Publication

• Title: Inhaled Sedation in Acute Respiratory Distress Syndrome: The SESAR Randomized Clinical Trial
• Acronym: SESAR
• Year: 2025
• Journal published in: JAMA
• Citation: Jabaudon M, Quenot JP, Badie J, Audard J, Jaber S, Rieu B, et al; for the SESAR investigators. Inhaled sedation in acute respiratory distress syndrome: the SESAR randomized clinical trial. JAMA. 2025;333(18):1608-1617.

Context & Rationale

• Background

  • Volatile anaesthetics (eg, sevoflurane/isoflurane) can be delivered for ICU sedation using anaesthetic-conserving devices; proposed advantages include rapid titratability, bronchodilation, and putative anti-inflammatory effects.
  • ARDS management frequently requires deep early sedation (often alongside neuromuscular blockade and prone positioning), making sedative choice a potentially important co-intervention affecting ventilation, haemodynamics, and downstream organ injury.
  • Pre-SESAR ARDS-specific evidence was limited: a pilot randomised trial of sevoflurane in ARDS suggested improvements in oxygenation/biomarkers but was not powered for patient-centred outcomes.¹
  • Across broader ICU populations, the evidence base for volatile sedation has been heterogeneous and of variable certainty, with meta-analytic signals (where present) not settling safety/benefit in high-risk subgroups such as ARDS.²
  • Specific safety questions remained clinically salient: device-related dead space/CO₂ burden, haemodynamic effects, and potential renal fluoride exposure with prolonged volatile use.³
• Research Question/Hypothesis

  • In adults with moderate-to-severe ARDS, does early inhaled sedation with sevoflurane (delivered via an anaesthetic conserving device) increase ventilator-free days through day 28 compared with intravenous propofol sedation?
• Why This Matters

  • Volatile sedation is increasingly feasible in ICU practice, particularly during periods of high sedative demand; ARDS is a biologically plausible setting for benefit, but also a high-risk setting for harm.
  • Ventilator-free days and survival are high-stakes endpoints in ARDS; an intervention that meaningfully changes these would directly influence modern ICU sedation pathways and equipment investment.
  • Guideline-consistent sedation strategies generally prioritise minimising iatrogenic harm; robust RCT data are necessary before adopting inhaled sedation as “standard” in ARDS.⁴

Design & Methods

• Research Question: In adults with moderate-to-severe ARDS, does inhaled sedation with sevoflurane (vs intravenous propofol) improve ventilator-free days through day 28?
• Study Type: Multicentre, parallel-group, open-label, randomised clinical trial (1:1), investigator-initiated; ICU setting (37 sites in France); enrolment May 2020 to October 2023; randomisation stratified by site, ARDS severity (PaO₂/FiO₂ <100 vs ≥100), suspected/proven COVID-19, and shock at randomisation.
• Population:
  • Adults in ICU receiving invasive mechanical ventilation with moderate-to-severe ARDS defined as PaO₂/FiO₂ <150 mmHg with PEEP ≥8 cm H₂O, enrolled within 24 hours of endotracheal intubation and initiation of mechanical ventilation.
  • Key exclusions reported in trial materials included pregnancy, suspected/proven intracranial hypertension, long QT syndrome, history of malignant hyperthermia, and prior liver injury attributed to halogenated anaesthetics.
• Intervention:
  • Inhaled sevoflurane administered via an anaesthetic-conserving device in the ventilator circuit; titrated to deep sedation (target Richmond Agitation–Sedation Scale [RASS] −5 to −4) with protocolised sedation/analgesia co-interventions.
  • Assigned sedation strategy intended for up to 7 days (with protocolised management of early ARDS co-interventions, including neuromuscular blockade and prone positioning as indicated).
• Comparison:
  • Intravenous propofol infusion titrated to the same deep sedation target (RASS −5 to −4), with otherwise protocolised co-interventions consistent with the intervention group.
• Blinding: Open-label (sedation route/device not amenable to blinding at the bedside); outcomes were largely objective but ventilator liberation and ICU discharge timing may be clinically mediated.
• Statistics: Planned sample size 700 patients (350 per group) to provide >80% power to detect a between-group difference of 2 ventilator-free days (assumed SD 8) at two-sided α=0.05 (allowing for ~2% withdrawal); primary analysis was modified intention-to-treat; ventilator-free days analysed as time alive and free of invasive ventilation through day 28 with death treated as a competing event (standardised hazard ratio reported).
• Follow-Up Period: 90 days (survival status).

Key Results

This trial was not stopped early. A blinded interim analysis after 350 participants resulted in a recommendation to continue the trial.

• Despite identical medians, the distribution of ventilator-free days favoured propofol: SHR 0.76 (0.50 to 0.97) and median difference −2.1 days (95% CI −3.6 to −0.7).
• By day 90, mortality was higher with sevoflurane: 52.9% vs 44.3% (risk difference 8.6%; 95% CI 1.2% to 16.1%; HR 1.31; 95% CI 1.05 to 1.62; log-rank P=0.02).
• Pre-specified subgroup analysis for day-90 survival showed effect modification by COVID-19 status: HR 1.06 (0.80 to 1.42) in confirmed COVID-19 pneumonia vs HR 1.79 (1.22 to 2.63) in non-COVID ARDS; interaction P=0.01.

Internal Validity

• Randomisation and allocation:
  • Central, web-based randomisation with stratification by site, ARDS severity, suspected/proven COVID-19, and shock at randomisation.
  • Randomisation occurred early: time from ICU admission to randomisation was 1 day (IQR 1–2) in both groups; randomisation on the same day as intubation occurred in 209/346 (60.4%) vs 221/341 (64.8%).
• Dropout/exclusions and follow-up completeness:
  • 687 patients were randomised (346 sevoflurane; 341 propofol); primary outcome denominators were preserved (346 vs 341).
  • Minor denominator attrition was present for some fixed-time mortality endpoints (eg, day-28 mortality 345 vs 340), consistent with small amounts of missingness/withdrawal; day-90 survival analyses included 346 vs 341.
• Performance/detection bias:
  • Open-label sedation introduces potential performance bias, particularly for ventilator liberation and ICU discharge decisions embedded in ventilator-free/ICU-free day outcomes.
  • Primary and key secondary outcomes included objective components (death; ventilation status), and ARDS co-interventions were protocolised (supporting consistency), but bedside decision-making could still influence time-to-liberation.
• Protocol adherence and separation of the exposure:
  • Deep sedation targets were specified (RASS −5 to −4) and the intervention was delivered early in the ARDS course.
  • Use of allocated study sedative was high early: day 1 “study drug sedation” occurred in 332/340 (97.6%) in the propofol group and 329/337 (97.6%) in the sevoflurane group (among those with recorded observations).
  • By day 7, recorded “study drug sedation” persisted in 80/122 (65.4%) vs 50/129 (38.8%), with more frequent interruptions in the sevoflurane group (day-7 interruption 76/127 [59.8%] vs 19/123 [15.4%]) among those with data.
• Baseline comparability and illness severity:
  • Key baseline features were broadly similar: age 64.9 (14.1) vs 64.4 (14.0) years; PaO₂/FiO₂ 111 (85–133) vs 107 (79–131) mmHg; tidal volume 6.1 (5.6–6.7) vs 6.0 (5.6–6.6) mL/kg predicted body weight.
  • Baseline continuous neuromuscular blockade was common in both groups but numerically higher in sevoflurane: 310/341 (90.9%) vs 286/334 (85.7%).
  • Confirmed COVID-19 pneumonia was frequent among those with available data: 187/291 (64.3%) vs 185/293 (63.1%).
• Timing and dose:
  • Early enrolment (≤24 hours from intubation/ventilation) supports biological plausibility for affecting early ARDS trajectory.
  • Deep sedation exposure duration was substantial: median duration of assigned sedation strategy was 7 days (IQR 4–7) in both groups.
• Outcome assessment and statistical rigour:
  • Ventilator-free days and ICU-free days were analysed with competing-risk methods, aligning with the composite nature of “alive and free” endpoints in high-mortality syndromes.
  • Two-sided P<0.05 was considered statistically significant, with no multiplicity adjustment (secondary endpoints described as exploratory).

Conclusion on Internal Validity: Moderate-to-strong: early stratified randomisation with near-complete follow-up and protocolised co-interventions support causal inference, but open-label delivery and embedded clinician-mediated components of ventilator/ICU-free days introduce some residual risk of performance bias, and treatment delivery diverged later in the first week among those still observed.

External Validity

• Population representativeness:
  • Represents contemporary ICU ARDS practice in a high-income healthcare system, with a large proportion of COVID-19–associated ARDS during the enrolment period.
  • Participants were selected for moderate-to-severe ARDS early after intubation, reflecting a high-risk ARDS subgroup commonly managed with deep sedation early in illness.
• Applicability:
  • Intervention requires specific infrastructure (volatile delivery device compatible with ventilator circuits, scavenging, staff expertise, monitoring of CO₂ and haemodynamics), which may limit uptake in resource-limited environments.
  • Because the protocol targeted deep sedation (RASS −5 to −4), applicability to ICUs using lighter sedation strategies and earlier spontaneous breathing may be limited.
  • Findings are most directly applicable to early, moderate-to-severe ARDS where deep sedation is intentionally maintained; extrapolation to non-ARDS ICU sedation should be cautious.

Conclusion on External Validity: Generalisability is moderate: the population is clinically relevant (early moderate-to-severe ARDS), but implementation requirements and the deep-sedation protocol context mean translation depends on local sedation philosophy, staffing, and equipment capability.

Strengths & Limitations

• Strengths:
  • Large, multicentre RCT in a syndrome with historically limited sedation-specific outcome trials.
  • Early enrolment with stratified randomisation for key prognostic features (site, ARDS severity, COVID-19 status, shock).
  • Clinically meaningful primary endpoint with competing-risk methods appropriate for high mortality.
  • Protocolised ARDS co-interventions (including deep sedation targets) improving interpretability within the trial’s intended clinical context.
• Limitations:
  • Open-label design with potential influence on ventilator liberation and ICU discharge (embedded in the primary/secondary “free days” outcomes).
  • Later divergence in observed delivery of the assigned sedative strategy within the first week among those still in follow-up observations.
  • Secondary outcomes were treated as exploratory without multiplicity adjustment, increasing false-positive risk across multiple endpoints.
  • High proportion of COVID-19–associated ARDS during enrolment may complicate inference for non-COVID ARDS, despite stratification and subgroup analyses.

Interpretation & Why It Matters

• Practice signal

  In early moderate-to-severe ARDS managed with deep sedation, inhaled sevoflurane (via an anaesthetic conserving device) was associated with fewer ventilator-free days and lower day-90 survival than propofol, arguing against routine adoption of this strategy in this population.
• Safety signal

  The trial demonstrated clinically relevant safety signals aligned with physiologic plausibility for device/drug effects in ARDS (eg, increased severe hypercapnic acidosis and higher-stage AKI), reinforcing that sedation “route and delivery system” cannot be treated as neutral in severe respiratory failure.
• Programme-level implication

  For trialists and methodologists, SESAR illustrates the importance of formally testing ICU “equipment-plus-drug” bundles in large RCTs before widespread implementation, particularly when endpoints incorporate clinician-driven processes (eg, ventilator liberation).

Controversies & Subsequent Evidence

• The accompanying editorial highlighted that SESAR’s direction of effect (worse ventilator-free days and survival with sevoflurane) was unexpected relative to physiological hypotheses and earlier smaller studies, and emphasised the need to consider both drug and device-related mechanisms when interpreting causality.⁵
• Correspondence raised the issue that inhaled-sedation delivery systems alter the ventilator circuit (eg, dead space/resistance), potentially contributing to CO₂ burden and complicating attribution of harm to the volatile agent alone; the trialists’ reply reinforced interpretation within the prespecified protocol framework and reported outcomes.⁶⁷⁸
• Earlier randomised evidence in ARDS and ICU sedation provided mixed signals (physiologic benefit in a pilot ARDS trial; broader ICU sedative non-inferiority in a large phase 3 isoflurane trial), underscoring that extrapolation from smaller physiology-focused datasets to ARDS outcomes is unreliable.¹⁹
• Pre-SESAR meta-analytic work on volatile sedation suggested limited-certainty evidence for short-term outcome effects across heterogeneous ICU populations; SESAR provides high-weight ARDS-specific outcome data that is likely to materially influence future pooled estimates and guideline discussions.²
• Renal safety remains a key interpretive axis: systematic review evidence on volatile anaesthetic renal outcomes exists, but SESAR’s observed AKI signal in ARDS highlights the need for syndrome- and delivery-specific safety assessment rather than reliance on operating-theatre or mixed-ICU extrapolations.³

Summary

• SESAR randomised 687 adults with early moderate-to-severe ARDS to inhaled sevoflurane vs intravenous propofol (deep sedation target RASS −5 to −4).
• Primary outcome favoured propofol: ventilator-free days through day 28 were lower with sevoflurane (SHR 0.76; 95% CI 0.50 to 0.97; median difference −2.1 days).
• Day-90 mortality was higher with sevoflurane (52.9% vs 44.3%; HR 1.31; 95% CI 1.05 to 1.62; log-rank P=0.02).
• Safety signals included more severe hypercapnic acidosis (9.1% vs 5.0%; RR 1.82; 95% CI 1.04 to 3.18) and more KDIGO stage 3 AKI (33.5% vs 27.0%).
• Subgroup analysis suggested effect modification by COVID-19 status for day-90 survival (interaction P=0.01), with a larger hazard in non-COVID ARDS.

Overall Takeaway

SESAR is a landmark ARDS sedation trial because it tested an increasingly feasible “drug-plus-device” ICU sedation strategy against a standard intravenous comparator using patient-centred endpoints in a large multicentre population. The results signal that, in early moderate-to-severe ARDS managed with deep sedation, inhaled sevoflurane should not be adopted as routine practice and that future volatile-sedation research must explicitly disentangle pharmacology from delivery-system effects.

Bibliography

• 1Jabaudon M, Boucher P, Imhoff E, et al. Sevoflurane for sedation in acute respiratory distress syndrome: a randomized controlled pilot study. Am J Respir Crit Care Med. 2017;195(6):792-800.
• 2Phillips MR, Weinberg JA, O’Gara B, et al. Volatile anesthetic sedation in intensive care units. NEJM Evid. 2024;3(5):EVIDra2300142.
• 3Taylor A, et al. Renal safety of volatile anesthetics: a systematic review and meta-analysis. J Anesth. 2023;37(5):806-819.
• 4Devlin JW, Skrobik Y, Gélinas C, et al. Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU. Crit Care Med. 2018;46(9):e825-e873.
• 5Venkatesh B. Sevoflurane sedation in acute respiratory distress syndrome. JAMA. 2025;333(18):1586-1588.
• 6Berra L, Slessarev M, Bittner EA, et al. Acute respiratory distress syndrome and inhaled sedation. JAMA. Published online July 10, 2025.
• 7De Bus L, et al. Acute respiratory distress syndrome and inhaled sedation. JAMA. Published online July 10, 2025.
• 8Jabaudon M, et al. In reply. JAMA. Published online July 10, 2025.
• 9Meiser A, Bellgardt M, Belda J, et al. Inhaled isoflurane via the anaesthetic conserving device versus propofol for sedation of invasively ventilated patients in intensive care units in Germany and Slovenia: an open-label phase 3 randomised controlled trial. Lancet Respir Med. 2021;9(11):1231-1240.