Context & Rationale

• Background

  • Before PROSEVA, multiple RCTs of prone positioning in ARDS improved oxygenation but not survival, likely due to heterogeneous populations (including mild ARDS), shorter/prn proning sessions, and less stringent lung‑protective ventilation.
  • Meta‑analyses hinted that patients with severe hypoxaemia might benefit, creating equipoise for a focused, protocolised trial.
• Why This Matters

  • Prone positioning is low‑cost and physiologically compelling (homogenises stress/strain; improves V/Q matching), but definitive mortality benefit had not been shown in a high‑severity cohort managed with strict protective ventilation.
  • Demonstrating benefit would change routine ICU practice and guideline recommendations.
• Question

  • Does early, prolonged prone positioning, delivered within a lung‑protective ventilation strategy, reduce mortality in severe ARDS compared with continued supine ventilation?

Design & Methodology

Trial Design

• Design

  • Multicentre (26 ICUs France, 1 Spain), investigator‑initiated, stratified, open‑label RCT with concealed web‑based randomisation
• Setting

  • ICU

Population

• Inclusion Criteria

  • Adults intubated <36 h for ARDS (AECC criteria), with severe ARDS defined as PaO₂/FiO₂ <150 mmHg on FiO₂ ≥ 0.6 and PEEP ≥ 5 cmH₂O, ventilated with tidal volume ≈6 mL/kg predicted body weight; criteria reconfirmed after a 12–24 h stabilisation period
• Exclusion Criteria

  • Contra‑indications to proning, raised ICP, unstable fractures, recent abdominal surgery, Pregnancy, BMI > 40 kg m⁻² and others (see protocol).

Intervention

• Prone Group

  • Prone positioning within 1 h of randomisation; ≥16 consecutive hours per day up to day 28, with standardised turning procedures on usual ICU beds.
  • Ventilation: volume‑controlled, Vt target 6 mL/kg PBW; PEEP selected from an ARDSNet PEEP–FiO₂ table; plateau pressure goal ≤30 cmH₂O; arterial pH target 7.20–7.45.
  • Repeated daily until PaO₂/FiO₂ ≥ 150 mm Hg for ≥4 h while supine on PEEP ≤ 10 & FiO₂ ≤ 0.6

Control

• Supine Group

  • Semi‑recumbent supine care
  • Same ventilation targets and protocol.
  • Crossover to proning allowed only as rescue for life‑threatening hypoxaemia (strict physiologic criteria including P/F <55 on FiO₂ 1.0, maximal PEEP per table, plus inhaled NO and almitrine, and recruitment manoeuvres).

Statistical Plan

• Effect Size

  • 456 patients were required to detect a 15% absolute reduction in 28‑day mortality (from 60% to 45%) with 90% power at a one‑sided 5% alpha.
  • A single interim analysis at 50% enrolment had a 2.5% alpha; early stopping required ≥25% absolute mortality difference.
• Analysis

  • Primary analysis was intention‑to‑treat.

Other

• Blinding

  • Open‑label for clinicians
  • Outcome assessors blinded
  • Primary/secondary outcomes predominantly objective (mortality, time‑based outcomes), mitigating detection bias
• Follow Up

  • Outcomes at 28 and 90 days (mortality, extubation, ICU length of stay, VFDs), with serial physiology during the first week.

Key Results

• Early Stopping

  The trial did not stop early; DSMB recommended continuation at interim.

Primary Outcome

• 28 Day Mortality

  • 16.0% (38/237) vs 32.8% (75/229)
  • HR 0.3; 95% CI, 0.25 - 0.63; P<0.001
  • Adjusted HR 0.42; 95% CI, 0.26 - 0.66; P<0.001

Secondary Outcomes

• 90 Day Mortality

  • 23.6% (56/237) vs 41.0% (94/229)
  • HR 0.44; 95% CI, 0.29 – 0.67; P<0.001
  • Adjusted HR 0.48;  0.32 – 0.72; P<0.001
• Successful extubation by Day 90

  • 80.5% (186/231) vs 65.0% (145/223)
  • OR/HR 0.45; 95% CI, 0.29 – 0.70; P<0.001
• Ventilator‑Free Days (VFDs) to Day 28

  • 14 ± 9 vs 10 ± 10
  • Mean difference +4
  • P<0.001
• VFDs to Day 90 (days alive & off ventilator)

  • 57 ± 34 vs 43 ± 38
  • Mean difference +14
  • P<0.001
• Pneumothorax by Day 90

  • 6.3% (15/236) vs 5.7% (13/229)
  • OR, 0.89; 95% CI, 0.39 – 2.02; P=0.85
• Cardiac Arrest

  • 16 events vs 31 events
  • P=0.02
• Tracheostomy by Day 90

  • 6.4% (15/235) vs 8.1% (18/223)
  • OR, 0.78; 95% CI, 0.36 – 1.67; P=0.59

Notes

Internal Validity

• Randomisation & Allocation

  • Centralised, computer‑generated randomisation stratified by ICU; allocation concealment at assignment is described; outcome assessors were blinded.
  • Selection bias risk appears low
• Performance/Detection Bias

  • Clinicians were unblinded (inevitable), but primary outcomes were objective.
  • Outcome assessors were masked.
• Protocol Adherence

  • Time to first prone episode 55 ± 55 min
  • Sessions 4 ± 4 per patient; 17 ± 3 h/session; 73% of 22,334 patient‑hours in prone during the active window vs 0 % in controls
• Outcome Assessment

  • Mortality and VFDs are hard endpoints
  • Extubation success prespecified
• Statistical Rigor

  • Predefined ITT analysis; alpha spending and interim plan prespecified
  • Adjusted analyses (for SOFA, vasopressors, NMB) confirmed robustness
  • Power target met/exceeded
• Separation of the Variable of Interest

  • Pronated within 55 ± 55 min; mean sessions 4 ± 4; 17 ± 3 h/session; patients spent 73% of “prone‑period” ICU time actually prone.
  • Supine arm remained supine except for tightly restricted rescue criteria.
  • This is excellent separation of the exposure
• Key Delivery Aspects

  • Implementation hinged on nurse and physician comfort with sedation breaks. The protocol required vigilance to ensure patient safety and comfort.
• Drop Outs or Exclusions

  • 8 (<2%) excluded post‑randomisation for prespecified reasons;
  • 466 analysed by intention‑to‑treat.
  • Loss to follow‑up not material.
• Baseline Characteristics

  • Groups broadly similar for lung mechanics and gas exchange at enrolment (e.g., Vt 6.1 ± 0.6 vs 6.1 ± 0.6 mL/kg PBW; PEEP 10 ± 4 vs 10 ± 3 cmH₂O; FiO₂ 0.79 ± 0.16 vs 0.79 ± 0.16; Pplat 23 ± 5 vs 24 ± 5 cmH₂O; PaO₂/FiO₂ 100 ± 20 vs 100 ± 30 mmHg).
  • There were differences in SOFA score and the proportions receiving vasopressors and neuromuscular blockers at inclusion, addressed in adjusted analyses.
• Heterogeneity

  • Multicentre, but all sites were experienced with proning
  • Ventilatory management was protocolised (PEEP–FiO₂ table, Vt target), limiting practice heterogeneity.
• Timing

  • Enrolment was early (mean 31–33 h after intubation) and proning began within ~1 h, consistent with the hypothesised window to limit VILI.
• Dose

  • ≥16 h/day; proning days up to day 28
  • Plateau pressure maintained ≤30 cmH₂O
  • By day 3, PEEP and FiO₂ were lower and Pplat ~2 cmH₂O lower in the prone arm, consistent with effective “dose” and physiological effect
• Adjunctive Therapy Use

  • Rescue therapies used more often in supine (e.g., inhaled NO 15.7% vs 9.7%; almitrine 6.6% vs 2.5%), suggesting control clinicians escalated adjuncts rather than cross over to proning
  • Unlikely to bias in favour of prone
• Blinding

  • Not possible
  • Outcomes objective (mortality, days on ventilator, LOS) → minimal detection bias
• Crossover

  • Supine ➜ Prone (rescue)
    • 17 / 229 controls (7.4 %)
    • Strict “life‑threatening hypoxaemia” bundle: PaO₂/FiO₂; 55 mm Hg on FiO₂ 1.0 despite maximal PEEP, inhaled nitric oxide 10 ppm, i.v.
  • Prone ➜ Supine‑only
    • None reported †
    • Not permitted by protocol; prone sessions could be terminated early for complications but patients remained in the intervention arm
  • Magnitude of crossover is small
    • Only 17 rescue crossovers in the control arm means the primary 28‑ and 90‑day mortality results are unlikely to be driven by differential contamination.
  • Direction favours conservatism.
    • Because crossovers moved from the control to the intervention but were still analysed as controls, any benefit they derived from proning would diminish, not exaggerate, the mortality separation.
  • Residual bias remains possible.
    • These 17 patients were by definition the sickest control patients; if rescue proning failed to reverse impending death, their inclusion in the control arm could magnify the apparent benefit.
    • Sensitivity analyses excluding them were not published, so a minor influence cannot be excluded—but the effect size (15 % absolute risk reduction) is far larger than what 17 patients could plausibly explain.
  • Overall, crossover in PROSEVA was minimal, unidirectional, protocol‑driven, and analytically conservative, reinforcing confidence that the pronounced survival benefit reflects the intervention rather than cross‑contamination.

• Conclusion

  Strong - Allocation was robust; separation excellent; outcomes objective; results consistent in adjusted analyses.

External Validity

• Population Representativeness

  • Adults with severe ARDS (P/F < 150 on high FiO₂ and PEEP) early after intubation—typical of severe ICU ARDS
  • However, centres were highly experienced with proning and followed a rigorous protocol
• Applicability

  • Findings generalise to ICUs capable of (i) early identification of severe ARDS, (ii) safe proning ≥16 h/day, and (iii) strict lung‑protective ventilation.
  • Adoption has been variable internationally
    • In LUNG SAFE (2014), proning was used in only ~16% of severe ARDS, reflecting system‑level constraints rather than biological limits.

• Conclusion

  • Good - for comparable ICUs with training and staffing to deliver prolonged proning within lung‑protective ventilation
  • Generalisability may be limited where such resources or adherence are lacking.

Strengths & Limitations

Strengths

• Strong Methodology

  • Early enrolment;
  • Severe ARDS focus
  • Strict lung‑protective ventilation
  • Prolonged daily proning
  • Excellent exposure separation
  • Multicentre
  • Blinded outcome assessment
  • Prespecified analyses with sufficient power

Limitations

• Few, not critical

  • Open‑label care
  • Experienced centres (learning‑curve effect limits generalisability)
  • Small baseline imbalances (SOFA, vasopressors/NMB) though adjusted analyses align
  • Success may depend on protocol fidelity and staffing skill mix

Interpretation / Why This Matters

• Transformational Trial

  • PROSEVA demonstrates that how and to whom proning is applied is crucial: early, prolonged sessions in severe ARDS, within a lung‑protective strategy, reduce mortality and improve liberation from ventilation.
  • It confirms and operationalises prior physiological rationale into practice‑changing evidence.
  • It also provides the largest, robust mortality impact (~50% reduction) of all critical care interventions.

Controversies & Subsequent Evidence

• Methodology / Interpretation Debates

  • Questions centred on generalisability beyond expert centres and whether co‑interventions (e.g., neuromuscular blockade) or lower Pplat in the prone group explained the effect.
  • The protocolised management, adjusted analyses, and concordant physiological signals argue the survival benefit is attributable to proning.
• Relation to adjuncts

  • Early NMB was common in both groups (trial‑level), and prior RCT evidence (ACURASYS) suggested benefit in severe ARDS
  • The later ROSE trial (2019) did not confirm routine early NMB for all
  • Proning’s benefit in PROSEVA stands independent of mandated NMB
• Meta Analyses

  • Post‑PROSEVA summaries show mortality reduction with proning in severe ARDS, especially with low Vt and ≥12–16 h sessions, aligning with PROSEVA’s protocol.
  • Cochrane and CMAJ reviews support benefit when protective ventilation is used.
• Guidelines

  • The 2017 ATS/ESICM/SCCM recommended proning >12 h/day in severe ARDS
  • The 2023 ESICM and 2024 ATS guidelines continue to endorse proning for moderate‑to‑severe ARDS
• Guidelines Stance

Summary

• Dutration of Ventilation Reduction

  • Early, prolonged prone positioning (≥16 h/day) halved 28‑day mortality (16.0% vs 32.8%) and reduced 90‑day mortality (23.6% vs 41.0%).
  • Pron­ing increased ventilator‑free days and successful extubation without excess serious complications; cardiac arrests were fewer vs supine.
  • Separation was excellent (first session within ~1 h; 17 h/session; ~73% of active period actually prone).
  • Benefit was observed within a strict lung‑protective ventilation protocol and experienced centres—critical for replication and implementation.

Conclusion

• Landmark Trial

  The Kress sedation break trial changed how sedation was delivered worldwide and lead to a series of trials seeking to optimise this new paradigm of lighter sedation

Further Reading

Other Trials

• Pre-PROSEVA

  • Gattinoni L. Effect of prone positioning on survival of acute respiratory failure. N Engl J Med 2001;345:568‑573
  • Guérin C, et al. Effects of Systematic Prone Positioning in Hypoxaemic Acute Respiratory Failure: A Randomized Controlled Trial. JAMA. 2004;292:2379‑2387
  • Mancebo J, et al. A Multicenter Trial of Prolonged Prone Ventilation in Severe ARDS. Am J Respir Crit Care Med. 2006;173:1233‑1239
  • Taccone P, et al. Prone Positioning in Patients with Moderate and Severe ARDS: A Randomized Controlled Trial. JAMA. 2009;302:1977‑1984
• Post-PROSEVA

  • Alhazzani. Effect of Awake Prone Positioning on Endotracheal Intubation in Patients With COVID-19 and Acute Respiratory Failure (COVI-PRONE). JAMA 2022;327;(21):2104-2113

Systematic Review & Meta Analysis

• Intubated

  • Sud S, et al. Effect of Prone Positioning During Mechanical Ventilation on Mortality among ARDS Patients: Systematic Review & Meta‑analysis. CMAJ. 2014;186(10):E381‑E390
    
  • Bloomfield R, et al. Prone Position for Acute Respiratory Failure in Adults. Cochrane Database Syst Rev. 2015;(11):CD008095
    
  • Munshi L, et al. Prone Position for ARDS: Systematic Review & Meta‑analysis. Ann Am Thorac Soc. 2017;14(Suppl 4):S280‑S288
    
  • Park M, et al. Multivariate Meta‑analysis of Mortality Effect of Prone Positioning in ARDS. J Intensive Care Med. 2021;36(11):1323-1330
• Non-Intubated

  Luo. Awake Prone Positioning in Adults With COVID-19. An Individual Participant Data Meta-Analysis. JAMA Intern Med 2025;185;(5):572-581

Observational Studies

• Global Epidemiology

  • Bellani G, et al.; LUNG SAFE Investigators. Epidemiology, Patterns of Care, and Mortality for ARDS in 50 Countries. JAMA. 2016;315:788‑800

Guidelines

• Brazilian

  • Brazilian medical societies. Joint statement on evidence-based practices in mechanical ventilation: suggestions from two Brazilian medical societies. Crit Care Sci 2025;37:e20250242en
• American

  • An Update on Management of Adult Patients with Acute Respiratory Distress Syndrome. An Official American Thoracic Society Clinical Practice Guideline. Am J Respir Crit Care Med 2024;209(1):24–36
  • American Association for the Surgery of Trauma/American College of Surgeons Committee on Trauma Clinical Protocol for Management of Acute Respiratory Distress Syndrome and Severe Hypoxemia. J Trauma Acute Care Surg 2023;epublished June 12th
• European

  • Grasselli G, Calfee CS, Camporota L, et al. ESICM Guidelines on ARDS: Definition, Phenotyping and Respiratory Support Strategies. Intensive Care Med. 2023;49:727‑759
• Japanese

  • Tasaka. ARDS Clinical Practice Guideline 2021. J Intensive Care 2022;10:32
• French

  • Papazian. (French) Formal guidelines: management of acute respiratory distress syndrome. Ann Intensive Care 2019;9:69
• British

  • Griffiths. Guidelines on the management of acute respiratory distress syndrome. BMJ Open Respiratory Research 2019;6:e000420
    
• German

  • Fichtner. Mechanical Ventilation and Extracorporeal Membrane Oxygenation in Acute Respiratory Insufficiency. Dtsch Arztebl Int 2018;115(50):840-7