Isoflurane Sedation Did Not Lower Mortality in Ventilated Adults With Acute Hypoxemic Respiratory Failure
In mechanically ventilated adults with acute hypoxemic respiratory failure, an inhaled sedation strategy built around isoflurane did not improve the outcome that mattered most: survival to hospital discharge. The SAVE-ICU randomized trial enrolled a broad ICU population across 14 Canadian centers and found nearly identical hospital mortality in the inhaled and standard intravenous sedation groups. The result is clinically sobering because the biological appeal of inhaled anesthetic sedation has always been strong: rapid titratability, pulmonary delivery, and a potential ability to preserve hemodynamic stability while reducing the burden of deep intravenous sedation.
The trial included 786 randomized patients, with 780 analyzed in the primary population. Mean age was 61.1 years, 34.0% were female, and the majority had non-COVID respiratory failure. Baseline oxygenation was poor, with a mean Pao2:Fio2 ratio of 152.6 mm Hg, placing many patients squarely in the zone where sedation, ventilator synchrony, and hypoxemia management often become inseparable at the bedside.
One sentence captures the central result: hospital mortality was essentially the same in both groups.
At discharge, 42.5% of patients in the isoflurane-based arm had died, compared with 43.8% in the intravenous sedation arm. The risk ratio was 0.98, with a 95% confidence interval from 0.84 to 1.14 and a P value of .85. No prespecified subgroup analysis suggested a different signal for age, sex, or COVID-19 status. In practical terms, the inhaled strategy did not move the mortality needle in any direction that would justify routine substitution for standard intravenous sedation on the basis of survival alone.
Sedation Delivery, Adjunctive Drugs, and Why the Strategy May Have Blunted Its Own Signal
The trial design allowed clinicians to use intravenous adjuncts as needed in the inhaled sedation group, and that detail shapes interpretation. Between 49.9% and 55.4% of patients in the inhaled arm received concomitant intravenous sedation on any day during the first 14 days after randomization. In other words, the intervention was not a pure isoflurane-only exposure; it was a pragmatic hybrid approach meant to reflect real ICU practice, where analgesia, anxiety control, ventilator dyssynchrony, and delirium risk often force multiple sedative classes into the same treatment plan.
That hybridization may have diluted any unique pharmacologic advantage of inhaled sedation. Isoflurane has a fast on-off profile and can be useful when rapid neurologic assessment or quick titration is needed, but mechanically ventilated patients with severe hypoxemia often require layered sedation to tolerate lung-protective ventilation, prone positioning, or other adjunctive measures. Once intravenous medications are added back into the regimen, the theoretical separation between strategies narrows, and so does the chance of seeing a large mortality difference.
The enrolled population also had enough illness severity to overwhelm modest changes in sedative pharmacology. Acute hypoxemic respiratory failure is driven by diffuse alveolar injury, shunt physiology, inflammatory capillary leak, and prolonged ventilator dependence. Sedation can improve synchrony and reduce oxygen consumption, but it does not reverse the underlying parenchymal process. A sedative strategy may influence day-to-day management, yet still fail to alter the downstream cascade of ventilator days, ICU recovery, and death.
Hypoxemia Adjuncts, Delirium Burden, and the Tradeoffs of Inhaled Sedation
The inhaled sedation arm used more adjunctive therapies for hypoxemia: 63.9% versus 55.0% in the intravenous sedation group, an absolute between-group difference of 8.89 percentage points. This is not a trivial operational finding. More rescue therapy suggests that patients in the isoflurane arm may have required additional bedside interventions to achieve oxygenation goals, maintain ventilator compatibility, or respond to persistent respiratory instability. The signal raises a simple clinical question: whether inhaled sedation made some aspects of ICU care more complex rather than less.
The neurologic recovery signal was also unfavorable. Patients in the inhaled sedation group had fewer delirium- and coma-free days by day 14, with a between-group difference of -0.95 days. That finding does not prove harm from isoflurane itself, but it does show that the intervention did not deliver a cleaner neurologic course. In critically ill adults, delirium and coma are not administrative endpoints; they affect ventilator liberation, early mobilization, communication with family, and the pace of recovery after the acute insult has passed.
Quality-of-life and disability scores at 3 months did not differ between groups. This matters because a therapy that fails to improve survival might still be worthwhile if it produces a better recovery trajectory. Here, no such advantage emerged. The absence of a benefit across mortality, short-term liberation metrics, and later patient-centered outcomes makes it difficult to argue for routine adoption of this inhaled strategy as a superior default.
What the SAVE-ICU Trial Means for ICU Sedation Practice
SAVE-ICU was a pragmatic, open-label superiority trial, which is the right design for a question rooted in ICU workflow rather than pharmacology alone. It tested a realistic bedside strategy across 14 intensive care units in Canada and followed patients through June 2025. Adults were enrolled between September 2020 and March 2025, and eligibility required mechanical ventilation with sedation for less than 72 hours in the setting of non-COVID respiratory failure or COVID-19-related respiratory failure.
The trial’s core message is not that isoflurane has no physiologic effect. It almost certainly does. The point is narrower and more important: in this population, those physiologic properties did not translate into better survival. For intensivists, that leaves the inhaled approach as a tool with niche utility rather than a broadly superior sedation strategy for adults with AHRF.
Bedside adoption will now hinge on more selective questions. Can inhaled sedation reduce deep intravenous drug exposure in specific subgroups? Does it improve ventilator synchrony in patients with extreme dyssynchrony or difficult weaning? Could there be benefit in centers with substantial expertise and streamlined delivery systems? SAVE-ICU does not answer those questions, and it should not be stretched beyond the population it studied.
The most defensible reading of the data is direct and restrained: isoflurane-based inhaled sedation, used with intravenous adjuncts as needed, did not improve hospital mortality compared with usual intravenous sedation in mechanically ventilated adults with acute hypoxemic respiratory failure. For now, survival remains unchanged, and standard ICU sedation strategies keep their place at the center of care.