Publication

• Title: Neuromuscular Blockers in Early Acute Respiratory Distress Syndrome.
• Acronym: ACURASYS — ARDS et Curarisation Systématique.
• Year: 2010.
• Journal published in: New England Journal of Medicine.
• Citation: Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A, et al. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010;363:1107–1116.

Context & Rationale

• Background

  • Lung-protective ventilation had improved ARDS outcomes, but substantial mortality persisted despite limiting tidal volume and airway pressure.
  • Neuromuscular blockade could improve patient–ventilator interaction, suppress injurious respiratory effort and facilitate controlled ventilation; earlier small trials suggested improved oxygenation and reduced inflammation.
  • Potential benefits had to be balanced against immobility, the requirement for deep sedation and concern about ICU-acquired neuromuscular dysfunction.
• Research Question/Hypothesis

  • Would a 48-hour course of cisatracurium initiated early in hypoxaemic ARDS improve survival when added to deep sedation and lung-protective ventilation?
• Why This Matters

  • The trial moved beyond physiological outcomes to test whether a brief, readily deliverable intervention could alter mortality and duration of organ support.
  • Its findings could justify moving paralysis from selective rescue treatment to systematic early treatment in eligible patients.

Design & Methods

• Research Question:
  • In invasively ventilated adults with ARDS of less than 48 hours’ duration and PaO₂/FiO₂ <150 mmHg, does 48 hours of cisatracurium reduce 90-day in-hospital mortality compared with placebo?
• Study Type:
  • Investigator-initiated, multicentre, parallel-group, randomised, double-blind, placebo-controlled superiority trial in 20 French ICUs; recruitment ran from March 2006 to March 2008.
  • Computer-generated randomisation used blocks of four, stratified by centre, age ≤60 versus >60 years and baseline duration of mechanical ventilation ≤48 versus >48 hours.
  • Public funding came from Assistance Publique–Hôpitaux de Marseille and the French Ministry of Health; GlaxoSmithKline supplied the study drugs.
• Population:
  • Adults receiving invasive ventilation with bilateral pulmonary infiltrates, no clinical evidence of left atrial hypertension, and PaO₂/FiO₂ <150 mmHg on PEEP ≥5 cmH₂O and tidal volume 6–8 mL/kg predicted body weight.
  • The qualifying ARDS syndrome had to have developed within the preceding 48 hours; this did not require total ventilation duration to be less than 48 hours.
  • Important exclusions included existing continuous neuromuscular blockade, neuromuscular-blocker allergy, pregnancy, suspected or confirmed intracranial hypertension, chronic respiratory failure requiring home oxygen or ventilation, weight in kilograms exceeding height in centimetres, extensive burns, Child–Pugh C cirrhosis, bone-marrow transplantation or chemotherapy-associated aplasia, existing pneumothorax, treatment limitations and expected ventilation for less than 48 hours.
  • Of 1,326 screened patients, 340 were randomised: 178 to cisatracurium and 162 to placebo; 339 contributed to the analysis.
• Intervention:
  • Cisatracurium 15 mg intravenously, followed by a fixed infusion of 37.5 mg/hour for 48 hours.
  • The dose was neither weight-adjusted nor titrated using peripheral nerve stimulation; train-of-four monitoring was prohibited to preserve blinding.
  • Deep sedation to a Ramsay score of 6 was established before study-drug administration and targeted throughout the intervention, using a benzodiazepine and opioid, with additional agents when required.
• Comparison:
  • Matching saline placebo bolus and infusion for 48 hours, with the same deep-sedation target.
  • Both groups received volume assist-control ventilation, tidal volume 6–8 mL/kg predicted body weight, a plateau-pressure ceiling of 32 cmH₂O and an ARDS Network PEEP/FiO₂ table.
  • Oxygenation targets were SpO₂ 88–95% or PaO₂ 55–80 mmHg; ventilator adjustments targeted arterial pH 7.20–7.45.
  • Open-label cisatracurium boluses were permitted for plateau pressure >32 cmH₂O for more than 10 minutes despite increased sedation and ventilator adjustment, beginning with 20 mg intravenously.
  • Prone positioning, inhaled nitric oxide and intravenous almitrine were available for refractory hypoxaemia; corticosteroids and haemodynamic management remained at clinicians’ discretion.
  • Ventilator weaning followed a standardised approach from day 3, with assessment when FiO₂ was ≤0.60. These treatment details are specified in the protocol and statistical analysis plan.¹
• Blinding:
  • Identical study-drug presentations concealed allocation from patients, clinical staff, evaluators, monitors and data analysts; allocation used a central telephone system and consecutively numbered boxes.²
• Statistics:
  • A total of 340 patients were required to detect a 15-percentage-point absolute mortality reduction, from 50% to 35%, with 80% power (β=0.20) at a two-sided α=0.05.
  • The primary analysis used Cox regression adjusted for baseline SAPS II, plateau pressure and PaO₂/FiO₂; crude mortality comparisons were also presented.
  • Analysis was described as intention-to-treat, with one exception: a patient assigned cisatracurium withdrew consent before treatment and was excluded.
  • No interim analysis was planned. Secondary outcomes and subgroup comparisons were not protected by a multiplicity-adjusted testing hierarchy.
• Follow-Up Period:
  • The primary endpoint was death before hospital discharge and within 90 days of enrolment: 90-day in-hospital mortality, rather than unrestricted 90-day vital status after discharge.
  • Secondary endpoints included 28-day mortality, days alive without ventilation or organ failure, barotrauma, and muscle strength at day 28 and ICU discharge.

Key Results

This trial was not stopped early. The planned 340 patients were randomised, with no interim analysis; 177 cisatracurium and 162 placebo recipients were analysed.

Outcome	Cisatracurium	Placebo	Effect	p value / 95% CI	Notes
90-day in-hospital mortality: primary adjusted analysis	31.6%	40.7%	Adjusted HR 0.68	95% CI 0.48 to 0.98; P=0.04	Group percentages are crude risks; the HR adjusts for SAPS II, plateau pressure and PaO₂/FiO₂.
90-day in-hospital mortality: crude comparison	31.6% 95% CI 25.2 to 38.8	40.7% 95% CI 33.5 to 48.4	Not reported	P=0.08	Did not reach the conventional significance threshold.
28-day mortality	42/177 (23.7%)	54/162 (33.3%)	RR 0.71	95% CI 0.51 to 1.00; P=0.05	Secondary endpoint; borderline evidence.
Ventilator-free days to day 28	10.6 ± 9.7	8.5 ± 9.4	Not reported	P=0.04; effect CI not reported	Mean ± SD; deaths before the assessment horizon received zero ventilator-free days.
Ventilator-free days to day 90	53.1 ± 35.8	44.6 ± 37.5	Not reported	P=0.03; effect CI not reported	Mean ± SD; incorporates survival and liberation from ventilation.
Days without organ failure to day 28	15.8 ± 9.9	12.2 ± 11.1	Not reported	P=0.01; effect CI not reported	Mean ± SD; secondary endpoint.
Days outside ICU to day 90	47.7 ± 33.5	39.5 ± 35.6	Not reported	P=0.03; effect CI not reported	Mean ± SD; secondary endpoint.
Barotrauma	9/177 (5.1%)	19/162 (11.7%)	RR 0.43	95% CI 0.20 to 0.93; P=0.03	Includes pneumothorax and other extra-alveolar air.
Pneumothorax	7/177 (4.0%)	19/162 (11.7%)	RR 0.34	95% CI 0.15 to 0.78; P=0.01	A component of barotrauma; not an independent replication.
Patients without ICU-acquired paresis at day 28	68/96 (70.8%)	52/77 (67.5%)	Not reported	P=0.64; effect CI not reported	Assessable patients only; paresis defined by MRC score <48/60.
Patients without ICU-acquired paresis at ICU discharge	72/112 (64.3%)	61/89 (68.5%)	Not reported	P=0.51; effect CI not reported	Assessable patients only; higher percentages indicate less paresis.

• Primary interpretation:
  • The adjusted primary analysis favoured cisatracurium, but the crude 90-day comparison remained uncertain; HR 0.68 is an adjusted relative hazard, not an absolute mortality reduction.
  • Ventilator-free days and barotrauma favoured treatment, but these secondary findings require consideration of multiplicity and their dependence on survival.
• Hypoxaemia subgroup:
  • Among patients with baseline PaO₂/FiO₂ <120 mmHg, 90-day mortality was 30.8% versus 44.6% (P=0.04).
  • The numerical threshold was selected from the cohort’s oxygenation distribution; no treatment-by-subgroup interaction was reported. The supplementary survival comparison for this stratum gave log-rank P=0.051, whereas P=0.04 refers to the mortality-proportion comparison.¹²
• Harms:
  • One episode of bradycardia occurred during cisatracurium infusion.
  • No statistically significant difference in measured paresis was identified; the restricted assessment denominators prevent a conclusion that neuromuscular harm was excluded.

Internal Validity

• Randomisation, concealment and attrition:
  • Central allocation and identical vials provided credible protection against selection bias.
  • The 178:162 allocation imbalance is compatible with incomplete blocks across numerous centre-specific strata and does not itself indicate defective randomisation.
  • One post-randomisation consent withdrawal occurred in the intervention group before treatment; the remaining 339 patients had complete primary-outcome follow-up.
• Baseline severity and heterogeneity:
  • Mean age was 58 years in both groups; mean SAPS II was 50 ± 16 versus 47 ± 14.
  • Baseline PaO₂/FiO₂ was lower with cisatracurium: 106 ± 36 versus 115 ± 41 mmHg (P=0.03).
  • Direct lung injury accounted for 80.2% versus 75.9%; pneumonia and aspiration predominated, providing a clinically coherent but not biologically homogeneous ARDS population.
  • Centre stratification and common ventilation procedures limited avoidable variation, but the trial could not establish consistent treatment effects across causes of ARDS.
• Timing, dose and treatment separation:
  • Median time from ARDS onset to enrolment was 18 hours (IQR 6–31) versus 15 hours (IQR 7–27), within the intended early-treatment window.
  • The fixed 37.5 mg/hour infusion provided a substantial pharmacological contrast; individual blockade depth and the optimal minimum effective dose were not established.
  • Open-label cisatracurium was administered during the first 48 hours to 18/177 (10%) versus 36/162 (22%) patients (P=0.004).
  • During the entire ICU stay, additional cisatracurium was used in 89 (50%) versus 90 (56%) patients (P=0.33). Permitted rescue treatment makes this a comparison of systematic early blockade with selective blockade.²
• Ventilation, sedation and adjunctive therapies:
  • Baseline tidal volume was 6.55 ± 1.12 versus 6.48 ± 0.92 mL/kg predicted body weight; PEEP was 9.2 ± 3.2 versus 9.2 ± 3.5 cmH₂O, and plateau pressure 25.0 ± 5.1 versus 24.4 ± 4.7 cmH₂O.
  • At 24 hours, tidal volume was 6.3 ± 0.8 mL/kg in both groups and PEEP was 9.6 ± 2.7 versus 9.7 ± 2.8 cmH₂O.
  • Among midazolam recipients, median cumulative dose over 48 hours was 432 mg (IQR 240–720; n=157) versus 480 mg (IQR 269–770; n=140), P=0.20.
  • Prone positioning was used in 50 (28%) versus 47 (29%), inhaled nitric oxide in 50 (28%) versus 53 (33%), and corticosteroids specifically for ARDS in 28 (16%) versus 37 (23%) patients.²
• Performance bias and outcome assessment:
  • Blinding was carefully designed, but visible respiratory effort or movement could disclose treatment; blinding success was not formally demonstrated.
  • Mortality was objective and weaning procedures were standardised; muscle-strength testing required survival, awakening and cooperation.
  • Patient–ventilator dyssynchrony was not systematically measured, limiting assessment of the proposed mechanism and physiological treatment separation.
• Statistical rigour:
  • The recruitment target was met, and adjusted survival modelling was envisaged in the original plan.
  • The precise model specification, a later SAP amendment and multiple secondary comparisons qualify the strength of the mortality inference; these issues are examined below.

Conclusion on Internal Validity: Internal validity is moderate overall: allocation, follow-up and treatment standardisation were strong, but the mortality conclusion is tempered by imprecision, analytical flexibility and uncertainties surrounding physiological blinding and muscle-strength assessment.

External Validity

• Population representativeness:
  • The population resembles adults with early, substantial hypoxaemia requiring invasive ventilation, predominantly from pulmonary causes.
  • “Severe ARDS” used the trial’s pre-Berlin threshold of PaO₂/FiO₂ <150 mmHg; the cohort therefore overlaps contemporary moderate and severe ARDS.
  • Exclusion of existing pneumothorax, marked obesity, important chronic respiratory disease, selected immunocompromised states and patients already receiving continuous blockade restricts extrapolation to these groups.
• Setting and contemporary applicability:
  • Twenty ICUs improve transportability beyond a single centre, but all were in France and used a shared sedation and ventilation strategy.
  • The results apply most directly when deep sedation is already required; they do not directly compare routine paralysis with a strategy permitting lighter sedation.
  • The tested regimen presupposes reliable infusion delivery, sustained analgesia and sedation, and experienced invasive-ventilation care.
  • The trial does not establish benefit from longer courses, later initiation, other neuromuscular blockers or routine use during prone positioning.

Conclusion on External Validity: External validity is moderate for similarly selected patients with early ARDS managed under deep sedation, but limited for universal adoption within contemporary sedation and ventilation practice.

Strengths & Limitations

• Strengths:
  • A clinically important question addressed with a randomised, multicentre, placebo-controlled design.
  • Credible allocation concealment, extensive blinding arrangements and almost complete retention.
  • Early treatment, an explicit regimen and standardised background ventilation and weaning.
  • Mortality, organ-support duration and neuromuscular safety assessed alongside respiratory outcomes.
  • Completion of the recruitment target without early stopping.
• Limitations:
  • A relatively small mortality trial designed around a large anticipated treatment effect.
  • An adjusted mortality result close to the significance threshold and uncertainty about the timing of final analytical decisions relative to unblinding.
  • Deep sedation in both groups, with no direct test against lighter sedation.
  • No systematic measurement of dyssynchrony, inspiratory effort or individual blockade depth.
  • Exploratory subgroup inference, multiple secondary outcomes and incomplete neuromuscular assessment.
  • Limited evidence concerning long-term function or the minimum effective dose and duration.

Interpretation & Why It Matters

• Historical contribution

  • A brief course of paralysis produced encouraging adjusted survival and ventilator-free-day findings despite background lung-protective ventilation, making neuromuscular blockade a serious candidate for outcome-modifying treatment.³
• What the trial supports

  • Early cisatracurium may improve outcomes when added to an established strategy of deep sedation and controlled ventilation in selected hypoxaemic patients.
  • The reduction in barotrauma supports a lung-protective interpretation, without establishing the mechanism or proving that all eligible patients require paralysis.
• Clinical significance

  • The relevant bedside question is whether blockade is needed to achieve safe ventilation in the individual patient; oxygenation severity alone does not describe respiratory effort, dyssynchrony or treatment tolerance.

Controversies & Other Evidence

• Adjusted significance and the chronology of the analysis plan:
  • The adjusted HR of 0.68 (95% CI 0.48 to 0.98; P=0.04) should be interpreted alongside crude mortality of 31.6% versus 40.7% (P=0.08). Covariate adjustment can improve precision; differing significance does not itself invalidate a randomised trial.
  • The original SAP anticipated Cox modelling with SAPS II and one of lung-injury score, plateau pressure or PaO₂/FiO₂, plus possible additional variables selected from univariate analyses.
  • An amendment dated 20 October 2008 selected SAPS II and plateau pressure and added oxygenation-tertile analyses, after recruitment had ended in March 2008. The supplied documents do not establish whether this amendment preceded database lock and unblinding.
  • Consequently, describing the entire adjusted analysis as post hoc is inaccurate, but treating every modelling decision as fixed before recruitment is also unjustified.¹
  • PaO₂/FiO₂ was added because of baseline imbalance; the supplement states that benefit persisted after removing it from the model. Correction for hypoxaemia alone therefore does not explain the adjusted result.²
• Power, multiplicity and the PaO₂/FiO₂ <120 claim:
  • The trial targeted a large mortality reduction, while observed control mortality was 40.7% rather than the anticipated 50%. Meeting the recruitment target did not make the study precise for smaller, clinically worthwhile effects.
  • The 120 mmHg threshold separated the two more hypoxaemic tertiles from the least hypoxaemic tertile; it was not an independently validated treatment-selection threshold.
  • A significant result within one stratum and a nonsignificant result in another do not demonstrate effect modification. The absence of a reported interaction test, data-derived boundary and unadjusted multiplicity make this finding exploratory.
  • The SAP amendment also removed several secondary assessments and deferred others; the index publication does not provide the originally envisaged comprehensive longer-term functional assessment.¹
• Blinding, dyssynchrony and the control strategy:
  • Clinical evidence of paralysis could compromise blinding; unmeasured dyssynchrony could also leave important differences in delivered ventilation undetected by scheduled airway-pressure measurements.⁴
  • The investigators emphasised profound sedation and similar respiratory rates, while acknowledging that subclinical inspiratory effort could not be excluded. Similar respiratory rates do not establish equivalent effort or transpulmonary stress.⁵
  • Rescue blockade in the placebo group was permitted care, not automatically a protocol violation. It may reduce the contrast between strategies, while making the comparison clinically meaningful.
• Muscle safety and outcome interpretation:
  • MRC assessment omits patients who die or remain unable to cooperate; unequal survival also changes the populations available for testing. A nonsignificant comparison among assessable survivors cannot exclude important weakness.
  • Concern about incomplete ascertainment of weakness was raised in correspondence; the investigators confirmed that physical activity had not been measured.⁴⁵
  • Ventilator-free and organ-failure-free days combine survival with recovery. Their improvement is supportive but cannot be counted as independent confirmation that paralysis accelerated recovery among comparable survivors.
• ROSE: the decisive subsequent randomised comparison:
  • ROSE randomised 1,006 patients in 48 US hospitals to the same 48-hour cisatracurium regimen with deep sedation or usual care permitting lighter sedation; both groups received a higher-PEEP strategy.
  • It stopped for futility at the second interim analysis. Ninety-day in-hospital mortality was 42.5% versus 42.8%; absolute difference −0.3 percentage points, 95% CI −6.4 to 5.9; P=0.93.
  • Serious cardiovascular adverse events numbered 14 versus 4 (P=0.02), and early physical activity was lower with routine blockade.
  • ROSE did not demonstrate a survival advantage from routine early paralysis over contemporary usual care; it did not resolve the benefit of selective rescue treatment for refractory dyssynchrony.⁶
• Reconciling ACURASYS and ROSE:
  • The trials tested different care strategies: adding blockade when both groups were deeply sedated versus introducing blockade together with deeper sedation than the control strategy.
  • Differences in PEEP and enrolment timing also complicate comparison; they were not randomised between trials and cannot establish which factor explains the divergent findings.
  • One plausible mechanism is prevention of reverse-triggered breath stacking in deeply sedated patients. This explanation remains a hypothesis because neither trial systematically measured the relevant dyssynchrony.
  • The practical implication is to identify harmful respiratory effort and failure to achieve protective ventilation, while recognising that physiological plausibility does not establish a survival benefit for an untested subgroup.⁷
• Meta-analysis: comparator choice matters:
  • A systematic review of seven trials involving 1,598 patients found different mortality estimates against deep-sedation controls (RR 0.72; 95% CI 0.58 to 0.91) and the lighter-sedation control in ROSE (RR 0.99; 95% CI 0.86 to 1.15).
  • Barotrauma was reduced (RR 0.55; 95% CI 0.35 to 0.85), while the weakness estimate remained compatible with harm (RR 1.16; 95% CI 0.98 to 1.37).
  • These comparisons between trials cannot isolate a causal interaction with sedation depth, because sedation, PEEP, timing and trial era varied together.⁸
• Current guidelines: different populations and questions:
  • ESICM 2023: recommends against routine continuous NMBA infusions to reduce mortality in moderate-to-severe ARDS unrelated to COVID-19; this is a strong recommendation with moderate-certainty evidence.⁹
  • ATS 2024: conditionally suggests blockade in early severe ARDS, with low-certainty evidence; implementation specifies onset <48 hours and PaO₂/FiO₂ ≤100 mmHg, with treatment generally limited to 48 hours when possible.¹⁰
  • SCCM 2026: conditionally suggests NMBAs for adults with PaO₂/FiO₂ <150 who remain hypoxaemic and/or fail to achieve mechanical-ventilation targets despite sedation, based on low-certainty evidence; prone positioning alone does not mandate blockade.¹¹
  • These recommendations do not establish a universal oxygenation threshold for paralysis. Their differing scope supports distinguishing routine administration from selective use to achieve specific ventilation goals.

Summary

• ACURASYS tested 48 hours of fixed-dose cisatracurium in early ARDS with PaO₂/FiO₂ <150 mmHg; both groups received deep sedation and lung-protective ventilation.
• The adjusted primary analysis favoured treatment: HR 0.68; 95% CI 0.48 to 0.98; P=0.04. Crude 90-day mortality was 31.6% versus 40.7% (P=0.08).
• Ventilator-free days increased and barotrauma decreased; limited muscle-strength ascertainment prevents a definitive claim of neuromuscular safety.
• The later SAP amendment, exploratory PaO₂/FiO₂ <120 threshold and multiple secondary comparisons temper confidence in the magnitude and specificity of benefit.
• ROSE and subsequent guidance favour assessing the need for blockade within the whole ventilation and sedation strategy, rather than automatically treating every patient meeting ACURASYS eligibility.

Overall Takeaway

ACURASYS established an influential rationale for brief early cisatracurium in selected patients with ARDS, but its adjusted survival finding does not justify universal paralysis. Its enduring contribution is to focus attention on achieving protective ventilation; subsequent evidence makes the indication, sedation strategy and physiological response central to treatment decisions.

Bibliography

• 1Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A, et al. Trial protocol and statistical analysis plan, including amendment of 20 October 2008. Supplement to: Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010;363:1107–1116.
• 2Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A, et al. Supplementary appendix to: Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010;363:1107–1116.
• 3Slutsky AS. Neuromuscular blocking agents in ARDS. N Engl J Med 2010;363:1176–1180.
• 4Devlin JW, Garpestad E, Hill NS. Neuromuscular blockers and ARDS. N Engl J Med 2010;363:2562.
• 5Papazian L, Forel JM, Roch A. Neuromuscular blockers and ARDS: the authors reply. N Engl J Med 2010;363:2563–2564.
• 6National Heart, Lung, and Blood Institute PETAL Clinical Trials Network. Early neuromuscular blockade in the acute respiratory distress syndrome. N Engl J Med 2019;380:1997–2008.
• 7Slutsky AS, Villar J. Early paralytic agents for ARDS? Yes, no, and sometimes. N Engl J Med 2019;380:2061–2063.
• 8Tarazan N, Alshehri M, Sharif S, Al Duhailib Z, Møller MH, Belley-Cote E, et al. Neuromuscular blocking agents in acute respiratory distress syndrome: updated systematic review and meta-analysis of randomized trials. Intensive Care Med Exp 2020;8:61.
• 9Grasselli G, Calfee CS, Camporota L, Poole D, Amato MBP, Antonelli M, et al. ESICM guidelines on acute respiratory distress syndrome: definition, phenotyping and respiratory support strategies. Intensive Care Med 2023;49:727–759.
• 10Qadir N, Sahetya S, Munshi L, Summers C, Abrams D, Beitler J, 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;209:24–36.
• 11Erstad BL, Cobas MA, Qadir N, Alshamsi F, Sharif S, Devine EM, et al. Society of Critical Care Medicine guidelines for the administration of neuromuscular blockade in adults with acute respiratory distress syndrome. Crit Care Med 2026;54:634–643.