 2018; Modarresi 2017; Mori 2009; Orban 2015; Perkins 2009; Pollak 1993; PREDIAN 2011; Rabizadeh 2018; Sahraei 2015; Silveira 2019; Singer 2011; Singh 2000; Tan 2019; Turki 2016; Voroneanu 2017; XANTHIN 2008), and eight studies reported CrCl at the end of study (Moist 2010; Navarro 1999; Navarro 1999a; Norio 2003; Rabl 1993; Schramm 2002; Soliman 2014; Xu 2017). Thirty‐one studies reported change in eGFR (Ahmadi 2017; ATIC 2005; Bansal 2017; BEACON 2013; BEAM 2011; Bejaoui 2022; Blackhall 2005; Borges 2016; Giliberti 2022; Goicoechea 2012; Guo 2013; Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Mori 2009; NICE 2020; Pergola 2009a; Perkins 2009; PREDIAN 2011; Rabizadeh 2018; Sengupta 2022; Silveira 2019; Singer 2011; Singh 2000; Song 2006a; Tan 2019; TSUBAKI 2020; Turki 2016; Voroneanu 2017; XANTHIN 2008; Yan 2017a), and four studies change in CrCl (Moist 2010; Navarro 1999; Navarro 1999a; Xu 2017). We calculated change scores and imputed SD (correlation coefficient = 0.98) for change in 20 studies (ATIC 2005; Bejaoui 2022; Borges 2016; Giliberti 2022; Goicoechea 2012; Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Mori 2009; Perkins 2009; PREDIAN 2011; Rabizadeh 2018; Silveira 2019; Singer 2011; Singh 2000; Tan 2019; Turki 2016; Voroneanu 2017; XANTHIN 2008; Xu 2017). Eight studies could not be included in the meta‐analyses because they only reported their results narratively (Bansal 2017, Soliman 2014; Song 2006a), only reported the percentage of change (Pergola 2009a) or empirical distribution function (Sengupta 2022), or the study population only comprised patients undergoing dialysis (Ahmadi 2017; Guo 2013; Singh 2000).

We observed considerable heterogeneity in the methods used to measure eGFR. The Modification of Diet in Renal Disease (MDRD) equation was used to measure eGFR in 20 studies (ATIC 2005; BEACON 2013; BEAM 2011; Bejaoui 2022; Danilovic 2011; Goicoechea 2012; Guo 2013; Lin 2008; Modarresi 2018; Modarresi 2017; Moreillon 2013; Orban 2015; Perkins 2009; PREDIAN 2011; Silveira 2019; Singer 2011; TSUBAKI 2020; Turki 2016; Xu 2017; Yan 2017a), the Chronic Kidney Disease Epidemiology Collaboration equation in six studies (Borges 2016; Martinez 2020; NICE 2020; Rabizadeh 2018; Tan 2019; Voroneanu 2017), and the Cockroft and Gault equation in four studies (Blackhall 2005; Moist 2010; Schramm 2002; Singh 2000). Ahmadi 2017 used the arithmetic mean of creatinine and urea clearance over 24 hours to determine eGFR, and Pollak 1993 used the clearance of Tc‐DTPA (diethylene‐triamine‐pentaacetate). Nine studies did not report which method they used to estimate eGFR (Bansal 2017; Giliberti 2022; Jimenez‐Osorio 2016; Navarro 1999; Navarro 1999a; Noel 1997; Rabl 1993; Sahraei 2015; XANTHIN 2008).

Serum creatinine

Thirty‐two studies reported SCr at the end of the study (AIONID 2013; Argani 2014; Asemi 2016; Blackhall 2005; Danilovic 2011; Demir 2006a; DeVault 1994; Firuzi 2016; Friedman 2003; Gonzalez‐Espinoza 2012; Hosseini 2021; Jimenez‐Osorio 2016; Konigsrainer 1995; Modarresi 2018; Moist 2010; Moreillon 2013; Mori 2009; Navarro 1999a; Noel 1997; Norio 2003; Orban 2015; Perkins 2009; Pollak 1993; Rabizadeh 2018; Rabl 1993; Rassaf 2016; Shoskes 2005; Singh 2000; Tan 2019; Vincenti 1996; XANTHIN 2008; Xu 2017) and 28 studies reported change in SCr (AIONID 2013; Argani 2014; Asemi 2016; BEACON 2013; BEAM 2011; Biniaz 2014; Blackhall 2005; Demir 2006a; DeVault 1994; Firuzi 2016; Friedman 2003; Gonzalez‐Espinoza 2012; Hosseini 2021; Jimenez‐Osorio 2016; Modarresi 2018; Moist 2010; Moreillon 2013; Mori 2009; Navarro 1999a; Paniagua 1995; Perkins 2009; Rabizadeh 2018; Rassaf 2016; Singh 2000; Tan 2019; XANTHIN 2008; Xu 2017; Zachara 2009). We calculated change scores and imputed SD for change (correlation coefficient = 0.89) in 17 studies (BEACON 2013; Biniaz 2014; Firuzi 2016; Friedman 2003; Jimenez‐Osorio 2016; Modarresi 2018; Moist 2010; Moreillon 2013; Mori 2009; Navarro 1999a; Perkins 2009; Rabizadeh 2018; Rassaf 2016; Singh 2000; Tan 2019; XANTHIN 2008; Xu 2017). Thirteen studies could not be included in the meta‐analyses because they only reported their results narratively (Paniagua 1995; Zachara 2009), the study population only comprised dialysis patients (AIONID 2013; Argani 2014; Asemi 2016; Biniaz 2014; Hajian 2022; Firuzi 2016; Friedman 2003; Gonzalez‐Espinoza 2012; Hosseini 2021; Rassaf 2016), or because no measure was reported (Konigsrainer 1995).

Urinary albumin/creatinine ratio

Six studies reported UACR at the end of the study (Bansal 2017; Borges 2016; HOPE 1996; Mori 2009; Silveira 2019; Tan 2019) and seven reported changes in UACR (Bansal 2017; Borges 2016; HOPE 1996; Mori 2009; NICE 2020; Silveira 2019; Tan 2019). We only identified studies reporting UACR in patients with CKD 3‐5. We calculated change scores and imputed SD for change (correlation coefficient = 0.80) in four studies (Borges 2016; HOPE 1996; Silveira 2019; Tan 2019).

Proteinuria

Eleven studies reported proteinuria at the end of study (Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Moist 2010; Mori 2009; Navarro 1999a; Noel 1997; Perkins 2009; Silveira 2019; Song 2006a; Voroneanu 2017) and eight studies reported change in proteinuria (Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Moist 2010; Mori 2009; Perkins 2009; Silveira 2019; Voroneanu 2017). We only identified studies reporting urinary proteinuria in patients with CKD 3‐5. We calculated change scores and imputed SD for change (correlation coefficient = 0.80) in eight studies (Jimenez‐Osorio 2016; Lin 2008; Martinez 2020; Moist 2010; Mori 2009; Perkins 2009; Silveira 2019; Voroneanu 2017). Two studies could not be included in the meta‐analyses because it was not possible to convert proteinuria to g/day (Lin 2008) or because no measure was reported (Noel 1997).

Albuminuria

Four studies reported albuminuria at the end of the study (ATIC 2005; Goicoechea 2012; Mori 2009; PREDIAN 2011), and five studies reported change in proteinuria (ATIC 2005; Goicoechea 2012; Mori 2009; PREDIAN 2011; Thaha 2009). We only identified studies reporting albuminuria in patients with CKD 3‐5. We calculated change scores and imputed SD for change (correlation coefficient = 0.80) in two studies (Mori 2009; PREDIAN 2011). One study could not be included in the meta‐analyses (Thaha 2009) because they only reported their results narratively.

Thrombosis of vascular access

Three studies reported thrombosis of vascular access (Lu 2007; Singer 2011; Yang 2019a).

Graft loss

Eleven studies reported graft loss (Danilovic 2011; DeVault 1994; Konigsrainer 1995; Modarresi 2017; Modarresi 2018; Noel 1997; Norio 2003; Orban 2015; Schneeberger 1990; Schramm 2002; Vincenti 1996).

Graft failure

One study reported graft failure (XANTHIN 2008).

Delayed graft function

Twelve studies reported DGF (Danilovic 2011; Modarresi 2017; Modarresi 2018; Noel 1997; Norio 2003; Oko 1999; Orban 2015; Sahraei 2015; Schramm 2002; Shoskes 2005; Soliman 2014; Wlodarczyk 2000). One study could not be included in the meta‐analyses because they only reported their results narratively (Wlodarczyk 2000).

Malignancy

Three studies reported malignancies (BEAM 2011; Moreillon 2013; SPACE 2000). We included both solid tumours and haematological malignancies. There were no studies reporting malignancies in kidney transplant recipients. One study could not be included in the meta‐analysis because the authors only provided the total number of events and did not respond to our enquiries for further information (Moreillon 2013).

Major bleeding

Six studies reported major bleeding (Attallah 2006; Himmelfarb 2007; Lin 2008; Perkins 2009; Rassaf 2016; SPACE 2000). There were no studies reporting major bleeding in kidney transplant recipients.

Infection

Seventeen studies reported infection (ACTIVE 2014; Antunes 2014; Attallah 2006; BEACON 2013; BEAM 2011; Konigsrainer 1995; Noel 1997; Perkins 2009; PREDIAN 2011; Shema‐Didi 2012; Shojaei 2011; Singer 2011; SPACE 2000; Tan 2019; TSUBAKI 2020; Wlodarczyk 2000; Xu 2017). Three studies could not be included in the meta‐analyses because the authors only provided the total number of events and did not respond to our enquiries for further information (Shema‐Didi 2012; Shojaei 2011; Xu 2017). There was a large heterogeneity among studies which reported infection, but almost all required hospitalisation.

Excluded studies

We excluded 147 studies after full‐text assessment.

• Did not report outcomes relevant to this review (96) (Abdollahzad 2009; Agarwal 2004a; Aghadavod 2018; Ahmadi 2013; Ali 2003; Ali 2021; Alvarenga 2020; Alvarenga 2022; Aminorroaya 2005; Bakris 2012; Baldi 2013; Biniaz 2013; Biniaz 2015; Candan 1996; Candan 2002; Chang 2007a; Chao 2002; Coloma 2011; Conner 2008a; Coombes 2000; Daud 2013; Elham 2019a; Fallah 2018; Fukuda 2015; Fumeron 2005; Garneata 2015; Giancaspro 2000; Heidari 2018; Hodkova 2006; Holdaas 1994; Hosseini 2022; Imada 2001; Irijanto 2007; Jamal 2022; Kalani 2005; Kamgar 2009; Keven 2003a; Khajehdehi 2000a; Khajehdehi 2001; Kobayashi 2015; Lin 2016; MacCallum 2009; Mahajan 1979; Martins 2021; Matson 2003; Mazani 2013; Moffitt 2013; Mooraki 2007; Mortazavi 2012; Muchova 2014; Mune 1999; Murillo Ortiz 2019; Najafabadi 2012; Napolitano 1996; Nascimento 2010; NCT00440869; Omrani 2015; Pakfetrat 2013; Panah 2019; Perez 2004; Pirhadi‐Tavandashti 2020; Poulia 2011; Pratama 2020; Purwanto 2012; Ramos 2011; Rashidi 2009; Renke 2008a; Rodger 1989; Rodrigues 2021; Roob 2000; Roozbeh 2011; Rymarz 2009; Safa 2014; Sagheb 2012; Saldanha 2016; Salimian 2022; Samadian 2017; Sanchez Alvarez 2005; Schneeberger 1989; Scholze 2004; Sedaghattalab 2021; Sedaghattalab 2021a; Shahbazian 2016; Shahreki 2022; Sinclair 1997; Svensson 2004; Temple 2000; Turk 1998; Turk 2013; Uzum 2006; VIPER 2004; Williams 2001a; Wittstock 2009; Yavari 2017; Yeksan 1992; Zachara 2011)

• Did not include patients with CKD (41) (Badri 2013; Bahmani 2016a; Bursell 1999; El Mashad 2016; El‐Shazly 2015; Fallahzadeh 2012; Farvid 2005; Gaede 2001; Gholnari 2018; Ghorbani 2012; Golmohammadi 2022; Guerrero‐Romero 1995a; Han 2015; Harmankaya 2003; Jadhav 2014; Javaherforooshzadeh 2021; Kadhim 2006; Khajehdehi 2011; Khajehdehi 2012; Khan 2013; Khatami 2016; Klein 1995; Leyva‐Jimenez 2009; Mahmood 2018; McAuliffe 1998; Mooraki 2006; Navarro 2003; Navarro 2005; Nowak 2020; Oliaei 2011; Pantoja 2003; Parham 2008; Ponce de Leon‐Vargas 2013; PREVENT 2016; Rodriguez‐Moran 2006; Roozbeh 2010; Sattarinezhad 2019; Shahidi 2015; Sun 2017b; Tan 2018; Vanaie 2019; Wijnen 2002).

• Unable to retrieve the full text (4) (Modi 2001; Munguia 2003; Sprenger 1983; Yokoyama 2001)

• Did not report outcomes per study arm (3) (Garrote 2009; Zahed 2016; Zhang 2013)

• Did not study an antioxidant intervention (2) (Peng 2010; Rodhe 2013).

Risk of bias in included studies

The risk of bias in the included studies is summarised in Figure 2. A detailed overview of the risk of bias assessment for individual studies is provided in Figure 3, and the characteristics of included studies tables.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

We judged methods for random sequence generation to be low risk of bias in 47 studies (Ahmadi 2017; AIONID 2013; Asemi 2016; Atapour 2022; ATIC 2005; BEAM 2011; Biniaz 2014; Borges 2016; CoQ10 Biomarker 2016; Gholipour Baradari 2011; Giliberti 2022; Goicoechea 2012; Gonzalez‐Espinoza 2012; Haddadian‐Khouzani 2022; Hajian 2022; HERO 2008; Himmelfarb 2007; HOPE 1996; Hajian 2022; Khabbazi 2012; Lin 2008; Lu 2007; Modarresi 2018; Modarresi 2017; Moist 2010; Navarro 1999; NICE 2020; Norio 2003; Okamoto 2020; Orban 2015; PATH 2014; Perkins 2009; Pollak 1993; PREDIAN 2011; Rabizadeh 2018; Rassaf 2016; Sahraei 2015; Salehi 2013; Silveira 2019; Singer 2011; SPACE 2000; Tan 2019; Tonelli 2015; Voroneanu 2017; XANTHIN 2008; Xu 2017; Yang 2019a). For three studies, the risk of bias due to random sequence generation was deemed high because the random sequence generation involved a non‐random component (Demir 2006a), the baseline characteristics were severely imbalanced (Mori 2009), and the sample size between the intervention and control group was very unevenly distributed (TSUBAKI 2020). For the remaining 45 studies, the risk of bias was unclear due to a lack of information on the methods for random sequence generation.

Allocation concealment

We judged methods for allocation concealment to be low risk of bias in 21 studies (ATIC 2005; BEAM 2011; Borges 2016; Giliberti 2022; Gonzalez‐Espinoza 2012; HERO 2008; HOPE 1996; Khabbazi 2012; Lu 2007; Modarresi 2018; Moreillon 2013; Mori 2009; NICE 2020; Norio 2003; Pollak 1993; PREDIAN 2011; Rassaf 2016; Salehi 2013; Tan 2019; Tonelli 2015; XANTHIN 2008). For seven studies, the risk of bias due to allocation concealment was judged high (Attallah 2006; Demir 2006a; Guo 2013; Konigsrainer 1995; Lin 2008; Navarro 1999; Xu 2017) because allocation was not concealed. For the remaining 67 studies, the risk of bias was unclear due to a lack of information on allocation concealment.

Blinding

Blinding of participants and personnel

We judged blinding of participants and personal to be low risk of bias in 55 studies (AIONID 2013; Argani 2014; Asemi 2016; Atapour 2022; ATIC 2005; BEACON 2013; BEAM 2011; Bejaoui 2022; Biniaz 2014; Borges 2016; Jern 2000; CoQ10 Biomarker 2016; DeVault 1994; Friedman 2003; Gonzalez‐Espinoza 2012; Haddadian‐Khouzani 2022; Hajian 2022; HERO 2008; Himmelfarb 2007; HOPE 1996; Hosseini 2021; Jimenez‐Osorio 2016; Khabbazi 2012; Lu 2007; Modarresi 2018; Modarresi 2017; Moist 2010; Moreillon 2013; Mori 2009; NICE 2020; Norio 2003; PATH 2014; Perkins 2009; Pollak 1993; PREDIAN 2011; Rassaf 2016; Roozbeh 2009; Salehi 2013; Schneeberger 1990; Shema‐Didi 2012; Shojaei 2011; Silveira 2019; Singer 2011; Singh 2000; Soliman 2014; SPACE 2000; Tan 2019; Tonelli 2015; TSUBAKI 2020; Turki 2016; Vafadar Afshar 2020; Vincenti 1996; XANTHIN 2008; Yang 2019a; Zachara 2009). For 18 studies, the risk of bias due to the blinding of participants and personnel was high because either one or both were unblinded (Attallah 2006; Blackhall 2005; Giliberti 2022; Goicoechea 2012; Guo 2013; Hajji 2021; Konigsrainer 1995; Lin 2008; Navarro 1999; Navarro 1999a; Okamoto 2020; Omar 2022; Rabizadeh 2018; Rivara 2015; Schramm 2002; Tepel 2003; Voroneanu 2017; Yan 2017a). For the remaining 22 studies, the risk of bias was unclear due to a lack of information on blinding procedures.

Blinding of outcome assessors

Eighteen studies did not blind outcome assessors (Blackhall 2005; Giliberti 2022; Goicoechea 2012; Guo 2013; Hajji 2021; Konigsrainer 1995; Lin 2008; Navarro 1999; Navarro 1999a; Okamoto 2020; Omar 2022; Rabizadeh 2018; Rivara 2015; Schramm 2002; Tepel 2003; Voroneanu 2017; Wlodarczyk 2000; Yan 2017a), and 18 did not specify whether outcome assessors were blinded (ACTIVE 2014; Ahmadi 2017; Antunes 2014; Bansal 2017; Danilovic 2011; Firuzi 2016; Martinez 2020; Oko 1999; Orban 2015; Paniagua 1995; Pergola 2009a; Rabl 1993; Sahraei 2015; Sengupta 2022; Shoskes 2005; Song 2006a; Thaha 2009; Xu 2017). Nevertheless, we assumed a low risk of bias for all studies since the results are unlikely to be affected by blinding of outcome assessors.

Incomplete outcome data

We judged the risk of bias due to incomplete outcome data low for 54 studies (Ahmadi 2017; Argani 2014; Asemi 2016; Atapour 2022; ATIC 2005; Attallah 2006; BEACON 2013; BEAM 2011; Biniaz 2014; Blackhall 2005; Borges 2016; CoQ10 Biomarker 2016; Friedman 2003; Gholipour Baradari 2011; Giliberti 2022; Goicoechea 2012; Gonzalez‐Espinoza 2012; Haddadian‐Khouzani 2022; HERO 2008; Himmelfarb 2007; HOPE 1996; Hosseini 2021; Khabbazi 2012; Lin 2008; Lu 2007; Modarresi 2017; Modarresi 2018; Moist 2010; Moreillon 2013; NICE 2020; Noel 1997; Norio 2003; Omar 2022; Orban 2015; Perkins 2009; PREDIAN 2011; Rabl 1993; Rassaf 2016; Sahraei 2015; Sengupta 2022; Shoskes 2005; Silveira 2019; Singer 2011; Singh 2000; SPACE 2000; Tepel 2003; Tonelli 2015; Turki 2016; Vafadar Afshar 2020; Vincenti 1996; Voroneanu 2017; XANTHIN 2008; Yan 2017a; Yang 2019a). For 10 studies, the risk of bias due to incomplete outcome data was high due to significant loss to follow‐up during the study (ACTIVE 2014; Guo 2013; Hajji 2021; Okamoto 2020; PATH 2014; Shojaei 2011) or unequal loss to follow‐up in intervention and control group (Salehi 2013; Shema‐Didi 2012; Tan 2019; TSUBAKI 2020). For the remaining 31 studies, the risk of bias was unclear because loss to follow‐up was not reported.

Selective reporting

We judged the risk of bias due to selective reporting low in 19 studies (ATIC 2005; BEACON 2013; BEAM 2011; Bejaoui 2022; Haddadian‐Khouzani 2022; HERO 2008; Jimenez‐Osorio 2016; Lin 2008; Moist 2010; Orban 2015; Rabizadeh 2018; Sengupta 2022; Silveira 2019; Singh 2000; Tan 2019; Thaha 2009; Tonelli 2015; TSUBAKI 2020; XANTHIN 2008). For three studies, the risk of reporting bias was deemed high because relevant outcomes reported in the methods were missing in the results (AIONID 2013; Pollak 1993) or because the SCr outcome appeared to have been changed to eGFR without explanation (Sahraei 2015). For 73 studies risk of bias was unclear because we were unable to retrieve a study protocol (58) (ACTIVE 2014; Ahmadi 2017; Antunes 2014; Argani 2014; Asemi 2016; Attallah 2006; Bansal 2017; Biniaz 2014; Blackhall 2005; Borges 2016; Danilovic 2011; Demir 2006a; DeVault 1994; Firuzi 2016; Friedman 2003; Gholipour Baradari 2011; Goicoechea 2012; Gonzalez‐Espinoza 2012; Guo 2013; Himmelfarb 2007; HOPE 1996; Hosseini 2021; Khabbazi 2012; Konigsrainer 1995; Schramm 2002; Martinez 2020; Modarresi 2018; Modarresi 2017; Moreillon 2013; Mori 2009; Navarro 1999; Navarro 1999a; NICE 2020; Noel 1997; Norio 2003; Oko 1999; Paniagua 1995; PATH 2014; Pergola 2009a; Perkins 2009; PREDIAN 2011; Rabl 1993; Rassaf 2016; Schneeberger 1990; Shema‐Didi 2012; Shoskes 2005; Singer 2011; Soliman 2014; Song 2006a; SPACE 2000; Tepel 2003; Turki 2016; Vincenti 1996; Voroneanu 2017; Xu 2017; Yan 2017a; Yang 2019a; Zachara 2009) or because the extracted outcomes was not specified in the methods or protocol (15) (Atapour 2022; Jern 2000; CoQ10 Biomarker 2016; Giliberti 2022; Hajian 2022; Hajji 2021; Lu 2007; Okamoto 2020; Omar 2022; Rivara 2015; Roozbeh 2009; Salehi 2013; Shojaei 2011; Vafadar Afshar 2020; Wlodarczyk 2000).

Other potential sources of bias

Seventy‐one studies appeared to be free from other sources of bias (Ahmadi 2017; AIONID 2013; Antunes 2014; Argani 2014; Atapour 2022; ATIC 2005; Attallah 2006; BEAM 2011; Bejaoui 2022; Biniaz 2014; Blackhall 2005; Borges 2016; CoQ10 Biomarker 2016; Danilovic 2011; Demir 2006a; DeVault 1994; Firuzi 2016; Friedman 2003; Gholipour Baradari 2011; Goicoechea 2012; Gonzalez‐Espinoza 2012; Guo 2013; Haddadian‐Khouzani 2022; Hajian 2022; Hajji 2021; Himmelfarb 2007; Hosseini 2021; Jern 2000; Jimenez‐Osorio 2016; Konigsrainer 1995; Lin 2008; Lu 2007; Modarresi 2017; Modarresi 2018; Moist 2010; Navarro 1999; Navarro 1999a; NICE 2020; Noel 1997; Norio 2003; Omar 2022; Orban 2015; Paniagua 1995; Perkins 2009; PREDIAN 2011; Rabizadeh 2018; Rabl 1993; Rassaf 2016; Rivara 2015; Roozbeh 2009; Salehi 2013; Schramm 2002; Sengupta 2022; Shema‐Didi 2012; Shojaei 2011; Silveira 2019; Singer 2011; Singh 2000; SPACE 2000; Tan 2019; Tepel 2003; Tonelli 2015; Turki 2016; Vafadar Afshar 2020; Vincenti 1996; Voroneanu 2017; Wlodarczyk 2000; Xu 2017; Yan 2017a; Yang 2019a; Zachara 2009). In 21 studies we identified other sources of bias which were judged as high risk of bias such as only publication of a conference abstract, involvement of the funder in the design, conduct, and analysis of the study, and problems with randomization (ACTIVE 2014; Bansal 2017; BEACON 2013; Giliberti 2022; HERO 2008; HOPE 1996; Martinez 2020; Mori 2009; Okamoto 2020; Oko 1999; PATH 2014; Pergola 2009a; Pollak 1993; Sahraei 2015; Schneeberger 1990; Shoskes 2005; Soliman 2014; Song 2006a; Thaha 2009; TSUBAKI 2020; XANTHIN 2008). For the remaining three studies, the risk of other sources of bias was unclear.

Effects of interventions

See: Table 1

Antioxidants versus placebo or standard care

See Table 1: antioxidants versus control for adults with chronic kidney disease.

Cardiovascular death

Antioxidants may have little to no effect on CVD in adults with CKD (Analysis 1.1 (8 studies, 3813 patients): RR 0.94, 95% CI 0.64 to 1.40; I² = 33%; low certainty evidence). The evidence was downgraded twice for risk of bias. The effect did not appear to differ between adults with CKD 3‐5, CKD 5D, and CKD 5‐transplant (test for subgroup differences: Chi² = 1.31, I² = 0, P = 0.52). No differences were observed between interventions with and without vitamins, risk of bias or duration of follow‐up (Analysis 4.1; Analysis 5.1; Analysis 6.1; Analysis 7.1).

1.1. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 1: Cardiovascular death

4.1. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 1: Cardiovascular death

5.1. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 1: Cardiovascular death

6.1. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 1: Cardiovascular death

7.1. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 1: Cardiovascular death

Death (any cause)

Antioxidants probably do not reduce death in adults with CKD (Analysis 1.2 (45 studies, 7530 patients): RR 0.95, 95% CI 0.82 to 1.11; I² = 0%; moderate certainty evidence). The evidence was downgraded for risk of bias. The effect did not differ between adults with CKD 3‐5, CKD 5D, and CKD 5‐transplant (test for subgroup differences: Chi² = 0.70, I² = 0%, P = 0.70). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.2; Analysis 5.2; Analysis 6.2; Analysis 7.2).

1.2. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 2: Death (any cause)

4.2. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 2: Death (any cause)

5.2. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 2: Death (any cause)

6.2. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 2: Death (any cause)

7.2. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 2: Death (any cause)

Cardiovascular disease

Antioxidants probably reduce the risk of CVD in adults with CKD (Analysis 1.3 (16 studies, 4768 patients): RR 0.79, 95% CI 0.63 to 0.99; I² = 23%; moderate certainty evidence). The evidence was downgraded for risk of bias. This effect was observed in adults with CKD 3‐5 (RR 0.76, 95% CI 0.59 to 0.99) but not in adults undergoing dialysis (RR 0.87, 95% CI 0.47 to 1.62) or kidney transplant recipients (RR 5.00, 95% CI 0.25 to 101.58), even though the test for subgroup differences was not statistically significant (Chi² = 1.60, I² = 0%, P = 0.45). In studies with a low risk of bias, we observed a larger protective effect of antioxidants (Analysis 6.3 (6 studies, 808 patients): RR 0.61, 95% CI 0.40 to 0.91; I² = 1%). No differences were observed between interventions with and without vitamins or duration of follow‐up (Analysis 4.3; Analysis 5.3; Analysis 7.3).

1.3. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 3: Cardiovascular disease

6.3. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 3: Cardiovascular disease

4.3. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 3: Cardiovascular disease

5.3. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 3: Cardiovascular disease

7.3. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 3: Cardiovascular disease

Coronary heart disease

Antioxidants may not reduce coronary heart disease in adults with CKD (Analysis 1.4 (12 studies, 4372 patients): RR 0.85, 95% CI 0.64 to 1.12; I² = 8%; low certainty evidence). The evidence was downgraded twice for risk of bias. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5D (test for subgroup differences: Chi² = 1.87, I² = 46.4%, P = 0.17). The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In studies with a low risk of bias, we observed a protective effect of antioxidants on coronary heart disease (Analysis 6.4. (6 studies, 818 patients): RR 0.40, 95% CI 0.21 to 0.77; I² = 0%). No differences were observed between interventions with and without vitamins or duration of follow‐up (Analysis 4.4; Analysis 5.4; Analysis 7.4).

1.4. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 4: Coronary heart disease

6.4. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 4: Coronary heart disease

4.4. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 4: Coronary heart disease

5.4. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 4: Coronary heart disease

7.4. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 4: Coronary heart disease

Heart failure

Antioxidants probably increase the risk of heart failure in adults with CKD (Analysis 1.5 (6 studies, 3733 patients): RR 1.40, 95% CI 1.11 to 1.76; I² = 0%; moderate certainty evidence). The evidence was downgraded for risk of bias. In the studies with a low risk of bias, we observed no effect of antioxidants on heart failure. No differences were observed between interventions with and without vitamins or duration of follow‐up (Analysis 4.5; Analysis 5.5; Analysis 6.5). The efficacy for patients receiving dialysis and kidney transplant recipients could not be estimated as none of the studies included these patients.

1.5. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 5: Heart failure

4.5. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 5: Heart failure

5.5. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 5: Heart failure

6.5. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 5: Heart failure

Cerebrovascular disease

Antioxidants probably do not reduce cerebrovascular disease in adults with CKD (Analysis 1.6 (8 studies, 1769 patients): RR 0.97, 95% CI 0.71 to 1.32; I² = 0%; moderate certainty evidence). The evidence was downgraded for risk of bias. The effect did not appear to differ between adults with CKD 3‐5, CKD 5D, and CKD 5‐transplant (test for subgroup differences: Chi² = 0.70, I² = 0%, P = 0.71). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.6; Analysis 5.6; Analysis 6.6; Analysis 7.6).

1.6. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 6: Cerebrovascular disease

4.6. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 6: Cerebrovascular disease

5.6. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 6: Cerebrovascular disease

6.6. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 6: Cerebrovascular disease

7.6. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 6: Cerebrovascular disease

Peripheral vascular disease

Antioxidants may have little to no effect on peripheral vascular disease in adults with CKD, but the evidence is very uncertain (Analysis 1.7 (4 studies, 1492 patients): RR 0.77, 95% CI 0.41 to 1.47; I² = 48%; very low certainty evidence). The evidence was downgraded once for imprecision and twice for risk of bias. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5D (test for subgroup differences: Chi² = 0.83, I² = 0%, P = 0.36). The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.7; Analysis 5.7; Analysis 6.7; Analysis 7.7).

1.7. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 7: Peripheral vascular disease

4.7. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 7: Peripheral vascular disease

5.7. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 7: Peripheral vascular disease

6.7. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 7: Peripheral vascular disease

7.7. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 7: Peripheral vascular disease

Kidney failure

Antioxidants may reduce the risk of kidney failure in adults with CKD (Analysis 1.8 (10 studies, 3201 patients): RR 0.65, 95% CI 0.41 to 1.02; I² = 41%; moderate certainty evidence). The evidence was downgraded for risk of bias. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5‐transplant (test for subgroup differences: Chi² = 0.28, I² = 0%, P = 0.59). We observed a protective effect from non‐vitamin antioxidants against kidney failure (Analysis 5.8. (9 studies, 3108 patients): RR 0.60, 95% CI 0.38 to 0.96; I² = 41%) but not for vitamin antioxidants (Analysis 4.8. (1 study, 93 patients): RR 1.63, 95% CI 0.41 to 6.43), although the latter was only based on a single study. No differences were observed in the risk of bias or duration of follow‐up (Analysis 6.8; Analysis 7.8).

1.8. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 8: Kidney failure

5.8. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 8: Kidney failure

4.8. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 8: Kidney failure

6.8. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 8: Kidney failure

7.8. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 8: Kidney failure

Estimated glomerular filtration rate

Antioxidants may result in a higher eGFR in adults with CKD, but the evidence is very uncertain (Analysis 1.9 (32 studies, 2435 patients): MD 3.85 mL/min/1.73 m², 95% CI 2.13 to 5.56; I² = 96%; very low certainty evidence). The evidence was downgraded once for risk of bias and twice for inconsistency. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5‐transplant (test for subgroup differences: Chi² = 0.39, I² = 0%, P = 0.53). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.9; Analysis 5.9; Analysis 6.9; Analysis 7.9).

1.9. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 9: eGFR at end of study

4.9. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 9: eGFR at end of study

5.9. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 9: eGFR at end of study

6.9. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 9: eGFR at end of study

7.9. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 9: eGFR at end of study

Antioxidants may increase change in eGFR in adults with CKD, but the evidence is very uncertain (Analysis 1.10 (28 studies, 4128 patients): MD 3.65 mL/min/1.73 m², 95% CI 2.81 to 4.50; I² = 99%; very low certainty evidence). The evidence was downgraded once for risk of bias, possible publication bias, and twice for inconsistency. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5‐transplant (test for subgroup differences: Chi² = 0.00 I² = 0%, P = 0.95). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.10; Analysis 5.10; Analysis 6.10; Analysis 7.10).

1.10. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 10: Change in eGFR

4.10. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 10: Change in eGFR

5.10. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 10: Change in eGFR

6.10. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 10: Change in eGFR

7.10. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 10: Change in eGFR

Serum creatinine

Antioxidants may have little to no effect on SCr in adults with CKD, but the evidence is very uncertain (Analysis 1.11 (22 studies, 1587 patients): MD ‐6.24 µmol/L, 95% CI ‐15.71 to 3.23; I² = 82%; very low certainty evidence). The evidence was downgraded once for risk of bias and imprecision and twice for inconsistency. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5‐transplant (test for subgroup differences: Chi² = 2.34, I² = 57%, P = 0.13). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.11; Analysis 5.11; Analysis 6.11; Analysis 7.11).

1.11. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 11: Serum creatinine at end of study

4.11. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 11: Serum creatinine at end of study

5.11. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 11: Serum creatinine at end of study

6.11. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 11: Serum creatinine at end of study

7.11. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 11: Serum creatinine at end of study

Antioxidants may slightly reduce change in SCr levels, but the evidence is very uncertain (Analysis 1.12 (16 studies, 3180 patients): MD ‐13.35 µmol/L, 95% CI ‐23.49 to ‐3.20; I² = 98%; very low certainty evidence). The evidence was downgraded once for possible publication bias and twice for risk of bias and inconsistency. The effect did not appear to differ between adults with CKD 3‐5 and CKD 5‐transplant (test for subgroup differences: Chi² = 0.12, I² = 0%, P = 0.73). The reduction in change in SCr was only observed in non‐vitamin antioxidants (Analysis 5.12 (14 studies, 3109 patients): MD ‐16.32 µmol/L, 95% CI ‐26.87 to ‐5.78; I² = 98%) but not for vitamin antioxidants (Analysis 4.12 (2 studies, 71 patients): MD 4.85 µmol/L, 95% CI ‐4.14 to 13.84; I² = 0%). The change in SCr was also larger in studies with a low risk of bias (Analysis 6.12 (4 studies, 401 patients): MD ‐35.01 µmol/L, 95% CI ‐68.93 to ‐1.09; I² = 0%). No difference was observed in the duration of follow‐up (Analysis 7.12).

1.12. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 12: Change in serum creatinine

5.12. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 12: Change in serum creatinine

4.12. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 12: Change in serum creatinine

6.12. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 12: Change in serum creatinine

7.12. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 12: Change in serum creatinine

Urinary albumin/creatinine ratio

Antioxidants may have little or no effect on UACR in adults with CKD, but the evidence is very uncertain. There was no difference in UACR in adults receiving antioxidants and those receiving placebo or usual care (Analysis 1.13 (5 studies, 1174 patients): MD ‐0.59 mg/mmol, 95% CI ‐12.78 to 11.59; I² = 57%; very low certainty evidence). The evidence was downgraded once for inconsistency and imprecision and twice for risk of bias. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In studies with a low risk of bias, we observed lower UACR levels in adults with CKD treated with antioxidants (Analysis 6.13 (2 studies, 78 patients): MD ‐111.94 mg/mmol, 95% CI ‐188.36 to ‐35.53; I² = 0%). No differences were observed between interventions with and without vitamins or duration of follow‐up (Analysis 4.13; Analysis 5.13; Analysis 7.13).

1.13. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 13: Urinary albumin/creatinine ratio at end of study

6.13. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 13: Urinary albumin/creatinine ratio at end of study

4.13. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 13: Urinary albumin/creatinine ratio at end of study

5.13. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 13: Urinary albumin/creatinine ratio at end of study

7.13. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 13: Urinary albumin/creatinine ratio at end of study

Antioxidants may have little to no effect on change in UACR in adults with CKD, but the evidence is very uncertain (Analysis 1.14 (7 studies, 1279 patients): MD ‐0.02 mg/mmol, 95% CI ‐0.70 to 0.66; I² = 53%; very low certainty of evidence). The evidence was downgraded once for imprecision and twice for risk of bias. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In studies with a low risk of bias, antioxidants may reduce the change in UACR, but the evidence is very uncertain (Analysis 6.14 (3 studies, 115 patients): MD ‐37.54 mg/mmol, 95% CI ‐101.00 to 25.92; I² = 66%). No differences were observed between interventions with and without vitamins or duration of follow‐up (Analysis 4.14; Analysis 5.14; Analysis 7.14).

1.14. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 14: Change in urinary albumin/creatinine ratio

6.14. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 14: Change in urinary albumin/creatinine ratio

4.14. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 14: Change in urinary albumin/creatinine ratio

5.14. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 14: Change in urinary albumin/creatinine ratio

7.14. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 14: Change in urinary albumin/creatinine ratio

Proteinuria

Antioxidants may have little to no effect on proteinuria in adults with CKD, but the evidence is very uncertain (Analysis 1.15 (9 studies, 536 patients): MD ‐0.24 g/24 hours, 95% CI ‐0.78 to 0.31; I² = 90%; very low certainty evidence). The evidence was downgraded once for risk of bias and twice for inconsistency and imprecision. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In studies with a follow‐up of six months or more, we observed lower proteinuria levels (Analysis 7.15 (5 studies, 233 patients): MD ‐0.95 g/24 hours, 95% CI –1.88 to ‐0.02; I² = 71%) No differences were observed between studies with low risk of bias (Analysis 6.15). We could not compare vitamin and non‐vitamin interventions because all studies included interventions without vitamins.

1.15. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 15: Proteinuria at end of study

7.15. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 15: Proteinuria at end of study

6.15. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 15: Proteinuria at end of study

Antioxidants may result in a small reduction in change in proteinuria in adults with CKD, but the evidence is very uncertain (Analysis 1.16 (8 studies, 468 patients): MD ‐0.51 g/24 hours, 95% CI ‐1.11 to 0.09; I² = 97%; very low certainty evidence). The evidence was downgraded once for risk of bias and twice for inconsistency and imprecision. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In studies with a follow‐up of six months or more, we observed a larger reduction in proteinuria (Analysis 7.16 (3 studies, 132 patients): MD ‐0.90 g/24 hours, 95% CI –1.80 to 0.01; I² = 94%) No differences were observed between studies with low risk of bias (Analysis 6.16). We could not compare vitamin and non‐vitamin interventions because all studies included interventions without vitamins.

1.16. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 16: Change in proteinuria

7.16. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 16: Change in proteinuria

6.16. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 16: Change in proteinuria

Albuminuria

Antioxidants may increase albuminuria in adults with CKD, but the evidence is very uncertain (Analysis 1.17 (4 studies, 395 patients): MD 94.75 mg/24 hours, 95% CI 0.57 to 188.94; I² = 0%; very low certainty evidence). This effect was mainly driven by Mori 2009 which was the only study reporting an increase in albuminuria. The evidence was downgraded twice for risk of bias and twice for imprecision. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. In the analysis with studies with low risk of bias and longer duration of follow‐up, the increase in albuminuria was not observed (Analysis 6.17 (2 studies, 262 patients): MD ‐40.57 mg/24 hours, 95% CI –274.19 to 193.04; I² = 0%); (Analysis 7.17 (3 studies, 353 patients): MD ‐2.78 mg/24 hours, 95% CI –151.55 to 145.99; I² = 0%). The effects of vitamin and non‐vitamin antioxidants may differ. In the vitamin antioxidants, we observed little or no effect on albuminuria (Analysis 4.15 (1 study, 93 patients): MD ‐4.0 mg/24 hours, 95% CI –217.48 to 209.48), and in the non‐vitamin antioxidants, we observed higher levels albuminuria (Analysis 5.17 (3 studies, 302 patients): MD 99.92 mg/24 hours, 95% CI –10.95 to 210.78; I² = 12%).

1.17. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 17: Albuminuria at end of study

6.17. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 17: Albuminuria at end of study

7.17. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 17: Albuminuria at end of study

4.15. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 15: Albuminuria at end of study

5.17. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 17: Albuminuria at end of study

Antioxidants may have little or no effect on change in albuminuria in adults with CKD, but the evidence is very uncertain (Analysis 1.18 (4 studies, 395 patients): MD ‐87.08 mg/24 hours, 95% CI ‐221.12 to 46.97; I² = 97%; very low certainty evidence). The evidence was downgraded once for risk of bias and inconsistency and twice for imprecision. The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.15; Analysis 5.17; Analysis 6.17; Analysis 7.17).

1.18. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 18: Change in albuminuria

Thrombosis of vascular access

Antioxidants may result in little to no difference in thrombosis of vascular access in adults undergoing dialysis (Analysis 1.19 (3 studies, 209 patients): RR 1.08, 95% CI 0.41 to 2.84; I² = 11%; low certainty evidence). The evidence was downgraded for risk of bias and imprecision. We could not compare vitamin and non‐vitamin interventions, risk of bias, and duration of follow‐up because all studies included interventions with vitamins.

1.19. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 19: Vascular access thrombosis

Graft loss

Antioxidants are likely to result in little to no difference in graft loss in kidney transplant recipients (Analysis 1.20 (11 studies, 1053 patients): RR 0.88, 95% CI 0.67 to 1.17; I² = 0%; moderate certainty evidence). The evidence was downgraded for risk of bias. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.18; Analysis 5.19; Analysis 6.20; Analysis 7.20).

1.20. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 20: Graft loss

4.18. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 18: Graft loss

5.19. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 19: Graft loss

6.20. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 20: Graft loss

7.20. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 20: Graft loss

Graft failure

Antioxidants may result in little to no difference in graft failure in kidney transplant recipients, but the evidence is very uncertain (Analysis 1.21. (1 study, 61 patients): RR 0.85, 95% CI 0.06 to 12.95; very low certainty evidence). The evidence was downgraded twice for risk of bias and twice for imprecision. We could not compare vitamin and non‐vitamin interventions, risk of bias, and duration of follow‐up because all studies included interventions with vitamins.

1.21. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 21: Graft failure

Delayed graft function

Antioxidants may result in little to no difference in delayed graft function in kidney transplant recipients, but the evidence is very uncertain (Analysis 1.22 (11 studies, 859 patients): RR 0.90, 95% CI 0.75 to 1.09; I² = 0%; moderate certainty of evidence). The evidence was downgraded for risk of bias. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.19; Analysis 5.21; Analysis 6.21; Analysis 7.22).

1.22. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 22: Delayed graft function

4.19. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 19: Delayed graft function

5.21. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 21: Delayed graft function

6.21. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 21: Delayed graft function

7.22. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 22: Delayed graft function

Malignancy

Antioxidants may result in little to no difference in the risk for malignancies in adults with CKD (Analysis 1.23 (3 studies, 507 patients): RR 0.67, 95% CI 0.12 to 3.36; I² = 0%; low certainty evidence). The evidence was downgraded for risk of bias and imprecision. The effect did not appear to differ between patients with CKD 3‐5, and CKD 5D (test for subgroup differences: Chi² = 0.65, I² = 0%, P = 0.42). The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.20; Analysis 5.22; Analysis 6.22; Analysis 7.23).

1.23. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 23: Malignancy

4.20. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 20: Malignancy

5.22. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 22: Malignancy

6.22. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 22: Malignancy

7.23. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 23: Malignancy

Major bleeding

Antioxidants may result in little to no difference in major bleeding in adults with CKD (Analysis 1.24 (6 studies, 449 patients): RR 1.13, 95% CI 0.41 to 3.12; I² = 0%; low certainty evidence). The evidence was downgraded twice for imprecision. The effect did not appear to differ between patients with CKD 3‐5 and CKD 5D (test for subgroup differences: Chi² = 0.45, I² = 0%, P = 0.50). The efficacy for kidney transplant recipients could not be estimated as none of the studies included these patients. No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.21; Analysis 5.23; Analysis 6.23; Analysis 7.24).

1.24. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 24: Major bleeding

4.21. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 21: Major bleeding

5.23. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 23: Major bleeding

6.23. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 23: Major bleeding

7.24. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 24: Major bleeding

Infection

Antioxidants probably increase the risk of infection in adults with CKD (Analysis 1.25 (14 studies, 3697 patients): RR 1.30, 95% CI 1.14 to 1.50; I² = 3%; moderate certainty evidence). The evidence was downgraded for risk of bias. The increased risk for infection was only observed for adults with CKD 3‐5 (RR 1.46, 95% CI 1.28 to 1.66) but not for those undergoing dialysis (RR 1.39, 95% CI 0.73 to 2.65) or kidney transplant recipients (RR 1.00, 95% CI 0.78 to 1.29). No differences were observed between interventions with and without vitamins, risk of bias, or duration of follow‐up (Analysis 4.22; Analysis 5.24; Analysis 6.24; Analysis 7.25).

1.25. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 25: Infection

4.22. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 22: Infection

5.24. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 24: Infection

6.24. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 24: Infection

7.25. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 25: Infection

Sensitivity analyses: impact of imputing standard deviations for change

Imputing SD for change in eGFR, change in SCr, and change in UACR is unlikely to have introduced bias. The results of the sensitivity analysis, including only the studies without imputed SD for change, were comparable to the analyses with studies with imputed SD (Analysis 1.26; Analysis 1.27; Analysis 1.28). We could not perform a sensitivity analysis for studies without imputed SD for change in proteinuria and change in albuminuria because we had to impute SD for all included studies.

1.26. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 26: Sensitivity analysis: change in eGFR without imputed SD

1.27. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 27: Sensitivity analysis: change in serum creatinine without imputed SD

1.28. Analysis.

Comparison 1: Antioxidants versus placebo or standard therapy, Outcome 28: Sensitivity analysis: change in urinary albumin/creatinine ratio without imputed SD

Pomegranate juice versus pomegranate extract

Death (any cause)

We could not estimate the effect of pomegranate juice versus pomegranate extract because no events occurred in either study arm during the conduct of the study.

Zinc acetate versus polaprezinc

Death (any cause)

Zinc acetate may increase the risk of death compared to polaprezinc in adults with CKD, but the evidence is very uncertain (Analysis 3.1 (1 study, 91 patients): RR 11.73, 95% CI 0.67 to 206.19; very low certainty evidence). The evidence was downgraded once for risk of bias and twice for imprecision.

3.1. Analysis.

Comparison 3: Zinc acetate versus polaprezinc, Outcome 1: Death (any cause)

Discussion

Summary of main results

This updated review of the evidence of antioxidant therapy for the prevention of CVD and deterioration of kidney disease in adults with CKD included 95 studies involving 10,468 randomised patients and 49 different antioxidant interventions. Risk of bias, imprecision, and inconsistencies between studies generally limited the quality of evidence, and the large number of interventions hampers drawing conclusions on the efficacy and safety of single antioxidant agents. Overall, antioxidants probably do not reduce all‐cause or cardiovascular death in adults with CKD. For cardiovascular outcomes, antioxidants possibly reduce overall CVD in adults with CKD, particularly in those not undergoing dialysis or who have received a kidney transplant, but are likely to increase the risk of heart failure. Overall, antioxidants may not reduce coronary heart disease in adults with CKD and also have little to no effect on the risk of cerebrovascular disease and peripheral arterial disease.

The effects of antioxidants on kidney endpoints were heterogeneous. Antioxidants probably reduce the risk of developing kidney failure in adults with CKD 3‐5 and kidney transplant recipients. In addition, antioxidants may increase eGFR and reduce SCr. However, no effect was observed on markers for kidney damage (UACR, proteinuria, or albuminuria), thrombosis of vascular access, or kidney transplant outcomes (graft loss, graft failure or DGF).

No difference was observed in the risk of cancer or major bleeding in adults with CKD. In contrast, antioxidants likely increase the risk of infection.

Overall completeness and applicability of evidence

We studied the effects of antioxidants on a broad range of cardiovascular and kidney endpoints. The overlap in our outcomes of interest allows us to cross‐validate the results. Although our results suggest that antioxidants overall reduce CVD in adults with CKD, it remains unclear where this reduction stems from as we did not observe a reduced risk for any of the individual cardiovascular endpoints.

A possible cardiovascular concern was an observed increase in the risk of heart failure. This risk was driven by BEACON 2013 and HOPE 1996, and was the reason why BEACON 2013 was prematurely terminated. For bardoxolone methyl, an increased heart failure risk was mainly observed in the first four weeks after initiation in patients with a history of heart failure or elevated B‐type natriuretic peptide at baseline, which can be explained by inhibition of the endothelin pathway leading to sodium and fluid retention (Chin 2014; Kohan 2006). Nevertheless, results across the different trials are inconsistent, and a dose‐response relation is lacking as BEAM 2011 and TSUBAKI 2020 report no elevated risks for heart failure despite the prescription of significantly higher doses of bardoxolone methyl. The increased risk for heart failure with vitamin E in HOPE 1996 also appears to be an anomaly, as an increased heart failure risk is not reported in any of the other studies with vitamin E (Shah 2021; Saremi 2010).

The results on kidney outcomes are inconclusive. The reduced risk for kidney failure is corroborated by an increase in eGFR and a decrease in SCr. This increase in eGFR and decrease in SCr were in the same order of magnitude. A reduction in SCr of 13 µmol/L in an average patient in the included studies (aged 60 with a baseline SCr of 150 to 250 µmol/L) would constitute an increase in eGFR of 2 to 5 mL/min/1.73 m² which is similar to the effect we observed in our analyses. Improvements in GFR and SCr are clinically significant because they may prevent progression to kidney failure or delay dialysis initiation, and for some patients, dialysis can be delayed until GFR is as low as 7.0 mL/min/1.73 m² (Cooper 2010). No difference in change of proteinuria was observed. The underlying mechanism for this improvement in kidney function with antioxidant therapy is poorly understood. A lack of change in proteinuria suggests that the renoprotective effect of antioxidants might not be mediated by improving the structural integrity of the glomerulus. An alternative explanation could be a reduction in kidney fibrosis from free radicals (Oberg 2004).

Surprisingly, we observed an increase in the risk of infection. Although the included infections were heterogeneous in nature, almost all were serious, requiring hospitalisation. This finding was unexpected since previous studies have reported that antioxidants enhance the immune system (Wintergerst 2007). Further research should address this matter and identify possible pathophysiological mechanisms behind this association.

We stratified all analyses by CKD stage to assess potential differences in the effects of antioxidants in adults with different stages of CKD. Unfortunately, we could include a limited number of studies with adults undergoing dialysis or kidney transplant recipients. Consequently, results for these patients must be interpreted cautiously since the observed effects are mainly based on studies in adults with earlier stages of CKD.

The generally short duration of follow‐up and heterogeneity in included interventions challenge the applicability of our results in clinical practice. Only 32 studies (36%) had a follow‐up of at least six months. The large number of different antioxidant therapies makes it difficult to compare antioxidant regimens head‐to‐head and conclude which intervention is the most efficacious. More importantly, we did not associate any of the observed effects with changes in oxidative stress markers or antioxidant levels. Oxidative stress markers were not included in the set of outcomes because we are only interested in hard endpoints directly relevant to patients. Since antioxidants affect various in vivo pathways unrelated to their antioxidant effects, any observed effect of antioxidants might be driven by mechanisms unrelated to their antioxidant effects.

Quality of the evidence

The quality of the included studies varied but was generally suboptimal. Many studies did not report in detail if or how they generated the randomisation sequences, concealed the allocation of treatment, and blinded the patients and personnel. Selective outcome reporting could not be assessed for the majority of studies because they did not publish a protocol in a clinical trial registry or as a separate publication.

To explore the impact of biased studies on the overall effect, we performed a subgroup analysis including only low risk of bias studies and studies with longer follow‐up. To account for improved standards of reporting trial methodology, we formulated a more lenient definition of low risk of bias studies, including only low risk of bias in the randomisation sequence generation, blinding of patients and personnel, and attrition bias. Overall, the results of our sensitivity analysis corroborated our findings from the main results: apart from a reduction in proteinuria and a reduction in the risk for coronary heart disease in the studies with a long follow‐up, results in the studies with longer follow‐up were comparable to the overall results for cardiovascular and kidney outcomes. For CVD, we even observed a stronger protective effect from antioxidants in the low risk of bias studies. These results suggest that the observed effects of antioxidants are robust and are not driven by studies with a flawed study design or a short period of follow‐up.

Potential biases in the review process

This review was conducted in line with the Cochrane Handbook (Higgins 2022). To minimise bias in the review process, screening of search results on title and abstract, screening of full texts, data extraction, risk of bias assessment, and GRADE assessment were performed independently by two authors. This approach limited the influence of individual authors on the review process. We completed a highly sensitive search of the Cochrane Kidney and Transplant's Specialised Register without language restriction up to November 2022. The registry contains hand‐searched literature and conference proceedings, maximising the inclusion of grey literature in this review. We additionally requested data from authors in studies for which information was incomplete.

Potential biases in this review mainly stem from limitations in the available data. Although the results of our meta‐analysis remain consistent in the sensitivity analyses, imputing SD for change may have biased effect estimates. Another limitation is the lack of studies which included adults undergoing dialysis or kidney transplant recipients, which prevents us from obtaining robust effect estimates of the effect of antioxidants in these patient populations.

Agreements and disagreements with other studies or reviews

We found no other systematic reviews that assessed the effect of antioxidants on cardiovascular outcomes in adults with CKD. Two systematic reviews of healthy participants and people with various chronic conditions report no effect of antioxidants on death (any cause) (Bjelakovic 2012; Jenkins 2020), cardiovascular death, CVD in general, and individual cardiovascular endpoints (Jenkins 2020). The difference in results ‐ we did find a possible reduction in fatal and non‐fatal cardiovascular events due to antioxidants in adults with CKD – may be explained by differences in study populations. Contrary to our review, Bjelakovic 2012 and Jenkins 20200 also included studies in healthy people and people with a broad range of chronic diseases who have a lower baseline risk and consequently may benefit less. Furthermore, Bjelakovic 2012 and Jenkins 2020, primarily focussed on the effects of vitamins and minerals like zinc and selenium.

We identified two systematic reviews that assessed the effect of antioxidants on kidney endpoints: a systematic review on the nephroprotective effects of antioxidants in people with CKD found an increase in eGFR and reduction in albuminuria (Casanova 2021) and a systematic review on the effect of antioxidants for delaying kidney disease in people with diabetic kidney disease observed a reduction in urinary albumin excretion but no effect on eGFR (Bolignano 2017). Our results differ from both Bolignano 2017 and Casanova 2021 in multiple aspects. Unlike Casanova 2021, we did not observe an effect of antioxidants on albuminuria, while Bolignano 2017 did not find an effect on SCr and eGFR, and Casanova 2021 found an increase in eGFR but no effect on SCr as would be expected. These discrepancies might be due to heterogeneity in study populations. Bolignano 2017 and Casanova 2021 focussed not only on studies in people with CKD but also included studies in people with diabetic nephropathy, which is characterised by proteinuria rather than a reduced eGFR. We chose to exclude most of these studies since the eGFR of included patients was often above 60 mL/min/1.73 m². Therefore, the patients included in these reviews had a higher eGFR than the patients in this study and lower levels of proteinuria. Additionally, the included interventions differed in some aspects. Bolignano 2017 primarily focussed on the effects of vitamins and minerals like zinc and selenium, while Casanova 2021 also included regular medications with possible antioxidant effects like allopurinol and pravastatin.

Authors' conclusions

Implications for practice.

We found no evidence that antioxidants reduce (cardiovascular) death in adults with CKD, which was our primary endpoint, reduce markers of kidney damage, or improve kidney transplant outcomes. The results of this review show some effects of antioxidants in reducing combined fatal and non‐fatal cardiovascular events and kidney failure and improving kidney function in adults with CKD. On the other hand, antioxidants may not be harmless as they possibly increase the risk of serious infections and heart failure. Information on the effects of antioxidants on cardiovascular and kidney outcomes in adults undergoing dialysis or kidney transplant recipients remains limited since we found few studies that included these patient populations.

Our results must be interpreted with caution as their interpretation is hampered by the lack of methodological quality of most of the included studies, heterogeneity in the included interventions, and limited duration of follow‐up. Analyses in studies with a low risk of bias support possible benefits of antioxidants in adults with CKD, but the number of participants and events was too small to be conclusive. Without studies of adequate quality and follow‐up, it is premature to include recommendations in clinical guidelines about antioxidants for the prevention of cardiovascular and kidney disease in adults with CKD.

Implications for research.

The lack of high‐quality RCTs prevents definite conclusions about the effects of antioxidant therapy on death, CVD and (the progression of ) kidney disease in adults with CKD. Particularly for patients undergoing dialysis and kidney transplant recipients, information is currently lacking. A possible increased risk of infection also requires further investigation. Any further attempts to obtain clarity about the effect of antioxidants on cardiovascular endpoints, kidney endpoints and possible adverse events require further well‐designed RCTs with adequate sample size and follow‐up in which cardiovascular endpoints, kidney endpoints, and adverse events are assessed. To strengthen evidence that any effects are due to the antioxidant actions of the supplements, markers of oxidative stress should be measured to associate outcomes with changes in oxidative stress levels. Such RCTs should include patients from the whole CKD spectrum (i.e. CKD 3‐5, CKD 5D and CKD 5‐transplant). The diversity in antioxidant supplements makes it challenging to select an antioxidant to study for an RCT. A multi‐arm placebo‐controlled RCT, therefore, may be the most suitable design as it allows the study of the effects of multiple interventions simultaneously.

What's new

Date	Event	Description
2 November 2023	New citation required and conclusions have changed	New studies (86) comparisons and outcomes included
2 November 2023	New search has been performed	New studies, interventions and outcomes included

History

Protocol first published: Issue 4, 2009
Review first published: Issue 10, 2012

Date	Event	Description
3 December 2021	Amended	Search strategies in Appendix 1 amended
7 May 2014	Amended	Minor copy edit made to study names

Appendices

Appendix 1. Electronic search strategies

Database	Search terms
CENTRAL	MeSH descriptor: [Antioxidants] explode all trees MeSH descriptor: [Carotenoids] 1 tree(s) exploded tocopherol*:ti,ab,kw acetylcysteine:ti,ab,kw antioxidant*:ti,ab,kw carotene*:ti,ab,kw carotenoid*:ti,ab,kw "ascorbic acid":ti,ab,kw "vitamin A":ti,ab,kw "vitamin C":ti,ab,kw "vitamin E":ti,ab,kw selenium:ti,ab,kw coenzyme q10:ti,ab,kw silymarin:ti,ab,kw milk thistle:ti,ab,kw curcumin:ti,ab,kw turmeric:ti,ab,kw astaxanthin:ti,ab,kw resveratrol:ti,ab,kw sodium nitrite*:ti,ab,kw pomegranate*:ti,ab,kw grape seed extract*:ti,ab,kw ginsenoside rb*:ti,ab,kw ginseng*:ti,ab,kw zinc:ti,ab,kw polyphenol*:ti,ab,kw {or 1‐26} "renal replacement therapy":ti,ab,kw dialysis:ti,ab,kw (hemodialysis or haemodialysis):ti,ab,kw (PD or CAPD or CCPD or APD):ti,ab,kw "renal insufficiency":ti,ab,kw MeSH descriptor: [Renal Insufficiency, Chronic] explode all trees (kidney next disease*):ti,ab,kw "kidney failure":kw uremic:ti,ab,kw or uraemia:ti,ab,kw or uremia:ti,ab,kw ((endstage next renal) or (endstage next kidney) or (end next stage next renal) or (end next stage next kidney)):ti,ab,kw ((chronic next kidney) or (chronic next renal)):ti,ab,kw (CKD or CKF or CRD or CRF):ti,ab,kw (ESRF or ESKF or ESRD or ESKD):ti,ab,kw ((kidney next transplant*) or (renal next transplant*) or (kidney next graft*) or (renal next graft*)):ti,ab,kw {or 28‐41} {AND 27,42}
MEDLINE	Kidney Diseases/ exp Renal Replacement Therapy/ Renal Insufficiency/ exp Renal Insufficiency, Chronic/ Diabetic Nephropathies/ diabetic kidney disease$.tw. diabetic nephropath$.tw. exp Hypertension, Renal/ dialysis.tw. (hemodialysis or haemodialysis).tw. (hemofiltration or haemofiltration).tw. (hemodiafiltration or haemodiafiltration).tw. (kidney disease* or renal disease* or kidney failure or renal failure).tw. (ESRF or ESKF or ESRD or ESKD).tw. (CKF or CKD or CRF or CRD).tw. (CAPD or CCPD or APD).tw. (predialysis or pre‐dialysis).tw. Uremia/ (uremic or ur?emia).tw. or/1‐19 exp Antioxidants/ exp Carotenoids/ exp Tocopherols/ exp Flavonoids/ Acetylcysteine/ Selenium/ antioxidant$.tw. acetylcysteine.tw. carotene$.tw. carotenoid$.tw. ascorbic acid.tw. vitamin A.tw. vitamin E.tw. vitamin C.tw. tocopherol$.tw. flavonoid$.tw. bardoxolone methyl.tw. selenium.tw. coenzyme q10.tw. exp Silymarin/ silymarin.tw. milk thistle$.tw. Curcumin/ curcumin.tw. turmeric.tw. astaxanthin.tw. Resveratrol/ resveratrol.tw. Sodium Nitrite/ sodium nitrite$.tw. Pomegranate/ pomegranate$.tw. Grape Seed Extract/ grape seed extract$.tw. Ginsenosides/ ginsenoside rb$.tw. ginseng.tw. Zinc/ exp Zinc Compounds/ zinc.tw. exp Polyphenols/ polyphenol$.tw. or/21‐62 and/20,63
EMBASE	exp renal replacement therapy/ kidney disease/ chronic kidney disease/ kidney failure/ chronic kidney failure/ mild renal impairment/ stage 1 kidney disease/ moderate renal impairment/ severe renal impairment/ end stage renal disease/ renal replacement therapy‐dependent renal disease/ diabetic nephropathy/ kidney transplantation/ renovascular hypertension/ (hemodialysis or haemodialysis).tw. (hemofiltration or haemofiltration).tw. (hemodiafiltration or haemodiafiltration).tw. dialysis.tw. (CAPD or CCPD or APD).tw. (kidney disease* or renal disease* or kidney failure or renal failure).tw. (CKF or CKD or CRF or CRD).tw. (ESRF or ESKF or ESRD or ESKD).tw. (predialysis or pre‐dialysis).tw. ((kidney or renal) adj (transplant* or graft* or allograft*)).tw. or/1‐24 exp Antioxidant/ exp Ascorbic Acid/ exp Carotenoid/ exp Tocopherol/ Acetylcysteine/ exp Flavonoid/ antioxidant$.tw. acetylcysteine.tw. carotene$.tw. carotenoid$.tw. ascorbic acid.tw. vitamin A.tw. vitamin C.tw. vitamin E.tw. tocopherol$.tw. flavonoid$.tw. selenium/ selenium.tw. ubidecarenone/ coenzyme q10.tw. astaxanthin/ astaxanthin.tw. silymarin/ silymarin.tw. milk thistle$.tw. sodium nitrite/ sodium nitrite.tw. resveratrol/ resveratrol.tw. pomegranate/ or pomegranate extract/ or pomegranate juice/ pomegranate$.tw. or/26‐56 and/25,5

Appendix 2. Risk of bias assessment tool

Potential source of bias	Assessment criteria
Random sequence generation Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence	Low risk of bias: Random number table; computer random number generator; coin tossing; shuffling cards or envelopes; throwing dice; drawing of lots; minimization (minimization may be implemented without a random element, and this is considered to be equivalent to being random).
	High risk of bias: Sequence generated by odd or even date of birth; date (or day) of admission; sequence generated by hospital or clinic record number; allocation by judgement of the clinician; by preference of the participant; based on the results of a laboratory test or a series of tests; by availability of the intervention.
	Unclear: Insufficient information about the sequence generation process to permit judgement.
Allocation concealment Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment	Low risk of bias: Randomisation method described that would not allow investigator/participant to know or influence intervention group before eligible participant entered in the study (e.g. central allocation, including telephone, web‐based, and pharmacy‐controlled, randomisation; sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes).
	High risk of bias: Using an open random allocation schedule (e.g. a list of random numbers); assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or non‐opaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure.
	Unclear: Randomisation stated but no information on method used is available.
Blinding of participants and personnel Performance bias due to knowledge of the allocated interventions by participants and personnel during the study	Low risk of bias: No blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding; blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding; blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Blinding of outcome assessment Detection bias due to knowledge of the allocated interventions by outcome assessors.	Low risk of bias: No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Incomplete outcome data Attrition bias due to amount, nature or handling of incomplete outcome data.	Low risk of bias: No missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods.
	High risk of bias: Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; ‘as‐treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.
	Unclear: Insufficient information to permit judgement
Selective reporting Reporting bias due to selective outcome reporting	Low risk of bias: The study protocol is available and all of the study’s pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way; the study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre‐specified (convincing text of this nature may be uncommon).
	High risk of bias: Not all of the study’s pre‐specified primary outcomes have been reported; one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. subscales) that were not pre‐specified; one or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); one or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta‐analysis; the study report fails to include results for a key outcome that would be expected to have been reported for such a study.
	Unclear: Insufficient information to permit judgement
Other bias Bias due to problems not covered elsewhere in the table	Low risk of bias: The study appears to be free of other sources of bias.
	High risk of bias: Had a potential source of bias related to the specific study design used; stopped early due to some data‐dependent process (including a formal‐stopping rule); had extreme baseline imbalance; has been claimed to have been fraudulent; had some other problem.
	Unclear: Insufficient information to assess whether an important risk of bias exists; insufficient rationale or evidence that an identified problem will introduce bias.

Data and analyses

Comparison 1. Antioxidants versus placebo or standard therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Cardiovascular death	8	3813	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.64, 1.40]
1.1.1 CKD stages 3‐5	3	3269	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.46, 1.74]
1.1.2 CKD stage 5D	4	444	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.49, 1.44]
1.1.3 Kidney transplant	1	100	Risk Ratio (M‐H, Random, 95% CI)	5.00 [0.25, 101.58]
1.2 Death (any cause)	45	7530	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.82, 1.11]
1.2.1 CKD stages 3‐5	13	4229	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.82, 1.20]
1.2.2 CKD stage 5D	21	2261	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.69, 1.22]
1.2.3 Kidney transplant	11	1040	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.53, 1.28]
1.3 Cardiovascular disease	16	4768	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.63, 0.99]
1.3.1 CKD stages 3‐5	8	4050	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.59, 0.99]
1.3.2 CKD stage 5D	6	557	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.47, 1.62]
1.3.3 Kidney transplant	2	161	Risk Ratio (M‐H, Random, 95% CI)	5.00 [0.25, 101.58]
1.4 Coronary heart disease	12	4372	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.64, 1.12]
1.4.1 CKD stages 3‐5	7	3811	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.76, 1.24]
1.4.2 CKD stage 5D	6	561	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.30, 1.15]
1.5 Heart failure	6	3733	Risk Ratio (M‐H, Random, 95% CI)	1.40 [1.11, 1.76]
1.5.1 CKD stages 3‐5	6	3733	Risk Ratio (M‐H, Random, 95% CI)	1.40 [1.11, 1.76]
1.6 Cerebrovascular disease	8	1768	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.71, 1.32]
1.6.1 CKD stages 3‐5	3	1147	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.71, 1.38]
1.6.2 CKD stage 5D	4	521	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.31, 1.98]
1.6.3 Kidney transplant	1	100	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.13, 71.92]
1.7 Peripheral vascular disease	4	1492	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.41, 1.47]
1.7.1 CKD stages 3‐5	2	1162	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.34, 2.90]
1.7.2 CKD stage 5D	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.26, 1.12]
1.8 Kidney failure	10	3201	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.41, 1.02]
1.8.1 CKD stages 3‐5	9	3024	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.40, 1.17]
1.8.2 Kidney transplant	1	177	Risk Ratio (M‐H, Random, 95% CI)	0.53 [0.26, 1.11]
1.9 eGFR at end of study	32	2435	Mean Difference (IV, Random, 95% CI)	3.85 [2.13, 5.56]
1.9.1 CKD stages 3‐5	22	1597	Mean Difference (IV, Random, 95% CI)	3.51 [1.47, 5.54]
1.9.2 Kidney transplant	10	838	Mean Difference (IV, Random, 95% CI)	4.75 [1.46, 8.04]
1.10 Change in eGFR	28	4128	Mean Difference (IV, Random, 95% CI)	3.65 [2.81, 4.50]
1.10.1 CKD stages 3‐5	26	4057	Mean Difference (IV, Random, 95% CI)	3.72 [2.85, 4.58]
1.10.2 Kidney transplant	2	71	Mean Difference (IV, Random, 95% CI)	3.97 [‐4.18, 12.13]
1.11 Serum creatinine at end of study	22	1587	Mean Difference (IV, Random, 95% CI)	‐6.24 [‐15.71, 3.23]
1.11.1 CKD stages 3‐5	9	562	Mean Difference (IV, Random, 95% CI)	1.70 [‐12.38, 15.79]
1.11.2 Kidney transplant	13	1025	Mean Difference (IV, Random, 95% CI)	‐13.87 [‐28.02, 0.28]
1.12 Change in serum creatinine	16	3180	Mean Difference (IV, Random, 95% CI)	‐13.35 [‐23.49, ‐3.20]
1.12.1 CKD stages 3‐5	11	2970	Mean Difference (IV, Random, 95% CI)	‐13.64 [‐24.75, ‐2.53]
1.12.2 Kidney transplant	5	210	Mean Difference (IV, Random, 95% CI)	‐18.90 [‐46.82, 9.01]
1.13 Urinary albumin/creatinine ratio at end of study	5	1174	Mean Difference (IV, Random, 95% CI)	‐0.59 [‐12.78, 11.59]
1.13.1 CKD stages 3‐5	5	1174	Mean Difference (IV, Random, 95% CI)	‐0.59 [‐12.78, 11.59]
1.14 Change in urinary albumin/creatinine ratio	7	1286	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.55, 0.47]
1.14.1 CKD stages 3‐5	7	1286	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.55, 0.47]
1.15 Proteinuria at end of study	9	536	Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.78, 0.31]
1.15.1 CKD stages 3‐5	9	536	Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.78, 0.31]
1.16 Change in proteinuria	8	468	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐1.11, 0.09]
1.16.1 CKD stages 3‐5	8	468	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐1.11, 0.09]
1.17 Albuminuria at end of study	4	395	Mean Difference (IV, Random, 95% CI)	94.75 [0.57, 188.94]
1.17.1 CKD stages 3‐5	4	395	Mean Difference (IV, Random, 95% CI)	94.75 [0.57, 188.94]
1.18 Change in albuminuria	4	395	Mean Difference (IV, Random, 95% CI)	‐87.08 [‐221.12, 46.97]
1.18.1 CKD stages 3‐5	4	395	Mean Difference (IV, Random, 95% CI)	‐87.08 [‐221.12, 46.97]
1.19 Vascular access thrombosis	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
1.19.1 CKD stage 5D	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
1.20 Graft loss	11	1053	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.67, 1.17]
1.20.1 Kidney transplant	11	1053	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.67, 1.17]
1.21 Graft failure	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
1.21.1 Kidney transplant	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
1.22 Delayed graft function	11	859	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.75, 1.09]
1.22.1 Kidney transplant	11	859	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.75, 1.09]
1.23 Malignancy	3	507	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.12, 3.63]
1.23.1 CKD stages 3‐5	1	227	Risk Ratio (M‐H, Random, 95% CI)	0.34 [0.05, 2.33]
1.23.2 CKD stage 5D	2	280	Risk Ratio (M‐H, Random, 95% CI)	1.33 [0.09, 20.26]
1.24 Major bleeding	6	449	Risk Ratio (M‐H, Random, 95% CI)	1.13 [0.41, 3.12]
1.24.1 CKD stages 3‐5	2	96	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.07, 5.97]
1.24.2 CKD stage 5D	4	353	Risk Ratio (M‐H, Random, 95% CI)	1.55 [0.45, 5.31]
1.25 Infection	14	3697	Risk Ratio (M‐H, Random, 95% CI)	1.30 [1.14, 1.50]
1.25.1 CKD stages 3‐5	6	2802	Risk Ratio (M‐H, Random, 95% CI)	1.46 [1.28, 1.66]
1.25.2 CKD stage 5D	5	616	Risk Ratio (M‐H, Random, 95% CI)	1.39 [0.73, 2.65]
1.25.3 Kidney transplant	3	279	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.78, 1.29]
1.26 Sensitivity analysis: change in eGFR without imputed SD	9	3024	Mean Difference (IV, Random, 95% CI)	4.17 [2.93, 5.42]
1.26.1 CKD stages 3‐5	8	3014	Mean Difference (IV, Random, 95% CI)	4.09 [2.84, 5.34]
1.26.3 Kidney transplant	1	10	Mean Difference (IV, Random, 95% CI)	9.60 [0.12, 19.08]
1.27 Sensitivity analysis: change in serum creatinine without imputed SD	3	297	Mean Difference (IV, Random, 95% CI)	‐0.12 [‐13.90, 13.65]
1.27.1 CKD stages 3‐5	2	287	Mean Difference (IV, Random, 95% CI)	‐5.43 [‐25.61, 14.74]
1.27.3 Kidney transplant	1	10	Mean Difference (IV, Random, 95% CI)	11.30 [‐7.76, 30.36]
1.28 Sensitivity analysis: change in urinary albumin/creatinine ratio without imputed SD	2	145	Mean Difference (IV, Random, 95% CI)	0.00 [‐0.14, 0.14]
1.28.1 CKD stages 3‐5	2	145	Mean Difference (IV, Random, 95% CI)	0.00 [‐0.14, 0.14]

Comparison 2. Pomegranate juice versus pomegranate extract.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Death (any cause)	1	24	Risk Ratio (M‐H, Random, 95% CI)	Not estimable

2.1. Analysis.

Comparison 2: Pomegranate juice versus pomegranate extract, Outcome 1: Death (any cause)

Comparison 3. Zinc acetate versus polaprezinc.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Death (any cause)	1	91	Risk Ratio (M‐H, Random, 95% CI)	11.73 [0.67, 206.19]

Comparison 4. Vitamin antioxidants versus placebo or standard therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Cardiovascular death	3	1223	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.59, 1.28]
4.1.1 CKD stages 3‐5	1	993	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.70, 1.40]
4.1.2 CKD stage 5D	2	230	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.27, 1.21]
4.2 Death (any cause)	13	2469	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.74, 1.14]
4.2.1 CKD stages 3‐5	2	1054	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.69, 1.18]
4.2.2 CKD stage 5D	10	1353	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.67, 1.42]
4.2.3 Kidney transplant	1	62	Risk Ratio (M‐H, Random, 95% CI)	0.50 [0.10, 2.53]
4.3 Cardiovascular disease	7	1658	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.49, 1.32]
4.3.1 CKD stages 3‐5	3	1282	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.38, 1.30]
4.3.2 CKD stage 5D	4	376	Risk Ratio (M‐H, Random, 95% CI)	1.22 [0.41, 3.59]
4.4 Coronary heart disease	6	1457	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.29, 1.26]
4.4.1 CKD stages 3‐5	3	1110	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.50, 1.52]
4.4.2 CKD stage 5D	4	347	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.12, 1.75]
4.5 Heart failure	1	993	Risk Ratio (M‐H, Random, 95% CI)	1.30 [0.96, 1.77]
4.5.1 CKD stages 3‐5	1	993	Risk Ratio (M‐H, Random, 95% CI)	1.30 [0.96, 1.77]
4.6 Cerebrovascular disease	5	1454	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.72, 1.37]
4.6.1 CKD stages 3‐5	3	1147	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.71, 1.38]
4.6.2 CKD stage 5D	2	307	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.35, 3.06]
4.7 Peripheral vascular disease	2	1189	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.28, 2.37]
4.7.1 CKD stages 3‐5	1	993	Risk Ratio (M‐H, Random, 95% CI)	1.20 [0.93, 1.55]
4.7.2 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.10, 1.40]
4.8 Kidney failure	1	93	Risk Ratio (M‐H, Random, 95% CI)	1.63 [0.41, 6.43]
4.8.1 CKD stages 3‐5	1	93	Risk Ratio (M‐H, Random, 95% CI)	1.63 [0.41, 6.43]
4.9 eGFR at end of study	6	387	Mean Difference (IV, Random, 95% CI)	0.59 [‐5.97, 7.16]
4.9.1 CKD stages 3‐5	4	241	Mean Difference (IV, Random, 95% CI)	‐1.70 [‐9.28, 5.88]
4.9.2 Kidney transplant	2	146	Mean Difference (IV, Random, 95% CI)	6.45 [4.88, 8.02]
4.10 Change in eGFR	5	251	Mean Difference (IV, Random, 95% CI)	2.42 [‐0.26, 5.10]
4.10.1 CKD stages 3‐5	4	241	Mean Difference (IV, Random, 95% CI)	1.96 [‐0.78, 4.70]
4.10.2 Kidney transplant	1	10	Mean Difference (IV, Random, 95% CI)	9.60 [0.12, 19.08]
4.11 Serum creatinine at end of study	3	133	Mean Difference (IV, Random, 95% CI)	0.51 [‐11.36, 12.38]
4.11.1 CKD stages 3‐5	1	61	Mean Difference (IV, Random, 95% CI)	7.00 [‐12.83, 26.83]
4.11.2 Kidney transplant	2	72	Mean Difference (IV, Random, 95% CI)	‐3.12 [‐17.93, 11.70]
4.12 Change in serum creatinine	2	71	Mean Difference (IV, Random, 95% CI)	4.85 [‐4.14, 13.84]
4.12.1 CKD stages 3‐5	1	61	Mean Difference (IV, Random, 95% CI)	3.00 [‐7.20, 13.20]
4.12.2 Kidney transplant	1	10	Mean Difference (IV, Random, 95% CI)	11.30 [‐7.76, 30.36]
4.13 Urinary albumin/creatinine ratio at end of study	2	1054	Mean Difference (IV, Random, 95% CI)	0.17 [‐2.74, 3.08]
4.13.1 CKD stages 3‐5	2	1054	Mean Difference (IV, Random, 95% CI)	0.17 [‐2.74, 3.08]
4.14 Change in urinary albumin/creatinine ratio	2	1054	Mean Difference (IV, Random, 95% CI)	2.73 [‐6.87, 12.32]
4.14.1 CKD stages 3‐5	2	1054	Mean Difference (IV, Random, 95% CI)	2.73 [‐6.87, 12.32]
4.15 Albuminuria at end of study	1	93	Mean Difference (IV, Random, 95% CI)	‐30.00 [‐342.31, 282.31]
4.15.1 CKD stages 3‐5	1	93	Mean Difference (IV, Random, 95% CI)	‐30.00 [‐342.31, 282.31]
4.16 Change in albuminuria	1	93	Mean Difference (IV, Random, 95% CI)	‐4.00 [‐217.48, 209.48]
4.16.1 CKD stages 3‐5	1	93	Mean Difference (IV, Random, 95% CI)	‐4.00 [‐217.48, 209.48]
4.17 Vascular access thrombosis	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
4.17.1 CKD stage 5D	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
4.18 Graft loss	1	62	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.08, 1.91]
4.18.1 Kidney transplant	1	62	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.08, 1.91]
4.19 Delayed graft function	2	146	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.49, 1.60]
4.19.1 Kidney transplant	2	146	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.49, 1.60]
4.20 Malignancy	1	196	Risk Ratio (M‐H, Random, 95% CI)	5.10 [0.25, 104.92]
4.20.1 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	5.10 [0.25, 104.92]
4.21 Major bleeding	2	238	Risk Ratio (M‐H, Random, 95% CI)	1.29 [0.22, 7.52]
4.21.1 CKD stage 5D	2	238	Risk Ratio (M‐H, Random, 95% CI)	1.29 [0.22, 7.52]
4.22 Infection	5	606	Risk Ratio (M‐H, Random, 95% CI)	1.39 [0.73, 2.64]
4.22.1 CKD stages 3‐5	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.06, 14.78]
4.22.2 CKD stage 5D	4	545	Risk Ratio (M‐H, Random, 95% CI)	1.42 [0.73, 2.74]

4.16. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 16: Change in albuminuria

4.17. Analysis.

Comparison 4: Vitamin antioxidants versus placebo or standard therapy, Outcome 17: Vascular access thrombosis

Comparison 5. Non‐vitamin antioxidants versus placebo or standard therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Cardiovascular death	5	2590	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.52, 2.19]
5.1.1 CKD stages 3‐5	2	2276	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.08, 4.19]
5.1.2 CKD stage 5D	2	214	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.58, 2.83]
5.1.3 Kidney transplant	1	100	Risk Ratio (M‐H, Random, 95% CI)	5.00 [0.25, 101.58]
5.2 Death (any cause)	32	5061	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.80, 1.21]
5.2.1 CKD stages 3‐5	11	3175	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.84, 1.45]
5.2.2 CKD stage 5D	11	908	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.55, 1.31]
5.2.3 Kidney transplant	10	978	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.54, 1.35]
5.3 Cardiovascular disease	9	3110	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.56, 0.93]
5.3.1 CKD stages 3‐5	5	2768	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.55, 0.96]
5.3.2 CKD stage 5D	2	181	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.34, 1.29]
5.3.3 Kidney transplant	2	161	Risk Ratio (M‐H, Random, 95% CI)	5.00 [0.25, 101.58]
5.4 Coronary heart disease	6	2915	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.61, 1.48]
5.4.1 CKD stages 3‐5	4	2701	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.63, 1.96]
5.4.2 CKD stage 5D	2	214	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.36, 1.50]
5.5 Heart failure	5	2740	Risk Ratio (M‐H, Random, 95% CI)	1.55 [1.09, 2.20]
5.5.1 CKD stages 3‐5	5	2740	Risk Ratio (M‐H, Random, 95% CI)	1.55 [1.09, 2.20]
5.6 Cerebrovascular disease	3	314	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.16, 4.11]
5.6.1 CKD stage 5D	2	214	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.08, 4.79]
5.6.2 Kidney transplant	1	100	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.13, 71.92]
5.7 Peripheral vascular disease	2	303	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.25, 1.35]
5.7.1 CKD stages 3‐5	1	169	Risk Ratio (M‐H, Random, 95% CI)	0.21 [0.01, 4.35]
5.7.2 CKD stage 5D	1	134	Risk Ratio (M‐H, Random, 95% CI)	0.64 [0.27, 1.52]
5.8 Kidney failure	9	3108	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.38, 0.96]
5.8.1 CKD stages 3‐5	8	2931	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.35, 1.10]
5.8.2 Kidney transplant	1	177	Risk Ratio (M‐H, Random, 95% CI)	0.53 [0.26, 1.11]
5.9 eGFR at end of study	26	2048	Mean Difference (IV, Random, 95% CI)	4.71 [2.88, 6.54]
5.9.1 CKD stages 3‐5	18	1356	Mean Difference (IV, Random, 95% CI)	4.83 [2.75, 6.91]
5.9.3 Kidney transplant	8	692	Mean Difference (IV, Random, 95% CI)	4.55 [0.27, 8.83]
5.10 Change in eGFR	23	3877	Mean Difference (IV, Random, 95% CI)	3.98 [3.09, 4.87]
5.10.1 CKD stages 3‐5	22	3816	Mean Difference (IV, Random, 95% CI)	4.12 [3.21, 5.03]
5.10.3 Kidney transplant	1	61	Mean Difference (IV, Random, 95% CI)	0.90 [‐1.25, 3.05]
5.11 Serum creatinine at end of study	19	1454	Mean Difference (IV, Random, 95% CI)	‐7.01 [‐17.44, 3.42]
5.11.1 CKD stages 3‐5	8	501	Mean Difference (IV, Random, 95% CI)	1.01 [‐13.92, 15.94]
5.11.3 Kidney transplant	11	953	Mean Difference (IV, Random, 95% CI)	‐15.59 [‐32.52, 1.34]
5.12 Change in serum creatinine	14	3109	Mean Difference (IV, Random, 95% CI)	‐16.32 [‐26.87, ‐5.78]
5.12.1 CKD stages 3‐5	10	2909	Mean Difference (IV, Random, 95% CI)	‐15.48 [‐26.84, ‐4.12]
5.12.3 Kidney transplant	4	200	Mean Difference (IV, Random, 95% CI)	‐31.39 [‐68.88, 6.09]
5.13 Urinary albumin/creatinine ratio at end of study	3	120	Mean Difference (IV, Random, 95% CI)	‐61.20 [‐152.33, 29.94]
5.13.1 CKD stages 3‐5	3	120	Mean Difference (IV, Random, 95% CI)	‐61.20 [‐152.33, 29.94]
5.14 Change in urinary albumin/creatinine ratio	5	232	Mean Difference (IV, Random, 95% CI)	‐0.92 [‐7.59, 5.76]
5.14.1 CKD stages 3‐5	5	232	Mean Difference (IV, Random, 95% CI)	‐0.92 [‐7.59, 5.76]
5.15 Proteinuria at end of study	9	536	Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.78, 0.31]
5.15.1 CKD stages 3‐5	9	536	Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.78, 0.31]
5.16 Change in proteinuria	8	468	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐1.11, 0.09]
5.16.1 CKD stages 3‐5	8	468	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐1.11, 0.09]
5.17 Albuminuria at end of study	3	302	Mean Difference (IV, Random, 95% CI)	99.92 [‐10.95, 210.78]
5.17.1 CKD stages 3‐5	3	302	Mean Difference (IV, Random, 95% CI)	99.92 [‐10.95, 210.78]
5.18 Change in albuminuria	3	302	Mean Difference (IV, Random, 95% CI)	‐111.27 [‐280.30, 57.75]
5.18.1 CKD stages 3‐5	3	302	Mean Difference (IV, Random, 95% CI)	‐111.27 [‐280.30, 57.75]
5.19 Graft loss	10	991	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.68, 1.20]
5.19.1 Kidney transplant	10	991	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.68, 1.20]
5.20 Graft failure	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
5.20.1 Kidney transplant	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
5.21 Delayed graft function	9	713	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.75, 1.10]
5.21.1 Kidney transplant	9	713	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.75, 1.10]
5.22 Malignancy	2	311	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.06, 1.73]
5.22.1 CKD stages 3‐5	1	227	Risk Ratio (M‐H, Random, 95% CI)	0.34 [0.05, 2.33]
5.22.2 CKD stage 5D	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.32 [0.01, 7.59]
5.23 Major bleeding	4	211	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.27, 4.48]
5.23.1 CKD stages 3‐5	2	96	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.07, 5.97]
5.23.2 CKD stage 5D	2	115	Risk Ratio (M‐H, Random, 95% CI)	3.05 [0.33, 28.45]
5.24 Infection	9	3091	Risk Ratio (M‐H, Random, 95% CI)	1.22 [0.92, 1.62]
5.24.1 CKD stages 3‐5	5	2741	Risk Ratio (M‐H, Random, 95% CI)	1.46 [1.28, 1.66]
5.24.2 CKD stages 5D	1	71	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.07, 15.81]
5.24.3 Kidney transplant	3	279	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.78, 1.29]

5.15. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 15: Proteinuria at end of study

5.16. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 16: Change in proteinuria

5.18. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 18: Change in albuminuria

5.20. Analysis.

Comparison 5: Non‐vitamin antioxidants versus placebo or standard therapy, Outcome 20: Graft failure

Comparison 6. Antioxidants versus placebo or standard therapy: low risk of bias.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Cardiovascular death	2	230	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.27, 1.21]
6.1.1 CKD stage 5D	2	230	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.27, 1.21]
6.2 Death (any cause)	16	1592	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.66, 1.33]
6.2.1 CKD stages 3‐5	4	498	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.11, 3.69]
6.2.2 CKD stage 5D	10	970	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.69, 1.44]
6.2.3 Kidney transplant	2	124	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.11, 1.95]
6.3 Cardiovascular disease	6	808	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.40, 0.91]
6.3.1 CKD stages 3‐5	4	685	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.40, 0.92]
6.3.2 CKD stage 5D	2	123	Risk Ratio (M‐H, Random, 95% CI)	0.48 [0.04, 6.49]
6.4 Coronary heart disease	6	818	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.21, 0.77]
6.4.1 CKD stages 3‐5	4	513	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.24, 1.59]
6.4.2 CKD stage 5D	3	305	Risk Ratio (M‐H, Random, 95% CI)	0.27 [0.11, 0.67]
6.5 Heart failure	3	435	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.39, 2.86]
6.5.1 CKD stages 3‐5	3	435	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.39, 2.86]
6.6 Cerebrovascular disease	2	289	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.26, 2.25]
6.6.1 CKD stages 3‐5	1	93	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.81]
6.6.2 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.27, 2.70]
6.7 Peripheral vascular disease	2	365	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.11, 1.15]
6.7.1 CKD stages 3‐5	1	169	Risk Ratio (M‐H, Random, 95% CI)	0.21 [0.01, 4.35]
6.7.2 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.10, 1.40]
6.8 Kidney failure	4	529	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.39, 2.37]
6.8.1 CKD stages 3‐5	4	529	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.39, 2.37]
6.9 eGFR at end of study	11	831	Mean Difference (IV, Random, 95% CI)	3.24 [0.25, 6.22]
6.9.1 CKD stages 3‐5	8	659	Mean Difference (IV, Random, 95% CI)	3.34 [‐0.10, 6.77]
6.9.3 Kidney transplant	3	172	Mean Difference (IV, Random, 95% CI)	1.69 [0.05, 3.33]
6.10 Change in eGFR	9	729	Mean Difference (IV, Random, 95% CI)	3.46 [2.32, 4.60]
6.10.1 CKD stages 3‐5	9	729	Mean Difference (IV, Random, 95% CI)	3.46 [2.32, 4.60]
6.11 Serum creatinine at end of study	4	240	Mean Difference (IV, Random, 95% CI)	‐1.98 [‐11.27, 7.32]
6.11.1 CKD stages 3‐5	2	104	Mean Difference (IV, Random, 95% CI)	‐1.61 [‐13.12, 9.89]
6.11.3 Kidney transplant	2	136	Mean Difference (IV, Random, 95% CI)	‐14.96 [‐61.72, 31.80]
6.12 Change in serum creatinine	4	401	Mean Difference (IV, Random, 95% CI)	‐35.01 [‐68.93, ‐1.09]
6.12.1 CKD stages 3‐5	3	327	Mean Difference (IV, Random, 95% CI)	‐18.84 [‐47.75, 10.08]
6.12.3 Kidney transplant	1	74	Mean Difference (IV, Random, 95% CI)	‐97.00 [‐137.94, ‐56.06]
6.13 Urinary albumin/creatinine ratio at end of study	2	78	Mean Difference (IV, Random, 95% CI)	‐111.94 [‐188.36, ‐35.53]
6.13.1 CKD stages 3‐5	2	78	Mean Difference (IV, Random, 95% CI)	‐111.94 [‐188.36, ‐35.53]
6.14 Change in urinary albumin/creatinine ratio	3	115	Mean Difference (IV, Random, 95% CI)	‐37.54 [‐101.00, 25.92]
6.14.1 CKD stages 3‐5	3	115	Mean Difference (IV, Random, 95% CI)	‐37.54 [‐101.00, 25.92]
6.15 Proteinuria at end of study	3	132	Mean Difference (IV, Random, 95% CI)	‐0.30 [‐1.00, 0.39]
6.15.1 CKD stages 3‐5	3	132	Mean Difference (IV, Random, 95% CI)	‐0.30 [‐1.00, 0.39]
6.16 Change in proteinuria	3	132	Mean Difference (IV, Random, 95% CI)	‐0.90 [‐1.80, 0.01]
6.16.1 CKD stages 3‐5	3	132	Mean Difference (IV, Random, 95% CI)	‐0.90 [‐1.80, 0.01]
6.17 Albuminuria at end of study	2	262	Mean Difference (IV, Random, 95% CI)	‐40.57 [‐274.19, 193.04]
6.17.1 CKD stages 3‐5	2	262	Mean Difference (IV, Random, 95% CI)	‐40.57 [‐274.19, 193.04]
6.18 Change in albuminuria	2	262	Mean Difference (IV, Random, 95% CI)	‐172.03 [‐505.20, 161.15]
6.18.1 CKD stages 3‐5	2	262	Mean Difference (IV, Random, 95% CI)	‐172.03 [‐505.20, 161.15]
6.19 Vascular access thrombosis	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
6.19.1 CKD stage 5D	3	209	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.41, 2.84]
6.20 Graft loss	3	198	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.19, 2.30]
6.20.1 Kidney transplant	3	198	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.19, 2.30]
6.21 Delayed graft function	3	172	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.77, 1.54]
6.21.1 Kidney transplant	3	172	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.77, 1.54]
6.22 Malignancy	2	423	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.07, 15.39]
6.22.1 CKD stages 3‐5	1	227	Risk Ratio (M‐H, Random, 95% CI)	0.34 [0.05, 2.33]
6.22.2 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	5.10 [0.25, 104.92]
6.23 Major bleeding	3	288	Risk Ratio (M‐H, Random, 95% CI)	1.22 [0.18, 8.34]
6.23.1 CKD stages 3‐5	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.27 [0.03, 2.40]
6.23.2 CKD stage 5D	2	248	Risk Ratio (M‐H, Random, 95% CI)	3.96 [0.45, 35.12]
6.24 Infection	5	728	Risk Ratio (M‐H, Random, 95% CI)	1.32 [0.70, 2.48]
6.24.1 CKD stages 3‐5	3	436	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.48, 3.40]
6.24.2 CKD stage 5D	2	292	Risk Ratio (M‐H, Random, 95% CI)	1.35 [0.59, 3.09]

6.18. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 18: Change in albuminuria

6.19. Analysis.

Comparison 6: Antioxidants versus placebo or standard therapy: low risk of bias, Outcome 19: Vascular access thrombosis

Comparison 7. Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Cardiovascular death	6	3633	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.57, 1.36]
7.1.1 CKD stages 3‐5	3	3269	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.46, 1.74]
7.1.2 CKD stages 5D	3	364	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.42, 1.46]
7.2 Death (any cause)	23	5866	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.83, 1.14]
7.2.1 CKD stages 3‐5	8	3860	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.82, 1.20]
7.2.2 CKD stage 5D	7	1151	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.74, 1.39]
7.2.3 Kidney transplant	8	855	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.51, 1.28]
7.3 Cardiovascular disease	12	4446	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.58, 0.98]
7.3.1 CKD stages 3‐5	7	3997	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.57, 1.00]
7.3.2 CKD stage 5D	4	388	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.33, 2.03]
7.3.3 Kidney transplant	1	61	Risk Ratio (M‐H, Random, 95% CI)	Not estimable
7.4 Coronary heart disease	9	4073	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.57, 1.13]
7.4.1 CKD stages 3‐5	5	3667	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.75, 1.24]
7.4.2 CKD stage 5D	4	406	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.25, 1.22]
7.5 Heart failure	5	3613	Risk Ratio (M‐H, Random, 95% CI)	1.40 [1.11, 1.76]
7.5.1 CKD stages 3‐5	5	3613	Risk Ratio (M‐H, Random, 95% CI)	1.40 [1.11, 1.76]
7.6 Cerebrovascular disease	6	1588	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.69, 1.29]
7.6.1 CKD stages 3‐5	3	1147	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.71, 1.38]
7.6.2 CKD stage 5D	3	441	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.25, 2.02]
7.7 Peripheral vascular disease	4	1492	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.41, 1.47]
7.7.1 CKD stages 3‐5	2	1162	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.34, 2.90]
7.7.2 CKD stage 5D	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.26, 1.12]
7.8 Kidney failure	9	3117	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.53, 1.09]
7.8.1 CKD stages 3‐5	8	2940	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.55, 1.25]
7.8.2 Kidney transplant	1	177	Risk Ratio (M‐H, Random, 95% CI)	0.53 [0.26, 1.11]
7.9 eGFR at end of study	20	1751	Mean Difference (IV, Random, 95% CI)	3.48 [1.09, 5.87]
7.9.1 CKD stages 3‐5	15	1239	Mean Difference (IV, Random, 95% CI)	3.42 [0.83, 6.01]
7.9.2 Kidney transplant	5	512	Mean Difference (IV, Random, 95% CI)	2.16 [‐10.20, 14.52]
7.10 Change in eGFR	18	3496	Mean Difference (IV, Random, 95% CI)	3.71 [2.77, 4.65]
7.10.1 CKD stages 3‐5	16	3425	Mean Difference (IV, Random, 95% CI)	3.82 [2.85, 4.78]
7.10.2 Kidney transplant	2	71	Mean Difference (IV, Random, 95% CI)	3.97 [‐4.18, 12.13]
7.11 Serum creatinine at end of study	11	1027	Mean Difference (IV, Random, 95% CI)	‐6.11 [‐21.23, 9.01]
7.11.1 CKD stages 3‐5	4	326	Mean Difference (IV, Random, 95% CI)	‐6.27 [‐32.00, 19.46]
7.11.2 Kidney transplant	7	701	Mean Difference (IV, Random, 95% CI)	‐5.86 [‐24.78, 13.07]
7.12 Change in serum creatinine	9	2847	Mean Difference (IV, Random, 95% CI)	‐16.21 [‐26.43, ‐6.00]
7.12.1 CKD stages 3‐5	6	2734	Mean Difference (IV, Random, 95% CI)	‐27.38 [‐37.80, ‐16.96]
7.12.2 Kidney transplant	3	113	Mean Difference (IV, Random, 95% CI)	7.86 [‐2.78, 18.50]
7.13 Urinary albumin/creatinine ratio at end of study	3	1119	Mean Difference (IV, Random, 95% CI)	‐2.63 [‐29.36, 24.11]
7.13.1 CKD stages 3‐5	3	1119	Mean Difference (IV, Random, 95% CI)	‐2.63 [‐29.36, 24.11]
7.14 Change in urinary albumin/creatinine ratio	3	1086	Mean Difference (IV, Random, 95% CI)	0.35 [‐16.76, 17.45]
7.14.1 CKD stages 3‐5	3	1086	Mean Difference (IV, Random, 95% CI)	0.35 [‐16.76, 17.45]
7.15 Proteinuria at end of study	5	233	Mean Difference (IV, Random, 95% CI)	‐0.95 [‐1.88, ‐0.02]
7.15.1 CKD stages 3‐5	5	233	Mean Difference (IV, Random, 95% CI)	‐0.95 [‐1.88, ‐0.02]
7.16 Change in proteinuria	4	209	Mean Difference (IV, Random, 95% CI)	‐1.65 [‐3.26, ‐0.03]
7.16.1 CKD stages 3‐5	4	209	Mean Difference (IV, Random, 95% CI)	‐1.65 [‐3.26, ‐0.03]
7.17 Albuminuria at end of study	3	353	Mean Difference (IV, Random, 95% CI)	‐2.78 [‐151.55, 145.99]
7.17.1 CKD stages 3‐5	3	353	Mean Difference (IV, Random, 95% CI)	‐2.78 [‐151.55, 145.99]
7.18 Change in albuminuria	3	353	Mean Difference (IV, Random, 95% CI)	‐88.37 [‐319.83, 143.09]
7.18.1 CKD stages 3‐5	3	353	Mean Difference (IV, Random, 95% CI)	‐88.37 [‐319.83, 143.09]
7.19 Vascular access thrombosis	1	34	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.28, 1.81]
7.19.1 CKD stage 5D	1	34	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.28, 1.81]
7.20 Graft loss	7	794	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.66, 1.16]
7.20.1 Kidney transplant	7	794	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.66, 1.16]
7.21 Graft failure	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
7.21.1 Kidney transplant	1	61	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.06, 12.95]
7.22 Delayed graft function	5	571	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.69, 1.07]
7.22.1 Kidney transplant	5	571	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.69, 1.07]
7.23 Malignancy	2	423	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.07, 15.39]
7.23.1 CKD stages 3‐5	1	227	Risk Ratio (M‐H, Random, 95% CI)	0.34 [0.05, 2.33]
7.23.2 CKD stage 5D	1	196	Risk Ratio (M‐H, Random, 95% CI)	5.10 [0.25, 104.92]
7.24 Major bleeding	5	397	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.35, 2.95]
7.24.1 CKD stages 3‐5	2	96	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.07, 5.97]
7.24.2 CKD stages 5D	3	301	Risk Ratio (M‐H, Random, 95% CI)	1.38 [0.36, 5.24]
7.25 Infection	9	3268	Risk Ratio (M‐H, Random, 95% CI)	1.45 [1.28, 1.65]
7.25.1 CKD stages 3‐5	5	2682	Risk Ratio (M‐H, Random, 95% CI)	1.46 [1.28, 1.67]
7.25.2 CKD stage 5D	3	449	Risk Ratio (M‐H, Random, 95% CI)	1.38 [0.69, 2.74]
7.25.3 Kidney transplant	1	137	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.45, 2.70]

7.5. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 5: Heart failure

7.18. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 18: Change in albuminuria

7.19. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 19: Vascular access thrombosis

7.21. Analysis.

Comparison 7: Antioxidants versus placebo or standard therapy: follow‐up for 6 months or more, Outcome 21: Graft failure

Characteristics of studies

Characteristics of included studies [ordered by study ID]

ACTIVE 2014.

Study characteristics
Methods	Study design Parallel RCT Time frame Time period: November 2005 to March 2008 Follow‐up: 3 years
Participants	Study characteristics Country: Japan Setting: multicentre (number of sites not reported) Inclusion criteria: HD > 1 year Exclusion criteria: > 75 years; history of chronic inflammation, malignancy, haematological disorders, or severe liver dysfunction; use of anti‐inflammatory drugs and immunosuppressive agents Baseline characteristics Number (randomised/analysed): intervention group (176/55); control group (96/67) Mean age ± SD (years): intervention group (60 ± 12 ); control group (53 ± 18) Sex (male, %): treatment group (58/176, 33%); control group (28/96, 29%) CKD stage/kidney function: 5D
Interventions	Intervention group EPO, iron, and vitamin C adjusted to ferritin and Hb levels according to protocol Control group Doses of EPO and iron determined by physician
Outcomes	Outcomes relevant to this review Death (any cause) CVD Cerebrovascular disease Infection
Notes	Additional information Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Unclear if allocation was concealed
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but lack of blinding unlikely to cause detection bias due to all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow‐up (21%)
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	High risk	50% of patients from the treatment group deviated from protocol

Ahmadi 2017.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: August to December 2014 Follow‐up: 3 months
Participants	Study characteristics Country: Iran Setting: multicentre (3 sites) Inclusion criteria: HD > 3 months and residual diuresis > 100 mL/day Exclusion criteria: AKI as the cause of dialysis Baseline characteristics Number (randomised/analysed): intervention group (27/26); control group (27/21) Mean age ± SD: 58 ± 13 years Sex, male (%): 30/54 (63%) CKD stage/kidney function: 5D
Interventions	Intervention group N‐acetylcysteine: 1200 mg twice/day Control group Usual care
Outcomes	Outcomes relevant to this review Death (any cause) eGFR at end of study Change in eGFR (SD for change imputed)
Notes	Additional information Funding sources: Tehran University
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation with simple randomisation software in Microsoft Excel, numbers 1‐55 generated in unsorted and unrepeated fashion. Number < 27 assigned to treatment group and rest to control group
Allocation concealment (selection bias)	Unclear risk	Unclear if allocation was concealed
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but lack of blinding unlikely to cause detection bias due to all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Numbers and reasons for loss to follow‐up reported
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	Low risk	None

AIONID 2013.

Study characteristics
Methods	Study design Parallel 2 x 2 factorial RCT Time frame Follow‐up 16 weeks
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: HD > 3 months, average monthly Kt/V > 1.2, serum albumin > 40 g/L; functioning AV graft (fistula or tunnelled) or catheter that will not switch within 6 months Exclusion criteria: PD or HD < 6 months; life expectancy < 6 months; use of intradialytic parenteral nutrition in the past 2‐3 months, acute wasting or active systemic disease; use of pulse chemotherapy Baseline characteristics Number (randomised/analysed): (93/84); intervention group 1 (19); intervention group 2 (22); intervention group 3 (22) control group (21) Mean age ± SD (years): intervention group 1 (59 ± 14); intervention group 2 (58 ± 14); intervention group 3 (55 ± 10) control group 4 (64 ± 9) Sex (male, %): intervention group 1 (9/19, 47%); intervention group 2 (10/22, 45%); intervention group 3 (12/22, 55%); control group 4 (8/21, 38%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Pentoxifylline: 300 mg 3 times/week Oral nutritional supplements: 2 cans Intervention group 2 Pentoxifylline:400 mg 3 times/week Oral nutritional supplements placebo Intervention group 3 Pentoxifylline placebo Oral nutritional supplements: 2 cans Control group Pentoxifylline placebo Oral nutritional supplements placebo
Outcomes	Outcomes relevant to this review Death (any cause) SCr at end of study Change in SCr (SD for change imputed) Cancer
Notes	Additional information Intervention group comprises patients receiving pentoxifylline with oral nutritional supplements or oral nutritional supplement placebo pooled and control group patients receiving pentoxifylline placebo and oral nutritional supplements or oral nutritional supplement placebo Funding sources: DaVita Clinical Research (clinical data and resources for conducting the research project and DaVita dietitians for the help in this study), Abbott Nutrition (provide nutritional products and placebos), National Institute of Diabetes, Digestive and Kidney Disease of the National Institutes of Health (grant no. R21‐DK078012‐01), and a philanthropic grant from Mr. Harold Simmons
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation with permutated block randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Blinding with placebo in unidentifiable identical containers
Blinding of outcome assessors (detection bias)	Low risk	Blinding with placebo in unidentifiable identical containers and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Low loss to follow‐up but reasons not specified per arm
Selective reporting (reporting bias)	High risk	Dialysis access problems mentioned as outcome in methods but not reported
Other bias	Low risk	Study appears free of other biases

Antunes 2014.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 16 weeks
Participants	Study characteristics Country: Brazil Setting: multicentre (number of sites not reported) Inclusion criteria: CKD patients with inflammation undergoing dialysis Exclusion criteria: < 18 years; dialysis vintage < 3 months, temporary vascular access, active bleeding or blood transfusion in past 3 months, HIV, haematologic, neoplastic, and hepatic disease; inadequate iron stores (TSAT < 20% and/or ferritin < 100U/L), or PTH > 0.11 pmol/L Baseline characteristics Number: intervention group (35); control group (36) Median age, IQR (years): intervention group (52, 46‐61); control group (60, 39‐68) Sex (male, %): intervention group (29/35, 83%); control group (17/36, 47%) CKD stage/kidney function: 5D
Interventions	Intervention group Pentoxifylline: 400 mg 3 times/week Control group Usual care
Outcomes	Outcomes relevant to this review Infection
Notes	Additional information Funding sources: Conselho Nacional de Desenvolvimento Científico e TecnológicoCNPq
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified, but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Low loss to follow‐up but reasons not specified per arm
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Argani 2014.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up 60 days
Participants	Study characteristics Country: Iran Setting: single‐centre Inclusion criteria: chronic HD; BMI ≤ 25; no GI disorders Exclusion criteria: transplant candidate; smoker; use of penicillamine, glucocorticoids, oestrogens, antibiotics or oral contraceptives; pregnant and lactating women Baseline characteristics Number (randomised/analysed): (66/60); intervention group (30); control group (30) Mean age ± SD (years): intervention group (56 ± 4); control group (56 ± 8) Sex (male, %): intervention group (19/30, 63%); control group (17/30, 57%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 440 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Drug Applied Research Center at Tabriz University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method and no protocol available
Other bias	Low risk	STudy apppears free of other biases

Asemi 2016.

Study characteristics
Methods	Study design Parallel, 2 x 2 factorial RCT Time frame Recruitment: August to November 2014 Follow‐up: 4 months
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 18 to 80 years; HD > 1 year Exclusion criteria: hospitalisation < 1 month; active infection; history of steatorrhoea; malignancy; thrombocytopenia; abnormal coagulation profile; psychosomatic disorders; need for anticoagulation therapy Baseline characteristics Number: intervention group 1 (30); intervention group 2 (30); control group 1 (30); control group 2 (30) Mean age ± SD (years): intervention group 1 (61.2 ± 16.6 years); intervention group 2 (54.9 ± 14.3); control group 1 (59.9 ± 15.7); control group 2 (55.2 ± 17.0) Sex (male, %): intervention group 1 (20/30, 67%); intervention group 2 (20/30, 67%); control group 1(20/30, 67%); control group 2 (20/30, 67%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Vitamin E: 400 IU/day Omega‐3 fatty acid placebo Intervention group 2 Vitamin E: 400 IU/day Omega‐3 fatty acid: 1250 mg/day Control group 1 Omega‐3 fatty acid placebo Vitamin E placebo Control group 2 Omega‐3 fatty acid: 1250 mg/day Vitamin E placebo The combined intervention group received vitamin E with omega‐3 fatty acids or omega‐3 fatty acid placebo The combined control group received vitamin E placebo and omega‐3 fatty acids or omega‐3 fatty acid placebo
Outcomes	Outcomes of interest to this review Death (any cause) SCr at end of study Change in SCr
Notes	Additional information Funding sources: Vice‐chancellor for Research, KUMS, and Iran and Kashan University of Medical Sciences (grant no. 93145)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by one of the investigators who had no clinical involvement in the study using random allocation software
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	Unclear risk	None

Atapour 2022.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 2019 Follow‐up: 3 months
Participants	Study characteristics Country: Iran Setting: multicentre (number of sites not reported) Inclusion criteria: ≥ 18 years; chronic HD; selenium deficiency Exclusion criteria: hepatitis B,; hepatitis C; skeletal abnormalities that could affect sit‐to‐stand test; normal selenium levels; use of corticosteroids or NSAIDs in the past 2 months Baseline characteristics Number (randomised/analysed): intervention group (44/40); control group (44/38) Mean age ± SD (years): intervention group (57 ± 12); control group (57 ± 14) Sex (male, %): intervention group (28/40, 70%); control group (20/38, 53%) CKD stage/kidney function: 5D
Interventions	Intervention group Selenium: 400 µg, 3 times/week Control group Placebo
Outcomes	Outcomes of interest to this review Death (any cause)
Notes	Additional information Funding sources: Isfahan University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with random allocation software
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

ATIC 2005.

Study characteristics
Methods	Study design Parallel RCT Stratification for prior use of ACEi or ARBs, CrCl (15 to 39 and 40 to 70 mL/min/1.73 m2), and age (20 to 49 and 50 to 80 years) Time frame Recruitment: May 2001 to December 2002 Follow‐up: 24 months
Participants	Study characteristics Country: The Netherlands Setting: multicentre (7 sites) Inclusion criteria: CrCl 15 to 70 mL/min/1.73m2 Exclusion criteria: diabetes mellitus, active vasculitis, nephrotic syndrome; kidney transplant; coronary artery disease; peripheral artery disease; cerebrovascular disease; fasting cholesterol > 7mmol/L or cholesterol‐lowering therapy in past 3 months Baseline characteristics Number: intervention group (47); control group (46) Mean age ± SD (years): intervention group (54 ± 11) control group (52 ± 13) Sex (male,%): intervention group (24/47, 51%), control group (29/46, 63%) CKD stage/kidney function: CKD 2‐4 Intervention group (60 ± 24 mL/min/1.73m2); control group (55 ± 23 mL/min/1.73m2)
Interventions	Intervention group Pravastatin: 40 mg/day Vitamin E: 300 mg/day added at 6 months Homocysteine‐lowering therapy: added at 12 months Control group Placebo
Outcomes	Outcomes relevant to this review CVD Coronary heart disease Cerebrovascular disease Kidney failure eGFR at end of study Change in eGFR (SD for change imputed) Albuminuria at end of study Change in albuminuria (SD for change imputed)
Notes	Additional information Funding sources: The Dutch Kidney Foundation (project number C97‐1707) and Bristol‐Myers Squibb
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation by computer‐generated sequence
Allocation concealment (selection bias)	Low risk	Concealed envelopes kept by hospital pharmacist
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	Outcomes reported in methods and protocol on Clinicaltrials.gov
Other bias	Low risk	Study appears free of other biases

Attallah 2006.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: HD > 6 months; ESA > 450 U/kg/week for > 6 months; IV iron administration; Hb < 6.8 mmol/L in past 3 months, ferritin > 500 µg/L; TSAT < 50% Exclusion criteria: bone marrow malignancy or possible underlying malignancy; myelodysplastic syndrome; chronic infection; haemochromatosis; haemoglobinopathies; significant bleeding in past 3 months; MCV > 100 fL; CRP > 20 mg/dL; iPTH > 53 pmol/L; aluminium > 20 µg/L Baseline characteristics Number: intervention group (20); control group (22) Mean age ± SD (years): intervention group (51 ± 5); control group (49 ± 6) Sex (male, %): intervention group (9/20, 45%); control group (10/22, 45%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin C: 300 mg 3 times/week Control group Usual care
Outcomes	Outcomes relevant to this review CVD Coronary heart disease Major bleeding Infection
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation described as "concealed randomisation"
Allocation concealment (selection bias)	High risk	Explicitly described that allocation was not concealed "Ideally, the study should have been conducted in a blinded manner (nurses and staff), with a placebo solution administered to the control group. However, the cost of having an unblinded pharmacist formulate and supply the experimental and placebo solutions was prohibitive"
Blinding of participants and personnel (performance bias)	High risk	No blinding
Blinding of outcome assessors (detection bias)	Low risk	No blinding but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Bansal 2017.

Study characteristics
Methods	Study design Parallel 5‐arm RCT Time frame Follow‐up: 3 months
Participants	Study characteristics Country: USA Setting: not reported Inclusion criteria: diabetic kidney disease; eGFR 15 to 60 mL/min/1.73m2, UACR > 16.5 mg/mmol Exclusion criteria: not reported Number: intervention group 1 (15); intervention group 2(16); intervention group 3 (12); intervention group 4 (16); control group (16) Mean age ± SD (all patients): 63 ± 7 years Sex (male. %; all patients): 67/75, 89% CKD stage/kidney function (all patients): 36 ± 13 mL/min/1.73m2
Interventions	Intervention group 1 N‐acetylcysteine: 600 mg twice/day Silybin: 960 mg/day Intervention group 2 N‐acetylcysteine: 600 mg twice/day Silybin: 480 mg/day Intervention group 3 N‐acetylcysteine: 600 mg twice/day Silybin placebo Intervention group 4 N‐acetylcysteine placebo Silybin: 480 mg/day Control group Double placebo
Outcomes	Outcomes relevant to this review Change in eGFR UACR at end of study Change in UACR
Notes	Additional information Funding sources: NIH support; Veterans Affairs Support Abstract‐only publication
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Double blind but method of blinding not specified
Blinding of outcome assessors (detection bias)	Low risk	Double blind but method of blinding not specified. Only objective outcomes included so blinding unlikely to affect outcomes
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	High risk	Conference abstracts only

BEACON 2013.

Study characteristics
Methods	Study design Parallel RCT Stratified by enrollment site Time frame Recruitment: June 2011 to September 2012 Follow‐up: 9 months
Participants	Study characteristics Country: international Setting: multicentre (319 sites) Inclusion criteria: eGFR 15 to 30 mL/min/1.73m2: type 2 DM Exclusion criteria: type 1 DM; non‐diabetic kidney disease; history of transplant or planned transplant from living donor; AKI; acute dialysis in past 12 weeks; UACR > 3500 mg/g Baseline characteristics Number: intervention group (1088); control group (1097) Mean age ± SD (years): intervention group (68.9 ± 9.7); control group (68.2 ± 9.4) Sex (male, %): intervention group (626/1088, 57%); control group (625/1097, 58%) CKD stage/kidney function: CKD 4 Interventiongroup (22 ± 4 mL/min/1.73m2); control group (23 ± 5 mL/min/1.73m2)
Interventions	Intervention group Bardoloxone methyl: 20 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Cardiovascular death Death (any cause) CVD: defined as a composite of nonfatal MI, nonfatal stroke, hospitalisation for heart failure, or death from cardiovascular causes Coronary heart disease Heart failure: defined as hospitalisation for heart failure or death due to heart failure Kidney failure: defined as the need for maintenance dialysis for 12 weeks or more or kidney transplantation Change in eGFR Change in SCr Infection
Notes	Additional information Study prematurely terminated due to increased incidence of heart failure in the bardoxolone methyl group Funding sources: Reata Pharmaceuticals
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	All relevant outcomes protocol reported
Other bias	High risk	Funding: sponsor (Reata pharmaceuticals) involved in design, analysis, interpretation of data, and writing of manuscript

BEAM 2011.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame Follow‐up: 52 weeks
Participants	Study characteristics Country: USA Setting: multicentre (43 sites) Inclusion criteria: adults; moderate‐to‐severe CKD (defined as eGFR of 20 to 45 mL/min/1.73 m²) and type 2 DM Exclusion criteria: type 1 DM; non‐diabetic kidney disease Baseline characteristics Number: intervention group 1 (57); intervention group 2 (57); intervention group 3 (56); control group (57) Mean age ± SD (years): intervention group 1 (67 ± 9); intervention group 2 (66 ± 9); intervention group 3 (67 ± 9); control group (68 ± 10) Sex (male, %): intervention group 1 (34, 60%); intervention group 2 (33, 58%); intervention group 3 (33, 59%); control group (28, 49%) CKD stage: CKD 3‐4 Intervention group 1 (33 ± 7 ml/min/1.73m2); intervention group 2 (33 ± 7 ml/min/1.73m2); intervention group 3 (67 ± 9 ml/min/1.73m2); control group (32 ± 8 ml/min/1.73m2)
Interventions	Intervention group 1 Bardoxolone methyl: 25 mg Intervention group 2 Bardoxolone methyl: 75 mg Intervention group 3 Bardoxolone methyl: 150 mg Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) CVD: defined as MI, stroke, TSA, coronary or peripheral revascularisation procedure, or hospitalisation for unstable angina or congestive heart failure Coronary heart disease Heart failure: defined as hospitalisation for heart failure Kidney failure: chronic dialysis or kidney transplant eGFR at end of study Change in eGFR SCr at end of study Change in SCr Cancer Infection
Notes	Additional information Funding sources: Reata Pharmaceuticals
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with stratified randomisation
Allocation concealment (selection bias)	Low risk	Central randomisation; procedure described
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Loss to follow up reported and reasons specified
Selective reporting (reporting bias)	Low risk	All outcomes in method and protocol reported
Other bias	Low risk	Funding: This study was sponsored by Reata Pharmaceuticals and was designed by the first author and representatives of the sponsor. Study investigators and coordinators jointly managed the study with the sponsor

Bejaoui 2022.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Tunisia Setting: multicentre (2 sites) Inclusion criteria: diabetic kidney disease and hypertension; use of insulin, captopril, furosemide, folic acid, and iron Exclusion criteria: HD; smoking Baseline characteristics Number: intervention group (30); control group (10) Mean age ± SEM (years): intervention group (57.9 ± 2.1); control group (58.4 ± 1.72) Sex (male, %): intervention group (16/30, 53%); control group (6/10, 60%) CKD stage/kidney function: CKD 3‐4 Intervention group (38 ± 11 mL/min/1.73m2); control group (39 ± 15 mL/min/1.73m2)
Interventions	Intervention group Grape seed flour: 1g/kg /day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed)
Notes	Additional information Funding sources: no funding sources were involved in the study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported
Other bias	Low risk	Study appears free of other biases

Biniaz 2014.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame Recruitment: October 2012 to January 2013 Follow‐up: 8 weeks
Participants	Study characteristics Country: Iran Setting: multi‐centre Inclusion criteria: undergoing HD > 6 months Exclusion criteria: transplant in the past year; active infection; cancer; smoking or passive smoke exposure; alcohol consumption; use of anti‐inflammatory medications Baseline characteristics Number (randomised/analysed): intervention group (59/55); control group 1 (58/55); control group 2 (55/55) Mean age ± SD (years): intervention group (60 ± 12); control group 1 (63 ± 11); control group 2 (62 ± 15) Sex (male, %): intervention group (33/59, 56%); control group 1 (34/58, 59%); control group 2 (35/55, 64%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin C: 250 mg 3 times/week Control group 1 Placebo Control group 2 Usual care:
Outcomes	Outcomes relevant to this review Death (any cause) SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Nephrology and Urology Research Center of Baqiyatallah University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly distributed by a lottery method (simple random sampling)
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up, reasons briefly described
Selective reporting (reporting bias)	Unclear risk	All outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Blackhall 2005.

Study characteristics
Methods	Study design Cross‐over RCT Time frame Follow‐up: 12 months (6 months intervention, 6 months placebo)
Participants	Study characteristics Country: Australia Setting: single centre Inclusion criteria: kidney transplant Exclusion criteria: already taking an antioxidant Baseline characteristics Number: 10 Mean age ± SD: 54 ± 13 years Sex (male/ %): 6/10, 60% CKD stage/kidney function: transplant Intervention group (67 ± 21 mL/min/1.73m2); control group (67 ± 12 mL/min/1.73m2)
Interventions	Intervention group Vitamin E: 400 IU/day Vitamin C: 500 mg/day Betacarotene: 6 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Change in eGFR SCr at end of study Change in SCr
Notes	Additional information Funding sources: Clifford Craig Medical Research Trust
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	Single blind, method of blinding not specified
Blinding of outcome assessors (detection bias)	Low risk	No blinding (single blind) but all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Borges 2016.

Study characteristics
Methods	Study design Parallel RCT Time frame November 2013 to December 2014 Follow‐up: 12 weeks
Participants	Study characteristics Country: Brazil Setting: single centre Inclusion criteria: > 18 years; diabetes; persistent macroalbuminuria (UACR > 3.4 mg/mmol); HbA1C < 86 mmol/mol; use of antiglycaemic drugs Exclusion criteria: auto‐immune disease; HIV; viral hepatitis; neoplasia; chronic UTI; chronic heart failure (NYHA class III or IV); recent history of coronary artery or cerebrovascular disease; eGFR < 30 mL/min/1.73 m2 Baseline characteristics Number*: intervention group (9); control group (4) Mean age ± SD (years): intervention group (61 ± 2); control group (65 ± 8) Sex (male, %): intervention group (5/9, 55%); control group (2/4, 50%) CKD stage/kidney function: intervention group (51 ± 2 mL/min/1.73 m2); control group (53 ± 3 mL/min/1.73 m2) *Only patients with eGFR < 60 mL/min/1.73 m2
Interventions	Intervention group Green tea polyphenols: 800 mg/day Control group Placebo
Outcomes	Outcomes of relevance to this review eGFR at end of study Change in eGFR (SD for change imputed) UACR at end of study Change in UACR (SD for change imputed)
Notes	Additional information Funding sources: Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, Grant no. 2008/57560‐0 and 2014/22687‐0) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Grant no. 304026/2013‐1)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a randomisation list generated with http://www.randomization.com by personnel at different location not involved in study
Allocation concealment (selection bias)	Low risk	Allocation was concealed with prepackaged bottles numbered according to randomisation sequence
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	None

CoQ10 Biomarker 2016.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame Recruitment: September 2011 to April 2013 Follow‐up: 4 months
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: 18 to 85 years; HD >3 months and life expectancy > 1 year Exclusion criteria: transplant in past 6 months; major cardiovascular event in past 6 months; current dialysis access through catheter Baseline characteristics Number (randomised/analysed): intervention group 1 (26/21); intervention group 2 (26/18) control group (28/26) Mean age ± SD (years): intervention group 1 (56 ± 12); intervention group 2 (49 ± 16); control group (57± 13) Sex (male, %): intervention group 1 (13/21, 62%); intervention group 2 (12/18; 67%); control group (19/26, 73%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Coenzyme Q10: 1200 mg/day Intervention group 2 Coenzyme Q10: 600 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: the National Institutes of Health, including National Center for Complementary and Alternative Medicine (grant no. R21 AT004265 and R21 AT3844), the National Institute of Diabetes and Digestive and Kidney Diseases (grant no. T32 DK007467 and K24 DK62849), National Center for Advancing Translational Sciences (grant no KL2 TR000421 and UL1 TR000445, National Institute of Environmental Health Sciences (grant no. P30 ES000267), and National Heart, Lung and Blood Institute (grant no. R01 HL070938), and an unrestricted gift from the Northwest Kidney Centers to the Kidney Research Institute
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly distributed by permutated block randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method nor in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Danilovic 2011.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: April 2005 to June 2008 Follow‐up: 12 months
Participants	Study characteristics Country: Brazil Setting: single centre Inclusion criteria: >18 years; transplant recipient from heart‐beating donor Exclusion criteria: not reported Baseline characteristics Number: intervention group (38); control group (36) Mean age ± SD (years): intervention group (50.7 ± 11.8); control group (49 ± 11) Sex (male, %): intervention group (20/38, 53%); control group (13/36, 36%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group N‐acetylcysteine: 600 mg twice/day for 7 days postoperatively Control group Usual care
Outcomes	Outcomes relevant to this review eGFR at end of study SCr at end of study Graft loss DGF
Notes	Additional information Funding sources: FAPESP (grant no. 06/52046‐0)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Investigators were blinded (but method of blinding not specified). Unclear if patients were blinded
Blinding of outcome assessors (detection bias)	Low risk	Outcome assessors were blinded (but method of blinding not specified) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Demir 2006a.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 3 months
Participants	Study characteristics Country: Turkey Setting: single centre Inclusion criteria: received transplant > 6 months ago Exclusion criteria: uncontrolled DM; hypertension; MI; infection; surgical procedure in the past 3 months; use anticoagulants Baseline characteristics Number: intervention group (22); control group (20) Mean age ± SD (years): intervention group (33 ± 12); control group (35 ± 12) Sex (male, %): intervention group (14/22, 64%); control group (14/20, 70%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 1200 mg/day Control group Usual care
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Patients were randomised with alternate allocation to successive patients
Allocation concealment (selection bias)	High risk	No allocation concealment (alternate allocation to successive patients)
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Assessors were blinded (method of blinding nog specified) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

DeVault 1994.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: to 05/01/1993
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics Number: intervention group (11); control group (12) Mean age ± SD (years): intervention group (44 ± 5); control group (38 ± 4) Sex (male, %): treatment group (7/11, 64%); control group (7/12, 58%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 800 mg before transplant and 800 mg 3 times/day for 3 days Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed) Graft loss
Notes	Additional information Funding sources: National Center of Research Resources Public Health Services research grant (grant no. MO1‐RR00064) and educational grants from Ortho Biotech, Inc., Raritan, NJ, and Hoechst‐Roussel Pharmaceuticals, Inc., Sommervile, NJ
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo with identical matching capsules)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Firuzi 2016.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 2 months
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 18 to 60 years; PD Exclusion criteria: bacterial peritonitis in the past 2 months; active hepatitis; smoking; use of oral contraception and levothyroxine Baseline characteristics Number: intervention group (28); control group (22) Mean age ± SD (years): intervention group (44.8 ± 11.6); control group (50.1± 11.4) Sex (male, %): intervention group (13/28, 46%); control group (13/22, 59%) CKD stage/kidney function: 5D
Interventions	Intervention group Silymarin: 140 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Vice‐Chancellor for Research, Shiraz University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified (participants received placebo but unclear if they were aware which arm they were allotted to)
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	High loss to follow‐up (17%), reasons globally specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Friedman 2003.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 4 weeks
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: HD Exclusion criteria: acute illness; liver cirrhosis Baseline characteristics Number (randomised/analysed): intervention group (18/18); control group (17/16) Mean age ± SD (years): intervention group (68 ± 3); control group (70 ± 4) Sex (male, %): intervention group (10/18, 56%); control group (8/17, 47%) CKD stage/kidney function: 5D
Interventions	Intervention group N‐acetylcysteine: 1200 mg twice/day Control group: Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Tom Johnson and Fleming & Co (donation of Nephrocap supplements)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo; Treatment assignments were made blinded to other aspects of the study)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo; laboratory analyses, data entry, and data analyses were performed by code so that treatment assignments remained concealed) and all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Gholipour Baradari 2011.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 22 May 2010 to 10 August 2010 Follow‐up: 8 weeks
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 20 to 70 years; HD > 6 months; serum phosphorus > 1.78 mmol/L Exclusion criteria: history of malignancy, malnutrition, and severe cardiac, respiratory, or liver disease; use of immunosuppressive drugs in past 2 months Baseline characteristics Number (randomised/analysed): intervention group (30/29); control group (30/29) Mean age ± SD (years): intervention group (59.7 ± 14.6); control group (60.6 ± 13.3) Sex (male, %): intervention group (17/29, 59%); control group (14/29, 48%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin C: 500 mg 3 times/week Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Addtional information Funding sources: Research Deputy of Mazandaran University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a random number table
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Double blind but method of blinding not specified
Blinding of outcome assessors (detection bias)	Low risk	Double blind (method of blinding not specified) but only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Giliberti 2022.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Italy Setting: multicentre (3 sites) Inclusion criteria: > 18 years; CKD stage 3; selenium deficiency Exclusion criteria: myelodysplasia; leukaemia; oncological disease; CKD 4‐5; dialysis; use of ESAs Baseline characteristics Number: intervention group 1 (22); intervention group 2 (20); intervention group 3 (20) Mean age ± SD (years): intervention group 1 (72.3 ± 7.5); intervention group 2 (68.3 ± 2.8); intervention group 3 (68.05 ± 1.7) Sex (male, %): intervention group 1 (11/22, 50%); intervention group 2 (10/20, 50%); intervention group 3 (11/20, 55%) CKD stage/kidney function: CKD stages 3‐4 Intervention group 1 (36 ± 8.5 mL/min/1.73 m2); intervention group 2 (46 ± 2.2 mL/min/1.73 m2); intervention group 3 (47.8 ± 3.5 mL/min/1.73 m2)
Interventions	Intervention group 1 Ferric sodium EDTA: 30 mg/day in combination with vitamin C, folic acid, copper gluconate, zinc gluconate and selenomethionine Intervention group 2 Ferrous sulfate prolonged release: 105 mg/day Intervention group 3 Ferric liposomal formulation: 30 mg/day Intervention groups 2 and 3 were combined for the analyses
Outcomes	Outcomes relevant to this review eGFR at end of study change in eGFR (SD for change imputed)
Notes	Additional information Funding sources: Aqma Italia S.p.A (Milan, Italy)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with the Excel randomisation function
Allocation concealment (selection bias)	Low risk	Allocation was not concealed (open label)
Blinding of participants and personnel (performance bias)	High risk	No blinding (open label)
Blinding of outcome assessors (detection bias)	Low risk	No blinding (open label) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	High risk	Randomisation yielded unbalanced groups, particularly for eGFR

Goicoechea 2012.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: January to May 2007 Median follow‐up: 2.25 years
Participants	Study characteristics Country: Spain Setting: single centre Inclusion criteria: eGFR < 60 mL/min/1.73 m2 Exclusion criteria: cardiovascular event or hospitalisation in the past 3 months; active infection/inflammatory disease or HIV; chronic liver disease; use of immunosuppression Baseline characteristics Number (randomised/analysed): intervention group (46/34); control group (45/36) Mean age ± SD (years): intervention group (70 ± 14); control group (70 ± 8) Sex (male, %): not reported CKD stage/kidney function: CKD 3‐5 Intervention group (42 ± 10 mL/min/1.73 m2); control group (41 ± 12 mL/min/1.73 m2)
Interventions	Intervention group Pentoxifylline: 400 mg twice/day Control group Usual care
Outcomes	Outcomes relevant to this review Cardiovascular death Death (any cause) Kidney failure eGFR at end of study Change in eGFR (SD for change imputed) Albuminuria at end of study Change in albuminuria (SD for change imputed)
Notes	Additional information Funding sources: ISCIII RETIC REDINREN (grant no. RD016/009) and FEDER funds
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated list
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	Single blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Outcome assessors were not blinded but only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	High loss to follow mostly due to mortality. Other causes for loss to follow‐up specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Gonzalez‐Espinoza 2012.

Study characteristics
Methods	Study design Parallel RCT Time frame January 2006 to May 2008 Follow‐up: 4 months
Participants	Study characteristics Country: Mexico Setting: single centre Inclusion criteria: HD > 2 months with AV fistula as vascular access Exclusion criteria: inflammatory cause of kidney failure; liver disease; cancer; AIDS; infectious disease in the past 2 months before study; failed transplant; haemorrhage/clotting disorders; pre‐existing cardiac arrhythmias, or arterial hypotension; treatment with antibiotics, NSAIDs, steroids, immunosuppressives or statins Baseline characteristics Number: intervention group (18); control group (18) Mean age ± SD (years): intervention group (31 ± 13); control group (39 ± 15) Sex (male, %): intervention group (14/18, 78%); control group (11/18, 61%) CKD stage/kidney function: 5D
Interventions	Intervention group Pentoxifylline: 400 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Fondo de Fomento a la Investigacio´n (grant no. 2006‐1A‐I‐007)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated randomisation list
Allocation concealment (selection bias)	Low risk	Patients were computer‐allocated
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo with identical starch tablet)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but extracted outcome not reported in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Guo 2013.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: January 2008 to June 2009 Follow‐up 8 weeks
Participants	Study characteristics Country: Taiwan Setting: single centre Inclusion criteria: chronic HD Exclusion criteria: > 70 years; follow‐up < 6 months Baseline characteristics Number: randomised/analysed): intervention group (40/40); control group (40/25) Mean age ± SD (years): intervention group (59 ± 10); control group (61 ± 8) Sex (male, %): intervention group (22/40, 55%); control group (14/25, 56%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc: 11 mg once/day Control group Usual care (no supplement)
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed)
Notes	Additional information Funding sources: Kuang‐Tien General Hospital, Taichung, Taiwan
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	High risk	Allocation not concealed
Blinding of participants and personnel (performance bias)	High risk	No blinding of patients and personnel
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	High risk	All patients with follow‐up < 6 months excluded, loss to follow‐up high in control group (30%, reasons specified) and loss to follow‐up in intervention group unclear
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Haddadian‐Khouzani 2022.

Study characteristics
Methods	Study design Parallel RCT Time frame October 2020 to October 2021 Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: multicentre (2 sites) Inclusion criteria: > 18 years; chronic HD twice/week; no smoking, history of cancer or liver disease; not on parenteral or enteral feeding; follow HD diet Exclusion criteria: kidney transplant or death; switching to PD; compliance < 85%; Baseline characteristics Number: intervention group (44); control group (43) Mean age ± SD (years): intervention group (48 ± 15); control group (50 ± 13) Sex (male, %): intervention group (30/44, 69%); control group (28/43, 65%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc gluconate: 30 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: Isfahan University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated randomisation sequence
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Hajian 2022.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 2019 Follow‐up: 3 months
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: chronic HD; iron deficiency anaemia Exclusion criteria: history of inflammatory disease; IV administration of ferric oxide saccharate in the past 2 months Baseline characteristics Number: intervention group (16); control group (16) Mean age ± SD (years): intervention group (61 ± 14); control group (50 ± 22) Sex (male, %): intervention group (7/16, 44%); control group (11/16, 69%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin C: 500 mg 3 times/week Control group Placebo
Outcomes	Outcomes relevent to this review Death (any cause)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with WinPepsi software
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Hajji 2021.

Study characteristics
Methods	Study design Parallel RCT Time frame February 2017 to March 2018 Follow‐up: 60 days
Participants	Study characteristics Country: Tunisia Setting: multicentre (5 sites) Inclusion criteria: > 18 years; HD for at least 6 months Exclusion criteria: infection; GI and liver diseases; congestive heart failure, cancer; psychiatric illness, pregnancy; use of immunosuppressants, corticoids, oestrogens or contraceptives; active smoking; alcoholism Baseline characteristics Number (randomised/analysed): intervention group (43/17); control group (34/20) Mean age ± SD (years): intervention group (53 ± 14); control group (54 ± 16) Sex (male, %): intervention group (12/17, 71%); control group (11/20, 55%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 100 mg once/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Age and sex only for participants who completed study Funding sources: Research Laboratory (grant no. LR99ES11), Ministry of Higher Education and Scientific Research of Tunisia
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	Single blind (patients blinded with matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow‐up (54%). Reasons for loss to follow‐up specified
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not reported in method
Other bias	Low risk	Study appears free of other biases

HERO 2008.

Study characteristics
Methods	Study design Parallel RCT Stratified by centre Time frame Recruitment: June 2009 to December 2011 Follow‐up: 4 months
Participants	Study characteristics Country: Australia and New Zealand Setting: multicentre (number of sites not reported) Inclusion criteria: CKD 4‐5 (eGFR < 30 mL/min/1.73 m2, including dialysis); Hb < 110 g/L for at least 3 months Exclusion criteria: peptic ulcer disease; haematological disease; known haemoglobinopathy; major surgery; infection; acute MI; malignancy; haemorrhagic stroke or severe haemorrhage in the past 3 months; iron deficiency (ferritin < 100 μg/L or TSAT < 20%); vitamin B12 deficiency; folate deficiency; PTH > 100 pmol/L; aluminum > 2 μmol/L; Kt/V < 1.0 (HD) or < 1.7 (PD); use of melatonin, androgen; blood transfusion in the past 1 month; Vitamin C > 100 mg/day Baseline characteristics Number: intervention group (26); control group (27) Mean age ± SD (years): intervention group (59 ± 15); control group (65 ± 16) Sex (male, %): intervention group (14/26, 54%); control group (10/27, 37%) CKD stage/kidney function: intervention group (CKD 4: 0/26, CKD 5: 26/26); control group (CKD 4: 1/27, CKD 5: 26/27)
Interventions	Intervention group Pentoxifylline: 400 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) CVD
Notes	Additional information Funding sources: Roche Foundation for Anaemia Research (RoFAR), Amgen, Janssen‐Cilag and the National Health and Medical Research Council of Australia
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with central computer‐generated randomisation
Allocation concealment (selection bias)	Low risk	Patients were computer‐allocated
Blinding of participants and personnel (performance bias)	Low risk	Patients, personnel were blinded (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Outcome assessors were blinded (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	Although adverse events not specified in protocol all relevant outcomes in protocol reported
Other bias	High risk	Definition of ESA‐hyporesponsive anaemia revised due to slow recruitment and screen failures

Himmelfarb 2007.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: March to October 2002 Follow‐up: 8 weeks
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: HD; normal hepatic function and WCC; BMI < 35 kg/m2 Exclusion criteria: cancer in the past 5 years or recent bacterial infection; ECG abnormalities Baseline characteristics Number: intervention group (31); control group (32) Mean age ± SEM (years): intervention group (58.0 ± 2.0); control group (61.8 ± 1.8) Sex (male, %): intervention group (23/31, 74%); control group (17/32, 53%) CKD stage/kidney function: 5D
Interventions	Intervention group Gamma‐tocopherol: 308 mg/day Docosahexaenoic acid: 800 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Major bleeding
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients randomised with stratified permutated block randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Subjects, side staff, and investigators were blinded to treatment assignment (matched placebo)
Blinding of outcome assessors (detection bias)	Low risk	Subjects, side staff, and investigators were blinded to treatment assignment (matched placebo) and all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

HOPE 1996.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: mean 1.7 years
Participants	Study characteristics Countries: North and South America, Europe Setting: International (19 countries; number of sites not reported) Inclusion criteria: > 55 years; history of CVD or diabetes; ≥ 1 cardiovascular risk factor: total cholesterol > 5.2 mmol/L; HDL cholesterol ≤ 0.9 mmol/L; hypertension; microalbuminuria; smoking Exclusion criteria: MI; unstable angina or stroke in past month; history of congestive heart failure (or left ventricular ejection fraction < 40%); overt proteinuria; SCr > 200 µmol/L; hyperkalaemia; uncontrolled hypertension Baseline characteristics Number: intervention group (499); control group (494) Mean age ± SD (years): intervention group (67 ± 7); control group (68 ± 7) Sex (male, %): intervention group (441/499, 88%); control group (426/494, 86%) CKD stage: not reported
Interventions	Intervention group Vitamin E (RRR‐ɑ‐tocopheryl acetate): 400 IU/day Control group Placebo
Outcomes	Outcomes relevant to this review CVD Death (any cause) CVD: defined as a composite of MI, stroke, or death from CVD Coronary heart disease Heart failure: defined as hospitalisation for heart failure with clinical and radiologic signs of congestion Cerebrovascular disease Peripheral vascular disease UACR at end of study Change in UACR (SD for change imputed)
Notes	Additional information Post‐hoc analysis of an international, multicentre RCT‐ Heart Outcomes Prevention Evaluation (HOPE) Study Funding sources: Medical Research Council of Canada (Grants MT12790 and UI12362); Hoechst‐Marion Roussel; AstraZeneca; King Pharmaceuticals; Natural Source Vitamin E Association; NEGMA and the Heart and Stroke Foundation of Ontario
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central randomisation using a 4 digit code, performed in blocks of 8 and stratified per centre
Allocation concealment (selection bias)	Low risk	Study protocol has been published and has reported the use of central randomisation
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Compliance reported, all analyses have been reported to be conducted based on the intention‐to‐treat principle
Selective reporting (reporting bias)	Unclear risk	Study protocol has been published and has specified study design. All pre‐specified outcomes have been published
Other bias	High risk	Funding: This study was funded by the Medical Research Council of Canada (Grants MT12790 and UI12362); Hoechst‐Marion Roussel; AstraZeneca; King Pharmaceuticals; Natural Source Vitamin E Association; NEGMA and the Heart and Stroke Foundation of Ontario. Salim Yusuf was supported by a Senior Scientist award of the Medical Research Council of Canada, and a Heart and Stroke Foundation of Ontario research chair. John K. Olynyk received a grant from the National Health and Medical Research Council of Australia Practitioner Fellowship (1042370).

Hosseini 2021.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: August to September 2019 Follow‐up: 8 weeks
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: chronic HD; history of diabetes and zinc deficiency Exclusion criteria: history of inflammatory disease or hospitalisation in the past 3 months; use of steroids and NSAIDs Baseline characteristics Number: intervention group (21); control group (25) Mean age ± SD (years): intervention group (54 ± 5); control group (56 ± 7) Sex (male, %): intervention group (13/21, 62%); control group (14/25, 56%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 220 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr
Notes	Additional information Funding sources: Zahedan University of Medical Sciences (grant no. 9510)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a random numbers table
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Jern 2000.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 90 days
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: chronic HD; no hospitalisation in past 3 months; protein catabolic rate < 0.9 g/kg Exclusion criteria: not reported Baseline characteristics Number: intervention group (10); control group (10) Mean age (years): intervention group (47); control group (57) Sex (male, %): intervention group (1/10, 10%); control group (6/10, 60%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 50 mg/day Control group Placebo
Outcomes	Outcomes relevant to this reivew Coronary heart disease
Notes	Additional information Baseline table reports 11/10 patients are male in the intervention group (also reports 9/10 are female) Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	High loss to follow‐up (30%) but general reasons for dropping out described
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method and no protocol available
Other bias	Low risk	Study appears free of other biases

Jimenez‐Osorio 2016.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 8 weeks
Participants	Study characteristics Country: Mexico Setting: single centre Inclusion criteria: 20 to 70 years; diabetic or non‐diabetic proteinuric CKD; no evidence of UTI or heart failure (NYHA class III or IV) Exclusion criteria: PD; transplant; liver disease; chemotherapy in past 2 years Baseline characteristics Number: intervention group (52); control group (49) Mean age ± SD (years): intervention group (47 ± 10); control group (49 ± 6) Sex (male, %): intervention group (29/52, 56%0;) control group (32/49, 65%) CKD stage/kidney function: intervention group (43 ± 8 mL/min/1.73 m2); control group (36 ± 7 mL/min/1.73m2 )*
Interventions	Intervention group Curcumin: 107 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed)
Notes	Additional information Pooled mean for patients with non‐diabetic and diabetic CKD for intervention and control group Funding sources: CONACyT‐Mexico (grant no. SALUD‐2013‐01‐201519)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Low risk	Outcomes reported in methods and protocol on Clinicaltrials.gov
Other bias	Low risk	Study appears free of other biases

Khabbazi 2012.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 8 weeks
Participants	Study characteristics Country: Iran Setting: not reported Inclusion criteria: HD > 6 months Exclusion criteria: GI disorders; hepatitis; infectious disease; smokers transplant candidate; use of NSAIDs, SAIDs or valproic acid in the past month Baseline characteristics Number (randomised/analysed): intervention group (31/24); control group (32/28) Mean age ± SD (years): intervention group (54 ± 13); control group (54 ± 14) Sex (male, %): intervention group (16/24, 67%); control group (18/28, 64%) CKD stage/kidney function: 5D
Interventions	Intervention group Alpha‐lipoic acid: 600 mg Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: Vice Chancellor, and Drug Applied Research Center, Tabriz University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated random sequence which was kept in a remote secure location and administered by an independent third party who was not involved with the clinical conduct of study until all study data were collected and verified.
Allocation concealment (selection bias)	Low risk	Patients were randomised with a computer‐generated random sequence which was kept in a remote secure location and administered by an independent third party who was not involved with the clinical conduct of study until all study data were collected and verified.
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in method reported but no protocol available
Other bias	Unclear risk	Only per protocol analysis

Konigsrainer 1995.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 90 days
Participants	Study characteristics Country: Austria Setting: not reported Inclusion criteria: transplant recipient Exclusion criteria: not reported Baseline characteristics Number: intervention group (50); control group (50) Mean age (years): intervention group (47); control group (44) Sex (male, %): intervention group (27/50, 54%); control group (26/50, 52%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 800 mg twice/day Control group No treatment
Outcomes	Outcomes relevant to this review Cardiovascular death CVD Cerebrovascular disease SCr at end of study (no SD reported) Graft loss Infection
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	High risk	Allocation not concealed
Blinding of participants and personnel (performance bias)	High risk	No blinding (open label)
Blinding of outcome assessors (detection bias)	Low risk	No blinding (open label) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Lin 2008.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 12 months
Participants	Study characteristics Country: Taiwan Setting: single centre Inclusion criteria: 20 to 70 years; eGFR 10 to 60 mL/min/1.73 m2; proteinuria > 500 mg/g; treatment on losartan > 6 months Exclusion criteria: MI, coronary artery bypass graft, percutaneous coronary angioplasty, stroke or retinal haemorrhage in the past 6 months; abnormal liver function; congestive heart failure (NHYA class III or IV); obstructive uropathy or active malignancy; inability to discontinue chronic immunosuppressive drugs or NSAIDs Baseline characteristics Number: intervention group (29); control group (27) Mean age ± SD (years): intervention group (57 ± 2); control group (59 ± 2) Sex (male, %): intervention group (9/29, 31%); control group (10/27, 37%) CKD stage/kidney function: CKD 3‐5 Intervention group: eGFR (41 ± 3 mL/min/1.73 m2); CKD 3 (24/29), (CKD 4 3/29), CKD 5 (2/29) Control group: eGFR (39 ± 3 ml/min/1.73 m2 ); CKD 3 (19/27), CKD 4 (5/27) CKD 5 (3/27)
Interventions	Intervention group eGFR 30 to 60 mL/min/1.73 m2: pentoxifylline 400 mg twice/day eGFR 29 to 10 mL/min/1.73 m2: pentoxifylline 400 mg/day Control group Usual care
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) Major bleeding
Notes	Additional information Funding sources: National Science Council (grant no. NSC‐94‐2314‐B‐002‐186 and NSC‐95‐2314‐B‐002‐104‐MY3) and the National Taiwan University (Ta‐Tung Kidney Foundation, Mrs. Hsiu‐Chin Lee Kidney Research Foundation), Taipei, Taiwan
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with computer generated random numbers
Allocation concealment (selection bias)	High risk	There was no concealment of allocation
Blinding of participants and personnel (performance bias)	High risk	No blinding of participants and personnel
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Lu 2007.

Study characteristics
Methods	Study design Parallel RCT Time frame Mean follow‐up 9.8 ± 4.7 months
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: HD patients with AV fistula or polytetrafluoroethylene graft Exclusion criteria: clotting disorders; lower extremity vascular access, central venous catheter access, central venous stenosis; current use of anticoagulant therapy Baseline characteristics Number (randomised/analysed): intervention group (17/14); control group (17/13) Mean age, range (years): intervention group (54.7, 31.7 to 86.1); control group (53.0, 31.7 to 72.5) Sex (male, %): intervention group (4/14, 29%); control group (8/13, 62%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin E: 800 IU/day Control group Placebo
Outcomes	Outcomes relevant to this review Cardiovascular death Death (any cause) CVD Coronary heart disease Thrombosis of vascular access
Notes	Additional information Funding sources: Leonard B. Rosenberg Renal Research Foundation of the Center for Dialysis Care, Cleveland, OH. Dr Weiss was supported by grants ES11461 from the National Institute of Environmental Health Sciences (grant no. ES11461)and the National Institute of Diabetes and Digestive Kidney Diseases (grant no. DK57733 and DK45619) and Dr. Salomon by grant from National Institute of General Medical Sciences (grant no. GM21249)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a random number generated by the hospital pharmacy
Allocation concealment (selection bias)	Low risk	Patients were randomised with a random number generated by the hospital pharmacy
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Relevant outcomes not specified in methods
Other bias	Low risk	Study appears free of other biases

Martinez 2020.

Study characteristics
Methods	Study design Parallel RCT Time frame May 2016 to September 2019 Follow‐up 24 weeks
Participants	Study characteristics Country: Mexico Setting: not reported Inclusion criteria: DM with eGFR > 15 mL/min/1.73 m2; proteinuria > 1 g/day despite RAAS blockade Exclusion criteria: no DM; no KRT or transplant; pregnancy Baseline characteristics Number: intervention group (54); control group (46) Mean age ± SD (years): intervention group (59 ± 3); control group (57 ± 11) Sex (male, %): intervention group (31/54, 57%); control group (36/46, 78%) CKD stage/kidney function: intervention group (34 ± 16 mL/min/1.73 m2); control group (38 ± 17 mL/min/1.73 m2)
Interventions	Intervention group Curcumin: 560 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed)
Notes	Additional information Abstract‐only publication Funding sources: Government Support ‐ Non‐U.S.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding not specified but only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes reported in method but no protocol available
Other bias	High risk	Conference abstract only

Modarresi 2017.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 1 month
Participants	Study characteristics Country: Iran Setting: multicentre (2 sites) Inclusion criteria: 18 to 75 years; transplant recipients previously on dialysis Exclusion criteria: pre‐emptive transplant; multiple organ transplant; second transplantation; n‐acetylcysteine use in previous 6 months; history of sensitivity to sulfa drugs Baseline characteristics Number (randomised/analysed): intervention group (31/31); control group (31/26) Mean age ± SD (years): intervention group (43 ± 12); control group (47 ± 12) Sex (male, %): intervention group (21/31, 32%); control group (16/31, 62%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group N‐acetylcysteine: 600 mg before transplant and 600 mg twice/day for 5 days postoperatively Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Graft loss DGF
Notes	Additional information Funding sources: Shahid Beheshti University of Medical Sciences, Tehran, Iran
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients randomised with simple randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in method reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Modarresi 2018.

Study characteristics
Methods	Study design Parallel RCT Time frame May 2014 to December 2015 Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 18 to 75 years; deceased donor transplant recipient Exclusion criteria: coronary artery bypass graft or stroke in 6 months prior to transplant Baseline characteristics Number (randomise/analysed): intervention group (37/36); control group (37/34) Mean age ± SD (years): intervention group (39.7 ± 14.3); control group (42.4 ± 14.7) Sex (male, %): intervention group (25/36, 70%); control group (30/34, 88%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group N‐acetylcysteine: 600 mg before transplant and 600 mg twice/day for 5 days postoperatively Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study SCr at end of study Change in SCr (SD for change imputed) Graft loss DGF
Notes	Additional information Funding sources: Shahid Labbafinejad Medical Center and Shahid Beheshti University of Medical Sciences, School of Pharmacy (grant no. 313/363)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with computer random number generator
Allocation concealment (selection bias)	Low risk	Allocation concealed with opaque sealed envelopes
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Moist 2010.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: November 2007 to October 2008 Follow‐up: 48 hours
Participants	Study characteristics Country: Canada Setting: multicentre (2 sites) Inclusion criteria: > 18 years; CrCl 30 to 60 mL/min Exclusion criteria: radiocontrast procedure scheduled; medications that interfere with tubular secretion or creatinine production Baseline characteristics Number: intervention group (30); control group (30) Mean age ± SD (years): intervention group (69 ± 13); control group (72 ± 9 years) Sex (male, %): intervention group (21/30, 70%); control group (25/30, 83%) CKD stage/kidney function: CKD 3 Intervention group (43 ± 19 mL/min/1.73 m2); control group (45 ± 19 mL/min/1.73 m2)
Interventions	Intervention group N‐acetylcysteine: 1200 mg twice/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR Change in eGFR (SD for change imputed) SCr at end of study Change in SCr Change in SCr (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed)
Notes	Additional information Funding sources: The Lawson Health Research Institute, London, Canada, and Physicians Services Inc, Ontario, Canada
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a random‐numbers chart
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo and patients instructed to take intervention or placebo with ginger ale to mask taste and smell)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	Outcomes reported in methods and protocol on Clinicaltrials.gov
Other bias	Low risk	Study appears free of other biases

Moreillon 2013.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up 8 weeks
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: > 18 years; CKD stages 1‐5 Exclusion criteria: illness requiring hospitalisation, malabsorption syndromes; life expectancy < 3 months; change in body weight in past 6 months Baseline characteristics Number: intervention group (9); control group (7) Mean age ± SD (years): intervention group (53 ± 18); control group (59 ± 13) Sex (male, %): not reported CKD stage/kidney function: intervention group (46 ± 11 mL/min/1.73 m2); control group (48 ± 23 mL/min/1.73 m2)
Interventions	Intervention group Herbal supplement composed of curcumin and Boswellia serrata (824 mg purified turmeric extract, 95% curcuminoids, and 516 mg Boswellia serrata extract, 10% 3‐acetyl‐11‐keto‐β‐boswellic acid) twice/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) SCr at end of study Change in SCr (SD for change imputed) Cancer
Notes	Additional information Funding sources: supplements were donated by Life Extension Inc.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Low risk	Allocation concealment with identical medication bottles dispensed by pharmacist
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo in identical medication bottles)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo in identical medication bottles) and only objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes reported in method but no protocol available
Other bias	Unclear risk	Study appears free of other biases

Mori 2009.

Study characteristics
Methods	Study design 2x2 factorial RCT Time frame Follow‐up: 8 weeks
Participants	Study characteristics Country: Australia Setting: multicentre (2 sites) Inclusion criteria: 25 to 75 years; eGFR 15 to 60 mL/min/1.73 m2; SCr < 350 µmol/L Exclusion criteria: DM; angina pectoris; major surgery; a cardiovascular event or symptoms of CVD in the past 3 months; liver disease; nephrotic syndrome; BP > 170/100 mm Hg; Hb < 6.8 mmol/L; use of NSAIDs, immunosuppressants, nitrates, sildenafil; > 1 fish meal/week or regular fish oil supplements; > 4 alcoholic drinks/day Baseline characteristics Number (randomised/analysed): intervention group 1 (22/18); intervention group 2 (21/20); intervention group 3 (23/21); control group (19/15) Mean age ± SD (years): intervention group 1 (56.9 ± 3.9); intervention group 2 (53.3 ± 3.2); intervention group 3 (55.4 ± 2.7); control group (58.6 ± 2.6) Sex (male, %): intervention group 1 (17/18, 94%) intervention group 2 (12/20, 60%); intervention group 3 (17/21, 81%); control group (8/15, 53%) CKD stage/kidney function: intervention group 1 (30.5 ± 2.2 mL/min/1.73 m2); intervention group 2 (36.4 ± 2.8 mL/min/1.73 m2); intervention group 3 (38.8 ± 2.2 mL/min/1.73 m2); control group (34.6 ± 2.3 mL/min/1.73 m2)
Interventions	Intervention group 1 Omega‐3 fatty acids: 4g/day Coenzyme Q 10: 200 mg/day Intervention group 2 Omega‐3 fatty acids: 4g/day Intervention group 3 Coenzyme Q 10: 200 mg/day Control group Placebo with olive oil
Outcomes	Outcomes relevant to this review Death (any cause) eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study ACR at end of study Change in UACR (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed) Albuminuria at end of study Change in albuminuria (SD for change imputed)
Notes	Additional information Funding sources: National Health and Medical Research Council of Australia project grants (grant no. APP303151 and APP1010495)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Patients were randomised with computer‐generated random numbers.
Allocation concealment (selection bias)	Low risk	Allocation was concealed with numbered containers
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unequal loss to follow‐up (21%, 5%, 9%, and 8% respectively)
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported. Unable to retrieve protocol from the Australian New Zealand Clinical Trials registry
Other bias	High risk	Randomisation appears to have failed (Omega‐3 + coenzyme Q10 17/18 male, higher SCr and lower eGFR than other groups)

Navarro 1999.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Spain Setting: single centre Inclusion criteria: DM; CrCl < 30 mL/min Exclusion criteria: blood loss or abnormal iron status; treatment with ACEi, theophylline, androgens or ESA in the past 6 months Baseline characteristics Number: intervention group (7); control group (5) Mean age: ± SD (years): intervention group (64 ± 7); control group (65 ± 9) Sex (male, %): intervention group (5/7, 71%); control group (2/5, 40%) CKD stage/kidney function: intervention group (25 ± 4 mL/min/1.73 m2); control group (25 ± 4 mL/min/1.73 m2)
Interventions	Intervention group Pentoxifylline: 400 mg/day Control group No treatment
Outcomes	Outcomes relevant to this review CrCl at end of study Change in CrCl (SD for change imputed)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated randomisation table
Allocation concealment (selection bias)	High risk	There was no allocation of treatment
Blinding of participants and personnel (performance bias)	High risk	No blinding of participants and personnel
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Navarro 1999a.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Spain Setting: single centre Inclusion criteria: insulin‐dependent DM; CrCl < 35 mL/min Exclusion criteria: immunologic disease; malignancy; infection Number: intervention group (14); control group (10) Baseline characteristics Mean age ± SD (years): intervention group (64 ± 7); control group (65 ± 9) Sex (male, %): intervention group (10/14, 71%); control group (4/10, 40%) CKD stage/kidney function: intervention group (CrCl: 26 ± 5 mL/min); control group (CrCl: 24 ± 2 mL/min)
Interventions	Intervention group Pentoxifylline: 400 mg Control group No treatment
Outcomes	Outcomes relevant to this review CrCl at end of study Change in CrCl (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) Proteinuria at end of study
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	There was no allocation of treatment
Blinding of participants and personnel (performance bias)	High risk	No blinding of participants and personnel
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

NICE 2020.

Study characteristics
Methods	Study design Parallel RCT Time frame March 2016 to June 2017 Follow‐up: 12 weeks
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: eGFR < 60 mL/min or UACR > 30mg/g or protein/creatinine > 150 mg/g and eGFR > 60 mL/min Exclusion criteria: HIV; organ or bone marrow transplant; recent MI; stroke; TIA; AKI; hospitalisation in the past 3 months; infection; heart failure (NYHA III or IV); haemolytic anaemia; COPD, glucose‐6 phosphate dehydrogenase deficiency; dialysis Hb < 6.2 mmol/L; use or taking immunosuppression, chemotherapy, DPP‐5 inhibitor, nitrates, NSAIDs, allopurinol meperidine depressants, tobacco, alcohol or substances Baseline characteristics Number: intervention group (35); control group (35) Mean age ± SD (years): intervention group (61 ± 10); control group (62 ± 10) Sex (male, %): intervention group (22/35, 63%); control group (22/35, 63%) CKD stage/kidney function: intervention group: (51 ± 22 mL/min/1.73 m2); control group: (45 ± 16 mL/min/1.73 m2)
Interventions	Intervention group Sodium nitrite: 40 mg twice/day Isoquercetin: 225 mg once/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Change in eGFR Change in UACR
Notes	Additional infomation Funding sources: Louisiana Clinical and Translational Science Center National Institute of General Medical Sciences of the National Institutes of Health (grant no. U54 GM104940), Tulane Centers of Biomedical Research Excellence for Clinical and Translational Research in Cardiometabolic Diseases National Institute of Health (NIH), National Institute of General Medical Sciences (NIGMS) grant no (P20 GM109036) and Tulane Translational Research in Hypertension and Renal Biology NIH, NIGMS (grant no. P30‐GM103337)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with SAS randomisation software
Allocation concealment (selection bias)	Low risk	Allocation was concealed with concealed envelopes only opened by study pharmacist
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Blinding of investigators and lab technicians (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods and in protocol on ClinicalTrials.gov reported
Other bias	Low risk	Study appears free of other biases

Noel 1997.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: January 1993 to November 1994 Follow‐up 12 months
Participants	Study characteristics Country: France Setting: single centre Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics Number: intervention group (69); control group (68) Mean age ± SD (years): intervention group (39 ± 12); control group (39 ± 12) Sex (male, %): intervention group (33/69, 47%); control group (39/68, 57%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 400 mg twice/day; when SCr < 265 µmol/L, 400 mg 3 times/day for 6 months Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Graft loss DDF Infection
Notes	Additional information Funding sources: Direction de la Recherche et des Etudes Doctorales at the Centre Hospitalier Regional Universitaire of Lille, and Region Nord Pas de Calais (grant no. 2693 and 2686)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Double blind (method of blinding not specified)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (method of blinding not specified) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Norio 2003.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up 12 months
Participants	Study characteristics Country: Finland Setting: single centre Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics Number: intervention group (31); control group (31) Mean age ± SD (years): intervention group (48 ± 14); control group (51 ± 12) Sex (male, %): not reported CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Perfusion of graft with 500 mg vitamin C before transplantation Control group No treatment
Outcomes	Outcomes relevant to this review CrCl at end of study SCr at end of study Graft loss DDF
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation of patients was performed before surgery by means of a coded card in a sealed envelope
Allocation concealment (selection bias)	Low risk	Randomisation of patients was performed before surgery by means of a coded card in a sealed envelope
Blinding of participants and personnel (performance bias)	Low risk	Double blind (method of blinding not specified)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (method of blinding not specified) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Okamoto 2020.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Japan Setting: single centre Inclusion criteria: ≥ 20 years; chronic HD; zinc deficiency Exclusion criteria: end‐stage malignancy; history of blood transfusion; active inflammation; severe liver damage; low ESA resistance Baseline characteristics Number (randomised/analysed): intervention group 1 (47/44); intervention group 2 (47/47) Mean age ± SD (years): intervention group 1 (67 ± 11); intervention group 2 (68 ± 11) Sex (male, %): intervention group 1 (30/44, 68%); intervention group 2 (29/47, 62%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Zinc acetate aldehyde: 50 mg/day Intervention group 2 Polaprezinc: 34 mg/day
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: Oyokyo Kidney Institute
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with the excel randomisation function
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	No blinding (open label)
Blinding of outcome assessors (detection bias)	Low risk	Outcome assessors not blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow‐up, particularly in zinc acetate aldehyde group (30% vs 11%)
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	High risk	Study prematurely terminated due to safety concerns

Oko 1999.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: not reported
Participants	Study characteristics Country: Poland Setting: single centre Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics Number: 24 (number per group not reported) Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline for 7 days after transplant (dose not reported) Control group Placebo
Outcomes	Outcomes relevant to this review DGF
Notes	Addtional information Abstract‐only publication Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Not specified if participants and personnel were blinded
Blinding of outcome assessors (detection bias)	Low risk	Not specified if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	High risk	Conference abstract only

Omar 2022.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame August 2021 to November 2021 Follow‐up: 4 months
Participants	Study characteristics Country: Egypt Setting: single centre Inclusion criteria: > 18 years; HD ≥ twice/week for ≥ 3 months Exclusion criteria: active liver disease; treatment with antioxidants in the past 2 months Baseline characteristics Number (randomised/analysed): intervention group 1 (35/31); intervention group 2 (35/31); control group (35/32) Median age, range (years): intervention group 1 (44, 19 to 70); intervention group 2 (55, 18 to 74); control group (50, 18 to 70) Sex (male, %): intervention group 1 (12/31, 39%); intervention group 2 (13/31, 42%); control group (12/32, 38%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Rutin: 120 mg Vitamin C: 320 mg 3 times/day Usual care Intervention group 2 Vitamin C: 500 mg/day Usual care Control group Usual care: calcium acetate 700 mg/day, cinacalcet 30 mg/day, epoetin alpha, alfacalcidol 0.25 µg/2 days, acetylsalicylic acid 75 mg/day and antihypertensives
Outcomes	Outcomes relevant to rhis review Death (any cause)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	No blinding (open label)
Blinding of outcome assessors (detection bias)	Low risk	No blinding but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Orban 2015.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 12 months
Participants	Study characteristics Country: France Setting: single centre Inclusion criteria donors: > 18 years; clinical signs of brain death Exclusion criteria donors: medical contraindications for organ procurement; kidney graft stemming from another centre; medical complications before or during procurement leading to stopping the procedure Baseline characteristics (recipients) Number: intervention group (118); control group (118) Mean age ± SD (mean): intervention group (50 ± 12); control group (46 ± 14) Sex (male, %): intervention group (73/118, 62%); control group (70/118, 59%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Donors received n‐acetylcysteine 600 mg 3 times before transplant Control group Standard care
Outcomes	Outcomes relevant to this review eGFR at end of study SCr at end of study Graft loss DGF
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computerized random‐number generator
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Authors report conflicting information. Trial is open label but "the investigators, the allograft recipients, and the physicians, who provided cares and renal function assessment, were blinded regarding donor’s pretreatment"
Blinding of outcome assessors (detection bias)	Low risk	Authors report conflicting information. Trial is open label but "the investigators, the allograft recipients, and the physicians, who provided cares and renal function assessment, were blinded regarding donor’s pretreatment" but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	Outcomes reported in methods and protocol on Clinicaltrials.gov
Other bias	Low risk	Study appears free of other biases

Paniagua 1995.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 2 weeks
Participants	Study characteristics Country: Mexico Setting: single centre Inclusion criteria: chronic HD; clinical zinc deficiency Exclusion criteria: not reported Baseline characteristics Number: intervention group (5); control group (5) Mean age ± SEM (years): intervention group (23 ± 2); control group (26 ± 4) Sex (male, %): intervention group (2/5, 40%); control group (1/5, 10%) CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 100 mg Control group Placebo
Outcomes	Outcomes relevant to this review Change in SCr
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Patients were given placebo but unclear if patients or personnel were blinded
Blinding of outcome assessors (detection bias)	Low risk	Not specified if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

PATH 2014.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: May 2006 to July 2009 Follow‐up: 6 months
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: > 18 years; HD > 4 months; life expectancy > 1 year Exclusion criteria: AIDS; active malignancy (excluding basal cell carcinoma); functional transplant in the past 6 months; hospitalisation > 1 in the past 30 days or > 2 the past 90 days; parenteral nutrition or anti‐inflammatory medication Baseline characteristics Number (randomised/analysed): intervention group (174/160); control group (179/165) Mean age ± SD (years): intervention group (58 ± 13); control group (58 ± 12) Sex (male, %): intervention group (96/165, 58%); control group (86/160, 54%) CKD stage/kidney function: 5D
Interventions	Intervention group Mixed tocopherols: 660 IU Alpha‐lipoic acid: 600 IU Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: Clinical Translational Science Award from the National Center for Research Resources (grant no 1UL‐1RR024975), the National Institute of Diabetes and Digestive and Kidney Diseases (grant no. K24 DK 62849 and T32DK007467), the National Heart, Lung, and Blood Institute and Center in Molecular Toxicology (grant no R01 HL070938), the National Institute of Environmental Health Sciences (grant no. P30 ES000267), and an unrestricted gift from the Northwest Kidney Centers Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with permutated block randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow‐up (27%)
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but not in protocol in ClinicalTrials.gov
Other bias	High risk	One study site excluded from participation due to protocol violations

Pergola 2009a.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame Follow‐up: 28 days
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: DM with CKD Exclusion criteria: not reported Baseline characteristics Number: intervention group 1 (19); intervention group 2 (19); intervention group 3 (19) Mean age: 62 years Sex (male, %): not reported CKD stage/kidney function: intervention group 1 (37 mL/min/1.73 m2); intervention group 2 (33 mL/min/1.73 m2); intervention group 3 (39 mL/min/1.73 m2)
Interventions	Intervention group 1 Bardoxolone methyl: 25 mg Intervention group 2 Bardoxolone methyl: 75 mg Intervention group 3 Bardoxolone methyl: 150 mg
Outcomes	Outcomes relevant to this review Change in eGFR
Notes	Additional information Abstract‐only publications Funding sources: Affymax, Biogen Idec, Cytochroma, Amgen, Johnson & Johnson,Hoffman La Roche, Solvay; consultant: Reata Pharmaceuticals; other: Novartis; other relationship: Speaker’s Bureau
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Not specified if participants and personnel were blinded
Blinding of outcome assessors (detection bias)	Low risk	Not specified if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	High risk	Conference abstract only

Perkins 2009.

Study characteristics
Methods	Study design Parallel RCT Time frame January 2006 to April 2008 Follow‐up: 12 months
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: proteinuria ≥ 1 g/24 hours, hypertension and eGFR 20 to 40 mL/min; proteinuria ≥ 3 g/24 hours and eGFR ≥ 20 mL/min; use of ACEi or ARB for at least 2 weeks Exclusion criteria: AKI; transplant; retinal or cerebral haemorrhage; use of theophylline, cytotoxic drugs or steroids Baseline characteristics Number (randomised/analysed): intervention group (22/22); control group (18/17) Mean age ± SD (years): intervention group (63 ± 10); control group (65 ± 12) Sex (male, %): intervention group (14/22, 64%); control group (10/18, 59%) CKD stage/kidney function: intervention group (29 ± 10 mL/min/1.7 3 m2); control group (34 ± 14 mL/min/1.73 m2)
Interventions	Intervention group Pentoxifylline: 400 mg twice/day Control group Placebo
Outcomes	Outcomes relevant to this review Heart failure: defined as hospitalisation for heart failure Kidney failure eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed) Major bleeding Infection
Notes	Additional information Funding sources: Washington, DC, Chapter of the National Kidney Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated random number table
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported and in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Pollak 1993.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 48 hours
Participants	Study characteristics Country: USA Setting: multicentre (3 sites) Inclusion criteria: > 18 years; transplant recipient Exclusion criteria: planned transplantation with a cadaver kidney preserved in excess of 72 hours Baseline characteristics Number: intervention group (58); control group (58) Mean age ± SD (years): intervention group (41.5 ± 1.6); control group (41.5 ± 1.9) Sex (male, %): intervention group (35/58, 60%); control group (43/58, 74%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Human recombinant superoxide dismutase: parenteral vial containing 500 mg Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study CrCl at end of study SCr at end of study
Notes	Additional information Funding sources: Bristol‐Myers Squibb Company
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated assignment schedule
Allocation concealment (selection bias)	Low risk	Allocation assignment was concealed. The only randomisation master code was held by the Bristol‐Myers Company. No investigator had access to randomisation codes until study termination
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	High risk	Study reported in the methods that routine "serum chemistry, cyclosporine blood levels, and hematology" were collected, however, these were not reported in the results
Other bias	High risk	Study reported early termination by the monitor citing that it was "unlikely that a benefit would be shown for rh‐SOD by the addition of 84 extra subjects" Bristol‐Myers Squibb Company supported the study and provided the recombinant SOD. Bristol‐Myers Squibb Company held the only master code.

PREDIAN 2011.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 24 months
Participants	Study characteristics Country: Spain Setting: single centre Inclusion criteria: > 40 years; CKD 3‐4 with diabetic nephropathy; using ACEi or ARB Exclusion criteria: non‐diabetic kidney disease; chronic inflammation; immunologic or tumoral disease; acute inflammatory or infectious episode in past 3 months; use of immunotherapy or immunosuppressive; smoking; ACEi/ARB combined with aldosterone antagonist or direct renin inhibitor; BP ≥ 180/110 mm Hg Baseline characteristics Number (randomised/completed): intervention group (82/78); control group (87/82) Mean age ± SD (years): intervention group (70 ± 10); control group (70 ± 9) Sex (male, %): intervention group (45/82, 55%); control group (46/87, 53%) CKD stage/kidney function: intervention group (37 ± 12.4 mL/min/1.73 m2); control group (37 ± 11.9 mL/min/1.73 m2)
Interventions	Intervention group Pentoxifylline: 600 mg twice/day Control group Routine care
Outcomes	Outcomes relevant to this review Death (any cause) CVD: defined as nonfatal MI, heart failure resulting in hospitalisation, cerebrovascular event, or lower limb revascularisation or amputation Coronary heart disease Heart failure Peripheral vascular disease Kidney failure eGFR at end of study Change in eGFR Albuminuria at end of study Change in albuminuria (SD for change imputed) Infection
Notes	Additional information Funding sources: Institute of Health Carlos III (ISCIII), Spanish Ministry of Economy and Competitiveness, and Spanish Ministry of Health, Social Services and Equality (grant no E07/90021). Juan F. Navarro González is supported by Programa de Intensificación de la Actividad Investigadora (ISCIII/Comunidad Autónoma de Canarias)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with computer random numbers
Allocation concealment (selection bias)	Low risk	Allocation was concealed with sequentially numbered, opaque, sealed, and stapled envelopes
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Rabizadeh 2018.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 1 April to 30 September 2015 Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 30 to 80 years; type 2 diabetes and urinary albumin excretion ≥ 150 mg/day; use of losartan 50 mg > 3 months Exclusion criteria: type 1 diabetes; infectious or malignant disease; non‐diabetic kidney disease; retinal haemorrhage; acute MI; unstable angina; history of cardio‐cerebrovascular or peripheral arterial disease; uncontrolled hypertension (≥ 140/90mm Hg); anaemia; hyperthyroidism; history of HD; eGFR < 30 mL/min/1.73 m2; serum potassium ≥ 5.5 mmol/L Baseline characteristics Number*: intervention group (9); control group (8) Mean age ± SD (years): intervention group (57 ± 9); control group (57 ± 10) Sex (male, %): intervention group (7/9, 78%); control group (3/8, 38%) CKD stage/kidney function: intervention group (45 ± 8 mL/min/1.73 m2); control group (46 ± 11 mL/min/1.73 m2) *Only patients with eGFR < 60 mL/min/1.73 m2
Interventions	Intervention group Pentoxifylline: 400 mg twice/day Losartan: 50 mg/day Control group Losartan: 50 mg twice/day
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed)
Notes	Additional information Funding sources: Tehran University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with random allocation software
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	No blinding of participants and personnel (open label)
Blinding of outcome assessors (detection bias)	Low risk	No blinding of participants and personnel (open label) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Low loss to follow‐up but reasons not specified
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Rabl 1993.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 7 days
Participants	Study characteristics Country: Austria Setting: single centre Inclusion criteria: surgical revascularisation operations for transplantation Exclusion criteria: not reported Baseline characteristics Number: intervention group (16); control group (14) Mean age ± SD (years): intervention group (42 ± 14); control group (43 ± 10) Sex (male, %): 22/30, 73% CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Omnibionta (antioxidant solution): 2 ampoules (10 mL each), infused before transplant Control group Usual care
Outcomes	Outcomes relevant to this review CrCl at end of study SCr at end of study
Notes	Additional infromation Funding sources: Association for International Cancer Research
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Not specified if patients and personnel were blinded
Blinding of outcome assessors (detection bias)	Low risk	Not specified if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All outcomes in methods reported but not protocol available
Other bias	Low risk	Study appears free of other biases

Rassaf 2016.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 2012 to 2013 Follow‐up 30 days
Participants	Study design Country: Germany Setting: single centre Inclusion criteria: > 18 years; HD with non‐fistula brachial artery Exclusion criteria: AKI; acute infection; heart failure; anuria Baseline characteristics Number (randomised/analysed): intervention group (26/24); control group (26/25) Mean age ± SD (years): intervention group (65 ± 14); control group (66 ± 14) Sex (male, $%): intervention group (18/26, 69%); control group (20/26, 77%) CKD stage/kidney function: 5D
Interventions	Intervention group Cocoa‐flavanol‐rich supplements: 900 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Change in SCr (SD for change imputed) Major bleeding
Notes	Additional information Funding sources: European Commission Flaviola (grant no. FP7‐KBBE 226588). Mars Symbioscience provided the cocoa flavanol test products and analytical standards. Tienush Rassaf is a Heisenberg Professor funded by the German Research Foundation (grant no Ra969/7‐2)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐randomised generation sequence
Allocation concealment (selection bias)	Low risk	Allocation was concealed with opaque sachets
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but not in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Rivara 2015.

Study characteristics
Methods	Study design Cross‐over RCT Time frame Recruitment: March to October 2012 Follow‐up: 4 weeks
Participants	Study characteristics Country: USA Setting: multicentre (number of sites not reported) Inclusion criteria: 18 to 85 years; chronic HD; life expectancy > 1 years Exclusion criteria: AIDS; active malignancy (excluding basal cell carcinoma); history of major atherosclerotic event functional transplant in the past 6 months; hospitalisation in the past 30 days; parenteral nutrition Baseline characteristics Number: intervention group 1 (11); intervention group 2 (13) Mean age ± SD (years): intervention group 1 (62 ± 10); intervention group 2 (59 ± 17) Sex (male, %): intervention group 1 (4/11, 36%); intervention group 2 (7/13, 54%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Pomegranate juice: 100 mL 3 times/week Intervention group 2 Pomegranate extract: 1050 mg/day
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: POM Wonderful, LLC to the Kidney Research Institute, NIH, National Heart, Lung, and Blood Institute, NIH (grant no R01 HL070938), National Institute of Environmental Health Sciences, and NIH (grant no P30 ES000267), and the National Institute of Diabetes and Digestive and Kidney Diseases (grant no. 5T32DK007467‐30)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	No blinding of patients and personnel (open label)
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in methods or protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Roozbeh 2009.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 weeks
Participants	Study characteristics Country: Iran Setting: multicentre (2 sites) Inclusion criteria: > 18 years' chronic HD; zinc deficiency Exclusion criteria: not reported Baseline characteristics Number (randomised/analysed): 60/53; intervention group (27); control group (26) Mean age: 56 years Sex (male, %): 28/53, 53% CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 220 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up specified, general reasons for loss to up specified
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Sahraei 2015.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame January 2011 to February 2013 Follow‐up: 60 days
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: > 18 years; transplant from living donor Exclusion criteria: active infectious disease; malignancy; inflammatory disease; Cushing syndrome; sepsis; sickle cell anaemia; cardiorenal syndrome; multiple sclerosis; acute pancreatitis; coronary artery bypass grafting; liver cirrhosis; Alzheimer disease; stroke; hyperoxaluria; mood disorders; schizophrenia; use of cimetidine Baseline characteristics Number (randomised/analysed): 89/84; intervention group 1 (19); intervention group 2 (33); control group (32) Mean age ± SD (years): intervention group 1 (37 ± 17); intervention group 2 (37 ± 17); control group (36 ± 17) Sex (male, %): intervention group 1 (12/19, 63%); intervention group 2 (16/33, 48%); control group (20/32, 63%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group 1 N‐acetylcysteine 600 mg + vitamin C 75 mg before transplant, and 12 and 18 hours after transplant Intervention group 2 N‐acetylcysteine 600 mg before transplant, and 12 and 18 hours after transplant Control group Routine care
Outcomes	Outcomes relevant to this review eGFR at end of study DGF
Notes	Additional information Funding sources: School of Pharmacy, and Urology and Nephrology Research Center of Shahid Beheshti University of Medical Sciences, Tehran, Iran
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a Microsoft Excel random number generator
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of patients and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	High risk	Almost all outcomes in methods reported. Serum creatine appears to have been substituted for eGFR
Other bias	High risk	Large variation in sample size different study arms (19, 33, and 32 patients respectively)

Salehi 2013.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: March to June 2009 Follow‐up: 12 weeks
Participants	Study design Country: Iran Setting: single centre Inclusion criteria: 18 to 80 years; HD Exclusion criteria: active infection or hospitalisation in the past month; use of immunosuppressive medications Baseline characteristics Number (randomised/analysed): intervention group (40/29); control group (40/36) Mean age ± SD (years): intervention group (50 ± 15); control group (55 ± 13) Sex (male, %): intervention group (16/40, 40%); control group (20/40, 50%) CKD stage/kidney function: 5D
Interventions	Intervention group Selenium: 200 µg/day Control group Placebo
Outcomes	Outcomes relevant to this review Cardiovascular death Death (any cause) CVD Coronary heart disease Cerebrovascular disease
Notes	Additional information Funding sources: Shiraz University of Medical Sciences, Shiraz, Iran
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with random allocation software, performed by third party
Allocation concealment (selection bias)	Low risk	Allocation was concealed with prepackaged bottles numbered according to randomisation sequence
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	Unequal loss to follow up; much higher in treatment arm (23 vs 5%). Mostly due to nTx
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not reported in methods or protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Schneeberger 1990.

Study characteristics
Methods	Study design Parallel RCT Time frame January 1987 to October 1988 Follow‐up: 5 years
Participants	Study characteristics Country: Germany Setting: single centre Inclusion criteria: cadaveric transplant recipient Exclusion criteria: not reported Baseline characteristics Number: intervention group (81); control group (96) Mean age ± SD (years): intervention group (46 ± 12); control group (46 ± 13) Sex (male, %): intervention group (53/81, 65%); control group (77/96, 80%) CKD stage/kidney function: 5D
Interventions	Intervention group Recombinant human superoxide dismutase: 200 mg Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Kidney failure: defined as a return to dialysis Graft loss
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Unclear if allocation was concealed. One of the authors (HS) of the study (although not involved in the patients' care) received the "randomisation code". This author was responsible for collecting all information about study. Study reports that the physicians in charge and the patients were not informed about the initial treatment. No information was provided regarding randomisation methodology
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not reported
Selective reporting (reporting bias)	Unclear risk	All outcomes in methods reported but no protocol available
Other bias	High risk	Funding: Two authors employees of 1) Shaman Pharmaceuticals (San Carlos, CA) and 2) The Lipsome Company Inc (Princeton, NJ)

Schramm 2002.

Study characteristics
Methods	Study design Parallel RCT Time frame Mean follow‐up: 57 ± 7 months
Participants	Study characteristics Country: Germany Setting: single centre Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics (recipients) Number: intervention group (28); control group (26) Mean age ± SD (years): intervention group (45.6 ± 1.94); control group (49.4 ± 2.33) Sex (male, %): 41/62, 66% CKD stage/kidney function: 5‐transplant
Interventions	Intervention group L‐arginine: 0.75 mg/kg 3 days post‐operatively Control group Placebo
Outcomes	Outcomes relevant to this review CrCl at end of study Graft loss DGF
Notes	Additional information Funding sources: grants from Deutsche Forschungsgeneinschaft (SFB 355/C7; Schr 452/1‐1) and the Thyssen Foundation (1994/1)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	Single blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	No blinding (single blind) but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Sengupta 2022.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame Follow‐up: 180 days
Participants	Study characteristics Country: India Setting: single centre Inclusion criteria: 18 to 80 years; CKD 3‐4; life expectancy > 1 year; MAP < 125 mm Hg; no diabetes Exclusion criteria: ongoing inflammation or AKI; dialysis or expected kidney transplant within 6 months or a history of kidney transplant; GI dysfunction requiring parental nutrition; serious disease (e.g. cirrhosis, heart failure NYHA stage IV, stroke) in the past 3 months; severe oedema or serious cavity effusion; malignancy in past 6 months; proteinuria > 10 mg/day; Hb < 5 or > 10 mmol/L; BMI < 18 or > 30 kg/m2; ongoing treatment for chronic infections; systemic steroids or immunosuppressive agents; intolerance to N‐acetylcysteine, taurine or pyridoxamine dihydrochloride Baseline characteristics Number: intervention group 1 (23); intervention group 2 (24); control group (22) Mean age ± SD (years): intervention group 1 (57 ± 14); intervention group 2 (57 ± 14); control group (59 ± 12) Sex (male, %): intervention group 1 (16/23, 70%); intervention group 2 (14/24, 58%); control group (15/22, 68%) CKD stage/kidney function: CKD 3 Intervention group 1 (37 ± 14 mL/min/1.73 m2); intervention group 2 (34 ± 12 mL/min/1.73 m2); control group (40 ± 18 mL/min/1.73 m2)
Interventions	Intervention group 1 Taurine: 500 mg N‐acetylcysteine: 150 mg Low‐protein diet: 0.6 g/kg body weight/day Intervention group 2 Pyridoxamine dihydrochloride: 50 mg N‐acetylcysteine: 300 mg Low‐protein diet: 0.6 g/kg body weight/day Control group Low‐protein diet: 0.6 g/kg body weight/day
Outcomes	Outcomes relevant to this review Change in eGFR
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Participants were presumably blinded (matching placebo); unclear if personnel was blinded
Blinding of outcome assessors (detection bias)	Low risk	Unclear if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported
Other bias	Low risk	Study appears free of other biases

Shema‐Didi 2012.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 12 months
Participants	Study characteristics Country: Israel Setting: single centre Inclusion criteria: > 18 years; chronic HD Exclusion criteria: HD < 3 months; participation in other studies; pregnant or planning a pregnancy Baseline characteristics Number (randomised/analysed): intervention group (66/41); control group (35/26) Mean age ± SD (years): intervention group (66 ± 11); control group (68 ± 13) Sex (male, %): intervention group (37/66, 56%); control group (18/35, 51%) CKD stage/kidney function: 5D
Interventions	Intervention group Pomegranate juice: 100 mL 3 times/week Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) CVD: defined as hospitalisations due to acute MI (fatal and non‐fatal); ischaemic stroke; new events of peripheral vascular disease (excluding the AV fistula) and unstable angina Infection
Notes	Additional information Funding sources: Chief Scientist Office of the Ministry of Health, Israel (grant no. 6186), the Jess & Midred Fisher Family Cardiology Research Fund (grant no. 2012255), the Office of the Executive Vice President for Research, Technion, Israel, and Iscar Ltd
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow‐up (consent withdrawal rate 15% and 6% in intervention and control group)
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported. Outcomes not reported in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Shojaei 2011.

Study characteristics
Methods	Study design Parallel, factorial RCT Time frame Follow‐up 3 months
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: chronic HD; elevated cholesterol treated with statins Exclusion criteria: acute inflammation; taking anti‐inflammatory or anti‐oxidative medication Baseline characteristics Number (randomised/analysed): total (64/52); intervention group 1 (14); intervention group 2 (12); intervention group 3 (13); control group (13) Mean age ± SD (years): intervention group 1 (53 ± 10); intervention group 2 (55 ± 16); intervention group 3 (54 ± 12); control group (52 ± 19) Sex (male, %): intervention group 1 (6/14, 43%); intervention group 2 (6/12, 50%); intervention group 3 (7/13, 54%); control group (6/13, 46%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Carnitine: 1000 mg 3 times/week Coenzyme Q10: 100 mg/day Intervention group 2 Carnitine: 1000 mg 3 times/week Intervention group 3 Coenzyme Q10: 100 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review CVD Infection
Notes	Additional information Funding sources: Tehran University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	High loss to follow up (almost 20%). Unclear if there are difference in rate of loss to follow‐up between study arms
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Shoskes 2005.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame September 2002 to August 2004 Follow‐up: 30 days
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: cadaveric tranplant Exclusion criteria: not reported Baseline characteristics Number: intervention group 1 (14); intervention group 2 (14); control group (15) Median age, range (years): intervention group 1 (46, 33 to 71); intervention group 2 (53, 20 to 74); control group (44, 19 to 74) Sex (male, %): intervention group 1 (7/14, 50%); intervention group 2 (7/14, 50%); control group (11/15, 71%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group 1 Oxy‐Q (480 mg of curcumin and 20 mg of quercetin): twice/day Intervention group 2 Oxy‐Q (480 mg of curcumin and 20 mg of quercetin): once/day Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study DGF
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Patients were blinded (matching placebo). Unclear if personnel was blinded
Blinding of outcome assessors (detection bias)	Low risk	Unclear if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	High risk	Control group had a higher proportion of men (71%) compared with the 2 intervention groups (50%) Funding: not stated. Bioflavonoid preparation Oxy‐Q (Farr Labs, Santa Monica, CA)

Silveira 2019.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 12 months
Participants	Study characteristics Country: Brazil Setting: single centre Inclusion criteria: 18 to 90 years; eGFR 25 to 70 mL/min; proteinuria > 300 mg/day Exclusion criteria: transplant; malignancy; use of immunosuppressive therapy; pregnancy Baseline characteristics Number (randomised/analysed): intervention group (18/18); control group (16/14) Mean age ± SD (years): intervention group (61 ± 11); control group (62 ± 11) Sex (male, %): intervention group (11/18, 61%); control group (7/14, 50%) CKD stage/kidney function: CKD 3‐4 Intervention group (37 ± 15 mL/min/1.73 m2); control group (35 ± 15 mL/min/1.73 m2)
Interventions	Intervention group Propolis: 500 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) eGFR at end of study Change in eGFR (SD for change imputed) UACR at end of study Change in UACR (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed)
Notes	Additional information Funding sources: L. Andrade received a grant from the Brazilian Conselho Nacional de Desenvolvimento Científico e Tecnológico grant no. 301193/2016–9) and AC Seguro received a grant from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP; grant no 2015/11933–3)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised by an external investigator
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and in protocol on ClinicalTrials.gov
Other bias	Low risk	None

Singer 2011.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 3 months
Participants	Study characteristics Country: Australia Setting: single centre Inclusion criteria: maintenance dialysis (HD or PD) or eGFR < 20 mL/min Exclusion criteria: life expectancy < 3 months; hyperoxaluria Baseline characteristics Number (randomised/analysed): intervention group (49/48); control group (51/48) Mean age ± SEM (years): intervention group (65 ± 13); control group (63 ± 15) Sex (male, %): intervention group (32/49, 65%); control group (36/50, 73%) CKD stage/kidney function: CKD 4‐5D
Interventions	Intervention group Vitamin C: 250 mg 3 times/week Control group Placebo
Outcomes	Outcomes relevant to this review Coronary heart disease Kidney failure eGFR at end of study Change in eGFR (SD for change imputed) Thrombosis of vascular access Infection
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated number sequence
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Singh 2000.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 28 days
Participants	Study characteristics Country: India Setting: single‐centre Inclusion criteria: chronic dialysis or advised to start with dialysis Exclusion criteria: AKI; obstructive uropathy; cancer; seriously ill patients with marked acidosis or shock Baseline characteristics Number: intervention group (11); control group (10) Mean age ± SD (years): intervention group (44 ± 10); control group (44 ± 9) Sex (male, %): intervention group (8/11, 72%); control group (7/10, 70%) CKD stage/kidney function: 5D
Interventions	Intervention group Coenzyme Q10: 60 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed)
Notes	Addtional information Funding sources: Centre of Nutrition and Tishcon Corporation, Westbury, NY, USA (provided coenzyme Q10)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients were randomised by the pharmacist but it is unclear how the random sequence was generated. Patients were asked to select one of two enclosed in sealed envelopes marked with group A or group B
Allocation concealment (selection bias)	Unclear risk	Allocation was concealed with enclosed sealed envelopes but is unclear if these were opaque
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Funding: financial support from the Centre of Nutrition Tishcon Corporation, Westbury, NY provided coenzyme Q10

Soliman 2014.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 90 days
Participants	Study characteristics Country: Egypt Setting: single centre Inclusion criteria: > 60 years; transplant recipient Exclusion criteria: unable to drink N‐acetylcysteine during the 1st 7 days postoperatively Baseline characteristics Number: intervention group (16); control group (11) Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5‐transplant
Interventions	Intervention group N‐acetylcysteine: 600 mg twice/day for 8 days postoperatively Control group Placebo
Outcomes	Outcomes relevant to this review CrCl at end of study DGF
Notes	Additional information Abstract‐only publication Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not reported
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	High risk	Conference abstract only

Song 2006a.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame Follow‐up: 12 months
Participants	Study characteristics Country: South Korea Setting: not reported Inclusion criteria: CKD 3‐4 with diabetic nephropathy Exclusion criteria: not reported Baseline characteristics Number: intervention group 1 (18); intervention group 2 group (18); control group (17) Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: CKD 3‐4
Interventions	Intervention group 1 Alpha‐lipoic acid: 600 mg Eprosartan: 600 mg Irbesartan: 300 mg Intervention group 2 Alpha‐lipoic acid: 600 mg Control group Eprosartan: 600 mg Irbesartan: 300 mg
Outcomes	Outcomes relevant to this review Proteinuria at end of study
Notes	Additional information Abstract‐only publication Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of participants and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	High risk	Conference abstract only

SPACE 2000.

Study characteristics
Methods	Study design Parallel RCT Time frame 1 November 1997 to 31 December 1999 Mean follow‐up: 1.4 years
Participants	Study characteristics Country: Israel Setting: multicentre (6 sites) Inclusion criteria: 40 to 75 years; stable chronic HD with history of CVD Exclusion criteria: known history of malignant disease (except non‐melanoma skin cancer); active liver disease anticoagulant therapy with warfarin or hypoglycaemic agents in the past 8 weeks Baseline characteristics Number: intervention group (97); control group (99) Mean age ± SD (years): intervention group (65 ± 8); control group (64 ± 9) Sex (male, %): intervention group (67/97, 69%); control group (68/99, 69%) CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin E (natural ɑ‐tocopherol): 800 IU/day Control group Placebo
Outcomes	Outcomes relevant to this review Cardiovascular death: defined as a composite of fatal myocardial infarction, ischaemic stroke or sudden death Death (any cause) CVD: defined as a composite of acute MI (fatal and nonfatal); ischaemic stroke; peripheral vascular disease (excluding AV fistulas) in a limb not previously affected; and unstable angina Coronary heart disease Cerebrovascular disease Peripheral vascular disease Cancer Major bleeding Infection
Notes	Additional information Funding sources: Chief Scientist’s Office, Ministry of Health, Israel (grant no 4204). Solgar, Inc, New York, USA and Henkel Corp, La Grange, IL, USA provided vitamin E
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer generated coin toss. Each participating centre randomised separately
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up reported
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Funding: This research was funded by grant 4204 from the Chief Scientist’s Office, Ministry of Health, Israel Vitamin E provided by Solgar, Inc, New York, USA, during the first year and Henkel Corp, La Grange, IL, USA, during the second year.

Tan 2019.

Study characteristics
Methods	Study design Parallel RCT Time frame 2019 to September 2020 Follow‐up: 12 months
Participants	Study characteristics Country: Malaysia Setting: multicentre (2 sites) Inclusion criteria: HbA1c 42 to 75 mmol/mol (6% to 9%); BP < 150/90 mm Hg; eGFR 30 to 60 mL/min/1.73 m2 or UACR >20 to 200mg/mmol Exclusion criteria: CKD 5; non‐diabetic kidney disease; severe chronic illness Baseline characteristics Number (randomised/analysed) intervention group (31/31); control group (30/28) Mean age ± SD (years): intervention group (66 ± 13); control group (70 ± 13) Sex (male, %): intervention group (20/31, 53%); control group (18/28, 47%) CKD stage/kidney function: CKD 3‐4 Intervention group (55 ± 18 mL/min/1.73 m2); control group (56 ± 20 mL/min/1.73 m2)
Interventions	Intervention group Tocopherol‐rich vitamin E: 400 mg Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Cerebrovascular disease eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) UACR at end of study Change in UACR (SD for change imputed) Infection
Notes	Additional information Funding sources: Malaysian Ministry of Higher Education (grant number FRGS/1/2018/SKK02/MUSM/02/1) and Monash University Malaysia under the Tropical Medicine and Biology (grant no. TMB‐2018‐CR3185140918‐SuzTMQ/KAK)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised by an independent party using computer‐generated random sequence
Allocation concealment (selection bias)	Low risk	Drug codes were assigned to patients by an independent party
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up intervention arm >20% compared to 7% control
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Tepel 2003.

Study characteristics
Methods	Study design Parallel RCT Time frame 1 October 1999 to 30 September 2001 Mean follow‐up: 1.2 years
Participants	Study characteristics Country: Germany Setting: single centre Inclusion criteria: HD > 3 months Exclusion criteria: know allergy to N‐acetylcysteine Baseline characteristics Number: intervention group (64); control group (70) Mean age ± SD (years): intervention group (63 ± 14); control group (62 ± 18) Sex (male, %): intervention group (33/64, 52%); control group (43/70, 61%) CKD stage/kidney function: 5D
Interventions	Intervention group N‐acetylcysteine: 600 mg twice/day Control group Matching placebo
Outcomes	Outcomes relevant to this review Cardiovascular death: defined as a composite of fatal MI, fatal stroke, and sudden death Death (any cause) CVD: defined as a composite of cardiac events including fatal and nonfatal MI, CVD death, need for coronary angioplasty or coronary bypass surgery, ischaemic stroke, peripheral vascular disease with amputation, or need for angioplasty Coronary heart disease: defined as a composite of fatal and nonfatal MI and need for coronary angioplasty or coronary bypass surgery Cerebrovascular disease Peripheral vascular disease
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	No blinding of patients and personnel
Blinding of outcome assessors (detection bias)	Low risk	No blinding of outcome assessors but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Thaha 2009.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 90 days
Participants	Study characteristics Country: Japan Setting: not reported Inclusion criteria: CKD 1‐4 Exclusion criteria: not reported Baseline characteristics Number: intervention group (11); control group (8) Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 1‐4
Interventions	Intervention group N‐acetylcysteine: 600 mg twice/day ACEi or ARB Control group ACEi or ARB
Outcomes	Outcomes relevant to this review Change in proteinuria
Notes	Additional information Abstract‐only information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of patients and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods and in interim analysis
Other bias	High risk	Conference abstract only

Tonelli 2015.

Study characteristics
Methods	Study design Parallel multi‐arm RCT Time frame Recruitment: November 2012 to June Follow‐up: 180 days
Participants	Study characteristics Country: Canada Setting: single centre Inclusion criteria: > 18 years; chronic HD; receiving renal vitamins Exclusion criteria: scheduled kidney transplant or GI surgery; short gut syndrome; life expectancy < 6 months; head and neck cancer Baseline characteristics Number (randomised/analysed): intervention group 1 (52/47); intervention group 2 (47/42); control group (51/45) Mean age, 95% CI (years): intervention group 1 (63, 59 to 67); intervention group 2 (60, 57 to 64); control group (58, 54 to 62) Sex (male, %): intervention group 1 (39/52, 75%); intervention group 2 (35/47, 75%); control group (37/51, 73%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Vitamin E: 250 IU Zinc: 50 mg Selenium: 75 µg Standard kidney vitamins Intervention group 2 Vitamin E: 250 IU Zinc 25 mg Selenium 50 µg Standard kidney vitamins Control group Standard kidney vitamins (biotin 300 µg, folic acid 1 mg, niacinamide 20 mg, thiamine 1.5 mg, cyanocobalamin 6 µg, riboflavin 1.7 mg, pyridoxine 10 mg, ascorbic acid 100 mg)
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: Northern Alberta Renal Program, the Canadian Institutes for Health Research (200703MOP‐171365‐POP‐CBAA‐62156), Canada Foundation for Innovation (Leaders Opportunity Fund grant). Marcello Tonelli has received an AHFMR Population Health Schilar award and Michael Walsh has received a New Investigator Award from the Canadian Institutes of Health Research
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with randomly generated lists, generated by the statician
Allocation concealment (selection bias)	Low risk	Allocation was concealed with identical medication bottles
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free from other biases

TSUBAKI 2020.

Study characteristics
Methods	Study design Parallel RCT Time frame 15 December 2015 to 28 July 2016 Follow‐up: 16 weeks
Participants	Study characteristics Country: Japan Setting: multicentre (36 sites) Inclusion criteria: 20 to 79 years; CKD 3 or 4; using ACEi or ARB Exclusion criteria: BNP > 200 pg/mL; history of CVD Baseline characteristics Number Intervention group: CKD 3 (41); CKD 4 (24) Control group: CKD 3 (41); CKD 4 (14) Mean age ± SD (years) Intervention group: CKD 3 (68 ± 7); CKD 4 (68 ± 9) Control group: CKD 3 (71 ± 7); CKD 4 (66 ± 7) Sex (male, %) Intervention group: CKD 3 (26/41, 63%); CKD 4 (17/24, 71%) Control group: CKD 3 (30/41, 73%); CKD 4 (9/14, 64%) CKD stage/kidney function Intervention group: CKD 3 (47 ± 8 mL/min/1.73 m2); CKD 4 (26 ± 4 mL/min/1.73 m2) Control group: CKD (3 47 ± 7 mL/min/1.73 m2); CKD 4 (23 ± 3 mL/min/1.73 m2)
Interventions	Intervention group Bardoloxone methyl: 5 mg/day up titrated to 10 mg/day at week 4 and 15 mg/day at week 8 Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Coronary heart disease Heart failure Change in eGFR Infection
Notes	Addtional information Funding sources: Kyowa Kirin
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Method of randomisation not specified. Sample size for CKD 4 cohort between intervention and control very uneven
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up intervention arm 20% compared to 3% control
Selective reporting (reporting bias)	Low risk	All relevant outcomes in methods reported and protocol on ClinicalTrials.gov
Other bias	High risk	Study was sponsored by Kyowa Kirin which was also involved in the interpretation of the data

Turki 2016.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Tunisia Setting: single centre Inclusion criteria: CKD 2‐4 Exclusion criteria: HD; smoking Baseline characteristics Number: intervention group (23); control group (10) Mean age ± SD (years): intervention group (63 ± 2); control group (62 ± 2) Sex (male, %): intervention group (13/23, 57%); control group (6/10, 60%) CKD stage/kidney function Intervention group: CKD 2 (6); CKD 3 (11); CKD 4 (6) Control group: CKD 2 (3); CKD 3 (4); CKD 4 (3)
Interventions	Intervention group Grape seed extract: 350 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Kidney failure eGFR at end of study Change in eGFR (SD for change imputed)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow up (1 patient), although reason not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Vafadar Afshar 2020.

Study characteristics
Methods	Study design Parallel RCT Time frame December 2017 to October 2018 Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: single centre Inclusion criteria: 18 to 80 years; chronic HD Exclusion criteria: thyroid disease; liver disease; active infection; chronic infection or inflammation; surgery or hospitalisation in the past month; use of steroids or NSAIDs Baseline characteristics Number: intervention group (30); control group (30) Mean age ± SD (years): intervention group (55 ± 13); control group (59 ± 8) Sex (male, %): intervention group (18/27, 67%); control group (16/27, 59%) CKD stage/kidney function: 5D
Interventions	Intervention group Nanocurcumin: 40 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause)
Notes	Additional information Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

Vincenti 1996.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: USA Setting: multicentre (4 sites) Inclusion criteria: transplant recipients Exclusion criteria: not reported Baseline characteristics Number: intervention group (22); control group (24) Mean age ± SD (years): intervention group (37 ± 10); control group (37 ± 11) Sex (male, %): intervention group (17/22, 77%); control group (13/24, 54%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 800 mg 3 times/day 24 to 72 hours before OKT3 (immunosuppressant) to 3 days after OKT3 Control group Placebo
Outcomes	Outcomes relevant to this review SCr at end of study Graft loss
Notes	Additional information Funding sources: Ortho Biotech
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Voroneanu 2017.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 24 months
Participants	Study characteristics Country: Romania Setting: single centre Inclusion criteria: > 18 years; type 2 diabetes; proteinuria > 500 mg/24 hours; treatment with ACEi or ARB > 6 months Exclusion criteria: advanced CKD (eGFR < 15 mL/min/1.73 m2); severely uncontrolled type 2 DM; uncontrolled hypertension; organ transplant history; chronic heart failure (NYHA class III or IV); hepatitis virus or HIV; active malignancy; pregnancy Baseline characteristics Number: intervention group (51); control group (51) Mean age ± SD (years): intervention group (64 ± 10); control group (65 ± 10) Sex (male, %): intervention group (34/51, 67%); control group (36/51, 71%) CKD stage/kidney function (median, IQR): intervention group (46, 32 to 64 mL/min/1.73 m2); control group (40, 26 to 52 mL/min/1.73 m2)
Interventions	Intervention group Silymarin: 150 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Kidney failure: defined as the need for maintenance dialysis for 12 weeks or more or kidney transplantation eGFR at end of study Change in eGFR (SD for change imputed) Proteinuria at end of study Change in proteinuria (SD for change imputed)
Notes	Additional information Funding sources: European Social Fund, Human Resources Development Operational Program 2007–2013 (grant no POSDRU/159/1.5/S/135760) and the University of Medicine and Pharmacy, Iasi (grant no. 1643/01.02.2013)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with block randomisation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	High risk	Single blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	No blinding (single blind) but all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Wlodarczyk 2000.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 7 days
Participants	Study characteristics Country: Poland Setting: single centre Inclusion criteria: transplant recipient Exclusion criteria: not reported Baseline characteristics Number: intervention group (21); control group (21) Mean age ± SEM (years): intervention group (45 ± 11); control group (41 ± 9) Sex (male, %): intervention group (9/21, 43%); control group (12/21, 57%) CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Pentoxifylline: 2 x 150 mg and 1 x 300 mg before transplant; 400 mg 3 times/day 7 days Control group Placebo
Outcomes	Outcomes relevant to this review DGF Infection
Notes	Additional information Funding sources: National Committee for Scientific Research (grant no. 4 PO 5B 07214)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding of patients and personnel not specified
Blinding of outcome assessors (detection bias)	Low risk	Blinding of outcome assessors not specified but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	Extracted outcome not specified in method
Other bias	Low risk	Study appears free of other biases

XANTHIN 2008.

Study characteristics
Methods	Study design Parallel RCT Time frame Recruitment: 2007 to 2013 Follow‐up: 12 months
Participants	Study characteristics Country: Australia Setting: multicentre (2 sites) Inclusion criteria: 18 to 85 years; transplant recipient Exclusion criteria: taking antioxidants; unable to take glyceryl trinitrate; participating in study Baseline characteristics Number: intervention group (33); control group (28) Mean age ± SD (years): intervention group (49 ± 11); control group (51 ± 13) Sex (male, %): intervention group (25/33, 76%); control group (20/28, 71%) CKD stage/kidney function: 5‐transplant Intervention group (51 ± 16 mL/min/1.73 m2); control group (52 ± 16 mL/min/1.73 m2)
Interventions	Intervention group Astaxanthin: 4 mg 3 times/day Control group Placebo
Outcomes	Outcomes relevant to this review CVD eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) Graft failure
Notes	Additional information Funding sources: Cyanotech Corporation (provided astaxanthin)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with computer‐generated random numbers by independent pharmacy‐based randomisation office
Allocation concealment (selection bias)	Low risk	Allocation was concealed by randomisation by an independent pharmacy‐based randomisation office and identical drug containers
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo and identical drug containers)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo and identical drug containers) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Low risk	All relevant outcomes reported in protocol
Other bias	High risk	One of the authors used to be director of company selling astaxanthin products

Xu 2017.

Study characteristics
Methods	Study design Parallel RCT Time frame 2012 to 2015 Follow‐up: 6 months
Participants	Study characteristics Country: China Setting: single centre Inclusion criteria: 18 to 90 years; eGFR 30 to 90 mL/min Exclusion criteria: inflammatory disease Baseline characteristics Number (randomised/analysed): intervention group (103/91); control group (94/86) Mean age ± SD (years): intervention group (59 ± 9); control group (58 ± 8) Sex (male, %): intervention group (66/91, 73%); control group (61/86, 71%) CKD stage/kidney function: CKD 2‐4 Intervention group (59 ± 9 mL/min/1.73 m2); control group (58 ± 8 mL/min/1.73 m2)
Interventions	Intervention group Ginsenoside Rb1: 500 mg Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR (SD for change imputed) SCr at end of study Change in SCr (SD for change imputed) Infection
Notes	Additional information Funding sources: Developing Medicine Funds
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with a computer‐generated block randomisation scheme
Allocation concealment (selection bias)	High risk	The assistant allocated sequential numbers to individual patients
Blinding of participants and personnel (performance bias)	Unclear risk	Blinding with concealed labels. It is unclear if personnel was blinded
Blinding of outcome assessors (detection bias)	Low risk	Unclear if outcome assessors were blinded but only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up specified, general reasons for loss to up specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

Yan 2017a.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame 1 January 2014 to 31 May 2015 Mean follow‐up: 17.3 weeks
Participants	Study characteristics Country: China Setting: single centre Inclusion criteria: > 18 years; eGFR 15 to 60 mL/min/1.73 m2; low serum T3, TSH max 1.2 fold above upper reference limit Exclusion criteria: history of KRT; thyroid or pituitary disease; hypertension or diabetes Baseline characteristics Number: intervention group 1 (42); intervention group 2 (42); control group (42) Mean age ± SD (years): intervention group 1 (54 ± 5); intervention group 2 (53 ± 6); control group (53 ± 9) Sex (male, %): intervention group 1 (23/42, 55%); intervention group 2 (25, 60%); control group (24/42, 57%) CKD stage/kidney function: intervention group 1 (19 ± 3 mL/min/1.73 m2); intervention group 2 (19 ± 2 mL/min/1.73 m2); control group (18 ± 2 mL/min/1.73 m2)
Interventions	Intervention group 1 N‐acetylcysteine: 600 mg/day Intervention group 2* Sodium bicarbonate: dose determined by serum HCO3‐ Control group Usual care *Data for patients receiving sodium bicarbonate not extracted or included in the meta‐analysis
Outcomes	Outcomes relevant to this review Death (any cause) Kidney failure: defined as eGFR < 15 mL/min/1.73 m2 Change in eGFR
Notes	Additional information Funding sources: National Natural Science Foundation of China (grant no. 81360122/H0518), and the Natural Science Foundation of Jiangxi Province (grant no. 20143ACB21010)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation was performed by an independent biostatistical
Blinding of participants and personnel (performance bias)	High risk	Singe blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	No blinding (single blind) but all objective outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported
Other bias	Low risk	Study appears free of other biases

Yang 2019a.

Study characteristics
Methods	Study design Parallel, multi‐arm RCT Time frame April to October 2011 Follow‐up: 12 weeks
Participants	Study characteristics Country: Taiwan Setting: single centre Inclusion criteria: chronic HD with failing AV fistula or grafts Exclusion criteria: hospitalisation for infection; heart failure; acute coronary syndrome in the past 3 months Baseline characteristics Number (randomised/analysed): intervention group 1 (31/30); intervention group 2 (31/29); control group (31/30) Mean age ± SD (years): intervention group 1 (65 ± 13); intervention group 2 (64 ± 12); control group (67 ± 13) Sex (male, %): intervention group 1 (14/30, 47%); intervention group 2 (17/29, 59%); control group (11/30, 37%) CKD stage/kidney function: 5D
Interventions	Intervention group 1 Vitamin C: 600 mg 3 times/week Intervention group 2 Vitamin C: 300 mg 3 times/week Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) CVD Thrombosis of vascular access
Notes	Additional information Funding sources: National Taiwan University Hospital, Hsinchu Branch (grant no. 99G002, HCH106‐01), the Ministry of Science and Technology (grant no. MOST‐106‐2314‐B‐002‐173‐MY3), the Foundation for Poison Control, and in part by the Novel Bioengineering and Technological Approaches to Solve Two Major Health Problems in Taiwan sponsored by the Taiwan Ministry of Science and Technology Academic Excellence Program (grant no. MOST 107‐2633‐B‐009‐003)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomised with computer‐generated numbers
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low loss to follow‐up and reasons specified per arm
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but not in protocol on ClinicalTrials.gov
Other bias	Low risk	Study appears free of other biases

Zachara 2009.

Study characteristics
Methods	Study design Parallel RCT Time frame Follow‐up: 3 months
Participants	Study characteristics Country: Poland Setting: not reported Inclusion criteria: chronic HD Exclusion criteria: not reported Baseline characteristics Number: intervention group (30); control group (28) Mean age ± SD: 61 ± 12 years Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Selenium: 200 µg/day Control group Placebo
Outcomes	Outcomes relevant to this review Change in SCr
Notes	Additional information Funding sources: State Committee for Scientific Research (KBN), Warsaw, Poland (grant No. 2 P05D 097 27). Bronislaw A Zachara has received an individual grant from the Foundation for Polish Science. Pharma Nord, Denmark, provided selenium‐enriched yeast and placebo
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not specified
Blinding of participants and personnel (performance bias)	Low risk	Double blind (matching placebo)
Blinding of outcome assessors (detection bias)	Low risk	Double blind (matching placebo) and only objective outcomes included
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Loss to follow‐up not specified
Selective reporting (reporting bias)	Unclear risk	All relevant outcomes in methods reported but no protocol available
Other bias	Low risk	Study appears free of other biases

ACEi: angiotensin‐converting enzyme inhibitors; AKI: acute kidney injury; ARB: angiotensin receptor blockers; AV: arteriovenous; BMI: body mass index; BP: blood pressure; CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; CrCl: creatinine clearance; CRP: C‐reactive protein; CVD: cardiovascular disease; DGF: delayed graft function; DM: diabetes mellitus; ECG: electrocardiograph; eGFR: estimated glomerular filtration rate; EPO: erythropoietin; ESA: erythropoietin stimulating agent; GFR: glomerular filtration rate; GI: gastrointestinal; Hb: haemoglobin; HD: haemodialysis; HDL: high‐density lipoprotein; HIV: human immunodeficiency virus; IQR: interquartile range; IV: intravenous; MAP: mean arterial pressure; MCV: mean corpuscular volume; MI: myocardial infarction; NSAIDs: nonsteroidal anti‐inflammatory drugs; NYHA: New York Heart Association; PTH: parathyroid hormone; RAAS: renin‐angiotensin‐aldosterone system; RCT: randomised controlled trial; SCr: serum creatinine; SD: standard deviation; TIA: transient ischaemic attack; TSAT: transferrin saturation; UACR: urinary albumin/creatinine ratio; UTI: urinary tract infection; WCC: white cell count

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abdollahzad 2009	No outcomes of interest
Agarwal 2004a	No outcomes of interest
Aghadavod 2018	No outcomes of interest
Ahmadi 2013	No outcomes of interest
Ali 2003	No outcomes of interest
Ali 2021	No outcomes of interest
Alvarenga 2020	No outcomes of interest
Alvarenga 2022	No outcomes of interest
Aminorroaya 2005	No outcomes of interest
Badri 2013	Wrong population
Bahmani 2016a	Wrong population
Bakris 2012	No outcomes of interest
Baldi 2013	No outcomes of interest
Biniaz 2013	No outcomes of interest
Biniaz 2015	No outcomes of interest
Bursell 1999	Wrong population
Candan 1996	No outcomes of interest
Candan 2002	No outcomes of interest
Chang 2007a	No outcomes of interest
Chao 2002	No outcomes of interest
Coloma 2011	No outcomes of interest
Conner 2008a	No outcomes of interest
Coombes 2000	No outcomes of interest
Daud 2013	No outcomes of interest
El Mashad 2016	Wrong population
El‐Shazly 2015	Wrong population
Elham 2019a	No outcomes of interest
Fallah 2018	No outcomes of interest
Fallahzadeh 2012	Wrong population
Farvid 2005	Wrong population
Fukuda 2015	No outcomes of interest
Fumeron 2005	No outcomes of interest
Gaede 2001	Wrong population
Garneata 2015	No outcomes of interest
Garrote 2009	Unable to obtain stratified outcomes for study arms
Gholnari 2018	Wrong population
Ghorbani 2012	Wrong population
Giancaspro 2000	No outcomes of interest
Golmohammadi 2022	Wrong population
Guerrero‐Romero 1995a	Wrong population
Han 2015	Wrong population
Harmankaya 2003	Wrong population
Heidari 2018	No outcomes of interest
Hodkova 2006	No outcomes of interest
Holdaas 1994	No outcomes of interest
Hosseini 2022	No outcomes of interest
Imada 2001	No outcomes of interest
Irijanto 2007	No outcomes of interest
Jadhav 2014	Wrong population
Jamal 2022	No outcomes of interest
Javaherforooshzadeh 2021	Wrong population
Kadhim 2006	Wrong population
Kalani 2005	No outcomes of interest
Kamgar 2009	No outcomes of interest
Keven 2003a	No outcomes of interest
Khajehdehi 2000a	No outcomes of interest
Khajehdehi 2001	No outcomes of interest
Khajehdehi 2011	Wrong population
Khajehdehi 2012	Wrong population
Khan 2013	Wrong population
Khatami 2016	Wrong population
Klein 1995	Wrong population
Kobayashi 2015	No outcomes of interest
Leyva‐Jimenez 2009	Wrong population
Lin 2016	No outcomes of interest
MacCallum 2009	No outcomes of interest
Mahajan 1979	No outcomes of interest
Mahmood 2018	Wrong population
Martins 2021	No outcomes of interest
Matson 2003	No outcomes of interest
Mazani 2013	No outcomes of interest
McAuliffe 1998	Wrong population
Modi 2001	Unable to retrieve full‐text
Moffitt 2013	No outcomes of interest
Mooraki 2006	Wrong population
Mooraki 2007	No outcomes of interest
Mortazavi 2012	No outcomes of interest
Muchova 2014	No outcomes of interest
Mune 1999	No outcomes of interest
Munguia 2003	Unable to retrieve full text
Murillo Ortiz 2019	No outcomes of interest
Najafabadi 2012	No outcomes of interest
Napolitano 1996	No outcomes of interest
Nascimento 2010	No outcomes of interest
Navarro 2003	Wrong population
Navarro 2005	Wrong population
NCT00440869	No outcomes of interest
Nowak 2020	Wrong population
Oliaei 2011	Wrong population
Omrani 2015	No outcomes of interest
Pakfetrat 2013	No outcomes of interest
Panah 2019	No outcomes of interest
Pantoja 2003	Wrong population
Parham 2008	Wrong population
Peng 2010	Wrong intervention
Perez 2004	No outcomes of interest
Pirhadi‐Tavandashti 2020	No outcomes of interest
Ponce de Leon‐Vargas 2013	Wrong population
Poulia 2011	No outcomes of interest
Pratama 2020	No outcomes of interest
PREVENT 2016	Wrong population
Purwanto 2012	No outcomes of interest
Ramos 2011	No outcomes of interest
Rashidi 2009	No outcomes of interest
Renke 2008a	No outcomes of interest
Rodger 1989	No outcomes of interest
Rodhe 2013	Wrong intervention
Rodrigues 2021	No outcomes of interest
Rodriguez‐Moran 2006	Wrong population
Roob 2000	No outcomes of interest
Roozbeh 2010	Wrong population
Roozbeh 2011	No outcomes of interest
Rymarz 2009	No outcomes of interest
Safa 2014	No outcomes of interest
Sagheb 2012	No outcomes of interest
Saldanha 2016	No outcomes of interest
Salimian 2022	No outcomes of interest
Samadian 2017	No outcomes of interest
Sanchez Alvarez 2005	No outcomes of interest
Sattarinezhad 2019	Wrong population
Schneeberger 1989	No outcomes of interest
Scholze 2004	No outcomes of interest
Sedaghattalab 2021	No outcomes of interest
Sedaghattalab 2021a	No outcomes of interest
Shahbazian 2016	No outcomes of interest
Shahidi 2015	Wrong population
Shahreki 2022	No outcomes of interest
Sinclair 1997	No outcomes of interest
Sprenger 1983	Unable to retrieve full text
Sun 2017b	Wrong population
Svensson 2004	Intervention not of interest
Tan 2018	Wrong population
Temple 2000	No outcomes of interest
Turk 1998	No outcomes of interest
Turk 2013	No outcomes of interest
Uzum 2006	No outcomes of interest
Vanaie 2019	Wrong population
VIPER 2004	No outcomes of interest
Wijnen 2002	Wrong population
Williams 2001a	No outcomes of interest
Wittstock 2009	No outcomes of interest
Yavari 2017	No outcomes of interest
Yeksan 1992	No outcomes of interest
Yokoyama 2001	Unable to retrieve full text
Zachara 2011	No outcomes of interest
Zahed 2016	Unable to obtain stratified outcomes for study arms
Zhang 2013	Unable to obtain stratified outcomes for study arms

Characteristics of ongoing studies [ordered by study ID]

CARDINAL 2021.

Study name	CARDINAL
Methods	Study design Parallel RCT Time frame Follow‐up: 104 weeks
Participants	Study characteristics Country: USA Setting: multicentre Inclusion criteria: 12 to 70 years; history of Alport syndrome; eGFR 30 to 90 mL/min; UACR < 3500 mg/g; maximum tolerated dose of ACEi or ARB Exclusion criteria: clinically significant CVD; uncontrolled diabetes or hypertension; B‐type natriuretic peptide > 24 pmol/L Baseline characteristics (for those with eGFR < 60 mL/min) Number: 66 Mean age ± SD: 44.5 ± 14.9 years Sex (male, %): 29/66, 44% CKD stage/kidney function (eGFR): CKD 3 (44.5 ± 14.9)
Interventions	Intervention group 1 Bardoxolone methyl: 30 mg Intervention group 2 Bardoxolone methyl: 20 mg Control group Placebo
Outcomes	Outcomes relevant to this review Kidney failure: defined as initiation of maintenance dialysis or kidney transplant eGFR at end of study change in eGFR
Starting date	August 2017
Contact information	Glenn M. Chertow: gchertow@stanford.edu
Notes

IRCT20100102002954N.

Study name	Evaluation of the effect of vitamin C on some blood indicators of dialysis patients
Methods	Study design Parallel RCT Time frame Follow‐up 12 weeks
Participants	Study characteristics Country: Iran Setting: unclear Inclusion criteria: 15 to 65 years; HD > 6 months; stable vital signs; controlled and consistent diet throughout study Exclusion criteria: severe digestive problems, (relatively) severe coagulation problems; history of chronic heart and liver disease; use of vitamin C in past month; change in diet without notice Baseline characteristics Sample size: 106 (estimated) Mean age ± SD (years): not reported Sex (male, 5): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin C: 250 mg/day Control group Usual care (dialysis, medication, diet)
Outcomes	No outcomes of interest specified in protocol
Starting date	June 2022
Contact information	Dr. Hoofar Rafiei: rafiei@shmu.ac.ir Dr. Mehdi Feyzabadi: mehdifyzabadi082@gmail.com
Notes

IRCT20110123005670N.

Study name	Evaluation of the effect of curcumin in patients after kidney transplantation
Methods	Study design Parallel RCT Time frame Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: unclear Inclusion criteria: 18 to 90 years; recipient of cadaveric kidney; no history of inflammatory bowel disease, liver cirrhosis, or malignancy; do not take antiepileptic drugs, curcumin, turmeric, omega 3, vitamin C or E supplements one month before start study Exclusion criteria: cognitive disorders Baseline characteristics Sample size: 40 patients (estimated) Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5‐transplant
Interventions	Intervention group Curcumin: 500 mg/day Piperine: 5 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR Graft loss DGF
Starting date	2022
Contact information	Dr. Ali Tarighat Esfanjani: tarighata@tbzmed.ac.ir
Notes

IRCT20150706023084N.

Study name	Trial of vitamin E on pruritus, appetite and muscle cramps in hemodialysis patients
Methods	Study design Parallel RCT Time frame Follow‐up: 12 weeks
Participants	Study characteristics Country: Iran Setting: unclear Inclusion criteria: 18 to 90 years; HD > 3 months Exclusion criteria: history of GI, dermatologic or liver disease; use of vitamin E or allergy to vitamin E Baseline characteristics Sample size: 40 patients Mean age ± (SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Vitamin E: 400 IE/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	October 2021
Contact information	Maryam Shiehmorteza" shiehmorteza@iaups.ac.ir
Notes

IRCT20160412027346N.

Study name	Evaluating the effect of zinc supplement on CRP and NLR inflammatory factor in MHD patients
Methods	Study design Parallel RCT Time frame Follow‐up: 1 month
Participants	Study characteristics Country: Iran Setting: unclear Inclusion criteria: > 18 years; dialysis twice/week for > 6 months; zinc < 70 µg/dL Exclusion criteria: hospitalisation in past 3 months; use of zinc or selenium supplementation in past 2 weeks Baseline characteristics Sample size: 40 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Zinc sulphate: 50 mg/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	August 2022
Contact information	Dr Shadi Ziaie: shadiziaie@sbmu.ac.ir
Notes

IRCT20220808055636N.

Study name	The effect of zinc sulfate on taste and dry mouth in dialysis patients
Methods	Study design Parallel RCT Time frame Follow‐up: 8 weeks
Participants	Study characteristics Country: Iran Setting: unclear Inclusion criteria: 18 to 70 years; dialysis > 6 months; no history of active infection, cancer, or heart disease Exclusion criteria: not reported Baseline characteristics Sample size: 50 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Zinc: 100 mg/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	September 2022
Contact information	Sakine Hashemi: akihashemi@yahoo.com
Notes

MPAC‐CKD‐1 2018.

Study name	MPAC‐CKD‐1 2018
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: Canada Setting: multicentre Inclusion criteria: > 18 years; eGFR 15 to 60 mL/min/1.73 m²; 24‐hour protein > 300 mg or UACR >30 mg/mmol; stable dose of ACEi or ARB Exclusion criteria: life expectancy < 1 year; KRT in the past 3 months; planned kidney transplant in study period; active peptic ulcer disease; hepatobiliary disease; AKI in the past 30 days; significant bleeding in the past 6 months; ongoing use of drugs that may interact with curcumin; allergy to turmeric or its derivatives or components of the investigational products Baseline characteristics Sample size: 500 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: CKD 3‐4
Interventions	Intervention group Microcurcumin: 90 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review eGFR at end of study Change in eGFR UACR at end of study Change in UACR
Starting date	October 2015
Contact information	Matthew Weir matthew.weir@lhsc.on.ca
Notes

NCT05183737.

Study name	Effects of supplementation with microcapsules of tumeric and propolis on inflammatory markers in patients on hemodialysis
Methods	Study design Parallel RCT Time frame Follow‐up: 12 weeks
Participants	Study characteristics Country: Brazil Setting: single centre Inclusion criteria: HD > 6 months; AV fistula as vascular access Exclusion criteria: auto‐immune disease; infectious disease; cancer; liver disease; AIDS; smokers; use of antibiotics in the past 3 months or antioxidants; smokers; habitual intake of propolis, curcumin and turmeric Baseline characteristics Sample size: 34 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Microcapsule with turmeric (0.50 mg) + green propolis (0.50 mg)/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	March 2022
Contact information	Denise Mafra
Notes

NCT05350124.

Study name	NCT05350124
Methods	Study design Factorial RCT Time frame Follow‐up: 12 weeks
Participants	Study characteristics Country: Saudi Arabia Setting: unclear Inclusion criteria: HD; history of restless leg syndrome Exclusion criteria: secondary causes of restless leg syndrome (medication, iron deficiency, peripheral neuropathy); use of medications that interfere with vitamin C and E absorption or contra‐indications for vitamin C or E Baseline characteristics Sample size: 160 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group 1 Vitamin C: 200 mg/day Vitamin E: 400 IU/day Intervention group 2 Vitamin C: 200 mg/day Intervention group 3 Vitamin E: 400 IU/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	November 2022
Contact information	Siraj Wali sowali@kau.edu.sa
Notes

NCT05422534.

Study name	NCT05422534
Methods	Study design Factorial RCT Time frame Follow‐up: 12 weeks
Participants	Study characteristics Country: USA Setting: single centre Inclusion criteria: 18 to 75 years; HD 3 times/week for > 6 months; clinically stable with single‐pool Kt/V > 1.2 in past 3 months Exclusion criteria: BMI > 35 kg/m²; history of function transplant in past 6 months; active acute connective tissue disease; acute infectious disease (past month); AIDS; acute MI or cerebrovascular event (3 months); uncontrolled BP; new or worsening mitral regurgitation murmur; hypotension; bradycardia or tachycardia' advanced liver disease; GI dysfunction requiring parental nutrition; active malignancy (excluding basal cell carcinoma); ejection fraction < 30%; anticipated live donor kidney transplant; inability to perform exercise Baseline characteristics Sample size: 156 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group 1 Coenzyme Q10: 1800 mg/day Exercise Intervention group 2 Coenzyme Q10: 1800 mg/day Intervention group 3 Exercise Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	January 2023
Contact information	Delia M Woods delia.woods@vumc.org Patricia Wright patricia.wright@vumc.org
Notes

NCT05564676.

Study name	NCT05564676
Methods	Study design Parallel RCT Time frame Follow‐up: 8 weeks
Participants	Study characteristics Country: Brazil Setting: single centre Inclusion criteria: > 19 years; HD 3 times/week for > 3 months Exclusion criteria: infection or hospitalisation at the beginning or during the study; HIV seropositivity or active malignancy; allergy to flaxseed or pomegranate Baseline characteristics Sample size: 50 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Treatment group Flaxseed oil: 2 g/day Pomegranate dry extract: 1.2 g/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	May 2021
Contact information	Aline Miroski de Abreu, Santa Catarina Federal University
Notes

Reed 2009.

Study name	Time course and dose response of alpha tocopherol on oxidative stress in haemodialysis patients
Methods	Study design Parallel RCT Time frame Follow‐up: 20 weeks
Participants	Study characteristics Country: Australia Setting: single centre Inclusion criteria: 18 to 85 years; hospital‐based HD 3 times/week; elevated oxidative stress Exclusion criteria: use of alpha tocopherol in the past 3 months; use of warfarin, desimpramine, chlorpromazine, and chloroquine Baseline characteristics Sample size: 20 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Alpha‐tocopherol: 1600 IU/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	Not reported
Contact information
Notes

VA PTXr 2021.

Study name	Pentoxifylline in diabetic kidney disease (VA PTXRx): protocol for a pragmatic randomised controlled trial
Methods	Study design Parallel RCT Time frame Follow‐up: 2 to 8 years
Participants	Study characteristics Country: USA Setting: multicentre Inclusion criteria: type 2 diabetes; CKD stage 3 or 4 Exclusion criteria: type 1 diabetes; history of non‐diabetic kidney disease; life expectancy < 1 year; history of organ or bone marrow transplant; dialysis; cerebral haemorrhage in past 3 months; current use of pentoxifylline; hypersensitivity to pentoxifylline current use of ketorolac or riociguat Baseline characteristics Sample size: 2510 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: CKD 3‐4
Interventions	Intervention group Pentoxifylline extended release: 800 mg/day Control group Placebo
Outcomes	Outcomes relevant to this review Death (any cause) Cardiovascular death CVD: defined as cardiovascular death, non‐fatal MI, non‐fatal cerebrovascular event Heart failure (hospitalisation) Peripheral arterial disease Kidney failure: defined as chronic dialysis or kidney transplantation Change in eGFR Change in UACR
Starting date	November 2019
Contact information	Dr David J Leehey: david.leehey@va.gov
Notes

Warnock 2012.

Study name	Prospective safety study of bardoxolone methyl in patients with type 2 diabetes mellitus, end‐stage renal disease and peritoneal dialysis
Methods	Study design Parallel RCT Time frame Follow‐up: 6 months
Participants	Study characteristics Country: USA and Canada Setting: multicentre Inclusion criteria: adult PD patients; mean systolic BP 90 to 160 mm Hg and mean diastolic BP 40 to 100 mm Hg Exclusion criteria: autosomal dominant polycystic kidney disease; amyloidosis or light chain nephropathy; active systematic lupus erythematosus; history of CVD; infection; malignancy < 5 years; use of systemic immunosuppressive agents Sample size: 45 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Bardoxolone methyl: 20 mg Control group Placebo
Outcomes	Outcomes relevant to this review Change in CrCl
Starting date	Not reported
Contact information	David G Warnock: dwarnock@uab.edu
Notes

Xu 2020a.

Study name	Effects of coenzyme Q10 on endothelial and cardiac function in patients undergoing haemodialysis: study protocol for a pilot randomised controlled trial
Methods	Study design Parallel RCT Time frame Follow‐up 12 months
Participants	Study characteristics Country: China Setting: single centre Inclusion criteria: 18 to 35 years; HD 3 times/week for > 3 months; life expectancy >1 year Exclusion criteria: severe systemic or local infection; malignancy; planned kidney transplant within 12 months; hospitalisation in the past 30 days; major atherosclerotic event in the past 3 months; use of antioxidants (excluding vitamin C); use of HD catheter Baseline characteristics Sample size: 60 patients Mean age ± SD (years): not reported Sex (male, %): not reported CKD stage/kidney function: 5D
Interventions	Intervention group Coenzyme Q10: 400 mg/day Control group Placebo
Outcomes	No outcomes of interest specified in protocol
Starting date	Not reported
Contact information	Dr Jianjun Gao: gao306kidney@126.com
Notes

ACEi: angiotensin‐converting enzyme inhibitor; AIDS: acquired immune deficiency syndrome; AKI: acute kidney injury; ARB: angiotensin receptor blocker; AV: arteriovenous; BMI: body mass index; BP: blood pressure; CrCl: creatinine clearance; CVD: cardiovascular disease; DGF: delayed graft function; eGFR: estimated glomerular filtration rate; GI: gastrointestinal; HD: haemodialysis; HIV: human immunodeficiency virus; KRT: kidney replacement therapy; MI: myocardial infarction; PD: peritoneal dialysis; RCT: randomised controlled trial; UACR: urinary albumin/creatinine ratio

Differences between protocol and review

We excluded one study which was included in the 2012 review (Wijnen 2002) because the study population did not consist of patients with CKD. Other differences between the 2012 review and this update are that we used the GRADE methodology to rate the quality of evidence and expanded the number of outcomes with heart failure, albuminuria, proteinuria, graft loss, graft failure, and delayed graft function. We split up eGFR, SCr, and UACR in absolute value at the end of the study and change during the study period. We changed haematological events to major bleeding to avoid confusion about whether haematological malignancies were included in the definition and excluded gastrointestinal and self‐reported adverse events. We expanded the number of antioxidant interventions.

Contributions of authors

2023 update

1. Study selection: JC, RV

2. Extract data from studies: JC, RV

3. Enter data into RevMan: JC, RV

4. Carry out analysis: JC, RV

5. Interpret analysis: all authors

6. draft the final review: all authors

2012 review

1. Draft the protocol: VV, BC, MJ, SZ, AW, VP

2. Study selection: MJ, VV

3. Extract data from studies: MJ, VV, BC

4. Enter data into RevMan: MJ, VV, BC

5. Carry out the analysis: MJ, VV

6. Interpret the analysis: MJ, VV, MR, BC, TN, SZ, AW, VP

7. Draft the final review: MJ, VV, VP

8. Disagreement resolution: VP

9. Update the review: MJ, VV, VP

Sources of support

Internal sources

• No sources of support provided

External sources

• No sources of support provided

Declarations of interest

• Julia MT Colombijn: no relevant interests were disclosed

• Lotty Hooft: no relevant interests were disclosed

• MinJun: no relevant interests were disclosed

• Angela C Webster: no relevant interests were disclosed

• Michiel L Bots: no relevant interests were disclosed

• Marianne C Verhaar: no relevant interests were disclosed

• Robin Vernooij: no relevant interests were disclosed

New search for studies and content updated (conclusions changed)