High Inspired Oxygen Fraction and Surgical Site Infection

Evidence continues to accumulate on whether the use of high inspired oxygen fraction (FIO₂) is effective for preventing surgical site infections (SSIs). In 2000, a randomized trial by Greif et al¹ demonstrated that SSIs were significantly decreased following colon surgery in patients who received 80% oxygen intraoperatively and for the first 2 hours following surgery. Subsequent clinical trials by Belda et al² and by Myles et al³ supported the use of perioperative supplemental high inspired oxygen for reducing risk of surgical wound infection, whereas a clinical trial by Pryor et al⁴ suggested that perioperative hyperoxia was not effective in reducing SSIs (and in fact increased them). A trial by Mayzler et al⁵ reported only 5 of 38 patients as developing SSI overall and thus had inadequate power to draw conclusions. A meta-analysis⁶ of these trials, pooling the outcomes of 3001 patients, found that perioperative administration of high inspired oxygen (80% concentration) was associated with a 3% absolute reduction (crude infection rates of 12% in the control group and 9% in the group receiving 80% oxygen) and a 25% relative reduction in risk of SSI.

In this issue of JAMA, Meyhoff and colleagues⁷ report the results of the PROXI trial, in which 1400 patients in 14 Danish hospitals undergoing acute or elective laparotomy were randomized to receive either 80% oxygen or 30% oxygen during and for 2 hours following their operations. The primary finding was that there were no statistically significant differences in rates of SSI with high FIO₂ (131 patients [19%] in the 80% oxygen group vs 141 [20%] in the 30% oxygen group, P = .64). Likewise, there were no statistically significant differences in rates of pulmonary complications, including atelectasis, pneumonia, or respiratory failure between the groups. These findings add to the evidence base surrounding the potential role of high FIO₂ for prevention of SSIs.

Previous laboratory and clinical studies have shown that the partial pressure of oxygen (PO₂) in wounds is important for healing.⁸ PO₂ is low in surgical wounds as a result of injury, coagulation, inflammation, and, in large measure, the sympathetic nervous system stimulation and consequent vasoconstriction caused by hypothermia,⁹ hypovolemia,¹⁰ and pain.¹¹ This wound hypoxia slows healing but can be corrected. For instance, Hartmann et al¹² assessed wound PO₂ and collagen deposition in 2 groups of patients after major abdominal operations and found that, compared with control patients given fluids per an accepted formula, those given additional fluids in response to low wound PO₂ had significantly higher wound PO₂ levels and significantly more collagen deposition.

Moreover, oxygen also is an important factor for eradication of infection. Studies using experimental wound models have demonstrated that Pseudomonas aeruginosa, Staphylococcus aureus, and Escherichia coli injected into wounds could be eradicated at rates proportional to FIO₂ or PO₂^{13 - 14} and that antibiotics were increasingly effective at higher FIO₂.¹³ Superoxide production by leukocytes, the index of oxidative killing, has been shown to be proportional to PO₂, such that production is negligible at a PO₂ of 0, half maximal at 3 kPa (near the usual PO₂ of a surgical wound), and maximal at about 15 kPa.¹⁴ In a prospective, observational study of surgical patients, vulnerability to infection was inversely proportional to wound PO₂.¹⁵ Oxidative killing by leukocytes has been cited as a possible mechanism¹⁶ for this effect. Kohanski et al¹⁷ demonstrated another potential mechanism of action for oxygen: the main groups of bactericidal antibiotics require oxygen to exert their lethal effects in the pathogen.

However, in the discussion over whether high FIO₂ is good, toxic, effective, or immaterial for prevention of SSIs, the use of 80% oxygen has become the focus of the debate. Wound PO₂ can be increased by supplemental oxygen, but only in the absence of vasoconstriction¹⁰ and under conditions conducive for supplemental oxygen to reach the periphery, including rigorous temperature control and appropriate fluid repletion.

In a clinical trial by Kurz et al,¹⁸ patients in whom perioperative normothermia was maintained had a significant reduction in SSIs and a significant increase in collagen deposition compared with control patients with hypothermia. Owing in large part to that study, maintenance of euthermia was included in guidelines for SSI prevention.¹⁹

In the clinical trial by Greif et al,¹ in which liberal fluid replacement and maintenance of euthermia were mandatory, wound PO₂ increased significantly and was sufficient to account for the reduced SSIs in the group receiving 80% oxygen. The trial by Belda et al,² using an almost identical protocol, showed a 40% relative reduction in SSIs.

In the PROXI trial, anesthetists were allowed to practice within the range of anesthesia management standard at their institutions, with the exception of the treatment group (inspired oxygen). However, compared with the trials by Greif et al and Belda et al, fluid volumes administered in the PROXI trial were smaller, with a goal that postoperative weight gain would be less than 1 kg; furthermore, normothermia was not maintained in all patients, as evidenced by end-of-surgery temperatures apparently as low as 35°C in some patients. Because these issues raise the possibility of vasoconstriction, the study findings would have been bolstered if wound PO₂ had been measured to demonstrate that supplementary oxygen reached the wounds. On the other hand, the PROXI trial found no evidence that high FIO₂ increased the risk of pulmonary complications such as atelectasis or pneumonia, allaying concerns about potential pulmonary toxicity. However, as Meyhoff et al point out, there was a nonsignificant increase in rates of respiratory failure (5.5% vs 4.4%) and 30-day mortality (4.4% vs 2.9%) in the 80% oxygen group compared with the 30% oxygen group, findings in contrast to other trials using 80% vs 30% oxygen.

The PROXI trial provides additional evidence to help inform the ongoing debate over whether the use of high FIO₂ is effective for preventing SSIs. Whether high FIO₂ should become a standard for most operations remains to be determined. In the meantime, ensuring rigorous perioperative care, including adequate fluid administration, maintenance of normothermia and normoglycemia, and appropriate use of perioperative antimicrobial agents will remain essential approaches for prevention of SSIs.