Conservative versus liberal oxygen therapy for mechanically ventilated adults: a systematic review with meta-analysis and trial sequential analysis

Optimal oxygen targets for mechanically ventilated adults remain uncertain. Prior meta-analyses included both ventilated and non-ventilated patients, limiting applicability to invasive ventilation. We performed a PRISMA-compliant systematic review and meta-analysis of randomized trials exclusively enrolling mechanically ventilated adults or reporting extractable IMV-specific data, comparing conservative (lower SpO 2 /PaO 2 targets) versus liberal (higher targets) oxygen therapy. The primary outcome was mortality at longest follow-up; secondary outcomes included 28/30-day, 90-day, ICU mortality and predefined adverse-event composites. Risk of bias was assessed with RoB 2 and certainty with GRADE. Trial sequential analysis (TSA) evaluated information size and robustness. Fourteen trials enrolling 22,930 mechanically ventilated adults were included. For the primary outcome of mortality at longest follow-up, data from all 14 trials ( n = 17,858 patients; 77.9% of the total cohort) showed no benefit of conservative oxygen therapy (RR 1.01, 95% CI 0.97–1.05; I²=0%), with outcome denominators reflecting trial-specific follow-up availability and the use of intention-to-treat or best-available analyzed populations. Similarly, no mortality reduction was observed at 28/30 days (7 trials; n = 18,446; RR 1.02, 95% CI 0.98–1.07; I²=0%), at 90 days (8 trials; n = 20,603; RR 1.02, 95% CI 0.98–1.05; I²=0%), or at ICU discharge (6 trials; n = 16,453; RR 0.93, 95% CI 0.74–1.17; I²=72.58%). Conservative oxygen increased composite cardiovascular adverse events (6 trials; n = 18,948; RR 1.19, 95% CI 1.00–1.42; I²=12.29%); other composites showed no clear differences. TSA indicated that the required information size was reached for mortality at longest follow-up, 28/30 days, and 90 days, whereas ICU mortality and adverse-event analyses remained statistically inconclusive. In mechanically ventilated adults, conservative oxygen therapy did not reduce mortality compared with liberal oxygen therapy and was associated with a possible increase in composite cardiovascular adverse events. These findings do not identify a single optimal oxygen strategy for this population.