Tissue oxygenation is a strong predictor of surgical site infection (SSI). Mild intraoperative hypercapnia increases peripheral, gastrointestinal, and splanchnic tissue oxygenation and perfusion. Hypercapnia also has anti-inflammatory effects. However, it is unknown whether hypercapnia reduces SSI risk. We tested the hypothesis that mild intraoperative hypercapnia reduces the risk of SSI in patients having colon resection surgery.
With institutional review board approval and subject consent, patients having elective colon resection (e.g. hemicolectomy and low-anterior resection) expected to last >2 h were randomly assigned to intraoperative normocapnia (PE′CO2≈35 mm Hg; n=623) or hypercapnia (PE′CO2≈50 mm Hg; n=592). Investigators blinded to group assignment evaluated perioperative SSI (Center for Disease Control criteria) for 30 postoperative days. SSI rates were compared.
Patient and surgical characteristics were comparable among the groups. The SSI rate for normocapnia was 13.3%, and for hypercapnia, it was 11.2% (P=0.29). The Executive Committee stopped the trial after the first a priori determined statistical assessment point because of much smaller actual effect compared with the projected. However, because the actual difference found in the SSI rates (15–16%) were within the 95% confidence intervals (CIs) of the projected relative difference of 33% (95% CI −43 to +24%), our results cannot be considered as ‘no difference’, and cannot exclude a Type II error. Time to first bowel movement was half-a-day shorter in the hypercapnia group.
Mild hypercapnia appears to have little or'possibly'no ability to prevent SSI after colon resection. Other strategies for reducing SSI risk should thus take priority.
Editor's key points
- •The Bohr effect underpins why hypercapnia favours unloading of oxygen at the tissue level.
- •Permissive hypercapnia-enhanced tissue oxygenation could reduce surgical site infection.
- •This study found that mild hypercapnia could possibly reduce the risk of surgical site infection but such an effect appears small.
- •Obesity is associated with higher rates of surgical site infection.
Surgical site infections (SSIs) are common and serious complications of anaethesia and surgery.
3Even with the best sterile technique, surgical wounds may become infected. Oxidative killing by neutrophils'the primary defence against pathogens
5—requires molecular oxygen and critically depends on local tissue oxygen partial pressure.
7Consequently, SSI is inversely related to tissue oxygenation during and for a few hours after surgery.
10Tissue oxygenation, reflecting local perfusion and arterial oxygen partial pressure (PaO2), is enhanced by cardiac output,
13management of carbon dioxide,
15and epidural anaesthesia.
Mild hypercapnia also increases subcutaneous
18oxygenation and perfusion, primarily by increasing cardiac output and oxygen availability.
20Hypercapnia also appears to be anti-inflammatory when applied during acute lung injury models in animals.
22We therefore tested the hypothesis that mild hypercapnia [end-tidal partial pressure (PE′CO2) of 50 vs 35 mm Hg] during general anaethesia reduces the incidence of SSI (followed up for 30 days) in patients undergoing colon resection.
With institutional review board approval of participating centres and written consent, we enrolled adults 18–80 yr old who were undergoing elective colon resection expected to last between 2 and 6 h. Exclusion criteria included chemotherapy in the 6 months preceding surgery, anticipated secondary wound closure, fever (≥38°C) or infection at admission, severe COPD [i.e. forced expiratory volume 1 (FEV1)/forced vital capacity <70%; FEV1<50% predicted, with or without chronic symptoms], unstable angina pectoris, hypertensive cardiomyopathy, congestive heart failure (according to the Modified Framingham Criteria), myocardial infarction within 6 months, known bowel obstruction (confirmed or clinical suspicion of obstruction defined by the primary surgeon), and ASA Physical Status ≥IV.
Mild hypercapnia increases subcutaneous tissue Po2 by 26–32 mm Hg in patients undergoing major abdominal surgery compared with normocapnia at an FIO2 of 0.40.
15Previous work suggests that using the more conservative 26 mm Hg increase in tissue oxygen partial pressure would reduce infection risk by a third, from 8.2% in normocapnic patients to 5.4% in those assigned to hypercapnia.
8A total of 1994 patients would therefore provide 80% power to identify a statistically significant difference between the groups at a one-tailed alpha level of 0.05; we therefore planned to study 2000 patients with an interim analysis at 1000 patients.
Patients were given electrolyte bowel preparation without antibiotics the night before surgery. According to surgical routine, cefuroxime (1.5 g i.v.) and metronidazole (0.5 g i.v.), cefotetan (2 g i.v.), or ertapenem (1 g i.v.) were given during anaesthetic induction. Propofol (2–4 mg kg−1) were used for induction of anaesthesia and muscle relaxation provided by rocuronium (0.6 mg kg−1) or vecuronium (0.1 mg kg−1); anaesthesia subsequently was maintained with desflurane or sevoflurane, fentanyl, and neuromuscular blocking drugs as clinically required. Postoperative epidural analgesia was permitted.
After induction of anaesthesia, patients were assigned to normocapnia or hypercapnia, stratified by centre. Assignments were based on computer-generated randomizations that were kept in sealed, sequentially numbered envelopes until used. Envelopes were opened in the operating theatre just before the start of surgery. Subjects assigned to normocapnia were maintained at a target intraoperative PE′CO2 of 35 mm Hg; the target in patients assigned to hypercapnia was 50 mm Hg. In hypercapnia patients, CO2 absorbent was removed from the partial-rebreathing circle system. Tidal volume was set at 8–10 ml kg−1 with a respiratory rate of 8–10 bpm, and fresh gas flow was adjusted to maintain PE′CO2 at 50 mm Hg. Patients assigned to normocapnia were similarly ventilated, but with CO2 absorbent in the partial-rebreathing circle system as usual. Inspired oxygen concentration was controlled at 80% in the initial 620 patients because of the reported beneficial effects of supplemental oxygen.
1The final 585 patients were assigned to 30 or 80% intraoperative oxygen and to receive 4 mg i.v. dexamethasone or placebo in a factorial approach. During the study enrolment period, new evidence from another randomized controlled trial reported negative results challenging the potential benefits of supplemental oxygen.
23Therefore, we decided to randomize inspired oxygen concentrations to 30 or 80%. Corticosteroids impair innate immune responses and, as a result, may impair healing. As might be expected, chronic preoperative corticosteroid use increases wound infection risk,
25but there are only very limited data looking into the effect of single-dose corticosteroids on SSI and wound healing.
26Therefore, we decided to include a single-dose dexamethasone (or placebo) as another treatment factor, which was also assigned under factorial randomization. The current report is restricted to the hypercapnia vs normocapnia randomization.
Volatile anaesthetic administration was adjusted to maintain mean arterial pressure (MAP) at ∼90% of pre-induction value. Small boluses of phenylephrine (i.e. 50–100 µg) and fluid administration were also used as necessary. Intraoperative crystalloids were generally given at a rate of 6–10 ml kg−1 h−1, which is considered as a relatively restrictive approach;
27however, variations based on clinical judgement were permitted. Target minimum haematocrit (HCT) was determined before randomization based on the patient's age and cardiovascular status. The target HCT was 26% in patients aged <65 yr having no significant cardiovascular disease and 28% in patients aged ≥65 yr or having cardiovascular disease. Significant cardiovascular disease was defined as history of myocardial infarction or peripheral vascular disease. HCT was maintained ≥30% in patients having both significant cardiovascular disease and age ≥65 yr. Intraoperative core temperature was maintained near 36°C. Core temperature was maintained to the extent possible with upper body, forced-air covers. Fluids were warmed if necessary. Patients who were hypothermic (<36°C) upon arrival in the post-anaesthesia care unit were warmed with a full body, forced-air cover.
Anaesthesiologists were not blinded to group assignments, but gas monitors were shielded to prevent surgeons from determining randomized group assignment. After giving the report to the post-anaesthesia care nurse, the anaesthesia record was sealed in an envelope as a part of the blinding process until patient's discharge (except at the University of Vienna and the Cleveland Clinic, which both used electronic records). Postoperative pain relief was maintained by patient-controlled i.v. morphine or hydromorphone analgesia, or by epidural analgesia in patients who received epidural analgesia. Epidural catheters were most often inserted in the lumbar rather than thoracic region; local anaesthetic, epidural opioids, or both were only given after operation.
Attending surgeons, who were blinded to the randomized assignments, made hospital discharge decisions. Discharge timing was based on routine surgical considerations, including return of bowel function, control of infections (if any), adequate healing of the incision, and overall recovery during the postoperative period. Major complications including readmission, anastomotic leak, ileus, wound dehiscence, pneumonia, sepsis, and acute renal failure were monitored up to 30 days after operation (see Appendix 1).
The SSI outcome was monitored for up to 30 days. Patients were called on the 15th and 30th postoperative days for follow-up if they were not at the hospital or surgeon's clinic on those days. Patients were asked to return to their physician or to the hospital if their answers suggested they had an infection.
Factors potentially influencing infection risk were recorded, including patient comorbidity, laboratory values, and anaesthetic and surgical management. End-tidal carbon dioxide (PE′CO2) pressures were measured from the side-stream end-tidal gas measurement unit of Datex-Ohmeda systems. Intraoperative haemodynamic values and PE′CO2 were collected every 15 min, and these values were averaged for each patient. Patients were asked to rate their pain on a 10 cm visual analogue scale (VAS) during the first hour of recovery, and on the first postoperative morning. In the recovery period, pain VAS scores were recorded at 30 min intervals. On the first postoperative morning, patients were asked to provide a single VAS value that summarized their pain experience as leaving the recovery unit.
Infection risk was evaluated with the system from the Efficacy of Nosocomial Infection Control (SENIC) of the Centers for Disease Control and Prevention (CDC), which assigns one point for each of the following factors: three or more underlying diagnoses at discharge, surgery that lasts ≥2 h, an abdominal site of surgery, and the presence of a contaminated or infected wound.
28As in previous studies, we modified the system by using the number of diagnoses at admission rather than at discharge. Although the SENIC score was established well before laparoscopic surgery became common, we designated patients having laparoscopic-assisted colectomies as having an abdominal surgical site. Infection risk was also evaluated with the National Nosocomial Infection Surveillance System, which predicts risk on the basis of the type of surgery, the rating of physical status on a scale developed by the ASA, and the duration of surgery.
SSI was defined by CDC criteria
30and characterized as superficial incisional, deep incisional, or peritoneal. Wound healing was numerically scored using the ASEPSIS system.
31This is an established and validated system for quantifying surgical wound infections and evaluating wound healing. The score is derived from the weighted sum of points assigned for the following factors: (i) duration of antibiotic administration; (ii) drainage of pus under local anaethesia; (iii) debridement of the wound under general anaethesia; (iv) serous discharge; (v) erythema; (vi) purulent exudate; (vii) separation of deep tissues; (viii) isolation of bacteria from discharge; and (ix) hospitalization exceeding 14 days.
Wounds were evaluated daily throughout hospitalization by an investigator blinded to treatment. Subsequently, patients were phoned on the 15th and 30th days and a scripted interview was used to inquire about SSI. When answers suggested that infection was likely, efforts were made to obtain relevant records. An investigator blinded to treatment adjudicated infections and complications using all available data.
Our primary outcome was the incidence of SSI within 30 days of surgery. Secondary outcomes included wound-healing scores, return of bowel function, and the duration of hospitalization. Values obtained at intervals throughout surgery were averaged in each patient, and then averaged among the patients in each group. Results were compared using t or χ2 tests, and non-parametric t-tests as appropriate. Gaussian distributions were used in deciding between parametric and non-parametric test selection with the exception of ‘pain-VAS scores’ and ‘ASEPSIS’, which were both planned as non-parametric tests a priori. Other outcome parameters analysed with non-parametric tests were first flatus, first food, first bowel movement, and duration of hospitalization.
Results are presented as means (sds) and medians [inter-quartile ranges]; P<0.05 was considered statistically significant.
For administrative reasons related to data entry and validation, the initial interim analysis, which was to be at 1000 patients, was delayed until 1139 patients were enrolled. Based on a futility analysis, the study was stopped by the Executive Committee. An additional 76 patients were enrolled while the committee evaluated the initial results; consequently, 1215 patients were enrolled when the study was stopped and 1206 included in final analysis. Nine patients were removed from the study by the attending anaethesiologist after randomization, because they did not find the patients suitable to the study. Therefore, no additional data were collected from these patients. Six study centres enrolled patients: University of Vienna (n=369), Cleveland Clinic (n=295), Washington University (n=201), University of Louisville (n=195), Mater Misericordiae University Hospital Dublin (n=96), and University of Bern (n=59). Figure 1 shows the trial profile diagram.
Clinical characteristics, diagnoses, surgical procedures, duration of surgery, haemodynamic values, and anaesthetic management were generally similar in patients randomized to each intervention (Table 1). Prophylactic antibiotics were given 36 (40) min [mean (sd)] before the surgery started. Patients in the initial randomization (n=629) and the factorial randomization (n=586) were similar based on their patient characteristic, morphometrics, primary illnesses, type of surgery, and medical history.
Table 1Patient characteristic, morphometric, and potential confounding factors*. *Results presented as means (sds) or percentage. †Numbers in parentheses beside the values in the first column represent either the missing data (i.e. missing ‘n’/total ‘n’) or the ‘n’ of the data source. ‡Blood loss data presented as median and inter-quartile range. MAC, minimum alveolar concentration; MAP, mean arterial pressure; VAS, visual analoge scale.
|35 mm Hg (n=616)||50 mm Hg (n=590)|
|Sex (M/F) (35/1206)†||310/288||303/270|
|Weight (kg) (30/1206)||78 (19)||77 (20)|
|Height (cm) (77/1206)||171 (13)||171 (11)|
|BMI (kg) m−2 (99/1206)||26.7 (6.2)||26.6 (6.5)|
|Age (yr) (114/1206)||53 (16)||51 (16)|
|ASA Physical Status 1/2/3 (%) (90/1206)||23/56/21||23/59/18|
|History of smoking [n (%)] (109/1206)||166 (30%)||176 (33%)|
|Current smoker [n (%)]||137 (25%)||155 (29%)|
|Diabetes [n (%)]||55 (9%)||37 (6%)|
|Diagnosis (%) (92/1206)|
|Cancer||280 (50%)||247 (45%)|
|Inflammatory bowel disease||162 (29%)||181 (33%)|
|Other||118 (21%)||126 (23%)|
|Approach (%) (112/1206)|
|Open||410 (73%)||367 (69%)|
|Laparoscopic-assisted||153 (27%)||164 (31%)|
|Fentanyl (mg) (163/1206)||0.69 (0.93)||0.67 (0.85)|
|MAC (155/1206)||0.82(0.36)||0.82 (0.33)|
|MAP (mm Hg) (70/1206)||82 (10)||80 (10)|
|Heart rate (beats min−1) (69/1206)||76 (12)||78 (12)|
|Crystalloid (ml kg−1) (140/1206)||45 (23)||46 (23)|
|Crystalloid (ml kg−1 h−1)||12(4)||12(5)|
|Colloid (ml kg−1) (n=614)||10 (6)||11 (9)|
|Blood lost (ml)‡ (44/1206)‡||250 [150–500]||250 [100–500]|
|Red-cell transfusions (n=970)|
|Patients (%)||67/503 (13%)||64/467 (14%)|
|Units of red cells||2.0 (1.2)||2.1 (1.1)|
|First oesophageal temperature (°C) (61/1206)||36.2 (0.5)||36.2 (0.4)|
|Final oesophageal temperature (°C) (72/1206)||36.1 (0.5)||36.1 (0.5)|
|Glucose (mg dl−1) (n=487)||122(50)||124(39)|
|Arterial oxygen saturation (%) (n=439)||99 (2)||99 (2)|
|PE′CO2 (kPa) (32/1206)||4.7 (0.7)||6.3 (0.8)|
|pH (n=518)||7.37 (0.06)||7.29 (0.06)|
|PaCO2 (kPa) (n=518)||5.2 (1.3)||7.1 (0.9)|
|PaO2 (kPa) (n=518)||28.1 (16.8)||30.9 (15.3)|
|FIO2 (36/1206)||68 (19)||69 (18)|
|Postoperative care unit|
|Pain VAS (cm) (335/1206)||4.8 (2.7)||4.8 (3.1)|
|Nausea (%) (n=525)||125 (43%)||111 (44%)|
|Emesis (%) (n=556)||5 (2%)||4 (2%)|
|SENIC score 1/2/3 (%) (45/1206)||21/70/10||22/68/10|
|First postoperative day|
|Pain VAS (cm) (370/1206)||4.6 (3.0)||4.7 (3.1)|
|Nausea (%) (n=525)||125 (43%)||103 (43%)|
|Emesis (%) (n=542)||34 (12%)||19 (8%)|
PaCO2 was 5.2 (1.3) kPa in patients assigned to normocapnia and 7.1 (0.9) kPa in those assigned to hypercapnia. PE′CO2 values, which were recorded every 15 min throughout surgery, were 4.7 (0.7) kPa and 6.3 (0.8) kPa, respectively. PE′CO2 values corresponding to the randomized carbon dioxide levels are presented in Figure 2 to show compliance to randomization. The major reason clinicians obtained arterial blood analysis was to determine haemoglobin concentrations; PaCO2 values were thus available only in a subset of the study population (518 patients). In 72 of these patients (∼14%), at least one pH measurement was <7.25. There were a total of 14 values of pH<7.20 of 518 (2.7%). All but 1 of the 14 values were in the hypercapnia group (5%).
Overall, SSI rates in the centres ranged from 8 to 20% (P=0.098), but the rates were similar in each centre as a function of randomized treatments. There were also no significant differences among the treatment groups in ASEPSIS score, duration of hospitalization, major complications, or mortality. A total of 160 patients received epidural analgesia; their incidence of SSI was 10%. After accounting for randomized carbon dioxide management, there were no differences in SSI rates in patients who did and did not receive epidural analgesia (P=0.48).
Major complications included readmission in 49 patients, electrolyte imbalance in 17, anastomotic leak in 12, ileus in 6, wound dehiscence in 6, pneumonia in 6, sepsis in 5, and acute renal failure in 5. The first postoperative bowel movement was about half a day earlier in the hypercapnia compared with the normocapnia group (P=0.007, Table 2).
Table 2Postoperative Outcomes by Randomization Group (univariable)*. *Results presented as number of patients (percentage) or means (sds). †Any CDC infection numbers are not simple totals of all the above infections. Some patients experienced superficial and deep infections simultaneously. ‡ASEPSIS score data are presented in median [inter-quartile]. P-values represent results of two-tailed analyses
|4.7 kPa (n=616)||6.6 kPa (n=590)||P-value|
|SSI diagnosed by CDC criteria|
|Superficial||63 (10.2)||49 (8.3)||0.269|
|Deep||27 (4.4)||30 (5.1)||0.546|
|Peritoneal||7 (1.1)||5 (0.8)||0.188|
|Any SSI infection (CDC criteria)†||82 (13.3)||66 (11.2)||0.292|
|ASEPSIS score‡||1 [0–4]||1 [0–3]||0.087|
|Epidural analgesia||80 (13)||78 (13)||0.932|
|First flatus, days after surgery||3.3 (2.3)||3.1 (1.7)||0.201|
|First solid food, days after surgery||4.7 (3.9)||4.3 (2.6)||0.131|
|First bowel movement, days after surgery||4.5 (3.3)||4.0 (2.6)||0.007|
|Duration of hospitalization after surgery, days||8.5 (4.9)||8.4 (5.2)||0.594|
|Patients with 30 day major complications||43 (7.0)||34 (5.8)||0.411|
|Patients with 30 day mortality||4 (0.6)||1 (0.2)||0.375|
SSI developed in 11.2% of the 590 subjects who received hypercapnia and in 13.3% of the 616 who received normocapnia (two-tailed Fisher's exact test, P=0.29). The relative risk [interim-adjusted 95% confidence interval (CI)] was 0.84 (0.57, 1.24), P=0.24. The observed Z-statistic of −1.2 crossed neither the efficacy boundary (Z<−2.5) nor the futility boundary (Z>−0.238). However, the conditional power for finding a statistically significant difference if the study were to continue to completion with a similar trend was only 50%. If the study had continued with the hypercapnia and normocapnia groups having the same rate of SSI, 4168 patients would have been needed to reach an α (P) of 0.05 under 90% power. The Executive Committee thus stopped the trial.
As shown in Figure 3, the SSI rate was associated with body mass index (BMI, P<0.01 multivariate). Surgical approach (laparoscopic-assisted vs open), ASA, and SENIC scores did not significantly influence the incidence of SSI. Overall mortality was too low to be considered within any comparison, and, more importantly, the study was not powered for this outcome.
In this large multicentre randomized-controlled trial, we could not prove our hypothesis that mild hypercapnia reduces the incidence of SSI in colorectal resection surgery patients. The study was stopped by the Executive Committee at the interim analysis'without passing the futility boundary'because the chances of a positive outcome were low even if the trial continued to the planned sample size of 2000. We considered an effect of 33% relative decrease in the SSI rate in our sample size estimate. Such relative difference represented a 2.8% absolute difference between the projected SSI rates, because our estimated SSI rates, which were used as base levels in our sample-size estimate, were relatively low (8.2 vs 5.4%) compared with the current rates of 11–13% or even higher previously published rates of 13–28%.
32Our low SSI estimate represented the data from a large study we performed in colorectal surgery patients more than a decade ago.
In spite of the relatively large difference between the projected SSI rates and the current actual SSI frequency, our results showed an absolute difference of 2.1% between the study groups. Obviously, the observed 2.1% difference is not much different than the projected absolute difference of 2.8%. However, the relative difference is the parameter that drives the statistically significant difference and is, therefore, the essential component of sample-size estimates. The current difference of 2.1%, which converts to a relative difference of 15–16%, is less than half of the projected 33% relative difference (efficacy). On the other hand, our projected relative difference of 33% (i.e. 33% reduction) is still within the 95% CI of the actual difference found (44% reduction to 24% increase). This means the observed treatment effect's CI ranged from a relative risk of 0.57 to 1.24, is consistent with a true treatment effect ranging from a 43% reduction to a 24% increase in infection rates. Therefore, this study cannot totally be considered as a study with no difference, and so there is the possibility of a Type II error. Although our results did neither cross the futility border nor report a strongly negative trial, we could not justify more than doubling our study patients compared with the projected sample-size because of limited funding. As a result, the Executive Committee decided to stop the trial prematurely at the time of initial interim analysis.
Cardiac index increases 10–15% for each 10 mm Hg increment in PaCO2.
34Increasing PaCO2 also shifts the oxyhaemoglobin dissociation curve rightward and decreases systemic vascular resistance, thereby improving tissue oxygen availability.
35That mild hypercapnia improves cardiac output, increases systemic oxygen delivery, and increases both peripheral subcutaneous tissue and splanchnic oxygenation are well established facts in animals, human volunteers, and patients.
35Improved oxygenation enhances surgical outcomes.
36There was thus considerable reason to believe that mild-to-moderate hypercapnia would reduce the risk of SSI. Nonetheless, hypercapnia's effect was found much smaller than expected, and we therefore could not show a reduction in the incidence of SSI in our patients.
It is possible that a yet-to-be-detailed effect of hypercapnia in immune responses negated the putative benefits of improved tissue oxygenation. Hypercapnia may impair innate immunity or alter wound healing, especially during the perioperative period.
39However, the most likely explanations for our surprising finding are that either hypercapnia could not increase tissue oxygenation by 25–30 mm Hg as expected,
15which was possibly attributable to a perfusion compromise that blunted the tissue oxygen delivery, or hypercapnia caused increased tissue oxygenation increase, but possibly the rate of oxygenation increase was insufficient to produce a clinically important reduction in infection risk in the specific setting tested. Because tissue oxygen tensions were not measured in this study, these explanations are only hypothetical.
In this study, we did not measure tissue oxygenation, because the relationship between PaCO2 and tissue oxygen partial pressure is well established.
34However, there is still a chance that our intervention of application of mild hypercapnia did not alter tissue oxygenation at the same ratio we have shown before.
Another major factor influencing tissue oxygenation is arterial oxygen level itself. Similar ranges of inspired oxygen were applied in both study patient groups, and the resulting PaO2 values obtained were similar in patients in whom blood-gas analyses were done. Arterial blood gas measurements were not required by protocol, but the results were recorded when obtained for clinical reasons. In the light of our previous trials, the effects of FIO2 on arterial oxygenation in the colorectal surgery setting were considered to be predictable. However, our current data proved this wrong, because PaO2 levels for the FIO2 0.80 group were comparably lower than published previous trials.
40One explanation for this is a possible selection bias. Because arterial blood-gas analysis was not enforced in the current trial, it is likely that the analysis was done in sicker patients with expected oxygenation problems.
After discharge, patients were followed up with standardized phone interviews. Patients were asked to return to their physician or to the hospital if their answers suggested they had an infection. Although post-discharge SSI was uncommon, partial reliance on phone interviews for a major study outcome was a limitation of our trial.
Prevalence of obesity has gradually increased in the population since the first study of supplemental oxygen and infection was published a decade ago. Obesity reduces tissue oxygenation.
42It is also well established that obese patients experience more perioperative complications and increased SSI rates.
43Our results extend previous findings by presenting a clear ‘dose–response’ type relationship between BMI and SSI. The infection rate was more than doubled in patients with a BMI >35 kg m−2 compared with patients with BMI 18–25 kg m−2. Prevention measures, such as timely application of appropriate prophylactic antibiotics
44and maintaining perioperative normothermia,
2may thus be especially important in obese patients.
Laparoscopic colorectal surgery was shown to decrease SSI rate,
45but this effect was not always consistent.
46Laparoscopic colonic surgery significantly decreases tissue oxygenation early in the course of surgery.
47As tissue oxygenation is one of the most important predictors of SSI,
48inconsistency in results is unsurprising. Laparoscopic surgery was not an independent risk factor of SSI in our study.
The Executive Committee stopped the trial because the treatment effect appeared to be too small and unlikely to reach statistical significance even if the study was carried to completion. However, there was a slight apparent treatment effect, with infection rates of 11.2 and 13.3% in the hypercapnia and normocapnia groups, respectively. It thus remains possible that hypercapnia reduces the risk of SSI; but if it does, the effect appears to be small and could only be demonstrated by a very large trial.
In conclusion, it is well established that mild-to-moderate hypercapnia increases tissue Po2 by roughly 25–30 mm Hg. However, the risk of SSI was not significantly reduced in 1206 patients randomized to mild hypercapnia (PE′CO2=50 mm Hg, n=590, 11.2% SSI) compared with normocapnia (PE′CO2=35 mm Hg, n=616, 13.3% SSI). Mild hypercapnia appears to have little or possibly no ability to prevent surgical wound infections in the colorectal surgery setting tested.
Declaration of interest
This study was supported in part by the Gheens Foundation (Louisville, KY, USA) and the Mater College for Postgraduate Research (Ireland). Viasys Healthcare (Wheeling, IL, USA) provided the Hi-Ox Oxygen masks. All personnel financial interests have been disclosed.
We appreciate statistical expertise and guidance from Edward J. Mascha, PhD (Cleveland Clinic), further statistical assistance from Gilbert Haugh, MS, editorial contributions from Nancy Alsip, PhD, and data management from Rachel A. Sheppard, MBA, CCRC, CCRA, and N. Lale Akça, MBA, CCRA (all'but G.H. and E.J.M.—from Office for Clinical Research Services and Support, University of Louisville, Louisville, Kentucky). We also appreciate the effort and support provided from each study site's attending and resident physicians, perioperative environments, recovery areas, and patient wards.
Major Complications are given in Table A1.
Table A1Major complications. These complications are chosen to be serious and plausibly related to infection or wound healing. †RIFLE-F criteria is defined as either an increase in serum creatinine >3× of baseline or creatinine >4.0 mg dl−1 or urine output <0.3 ml kg−1 h−1 for 24 h or anuria for 12 h.
|Complication||Requirements for acceptance|
|Necrosis of stoma||Intraperitoneal necrosis needing surgery|
|Surgical wound infection||CDC criteria|
|Intra-abdominal abscess||Ultrasound or CT scan or surgical confirmation|
|Ileus||Surgical, X-ray, and CT scan confirmation|
|Peritonitis without leak||Surgery, excepting anastomotic leak|
|Anastomotic leak||Leak confirmed during secondary surgery|
|Wound dehiscence||Requiring secondary suture of fascia for treatment|
|Pneumonia||According to the modified CPIS or CDC criteria*|
|Respiratory insufficiency-Moderate-Severe||Requiring reintubation and mechanical ventilation|
|Acute MI||Increased levels of cardiac enzymes or new Q waves|
|Congestive heart failure||Proved with CXR'pulmonary congestion|
|Sepsis||Positive blood culture and at least two of the following: Hypo- or hyperthermia, tachycardia, tachypnoea, leucopenia/leucocytosis (DIC), or multi-organ dysfunction|
|Bleeding||Requiring transfusion and surgery|
|Gastrointestinal bleeding||Requiring transfusion, surgery, or both|
|Electrolyte imbalance||Electrolyte levels beyond normal laboratory levels|
|Acute renal failure||Requiring dialysis or RIFLE-F criteria†|
Hypercapnia Trial Investigators
University of Vienna, Vienna, Austria: Klaus Erdik; Erol Eredics, MD; Barbara Kabon, MD; Sara Kazerounian; Oliver Kimberger, MD; Andre Kugener, MD; Corinna Marschalek, MD; Pia Mikocki; Monika Niedermayer, MD; Eva Obewegeser; Ina Ratzenbock; Romana Rozum, MD; Sonja Sindhuber; Katja Schlemitz, MD; Karl Schebesta, Anton Stift, MD.
Cleveland Clinic, Cleveland, OH, USA: Adrian Alvarez, MD; Endrit Bala, MD; Samuel T. Chen, BA; Jagan Devarajan, MD; Ankit Maheshwari, MD, Ramatia Mahboobi, MD; Edward Mascha, PhD, Hassan Nagem, MD; Suman Rajogopalan, MD; Luke Reynolds, MSc; Erich Zirzow, EMT-P, CCRP.
Washington University, St. Louis, MO, USA: Helmut Hager, MD; Akiko Taguchi, MD; James W. Fleshman, MD; Thomas E. Read, MD.
University of Louisville, Louisville, KY, USA: Diane DeLong, RN, BSN, CCRC; Anthony G. Doufas, MD, PhD.; Raghavendra Govinda, MD; Yusuke Kasuya, MD; Ryu Komatsu, MD; Rainer Lenhardt, MD, MBA; Mukadder Orhan-Sungur, MD; Papiya Sengupta, MD; Anupama Wadhwa, MD.
Mater Misericordiae University Hospital, Dublin, Ireland: Mujeeb Arain, FCAI; Siun Burke, FCAI; Barry McGuire, MRCSI; Jackie Ragheb, FCAI.
University of Bern, Bern, Switzerland: Tobias Podranski, MD; Akikio Taguchi, MD; Oliver Kimberger, MD.
- Outcomes Research™ Group Supplemental perioperative oxygen to reduce the incidence of surgical wound infection.N Engl J Med. 2000; 342: 161-167
- Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization. Study of Wound Infection and Temperature Group.N Engl J Med. 1996; 334: 1209-1215
- Fraser VJ Hospital-associated costs due to surgical site infection after breast surgery.Arch Surg. 2008; 143 (discussion 1): 53-60
- Oxygen and surgical site infection: a study of underlying immunologic mechanisms.Anesthesiology. 2010; 113: 369-377
- Wound hypoxia and acidosis limit neutrophil bacterial killing mechanisms.Arch Surg. 1997; 132: 991-996
- Oxygen-dependent microbial killing by phagocytes (second of two parts).N Engl J Med. 1978; 298: 721-725
- Oxygen-dependent microbial killing by phagocytes (first of two parts).N Engl J Med. 1978; 298: 659-668
- Wound tissue oxygen tension predicts the risk of wound infection in surgical patients.Arch Surg. 1997; 132 (discussion 5): 997-1004
- Direct measurement of wound and tissue oxygen tension in postoperative patients.Ann Surg. 1983; 197: 470-478
- Oxygen as an antibiotic. The effect of inspired oxygen on infection.Arch Surg. 1984; 119: 199-204
- Hypercapnia improves tissue oxygenation.Anesthesiology. 2002; 97: 801-806
- Postoperative pain and subcutaneous oxygen tension.Lancet. 1999; 354: 41-42
- Supplemental perioperative fluid administration increases tissue oxygen pressure.Surgery. 2003; 133: 49-55
- Mild hypercapnia increases subcutaneous and colonic oxygen tension in patients given 80% inspired oxygen during abdominal surgery.Anesthesiology. 2006; 104: 944-949
- The effect of intraoperative end-tidal carbon dioxide partial pressure on tissue oxygenation.Anaesthesia. 2003; 58: 536-542
- The effects of epidural and general anesthesia on tissue oxygenation.Anesth Analg. 2003; 96: 1553-1557
- Hypercapnia improves tissue oxygenation in morbidly obese surgical patients.Anesth Analg. 2006; 103: 677-681
- Hypercapnia induces a concentration-dependent increase in gastric mucosal oxygenation in dogs.Intensive Care Med. 2008; 34: 1898-1906
- A comparison between transpulmonary, right and left indicator dilution curves during hypotension.Br J Anaesth. 1970; 42: 90
- Effect of injection and sampling site on dye dilution curves during hypotension.Cardiovasc Res. 1972; 6: 741-747
- Ambient Pco2 modulates intracellular pH, intracellular oxidant generation, and interleukin-8 secretion in human neutrophils.J Leukoc Biol. 2002; 71: 603-610
- Hypercapnic acidosis attenuates endotoxin-induced acute lung injury.Am J Respir Crit Care Med. 2004; 169: 46-56
- Surgical site infection and the routine use of perioperative hyperoxia in a general surgical population: a randomized controlled trial.JAMA. 2004; 291: 79-87
- Elective colon and rectal surgery differ in risk factors for wound infection: results of prospective surveillance.Ann Surg. 2006; 244: 758-763
- Multivariable predictors of postoperative surgical site infection after general and vascular surgery: results from the patient safety in surgery study.J Am Coll Surg. 2007; 204: 1178-1187
- Single dose dexamethasone for postoperative nausea and vomiting'a matched case-control study of postoperative infection risk.Anaesth Intensive Care. 2010; 38: 661-666
- Meta-analysis of standard, restrictive and supplemental fluid administration in colorectal surgery.Br J Surg. 2009; 96: 331-341
- Identifying patients at high risk of surgical wound infection. A simple multivariate index of patient susceptibility and wound contamination.Am J Epidemiol. 1985; 121: 206-215
- Surgical wound infection rates by wound class, operative procedure, and patient risk index. National Nosocomial Infections Surveillance System.Am J Med. 1991; 91: 152S-157S
- CDC definitions of nosocomial surgical site infections, 1992: a modification of CDC definitions of surgical wound infections.Infect Control Hosp Epidemiol. 1992; 13: 606-608
- Postoperative wound scoring.Biomed Pharmacother. 1989; 43: 669-673
- Ertapenem versus cefotetan prophylaxis in elective colorectal surgery.N Engl J Med. 2006; 355: 2640-2651
- Carbon dioxide and the critically ill'too little of a good thing?.Lancet. 1999; 354: 1283-1286
- Prophylactic antibiotics in surgery and surgical wound infections.Am Surg. 2000; 66: 105-111
- Biological effects of hypercapnia.Intensive Care Med. 2000; 26: 133-138
- Carbon dioxide and tissue oxygenation: is there sufficient evidence to support application of hypercapnia for hemodynamic stability and better tissue perfusion in sepsis?.Intensive Care Med. 2008; 34: 1752-1754
- Hypercapnic acidosis attenuates pulmonary epithelial wound repair by an NF-kappaB dependent mechanism.Thorax. 2009; 64: 976-982
- Hypercapnic acidosis attenuates lung injury induced by established bacterial pneumonia.Anesthesiology. 2008; 109: 837-848
- Hypercapnic acidosis attenuates severe acute bacterial pneumonia-induced lung injury by a neutrophil-independent mechanism.Crit Care Med. 2008; 36: 3135-3144
- Supplemental perioperative oxygen and the risk of surgical wound infection: a randomized controlled trial.JAMA. 2005; 294: 2035-2042
- Tissue oxygenation in obese and non-obese patients during laparoscopy.Obes Surg. 2005; 15: 813-819
- Obesity decreases perioperative tissue oxygenation.Anesthesiology. 2004; 100: 274-280
- The impact of obesity on outcome after major colorectal surgery.Dis Colon Rectum. 2007; 50: 2223-2237
- Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgical Infection Prevention Project.Am J Surg. 2005; 189: 395-404
- Laparoscopic versus open colorectal surgery: a randomized trial on short-term outcome.Ann Surg. 2002; 236 (disscussion 67): 759-766
- Moderate increase in intraabdominal pressure attenuates gastric mucosal oxygen saturation in patients undergoing laparoscopy.Anesthesiology. 2004; 100: 1081-1087
- Laparoscopic surgery impairs tissue oxygen tension more than open surgery.Br J Surg. 2007; 94: 362-368
- Early postoperative subcutaneous tissue oxygen predicts surgical site infection.Anesth Analg. 2010; 111: 946-952
- Acute renal failure definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group.Crit Care. 2004; 8: R204-R212
Accepted: May 6, 2013
© 2013 The Author(s). Published by Elsevier Inc.
User LicenseElsevier user license |
How you can reuse
Elsevier's open access license policy
Elsevier user license
For non-commercial purposes:
- Read, print & download
- Text & data mine
- Translate the article
- Reuse portions or extracts from the article in other works
- Redistribute or republish the final article
- Sell or re-use for commercial purposes
Elsevier's open access license policy