Isoflurane Sedation Did Not Lower Mortality in Severe Hypoxemic Respiratory Failure
In a pragmatic randomized trial spanning 14 Canadian intensive care units, isoflurane-based inhaled sedation did not improve survival for mechanically ventilated adults with acute hypoxemic respiratory failure. Hospital mortality was nearly identical between groups, despite the biologic appeal of a volatile anesthetic strategy in a population that often requires deep sedation, high ventilatory support, and prolonged ICU care.
The SAVE-ICU trial enrolled 786 adults and included 780 in the primary analysis. Mean age was 61.1 years, 34.0% were female, and 73.7% had non-COVID AHRF. The cohort was sick by any bedside measure: the mean Pao2:Fio2 ratio was 152.6 mm Hg, a level consistent with substantial gas-exchange impairment. Patients were randomized to isoflurane with intravenous adjuncts as needed or to usual intravenous sedation, and sedatives were titrated to target sedation scores until extubation or death.
Hospital discharge mortality was 42.5% in the inhaled sedation group and 43.8% in the intravenous sedation group. The risk ratio of 0.98, with a 95% confidence interval crossing unity, left no room for a clinically meaningful mortality signal. Prespecified subgroup analyses by age, sex, and COVID-19 status also failed to show effect modification.
A Sedation Strategy With No Survival Advantage
The absence of a mortality difference is striking because volatile anesthetics have long carried mechanistic appeal in the ICU. Isoflurane can be delivered directly through the ventilator circuit, can reduce sedative exposure from conventional intravenous agents, and has been discussed as a potential option for patients in whom sedation depth and ventilator synchrony are difficult to maintain. In theory, a more controllable inhaled agent could reduce respiratory drive mismatch, blunt agitation, and potentially limit the cascade of ventilator dyssynchrony, excess oxygen demand, and escalation of ventilatory support.
In this trial, those theoretical advantages did not translate into a survival benefit. The data argue against assuming that a different sedative delivery route will improve outcomes simply because it is pharmacologically distinctive. In critically ill patients with severe hypoxemia, mortality is usually driven by the underlying lung injury, multiorgan dysfunction, and downstream complications of critical illness rather than by sedative class alone.
More Hypoxemia Adjuncts and Fewer Delirium-Free Days
The inhaled sedation arm required more adjunctive therapies for refractory hypoxemia: 63.9% versus 55.0%, a between-group difference of 8.89%. This is an important bedside signal, even if it does not prove causation. Patients with worse oxygenation often need prone positioning, neuromuscular blockade, recruitment maneuvers, or other rescue interventions, and the greater use of adjunctive therapies suggests that the inhaled-sedation cohort may have carried a heavier burden of respiratory instability during the ICU course.
The trial also found fewer delirium- and coma-free days in the inhaled sedation group, with a between-group difference of -0.95 days. That finding matters because neurological recovery is central to liberation from mechanical ventilation, ICU throughput, and later functional status. Sedation strategy can influence the balance between comfort and oversedation, and any regimen that deepens the burden of coma or delirium can complicate extubation readiness, communication, and rehabilitation. Even so, those neurologic differences did not cascade into measurable gains or losses in the hard outcomes of ventilator-free days, ICU-free days, or hospital-free days.
Concomitant Intravenous Sedation Remained Common
One practical feature of the trial deserves attention: in the inhaled sedation group, 49.9% to 55.4% of patients received concomitant intravenous sedation on any day during the first 14 days after randomization. In other words, inhaled therapy was rarely a pure stand-alone strategy. That pattern reflects ICU reality. Clinicians often need to layer agents to meet sedation targets, blunt dyspnea, and maintain ventilator synchrony, especially when patients are severely hypoxemic or agitated.
Because of that overlap, SAVE-ICU tested a real-world sedation pathway rather than a sterile pharmacology experiment. The result is still informative. If a combined inhaled-plus-intravenous approach fails to outperform usual intravenous care in a large pragmatic cohort, routine adoption becomes difficult to justify outside specific local circumstances or research settings.
No Signal in Functional Recovery at 3 Months
Quality-of-life and disability scores at 3 months were similar between the two strategies. That finding closes off another possible argument for inhaled sedation: even if survival were unchanged, perhaps recovery might improve by reducing cumulative exposure to conventional sedatives or easing the course of mechanical ventilation. SAVE-ICU did not show that kind of downstream benefit.
The trial’s broad setting strengthens the relevance of the result. Fourteen Canadian ICUs, enrollment over several years, and inclusion of both non-COVID and COVID-19-related AHRF make this a clinically grounded test of a strategy that had generated interest but little rigorous comparative evidence. Pragmatic trials often disappoint enthusiasts because they capture the messy conditions of routine care rather than the cleaner environment of physiologic proof-of-concept studies. Here, that messiness is exactly what makes the negative result useful.
What the Trial Means for ICU Sedation Practice
For mechanically ventilated adults with AHRF, isoflurane did not outperform usual intravenous sedation on the outcome that matters most: hospital survival. The absence of benefit on ventilator-free days, ICU-free days, hospital-free days, and longer-term functional measures leaves little support for widespread use of inhaled sedation as a mortality-reducing strategy.
The trial also leaves open a narrower question: whether selected ICU subgroups, managed in centers with substantial inhaled-anesthetic expertise, might derive logistical or physiologic advantages not captured in this broad population. The current evidence does not answer that. What it does answer is more immediate. Routine isoflurane-based sedation should not be expected to rescue outcomes in severe hypoxemic respiratory failure, and the extra complexity of delivering inhaled therapy needs a stronger rationale than this trial provides.
ClinicalTrials.gov Identifier: NCT04415060.