Publication

• Title: Proportional-Assist Ventilation for Minimizing the Duration of Mechanical Ventilation
• Acronym: PROMIZING
• Year: 2025
• Journal published in: The New England Journal of Medicine
• Citation: Bosma KJ, Burns KEA, Martin CM, et al. Proportional-assist ventilation for minimizing the duration of mechanical ventilation. N Engl J Med. 2025;393:1088-1103.

Context & Rationale

• Background

  • Prolonged invasive mechanical ventilation (IMV) is associated with diaphragm disuse and weakness, patient–ventilator asynchrony, delirium, ventilator-associated complications, and longer ICU/hospital stay.
  • Pressure-support ventilation (PSV) is the standard assisted mode during weaning, but fixed pressure support can over-assist (diaphragm unloading/disuse) or under-assist (excessive respiratory effort), both of which can plausibly delay liberation.
  • Proportional assist ventilation with load-adjustable gain factors (PAV+) provides assist proportional to patient effort (and estimates respiratory system mechanics), aiming to improve synchrony and maintain a physiologic level of respiratory muscle activity.
  • Before PROMIZING, outcome-focused evidence was dominated by small trials and physiological studies, creating uncertainty about whether PAV+ meaningfully reduces IMV duration at scale.
• Research Question/Hypothesis

  • In invasively ventilated adults who were not yet ready for unassisted breathing trials (or had recently failed), does a structured weaning strategy using PAV+ shorten time to successful liberation from IMV compared with a structured PSV strategy?
  • Hypothesis: PAV+ would reduce duration of IMV by better matching ventilatory assistance to patient demand and preserving respiratory muscle function.
• Why This Matters

  • Even modest reductions in IMV duration can translate into fewer ventilator-associated complications, less exposure to sedatives, and improved ICU throughput.
  • PAV+ requires specific ventilator capability and clinician expertise; establishing patient-centred benefit (or lack thereof) is critical before widespread adoption.
  • A definitive pragmatic trial could recalibrate interpretation of prior smaller studies and inform future guidelines and implementation strategies.

Design & Methods

• Research Question: Among invasively ventilated adults requiring assisted ventilation, does PAV+ (targeting a predefined range of respiratory muscle pressure) reduce time to successful liberation from IMV versus PSV (targeting standard respiratory rate and tidal volume ranges)?
• Study Type: Multicentre, parallel-group, investigator-initiated, international, open-label, randomised controlled trial conducted in 23 ICUs across 7 countries; allocation stratified by site with variable block sizes; pragmatic co-interventions with protocolised weaning algorithms.
• Population:
  • Setting: adult ICUs; enrolled after ≥24 hours of invasive ventilation and when transitioning (or attempting to transition) to assisted ventilation.
  • Key inclusion features (operational): invasively ventilated; able to trigger breaths; requiring assisted ventilation with planned weaning attempts; not yet meeting criteria for unassisted breathing (or had recently failed a weaning assessment).
  • Key exclusions (operational): inability to trigger; clinical scenarios requiring full control or incompatible with protocolised assisted ventilation; other protocol-specified exclusions (including those discovered post-randomisation leading to early discontinuation of study ventilation in rare cases).
• Intervention:
  • PAV+ (load-adjustable gain factors) delivered as the primary assisted mode during weaning.
  • Ventilator gain adjusted to target peak respiratory muscle pressure (Pmus) 5–10 cm H₂O (with algorithmic responses to hypoventilation, hyperventilation, and respiratory distress).
  • Protocol permitted switching to assist-control ventilation if maximum protocol settings were reached or if clinical instability required it.
  • Daily screening for weaning readiness; spontaneous breathing trials (SBTs) performed on a T-piece with no ventilatory support; extubation intended within 2 hours after passing SBT (per protocol workflow).
• Comparison:
  • PSV delivered as the primary assisted mode during weaning.
  • PSV adjusted to maintain respiratory rate 12–35 breaths/min and tidal volume 5–10 mL/kg predicted body weight (PBW), with protocolised responses to hypoventilation, hyperventilation, and respiratory distress.
  • Same daily screening for weaning readiness and T-piece SBT/extubation workflow as the intervention group.
• Blinding: Open-label (clinicians and research staff not blinded); trial statistician and data and safety monitoring board were blinded; the primary outcome was objective (time to successful liberation) but clinician behaviours (sedation adjustment, timing of extubation, rescue mode selection) could still be influenced by lack of blinding.
• Statistics:
  • Power calculation: a minimum of 558 patients were required to detect a hazard ratio of 1.30 for successful liberation (80% power; two-sided α=0.05), corresponding to a reduction in median time to liberation from 7.70 to 5.92 days.
  • Primary analysis: intention-to-treat time-to-event analysis treating death as a competing risk (Fine–Gray); cause-specific Cox models used as sensitivity analyses.
  • Published protocol available.¹
• Follow-Up Period: Outcomes reported through day 28 (e.g., ventilator-free days) and day 90 (vital status); the primary endpoint was time to successful liberation defined by sustained freedom from invasive ventilation (≥7 consecutive days).

Key Results

This trial was not stopped early. Recruitment and follow-up were completed as planned; no interim analyses were conducted.

• PAV+ did not shorten time to successful liberation from IMV compared with PSV (HR 0.96; 95% CI 0.80 to 1.15).
• Safety and major clinical outcomes (reintubation, tracheostomy, 90-day mortality) were broadly similar between groups, with wide confidence intervals for several outcomes.
• PAV+ patients more frequently required assist-control ventilation at least once (58.9% vs 49.7%), suggesting either greater clinical instability, greater clinician concern, or protocol-driven rescue thresholds in the PAV+ strategy.

Internal Validity

• Randomisation and Allocation: Central, secure web-based randomisation with site stratification and variable block sizes; this supports allocation concealment and reduces selection bias.
• Drop out or exclusions:
  • Primary outcome data were available for 570/573 randomised patients (99.5%); day 90 vital status was available for 572/573 (99.8%).
  • Three patients withdrew consent and were censored at last contact (minimal impact on primary time-to-event endpoint).
  • Early discontinuation of study ventilation occurred more often in the PAV+ group (27/287) than the PSV group (9/286), predominantly due to transfer from ICU, protocol violations, withdrawal, discovery of exclusion criteria, or other administrative reasons.
• Performance/Detection Bias:
  • Open-label delivery introduces risk that clinicians altered sedation, extubation timing, or rescue strategies differently by group.
  • The primary endpoint (time to sustained liberation) is relatively objective, but some components remain clinician-mediated (timing of SBT initiation, extubation decision-making, rescue intubation thresholds).
• Protocol Adherence:
  • Non-protocol ventilation modes were used on 169/6795 trial days (2.5%), indicating generally high adherence.
  • In the PAV+ group, Pmus was within target (5–10 cm H₂O) on 80.3% of trial days; Pmus >10 cm H₂O on 12.8% and <5 cm H₂O on 7.0% of trial days (suggesting the “dose” of intervention was mostly achieved but not universally).
• Baseline Characteristics: Groups were broadly similar at baseline; patients were in a “not-yet-weanable or recently failed weaning” state, with substantial illness burden and a clinically meaningful window for a weaning-mode intervention.
• Heterogeneity: Multinational ICU mix improves representativeness; heterogeneity in case-mix and sedation/weaning culture could dilute small effects but also enhances pragmatic relevance.
• Timing: Median duration of IMV before randomisation was 4.8 vs 5.0 days (PAV+ vs PSV), consistent with targeting a clinically relevant weaning phase rather than very early ventilation.
• Dose: PAV+ gain targeting Pmus 5–10 cm H₂O operationalises an intermediate work-of-breathing target; adherence data suggest this target was largely met, supporting dose fidelity.
• Separation of the Variable of Interest:
  • At 3–6 hours after randomisation: Pmus 7.7 vs 7.0 cm H₂O (difference 0.7; 95% CI 0.1 to 1.3; P=0.02); work of breathing 5.5 vs 4.6 J/min (difference 0.9; 95% CI 0.2 to 1.6; P=0.01); estimated assistance 60.6% vs 64.0% (difference −3.4%; 95% CI −5.7 to −1.1; P=0.003).
  • At 24 hours: Pmus 7.5 vs 6.7 cm H₂O (difference 0.8; 95% CI 0.2 to 1.5; P=0.01); work of breathing 5.8 vs 4.6 J/min (difference 1.2; 95% CI 0.3 to 2.1; P=0.01); estimated assistance 62.6% vs 65.6% (difference −3.0%; 95% CI −5.1 to −0.8; P=0.007).
  • These data confirm physiologic separation: PAV+ produced higher patient effort and slightly lower estimated ventilator assistance than PSV.
• Key Delivery Aspects:
  • Rescue to assist-control was more frequent in the PAV+ group (58.9% vs 49.7% at least once), and median duration of assist-control on study ventilation was 0.44 vs 0.00 days (difference 0.44; 95% CI 0.03 to 0.67).
  • This rescue pattern may have reduced sustained exposure to the intended assisted mode (and therefore could attenuate any treatment effect attributable to PAV+).
• Outcome Assessment: Primary outcome definition (time to sustained liberation with a 7-day criterion) reduces noise from short-term extubation failures and recaptures clinically meaningful weaning success.
• Statistical Rigor: Competing-risk framework is appropriate (death precludes liberation); primary analysis aligned with design; sample-size target was achieved and effect estimate precision was clinically informative.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong given robust randomisation, minimal missing data, and objective outcome definition; key threats are open-label delivery and differential rescue/early discontinuation of study ventilation, which could dilute a modest true effect.

External Validity

• Population Representativeness: Adult ICU patients in the weaning phase (median ~5 days ventilated at randomisation), including substantial medical–surgical heterogeneity, improving applicability to typical tertiary ICUs.
• Applicability:
  • Intervention requires ventilators capable of PAV+ and clinicians trained in interpreting/targeting effort-based metrics (Pmus), which may limit applicability in resource-limited settings or units without PAV+ availability.
  • The structured algorithms for both arms may exceed “usual care” in settings without protocolised weaning; the marginal benefit of any mode may be smaller when both groups receive high-quality weaning processes.
  • Results are most directly generalisable to ICUs with established weaning workflows and capability to deliver either strategy with adherence similar to the trial.

Conclusion on External Validity: Generalisability is good for well-resourced ICUs delivering protocolised weaning, but more limited for settings lacking PAV+ technology or expertise, and for environments where usual care is less standardised than in this trial.

Strengths & Limitations

• Strengths:
  • Largest pragmatic randomised evaluation of PAV+ versus PSV focused on patient-centred weaning outcomes.
  • International, multicentre design with high follow-up completeness and broadly representative ICU case-mix.
  • Protocolised delivery in both groups reduced confounding by variable weaning practices and ensured meaningful separation in respiratory effort targets.
  • Appropriate competing-risk primary analysis for death as a non-ignorable intercurrent event.
• Limitations:
  • Open-label delivery risks performance bias (sedation, extubation timing, and rescue strategies are clinician-mediated).
  • Differential use of rescue assist-control ventilation and higher early discontinuation of study ventilation in the PAV+ group may have reduced exposure to the intended assisted mode.
  • Effect estimates for some outcomes (e.g., 90-day mortality) were imprecise, and clinically meaningful benefit or harm cannot be fully excluded.
  • Implementation complexity (PAV+ tuning and interpretation) may affect reproducibility outside experienced centres.

Interpretation & Why It Matters

• Clinical implication

  PAV+ should not be adopted solely to shorten IMV duration in broadly defined ICU weaning populations, as it did not improve time to sustained liberation compared with PSV when both were embedded within structured weaning algorithms.
• Mechanistic insight

  The intervention achieved physiologic separation (higher patient effort and slightly lower assistance on PAV+), yet this did not translate into shorter ventilation, suggesting that effort-targeting alone may be insufficient to overcome other determinants of weaning delay (sedation burden, disease trajectory, secretion load, weakness, delirium, extubation readiness assessment).
• Systems implication

  Given the technology/training requirements and the neutral result, priority may remain on high-yield weaning practices (consistent readiness screening, SBTs, minimising sedatives, early mobilisation) rather than mode substitution.

Controversies & Subsequent Evidence

• Prior outcome-focused trials of PAV+ were small and suggested potential benefit in weaning success and/or ventilation duration, including a randomised ICU weaning trial and a trial comparing PAV+ with PSV in critically ill patients; PROMIZING provides substantially more precise estimates and did not reproduce a clinically important reduction in ventilation duration.²
• Smaller RCT evidence comparing PAV+ with PSV emphasised improved patient–ventilator interaction and feasibility; PROMIZING demonstrates that feasibility and physiologic separation do not necessarily translate into faster liberation when embedded in high-quality protocolised weaning pathways.³
• Pre-PROMIZING meta-analyses reported improved weaning-related outcomes with PAV+ (and more broadly proportional modes) compared with PSV; PROMIZING’s neutral effect is likely to attenuate pooled estimates and increases uncertainty about patient-important benefit outside selected subgroups.⁴
• Meta-analyses aggregating proportional modes (PAV and NAVA) versus PSV have suggested potential clinical benefit, but they rely on heterogeneous populations, variable implementation, and often small single-centre trials; PROMIZING provides a pragmatic benchmark for expected effect size in contemporary ICU practice.⁵
• Large international observational data highlight that weaning practices and outcomes vary substantially across ICUs, and prolonged weaning is common; PROMIZING’s protocolised approach may have narrowed between-group differences by improving the quality of care in both arms (raising the bar for detecting incremental benefit from mode choice alone).⁶
• The “WIND” framework underscores heterogeneity in weaning trajectories and outcomes; PROMIZING’s neutral average effect does not exclude the possibility that more targeted application (e.g., specific weaning phenotypes) could be beneficial, but such strategies require new predictive enrichment approaches and prospective validation.⁷
• Physiological rationale remains plausible: patient–ventilator asynchrony is common during assisted ventilation and may be associated with worse outcomes; PAV+ aims to mitigate asynchrony by matching assistance to demand, but PROMIZING suggests that (in aggregate) this is not sufficient to shorten ventilation in a protocolised weaning ecosystem.⁸
• Ventilator-induced diaphragm atrophy can occur rapidly during mechanical ventilation; PROMIZING achieved higher patient effort with PAV+, but any diaphragm-protective effect did not translate into shorter ventilation at a population level, indicating competing causal pathways (neuromuscular weakness beyond diaphragm, delirium, pulmonary mechanics, secretion burden) likely dominate for many patients.⁹
• Established weaning guidance has historically prioritised readiness assessment, SBTs, and protocolised liberation strategies over specific assisted modes; PROMIZING supports continuing to emphasise structured weaning processes, as optimised PSV performed similarly to PAV+ in this setting.¹⁰
• More contemporary liberation guidance (rehabilitation, liberation protocols, cuff leak tests) similarly focuses on processes rather than mandating a specific assisted mode; PROMIZING is consistent with this process-centred paradigm and adds high-quality evidence that mode substitution alone is unlikely to be transformative.¹¹
• Evidence-based assessments for ventilator discontinuation emphasise structured evaluation and SBT execution; PROMIZING highlights that when such assessments are embedded for all patients, the incremental value of proportional assist strategies may be small.¹²
• Adjacent proportional-mode evidence (e.g., neurally adjusted ventilatory assist) suggests that improving synchrony and matching patient demand is feasible at scale; PROMIZING indicates that feasibility does not guarantee shorter ventilation without additional effective co-interventions or patient selection strategies.¹³

Summary

• In 573 invasively ventilated ICU adults, PAV+ did not reduce time to sustained liberation from IMV compared with PSV (median 7.3 vs 6.8 days; HR 0.96; 95% CI 0.80 to 1.15).
• Key secondary clinical outcomes were similar, including ventilator-free days to day 28 and 90-day mortality (29.6% vs 26.6%; RR 1.13; 95% CI 0.83 to 1.54).
• PAV+ achieved physiologic separation (higher patient effort and slightly lower estimated assist), confirming protocol delivery but not patient-centred benefit.
• PAV+ was associated with more frequent rescue to assist-control ventilation (58.9% vs 49.7%), potentially diluting exposure and highlighting real-world implementation challenges.
• When both arms received structured, algorithmic weaning and SBT/extubation workflows, mode substitution alone did not shorten ventilation duration.

Overall Takeaway

PROMIZING is a landmark, pragmatic, multinational RCT showing that an effort-targeted PAV+ weaning strategy does not shorten time to sustained liberation from invasive ventilation compared with an optimised PSV strategy when both are delivered within structured weaning algorithms. The trial confirms that physiologic separation and feasibility are achievable at scale, but also that mode substitution alone is unlikely to be a dominant lever for reducing ventilation duration in contemporary ICU practice.

Bibliography

• 1. Bosma KJ, Martin CM, Burns KEA, Mancebo J, Suárez Montero JC, Skrobik Y, et al. Proportional assist ventilation and pressure support ventilation for minimising the duration of invasive mechanical ventilation: study protocol for a randomized controlled trial. Trials. 2023;24:232.
• 2. Bosma KJ, Read BA, Bahr M, Pineda LA, Machado FR, Cavalcanti AB, et al. A randomized trial of weaning from mechanical ventilation on pressure support ventilation versus proportional assist ventilation. Crit Care Med. 2016;44:1098-1108.
• 3. Xirouchaki N, Kondili E, Vaporidi K, Xirouchakis G, Klimathianaki M, Georgopoulos D. Proportional assist ventilation with load-adjustable gain factors in critically ill patients: comparison with pressure support. Intensive Care Med. 2008;34:2026-2034.
• 4. Ou-Yang LJ, Zeng Y, Feng YQ, Lin B, Luo KY, Chen PX. Proportional assist ventilation versus pressure support ventilation for weaning from mechanical ventilation in adults: a meta-analysis and trial sequential analysis. Crit Care. 2020;24:556.
• 5. Kataoka J, Kuriyama A, Norisue Y, Fujitani S. Proportional modes versus pressure support ventilation: a systematic review and meta-analysis. Ann Intensive Care. 2018;8:122.
• 6. Burns KEA, Rizvi L, Cook DJ, et al. The WEAN SAFE study: an international cohort study of weaning practices and outcomes. Lancet Respir Med. 2023;11:65-83.
• 7. Béduneau G, Pham T, Schortgen F, Piquilloud L, Zogheib E, Jonas M, et al. Epidemiology of weaning outcome according to a new definition. The WIND Study. Am J Respir Crit Care Med. 2017;195:772-783.
• 8. Thille AW, Rodriguez P, Cabello B, Lellouche F, Brochard L. Patient-ventilator asynchrony during assisted mechanical ventilation. Intensive Care Med. 2006;32:1515-1522.
• 9. Levine S, Nguyen T, Taylor N, et al. Rapid disuse atrophy of diaphragm fibers in mechanically ventilated humans. N Engl J Med. 2008;358:1327-1335.
• 10. Boles JM, Bion J, Connors A, Herridge M, Marsh B, Melot C, et al. Weaning from mechanical ventilation. Eur Respir J. 2007;29:1033-1056.
• 11. Girard TD, Alhazzani W, Kress JP, Ouellette DR, Schmidt GA, Truwit JD, et al. An official American Thoracic Society/American College of Chest Physicians clinical practice guideline: liberation from mechanical ventilation in critically ill adults. Rehabilitation protocols, ventilator liberation protocols, and cuff leak tests. Am J Respir Crit Care Med. 2017;195:120-133.
• 12. Macintyre NR. Evidence-based assessments in the ventilator discontinuation process. Respir Care. 2012;57:1611-1618.
• 13. Kacmarek RM, Villar J, Parrilla D, Alba F, Solano R, Liu S, et al. Neurally adjusted ventilatory assist in acute respiratory failure: a randomized controlled trial. Intensive Care Med. 2020;46:2327-2337.