Updated Recommendations for Control of Surgical Site Infections

J. Wesley Alexander, MD, ScD; Joseph S. Solomkin, MD; Michael J. Edwards, MD


Annals of Surgery. 2011;253(6):1082-1093. 

In This Article

Effect of Oxygen Therapy

Killing of ingested microbes by phagocytic cells, especially neutrophils, involves oxidative processes (SDC-203-205). Neutrophils rapidly lose their ability to kill bacteria below a tissue PO2 of about 20 to 40 mm Hg (SDC204, SDC-206). As a consequence, reduction in the oxygen tension in wounds increases the incidence and severity of infections in humans (SDC-205-207). This is particularly important in surgical wounds (SDC-205) where oxygen tensions (Psq O2 10–55 mm Hg) are much lower than normal tissues (65 ± 7 mm Hg) and can reach near 0 mm Hg in infected wounds. In experimental animals, the incidence of wound infection is inversely related to tissue oxygen tension, and the effect is time-dependent, similar to the effectiveness of prophylactic antibiotics (SDC-208-210). The effects of oxygen and antibiotics are also additive. In one human study, an inverse correlation was observed between subcutaneous (Psq) O2 and the associated rate of postoperative wound infection.[15] When the maximum Psq O2 concentrations were 40 to 49 mm Hg, 43% of the patients developed postoperative infection, but when the subcutaneous oxygen tensions were at least 90 mm Hg, no patients developed wound infection (SDC-Fig. 2 at https://links.lww.com/SLA/A122). Another study (SDC-211) showed there was a significant difference in tissue oxygen saturation at the operative site 12 hours after operation in patients who developed an SSI (Psq O2 = 43.4%) compared to those who did not (Psq O2 = 55.8%). Thus, tissue oxygen concentration is an excellent predictor of wound infection. Subcutaneous Po2 is considerably lower in laparoscopic surgery compared to open procedures (SDC-212), consistent with a surprisingly high incidence of wound infections after laparoscopy. Obese patients also have considerably less Psq O2 with the same FIO2 compared to normal weight patients (SDC-213), perhaps contributing to the observed higher incidence of infections in obese patients.

These observational studies soon led to prospective randomized clinical trials comparing patients receiving 80% inspired oxygen versus those receiving 30% oxygen. In a well-controlled study of 500 patients with colorectal resections, 5.2% of the patients receiving 80% oxygen developed wound infections compared to 11.2% of patients receiving 30% oxygen.[16] The mean intraoperative subcutaneous PO2 were 109 ± 43 and 59 ± 15 mm Hg, respectively within each subgroup. In a double-blinded controlled trial of 300 patients undergoing elective colorectal surgery, patients were randomly assigned to receive either 80% or 30% FIO2 intraoperatively and for 6 hours after surgery (SDC-214). Surgical site infections occurred in 14.9% in those receiving 80% FIO2 compared to 24.4% of the patients receiving 30% FIO2 (RR = 0.46). Another study (SDC-215) compared complications in 2025 patients with surgeries expected to last 2 hours or more, randomized to receive inspired gas with 80% O2/20% nitrogen versus 30% O2/70% nitrous oxide. Patients with the higher FIO2 had fewer infections (7.7% versus 10%, RR = 0.72, P = 0.036). However, a role of nitrous oxide separate from the effect of FIO2 could not be excluded. Continuing positive airway pressure has been reported to increase oxygenation and reduce wound infection (RR = 0.27) in postoperative patients receiving 50% O2 (SDC-216).

In contrast to demonstrated effectiveness in these well-controlled studies, there have been 4 studies, which have failed to show reduction in postoperative infections with an increased FIO2. The first negative study (SDC-217) of 160 randomized patients, with a nonstandardized group of abdominal operations, compared FIO2 80% with FIO2 35%, but was underpowered and not homogenous with additional methodological problems. Importantly, more than 50% of these patients had laparoscopically assisted surgeries. The second study (SDC-218) had 143 patients undergoing Cesarean section but was stopped early. The third study (SDC-219) had 38 patients undergoing colectomy and showed a nonsignificant trend toward benefit but was underpowered. The fourth and most important of these studies enrolled 1400 patients undergoing laparotomy (SDC-220) and showed similar rates of overall infection. However, deep wound infections occurred in 2.9% of patients with 80% O2 versus 3.7% of patients with 30% O2, a reduction of 21.6%. More importantly, most patients received epidural anesthesia, which could increase subcutaneous blood flow. No significant adverse effects of 80% FIO2 have been observed in any of the studies.

Three meta-analyses using MEDLINE and Cochran databases have been published within the last 2 years (SDC-221-223), and all have shown a significant reduction of wound infection in patients receiving hyperoxia compared to those that did not (RR, 0.68–0.74), but these exclude the last negative study. The benefit was seen mostly in colorectal procedures but hyperoxia may be beneficial in other types of surgery as well. In a case-controlled study of spinal surgery, SSI occurred more frequently in patients receiving FIO2 less than 50% during the procedure (OR = 12 after controlling for other variables) (SDC-224).


There is little doubt that low subcutaneous concentrations of O2 at wound sites impairs the antibacterial functions of neutrophils and promotes the development of wound infections. Some studies show that increasing inspired O2 from 30% to 80% results in improved O2 tension at wound sites, increasing the ability to kill contaminating bacteria and decreasing the incidence and severity of wound infections. However, not all studies show a benefit, most likely because of other factors, which alter tissue PO2 that may alter the effectiveness of hyperoxia, including body temperature, blood pressure, smoking, type of anesthesia, noncomparable patient populations, differing practice of fluid replacement, use of vasopressors, type of operation and tissue handling. Supplemental O2 should start with induction, but optimal concentration and duration of therapy have not been established. Current data would suggest it should be given for at least 2 hours after closure.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.