Association between oxygen saturation ranges and mortality among mechanically ventilated sepsis-associated ARDS patients: a multicenter retrospective cohort analysis.

BACKGROUND: The SpO2 range associated with the best survival outcomes for acute respiratory distress syndrome (ARDS), particularly in sepsis-related cases, remains unclear. This study aimed to explore oxygenation range for patients with sepsis and ARDS during mechanical ventilation. METHODS: We analyzed 2951 ICU patients with sepsis and ARDS requiring mechanical ventilation. Using restricted cubic splines (RCS) and generalized additive models (GAM), we identified optimal oxygenation ranges, which were further cross-replicated across independent ICU datasets via logistic regression, propensity score matching, IPW, and double robustness analysis. The primary outcome was 28-day all-cause mortality. RESULTS: In this cohort, the range of oxygen saturation (SpO2) with 93%-99% was associated with the lowest 28-day mortality in patients with sepsis and ARDS, as determined by RCS and GAMs. These findings were verified via multivariate logistic regression, propensity score matching, IPW, and double-robust adjustment incorporating all available covariates. This range SpO2 window remained consistent across mild, moderate and severe ARDS subgroups. Patients with SpO2 < 93% had higher SOFA, APACHE II and SAPS II scores. Infection source did not significantly modify the SpO2-mortality association. CONCLUSION: SpO2 ranging from 93% to 99% was independently associated with reduced 28-day all-cause mortality in mechanically ventilated patients with sepsis-related ARDS in this retrospective cohort.