Skip to main content

Foundational Trials

ACURASY

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Neuromuscular blockers in early acute respiratory distress syndrome
  • Acronym: ACURASYS
  • Year: 2010
  • Journal published in: The New England Journal of Medicine
  • Citation: Papazian L, Forel JM, Gacouin A, et al. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med. 2010;363(12):1107-1116.

Context & Rationale

  • Background
    Early ARDS is characterised by severe hypoxaemia, high ventilatory demand, and prominent patient–ventilator dyssynchrony that can undermine lung-protective ventilation.
    Neuromuscular blocking agents (NMBAs) were commonly used to facilitate ventilation and reduce oxygen consumption, but their net effect on patient-important outcomes was uncertain.
    A key countervailing concern was ICU-acquired weakness/paresis, historically linked to prolonged paralysis and deep sedation.
  • Research Question/Hypothesis
    In patients with early, severe ARDS receiving protocolised lung-protective ventilation, does a time-limited (48-hour) cisatracurium infusion, compared with placebo, reduce 90-day mortality and improve clinically relevant outcomes without increasing ICU-acquired paresis?
  • Why This Matters
    If effective, early short-course NMBA would represent a rapidly deployable adjunct to reduce ventilator-induced lung injury (VILI), barotrauma, and potentially mortality in severe ARDS.
    If ineffective (or harmful via weakness, immobility, or deeper sedation), routine paralysis would represent avoidable iatrogenesis and resource burden.

Design & Methods

  • Research Question: Among adults with early severe ARDS ventilated with a lung-protective strategy, does 48-hour continuous cisatracurium (vs placebo) reduce 90-day in-hospital mortality?
  • Study Type: Randomised, multicentre, double-blind, placebo-controlled, investigator-initiated trial in 20 French ICUs (enrolment March 2006 to March 2008); stratified randomisation (centre, age, and duration of mechanical ventilation) with concealed allocation and identical study-drug preparation.1
  • Population:
    • Setting: mechanically ventilated ICU patients with early severe ARDS (enrolled within 48 hours of meeting ARDS criteria).
    • Key inclusion features: PaO2/FiO2 <150 with PEEP ≥5 cmH2O; bilateral pulmonary infiltrates; no clinical evidence of left atrial hypertension; lung-protective ventilation mandated.
    • Key exclusions (examples): pneumothorax at baseline; circumstances precluding protocolised ventilation/sedation; and other pre-specified safety/feasibility exclusions (as per trial flow diagram and protocol).
  • Intervention:
    • Cisatracurium besylate: 15 mg IV bolus then fixed-dose infusion 37.5 mg/hour for 48 hours (no peripheral nerve stimulator monitoring to preserve blinding).
    • Deep sedation targeted during the infusion period (Ramsay sedation scale 6) with analgesia as required.
    • Rescue bolus cisatracurium permitted (both groups) for unsafe ventilatory mechanics despite protocol measures (e.g., persistent high plateau pressure), per protocol.
  • Comparison:
    • Placebo infusion for 48 hours, with the same protocolised ventilation and deep sedation targets during the blinded infusion period.
    • Rescue bolus cisatracurium permitted using the same criteria as the intervention group.
  • Blinding: Double-blind (patients, treating clinicians, and outcome assessors/statisticians); neuromuscular monitoring was intentionally not used to limit unblinding risk.1
  • Statistics: A total of 340 patients were required to detect a 15% absolute reduction in 90-day mortality (from 50% to 35%) with 80% power at the 5% (two-sided) significance level; analyses were conducted by intention-to-treat; the primary endpoint was analysed with Cox regression adjusted for baseline PaO2/FiO2, SAPS II, and plateau pressure.
  • Follow-Up Period: Primary outcome assessed to day 90 (in-hospital death); the published protocol planned longer-term follow-up assessments to day 180 (including pulmonary function testing and functional/quality-of-life measures), which were not reported in the main NEJM publication.2

Key Results

This trial was not stopped early. Recruitment reached the planned sample size; no interim analysis was performed.

Outcome Cisatracurium (n=177) Placebo (n=162) Effect p value / 95% CI Notes
90-day in-hospital mortality (primary; adjusted time-to-event) 31.6% 40.7% HR 0.68 95% CI 0.48 to 0.98; P=0.04 Primary analysis: Cox model adjusted for baseline PaO2/FiO2, SAPS II, and plateau pressure
90-day mortality (subgroup: baseline PaO2/FiO2 <120) 30.8% 44.6% Not reported P=0.04 Threshold arose from tertile-based subgrouping described in the protocol
28-day mortality 23.7% 33.3% RR 0.71 95% CI 0.51 to 1.00; P=0.05 Absolute difference reported in paper: −9.6 percentage points
ICU mortality 29.4% 38.9% RR 0.76 95% CI 0.56 to 1.02; P=0.06 Not statistically significant at 0.05
Hospital mortality 32.2% 41.4% RR 0.78 95% CI 0.59 to 1.03; P=0.08 Not statistically significant at 0.05
Ventilator-free days (day 1 to 28) 10.6 ± 9.7 8.5 ± 9.4 Not reported P=0.04 Ventilator-free days scored as 0 after death
Ventilator-free days (day 1 to 90) 53.1 ± 35.8 44.6 ± 37.5 Not reported P=0.03 Longer-horizon ventilator-free time (death scored 0)
Days alive outside ICU (day 1 to 90) 47.7 ± 33.5 39.5 ± 35.6 Not reported P=0.03 Composite of survival and ICU length of stay
Days without nonpulmonary organ failure (day 1 to 28; none of 4) 15.8 ± 9.9 12.2 ± 11.1 Not reported P=0.01 Nonpulmonary components: cardiovascular, coagulation, hepatic, renal
Barotrauma 5.1% 11.7% RR 0.43 95% CI 0.20 to 0.93; P=0.03 Includes pneumothorax and other barotrauma events
Pneumothorax 4.0% 11.7% RR 0.34 95% CI 0.15 to 0.76; P=0.01 Clinically important harm signal in control group (higher event rate)
ICU-acquired paresis at ICU discharge (patients without paresis) 72/112 (64.3%) 61/89 (68.5%) Not reported P=0.51 Assessed only in cooperative patients (subset of cohort)
  • Mortality benefit met conventional statistical significance only in the pre-specified adjusted primary analysis (HR 0.68; 95% CI 0.48 to 0.98; P=0.04), while unadjusted mortality comparisons were less definitive.
  • Cisatracurium was associated with more ventilator-free time (day 1–28: 10.6 ± 9.7 vs 8.5 ± 9.4; P=0.04) and fewer barotrauma events (5.1% vs 11.7%; RR 0.43; 95% CI 0.20 to 0.93; P=0.03).
  • Among assessable survivors, there was no signal for increased ICU-acquired paresis (patients without paresis at ICU discharge: 64.3% vs 68.5%; P=0.51).

Internal Validity

  • Randomisation and Allocation: Centralised randomisation with concealed allocation and stratification (centre, age, and prior duration of mechanical ventilation) supported low risk of selection bias.1
  • Dropout/exclusions: 340 patients were randomised; 1 withdrew consent before receiving the study drug (analysis population: 177 vs 162). Follow-up for mortality to day 90 was complete (objective endpoint).
  • Performance/Detection Bias: Blinding was robust (placebo-controlled, identical infusions), but clinical inference of assignment was possible because paralysis suppresses spontaneous effort; protocol-mandated deep sedation (Ramsay 6) in both groups mitigated (but does not eliminate) this risk.
  • Protocol Adherence: Lung-protective ventilation was delivered with close between-group similarity in key mechanics (e.g., baseline tidal volume 6.55 ± 1.12 vs 6.48 ± 0.92 mL/kg predicted body weight; plateau pressure 24.5 ± 4.7 vs 25.0 ± 5.1 cmH2O).
  • Baseline Characteristics: Groups were generally comparable (age 58 ± 16 vs 58 ± 15; SAPS II 45.6 ± 15.1 vs 44.7 ± 16.2), with a small but statistically significant imbalance in baseline PaO2/FiO2 (106 ± 36 vs 115 ± 41; P=0.03), favouring the placebo group.
  • Timing: Intervention was initiated early (within 48 hours of meeting eligibility criteria), aligning with the biological plausibility of targeting the early injurious phase of ARDS.
  • Dose: Fixed-dose cisatracurium (37.5 mg/hour) without neuromuscular monitoring prioritised blinding but limited dose individualisation (potential for under- or over-paralysis in extremes of pharmacokinetic variability).
  • Separation of the Variable of Interest: Study-drug exposure was maximally separated by design (active infusion vs placebo for 48 hours); however, rescue cisatracurium boluses were permitted in both arms, and were used more frequently in the placebo arm early (first 48 hours: 18/177 [10%] vs 36/162 [22%], P=0.004).1
  • Adjunctive therapy use: Co-interventions used for refractory hypoxaemia were broadly similar; for example, use of prone positioning and inhaled nitric oxide occurred in both groups (detailed co-intervention frequencies reported in the supplementary appendix).1
  • Outcome Assessment: Mortality and barotrauma are objective outcomes; ICU-acquired paresis assessment required patient cooperation and therefore applied to a subset (e.g., paresis assessment at ICU discharge in 112 vs 89 patients), introducing potential ascertainment/selection bias for neuromuscular outcomes.
  • Statistical Rigor: The primary endpoint relied on an adjusted Cox model; while adjustment can increase precision, clinical interpretation depends on confidence that covariate selection and modelling were pre-specified and not data-driven.

Conclusion on Internal Validity: Overall, internal validity is moderate: randomisation and blinding were strong and major outcomes were objective, but interpretation hinges on the adjusted primary analysis and neuromuscular safety was assessed in a non-random subset of survivors.

External Validity

  • Population Representativeness: Enrolled patients had severe ARDS (PaO2/FiO2 <150) in well-resourced ICUs, with protocolised lung-protective ventilation and deep sedation during the infusion period.
  • Applicability: Findings are most directly applicable to patients with early, severe ARDS where dyssynchrony and ventilatory mechanics threaten lung-protective targets and where a short, protocolised NMBA course is feasible.
  • Practice evolution considerations: Contemporary ARDS care more commonly emphasises lighter sedation, higher utilisation of prone positioning in severe ARDS, and different ventilator management patterns; these shifts may modify the risk–benefit balance of routine early paralysis.

Conclusion on External Validity: Generalisability is reasonable for early severe ARDS in similar ICU settings, but may be limited in milder ARDS, in units prioritising light sedation, or where co-interventions (e.g., early proning) are routinely deployed.

Strengths & Limitations

  • Strengths: Double-blind placebo-controlled design; multicentre ICU trial; protocolised lung-protective ventilation; clinically important endpoints (90-day mortality, ventilator-free days, barotrauma); objective primary outcome.
  • Limitations: Modest sample size with lower-than-anticipated mortality (reduced power for the prespecified effect size); reliance on an adjusted primary model for statistical significance; deep sedation mandated during the intervention window (limits applicability to lighter-sedation paradigms); neuromuscular weakness outcomes measured only in cooperative survivors rather than the full randomised cohort.

Interpretation & Why It Matters

  • Clinical signal
    A short, early cisatracurium infusion in severe ARDS was associated with fewer barotrauma events and more ventilator-free days, alongside an adjusted reduction in 90-day mortality.
  • Mechanistic plausibility
    Potential pathways include improved patient–ventilator synchrony, reduced inspiratory effort and transpulmonary pressure swings, and attenuated VILI (consistent with fewer barotrauma events).
  • Pragmatic implication
    The trial supports considering time-limited NMBA in selected early severe ARDS when lung-protective ventilation is threatened; routine use requires integration with sedation strategy and weakness-monitoring practices.

Controversies & Subsequent Evidence

  • Primary endpoint dependence on adjustment: The headline mortality effect relied on the adjusted Cox model; unadjusted mortality comparisons were less definitive, prompting debate about the appropriate evidentiary weight of model-dependent “primary” results in RCTs.
  • Deep sedation as a design feature (and potential effect modifier): Deep sedation was protocolised during the 48-hour blinded infusion, raising questions about whether any observed benefit reflects paralysis itself, the mitigation of injurious spontaneous effort, or the broader sedation–ventilation package.
  • Neuromuscular safety ascertainment: ICU-acquired paresis was assessed only in cooperative patients, which may under-detect weakness in patients who died early, were delirious, or had prolonged coma; this complicates inference about the true population-level neuromuscular safety signal.
  • Editorial framing: Contemporary editorial commentary highlighted the biological plausibility (VILI mitigation) while emphasising the need to balance potential benefit against risks of weakness and the challenges of interpreting “routine paralysis” in evolving ventilatory practice.3
  • Correspondence: NEJM letters and author reply focused on interpretation of mortality analyses and generalisability, reinforcing the central methodological debate around adjustment, sedation, and neuromuscular outcomes.4
  • Subsequent RCT evidence (ROSE): A later, larger trial comparing early continuous cisatracurium plus deep sedation versus a lighter-sedation usual-care strategy (with intermittent NMBA allowed) found no difference in 90-day mortality, challenging routine early paralysis as a universal standard and highlighting sedation strategy as a plausible effect modifier.5
  • Updated synthesis: Post-ROSE meta-analytic updates incorporate control-arm sedation depth as an explicit heterogeneity driver, and overall support a more conditional, phenotype- and practice-dependent interpretation of early NMBA use in ARDS.6
  • Guideline evolution: Earlier professional society guidance conditionally supported short-course NMBA in selected severe ARDS; subsequent evidence has reinforced the need for individualised use within a broader lung-protective bundle, rather than automatic paralysis for all severe ARDS.7

Summary

  • ACURASYS tested 48-hour early cisatracurium versus placebo in early severe ARDS under protocolised lung-protective ventilation and deep sedation.
  • The primary adjusted analysis showed reduced 90-day mortality (HR 0.68; 95% CI 0.48 to 0.98; P=0.04), while unadjusted mortality comparisons were less definitive.
  • Cisatracurium increased ventilator-free days (10.6 ± 9.7 vs 8.5 ± 9.4 at day 28; P=0.04) and reduced barotrauma/pneumothorax (barotrauma 5.1% vs 11.7%; P=0.03).
  • Among assessable survivors, ICU-acquired paresis did not differ (patients without paresis at ICU discharge 64.3% vs 68.5%; P=0.51), but ascertainment was incomplete by design.
  • Later RCT evidence and updated syntheses suggest any benefit of early NMBA is context-dependent (particularly sedation strategy and severity phenotype), supporting selective rather than routine use.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • ACURASYS pre-dated the Berlin definition of ARDS and mandated deep sedation during the blinded infusion period; later trials and practice have shifted towards lighter sedation and earlier prone positioning, which may modify NMBA effect estimates.

Overall Takeaway

ACURASYS is a landmark because it rigorously tested short-course early paralysis in severe ARDS within a blinded, protocolised ventilation framework and demonstrated fewer barotrauma events and a model-adjusted mortality signal. Subsequent evidence has tempered routine use, but the trial remains foundational in defining when (and how) neuromuscular blockade might be beneficial as part of a lung-protective bundle—particularly in early, severe ARDS with high risk of injurious patient effort.

Overall Summary

  • In early severe ARDS, 48-hour cisatracurium reduced barotrauma and increased ventilator-free days, with an adjusted reduction in 90-day mortality.
  • Interpretation is constrained by reliance on adjusted primary analysis and incomplete ascertainment of ICU-acquired weakness (subset-based assessment).
  • Later RCTs and syntheses suggest NMBA benefit is not universal and may depend on sedation strategy and ARDS phenotype; selective, time-limited use is most defensible.

Bibliography