[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
PRISMA Flow Diagram
PRISMA Flow Diagram
Figure 2.
Network Geometry and Ranking Probabilities for the Association of Interventions With Outcomes
Network Geometry and Ranking Probabilities for the Association of Interventions With Outcomes

Network geometry shows the interventions as nodes and the direct comparisons as connections between the nodes. Larger node size indicates a larger total number of participants in arms including that intervention, and thicker connection indicates a larger number of trials investigating that comparison. HFOV indicates high-frequency oscillatory ventilation; INO, inhaled nitric oxide; LPV, lung protective ventilation; NMBA, neuromuscular blockade; RM, recruitment maneuver; and VV ECMO, venovenous extracorporeal membrane oxygenation.

aOpen lung strategy using RM and/or higher positive end-expiratory pressure.

Table 1.  
Summary of Identified 25 Studies
Summary of Identified 25 Studies
Table 2.  
Summary of Findings for 28-Day Mortality
Summary of Findings for 28-Day Mortality
Table 3.  
Summary of Findings for Barotrauma
Summary of Findings for Barotrauma
Supplement.

eAppendix 1. Summary of Inclusion and Exclusion Criteria

eAppendix 2. Updated Search Strategy in MEDLINE, Embase, Cochrane, PubMed, and CINAHL

eAppendix 3. Assessment of Heterogeneity, Consistency, and Intransitivity

eAppendix 4. Network Metaregression With Treatment by Covariate Interactions to Measure the Impact of Age and ARDS Severity

eFigure 1. Results of Network Metaregression for Age (a) and ARDS Severity (b)

eFigure 2. Forest Plots of Direct, Indirect, and Pooled Comparisons for Primary Outcome

eFigure 3. Forest Plots of Direct, Indirect, and Pooled Comparisons for Secondary Outcome

eFigure 4. Risk-of-Bias Graph for All Eligible Studies

eFigure 5. Risk-of-Bias Summary for All Eligible Studies

eFigure 6. Funnel Plots for Primary (a) and (b) Secondary Outcomes

eFigure 7. Gelman Plots for Model Convergence With 4 Chains and 2 000 000 Iterated Simulations, Discarding the Initial 1 500 000 Iterations as Burn-in on Primary (a) and (b) Secondary Outcomes

eFigure 8. Ranking Probabilities for the Association of Interventions With Outcomes

eTable 1. Detailed Explanation for Excluded Studies With Full-Text Assessment

eTable 2. Comparison of the Included Interventions in Risk Ratio (95% CrI) for 28-Day Mortality

eTable 3. Network Meta-analysis for 28-Day Mortality and Quality of Evidence Assessment

eTable 4. Results of Sensitivity Analysis for 28-Day Mortality

eTable 5. Results of Preliminary Analysis Excluding Studies Without Description of Cointerventions

eTable 6. Results of Analysis Using Poisson Models for the Primary Outcome Adjusting for Different Timing of the Measurement

eTable 7. Potential Scale Reduction Factor From Brooks-Gelman-Rubin Diagnostic for Model Convergence on Primary (a) and (b) Secondary Outcomes

eTable 8. Assessment of Model Fits for Primary (a) and (b) Secondary Outcomes

eTable 9. Results of Assessment on Small Sample Size Effects

eTable 10. Comparison of the Included Interventions in Risk Ratio (95% CrI) for Barotrauma

eTable 11. Network Meta-analysis for Barotrauma and Quality of Evidence Assessment

eTable 12. Results of Node-Splitting Models

eTable 13. Summary of Cointerventions in Included Studies

eReferences

1.
Bellani  G, Laffey  JG, Pham  T,  et al; LUNG SAFE Investigators; ESICM Trials Group.  Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.  JAMA. 2016;315(8):788-800. doi:10.1001/jama.2016.0291PubMedGoogle ScholarCrossref
2.
Ashbaugh  DG, Bigelow  DB, Petty  TL, Levine  BE.  Acute respiratory distress in adults.  Lancet. 1967;2(7511):319-323. doi:10.1016/S0140-6736(67)90168-7PubMedGoogle ScholarCrossref
3.
Fan  E, Brodie  D, Slutsky  AS.  Acute respiratory distress syndrome: advances in diagnosis and treatment.  JAMA. 2018;319(7):698-710. doi:10.1001/jama.2017.21907PubMedGoogle ScholarCrossref
4.
Thompson  BT, Chambers  RC, Liu  KD.  Acute respiratory distress syndrome.  N Engl J Med. 2017;377(6):562-572. doi:10.1056/NEJMra1608077PubMedGoogle ScholarCrossref
5.
Slutsky  AS, Ranieri  VM.  Ventilator-induced lung injury.  N Engl J Med. 2013;369(22):2126-2136. doi:10.1056/NEJMra1208707PubMedGoogle ScholarCrossref
6.
Fan  E, Del Sorbo  L, Goligher  EC,  et al; American Thoracic Society, European Society of Intensive Care Medicine, and Society of Critical Care Medicine.  An official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine clinical practice guideline: mechanical ventilation in adult patients with acute respiratory distress syndrome.  Am J Respir Crit Care Med. 2017;195(9):1253-1263. doi:10.1164/rccm.201703-0548STPubMedGoogle ScholarCrossref
7.
Cavalcanti  AB, Suzumura  ÉA, Laranjeira  LN,  et al; Writing Group for the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (ART) Investigators.  Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low PEEP on mortality in patients with acute respiratory distress syndrome: a randomized clinical trial.  JAMA. 2017;318(14):1335-1345. doi:10.1001/jama.2017.14171PubMedGoogle ScholarCrossref
8.
Combes  A, Hajage  D, Capellier  G,  et al; EOLIA Trial Group, REVA, and ECMONet.  Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome.  N Engl J Med. 2018;378(21):1965-1975. doi:10.1056/NEJMoa1800385PubMedGoogle ScholarCrossref
9.
Beitler  JR, Sarge  T, Banner-Goodspeed  VM,  et al; EPVent-2 Study Group.  Effect of titrating positive end-expiratory pressure (PEEP) with an esophageal pressure-guided strategy vs an empirical high PEEP-Fio2 strategy on death and days free from mechanical ventilation among patients with acute respiratory distress syndrome: a randomized clinical trial.  JAMA. 2019;321(9):846-857. doi:10.1001/jama.2019.0555PubMedGoogle ScholarCrossref
10.
Moss  M, Huang  DT, Brower  RG,  et al; National Heart, Lung, and Blood Institute PETAL Clinical Trials Network.  Early neuromuscular blockade in the acute respiratory distress syndrome.  N Engl J Med. 2019;380(21):1997-2008. doi:10.1056/NEJMoa1901686PubMedGoogle ScholarCrossref
11.
Higgins  J, Green  S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. http://handbook-5-1.cochrane.org/. Accessed June 27, 2019.
12.
Hutton  B, Salanti  G, Caldwell  DM,  et al.  The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations.  Ann Intern Med. 2015;162(11):777-784. doi:10.7326/M14-2385PubMedGoogle ScholarCrossref
13.
Ranieri  VM, Rubenfeld  GD, Thompson  BT,  et al; ARDS Definition Task Force.  Acute respiratory distress syndrome: the Berlin Definition.  JAMA. 2012;307(23):2526-2533. doi:10.1001/jama.2012.5669PubMedGoogle Scholar
14.
Cohen  J.  Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit.  Psychol Bull. 1968;70(4):213-220. doi:10.1037/h0026256PubMedGoogle ScholarCrossref
15.
Puhan  MA, Schünemann  HJ, Murad  MH,  et al; GRADE Working Group.  A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analysis.  BMJ. 2014;349:g5630. doi:10.1136/bmj.g5630PubMedGoogle ScholarCrossref
16.
Guyatt  GH, Oxman  AD, Kunz  R,  et al.  GRADE guidelines 6: rating the quality of evidence: imprecision.  J Clin Epidemiol. 2011;64(12):1283-1293. doi:10.1016/j.jclinepi.2011.01.012PubMedGoogle ScholarCrossref
17.
Brignardello-Petersen  R, Bonner  A, Alexander  PE,  et al; GRADE Working Group.  Advances in the GRADE approach to rate the certainty in estimates from a network meta-analysis.  J Clin Epidemiol. 2018;93:36-44. doi:10.1016/j.jclinepi.2017.10.005PubMedGoogle ScholarCrossref
18.
Brignardello-Petersen  R, Murad  MH, Walter  SD,  et al; GRADE Working Group.  GRADE approach to rate the certainty from a network meta-analysis: avoiding spurious judgments of imprecision in sparse networks.  J Clin Epidemiol. 2019;105:60-67. doi:10.1016/j.jclinepi.2018.08.022PubMedGoogle ScholarCrossref
19.
Warn  DE, Thompson  SG, Spiegelhalter  DJ.  Bayesian random effects meta-analysis of trials with binary outcomes: methods for the absolute risk difference and relative risk scales.  Stat Med. 2002;21(11):1601-1623. doi:10.1002/sim.1189PubMedGoogle ScholarCrossref
20.
van Valkenhoef  G, Dias  S, Ades  AE, Welton  NJ.  Automated generation of node-splitting models for assessment of inconsistency in network meta-analysis.  Res Synth Methods. 2016;7(1):80-93. doi:10.1002/jrsm.1167PubMedGoogle ScholarCrossref
21.
Caldwell  DM, Ades  AE, Higgins  JP.  Simultaneous comparison of multiple treatments: combining direct and indirect evidence.  BMJ. 2005;331(7521):897-900. doi:10.1136/bmj.331.7521.897PubMedGoogle ScholarCrossref
22.
van Valkenhoef  G, Lu  G, de Brock  B, Hillege  H, Ades  AE, Welton  NJ.  Automating network meta-analysis.  Res Synth Methods. 2012;3(4):285-299. doi:10.1002/jrsm.1054PubMedGoogle ScholarCrossref
23.
Franchini  AJ, Dias  S, Ades  AE, Jansen  JP, Welton  NJ.  Accounting for correlation in network meta-analysis with multi-arm trials.  Res Synth Methods. 2012;3(2):142-160. doi:10.1002/jrsm.1049PubMedGoogle ScholarCrossref
24.
Brooks  SP, Gelman  A.  General methods for monitoring convergence of iterative simulations.  J Comput Graph Stat. 1998;7(4):434-455. doi:10.1080/10618600.1998.10474787Google Scholar
25.
Spiegelhalter  DJ, Best  NG, Carlin  BP, Van Der Linde  A.  Bayesian measures of model complexity and fit.  J R Stat Soc Series B Stat Methodol. 2002;64(4):583-639. doi:10.1111/1467-9868.00353Google ScholarCrossref
26.
Salanti  G, Ades  AE, Ioannidis  JPA.  Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial.  J Clin Epidemiol. 2011;64(2):163-171. doi:10.1016/j.jclinepi.2010.03.016PubMedGoogle ScholarCrossref
27.
Dias  S, Sutton  AJ, Ades  AE, Welton  NJ.  Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials.  Med Decis Making. 2013;33(5):607-617. doi:10.1177/0272989X12458724PubMedGoogle ScholarCrossref
28.
Ferguson  ND, Cook  DJ, Guyatt  GH,  et al; OSCILLATE Trial Investigators; Canadian Critical Care Trials Group.  High-frequency oscillation in early acute respiratory distress syndrome.  N Engl J Med. 2013;368(9):795-805. doi:10.1056/NEJMoa1215554PubMedGoogle ScholarCrossref
29.
Young  D, Lamb  SE, Shah  S,  et al; OSCAR Study Group.  High-frequency oscillation for acute respiratory distress syndrome.  N Engl J Med. 2013;368(9):806-813. doi:10.1056/NEJMoa1215716PubMedGoogle ScholarCrossref
30.
Mohamed  SAR, Mohamed  NN.  Efficacy and adverse events of early high-frequency oscillatory ventilation in adult burn patients with acute respiratory distress syndrome.  Egypt J Anaesth. 2016;32(3):421-429. doi:10.1016/j.egja.2016.01.001Google ScholarCrossref
31.
Peek  GJ, Mugford  M, Tiruvoipati  R,  et al; CESAR trial collaboration.  Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial.  Lancet. 2009;374(9698):1351-1363. doi:10.1016/S0140-6736(09)61069-2PubMedGoogle ScholarCrossref
32.
Fernandez  R, Trenchs  X, Klamburg  J,  et al.  Prone positioning in acute respiratory distress syndrome: a multicenter randomized clinical trial.  Intensive Care Med. 2008;34(8):1487-1491. doi:10.1007/s00134-008-1119-3PubMedGoogle ScholarCrossref
33.
Taccone  P, Pesenti  A, Latini  R,  et al; Prone-Supine II Study Group.  Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2009;302(18):1977-1984. doi:10.1001/jama.2009.1614PubMedGoogle ScholarCrossref
34.
Guérin  C, Reignier  J, Richard  J-C,  et al; PROSEVA Study Group.  Prone positioning in severe acute respiratory distress syndrome.  N Engl J Med. 2013;368(23):2159-2168. doi:10.1056/NEJMoa1214103PubMedGoogle ScholarCrossref
35.
Gainnier  M, Roch  A, Forel  J-M,  et al.  Effect of neuromuscular blocking agents on gas exchange in patients presenting with acute respiratory distress syndrome.  Crit Care Med. 2004;32(1):113-119. doi:10.1097/01.CCM.0000104114.72614.BCPubMedGoogle ScholarCrossref
36.
Forel  J-M, Roch  A, Marin  V,  et al.  Neuromuscular blocking agents decrease inflammatory response in patients presenting with acute respiratory distress syndrome.  Crit Care Med. 2006;34(11):2749-2757. doi:10.1097/01.CCM.0000239435.87433.0DPubMedGoogle ScholarCrossref
37.
Papazian  L, Forel  J-M, Gacouin  A,  et al; ACURASYS Study Investigators.  Neuromuscular blockers in early acute respiratory distress syndrome.  N Engl J Med. 2010;363(12):1107-1116. doi:10.1056/NEJMoa1005372PubMedGoogle ScholarCrossref
38.
Guervilly  C, Bisbal  M, Forel  JM,  et al.  Effects of neuromuscular blockers on transpulmonary pressures in moderate to severe acute respiratory distress syndrome.  Intensive Care Med. 2017;43(3):408-418. doi:10.1007/s00134-016-4653-4PubMedGoogle ScholarCrossref
39.
Brower  RG, Lanken  PN, MacIntyre  N,  et al; National Heart, Lung, and Blood Institute ARDS Clinical Trials Network.  Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome.  N Engl J Med. 2004;351(4):327-336. doi:10.1056/NEJMoa032193PubMedGoogle ScholarCrossref
40.
Meade  MO, Cook  DJ, Guyatt  GH,  et al; Lung Open Ventilation Study Investigators.  Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2008;299(6):637-645. doi:10.1001/jama.299.6.637PubMedGoogle ScholarCrossref
41.
Mercat  A, Richard  JC, Vielle  B,  et al; Expiratory Pressure (Express) Study Group.  Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2008;299(6):646-655. doi:10.1001/jama.299.6.646PubMedGoogle ScholarCrossref
42.
Talmor  D, Sarge  T, Malhotra  A,  et al.  Mechanical ventilation guided by esophageal pressure in acute lung injury.  N Engl J Med. 2008;359(20):2095-2104. doi:10.1056/NEJMoa0708638PubMedGoogle ScholarCrossref
43.
Huh  JW, Jung  H, Choi  HS, Hong  S-B, Lim  C-M, Koh  Y.  Efficacy of positive end-expiratory pressure titration after the alveolar recruitment manoeuvre in patients with acute respiratory distress syndrome.  Crit Care. 2009;13(1):R22. doi:10.1186/cc7725PubMedGoogle ScholarCrossref
44.
Xi  XM, Jiang  L, Zhu  B; RM group.  Clinical efficacy and safety of recruitment maneuver in patients with acute respiratory distress syndrome using low tidal volume ventilation: a multicenter randomized controlled clinical trial.  Chin Med J (Engl). 2010;123(21):3100-3105.PubMedGoogle Scholar
45.
Hodgson  CL, Tuxen  DV, Davies  AR,  et al.  A randomised controlled trial of an open lung strategy with staircase recruitment, titrated PEEP and targeted low airway pressures in patients with acute respiratory distress syndrome.  Crit Care. 2011;15(3):R133. doi:10.1186/cc10249PubMedGoogle ScholarCrossref
46.
Kacmarek  RM, Villar  J, Sulemanji  D,  et al; Open Lung Approach Network.  Open lung approach for the acute respiratory distress syndrome: a pilot, randomized controlled trial.  Crit Care Med. 2016;44(1):32-42. doi:10.1097/CCM.0000000000001383PubMedGoogle ScholarCrossref
47.
Park  KJ, Lee  YJ, Oh  YJ, Lee  KS, Sheen  SS, Hwang  SC.  Combined effects of inhaled nitric oxide and a recruitment maneuver in patients with acute respiratory distress syndrome.  Yonsei Med J. 2003;44(2):219-226. doi:10.3349/ymj.2003.44.2.219PubMedGoogle ScholarCrossref
48.
Papazian  L, Gainnier  M, Marin  V,  et al.  Comparison of prone positioning and high-frequency oscillatory ventilation in patients with acute respiratory distress syndrome.  Crit Care Med. 2005;33(10):2162-2171. doi:10.1097/01.CCM.0000181298.05474.2BPubMedGoogle ScholarCrossref
49.
Wang  C, Wang  X, Chi  C,  et al.  Lung ventilation strategies for acute respiratory distress syndrome: a systematic review and network meta-analysis.  Sci Rep. 2016;6:22855. doi:10.1038/srep22855PubMedGoogle ScholarCrossref
50.
Goligher  EC, Tomlinson  G, Hajage  D,  et al.  Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome and posterior probability of mortality benefit in a post hoc Bayesian analysis of a randomized clinical trial.  JAMA. 2018;320(21):2251-2259. doi:10.1001/jama.2018.14276PubMedGoogle ScholarCrossref
51.
Munshi  L, Walkey  A, Goligher  E, Pham  T, Uleryk  EM, Fan  E.  Venovenous extracorporeal membrane oxygenation for acute respiratory distress syndrome: a systematic review and meta-analysis.  Lancet Respir Med. 2019;7(2):163-172. doi:10.1016/S2213-2600(18)30452-1PubMedGoogle ScholarCrossref
52.
Gebistorf  F, Karam  O, Wetterslev  J, Afshari  A.  Inhaled nitric oxide for acute respiratory distress syndrome (ARDS) in children and adults.  Cochrane Database Syst Rev. 2016;(6):CD002787. doi:10.1002/14651858.CD002787.pub3PubMedGoogle Scholar
53.
Alhazzani  W, Alshahrani  M, Jaeschke  R,  et al.  Neuromuscular blocking agents in acute respiratory distress syndrome: a systematic review and meta-analysis of randomized controlled trials.  Crit Care. 2013;17(2):R43. doi:10.1186/cc12557PubMedGoogle ScholarCrossref
54.
Barbaro  RP, Odetola  FO, Kidwell  KM,  et al.  Association of hospital-level volume of extracorporeal membrane oxygenation cases and mortality: analysis of the extracorporeal life support organization registry.  Am J Respir Crit Care Med. 2015;191(8):894-901. doi:10.1164/rccm.201409-1634OCPubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 3,167
    Original Investigation
    Critical Care Medicine
    July 31, 2019

    Assessment of Therapeutic Interventions and Lung Protective Ventilation in Patients With Moderate to Severe Acute Respiratory Distress Syndrome: A Systematic Review and Network Meta-analysis

    Author Affiliations
    • 1Department of Anesthesiology, Graduate School of Medicine, University of Tokyo, Tokyo, Japan
    • 2Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada
    • 3Department of Anesthesia and Pain Medicine, The Hospital for Sick Children, Toronto, Ontario, Canada
    • 4Peter Gilgan Centre for Research and Learning, The Hospital for Sick Children Research Institute, Toronto, Ontario, Canada
    • 5Institute for Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
    • 6Library and Information Services, University Health Network, Toronto, Ontario, Canada
    JAMA Netw Open. 2019;2(7):e198116. doi:10.1001/jamanetworkopen.2019.8116
    Key Points español 中文 (chinese)

    Question  What is the relative association of management strategies for adult patients with moderate to severe acute respiratory distress syndrome with mortality and barotrauma?

    Findings  In this systematic review and network meta-analysis of 25 randomized clinical trials including 7743 patients, venovenous extracorporeal membrane oxygenation and prone positioning were associated with significantly lower 28-day mortality compared with lung protective ventilation alone. Moreover, venovenous extracorporeal membrane oxygenation was the highest-ranked intervention associated with a reduction in 28-day mortality among the 9 interventions evaluated.

    Meaning  These findings support the use of prone positioning in patients with moderate to severe acute respiratory distress syndrome and venovenous extracorporeal membrane oxygenation in patients with severe acute respiratory distress syndrome.

    Abstract

    Importance  A number of interventions are available to manage patients with moderate to severe acute respiratory distress syndrome (ARDS). However, the associations of currently available ventilatory strategies and adjunctive therapies with mortality are uncertain.

    Objectives  To compare and rank different therapeutic strategies to identify the best intervention associated with a reduction in mortality in adult patients with moderate to severe ARDS.

    Data Sources  An electronic search of MEDLINE, MEDLINE In-Process/ePubs Ahead of Print, Embase, Cochrane Controlled Clinical Trial Register (Central), PubMed, and CINAHL was conducted, from database inception to May 29, 2019.

    Study Selection  Randomized clinical trials of interventions for adults with moderate to severe ARDS that used lung protective ventilation. No language restrictions were applied.

    Data Extraction and Synthesis  Data were independently extracted by 2 reviewers and synthesized with Bayesian random-effects network meta-analyses.

    Main Outcomes and Measures  The primary outcome was 28-day mortality. Barotrauma was a secondary outcome.

    Results  Among 25 randomized clinical trials evaluating 9 interventions, 2686 of 7743 patients (34.6%) died within 28 days. Compared with lung protective ventilation alone, prone positioning and venovenous extracorporeal membrane oxygenation were associated with significantly lower 28-day mortality (prone positioning: risk ratio, 0.69; 95% credible interval, 0.48-0.99; low quality of evidence; venovenous extracorporeal membrane oxygenation: risk ratio, 0.60; 95% credible interval, 0.38-0.93; moderate quality of evidence). These 2 interventions had the highest ranking probabilities, although they were not significantly different from each other. Among 18 trials reporting on barotrauma, 448 of 6258 patients (7.2%) experienced this secondary outcome. No intervention was superior to any other in reducing barotrauma, and each represented low to very low quality of evidence.

    Conclusions and Relevance  This network meta-analysis supports the use of prone positioning and venovenous extracorporeal membrane oxygenation in addition to lung protective ventilation in patients with ARDS. Moreover, venovenous extracorporeal membrane oxygenation may be considered as an early strategy for adults with severe ARDS receiving lung protective ventilation.

    Introduction

    Acute respiratory distress syndrome (ARDS) is a lethal condition whereby the lung is injured by direct (eg, pneumonia or aspiration) or indirect (eg, extrapulmonary sepsis) insults. Clinically, patients with ARDS develop severe hypoxemia and/or hypercapnia, and most die of sepsis or multiorgan failure rather than from refractory respiratory failure. Acute respiratory distress is an important public health problem. A global epidemiologic study1 reported that 10.4% of total intensive care unit admissions and 23.4% of all patients who were intubated had ARDS, with an associated hospital mortality of 40%.

    Since the first description of ARDS in 1967,2 a number of approaches to its management have been evaluated and used clinically.3,4 Despite more than 50 years of research, to our knowledge, none of the treatments currently available are aimed directly at the pathophysiological mechanism resulting in acute respiratory failure (ie, increased alveolar capillary permeability); current approaches are mainly supportive. Perhaps most important has been the recognition that although mechanical ventilation is critical for the survival of patients with ARDS, it can also be injurious (ie, ventilator-induced lung injury).5 Indeed, lung protective ventilation (LPV), using low tidal volumes and airway pressures to mitigate ventilator-induced lung injury, is considered the mainstay of management in patients with ARDS.6

    In 2017, clinical practice guidelines on mechanical ventilation in adult patients with ARDS were published for 6 individual interventions, as follows: (1) LPV, (2) higher positive end-expiratory pressure (PEEP), (3) lung recruitment maneuvers (RMs), (4) high-frequency oscillatory ventilation (HFOV), (5) prone positioning, and (6) venovenous extracorporeal membrane oxygenation (VV ECMO).6 However, for clinicians, choosing between potentially efficacious treatments can be challenging if the treatments have not been directly compared in clinical trials (eg, prone positioning vs VV ECMO). Moreover, 4 randomized clinical trials (RCTs)—the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (ART),7 the ECMO to Rescue Lung Injury in Severe ARDS (EOLIA) trial,8 the Esophageal Pressure-Guided Ventilation 2 (EPVent2) trial,9 and the Reevaluation of Systemic Early Neuromuscular Blockade (ROSE) trial10—have been published since the guidelines were completed. Therefore, we conducted a systematic review and network meta-analysis to compare different therapeutic strategies simultaneously to identify the best strategy associated with a reduction in mortality and to rank those therapeutic modalities for adult patients with moderate to severe ARDS.

    Methods
    Eligibility Criteria, Literature Search, and Study Selection

    We followed the steps outlined by the Cochrane Collaboration11 and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.12 We included RCTs or quasi-RCTs enrolling adult patients (aged ≥18 years) with moderate to severe ARDS who received mechanical ventilation in the intensive care unit.13 For studies using the American-European Consensus Conference to diagnose ARDS, we checked Pao2-to-fraction of inspired oxygen (Fio2) ratio of inclusion criteria or the mean and distribution of Pao2/Fio2 ratio in each trial to determine the severity of ARDS among enrolled patients. We included interventions available for moderate to severe ARDS, either alone or in combination with LPV or another intervention. We defined LPV as a mechanical ventilation using low tidal volume of 4 to 8 mL/kg of predicted body weight. The interventions that we considered a priori were LPV, open lung strategies (ie, RM or PEEP), neuromuscular blockade (NMBA), inhaled nitric oxide (INO), HFOV, prone positioning, and VV ECMO. Participants in the comparator group could also have received 1 cointervention (eg, NMBA), as described above. Our primary outcome was 28-day mortality. If not reported explicitly, we identified or calculated 28-day mortality from Kaplan-Meier curves (using Digitaliser v10.9 [Engauge]) or from the closest reported time point, assuming constant mortality rate over time. Our prespecified secondary outcome was barotrauma at any time point. Inclusion and exclusion criteria are summarized in eAppendix 1 in the Supplement.

    We performed an electronic search of MEDLINE, MEDLINE In-Process/ePub Ahead of Print, Embase, Cochrane Controlled Clinical Trial Register (Central) (via the Ovid search interface), PubMed (via the National Library of Medicine and excluding Medline records), and CINAHL (via EbscoHost) from database inception to May 29, 2019, using a sensitive search strategy (eAppendix 2 in the Supplement). We used controlled vocabulary terms (when available), text words, and keywords. No language restrictions were applied. We screened the reference lists of key articles for additional potentially relevant articles.

    Two reviewers (H.A. and K.U.) independently identified and assessed potentially eligible studies for inclusion in the review, and any disagreement and discrepancies were resolved by discussion with and adjudication by a third reviewer (E.F.). Cohen κ was reported for agreement between the 2 reviewers.14 Data extraction and risk-of-bias assessment were also performed independently in duplicate by 2 authors (H.A. and K.U.).

    Data Extraction and Risk-of-Bias Assessment

    A standardized, piloted data collection form designed for this systematic review was developed for data extraction. Risk of bias for each eligible study was determined using the Cochrane risk-of-bias tool.11 The certainty of evidence for the network meta-analysis was assessed and determined using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) tool for network meta-analysis.15 The risk of bias was graded as low, high, or unclear on the basis of each study’s randomization, allocation concealment, blinding of outcome assessment, losses to follow-up, treatment of withdrawals, and selective reporting. For performance and detection bias domains, we judged that, because mortality is objective, it was unlikely to be influenced by lack of blinding as long as a strict protocol for both groups was provided. To determine imprecision, a sample size required to detect a 30% relative risk reduction (optimal information size) was calculated for each comparison for each outcome based on a total event rate in the control group.16-18 Funnel plots were used to assess publication bias for each arm of the comparison, and further statistical analysis of funnel plot asymmetry was planned if there were more than 10 trials in each arm.11 Assessment of heterogeneity, consistency, and intransitivity are described in eAppendix 3 in the Supplement.

    Statistical Analysis

    Standard pairwise meta-analysis methods with random-effect models were used to analyze interventions of eligible RCTs directly, where forest plots were constructed with subsequent calculation of risk ratios and 95% CIs for effect size. The I2 statistic was used to assess statistical heterogeneity in pairwise meta-analysis (eAppendix 3 in the Supplement). For network meta-analysis, we calculated effect sizes by determining risk ratios and 95% credible intervals (CrI) by the Bayesian hierarchical random-effects model, using Markov chain Monte Carlo simulation with noninformative prior distributions.19-22 For multiarm trials, correction of the treatment effects between arms was taken into account.23 Generalized linear models with a log-link function were applied for the analysis, with 4 chains and 2 000 000 iterated simulations, discarding the initial 1 500 000 iterations as burn-in. Potential scale reduction factor derived from the Brooks-Gelman-Rubin diagnostic was used for assessment of model convergence.24 Model fit was assessed by residual deviance, leverage, and the deviance information criterion.25 Consistency (ie, between-trial differences in the underlying treatment effects between comparisons) was assessed by the node-splitting method (ie, exploring differences between the treatment effects estimated by direct evidence and treatment effects estimated using indirect evidence), and transitivity (ie, the assumption that all treatments are equally likely candidates for the patients in the network) between comparisons was assessed by inspection of differences in potential effect modifiers (eAppendix 3 in the Supplement).

    A rank statistic was determined and represented in a rankogram to illustrate the probability that a chosen treatment of all eligible interventions to be investigated was associated with the best, second best, and so on reduction in mortality.26 We also used the surface under cumulative ranking (SUCRA) curve to provide a numerical ranking of the association of all treatments with reduction in mortality, from 0 (certain to be the worst) to 100 (certain to be the best).26

    Sensitivity analyses were conducted by excluding trials with high risk of bias and trials without a description of cointerventions. Model fit excluding small-sized trials was assessed as a sensitivity analysis. The other sensitivity analysis was to use Poisson models in the network meta-analysis for the primary outcome to adjust for different follow-up periods.27 Preplanned subgroup analysis was not conducted for the primary and secondary outcomes to assess the association of the distribution of outcome modifiers, including age and ARDS severity (ie, Pao2/Fio2 ratio), because network metaregression with treatment by covariate interactions showed none of the interventions included in our study were affected by age and ARDS severity at the study level (eAppendix 4 and eFigure 1 in the Supplement).

    All statistical significance testing was 2-sided, and P < .05 was considered statistically significant. RevMan version 5.1 (The Nordic Cochrane Centre) was used to generate funnel plots and the risk-of-bias tables. Statistical analyses were conducted using R version 3.5.1 (The R Foundation) with R packages gemtc, coda, pcnetmeta, and rjags and using Just Another Giggs Samples version 4.3.0 (JAGS).

    Results

    We identified 10 195 records in our electronic search (Figure 1). After screening by title and abstract, we obtained full-text articles for 49 citations that were potentially eligible for inclusion. We included 25 studies (7753 participants; range, 20-1010 participants) in this review (Table 1),7-10,28-48 while 24 studies did not meet our inclusion criteria (eTable 1 in the Supplement). There was near-perfect agreement on study inclusion between the 2 reviewers (κ = 0.97). Overall, 9 interventions were investigated (Figure 2) (eFigure 2 and eFigure 3 in the Supplement): LPV (23 trials),7-10,28-46 an open lung strategy using RM and/or higher PEEP (10 trials),7,9,39-46 NMBA using a 48-hour infusion of cisatracurium (5 trials),10,35-38 INO (1 trial),47 INO with RM (1 trial),47 HFOV (3 trials),28-30 HFOV with prone positioning (1 trial),48 prone positioning (3 trials),32-34 and VV ECMO (2 trials).8,31

    There was high risk of bias for 1 or more key domains in 4 studies (eFigures 4-6 in the Supplement).30,39,44,47 All trials specified standardized protocols for the implementation of each intervention and ventilatory management; thus, blinding domains were judged low for our primary outcome. As a result, 14 of 25 studies (56%) had a low risk of bias across all domains.7,9,10,29,34-38,40-42,45,46

    Among the 25 studies, overall 28-day mortality was 34.6% (2686 of 7753 patients). Overall, LPV was the most frequently investigated intervention, whereas 3 interventions were investigated by only 1 trial (Figure 2). Compared with LPV alone, prone positioning (risk ratio, 0.69; 95% CrI, 0.48-0.98; low quality of evidence) and VV ECMO (risk ratio, 0.60; 95% CrI, 0.38-0.93; moderate quality of evidence) were associated with significantly lower 28-day mortality (Table 2) (eTable 2, eFigure 2, and eFigure 7 in the Supplement). Prone positioning prevented 124 more deaths per 1000 patients compared with LPV alone, and VV ECMO prevented 161 more deaths per 1000 patients compared with LPV alone (Table 2). The results of comparisons between all possible pairs of interventions with quality of evidence are summarized in eTable 3 in the Supplement. In a sensitivity analysis restricted to studies without a high risk of bias (12 studies; 4213 patients) and studies with a description of cointerventions (17 studies; 7196 patients), prone positioning and VV ECMO remained significantly associated with lower 28-day mortality (eTable 4 and eTable 5 in the Supplement). A sensitivity analysis using Poisson models showed that VV ECMO and prone positioning were significantly associated with reduced mortality (eTable 6 in the Supplement). Exclusion of small-sized trials did not change model fit in the network meta-analysis (eTables 7-9 in the Supplement).

    The incidence of barotrauma was 7.2% (448 of 6253 patients) from 17 trials evaluating 6 interventions (Figure 2). There were no significant differences between interventions in the risk of barotrauma, with variable quality of evidence (Table 3) (eTable 10, eTable 11, eFigure 3, and eFigure 7 in the Supplement).

    For mortality, both VV ECMO and prone positioning were ranked highly on the basis of SUCRA, with VV ECMO being highest (0.82). However, the ranking probability for VV ECMO did not differ significantly from prone positioning (eFigure 8 in the Supplement). For barotrauma, NMBA had the highest SUCRA (0.93), although no intervention was significantly different from any other in reducing barotrauma in studies reporting this outcome (eFigure 8 in the Supplement).

    There were 36 direct or indirect comparisons for the primary outcome among 25 studies (eTable 3 in the Supplement). Only 7 comparisons were not affected by moderate or high heterogeneity (eTable 3 in the Supplement). Comparisons of LPV, HFOV, and open lung strategy using RM and/or higher PEEP in an arm with any intervention in the other arm met the optimal information size for imprecision. Node splitting found no significant inconsistency in 3 comparisons (HFOV vs LPV, prone positioning vs LPV, and prone positioning vs HFOV) (eTable 12 in the Supplement). Intransitivity was not found based on network metaregression with treatment by covariate interactions to measure potential effect modifiers. Based on these findings in the primary outcome among all 36 comparisons, the quality of evidence for network effects estimates was judged as high in 1 comparison, moderate in 4, low in 6, and very low in 25 (eTable 3 in the Supplement).

    No comparison for barotrauma had a high risk of bias or met the optimal information size in imprecision (eTable 11 in the Supplement). All comparisons, except VV ECMO vs LPV, showed moderate to high heterogeneity. Because there was no closed loop of interventions for barotrauma, the assumption of consistency in any comparison cannot be violated. Intransitivity was not found. The quality of evidence for network effects estimates in barotrauma was judged as moderate in 1 comparison, low in 4, and very low in 10 (eTable 11 in the Supplement).

    Discussion

    Our systematic review and network meta-analysis of 25 RCTs, which included 7753 patients and 9 interventions, found that prone positioning was associated with significantly lower 28-day mortality in patients with moderate ARDS compared with LPV alone. In patients with severe ARDS, VV ECMO was associated with significantly lower 28-day mortality. Furthermore, these interventions were highly ranked, with VV ECMO being the highest, although there were no significant differences in ranking probabilities between these interventions. No intervention was superior to any other in reducing barotrauma.

    While the use of LPV remains the mainstay of supportive care in patients with ARDS,6 LPV alone may be insufficient to maintain adequate gas exchange or prevent ventilator-induced lung injury, and adjunctive interventions may be required. While many of these adjunctive interventions have been evaluated in RCTs or meta-analyses, to our knowledge, there are limited data comparing their relative efficacy with each other. Wang et al49 assessed 26 ventilatory strategies in their network meta-analysis for ARDS but divided the interventions into complex groupings, making clinical interpretation difficult. Moreover, they made comparisons with a strategy of higher tidal volumes that has little clinical relevance in the current management of patients with ARDS.49 In contrast, we decided to focus on a limited number of interventions that are commonly used in patients with moderate to severe ARDS1 rather than include all possibilities to provide the most relevant guidance for clinicians at the bedside.

    Our results are consistent with the strong recommendations from the American Thoracic Society, European Society of Intensive Care Medicine, and Society of Critical Care Medicine clinical practice guidelines6 based on conventional (pairwise) meta-analyses. Specifically, our study supports the use of prone positioning (ie, significant reduction in mortality and high ranking) as well as the strong recommendation against the routine use of HFOV (ie, increased mortality and low ranking). The conditional recommendations for RMs or higher PEEP did not include the results of the ART7 or the EPVent2 trial.9 Our pooled results did not find an association of RMs or higher PEEP with mortality. Importantly, the guideline could not make a recommendation on the use of VV ECMO in patients with severe ARDS owing to insufficient data. Our study supports the earlier consideration of VV ECMO along with LPV in patients with severe ARDS, consistent with the results of the EOLIA trial,8 the post hoc Bayesian reanalysis of that trial,50 and the updated meta-analysis of VV ECMO studies.51 The results for INO, which were not included in the guideline, are consistent with the 2016 meta-analysis52 that showed an increased risk for renal failure with no significant mortality benefit for INO. Finally, although the most recent meta-analysis of NMBA, to our knowledge, showed a reduction in mortality,53 our study results are consistent with the results of the ROSE trial,10 in which NMBA did not improve mortality of patients with moderate to severe ARDS.

    By including the ART,7 EPVent2,9 EOLIA,8 and ROSE trials10 in our network meta-analysis, our findings could inform future updates of the clinical practice guidelines and decision-making process at the bedside. Given the limited clinical interpretation of SUCRA (relative efficacy assessment on eligible interventions),26 our findings demonstrate that prone positioning and VV ECMO should be considered early, with implementation of these interventions influenced by individual patient characteristics or available clinical resources. Our preplanned metaregression confirmed that ARDS severity (ie, aggregated Pao2/Fio2 ratio at study level) did not affect the findings of these analyses. However, VV ECMO should be restricted to severe ARDS, since the 2 included RCTs of VV ECMO only recruited patients with severe ARDS. Moreover, in addition to the statistical heterogeneity in severity among eligible studies, considerations such as the invasiveness, availability of resources, and need for clinical experience may also limit the use of VV ECMO to the patients with the most severe ARDS in high-volume centers.54 Therefore, prone positioning should be considered as a first-line approach for patients with moderate to severe ARDS in addition to LPV. Timely transfer to an ECMO-capable center and VV ECMO should be considered for patients with severe ARDS who have contraindications or who fail prone positioning, similar to the EOLIA trial.8 The strategy of HFOV combined with prone positioning was associated with a nonsignificant reduction in mortality but with a high SUCRA; however, inferences about its clinical efficacy and relevance remain uncertain owing to limitations from a single study with very small sample size.48 The lower rankings of the other 5 interventions (NMBA, HFOV, INO, INO with RM, and open lung strategy using RM and/or higher PEEP) suggests against their routine use.

    Limitations

    This study has several limitations. First, the network meta-analysis depends on the assumption that population and intervention characteristics were largely similar across the included studies (ie, transitivity). Thus, an analysis including trials applying different thresholds of Pao2/Fio2 ratio (ie, severity of ARDS) for the patient recruitment is subject to bias. For example, the EOLIA trial only enrolled patients with severe ARDS. Furthermore, this assumption may also be violated by the difference in the period of those trials performed. We confirmed that there was no effect modification of Pao2/Fio2 ratio on treatment by covariate interaction by network metaregressions with Pao2/Fio2 ratio (eFigure 1 in the Supplement). However, the most robust method for evaluating the potential for effect modification by Pao2/Fio2 ratio would be with a meta-analysis of individual patient data. Second, some studies allowed or recommended cointerventions in some settings for both experimental and control groups (eg, prone positioning for certain Pao2/Fio2 ratios in a study of NMBA) but not in all participants. However, many studies did not explicitly describe the use of cointerventions in the protocol or results of their published articles. We conducted a sensitivity analysis excluding studies without a description of cointerventions, which was robust for the main analysis (eTable 5 and eTable 13 in the Supplement). Moreover, we attempted to address this limitation only by downrating the quality in GRADE assessment. Future trials are strongly encouraged to provide a detailed description of cointerventions used or to protocolize their implementation within the trial. Third, we only evaluated short-term mortality. However, we believe that this approach is theoretically justified because most interventions studied have the main rationale of reducing ventilator-induced lung injury and more immediate death. Future studies should include an evaluation of long-term mortality and functional outcomes. Fourth, the adverse effects of treatments were limited because of the nature of different interventions. We only assessed barotrauma, and there were differences in definition of barotrauma in each study. Fifth, although intervention from smaller studies (eg, INO, HFOV with prone positioning) can lead to imprecision of estimated effects (ie, wider CrIs), we confirmed that excluding such small-sized studies did not change overall model fitting of the network meta-analysis and the other main findings (eTable 7-9 in the Supplement). Sixth, we could not specify 28-day mortality in 5 of 25 trials (20%) and computed mortality assuming constant risk of death during the study period. An alternative way to address this issue was to use Poisson models in the network meta-analysis and use follow-up period as offset in the model. Using this approach, we found no differences between Poisson models and our primary analyses (eTable 6 in the Supplement). Seventh, we did not account for covariates (ie, age and Pao2/Fio2 ratio) in the final models of network meta-analysis because we did not find treatment by covariate interaction from network metaregressions with these covariates. However, these findings were based on study-level data instead of individual patient data, so aggregation bias still remains possible. Eighth, a meta-analysis with individual patient data may be desirable to further separate the effectiveness of each intervention and explore outcomes in important patient subgroups. Despite these limitations, the use of network meta-analysis enabled us to compare clinically relevant management strategies in patients with ARDS. Ninth, we used robust methods as recommended by the Cochrane Collaboration, PRISMA, and GRADE for network meta-analysis.11,12,15

    Conclusions

    This network meta-analysis supports the use of prone positioning and VV ECMO in addition to LPV in patients with moderate to severe ARDS and severe ARDS, respectively. In addition, our results are consistent with recent data suggesting that VV ECMO may be considered as an early strategy for adults with severe ARDS.

    Back to top
    Article Information

    Accepted for Publication: June 6, 2019.

    Published: July 31, 2019. doi:10.1001/jamanetworkopen.2019.8116

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Aoyama H et al. JAMA Network Open.

    Corresponding Author: Eddy Fan, MD, PhD, Interdepartmental Division of Critical Care Medicine, University of Toronto, 585 University Ave, PMB 11-123, Toronto, ON M5G 2N2, Canada (eddy.fan@uhn.ca).

    Author Contributions: Drs H. Aoyama and Fan had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: H. Aoyama, Uchida, K. Aoyama, Pechlivanoglou, Englesakis, Fan.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: H. Aoyama, K. Aoyama, Fan.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: H. Aoyama, K. Aoyama, Pechlivanoglou.

    Obtained funding: K. Aoyama.

    Administrative, technical, or material support: H. Aoyama, K. Aoyama, Englesakis, Fan.

    Supervision: Yamada, Fan.

    Conflict of Interest Disclosures: Dr Uchida reported receiving grants from the Japan Society for the Promotion of Science and a collaborative research agreement with Nihon Kohden Cooperation outside the submitted work. Dr Fan reported receiving personal fees from MC3 Cardiopulmonary and personal fees from ALung Technologies outside the submitted work. No other disclosures were reported.

    Funding/Support: Dr K. Aoyama is supported by a Fellowship Award from the Canadian Institutes of Health Research. Dr Fan is supported by a New Investigator Award from the Canadian Institutes of Health Research.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Bellani  G, Laffey  JG, Pham  T,  et al; LUNG SAFE Investigators; ESICM Trials Group.  Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.  JAMA. 2016;315(8):788-800. doi:10.1001/jama.2016.0291PubMedGoogle ScholarCrossref
    2.
    Ashbaugh  DG, Bigelow  DB, Petty  TL, Levine  BE.  Acute respiratory distress in adults.  Lancet. 1967;2(7511):319-323. doi:10.1016/S0140-6736(67)90168-7PubMedGoogle ScholarCrossref
    3.
    Fan  E, Brodie  D, Slutsky  AS.  Acute respiratory distress syndrome: advances in diagnosis and treatment.  JAMA. 2018;319(7):698-710. doi:10.1001/jama.2017.21907PubMedGoogle ScholarCrossref
    4.
    Thompson  BT, Chambers  RC, Liu  KD.  Acute respiratory distress syndrome.  N Engl J Med. 2017;377(6):562-572. doi:10.1056/NEJMra1608077PubMedGoogle ScholarCrossref
    5.
    Slutsky  AS, Ranieri  VM.  Ventilator-induced lung injury.  N Engl J Med. 2013;369(22):2126-2136. doi:10.1056/NEJMra1208707PubMedGoogle ScholarCrossref
    6.
    Fan  E, Del Sorbo  L, Goligher  EC,  et al; American Thoracic Society, European Society of Intensive Care Medicine, and Society of Critical Care Medicine.  An official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine clinical practice guideline: mechanical ventilation in adult patients with acute respiratory distress syndrome.  Am J Respir Crit Care Med. 2017;195(9):1253-1263. doi:10.1164/rccm.201703-0548STPubMedGoogle ScholarCrossref
    7.
    Cavalcanti  AB, Suzumura  ÉA, Laranjeira  LN,  et al; Writing Group for the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (ART) Investigators.  Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low PEEP on mortality in patients with acute respiratory distress syndrome: a randomized clinical trial.  JAMA. 2017;318(14):1335-1345. doi:10.1001/jama.2017.14171PubMedGoogle ScholarCrossref
    8.
    Combes  A, Hajage  D, Capellier  G,  et al; EOLIA Trial Group, REVA, and ECMONet.  Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome.  N Engl J Med. 2018;378(21):1965-1975. doi:10.1056/NEJMoa1800385PubMedGoogle ScholarCrossref
    9.
    Beitler  JR, Sarge  T, Banner-Goodspeed  VM,  et al; EPVent-2 Study Group.  Effect of titrating positive end-expiratory pressure (PEEP) with an esophageal pressure-guided strategy vs an empirical high PEEP-Fio2 strategy on death and days free from mechanical ventilation among patients with acute respiratory distress syndrome: a randomized clinical trial.  JAMA. 2019;321(9):846-857. doi:10.1001/jama.2019.0555PubMedGoogle ScholarCrossref
    10.
    Moss  M, Huang  DT, Brower  RG,  et al; National Heart, Lung, and Blood Institute PETAL Clinical Trials Network.  Early neuromuscular blockade in the acute respiratory distress syndrome.  N Engl J Med. 2019;380(21):1997-2008. doi:10.1056/NEJMoa1901686PubMedGoogle ScholarCrossref
    11.
    Higgins  J, Green  S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. http://handbook-5-1.cochrane.org/. Accessed June 27, 2019.
    12.
    Hutton  B, Salanti  G, Caldwell  DM,  et al.  The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations.  Ann Intern Med. 2015;162(11):777-784. doi:10.7326/M14-2385PubMedGoogle ScholarCrossref
    13.
    Ranieri  VM, Rubenfeld  GD, Thompson  BT,  et al; ARDS Definition Task Force.  Acute respiratory distress syndrome: the Berlin Definition.  JAMA. 2012;307(23):2526-2533. doi:10.1001/jama.2012.5669PubMedGoogle Scholar
    14.
    Cohen  J.  Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit.  Psychol Bull. 1968;70(4):213-220. doi:10.1037/h0026256PubMedGoogle ScholarCrossref
    15.
    Puhan  MA, Schünemann  HJ, Murad  MH,  et al; GRADE Working Group.  A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analysis.  BMJ. 2014;349:g5630. doi:10.1136/bmj.g5630PubMedGoogle ScholarCrossref
    16.
    Guyatt  GH, Oxman  AD, Kunz  R,  et al.  GRADE guidelines 6: rating the quality of evidence: imprecision.  J Clin Epidemiol. 2011;64(12):1283-1293. doi:10.1016/j.jclinepi.2011.01.012PubMedGoogle ScholarCrossref
    17.
    Brignardello-Petersen  R, Bonner  A, Alexander  PE,  et al; GRADE Working Group.  Advances in the GRADE approach to rate the certainty in estimates from a network meta-analysis.  J Clin Epidemiol. 2018;93:36-44. doi:10.1016/j.jclinepi.2017.10.005PubMedGoogle ScholarCrossref
    18.
    Brignardello-Petersen  R, Murad  MH, Walter  SD,  et al; GRADE Working Group.  GRADE approach to rate the certainty from a network meta-analysis: avoiding spurious judgments of imprecision in sparse networks.  J Clin Epidemiol. 2019;105:60-67. doi:10.1016/j.jclinepi.2018.08.022PubMedGoogle ScholarCrossref
    19.
    Warn  DE, Thompson  SG, Spiegelhalter  DJ.  Bayesian random effects meta-analysis of trials with binary outcomes: methods for the absolute risk difference and relative risk scales.  Stat Med. 2002;21(11):1601-1623. doi:10.1002/sim.1189PubMedGoogle ScholarCrossref
    20.
    van Valkenhoef  G, Dias  S, Ades  AE, Welton  NJ.  Automated generation of node-splitting models for assessment of inconsistency in network meta-analysis.  Res Synth Methods. 2016;7(1):80-93. doi:10.1002/jrsm.1167PubMedGoogle ScholarCrossref
    21.
    Caldwell  DM, Ades  AE, Higgins  JP.  Simultaneous comparison of multiple treatments: combining direct and indirect evidence.  BMJ. 2005;331(7521):897-900. doi:10.1136/bmj.331.7521.897PubMedGoogle ScholarCrossref
    22.
    van Valkenhoef  G, Lu  G, de Brock  B, Hillege  H, Ades  AE, Welton  NJ.  Automating network meta-analysis.  Res Synth Methods. 2012;3(4):285-299. doi:10.1002/jrsm.1054PubMedGoogle ScholarCrossref
    23.
    Franchini  AJ, Dias  S, Ades  AE, Jansen  JP, Welton  NJ.  Accounting for correlation in network meta-analysis with multi-arm trials.  Res Synth Methods. 2012;3(2):142-160. doi:10.1002/jrsm.1049PubMedGoogle ScholarCrossref
    24.
    Brooks  SP, Gelman  A.  General methods for monitoring convergence of iterative simulations.  J Comput Graph Stat. 1998;7(4):434-455. doi:10.1080/10618600.1998.10474787Google Scholar
    25.
    Spiegelhalter  DJ, Best  NG, Carlin  BP, Van Der Linde  A.  Bayesian measures of model complexity and fit.  J R Stat Soc Series B Stat Methodol. 2002;64(4):583-639. doi:10.1111/1467-9868.00353Google ScholarCrossref
    26.
    Salanti  G, Ades  AE, Ioannidis  JPA.  Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial.  J Clin Epidemiol. 2011;64(2):163-171. doi:10.1016/j.jclinepi.2010.03.016PubMedGoogle ScholarCrossref
    27.
    Dias  S, Sutton  AJ, Ades  AE, Welton  NJ.  Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials.  Med Decis Making. 2013;33(5):607-617. doi:10.1177/0272989X12458724PubMedGoogle ScholarCrossref
    28.
    Ferguson  ND, Cook  DJ, Guyatt  GH,  et al; OSCILLATE Trial Investigators; Canadian Critical Care Trials Group.  High-frequency oscillation in early acute respiratory distress syndrome.  N Engl J Med. 2013;368(9):795-805. doi:10.1056/NEJMoa1215554PubMedGoogle ScholarCrossref
    29.
    Young  D, Lamb  SE, Shah  S,  et al; OSCAR Study Group.  High-frequency oscillation for acute respiratory distress syndrome.  N Engl J Med. 2013;368(9):806-813. doi:10.1056/NEJMoa1215716PubMedGoogle ScholarCrossref
    30.
    Mohamed  SAR, Mohamed  NN.  Efficacy and adverse events of early high-frequency oscillatory ventilation in adult burn patients with acute respiratory distress syndrome.  Egypt J Anaesth. 2016;32(3):421-429. doi:10.1016/j.egja.2016.01.001Google ScholarCrossref
    31.
    Peek  GJ, Mugford  M, Tiruvoipati  R,  et al; CESAR trial collaboration.  Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial.  Lancet. 2009;374(9698):1351-1363. doi:10.1016/S0140-6736(09)61069-2PubMedGoogle ScholarCrossref
    32.
    Fernandez  R, Trenchs  X, Klamburg  J,  et al.  Prone positioning in acute respiratory distress syndrome: a multicenter randomized clinical trial.  Intensive Care Med. 2008;34(8):1487-1491. doi:10.1007/s00134-008-1119-3PubMedGoogle ScholarCrossref
    33.
    Taccone  P, Pesenti  A, Latini  R,  et al; Prone-Supine II Study Group.  Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2009;302(18):1977-1984. doi:10.1001/jama.2009.1614PubMedGoogle ScholarCrossref
    34.
    Guérin  C, Reignier  J, Richard  J-C,  et al; PROSEVA Study Group.  Prone positioning in severe acute respiratory distress syndrome.  N Engl J Med. 2013;368(23):2159-2168. doi:10.1056/NEJMoa1214103PubMedGoogle ScholarCrossref
    35.
    Gainnier  M, Roch  A, Forel  J-M,  et al.  Effect of neuromuscular blocking agents on gas exchange in patients presenting with acute respiratory distress syndrome.  Crit Care Med. 2004;32(1):113-119. doi:10.1097/01.CCM.0000104114.72614.BCPubMedGoogle ScholarCrossref
    36.
    Forel  J-M, Roch  A, Marin  V,  et al.  Neuromuscular blocking agents decrease inflammatory response in patients presenting with acute respiratory distress syndrome.  Crit Care Med. 2006;34(11):2749-2757. doi:10.1097/01.CCM.0000239435.87433.0DPubMedGoogle ScholarCrossref
    37.
    Papazian  L, Forel  J-M, Gacouin  A,  et al; ACURASYS Study Investigators.  Neuromuscular blockers in early acute respiratory distress syndrome.  N Engl J Med. 2010;363(12):1107-1116. doi:10.1056/NEJMoa1005372PubMedGoogle ScholarCrossref
    38.
    Guervilly  C, Bisbal  M, Forel  JM,  et al.  Effects of neuromuscular blockers on transpulmonary pressures in moderate to severe acute respiratory distress syndrome.  Intensive Care Med. 2017;43(3):408-418. doi:10.1007/s00134-016-4653-4PubMedGoogle ScholarCrossref
    39.
    Brower  RG, Lanken  PN, MacIntyre  N,  et al; National Heart, Lung, and Blood Institute ARDS Clinical Trials Network.  Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome.  N Engl J Med. 2004;351(4):327-336. doi:10.1056/NEJMoa032193PubMedGoogle ScholarCrossref
    40.
    Meade  MO, Cook  DJ, Guyatt  GH,  et al; Lung Open Ventilation Study Investigators.  Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2008;299(6):637-645. doi:10.1001/jama.299.6.637PubMedGoogle ScholarCrossref
    41.
    Mercat  A, Richard  JC, Vielle  B,  et al; Expiratory Pressure (Express) Study Group.  Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: a randomized controlled trial.  JAMA. 2008;299(6):646-655. doi:10.1001/jama.299.6.646PubMedGoogle ScholarCrossref
    42.
    Talmor  D, Sarge  T, Malhotra  A,  et al.  Mechanical ventilation guided by esophageal pressure in acute lung injury.  N Engl J Med. 2008;359(20):2095-2104. doi:10.1056/NEJMoa0708638PubMedGoogle ScholarCrossref
    43.
    Huh  JW, Jung  H, Choi  HS, Hong  S-B, Lim  C-M, Koh  Y.  Efficacy of positive end-expiratory pressure titration after the alveolar recruitment manoeuvre in patients with acute respiratory distress syndrome.  Crit Care. 2009;13(1):R22. doi:10.1186/cc7725PubMedGoogle ScholarCrossref
    44.
    Xi  XM, Jiang  L, Zhu  B; RM group.  Clinical efficacy and safety of recruitment maneuver in patients with acute respiratory distress syndrome using low tidal volume ventilation: a multicenter randomized controlled clinical trial.  Chin Med J (Engl). 2010;123(21):3100-3105.PubMedGoogle Scholar
    45.
    Hodgson  CL, Tuxen  DV, Davies  AR,  et al.  A randomised controlled trial of an open lung strategy with staircase recruitment, titrated PEEP and targeted low airway pressures in patients with acute respiratory distress syndrome.  Crit Care. 2011;15(3):R133. doi:10.1186/cc10249PubMedGoogle ScholarCrossref
    46.
    Kacmarek  RM, Villar  J, Sulemanji  D,  et al; Open Lung Approach Network.  Open lung approach for the acute respiratory distress syndrome: a pilot, randomized controlled trial.  Crit Care Med. 2016;44(1):32-42. doi:10.1097/CCM.0000000000001383PubMedGoogle ScholarCrossref
    47.
    Park  KJ, Lee  YJ, Oh  YJ, Lee  KS, Sheen  SS, Hwang  SC.  Combined effects of inhaled nitric oxide and a recruitment maneuver in patients with acute respiratory distress syndrome.  Yonsei Med J. 2003;44(2):219-226. doi:10.3349/ymj.2003.44.2.219PubMedGoogle ScholarCrossref
    48.
    Papazian  L, Gainnier  M, Marin  V,  et al.  Comparison of prone positioning and high-frequency oscillatory ventilation in patients with acute respiratory distress syndrome.  Crit Care Med. 2005;33(10):2162-2171. doi:10.1097/01.CCM.0000181298.05474.2BPubMedGoogle ScholarCrossref
    49.
    Wang  C, Wang  X, Chi  C,  et al.  Lung ventilation strategies for acute respiratory distress syndrome: a systematic review and network meta-analysis.  Sci Rep. 2016;6:22855. doi:10.1038/srep22855PubMedGoogle ScholarCrossref
    50.
    Goligher  EC, Tomlinson  G, Hajage  D,  et al.  Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome and posterior probability of mortality benefit in a post hoc Bayesian analysis of a randomized clinical trial.  JAMA. 2018;320(21):2251-2259. doi:10.1001/jama.2018.14276PubMedGoogle ScholarCrossref
    51.
    Munshi  L, Walkey  A, Goligher  E, Pham  T, Uleryk  EM, Fan  E.  Venovenous extracorporeal membrane oxygenation for acute respiratory distress syndrome: a systematic review and meta-analysis.  Lancet Respir Med. 2019;7(2):163-172. doi:10.1016/S2213-2600(18)30452-1PubMedGoogle ScholarCrossref
    52.
    Gebistorf  F, Karam  O, Wetterslev  J, Afshari  A.  Inhaled nitric oxide for acute respiratory distress syndrome (ARDS) in children and adults.  Cochrane Database Syst Rev. 2016;(6):CD002787. doi:10.1002/14651858.CD002787.pub3PubMedGoogle Scholar
    53.
    Alhazzani  W, Alshahrani  M, Jaeschke  R,  et al.  Neuromuscular blocking agents in acute respiratory distress syndrome: a systematic review and meta-analysis of randomized controlled trials.  Crit Care. 2013;17(2):R43. doi:10.1186/cc12557PubMedGoogle ScholarCrossref
    54.
    Barbaro  RP, Odetola  FO, Kidwell  KM,  et al.  Association of hospital-level volume of extracorporeal membrane oxygenation cases and mortality: analysis of the extracorporeal life support organization registry.  Am J Respir Crit Care Med. 2015;191(8):894-901. doi:10.1164/rccm.201409-1634OCPubMedGoogle ScholarCrossref
    ×