Supplemental perioperative oxygen for reducing surgical site infection: a meta-analysis


Dr Nasia Safdar
University of Wisconsin Medical School
H4/572 600 Highland Avenue
WI 53792


Objective  To assess the efficacy of supplemental perioperative oxygenation for prevention of surgical site infection (SSI).

Data sources  Computerized PUBMED and MEDLINE search supplemented by manual searches for relevant articles.

Study selection  Randomized, controlled trials evaluating efficacy of supplemental perioperative oxygenation versus standard care for prevention of SSI in patients’ undergoing colorectal surgery.

Data synthesis  Data on incidence of SSI were abstracted as dichotomous variables. Pooled estimates of the relative risk (RR) and 95% confidence interval (CI) were obtained using the DerSimonian and Laird random effects model and the Mantel-Haenzel fixed effects model. Heterogeneity was assessed using the Cochran Q statistic and I2.

Results  Four randomized controlled trials met the inclusion criteria. Supplemental perioperative oxygenation resulted in a reduced incidence of SSI [RR 0.70 (95% CI 0.52–0.94), P = 0.01], using a fixed effects model. Using the more conservative random effects model, the point estimate was similar [RR 0.74 (95% CI 0.39–1.43), P = 0.37], but the results failed to achieve statistical significance. The I2 test showed moderate heterogeneity.

Conclusions  Our analysis showed that supplemental perioperative oxygenation is beneficial in preventing SSI in patients undergoing colorectal surgery. Because of heterogeneity in study design and patient population, additional randomized trials are needed to determine whether this confers benefit in all patient populations undergoing other types of surgery. Supplemental perioperative oxygenation is a low-cost intervention that we recommend be implemented in patients undergoing colorectal surgery pending the results of further studies. Further research is needed to determine whether or not supplemental hyperoxia may cause unanticipated adverse effects.