Design: Pragmatic, open-label, randomised superiority trial
Population: 780 ventilated adults with acute hypoxaemic respiratory failure in 14 Canadian ICUs
Primary outcome: All-cause hospital mortality
Effect: 42.5% vs 43.8%; risk ratio 0.98 (95% CI 0.84–1.14)

SAVE-ICU was a pragmatic, open-label trial in 14 Canadian intensive care units. It randomised 780 mechanically ventilated adults with acute hypoxaemic respiratory failure, a quarter with COVID-19, to an isoflurane-based inhaled sedation strategy or usual intravenous sedation.

Hospital mortality was 42.5% with inhaled sedation and 43.8% with intravenous sedation (risk ratio 0.98, 95% CI 0.84–1.14). Ventilator-free and ICU-free days were similar. The inhaled group needed more rescue therapies for hypoxaemia (64% vs 55%) and had about one day fewer free of delirium or coma. About half of the inhaled group also needed intravenous sedatives on any given day.

Inhaled sedation had been promoted partly for possible lung-protective effects. This trial does not support adopting it as a routine strategy for this population.

Continue intravenous sedation as the standard for ventilated patients with hypoxaemic respiratory failure.
Do not expect inhaled isoflurane to reduce mortality in this group.
Watch for delirium if inhaled agents are used; delirium-free days were fewer.
Keep targeting light sedation with daily assessment, whatever the agent.

Why it matters

A promising sedation strategy failed its first large mortality test.

The statistics, in plain English

A risk ratio of 0.98 with an interval from 0.84 to 1.14 means no difference was found and a meaningful benefit or harm cannot be excluded entirely. The secondary findings on delirium and rescue therapies point the wrong way, but the trial was not powered for them.