Neuromuscular blocking agents for acute respiratory distress syndrome

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the effectiveness and safety of neuromuscular blocking agents for mechanically‐ventilated individuals with acute respiratory distress syndrome.

Background

Description of the condition

Acute respiratory distress syndrome (ARDS) is an inflammatory process of the lung parenchyma that impairs gas exchange and potentially leads to multi‐organ dysfunction (Anderson 2003). ARDS was first reported in 1967 (Ashbaugh 1967). The incidence of ARDS ranges from 6.3 to 78.9 per 100,000 person‐years (Caser 2014; Hernu 2013; Rubenfeld 2005; Villar 2011). A population‐based study in Olmsted County, Minnesota, USA suggests that there has been a decline in ARDS from 82.4 to 38.9 per 100,000 person‐years between 2001 and 2008 (Li 2011). However, an epidemiological study from 50 countries reported the prevalence of ARDS to be 10.4% of all intensive care admissions (Bellani 2016). There are numerous predisposing conditions associated with ARDS: direct lung injury (pneumonia, aspiration of gastric contents, lung contusion, toxic inhalation, and near‐drowning) and indirect lung injury (severe sepsis, blood transfusion, trauma, cardiopulmonary bypass, and pancreatitis) (Walkey 2012). Mortality of ARDS is high, ranging from 27% to 45% (ARDS Definition Task Force 2012). In addition, survivors require additional health resources, suffer long‐term functional limitations with cognitive and psychological deterioration, and have a lower quality of life (Fan 2014; Herridge 2011; Wang 2014). Only 50% of survivors return to work after 12 months (Kamdar 2018).

The definition of ARDS has changed over time. The most common criteria, from the American‐European Consensus Conference, are: 1) the presence of bilateral infiltrates on the chest radiograph, 2) without clinical evidence of elevated left atrial pressure or a pulmonary capillary wedge pressure (PCWP) of less than 18 mmHg, and 33) a PaO₂/FiO₂ ratio of 200 mmHg or less (Bernard 1994). These criteria have some limitations, such as a lack of explicit time of onset from the predisposing clinical insult, no specific guidelines on how the PaO₂/FiO₂ ratio is measured, poor inter‐rater reliability of the chest radiograph, and difficulties with excluding volume overload or congestive heart failure as the primary cause of respiratory failure (Ferguson 2012). Since pulmonary artery catheters are not routinely used in intensive care units (ICUs) in recent times, the lack of PCWP measurement is also a limitation of these criteria.

In 2012, an expert consensus panel proposed these new criteria, termed the 'Berlin definition' (ARDS Definition Task Force 2012):

• onset of 7 days or less from the predisposing clinical insult;

• bilateral opacities on chest radiograph or computed tomography scan not fully explained by effusion, atelectasis, or nodules;

• respiratory failure not fully explained by cardiogenic pulmonary oedema or volume overload;

• PaO₂/FiO₂ ratio of 300 mmHg or less with at least 5 cm H₂O positive end‐expiratory pressure (PEEP).

This definition further classifies ARDS into three categories based on the severity: mild (PaO₂/FiO₂ ratio of 201 mmHg to 300 mmHg under PEEP or continuous positive airway pressure (CPAP) of ≥ 5 cm H₂O), moderate (PaO₂/FiO₂ ratio of 101 mmHg to 200 mmHg under PEEP ≥ 5 cm H₂O), and severe (PaO₂/FiO₂ ratio of ≤ 100 mmHg under PEEP ≥ 5 cm H₂O) (ARDS Definition Task Force 2012).

Description of the intervention

There are two types of neuromuscular blocking agents (NMBAs): depolarising and non‐depolarising. We will focus on non‐depolarising NMBAs in this review, because these are the only NMBAs that have been assessed in studies of ARDS.

Non‐depolarising NMBAs are competitive acetylcholine antagonists that bind directly to postsynaptic nicotinic receptors in the neuromuscular junctions and result in paralysis of respiratory muscles. This reduces patient‐ventilator asynchrony (Jonsson Fagerlund 2009). Non‐depolarising NMBAs are classified into two categories based on their chemical structure: benzylisoquinolinium (d‐tubocurarine, gallamine, mivacurium, atracurium, cisatracurium) and steroidal (rocuronium, vecuronium, pancuronium).

How the intervention might work

There are several reasons for considering using NMBAs among individuals with ARDS. Their use minimises the risk of patient‐ventilator asynchrony, which is due to mismatch between the patient inspiratory demand and effort. Asynchrony is associated with ventilator‐associated lung injury such as barotrauma (lung injury due to application of high airway pressure with alveoli rupture), volutrauma (lung injury due to local distension of normal alveoli), atelectrauma (lung injury due to inadequate mechanical ventilation with alveolar collapse), and biotrauma (lung injury due to the inflammation of the lung by pro‐inflammatory cytokines and white cells that were triggered by unphysiological stress/strain) (Akoumianaki 2013; Bourenne 2019; Gattinoni 2010; Thille 2006). Patient‐ventilator asynchrony also prolongs the duration of mechanical ventilation and increases mortality (Beitler 2016; Blanch 2015; Thille 2006). NMBAs may also improve the mechanical viscoelastic properties of the chest wall. They also block contraction of respiratory muscles, decreasing oxygen consumption (Bishop 1984; Hunter 1995). In addition, NMBAs facilitate more homogenous regional inflation across the alveoli during ventilation, increase alveolar recruitment, and improve oxygenation (Slutsky 2010). Finally, cisatracurium, one type of NMBA, has a direct anti‐inflammatory effect (Fanelli 2016).

Why it is important to do this review

A 2013 systematic review, based on three randomised trials involving 431 individuals, suggested that NMBA use reduces mortality and improves oxygenation (Alhazzani 2013). However, the effect of NMBAs on mechanical ventilation and ICU length of stay were unclear. In addition, Alhazzani and colleagues found no evidence of ICU‐acquired weakness, a potential adverse effect of NMBAs that affects long‐term patient prognosis (Jolley 2016). However, early studies diagnosed ICU‐acquired weakness based on clinically detectable quadriparesis, which lacks sensitivity and specificity (Kress 2014). Subsequently, guidelines on the management of ARDS have weakly recommended the use of NMBAs (Claesson 2016; Griffiths 2019).  There has since been an additional large trial ‐ Moss 2019 ‐ that was not included in the 2013 systematic review. More recent systematic reviews that included Moss 2019 provided different conclusions regarding the efficacy of NMBAs in individuals with ARDS, depending on whether the reivew authors focused on clinical characteristics such as depth of sedation (Tarazan 2020; Torbic 2021). A recent rapid clinical practice guideline that included this randomized controlled trial provided recommendations regarding mortality, adverse events, and ICU‐acquired weakness, with low to moderate certainty (Alhazzani 2020). A complete, rigorous assessment of the impact of NMBAs may help with future guideline development and with updates of existing ones, highlighting the importance of this review.

We aim to conduct a systematic review on the effect of NMBAs for individuals with ARDS that incorporates several recently published randomised trials. The inclusion of these trials and a focus on the clinical characteristics of included studies may change the recommendations regarding the use of NMBAs for individuals with ARDS.

Objectives

To assess the effectiveness and safety of neuromuscular blocking agents for mechanically‐ventilated individuals with acute respiratory distress syndrome.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs) that assess the efficacy and safety of NMBAs in individuals with ARDS. We will exclude quasi‐randomised controlled trials, non‐randomised studies, cross‐over and observational studies.

Types of participants

We will include adults 18 years and older with ARDS who are receiving mechanical ventilation with sedation. We will impose no restriction on the setting of care. We will accept the definition of ARDS in each study, anticipating that the definitions of ARDS may include the 'Berlin' ARDS definition (PaO₂/FiO₂ ratio ≤ 300 mmHg at PEEP ≥ 5 cm H₂O) (ARDS Definition Task Force 2012), or the American‐European Consensus Conference criteria (Bernard 1994). We will exclude studies that do not use either of these criteria, if we can establish this from the study's methods.

Types of interventions

We will include studies that evaluate the use of non‐depolarising NMBAs as the main intervention in addition to conventional therapy. We will place no restrictions on the type or duration of non‐depolarising NMBAs used, or on the time, delivery, and duration of interventions with non‐depolarising NMBAs.

For the comparator, we will include studies that compare NMBAs to conventional therapy alone or conventional therapy plus placebo for individuals with ARDS. We will also include studies that compare NMBAs to any pharmacological or non‐pharmacological intervention.

We will define conventional therapy as mechanical ventilation with sedation, appropriate control of fluid therapy, and adequate enteral nutrition, as well as the treatment for underlying aetiology of ARDS wherever possible, including antimicrobial therapy.

While different NMBA drugs may not be expected to act in a meaningfully different way, there may be differences in the protocols used for the drugs (e.g. how or when clinicians choose to reverse paralysis during treatment, depth of paralysis targeted). We will extract information about the specific protocol for NMBA use from each included study.

Types of outcome measures

Primary outcomes

• 28‐day mortality (we will use mortality within 30 days, intensive care unit mortality, or hospital mortality, in that order, if 28‐day mortality is not reported)

• Any serious adverse events determined by each included study

• ICU‐acquired weakness (we will accept the definition used by each included study, but we will explore this outcome in a subgroup analysis if multiple definitions are identified)

Secondary outcomes

• 90‐day mortality

• Barotrauma (we will accept the definition used by each included study)

• Days free of mechanical ventilation at 28 days

• Days outside ICU at 28 days

• PaO₂/FiO₂ ratio at 24 hours and 48 hours

• Pulmonary morbidity as specified by the study authors

• Weaning success as specified by the study authors

Search methods for identification of studies

Electronic searches

We will search for studies in accordance with the guidance described in the Cochrane Handbook of Systematic Reviews of Interventions, Chapter 4 (Lefebvre 2019; hereafter referred to as the Cochrane Handbook). We will place no restrictions on language, publication year, or publication status.

We will search these databases, from their inception to the present, with the help of a medical librarian:

• Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library;

• PubMed or MEDLINE Ovid;

• Embase Ovid; and

• Web of Science.

We will use subject headings along with free‐text terms. The search strategy for MEDLINE can be found in Appendix 1. We will adapt it for the searches of the other databases for the full review. We will use the sensitivity maximising strategies described in the Cochrane Handbook to search for RCTs in MEDLINE and Embase (Lefebvre 2019). Where appropriate, we will use similar search strategies for identifying RCTs in other databases.

We will also search Google Scholar and conference proceedings of Critical Care Congress and International Symposium on Intensive Care and Emergency Medicine for grey literature that is not identified via the database searches.

We developed the search strategy in consultation with the Cochrane Emergency and Critical Care Information Specialist.

Searching other resources

We will scan the reference lists and citations of eligible studies and systematic reviews for further potentially relevant studies.

We will also search for unpublished and ongoing studies using these trial registries:

• ClinicalTrials.gov (clinicaltrials.gov)

• ISRCTN registry (http://isrctn.com)

• International Clinical Trials Registry Platform (ICTRP)

We will use free‐text terms to search in these databases, and will also screen the reference lists of potentially eligible ongoing studies.
We will check our included studies for retractions.

Data collection and analysis

Selection of studies

We will use reference manager software and will exclude duplicate publications. Two review authors (AK and JLJ) will independently and in duplicate assess the titles and abstracts of studies for eligibility. We will independently and in duplicate screen the full texts of studies deemed potentially eligible, and provide reasons for excluding studies. We will resolve any disagreements through consensus or discussion with a third person. If information in the study is insufficient to determine its eligibility, one review author (AK) will contact the corresponding author of the study.

Data extraction and management

Two review authors (AK and JLJ) will independently extract the data from each eligible study using a piloted data extraction form (Appendix 2). One review author will do the primary extraction, with results verified by the other. We will resolve disagreements through discussion. If additional data or other unpublished information missing from study reports are required or if clarification is needed, we will contact the corresponding author or sponsor of the study. We will extract these data from each included study:

• year and country of the study;

• study subject demographic information (age, gender, ethnicity);

• specific NMBA used, dose and duration of use;

• depth of paralysis target;

• illness severity;

• duration of follow‐up;

• mortality (28‐ and 90‐days);

• serious adverse events;

• ICU‐acquired weakness (including study definition);

• barotrauma (including definition);

• days off ventilation and not in ICU at 28 days;

• PaO₂/FiO ² ratio at 24 and 48 hours;

• weaning success.

Assessment of risk of bias in included studies

Two review authors (AK and JLJ) will independently assess all studies using the risk of bias tool described in the Cochrane Handbook (Higgins 2011). We will resolve any disagreements on the risk of bias assessment through discussion or by involving a third party. If there is insufficient information to assess the risk of bias, we will contact the corresponding author of the study for unpublished details. The form for this purpose is provided (see Appendix 3).

We will assess each study according to these domains:

• random sequence generation;

• allocation concealment;

• blinding of participants and personnel;

• blinding of outcome assessment;

• incomplete outcome data;

• selective reporting;

• other bias.

In assigning the potential impact of overall bias, we will use the following criteria.

Risk of Bias	Interpretation	Within a study	Across studies
Low risk of bias	Plausible bias unlikely to seriously alter the results	Low risk of bias for all key domains	Most information is from studies at low risk of bias
Unclear risk of bias	Plausible bias that raises some doubt about the results	Unclear risk of bias for one or more key domains	Most information is from studies at low or unclear risk of bias
High risk of bias	Plausible bias that seriously weakens confidence in the results	High risk of bias for one or more key domains	The proportion of information from studies at high risk of bias is sufficient to affect the interpretation of results

We will also consider additional risk of bias for cluster‐randomised trials; namely, recruitment bias, baseline imbalance, loss of clusters, incorrect analysis, and comparability.

We will generate a risk of bias graph and a risk of bias summary figure using Review Manager 5.4 (Review Manager 2020).

Measures of treatment effect

We will use risk ratios (RRs) with 95% confidence intervals (CIs) for dichotomous outcomes. For continuous outcomes, we will pool those with the same unit of measure using weighted mean differences (WMD). We will pool those with varying measurement scales using standardised mean differences (SMD).

Unit of analysis issues

Because we expect multiple intervention groups in a single study, we will consider the level at which randomisation occurred as the unit of analysis. We will pool the multiple groups into a single group when there are multiple intervention groups in a single study. We will include cluster‐randomised trials if the intracluster correlation coefficient is reported. If a study does not report the intracluster correlation coefficient, we will use external estimates obtained from similar studies to estimate the intracluster correlation coefficient.

Dealing with missing data

We will contact the corresponding author of a study to request missing data.

For dichotomous data, we will extract data on the assumption of an intention‐to‐treat analysis.

For continuous data, we will calculate missing standard deviations (SDs) based on the reported standard errors or P values (Follman 1992), if we cannot obtain such information from the author. For studies that report a median instead of a mean, and a range or interquartile range (IQR) instead of SDs for a continuous outcome, if we cannot obtain information from the study's authors, we will calculate the means and SDs using Wan's method (Wan 2014). We will make no assumptions about loss to follow‐up for continuous data, and pool the results for individuals who have complete results.

Assessment of heterogeneity

We will examine clinical heterogeneity based on the characteristics of the included studies. If there is clinical heterogeneity based on a certain clinical characteristic, we will examine its impact using subgroup analysis.

We will measure statistical heterogeneity using the I² statistic and follow Cochrane criteria for the significance of heterogeneity for randomised trials:

• 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.

If we identify significant statistical heterogeneity (P value < 0.1), we will investigate potential sources of heterogeneity through subgroup and sensitivity analysis as stated below. We will also check the accuracy of our data extraction and input.

Assessment of reporting biases

We will assess reporting bias if at least 10 studies are included in the meta‐analysis using funnel plot asymmetry visual inspection, as well as Egger's method for continuous outcomes (Egger 1997), and Peters's method for dichotomous ones (Peters 2008).

We will follow ORBIT (Outcome Reporting Bias In Trials) criteria to assess the risk for outcome reporting bias. The two review authors will independently apply the criteria to each study and reach consensus regarding the judgment. ORBIT‐1 assesses the risk of selective outcome reporting bias for benefit (Kirkham 2010), and ORBIT‐2 the risk of selective outcome reporting for harms (Kirkham 2018). Both classify studies as high, low or no risk of selective reporting bias. We will explore the potential impact of outcome reporting bias using Copas's method (Copas 2014).

We will also examine whether there is inconsistency in the direction of the effect.

Data synthesis

We will decide whether or not pooling is justified based on the number of studies (at least three) and the degree of heterogeneity (both statistical and qualitative study characteristics). We will describe narratively any outcome that cannot be pooled. For outcomes that can be pooled, we will use the mean difference (MD) for continuous outcomes, because these outcomes are typically measured and presented using an identical, standardised scale. However, if studies use different scales for continuous outcomes, we will pool the results using standardised mean differences (SMD). We will pool the data using a random‐effects model for all outcomes because this is the most conservative approach. Given the use of an inverse variance approach, studies with no significant heterogeneity statistically produce identical results with both fixed‐effect and random‐effects models.

ICU studies often have a substantial death rate. Our primary outcome is 28‐day mortality. If there is a significant impact on mortality, this will be the strongest evidence for benefit (or harm) for intervening with NMBAs. Mortality is potentially impacted by NMBAs if they decrease the risk of various types of trauma during ventilation. If there is no difference in mortality rates, then variables that could be impacted by mortality, such as time‐to‐event data, will not be different between the two groups. If possible, we will use survival meta‐analytic approaches to deal with censoring by death (Tierney 2007).

Subgroup analysis and investigation of heterogeneity

We will conduct the following subgroup analyses by interventions and participants.

The use of steroids is a risk factor for ICU‐acquired weakness. Further, deep sedation can be a confounding factor of mortality in mechanically‐ventilated individuals (Stephens 2018). Therefore, we will conduct the following subgroup analyses for the primary outcomes.

• Type of NMBAs: steroidal versus non‐steroidal NMBAs.

• Depth of sedation: deep or light sedation according to the scale of depth of sedation used in the original studies.

It is anticipated that a lower PaO₂/FiO₂ ratio at inclusion to a study or onset of ARDS may be associated with more frequent mortality outcomes and barotrauma as well as increased days free of mechanical ventilation and lower PaO₂/FiO₂ ratio at 24 hours and 48 hours. We will conduct the following subgroup analysis for the secondary outcomes.

• Severity of ARDS according to PaO₂/FiO₂ ratio at inclusion to a study or onset of ARDS.

• Presence or absence of titration of NMBA by the depth of paralysis assessed by the peripheral nerve stimulation.

Unfortunately, there may be important characteristics that cannot be teased out from aggregate data. For example, we would expect age to be an important predictor of mortality with ARDS, but if the studies all have participants of similar ages, the relationship between age and mortality described within each study may not be statistically appreciable. We will qualitatively present data that are provided within studies regarding such relationships. If there are data (such as correlation coefficients between age and mortality) that can be pooled presented in a sufficient number of articles, we will provide a pooled estimate of these relationships.

Sensitivity analysis

We will conduct the following sensitivity analyses on the primary outcomes to examine the robustness of our findings.

• Stratified analysis examining the impact of excluding studies that are at unclear or high risk of incomplete reporting bias.

• Stratified analysis to exclude cluster‐randomised trials.

• Stratified analysis to exclude definitions other than the 'Berlin' ARDS definition (ARDS Definition Task Force 2012), or the American‐European Consensus Conference criteria (Bernard 1994).

• We will assess the presence of publication bias by inspecting funnel plots for asymmetry. We will test for publication bias using using Egger's test for continuous outcomes (Egger 1997), and Peters's test for dichotomous outcomes (Peters 2008). We will explore the potential impact of publication bias using the meta‐trim approach (Duval 2000).

• We will assess the potential impact of outcome reporting bias using the ORBIT approach (Kirkham 2018), and analytic approaches suggested by Copas (Copas 2014).

• We will assess for the risk of Type 1 and 2 error using Trial Sequential Analysis (TSA) (Wetterslev 2017). We will calculate the required information size, which is the number of individuals in a meta‐analysis to confirm or refute the prespecified intervention effect. A monitoring boundary adjusts the P value that is required for obtaining a statistical significance according to the number of participants and events in a meta‐analysis. A cumulative Z‐curve is created by connecting the cumulative z‐values after addition of each new trial in a cumulative meta‐analysis. We will assess the risk of Type 1 and 2 errors based on the required information size, a monitoring boundary, and cumulative Z‐curve. We will use TSA software, version 0.9 beta (Copenhagen Trial Unit, Copenhagen, Denmark).

• We will perform the analyses using both fixed‐effect and random‐effects approaches.

Summary of findings and assessment of the certainty of the evidence

We will use GRADEpro to create a summary of findings table for the comparison of NMBA use in ventilated individuals.

We will use the GRADE approach, suggested by the GRADE working group (Schünemann 2013), to assessing the certainty of our findings for these outcomes (Guyatt 2008):

• 28‐day mortality;

• any serious adverse events;

• ICU‐acquired weakness;

• 90‐day mortality;

• barotrauma; and

• PaO₂/FiO₂ ratio at 24 hours and 48 hours.

We will asses the certainty across five GRADE domains:

• limitations in the design and implementation of available studies (i.e. unclear or high risk of bias of studies contributing to the respective outcome) (Guyatt 2011a);

• high probability of publication bias (i.e. high risk of selective outcome reporting bias for studies contributing to the outcome, based on funnel plot asymmetry, Egger's test, different results of published versus unpublished studies, and whether the evidence consisted of many small studies with potential conflicts of interest) (Guyatt 2011b);

• imprecision of results (i.e. small number of participants included in an outcome and wide CIs) (Guyatt 2011c);

• unexplained heterogeneity or inconsistency of results (i.e. heterogeneity based on variation of effect estimates, CIs, the statistical test of heterogeneity and I², but the subgroup analyses fail to identify a plausible explanation) (Guyatt 2011d); and

• indirectness of evidence (i.e. included studies limited to certain participants, intervention types, or comparators) (Guyatt 2011e).

Two review authors (AK, JLJ), working independently, will conduct the assessment of the certainty of the evidence in duplicate, resolving any disagreements by discussion or by consulting a third review author (Schünemann 2013).

We will rate the certainty of the evidence as high, moderate, low or very low based on GRADE criteria (Schünemann 2013). High certainty evidence indicates high confidence that the true effect lies close to that of the estimate of effect. Very low certainty evidence indicates that we have very little confidence in the effect estimate and that the true effect is likely to be substantially different from the estimate of effect.

Appendices

Appendix 1. MEDLINE search strategy

1 exp Neuromuscular Blocking Agents/
2 ((neuromuscular or neuro‐muscular) adj5 (non‐depolariz* or nondepolariz* or non depolariz* or block*)).mp.
3 (NMBA or NMBAs).mp.
4 (curariform* adj5 (drug* or agent*)).mp.
5 (muscle relaxant* and (non depolariz* or nondepolariz*)).mp.
6 (Alcuronium or Atracurium or Curare or dacuronium or doxacurium or Gallamine Triethiodide or metocurine or Mivacurium or Pancuronium or Pipecuronium or rapacuronium or Rocuronium or Toxiferine or Tubocurarine or Vecuronium Bromide).mp.
7 1 or 2 or 3 or 4 or 5 or 6
8 Respiratory Distress Syndrome, Adult/
9 exp Acute Lung Injury/
10 exp Respiratory Insufficiency/
11 exp Severe Acute Respiratory Syndrome/
12 (ARDS* or SARS).mp.
13 ((acute or adult* or syndrome*) and (respirat* adj1 distress)).mp.
14 (acute adj1 respirat* adj2 (syndrome* or distress or failure*)).mp.
15 ((Lung adj1 shock) or (acute adj1 lung adj1 injur*)).mp.
16 (respirat* adj2 (insufficien* or failure* or depression)).mp.
17 ((stiff or wet) adj1 lung*).mp.
18 exp coronavirus/
19 exp Coronavirus Infections/
20 (coronavirus* or corona virus* or Covid 19 or Covid19 or SARS CoV* or SARSCov* or ncov* or 19ncov*).mp.
21 18 or 19 or 20
22 21 and (201912* or 2020*).dt.
23 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 22
24 7 and 23

Appendix 2. Data Extraction Sheet

1. General Information

Name of the person extracting data
Report title (title of paper/ abstract/ report from which data are extracted)
Report ID (ID for this paper/ abstract/ report)
Report author contact details
Publication type
Study funding source (including role of funders)

2. Participant Characteristics

Participant characteristics	Description stated in report	Location in text
Age (mean, median, range)
Female participants (number/ %)
Severity of illness (SOFA/ APACHE II, etc)

3. Study Characteristics

Study characteristics	Description stated in report	Location in text
Country
Single centre/ Multicentre
Setting (including location and social context)
Population (from which study participants are drawn)
Number of participants randomised to each group
Number of withdrawals and dropouts after randomisation
Number of participants analysed in each group
Interventions
The name of drug used in the experimental group
The type of drug used in the control group
Intermittent or continuous administration?
Dose/ frequency of administration
Duration of treatment (days/ hours)
The timing when the paralysis initiated
Target depth of paralysis
Other treatment received (additional intervention)
Follow‐up period

4. Risk of bias assessment

Domain	Judgement	Support for judgement
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting of benefit outcomes based on ORBIT‐1 (reporting bias)
Selective reporting of harm outcomes on ORBIT‐2
Other bias

5. Outcomes

Primary Outcomes

1. 28‐day mortality (dichotomous outcomes)

Our definition	Mortality within 30 days, intensive care unit mortality, or hospital mortality will be used in that order, if 28‐day mortality is not reported.
Reported outcome (actually reported outcomes)	28‐day mortality, mortality within 30 days, intensive care unit mortality, or hospital mortality.
Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

2. Any serious adverse events determined by each included primary study (dichotomous outcomes)

Definition
Results	Interventions			Control
	No. of events		No. of participants	No. of events	No. of participants

3. ICU‐acquired weakness (dichotomous outcomes)

Definition
Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

Secondary Outcomes

1. 90‐day mortality (dichotomous outcomes)

Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

2. Barotrauma (dichotomous outcomes)

Definition
Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

3. Days free of mechanical ventilation at 28 days (continuous outcomes)

Results	Interventions			Control
	Mean/ Median	Standard deviations/ Interquartile range	No. of participants	Mean/ Median	Standard deviations/ Interquartile range	No. of participants

4. Days outside ICU at 28 days (continuous outcomes)

Results	Interventions			Control
	Mean/ Median	Standard deviations/ Interquartile range	No. of participants	Mean/ Median	Standard deviations/ Interquartile range	No. of participants

5. PaO₂/FiO₂ ratio at 24 hours and 48 hours (continuous outcomes)

Time point	24 hours
Results	Interventions			Control
	Mean/ Median	Standard deviations/ Interquartile range	No. of participants	Mean/ Median	Standard deviations/ Interquartile range	No. of participants

Time point	48 hours
Results	Interventions			Control
	Mean/ Median	Standard deviations/ Interquartile range	No. of participants	Mean/ Median	Standard deviations/ Interquartile range	No. of participants

6. Pulmonary morbidity (dichotomous outcomes)

Definition
Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

7. Weaning success (dichotomous outcomes)

Definition
Results	Interventions		Control
	No. of events	No. of participants	No. of events	No. of participants

Appendix 3. Quality Assessment Instrument

Study
Rater
Domain	Criteria	Review authors’ judgment
Sequence generation	Investigators describe a random component in the sequence generation such as: Referring to a random number table Using a computer random number generator Coin tossing Shuffling cards or envelopes Throwing dice Drawing of lots minimization	Was the allocation sequence adequately generated? YES / NO / UNCLEAR
Allocation concealment	Participants and investigators enrolling participants could not foresee assignment because one of the following or an equivalent method was used to conceal allocation: Central allocation (including telephone, web‐based and pharmacy‐controlled allocation) Sequential numbered drug containers of identical appearance Sequentially numbered, opaque, sealed envelopes.	Was allocation adequately concealed? YES / NO / UNCLEAR
Blinding of participants, personnel and outcome assessors Outcome:	Was knowledge of the interventions adequately prevented during the study: No blinding, but the review author’s judge that the outcome and outcome measurement are 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. Either the participants or some key study personnel were not blinded, but outcome assessment was blinded and the non‐blinding of others unlikely to introduce bias.	Was knowledge of the allocated intervention adequately prevented during the study? YES / NO / UNCLEAR
Incomplete outcome data Outcome:	Were incomplete data adequately addressed: No missing outcome data Reasons for missing outcome data unlikely to be related to true outcome. 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 among missing outcomes not enough to have a clinically relevant impact on observed effect size. Missing data have been imputed using appropriate methods.	Were incomplete outcome data adequately addressed? YES / NO / UNCLEAR
Selective outcome reporting	Are reports of the study free of suggestion of selective outcome reporting? 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).	Are reports of the study free of suggestion of selective outcome reporting? YES / NO / UNCLEAR
Other sources of bias	Was the study apparently free of other problems that could put it at risk of bias? (this is a wastebasket category for anything that we may identify that is unique to this group of studies that could potentially introduce bias). NOTE: Cochrane notes that it is likely that the majority of studies will fall in the unclear category for this quality item.	Was the study apparently free of other problems that could put it at a high risk of bias? YES / NO / UNCLEAR
Industry sponsorship		YES / NO / UNCLEAR

Contributions of authors

AK conceived the review topic and both authors contributed to the design and drafting of the protocol.

Sources of support

Internal sources

• New Source of support, Other

  None.

External sources

• No sources of support provided

Declarations of interest

Akira Kuriyama: None known

Jeffery L Jackson: None known

New