Peri-Operative Oxygen and the Risk of Surgical Infection

Abstract

Background:

In spite of all efforts, surgical site infection (SSI) continues to be a common and costly complication of surgical procedures and thus a major concern for surgeons, patients, and hospitals. Laboratory observations that high concentrations of oxygen in tissues enhance bacterial killing and improve wound healing have prompted the study of peri-operative oxygenation as a means of reducing SSI.

Methods:

Review of pertinent English-language literature.

Results:

Seven randomized, controlled studies of increased peri-operative oxygenation during and shortly after general anesthesia have been conducted since 2000, with four showing no benefit (including one terminated early because of possible harmful effects) and three showing positive effects, with risk reductions of 25%–50%. Three meta-analyses also have been conducted and concluded that there likely is a beneficial effect of increased peri-operative oxygenation.

Conclusions:

Given the divergent results of this relatively large number of randomized studies, no definitive consensus has emerged. At present, clinicians should proceed with caution, and the administration of greater amounts of supplemental oxygen probably should be restricted to well-designed and -conducted clinical trials.

Surgical site infection (SSI) is a relatively common complication of many operations and has been associated with higher cost, longer hospital stays, and higher mortality rates [1 –5]. Interest in cleansing wounds to promote healing was documented as early as 1750 B.C.E. in the Code of Hammurabi, where the ancient Sumerians used beer to cleanse wounds. This manner of treating wounds persisted into the era of Hippocrates (circa 400 B.C.E.), who recognized that pus was not part of normal wound healing. One can surmise that the alcohol content of the beverages was a contributing factor to the cleanliness of wounds. Galen (129–199 or 217 C.E.) was the first to write that the development of pus was “essential” for wound healing. He coined the term “laudable pus”; despite Hippocrates's recognition that pus was a complication, the concept of laudable pus persisted until late in the 19th Century, such that surgeons practicing as recently as the American Civil War (1861–1865) generally accepted the dogma that the presence of pus in a wound was evidence that the patient's outcome would be favorable.

Progress in the understanding and prevention of infection began in the mid-to-late 19th century [6, 7, 8]. In 1847, Semmelweis demonstrated in animal experiments that puerperal fever was transmitted by a lack of hospital hygiene. His introduction of handwashing using disinfectants reduced the incidence of puerperal fever dramatically at his Vienna hospital [9], yet few physicians adopted his innovative approach.

When Louis Pasteur proposed the germ theory of disease and introduced the concept that microorganisms were responsible for many diseases, physicians began to reevaluate the role of pus in wounds [9]. Joseph Lister, a prominent professor of surgery at the Glasgow Royal Infirmary, built on Pasteur's concept of the germ theory and published his seminal paper on antisepsis in 1867 [10]. In his experiments, Lister applied carbolic acid (phenol) to compound fractures. These wounds healed without suppuration, and he was able to reduce the mortality rate dramatically. Koch combined the new strands of knowledge and introduced his postulates on how to establish the role of specific pathogens as the cause of particular infectious diseases. These advances were adopted slowly by the surgical mainstream, but since then, with the aid of aseptic techniques (e.g., gloves, gowns, sterilization of instruments, sterile drapes) and (later) antibiotics, the incidence of SSI has been reduced substantially, although the complication has not been eliminated. A variety of interventions continue to be evaluated in the hope of limiting further the impact of SSI on patient outcomes [4].

In spite of all efforts, SSI continues to be a major concern for surgeons, patients, and hospitals and remains associated with higher cost, prolonged hospital stays, and higher morbidity and mortality rates [1 –5]. Increasingly, rates of SSI are being used as markers of quality as well as compliance with available evidence [11 –13]. In 1999, the U.S. Centers for Disease Control and Prevention (CDC) produced a report, “Guidelines for Prevention of Surgical Site Infection,” which identified approaches that could be used to minimize SSI [14], such as sterile technique, drapes, gloves, and gowns; peri-operative antimicrobial agents; patient preparation (hair, skin); surgical technique and care of tissues; sterilization of instruments; and post-operative wound care. There had been no high-quality studies on the effects of supplemental oxygen or hyperoxia published up to that time, and no recommendations were made by the CDC regarding the use of perioperative hyperoxia. A similar set of guidelines was produced in the United Kingdom by the National Institute for Health and Clinical Excellence (NICE) in 2008 [15]. “Clinical Guideline 74” provided healthcare professionals in England, Wales, Scotland, and Northern Ireland with a set of tactics to prevent SSI. Specifically mentioned under 1.3.11 was: “Maintain optimal oxygenation during surgery. In particular, give patients sufficient oxygen during major surgery and in the recovery period to ensure that a haemoglobin saturation >95% is maintained.”

Basic Science Considerations

The rationale for using hyperoxia to prevent SSI is based on numerous scientific observations. Several theoretical mechanisms explain the putative benefit of oxygen in controlling surgical infection [16 –18]. The process of incision healing itself is dependent on the partial pressure of oxygen in tissue [17,18]. In incisions, the partial pressure of oxygen at the tissue level is involved in the production and cross-linking of collagen, a step responsible for the tensile strength of the site. Also of note, the hydroxylation of proline and lysine is related directly to the amount of oxygen present.

In addition to its observed effects on incision repair, oxygenation influences immune function. Oxidative killing by neutrophils is dependent on a respiratory burst that involves the production of superoxide anion and other oxygen-free radicals. Wounds subjected to hyperoxia exhibit an 87% increase in these reactive oxygen species (ROS) [16]. Of particular interest to surgeons is the potentially deleterious effect of hypoxia at the site of a newly created anastomosis. Inadequate blood supply leading to decreased tissue oxygen tension may contribute to anastomotic leakage and subsequent infection. At the cellular level, oxygen concentrations in tissue affect collagen formation, neovascularization, and epithelialization, all important to healing of an incision [16,17]. Because optimizing cellular oxygen tension depends on both tissue perfusion and the gradient of oxygen from capillary to cell, efforts focusing solely on hyperoxia without attention to tissue perfusion may fail to achieve improvements in site healing and lower infection rates [19 –22]. However, caution is warranted, as not all effects of hyperoxia are beneficial, and ROS have the potential to induce tissue damage and hamper incision healing and host antibacterial defenses by much the same mechanisms as are involved in the improved immune effects noted above [22].

Review of the Literature

Since the publication of the CDC guidelines in 1999, there have been seven randomized studies of the impact of supplemental peri-operative oxygen on SSI in patients undergoing general anesthesia, as well as three meta-analyses on the subject [20 –28] (Table 1). No definitive consensus has emerged, as the studies showed mixed results. The meta-analyses pointed out several noteworthy findings: One single-institution study introduced a substantial portion of the variability, different studies had various populations and sample sizes, there was inconsistent control of confounding variables in at least one study, the studies had different lengths of follow up, and there were external validity concerns [1,20,25]. These 10 papers are reviewed in the order in which they appeared in the literature.

Table 1.

Summary of Randomized Studies in Patients Undergoing General Anesthesia with or without Supplemental Oxygen

Series	Study type, procedure	No. of control/treated subjects	Intraoperative O₂ continued for	Number of SSIs in control/treated patients (%)	Odds ratio (95% confidence interval) or p value	Outcome with additional peri-operative O₂
Grief et al. 2000 [21]	Multi-center, open colorectal surgery	250/250	2 h post-op	28 (11.2)/13 (5.2)	0.43 (0.22-0.86)	Decreased SSI
Pryor et al. 2004 [22]	Single-center, open and laparoscopic abdominal surgery	80/80	2 h post-op	9 (11.3)/20 (25.0)	1.82 (0.73-4.55)	No benefit; may be harmful
Belda et al. 2005 [23]	Multi-center, open colorectal surgery	143/148	6 h post-op	35 (24.5)/22 (14.9)	0.54 (0.30-0.97)	Decreased SSI
Mayzler et al. 2005 [24]	Single-center, open colorectal surgery	19/19	2 h post-op; nitrous oxide mixture used in control group	3 (15.8)/2 (10.5)	0.63 (0.09-4.26)	No benefit
Myles et al. 2007 [26]	Multi-center	1,015/997	Variable time post-op	106 (10.4)/77 (7.7)	0.74 (0.37-1.48)	No benefit
Meyhoff et al. 2009 [28]	Multi-center, colorectal, gynecologic oncology, small bowel surgery, appendectomy	701/685	2 h post-op	141 (20.1)/131 (19.1)	p=0.98	No benefit
Bickel et al. 2011 [29]	Single-center, open appendectomy	73/80	2 h post-op	14 (13.6)/6 (5.6)	p=0.04	Decreased SSI

SSI=surgical site infection.

The first randomized study appeared in The New England Journal of Medicine in 2000. Grief et al. studied 500 patients undergoing colorectal surgery at three institutions in Germany and Austria [21]. They excluded all patients with malnutrition and those who were having minor surgery or were believed to be infected already. Patients were randomized to either 30% or 80% oxygen given intra-operatively and for 2 h post-operatively. Surgeons were blinded to group assignment. All patients received standard bowel preparation, intravenous antibiotics, and routine anesthesia and fluid resuscitation, and had their temperature maintained at 36°C. A generous infusion of fluids at 3.5 mL/kg/h was administered for the first 24 h post-operatively. Leukocyte-depleted blood was given to replace intra-operative blood loss and also at the discretion of the attending surgeon. Standardized scales were used to estimate the risk of SSI (Study on the Efficacy of Nososcomial Infection Control [SENIC] and National Nosocomial Infection Surveillance System [NNISS]) and to evaluate post-operative wound healing and infection (additional treatment, Serous discharge, erythema, purulent exudate, separation of deep tissues, isolation of bacteremia, duration of inpatient stay [ASEPSIS score]). An SSI was defined as the presence of pus in the incision with a positive bacterial culture. Patients were followed for two weeks post-operatively. The authors documented significantly higher oxygen concentrations in blood and tissues in the group receiving 80% oxygen. The SSI rates were 11.2% (95% confidence interval [CI] 7.3-15.1) in patients receiving 30% oxygen versus 5.2% (95% CI 2.4-8.0) in patients receiving 80% oxygen (p=0.01; relative risk [RR] 0.46 [95% CI 0.25-0.86]). The authors concluded that the “administration of supplemental oxygen during colorectal resection and for two hours afterward halved the incidence of surgical wound infection” [21].

In a study published in 2004 in JAMA, Pryor et al. reported the results of a randomized study comparing two levels of oxygen administration on 160 adult patients undergoing abdominal surgery at one institution in New York City [22]. Exclusion criteria were severe pulmonary disease, American Society of Anesthesiologists (ASA) score of ≥5 points, or patients deemed to be unstable or unlikely to survive. The initial study was intended to randomize 300 patients to receive either 35% or 80% oxygen during surgery and for 2 h post-operatively. The surgeons were blinded to group assignment. The bowel preparation differed among patients per attending surgeon preference; standard intravenous prophylactic antibiotic was given, the fluid and anesthesia technique was according to attending anesthesiologist preference, and the patients' temperatures were maintained at 36°C. Nitrous oxide usage was not controlled for and was more frequent in the group receiving 35% oxygen (30%, compared with 10% in the 80% oxygen group). The risk of SSI was determined for each patient using NNISS criteria. The incision was evaluated initially by the surgery team and then through retrospective review by a blinded evaluator. The presence of SSI was defined by three criteria: The surgical team's documentation of clinical signs of SSI, a management change for SSI (e.g., antibiotics, wound manipulation), and the presence of at least three objective signs of infection supporting the clinical impression of SSI (i.e., leukocytosis, fever, pus in the incision). An interim analysis of the data after enrollment of 165 patients showed a higher SSI rate in the 80% oxygen group, and the study was stopped. Five patients were lost at the start of the study and excluded. In the remaining 160 patients, the SSI rates were 25% in the group receiving 80% oxygen versus 11.3% in the group receiving 35% oxygen (p<0.02; RR 1.39 [95% CI 1.04-2.32]). The investigators noted that there was a higher transfusion rate, crystalloid infusion rate, and body mass index in the 80% oxygen group. The authors concluded that the “results of this study do not support the routine use of a high F_IO₂ in patients undergoing major abdominal surgery to reduce SSI” [22].

In 2005, also in JAMA, Belda et al. reported the results of their randomized study of 300 patients undergoing colorectal surgery at 14 institutions in Spain [23]. The patients were randomized to either 30% or 80% oxygen given intra-operatively and for 6 h after surgery. Patients with malnutrition, human immunodeficiency virus infection, or substantial weight loss, as well as those having minor or infected procedures, were excluded. The surgical teams were blinded to group assignment. Standard bowel preparation, parenteral antibiotic prophylaxis, anesthesia, and fluid management were given; SENIC, NNISS, and ASEPSIS scores were calculated. Surgeons blinded to the treatment arm evaluated the incisions using CDC criteria to define SSI within the first 14 days after surgery. Reported SSI rates were 24.4% in the group receiving 30% oxygen versus 13.9% in the group receiving 80% oxygen (p=0.04; RR 0.61 [95% CI 0.38-0.98]). The authors concluded that “supplemental 80% (oxygen) during and for 6 h after major colorectal surgery reduced postoperative wound infection risk by roughly a factor of two” [23].

In a small study published in 2005, Mayzler et al. reported on 38 patients undergoing elective colorectal surgery [24]. Patients were randomized into two groups. The first group of 19 patients received a mixture of 80% oxygen and 20% nitrogen intra-operatively and for 2 h in the recovery room. The second group of 19 patients received a mixture of 70% nitrous oxide and 30% oxygen during anesthesia, and continued for 2 h in the post-anesthesia care unit. Surgical site infections were evaluated during the hospital stay and for one month of follow up. The incidence of SSI was 12.5% versus 17.6% (p=0.53) in the 80% oxygen and the 30% oxygen groups, respectively [24]. The small patient number and the variations in the anesthetic technique make this study of limited value because of its inadequate power and questionable external validity.

The first meta-analysis of the published literature was undertaken by Chura et al. in 2007 in Surgical Infections [25]. The authors evaluated the above-described randomized trials of supplemental peri-operative oxygen with an endpoint of SSI. The four studies included 943 patients, of whom 477 received supplemental oxygen and 466 served as controls. The baseline SSI rate ranged from 11.2% (Grief et al. [21]) to 24.5% (Belda et al. [23]). The pooled risk ratio in the fixed-effects model was 0.68 in favor of the higher oxygen concentration, but this statistically significant difference was not maintained in a random effects model (RR=0.73). The authors noted that there was significant heterogeneity among the studies, with one study (Pryor et al. [22]) contributing most of the variability. Based on their analysis, they recommended the use of supplemental oxygen therapy in patients undergoing colorectal surgery. A second meta-analysis appeared in 2009 by Al-Naimi et al. [20]. These meta-analyses were similar in design, evaluated the same four studies, and came to similar conclusions.

In 2007, Myles et al. reported the results of the ENIGMA trial, a randomized study of the effect of a nitrous oxide-free anesthetic on hospital length of stay and post-operative complications [26]. Because 80% oxygen was used in the nitrous oxide-free group, the investigators planned a priori to evaluate the effects of increased oxygen administration on post-operative complications, including SSI in a secondary analysis. The study randomized 2,050 patients undergoing a broad range of surgical procedures at 19 centers in several countries to receive either a nitrous oxide-free anesthetic (80% oxygen and 20% nitrogen) or a nitrous oxide-based anesthetic (70% nitrous oxide and 30% oxygen). Individual centers enrolled various numbers of patients (from one to 640 per center). The NNNIS scale was used to assess SSI risk. Antibiotic selection was per “institutional practice.” Surgical teams were blinded to group assignment, but anesthesia records were available in the chart during the hospitalization. Anesthesiologists were able to adjust the oxygen concentration during surgery, and post-operative oxygen administration was not standardized. After surgery, a research assistant visited patients on the first day and on an as-needed basis until hospital discharge. Patients were contacted at 30 days, and records were reviewed retrospectively. Surgical site infection was defined as pus with or without a positive culture. Although the rates of several post-operative complications, including SSI, were reduced in the nitrous oxide-free group, the authors concluded in the secondary analysis that there was no independent measurable effect of supplemental oxygen on the SSI rate. The authors pointed out that either the absence of nitrous oxide or the presence of high supplemental oxygen could explain the study outcomes [26]. This study should be interpreted with caution, given the complex interplay among numerous variables, such as nitrous oxide use, oxygen concentration, institutional antibiotic preferences, propofol use, post-operative incision evaluation, and non-standardized SSI assessment.

A third meta-analysis was performed by Qadan et al. and published in the Archives of Surgery in 2009 [27]. These investigators performed an extensive literature review that identified the five randomized controlled trials reviewed above. A total of 3,001 patients were included in their analysis, of whom 1,494 were in the hyperoxia group. In this review, infection rates were 9% in the group that received perioperative hyperoxia and 12% in the group that did not. This represented a relative risk reduction of 25.3% and an overall risk ratio of 0.742 (95% CI 0.60-0.92; p=0.006). A number of methodological issues were apparent, including differing patient populations, antibiotics, and lengths of followup, as well as the high baseline SSI rate. The authors concluded that peri-operative supplemental oxygen therapy may be beneficial in the prevention of SSI [27].

In JAMA in 2009, the results of the Perioperative Oxygen Fraction (PROXI) trial, a double-blind, multi-institutional, randomized trial, were published [28]. In this trial, Meyhoff et al. reported on 1,386 patients undergoing laparotomy for a variety of indications (including colorectal, gynecologic oncology, and small bowel surgery and appendectomy) at 14 Danish institutions. They excluded patients with a pre-operative oxygen saturation <90% and those who had surgery within 30 days, or chemotherapy within 90 days, prior to the study. Patients were randomized to receive either 30% or 80% oxygen during the procedure and for 2 h post-operatively. Patients, surgeons, and statisticians were blinded to treatment group assignment. Both groups of patients had standard bowel preparation, parenteral antibiotic prophylaxis, standard anesthesia with normothermia maintained during surgery, epidural analgesia, and standard fluid resuscitation. Infection risk was assessed with the NNISS and SENIC scales. Interestingly, nearly one third of patients received corticosteroids peri-operatively. The primary outcome was SSI defined by the CDC criteria in the 14 days after surgery. Blinded investigators evaluated patients for SSI and other infections and determined ASEPSIS scores. The authors reported SSI rates of 20.1% versus 19.1% (p=0.98) in the 30% and 80% groups, respectively, and concluded that the higher oxygen concentration “did not result in a difference in risk of SSI” [28]. Among the potential limitations of this study was the use of corticosteroids in an unexpectedly large percentage of patients, a relatively high baseline SSI rate, and the modest fluid administration, which, at least in theory, may have reduced the potentially beneficial effects of hyperoxia by impairing perfusion. In work presented recently at the American Society of Anesthesiologists, Meyhoff et al. reported that at two years of follow up, the patients who had received 80% oxygen had a higher long-term mortality rate than the group that received 30% oxygen.

Bickel et al. performed a randomized, controlled trial to address the concern that previous studies had included heterogeneous populations of patients and various procedures [29]. Their study included 210 patients, all of whom underwent open appendectomy. The patients in the study group received 80% oxygen during anesthesia that continued for 2 h in the recovery room. The control group received 30% oxygen. Nitrous oxide was utilized at the anesthesiologist's discretion. All patients had the same surgical incision in the right lower quadrant for a uniform diagnosis of acute appendicitis. Exclusion criteria were pediatric patients, patients with an ASA score of ≥3 points, and patients with severe malnutrition, chronic obstructive pulmonary disease, or immunodeficiency. The surgeons and patients were blinded. The primary endpoint was the presence of an SSI, as defined by the surgical team according to clinical criteria as well as ASEPSIS scores. A secondary endpoint was the duration of post-operative hospitalization. There were six cases of SSI in the study group (5.6%) compared with 14 cases in the control group (13.6%; p=0.04). The mean hospital stay was significantly longer in the control group as well (2.92 days vs. 2.51 days; p=0.01) [29]. The authors believed they “demonstrated that the use of supplemental oxygen is advantageous in operations for acute appendicitis.” As the authors noted, however, this study had some methodological limitations. Because nitrous oxide use was allowed, it is possible that the better results in the hyperoxia group were related to the decreased use of nitrous oxide rather than the hyperoxia itself. The relatively small number of patients also was of concern, with borderline statistical significance in some of the subgroup analyses. In addition, a two-tailed Fisher exact test would have yielded a p value of 0.06, nullifying the conclusions.

In addition to the seven studies in patients under general anesthesia, Gardella et al. reported in 2008 on the effects of hyperoxia in patients undergoing cesarean section under regional anesthesia [30]. A total of 143 patients at a single institution were randomized to receive either 80% or 30% oxygen via a non-rebreathing mask during surgery and for 2 h postoperatively. Incision sites were evaluated clinically by blinded clinicians (including house staff) without a standardized scoring system. The study was stopped for futility after 143 of a planned 550 patients had been enrolled. The authors concluded that their results “suggest that high-concentration supplemental perioperative oxygen is unlikely to be clinically useful in preventing SSI after cesarean surgery” [30].

Discussion

As is evident from the above review of the literature, the impact of high supplemental oxygen has received substantial attention by researchers. However, despite the seven randomized studies in patients undergoing surgery under general anesthesia and the three meta-analyses, there has not been widespread adoption of hyperoxia as a standard of care. With the conclusions of the studies on the subject divided almost in half, and with nearly all of the studies having some methodological weaknesses, most clinicians remain unconvinced of the value of supplemental oxygen. Performing randomized trials in this context is not straightforward, and the many groups that have contributed to the study of this subject deserve recognition for the effort they invested in attempting to resolve the controversy. Given the impact of SSI on resource utilization and patient outcomes, and the possibility that hyperoxia may contribute to undesirable outcomes, objective resolution of this question is worthwhile. Ideally, what is needed is a definitive randomized, multi-center study of sufficient size and methodological rigor that evaluates the impact of increased perioperative oxygen on SSI. Such a study would need either to restrict enrollment to patients having a single procedure, or to be of sufficient size to allow subgroup analyses for individual operations. In addition, follow-up and surgical site evaluation should be standardized and continued for longer than the 14 days used in most studies, to ensure capture and appropriate categorization of all SSI events. Preliminary investigations such as the subgroup analysis of obese patients from the PROXI trial [31] are a step in that direction, but as pointed out in an accompanying editorial, the debate is only “getting started” [32].

Given the divergent results of this relatively large number of randomized studies, and the methodological issues inherent in them, no consensus has emerged on the use of increased supplemental peri-operative oxygen to reduce SSI rates. At present, clinicians should proceed with caution, and administration of supplemental oxygen should be restricted to well-designed and -conducted clinical trials.

Footnotes

Author Disclosure Statement

The authors report no commercial associations or competing financial interests.

References

Chura

, Boyd

, Argenta

. Surgical site infections and supplemental perioperative oxygen in colorectal surgery patients: A systematic review. Surg Infect, 2007; 8:455–461.

Fry

. The economic cost of surgical site infections. Surg Infect, 2002; 3,Suppl 1:S37–S43.

Kirkland

, Briggs

, Trivette

et al. The impact of surgical-site infections in the 1990s: Attributable mortality, excess length of hospitalization, and extra costs. Infect Control Hosp Epidemiol, 1999; 20:725–730.

Alexander

, Solomkin

, Edwars

. Updated recommendations for control of surgical site infections. Ann Surg, 2011; 254:1082–1093.

Leaper

. Risk factors for and epidemiology of surgical site infections. Surg Infect, 2010; 11:283–287.

Miller

, Rahimi

, Lee

. History of infection control and its contributions to the development and success of brain tumor operations. Neurosurg Focus, 2005; 18:1–5.

Nespoli

, Geroulanos

, Nardone

et al. The history of surgical infections. Surg Infect, 2011; 12:3–13.

Fry

. In vino veritas. Surg Infect, 2001; 2:185–191.

Alexander

. The contributions of infection control to a century of surgical progress. Ann Surg, 1985; 201:423–428.

10.

Lister

. On a new method of treating compound fracture, abscess etc. with observations on the conditions of suppuration. Lancet, 1867; 1:326–329.

11.

Barie

. SCIP to the loo? Surg Infect, 2011; 12:161–162.

12.

Davis

, Kuo

, Ahmed

, Kuo

. Report card on Surgical Care Improvement Project (SCIP): Nationwide inpatient sample infection data 2001–2006. Surg Infect, 2011; 12:429–434.

13.

Edmiston

, Spencer

, Lewis

et al.

Reducing the risk of surgical site infections: Did we really think SCIP was going to lead us to the Promised Land?

Surg Infect, 2011; 12:169–177.

14.

Mangram

, Horan

, Pearson

et al. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol, 1999; 20:250–278.

15.

National Institute for Clinical Health and Excellence, Guideline Development Group of the National Collaborating Center for Women's and Children's Health. Surgical Site Infection: Prevention and Treatment of Surgical Site Infection. October 2008. www.nice.org.uk/CG074. 2011 May 11.

16.

Babior

. Oxygen-dependent microbial killing by phagocytes part 2. N Engl J Med, 1978; 298:721–725.

17.

Allen

, Maguire

, Mahdavian

et al. Wound hypoxia and acidosis limit neutrophil bacterial killing mechanisms. Arch Surg, 1997; 132:991–996.

18.

Hopf

, Hunt

, West

et al. Wound tissue oxygen tension predicts the risk of wound infection in surgical patients. Arch Surg, 1997; 132:997–1005.

19.

Lumb

, Nair

. Effects of increased inspired oxygen concentration on tissue oxygenation: Theoretical considerations. Eur J Anaesthesiol, 2010; 27:275–279.

20.

Al-Naimi

, Safdar

. Supplemental perioperative oxygen for reducing surgical site infection: A meta-analysis. J Eval Clin Pract, 2009; 15:360–365.

21.

Grief

, Ozan

, Horn

et al. Supplemental perioperative oxygen to reduce the incidence of surgical wound infection. N Engl J Med, 2000; 342:161–167.

22.

Pryor

, Fahey

, Lien

et al. Surgical site infection and the routine use of perioperative hyperoxia in a general surgical population: A randomized controlled trial. JAMA, 2004; 291:78–87.

23.

Belda

, Aguilera

, García de la Asunción

et al. Supplemental perioperative oxygen and the risk of surgical wound infection: A randomized controlled trial. JAMA, 2005; 294:2035–2042.

24.

Mayzler

, Weksler

, Domchik

et al.

Does supplemental perioperative oxygen administration reduce the incidence of wound infection in elective colorectal surgery?

Minerva Anesthesiol, 2005; 71:21–25.

25.

Chura

, Boyd

, Argenta

. Surgical site infections and supplemental perioperative oxygen in colorectal surgery patients: A systematic review. Surg Infect, 2007; 8:455–461.

26.

Myles

, Leslie

, Chan

MTV

et al. Avoidance of nitrous oxide for patients undergoing major surgery: A randomized controlled trial. Anesthesiology, 2007; 107:221–231.

27.

Qadan

, Akca

, Mahid

et al. Perioperative supplemental oxygen therapy and surgical site infection: A meta analysis of randomized controlled trials. Arch Surg, 2009; 144:359–366.

28.

Meyhoff

, Wetterslev

, Jorgensen

et al. Effect of perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: The PROXI randomized clinical trial. JAMA, 2009; 302:1543–1550.

29.

Bickel

, Gurevits

, Vamos

et al. Perioperative hyperoxygenation and wound site infection following surgery for acute appendicitis: A randomized, prospective, controlled trial. Arch Surg, 2011; 146:464–470.

30.

Gardella

, Goltra

, Laschansky

et al. High-concentration supplemental perioperative oxygen to reduce the incidence of postcesarean surgical site infection: A randomized controlled trial. Obstet Gynecol, 2008; 112:545–552.

31.

Stæhr

, Meyhoff

, Rasmussen

. the PROXI Trial Group. Inspiratory oxygen fraction and postoperative complications in obese patients: A subgroup analysis of the PROXI Trial. Anesthesiology, 2011; 114:1313–1319.

32.

Canet

, Belda

. Perioperative hyperoxia: The debate is only getting started. Anesthesiology, 2011; 114:1271–1273.