Hyperoxia is Dose-Dependently Associated with an Increase of Unfavorable Outcomes in Ventilated Patients with Aneurysmal Subarachnoid Hemorrhage: A Retrospective Cohort Study

Abstract

Abstract Background Adequate oxygenation in patients with aneurysmal subarachnoid hemorrhage (SAH) is imperative. However, hyperoxia increases formation of reactive oxygen species and may be associated with a dose-dependent toxicity. We postulated a threshold for arterial partial pressure of oxygen (paO2) above which toxicity effects precipitate and sought to study the effects on 30-day mortality, favorable outcome at discharge and at 3 months, and delayed cerebral ischemia. Methods In this retrospective single-center cohort study, patients with SAH and mechanical ventilation > 72 h were included. Oxygen integrals were calculated above the following thresholds: 80, 100, 120, and 150 mm Hg and time-weighted mean paO2. All calculations were done from admission to end of day 1, day 3, and day 14. We conducted multivariable logistic regression analyses adjusted for age, sex, duration of ventilation, and Hunt and Hess grade. Time-weighted mean paO2 was categorized by quartiles. Favorable outcome was defined as Glasgow Outcome Scale scores of 4 and 5. Results From November 2010 to February 2021, 282 of 549 patients fulfilled the inclusion criteria. Odds ratios for 30-day mortality increased dose dependently and were as follows: 1.07 (95% confidence interval [CI] 1.03–1.11; p = 0.001) for each 1 mm Hg per day above 80 mm Hg; 1.16 (95% CI 1.07–1.27), above 100 mm Hg; 1.36 (95% CI 1.15–1.61), above 120 mm Hg; and 1.59 (95% CI 1.22–2.08), above 150 mm Hg (all p < 0.001) at day 14. For favorable outcome at 3 months, odds ratios were 0.96 (95% CI 0.92–0.99) for each 1 mm Hg per day above 80 mm Hg; 0.90 (95% CI 0.84–0.98), above 100 mm Hg; 0.83 (95% CI 0.72–0.97), above 120 mm Hg; and 0.77 (95% CI 0.61–0.97), above 150 mm Hg (all p < 0.05). For time-weighted mean paO2, lowest 30-day mortality and highest favorable outcome at 3 months were found in the second quartile (78–85 mm Hg). Thirty-day mortality increased above 93 mm Hg (fourth quartile), with an odds ratio of 3.4 (95% CI 1.4–8.4, p = 0.007). Odds ratios for favorable outcome at 3 months were 0.28 (95% CI 0.12–0.69), 0.27 (95% CI 0.11–0.67), and 0.24 (95% CI 0.10–0.59) for the first, third, and fourth quartiles, respectively (all p < 0.01). No significant association was found at day 1 and day 3, for favorable outcome at discharge, or for delayed cerebral ischemia. Conclusions Integrals above the defined paO2 thresholds were dose-dependently associated with an increase in mortality in ventilated patients with SAH. When we considered time-weighted mean paO2, unfavorable outcomes and 30-day mortality were more frequent both below and above a certain range. Unfavorable outcomes increased in paO2 ranges usually defined as normoxia. This emphasizes the necessity to further characterize oxygenation thresholds in ventilated patients with SAH in prospective clinical studies.