Supplemental Oxygen Reduces Surgical Infection

Overview

Supplemental perioperative oxygen has been reported to halve or double the risk of surgical wound infection. We tested the hypothesis that supplemental oxygen reduces infection risk following colorectal surgery. Colorectal surgery patients (n=300) were randomly assigned to 30% or 80% FiO2 intraoperatively and 6 hours postoperatively. Wound infections were diagnosed by blinded investigators using Centers for Disease Control criteria. Baseline patient characteristics, anesthetic management, and potential confounding factors were recorded. Wound infection rates were compared with chi-square analysis. Logistic regression identified the contribution of potential confounding factors. Surgical wound infection occurred in 24.4% of patients receiving 30% oxygen, but only 14.9% of those receiving 80% oxygen (P\<0.039). According to logistic regression, the relative risk of infection in patients given supplemental oxygen was 0.46 (P=0.035). Supplemental inspired oxygen reduced wound infection risk by roughly a factor of two. We thus recommend that supplemental oxygen be provided to patients undergoing colorectal surgery.

Context: Supplemental perioperative oxygen has been variously reported to halve or double the risk of surgical wound infection. Objective: We tested the hypothesis that supplemental oxygen reduces infection risk following colorectal surgery. Design: Randomized, controlled trial. Setting: Fourteen Spanish hospitals. Patients: Three hundred patients undergoing elective colorectal surgery. Interventions: Patients were randomly assigned to either 30% or 80% FiO2 intraoperatively and for 6 hours after surgery. Anesthetic management and antibiotic administration were standardized. Main outcome measures: Wound infections were diagnosed by blinded investigators using Centers for Disease Control criteria. Baseline patient characteristics, anesthetic management, and potential confounding factors were recorded. The wound infection rate in the 30% and 80% oxygen groups was compared with chi-square analysis. A logistic regression was used to identify the contribution of potential confounding factors.