Hypothesis: To determine whether the application of a strict conservative protocol for oxygen supplementation to maintain PaO2 within physiologic limits could improve outcomes in critically ill ICU patients.
Methods: Four-hundred and thirty-four subjects were randomized to receive either conventional O2 therapy (PaO2 up to 150 mmHg) or conservative O2 therapy (PaO2 70-100 mmHg). Other treatment decisions were left to the individual physicians.
Results: The trial was stopped early after recruitment of 73% of the planned subjects due to an earthquake at the study site. ICU mortality was lower in the conservative O2 group vs the conventional group (11.6% vs 20.2%; RR, 0.57 [95% CI, 0.37-0.90]).
Strengths: RCT, leaving other treatment decisions to individual physicians increases generalizability, primary outcome (mortality) is highly relevant
Weaknesses: Single-center, unplanned interim analysis, stopped trial early
Key Point: A conservative oxygen strategy results in lower mortality in critically-ill patients in the ICU.