Invited Commentary 

August 7, 2024

Is Too Much Oxygen a Bad Thing?

Bryant Fisher, Kathirvel Subramaniam, Danny Chu

JAMA Surg. Published online August 7, 2024. doi:10.1001/jamasurg.2024.2921

The current referenced double-blind randomized clinical trial by Lopez and others¹ compared strategies of intraoperative hyperoxia vs normoxia during elective cardiac surgery to assess the impact on oxidative stress and clinical markers of end-organ damage including acute kidney injury (AKI). The primary end points of the study were changes in the levels of F2-isoprostanes and isofurans as markers of oxidative stress and serum creatinine from baseline to postoperative day 2. The authors conclude that intraoperative management with hyperoxia did not affect clinical outcomes in the postoperative period despite a significant increase in intraoperative serum oxidative stress markers.

The study is well designed with stringent protocols for management of oxygenation to ensure maximum treatment effects. However, the study may be limited in its ability to detect real clinical differences due to the decision of using oxidative stress serum markers as a mechanistic primary end point. Despite availability of validated criteria for AKI diagnosis, such as RIFLE,² the investigators chose numerical change in serum creatinine level as a primary clinical outcome. In addition, the effect size for the sample size calculation is based on a previous study³that was seemingly stopped for futility. While there appears to be a signal for increased risk of atrial fibrillation, delirium, and other compositive outcomes of end-organ damage with intraoperative hyperoxia, the study’s underpowered design prevented detection of potentially significant clinical differences in the treatment cohorts. The investigators conclude that a liberal oxygenation strategy significantly increased oxidative stress as measured by the levels of F2-isoprostanes and isofurans. Yet, of the 6 time points that these levels were measured, only 1 of the time points succeeded in reaching statistical significance with an increase of 9.2 pg/mL from baseline (eTable 10 in the Supplement in the Lopez et al article¹).

Despite these limitations, certain aspects of the study certainly warrant further investigation. Compared to those with levels of oxidative stress markers below the 25th percentile, participants with levels higher than the 75th percentile experienced significantly greater odds of AKI and delirium. These results corroborate previous findings by the investigators in their cohort study showing an independent association between levels of F2-isoprostanes and isofurans and postoperative delirium.⁴ These findings may reflect effects of hyperoxygenation or underlying impaired response of these patients to oxidative stress. Numerous other randomized trials have addressed this question with mixed results,^5-7 with 1 trial showing no effect on postoperative delirium and another showing potential reduction of incidence of AKI with a hyperoxygenation strategy. Nonetheless, the referenced study provides evidence of safety for hyperoxygenation strategy, but lingering questions may necessitate a larger study. Ideally, the oxygenation strategy should extend into the postoperative period to fully assess its effect on clinical outcomes. A multicenter, pragmatic trial may be beneficial in capturing these clinical end points and improving study power with increased sample size.