Publication

• Title: Prone Positioning in Severe Acute Respiratory Distress Syndrome
• Acronym: PROSEVA
• Year: 2013
• Journal published in: The New England Journal of Medicine
• Citation: Guérin C, Reignier J, Richard J-C, et al. Prone positioning in severe acute respiratory distress syndrome. N Engl J Med. 2013;368(23):2159-2168.

Context & Rationale

• Background

  • Prone positioning was well established to improve oxygenation in ARDS, plausibly by redistributing transpulmonary pressures and ventilation–perfusion relationships, and by reducing ventral overdistension with dorsal atelectrauma.
  • Earlier randomised trials in acute hypoxaemic respiratory failure/ARDS generally improved oxygenation but did not demonstrate a consistent mortality benefit; many enrolled patients with less severe hypoxaemia and/or used shorter daily prone “doses”, often before low tidal volume ventilation became standard.
  • It remained uncertain whether applying prone positioning early, for prolonged sessions, in a clearly defined severe ARDS population receiving lung-protective ventilation could deliver a survival benefit.
• Research Question/Hypothesis

  • In invasively ventilated adults with severe ARDS, does early initiation of prolonged prone positioning sessions (target ≥16 consecutive hours/day) reduce 28-day mortality compared with continued supine (semi-recumbent) positioning under a lung-protective ventilation strategy?
• Why This Matters

  • Prone positioning is a low-cost intervention but is labour-intensive and requires training and protocols; a robust mortality benefit would justify broad implementation.
  • Clarifying the severity threshold, timing, and “dose” needed for benefit would inform guidelines and standardise practice for a high-mortality syndrome.

Design & Methods

• Research Question: Whether early, prolonged prone positioning reduces 28-day mortality in severe ARDS compared with supine management under lung-protective ventilation.
• Study Type: Randomised, multicentre, investigator-initiated, parallel-group, open-label trial in ICUs; central computer-generated allocation stratified by ICU (27 ICUs: 26 in France, 1 in Spain).
• Population:
  • Adults with ARDS (acute onset, bilateral infiltrates, PaO₂/FiO₂ ≤300 mmHg, no evidence of left atrial hypertension), intubated and invasively ventilated for <36 hours.
  • Severe ARDS confirmed after a 12–24 hour stabilisation period: PaO₂/FiO₂ <150 mmHg with FiO₂ ≥0.60 and PEEP ≥5 cm H₂O, with tidal volume targeted at ~6 mL/kg predicted body weight.
  • Key exclusions (examples): age <18 years or pregnancy; contraindication to prone positioning (e.g., unstable spine, raised intracranial pressure, recent tracheal surgery, massive haemoptysis); chronic respiratory failure requiring long-term oxygen or home NIV; ECMO at screening; and anticipated end-of-life decision.
• Intervention:
  • Prone positioning started as soon as possible after randomisation (mean 55 ± 55 minutes) and delivered in sessions of ≥16 consecutive hours (mean 17 ± 3 hours), repeated daily until pre-specified stopping criteria or day 28.
  • Lung-protective ventilation in volume-controlled mode: tidal volume 6 mL/kg predicted body weight; plateau pressure ≤30 cm H₂O; PEEP/FiO₂ table; permissive hypercapnia with target pH 7.20–7.45.
  • Deep sedation targeted early; neuromuscular blockade was strongly recommended during the first 48 hours (and during prone sessions) to facilitate protective ventilation and proning.
• Comparison:
  • Supine (semi-recumbent) positioning with otherwise identical protocolised lung-protective ventilation and supportive care.
  • Crossover to prone positioning prohibited except as protocolised rescue for life-threatening hypoxaemia after optimisation and rescue measures (including FiO₂ 1.0, maximal PEEP per table, inhaled nitric oxide 10 ppm, almitrine 4 µg/kg/min, and recruitment manoeuvres).
• Blinding: Unblinded (positioning not feasibly blindable); mortality objective, but co-interventions and liberation-from-ventilation decisions could be influenced by lack of blinding.
• Statistics: A total of 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% significance level; interim analysis planned after 50% enrolment with two one-sided analyses at the 2.5% level; analysis described as intention-to-treat with Cox proportional hazards models stratified by centre for time-to-event outcomes and regression models for other outcomes.
• Follow-Up Period: 90 days (primary endpoint at day 28; vital status and key outcomes assessed to day 90).

Key Results

This trial was not stopped early. Recruitment continued to the planned sample size; 466 patients were analysed (237 prone; 229 supine) with follow-up to 90 days.

Outcome	Prone position	Supine position	Effect	p value / 95% CI	Notes
All-cause mortality (day 28)	38/237 (16.0%; 95% CI 11.3 to 20.7)	75/229 (32.8%; 95% CI 26.4 to 38.6)	HR 0.39	95% CI 0.25 to 0.63; P<0.001	Primary endpoint; adjusted HR 0.42; 95% CI 0.26 to 0.66; P<0.001 (as reported).
All-cause mortality (day 90)	56/237 (23.6%; 95% CI 18.2 to 29.0)	94/229 (41.0%; 95% CI 34.6 to 47.4)	HR 0.44	95% CI 0.29 to 0.67; P<0.001	Adjusted HR 0.48; 95% CI 0.32 to 0.72; P<0.001 (as reported).
Successful extubation (day 90)	186/231 (80.5%; 95% CI 75.4 to 85.6)	145/223 (65.0%; 95% CI 58.7 to 71.3)	HR 0.45	95% CI 0.29 to 0.70; P<0.001	Defined as extubation without re-intubation or NIV within 48 hours (as reported).
Ventilator-free days (day 28)	14 ± 9	10 ± 10	Not reported	P<0.001	Mean ± SD (higher is better).
Ventilator-free days (day 90)	57 ± 34	43 ± 38	Not reported	P<0.001	Mean ± SD (higher is better).
Pneumothorax (to day 90)	15/237 (6.3%; 95% CI 4.9 to 7.7)	13/229 (5.7%; 95% CI 3.9 to 7.5)	OR 0.89	95% CI 0.39 to 2.02; P=0.85	Odds ratios reported for this outcome.
Cardiac arrest (adverse event)	16/237 (6.8%)	31/229 (13.5%)	Not reported	P=0.02	No CI reported for between-group comparison.

• Mortality was lower with prone positioning at day 28 (16.0% vs 32.8%; HR 0.39; 95% CI 0.25 to 0.63; P<0.001) and remained lower at day 90 (23.6% vs 41.0%; HR 0.44; 95% CI 0.29 to 0.67; P<0.001).
• Successful extubation by day 90 was higher with prone positioning (80.5% vs 65.0%; HR 0.45; 95% CI 0.29 to 0.70; P<0.001) with more ventilator-free days at day 28 (14 ± 9 vs 10 ± 10; P<0.001).
• Major complications were uncommon; pneumothorax rates were similar (6.3% vs 5.7%; OR 0.89; 95% CI 0.39 to 2.02; P=0.85), while cardiac arrest occurred more often in the supine group (13.5% vs 6.8%; P=0.02).

Internal Validity

• Randomisation and allocation: Computer-generated randomisation stratified by ICU with central web-based allocation supports allocation concealment and minimises selection bias.
• Drop out or exclusions: 474 patients were randomised; 8 were excluded post-randomisation for eligibility/procedural reasons, leaving 466 in the analysed population (237 prone; 229 supine), consistent with a modified intention-to-treat approach; 90-day vital status was reported for the analysed cohort.
• Performance/detection bias: Unblinded positioning could influence co-interventions and extubation practices; the primary endpoint (mortality) is objective and less vulnerable to ascertainment bias.
• Protocol adherence and separation: First prone session occurred 55 ± 55 minutes after randomisation; mean number of sessions 4 ± 4 per patient; mean session duration 17 ± 3 hours; prone positioning comprised 73% of total 22,334 patient-hours in ICU for the prone group.
• Crossover: 17/229 (7.4%) in the supine group crossed over to prone positioning for refractory hypoxaemia per protocolised rescue criteria; this direction of crossover would be expected to bias estimates towards the null.
• Baseline comparability: Severe ARDS at inclusion (PaO₂/FiO₂ 100 ± 20 prone vs 100 ± 30 supine; FiO₂ 0.74 ± 0.18 vs 0.74 ± 0.17; PEEP 10 ± 3 vs 10 ± 3 cm H₂O; tidal volume 6.1 ± 0.6 vs 6.1 ± 0.6 mL/kg predicted body weight); illness severity broadly similar (SAPS II 45 ± 15 vs 47 ± 17; SOFA 9.6 ± 3.2 vs 10.4 ± 3.4).
• Timing and dose: Randomisation occurred after a 12–24 hour stabilisation period and within 36 hours of mechanical ventilation; time from intubation to randomisation was 33 ± 24 hours (prone) vs 31 ± 26 hours (supine); sessions were long-duration (≥16 hours; mean 17 hours) and repeated.
• Adjunctive therapy use: High use of vasopressors (72.6% prone vs 83.0% supine) and neuromuscular blockers (91.0% prone vs 82.3% supine) at inclusion, with low but present use of inhaled nitric oxide (11.0% vs 14.4%), almitrine (3.8% vs 2.6%), and ECMO (1.7% vs 2.2%).
• Outcome assessment and statistical rigour: Mortality endpoints were analysed with Kaplan–Meier/log-rank methods and Cox regression stratified by centre; target sample size was achieved (466 analysed vs 456 required by the trialists’ calculation) with large and consistent effect estimates for mortality.

Conclusion on Internal Validity: Overall, internal validity appears strong, supported by centralised stratified randomisation, objective primary outcome assessment, and excellent separation in prone “dose”; limitations include open-label care and small post-randomisation exclusions consistent with a modified intention-to-treat population.

External Validity

• Population representativeness: Participants were a selected severe ARDS subset (466 analysed from 3,449 ARDS cases screened for eligibility), with exclusions that remove patients with common proning barriers (e.g., major contraindications, chronic respiratory failure).
• Setting and expertise: Conducted in 27 ICUs (predominantly France) where all participating units reported routine prone positioning use for >5 years, which may not reflect implementation conditions in less experienced ICUs.
• Applicability: Findings are most applicable to invasively ventilated adults with severe ARDS receiving contemporary lung-protective ventilation, where a trained team can reliably deliver prolonged prone sessions and manage associated sedation, neuromuscular blockade, and line/airway risks.
• Resource-limited settings: The intervention is equipment-light but staff- and process-intensive; generalisability depends on staffing, training, and the ability to maintain safety during turns and prolonged sessions.

Conclusion on External Validity: Generalisability is moderate: PROSEVA provides high-confidence evidence for early prolonged proning in severe ARDS, but translation to milder ARDS, units without established proning expertise, or settings with limited staffing remains less certain.

Strengths & Limitations

• Strengths:
  • Clear severe ARDS phenotype with protocolised stabilisation before randomisation.
  • Early initiation and high “dose” of prone positioning with excellent protocol adherence (mean 17-hour sessions; repeated daily).
  • Contemporary lung-protective ventilation in both groups, supporting isolation of the positioning effect.
  • Objective primary endpoint with consistent mortality benefit at day 28 and day 90.
• Limitations:
  • Open-label design introduces potential performance bias for co-interventions and weaning decisions.
  • Modified intention-to-treat analysis (8 post-randomisation exclusions).
  • Highly experienced centres and selected population may limit real-world reproducibility without robust training and implementation programmes.
  • Rescue crossover to prone in the control arm (7.4%) and co-intervention imbalances (e.g., neuromuscular blockade, vasopressors) complicate causal attribution in pragmatic deployment.

Interpretation & Why It Matters

• Clinical implications

  • In severe ARDS managed with lung-protective ventilation, early prolonged prone positioning was associated with substantially lower mortality and more ventilator-free days.
  • The trial operationalised proning as an early, protocolised strategy rather than a late “rescue-only” manoeuvre, emphasising timing and duration as core determinants of effect.
  • Implementation requires system readiness: trained teams, deep sedation/neuromuscular blockade strategies where appropriate, and safety processes for airway and line management during turns and prolonged sessions.

Controversies & Subsequent Evidence

• Generalisability: the accompanying editorial highlighted that PROSEVA was conducted in ICUs with substantial prior experience in prone positioning, raising implementation concerns about achieving similar outcomes in less experienced units.¹
• Magnitude of benefit: the mortality effect size was larger than earlier proning trials, and the editorial contextualised this against differences in patient selection (more severe ARDS), earlier initiation, and longer daily prone exposure.¹
• Physiology and practicalities: published correspondence raised issues including the interpretation of PaCO₂ changes and the logistics of cardiopulmonary resuscitation in the prone position, underscoring the need for safety protocols and team training.²
• Subsequent evidence synthesis: post-PROSEVA meta-analysis in the low tidal volume era reported reduced mortality with prone positioning, with benefit concentrated in more severe ARDS and with longer prone sessions.³
• Guideline incorporation: international clinical practice guidelines after PROSEVA recommend prone positioning as a core therapy for moderate-to-severe/severe ARDS when deliverable safely by trained teams.⁴⁵

Summary

• In severe ARDS (PaO₂/FiO₂ <150 mmHg with high FiO₂ and PEEP), early prolonged prone positioning reduced 28-day mortality compared with supine management.
• Mortality benefit persisted to day 90, with higher successful extubation and more ventilator-free days in the prone group.
• Protocol delivery was high-dose and early (first session within 55 ± 55 minutes; mean 4 ± 4 sessions of 17 ± 3 hours; 73% of ICU time prone in the intervention group).
• Crossover from control to prone occurred in 7.4% (rescue for refractory hypoxaemia), tending to dilute between-group differences.
• Major complications were uncommon; pneumothorax was similar between groups, while cardiac arrest occurred more frequently in the supine group.

Overall Takeaway

PROSEVA established that, in a clearly defined severe ARDS population managed with lung-protective ventilation, early and prolonged prone positioning can substantially improve survival and key patient-centred outcomes. The trial reframed proning from an oxygenation “rescue” manoeuvre into a time-sensitive, protocolised, high-dose intervention, shaping modern guideline-based ARDS management and ICU implementation priorities.

Bibliography

• 1.Soo Hoo GW. In prone ventilation, one good turn deserves another. N Engl J Med. 2013;368(23):2227-2228. Link
• 2.Möller M, Neuzner J, Gradaus R, Wiedemann S, Strasser RH, Simonis G, et al. Prone positioning in the acute respiratory distress syndrome. N Engl J Med. 2013;369(10):979-980. Link
• 3.Sud S, Friedrich JO, Adhikari NKJ, et al. Effect of prone positioning during mechanical ventilation on mortality among patients with acute respiratory distress syndrome: a systematic review and meta-analysis. CMAJ. 2014;186(10):E381-E390. Link
• 4.Fan E, Del Sorbo L, Goligher EC, et al. 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. Link
• 5.Qadir N, Sahetya SK, Munshi L, et al. 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. Link