Perioperative oxygen therapy: an overview of systematic reviews and meta-analyses

Perioperative oxygen is routinely used, but evidence of its clinical impact remains inconsistent, leading to variable practice. The aim of this work was to provide a comprehensive overview of the effectiveness of perioperative oxygen therapy strategies.

We searched multiple databases for systematic reviews comparing perioperative oxygen strategies. Two reviewers independently extracted data. The co-primary outcomes were surgical site infection (SSI) and mortality. We classified systematic reviews with the largest number of trials as anchoring reviews. We updated anchoring reviews with data from more recent RCTs. We assessed the risk of bias of the anchoring reviews using the ROBIS tool, updated meta-analyses and subgroup analyses, and undertook exploratory meta-regression. We assessed the certainty in evidence using GRADE framework and conducted trial sequential analysis.

Fifty-nine systematic reviews met the inclusion criteria, from which five anchoring reviews were selected. Perioperative high fraction of inspired oxygen (Fio₂, 80%), compared with a low Fio₂ (30-35%), may reduce the incidence of SSI slightly (risk ratio [RR] 0.87, 95% confidence interval [CI] 0.76-1.01; risk difference [RD] 1.6% lower, 3% lower to 0.1% higher), but the evidence is very uncertain. High inspired oxygen may result in little to no difference in mortality (RR 1.17, 95% CI 0.77-1.78, RD 0.3% higher, 0.4% lower to 1.3% higher), based on low-certainty evidence. The evidence suggests that high Fio₂ results in a large increase in the incidence of atelectasis (RR 1.47, 95% CI 1.20-1.79, RD 6.5% higher, 2.8% higher to 10.9% higher, low-certainty evidence). Postoperative noninvasive ventilation (NIV) does not reduce mortality compared with conventional oxygen therapy (COT; RR 0.91, 95% CI 0.62-1.32, RD 0.1% lower, 0.6% lower to 0.5% higher), based on high-certainty evidence. Low-certainty evidence suggests that postoperative high-flow nasal oxygen (HFNO) compared with COT does not reduce mortality (RR 0.78, 95% CI 0.27-2.24, RD 0.4% lower, 1.4% lower to 2.4% higher). Low- to very low-certainty evidence indicates that postoperative NIV and HFNO may reduce some of the pulmonary adverse events compared with COT. Trial sequential analyses showed that further studies are required to determine which perioperative oxygen strategy is most clinically and cost effective.

We did not find evidence to support routine use of high inspired oxygen to reduce surgical site infection and improve patient outcomes. A small reduction in surgical site infection associated with high Fio₂ cannot be ruled out and possible effect modifiers require further investigation. Existing evidence favours postoperative noninvasive ventilation and high-flow nasal oxygen over conventional oxygen therapy, but the low to very low certainty of evidence limits our confidence in the findings.

PROSPERO CRD42021272361.