Does intra‐operative humidification with warmed CO2 reduce surgical site infection in open colorectal surgery? A randomized control trial

Surgical site infections (SSI) are a significant cause of postoperative morbidity and mortality, contributing to a considerable financial burden on the healthcare system. Insufflation of the open surgical wound with warm, humidified carbon dioxide (CO2) is a novel measure aimed to reduce SSI. The local atmosphere of warm, humidified CO2 within the open surgical wound is proposed to decrease airborne contamination, bacterial growth, desiccation, and heat loss while improving tissue oxygenation and perfusion. This randomized controlled trial evaluates the impact of the HumiGard™ surgical humidification system on the incidence of SSI in patients undergoing open colorectal surgery.


Introduction
Surgical site infections (SSI) are the most common cause of healthcare-associated infection, and can lead to prolonged hospital admission, extensive antibiotic treatment, or in severe cases, re-intervention. 1 Accordingly, SSI is also a significant burden on healthcare costs. 1 Surgical humidification with warm carbon dioxide (CO 2 ) is extensively used in laparoscopic surgery and has been previously demonstrated to reduce the risk of intra-operative hypothermia and infection. 2 The utilization of intraoperative insufflation with warm, humidified CO 2 with the HumiGard™ surgical humidification system is a novel measure to prevent SSI following laparotomy. Surgical humidification with CO 2 reduces airborne contamination. 3 and has a bacteriostatic effect at body temperature on Staphylococcus aureus, a common pathogen implicated in SSI. 4 Surgical humidification also maintains warmth and moisture in the surgical tissues, enhancing wound healing. 5,6 In laparoscopic colorectal surgery, insufflation with warm, humidified CO 2 has been shown to reduce postoperative pain and surgical site infections compared to the control group (5.7% v 13%, P = 0.04) in one study involving 246 patients. 2 Rodent studies have investigated the effect of warm, humidified CO 2 in open surgery and found it reduced peritoneal damage, increased tissue oxygenation and perfusion. 7,8 Increased mean core and wound temperature have also been demonstrated during open surgery in human trials. 5 However, there is limited data examining the effect of warm, humidified CO 2 on SSI incidence in open abdominal surgery. Our study is the largest human trial to date investigating the efficacy of warm, humidified CO 2 in reducing the incidence of SSI in open colorectal surgery. The primary outcome was the occurrence of SSI, while secondary outcomes included intensive care unit (ICU) length of stay (LOS), total LOS and mean core temperature. Our hypothesis is that the utilization of intraoperative warm, humidified CO 2 will significantly reduce SSI incidence in patients undergoing open colorectal surgery.

Study design
This study was a single-blinded multi-site randomized control trial conducted at a single tertiary Colorectal Surgery service in Melbourne, Australia, from July 2019 to August 2021. Participants were randomized into either the control group (current standard care) or treatment group (HumiGard™) by a computer generated single block randomisation method. The trial was single-blinded at the level of the patients, who were unaware of their allocation to control or treatment groups. As it is not feasible to blind surgeons, no subjective outcomes assessed by surgeons were included, to avoid introducing any biases. Postoperative assessment was conducted by assessors not involved with the procedure to maintain blinding to group allocation.

Patient recruitment and randomization
The study population included eligible patients undergoing either elective or emergency open colorectal surgery at various hospital sites throughout Monash Health. Inclusion criteria included patients aged 18-80 years undergoing either elective or emergency open colorectal surgery and who were able to give informed consent. Exclusion criteria included paediatric patients aged <18 years, elderly patients over 80 years of age, and inability to give informed consent, including if patients were too unwell to consent. Patients were also excluded if they had current abdominal wall infection or surgical site infection, immunocompromise, or a contraindication to CO 2 insufflation, or if they were pregnant, morbidly obese or had severe end organ dysfunction. The full list of inclusion and exclusion criteria is provided in Appendix 1. Elective patients received bowel preparation and enhanced recovery after surgery (ERAS) routinely, as per institutional protocols.
Written and verbal informed consent was obtained from the trial participants, who were then randomized to either 'HumiGard™' or 'control' through a computer-based system (single block randomisation).

Surgical procedures
Immediately prior to the surgical start, the primary surgeon opened an opaque unmarked envelope containing the group allocation. Patients were normothermic (>36 C) before anaesthetic induction, and core temperature was monitored via an oesophageal temperature probe. During anaesthetic induction, antimicrobial prophylaxis was administered as a single dose intravenously of cefazolin and metronidazole, with adjustment as needed for weight and allergies. Standard surgical antimicrobial skin preparation was used (alcoholic iodine or iodine).

Materials
The HumiGard™ device is a small filtration device that delivers warm, humidified CO 2 at a flow rate of 10 L per minute (Fig. 1). The sterile portion of the device is secured at the proximal end of the laparotomy wound. It warms the peritoneal cavity from the wound surface.
For the HumiGard™ group, the Fisher & Paykel Healthcare Ltd. HumiGard™ system was prepared according to the user instructions by the investigators and the sterile portion placed in the wound by the primary surgeon. The anaesthetist was advised of the commencement of the HumiGard™ system to allow for adjusted ventilation to compensate for CO 2 absorption. The control group received standard open surgery. A wound protector was not used during surgery to reduce bias. After wound closure, a hydrocolloid dressing was used in all patients.

Postoperative follow-up
Following the surgery, the wounds were assessed for evidence of SSI using the CDC/NHSH Surveillance definitions for Specific Types of Infection SSI Grading Scale. 9 The patients were initially intended to be reviewed in-person at 14 days, and by a phone call at 30 days post-operatively for wound assessment. However, due to disruptions to outpatient follow up caused by the COVID-19 pandemic, patients were instead assessed for signs of SSI immediately prior to discharge by unit clinicians. entry. Following completion of data entry, data was checked for duplicates and errors and amended accordingly. Identifying data was removed except for study ID number before undergoing statistical analysis.

Sampling
A priori power analysis was conducted using G*Power 3.1 10 to calculate an appropriate sample size. For a reduction in SSI from 15% to 5% at a power of 0.8 and type I error rate set at 0.05, a sample of 130 patients was estimated for analysis by Fisher Exact and Chi-Square. The target sample was increased to 156 patients to account for 20% attrition.

Statistical analysis
Statistical analysis was conducted using IBM SPSS 25 and on an Intention to Treat (ITT) basis. All variables were tested for normality using Shapiro-Wilks test. Missing temperature data was assessed with Littles MCAR and found to be missing completely at random (P = 0.904), permitting limited imputations by expectation maximization. BMI, surgical duration and total LOS were found to be nonparametric, subsequently analysed using Mann-Whitney U test and summarized as median (interquartile range). Remaining parametric data were analysed using independent t-test and summarized as mean (SEM). All categorical variables were analysed using Fishers Exact and Chi-Square tests and summarized by frequency (proportion). A binary logistic regression was conducted to explore group membership between treatment cohorts and clinically significant variables. Results from this analysis were summarized using classification accuracy and odds ratio (confidence interval). Between-subject sub-factor analysis was conducted that split data by urgency, smoking history, bowel prep, gross contamination, ICU admission, abdominal procedure and past abdominal history.

Patient characteristics
Between July 2019 and August 2021, a total of 156 patients were recruited who were planned to undergo open colorectal surgery. Thirty-six patients were subsequently excluded as they were either lost to follow up or did not receive the allocated intervention, including 13 patients who did not proceed with the trial due to the institutional COVID-19 response, and 10 patients who no longer met eligibility criteria, either due to the use of a wound protector (seven patients) or medical deterioration between randomisation and intervention (three patients). Two patients were excluded from the HumiGard™ group as the intervention was ceased prematurely; in one case, the device malfunctioned and was not able to be used, and in one case, the patient developed surgical emphysema and the humidifier was then stopped. One patient from the control group was excluded from analysis as they underwent an unrelated complication post-operatively and required a further surgical procedure prior to the assessment of outcomes. One hundred and twenty patients were included in the final analysis, comprising 66 patients in the HumiGard™ group and 54 patients in the control group (Fig. 2). Core body temperature data was available for 111 patients. Baseline characteristics are presented in Table 1, which were similar between groups.
There was a higher rate of gross contamination intraoperatively in the control group (33.3% versus 16.7% in the HumiGard™ group; P = 0.098), which was further investigated by sub-factor analysis. The remaining peri-operative data is presented in Table 2.
Post-operatively, fewer patients received corticosteroids in HumiGard™ group (6.1% versus 16.7% in the control group; P = 0.080), although this was not statistically significant, Postoperative care was otherwise similar between the two groups ( Table 2).

Incidence of SSI
Patients that received HumiGard™ had a lower incidence of SSI (4.5%) compared to those in the control cohort (13.0%), although this finding did not reach statistical significance (P = 0.092) ( Table 3).
Binary logistic regression was conducted to identify predictors of SSI based on treatment cohort and clinically relevant patient characteristics ( Table 4). The nested model found a high classification accuracy of 91.7 with no improvement or significance once incorporating all variables (R 2 MF = 0.214; R 2 CS = 0.116; R 2 N = 0.266, χ 2 (8) = 7.033, P = 0.533). This model found that HumiGard™ use was associated with reduced risk of SSI (OR = 0.223; CI = 0.05-1.10), although this again did not reach statistical significance (P = 0.066). ICU LOS, total LOS and core temperature changes HumiGard™ group had a shorter ICU LOS (68.11 h) compared to the control group (81.13 h), that was not statistically significant (P = 0.526) ( Table 6). Total LOS did not significantly vary between cohorts (Table 5).
There was a statistically significant increase in core temperature between the start and end of surgery in the HumiGard™ cohort (mean 0.31 C increase versus 0.02 C decrease in the control group; P < 0.001). There was no significant difference in core temperature between the HumiGard™ (35.99 C) and control (36.17 C) groups at the start of surgery (P = 0.130) or in core temperature at the end of surgery (HumiGard™ cohort 36.24 C versus control group 36.15 C; P = 0.073). Temperature analysis is presented in Table 5.

Sub-factor analysis
A sub-factor analysis was conducted between cohorts according to urgency, smoking history, bowel preparation, gross contamination, ICU admission and past abdominal procedures. Results are demonstrated in Table 6. In patients undergoing elective surgery, there was no SSI in the HumiGard™ group (0%) compared to the control group (17.6%), which did not reach statistical significance (P = 0.062). No significant difference was observed in patients without smoking history. Among patients without gross contamination intraoperatively, the HumiGard™ group had a statistically significant decrease in the incidence of SSI (3.8% in the treatment group versus 16.7% in the control group; P < 0.05). There was no statistically significant difference between groups when controlling for ICU admission. HumiGard™ patients with a past history of abdominal procedures were less likely to have SSI (2.6%) compared to the control patients (13.5%), although this was not significant (P = 0.082).

Discussion
Our study could not demonstrate that the use of HumiGard™ reduced the incidence of SSI during open colorectal surgery. Despite the nonsignificant results, there was a reduction in SSI appreciated (4.5% v 13%, P = 0.092) similar to past laparoscopic surgeries that utilized warm, humidified CO 2 (5.7% v 13%, P = 0.04). 2 These findings support the theory that, as in laparoscopic procedures, CO 2 insufflation may act as a protective barrier against airborne contamination in open surgical wounds, 3 acting as a denser-than-air barrier to deflect airborne particles according to Stokes' law for the settling velocity of gas/liquid particles. 11 This is particularly salient, given that airborne bacteria continue to contribute to surgical site contamination, despite modern advances in operating theatre ventilation. 12 In addition, CO 2 insufflation in open surgery decreases desiccation of internal tissues, reducing tissue trauma and the risk of SSI development 13 ; humidified CO 2 was associated with a lower evaporation rate compared to dry CO 2 in a wound model study, 14 and with reduced peritoneal damage during open surgery in both a rodent model study 15 and a human randomized control trial of 40 patients undergoing open colorectal surgery. 16 Interestingly, only 2.6% of patients with past abdominal surgery presented with an SSI, which is a finding requiring further Note: *Significant at 0.05. investigation to identify its cause. Elective patients (who undertook ERAS bowel preparation) were also observed to have a reduction in SSI.
Our study suggests that the use of warm, humidified CO 2 may be worthwhile as an additional, routine measure to reduce SSI incidence and protect against additional tissue damage and inflammation.
Increased wound and core temperature has also been found to reduce SSI incidence and the colonization of associated pathogens. 5 HumiGard™ has been demonstrated to increase core temperature intraoperatively, which may help reduce bacterial colonization and, subsequently SSI incidence. Surgical humidification with CO 2 has previously been found to be effective for wound warming in a randomized control trial, 4 as well as preventing heat loss from the wound surface in patients undergoing open colorectal surgery. 17 Increased wound and core temperature may lower the risk of SSI by improving tissue oxygenation and immune cell function; 13 surgical humidification has been shown to have a significant increase in oxygen tension within the surgical tissues in a rodent model study. 7 Another rodent model study demonstrated that warm, humidified CO 2 is associated with increased tissue perfusion. 8 The effects of humidified CO 2 on pathogen culture should also be considered; S. aureus, for example, is a common pathogen associated with SSI, upon which CO 2 has a bacteriostatic effect at body temperature, which exponentially increases with time. 4 SSI persist as common postoperative complications leading to increased morbidity and mortality, with a current SSI rate of 15% for open colorectal surgery at our institution. The demonstrated reduction in SSI associated with HumiGard™ may therefore be beneficial to our patients undergoing open laparotomy. Warm, humidified CO 2 offers many benefits intra-operatively, including reducing bacterial colonization, preventing wound damage and improving wound and core temperature. When coupled with ERAS protocols and bowel preparation, HumiGard™ could significantly improve patient outcomes post-operatively and warrants further robust investigation.
We accept that there were certain limitations of the study. The COVID-19 pandemic posed challenges in study commencement, as well as recruitment and the follow-up of patients, ultimately underpowering the study. Several patients were withdrawn from the clinical trial due to institutional concerns about the unknown effect of surgical wound humidification on transmission of COVID-19. Being a multi-site trial, there were further challenges with obtaining consent, ensuring availability of the device and troubleshooting the device during surgeries, despite adequate education and training. The diffuser was challenging to use and there was variability in its placement during surgery. The appropriate placement of the device and intraoperative gas insufflation during laparotomy can also sometimes pose concerns for colorectal anastomosis. HumiGard™ was ceased in two events, one in which the patient had surgical emphysema due to incorrect positioning of the device without any short or long-term adverse outcome, and another due to local skin irritation by the diffuser. Post-operatively, the timing of wound reviews was inconsistent and posed issues during data collection. The final sample size was less than power calculations estimated, as there were more patients lost to follow up or who did not undergo surgery than anticipated.
In addition, the exclusion of patients who are aged >80 years or morbidly obese limits the generalisability of our result to patients who are higher risk for wound infection for these reasons. Future investigation of the role of surgical humidification in open colorectal surgery could consider its effect in these patient groups, as well as comparison with existing wound protection devices, such as use of the Alexis port. Intra-operative wound insufflation with warm, humidified CO 2 with HumiGard™ is a novel technique that may be a valuable adjunct to reduce SSI in patients undergoing open colorectal surgery. There are many imperative learning points from this trial, including the effect of increased core temperature on SSI, which should be further investigated to assess the clinical significance of warm, humidified CO 2 in open surgery. Larger sample size should be incorporated in future studies to better assess the efficacy of this novel preventative method.