Hyperoxia Ineffective in Preventing Surgical Site Infections

Abstract & Commentary

Synopsis: Patients who received hyperoxia during general surgery had an increase in surgical site infections compared to those who received a lower oxygen concentration.

Source: Pryor KO, et al. JAMA. 2004;291:79-87.

In 2000, Greif and colleagues reported that delivery of 80% oxygen in the perioperative period produced a 50% reduction in the incidence of surgical wound infection in patients undergoing open colorectal resection, compared with patients who received 30% oxygen.1,2 Commentary was cautious and most recommended additional clinical trials. The present study was undertaken to further examine the effects of hyperoxia in this patient population.

Subjects were 165 patients scheduled to undergo a variety of major abdominal surgical procedures, eg, colectomy, abdominoperineal resection, gastrectomy, exploratory laparotomy, gynecologic staging/debulking, in an academic medical center. The study used a double-blind design. After consent was obtained, patients were randomized to receive a fraction of inspired oxygen (FIO2) of 0.35 or a FIO2 of 0.80 during and for 2 hours after surgery, using sequentially numbered envelopes generated from a random numbers table. The surgical team was blinded to group assignment, as was the investigator, who evaluated patients following surgery for evidence of infection. All patients were evaluated for evidence of infection for 14 days after surgery, regardless of discharge date.

A total of 29 patients (18.1%) developed surgical site infections, 9 (11.3%) in the FIO2 0.35 group and 20 (25%) in the FIO2 0.80 group (P = 0.02). Patients who developed an infection had a longer length of hospitalization than those who did not develop infection (mean, 13.3 ± 9.9 days vs 6.0 ± 4.2 days; P < 0.001). Hospitalization was longer in the FIO2 0.80 group, but the difference was not significant (mean, 8.3 ± 7.5 vs 6.4 ± 4.7 days; P = 0.07). Four patients who developed infection required re-operation; all were in the FIO2 0.80 group. In multivariate logistic regression analysis, FIO2 remained predictive of infection (P = 0.03). Five patients in the group with FIO2 0.80 and one in the group with FIO2 0.35 remained intubated at the end of surgery, which was the only other significant predictor of infection. All other examined variables were not predictive.

Comment by Leslie A. Hoffman, RN, PhD

As Pryor and colleagues note, the physiologic changes and interactions that result from a substantial increase in arterial oxygen tension are multiple, complex, and difficult to study in vivo. A high oxygen partial pressure will increase the production of reactive oxygen species and a number of these reactive oxygen species are components of bactericidal host defenses. Consequently, there is support for the potential benefit of hyperoxia as a means to prevent surgical site infections. However, reactive oxygen species are also involved in processes that produce tissue injury and inhibit antibacterial mechanisms. Thus, there is also support for potential harm from this therapy.

The protocol in this study was designed to model routine use of hyperoxia in abdominal surgery associated with a high risk of infection. A large number of patients and management characteristics were examined, eg, type of skin closure (suture, staples), duration of surgery, temperature at extubation, antibiotic use, fluid management, type of anesthetic agent, length of postoperative stay, and few differed between groups at baseline. There was a slightly higher body mass index in the group that received a FIO2 of 0.80 (P = 0.04), with a corresponding higher incidence of obesity, defined as a body mass index > 30 (P = 0.04), but these variables were not significant predictors in the multivariate analysis. There was also a higher estimated blood loss (P = 0.03) and higher crystalloid administration (P = 0.02). However, the absolute difference was not in the clinically relevant range and neither were significant predictors.

The possibility exists that an unexamined practice element led to the divergent findings reported from this study and the prior study that found benefits from hyperoxia, but this was not evident from study findings. Results of this study suggest no benefit, and the potential of harm, from using a high FIO2 during major abdominal surgery as a means to reduce the incidence of surgical site infections.


1. Greif R, et al. N Engl J Med. 2000;342(3):161-167.

2. Pierson D. Critical Care Alert. 2002(Feb);11(9):121-123.

Leslie A. Hoffman, PhD, RN, Department of Acute/Tertiary Care, School of Nursing, University of Pittsburgh, is Associate Editor of Critical Care Alert.