Surgical Site Infection Wound Bundles Should Become Routine in Colorectal Surgery: A Meta-Analysis

Background: Surgical Site Infections (SSI) are a major source of post-operative complications and potentially affect oncological outcomes. Reducing SSI is multi-factorial, best served by the additive affect of individual wound bundle elements. With changing strategies and novel innovations ongoing meta-analyses are needed to inform current practice. This study undertook a meta-analysis of existing wound bundles impact on SSI in colorectal surgery. Methods: A PROSPERO-registered (ID: CRD42018104923) meta-analysis following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and using databases PubMed, Scopus and Web of Science, from January 2008 to July 2018, was undertaken. Articles scoring ≥ 17 using Methodological Index for non-randomised Studies (MINORS) criteria were included. Results: 5,104 articles were reviewed, and 27 studies met inclusion criteria with a total cohort of 23851 patients. Wound bundles significantly decreased SSI rates from 17.5% to 9.7%. Sub-analysis identified greatest impact on superficial SSI (risk reduction of 54%; p<0.00001) and organ-space infections (risk reduction 42%; p=0.0006).Wound bundles also significantly reduced hospital length of stay (MD = −0.79; p<0.00001). Conclusions: Colorectal wound bundles significantly reduce the risk of SSI and length of hospital stay. They should become routine in colorectal surgery. Future work encompasses the need for standardisation of wound complications, standardised follow-up of patients and internationally agreed research definitions.


Introduction
The global impact of Surgical Site Infection (SSI) is increasingly recognised, both in terms of post-operative complications and the effect on patient's outcomes. SSI rates vary internationally, related in part to variable definitions, different populations, co-morbidities and strategies utilised to reduce surgical site infection [1,2]. Surgical site infection may cause distress and inconvenience to patients, delay their discharge, increase risk of incisional hernia and re-admission to hospital [3,4]. Furthermore, the hospital or patients may be financially penalised. Recently the negative oncological impact of SSI is becoming increasingly reported [5][6][7]. A key to reducing SSI is a team approach, involving all providers, in every phase of care, with a cumulative additive benefit of each aspect in the bundle. A wound bundle, in general, will have more than three components and extend from pre-operative care through to rehabilitation. Newer concepts in colorectal surgery wound care include negative pressure therapy [8] and wound protective devices [9,10].
While several meta-analyses have been performed looking at bundles and surgical site infection, with the exception of Pop-Vicas, et al. [11], most relate to publications and interventions before 2016. The search strategy used in this paper differed from that of Pop-Vicas, et al. [11] in that it used different keywords and databases. This study therefore undertook a meta-analysis of bundle impact on SSI.

Search Strategy and Study Eligibility
A detailed meta-analysis of the literature was undertaken to incorporate articles relating to colorectal surgery wound care, surgical wound infection, and surgical site care bundles. Existing research optimizing wound care in colorectal surgery was reviewed to determine current bundle strategies to improve wound outcomes. A systematic review and meta-analysis of all published English articles was conducted using PubMed, Scopus, Web of Science and Cochrane electronic databases from 2008 to July 2018. A literature search was conducted using keywords; colorectal surgery, surgical site infections, wound bundles, compliance, care pathway, and surgical outcomes. Additional studies were identified by searching the reference lists of included articles.

Inclusion and Exclusion Criteria
The methods of the analysis and inclusion criteria were specified in advance and registered with the International Prospective Register of Systematic Reviews (PROSPERO) on 23/07/2018. This meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The Centres for Disease Control and Prevention (CDC) definitions for surgical site infection were used. They are classified into superficial, deep or organ/space in this study [12]. A wound care bundle was defined as three or more items combined to reduce wound infection as per the Institute for Healthcare Improvement [13]. For this meta-analysis, only studies with pre-and postintervention SSI data for colorectal surgery were included, while studies that did not compare results to pre-intervention SSI rates were not included. Non-English articles were not included.

Eligibility Assessment and Data Extraction
Eligibility assessment was performed independently in a blinded standardised manner by two reviewers (DF and CMcI). Disagreements between reviewers were resolved by discussion between the two review authors and if no agreement could be reached, it was planned a third reviewer would decide (AJ), however a third reviewer was not required. Two reviewers (DF and CMcI) independently assessed each published study for the quality of study design by using the Methodological Index for nonrandomised Studies (MINORS) score [14]. A MINORS score of ≥ 17 was considered the standard for inclusion. Information was extracted from each included study on SSI classifications, bundle elements, length of stay, bundle adherence rates, study design, country, study length, cohort sizes, and SSI rates pre-and postintervention. The primary outcome was SSI rates following the use of wound bundles. Secondary outcomes were the effect of individual interventions included in the bundles and the SSI rates for superficial, deep and space organ infections.

Statistical Analysis
For comparison of SSI rates pre-and post-intervention risk ratios (RR) were calculated using Review Manager Version Five (RevMan5). Meta-analyses were performed by computing the RR using Mantel-Haenszel method and both fixed-effect models or random-effects models, depending on the heterogeneity of studies. Heterogeneity was assessed using the I 2 statistic where a value greater than 50% was considered high and a randomeffect model was then used to combine variables of interest. RR and 95% Confidence Intervals (CI) for each classification of SSI was calculated, along with the p-value for which a value < 0.05 represented statistical significance. For the analysis of wound bundle elements, individual bundle elements in each study were reported in three phases of care: pre, peri-and post-operative care. However, any perioperative intervention that was only used once was not included in the table and was reported separately. The individual elements of each wound bundle were reviewed, and random-effect models were used to further explore the underlying effects of specific methodological features and intervention aspects of the care bundles on the rate of SSI. Some wound bundle features were identified that may explain some of the heterogeneity in the risk of SSI between studies. 19/46 studies were excluded from the meta-analysis: Seven studies stratified cohorts based primarily on compliance in using a bundle [42][43][44][45][46][47][48], six did not state colorectal specific SSI rates [49][50][51][52][53][54], 3 were deemed of low quality [55][56][57] and three did not provide pre-intervention cohort sizes [58][59][60]. Characteristics of included studies are shown in Table 1   22 of the 25 studies had a statistically significant decrease in overall SSI rates following bundle implementation [16][17][18][19][20][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]. Two studies showed no effect [21,40] and Anthony, et al. 2011 [15] reported a statistically significant increase in SSI after bundle implementation.

Length of Stay Results:
There were seven studies that included data on the length of hospital stay in both pre-intervention and post-intervention cohorts [18,22,23,25,27,31,34]. The mean difference between the length of hospital stay pre-and post-intervention was calculated in a meta-analysis. Two studies [22,34] provided the mean and standard deviation (mean ± SD) for the number of hospital days. The other five studies provided the median with either the full range or interquartile range. For these five studies [18,23,25,27,31], the mean ± SD were calculated from the data provided, according to calculations set out in the following studies:

Discussion
This meta-analysis evaluated the efficacy of wound bundles in SSI reduction, based on a ten-year literature review, yielded 27 publications meeting quantitative criteria. Two were RCT and 25 were retrospective cohort studies. The overwhelming evidence supports the use of wound bundles, even with the recognised heterogeneity of the studies. Surgical site infection, including superficial, deep and organ space, is one of the most common complications following open and colorectal cancer surgery [64]. At the outset, there is a global challenge in relation to the definition and heterogeneity of both superficial and deep SSIs. The lack of standardization of wound event reporting is common both in colorectal and other areas of surgery [65]. DeBord, in an editorial review of the issue, looks at the concept of proposals to classify surgical site events and surgical site occurrences requiring procedural interventions [66]. Of the 27 papers used in our meta-analysis, 15 papers used the Centres for Disease Control and Prevention (CDC) definitions for SSI [12]. A further ten papers used the National Surgical Quality Improvement Program (NSQIP) which uses the CDC definition for the types of SSI. The two remaining papers used the European ECDC definition which is again the CDC definition.
Definitions and reporting of surgical site occurrences, first defined by the Ventral Hernia Working Group (VHWG) in 2010 [67] to include seroma, wound dehiscence, and enterocutaneous fistula have not been widely adopted thus far [66]. Only one study, Anthony et al. 2011, showed an increase in SSIs following application of wound bundles, which may have been due to their failure to include mechanical bowel preparation or oral antibiotic preparation.
There is a significant increase in SSI rate in urgent or emergency procedures due to a myriad of confounding factors such as poor preoperative preparation and both clean contaminated and dirty operations [17,68,69]. Watanabe, et al. [70] have suggested that in cases of colon perforation with generalised contamination, delayed primary skin closure or leaving an incision open to heal by secondary intention should be considered. This is increasingly challenged by more use of comprehensive wound bundles that include wound irrigation and incisional negative pressure therapies [71]. However, despite this, a significant proportion of dirty wounds (without fasciitis) are not closed primarily. In a study by Alkaaki, et al. [72], more than half (30/55 [54%]) of the infected patients in their study underwent emergency surgery and they found that emergency surgery increased the risk of SSI fivefold compared to elective surgery. Ensuring strict adherence of preventative wound bundles, especially in emergency procedures, may see a very significant reduction in SSI globally. Successful implementation of clinical guidelines to reduce hospital acquired infections is challenging. Some have evolved using protocol-driven reduction [73] and others have looked at multiple different implementation strategies [74]. The Institute of Healthcare Improvement (IHI) developed a concept of bundles. A bundle generally uses more than three evidence-based measures which implemented together are more effective than in isolation. Recently, Tomsic, et al. suggested that bundle size itself is important and in their analysis suggested that a bundle with more than eleven items have additional standalone benefit in surgical site reduction [75] .
In our meta-analysis we identified that surgical site infections were significantly reduced with the use of wound bundles. With sub-analysis of SSI into superficial SSI, deep SSI and organ space SSI, there were differences in outcome. Superficial SSI and organ space SSI were significantly reduced by the bundle, whereas there was only a trend for deep SSI. The reason for this is not entirely clear and may relate to the variability in bundle elements used. Many studies did not use negative wound pressure dressings. Recently, Murphy and colleagues [76] in Canada identified that negative pressure in the Neptune study had no associated effect on SSI. They report a very high SSI rate, approaching 32-34%. However, in their study they did not report or use any wound bundle. This may account for the failure to obtain a significant reduction in infection. Ideally, bundles target areas for reduction in variation in the delivery of care focusing on three key phases pre-operative, intra-operative and post-operative. Bundles should not just involve the patient but also their family. Pop-Vicas and colleagues [11] published a recent meta-analysis on colorectal bundles for surgical site infection prevention in the journal of Infection Control and Hospital Epidemiology. Multiple papers on the same topic are important to reinforce an important clinical issue. Given the potential implications in terms of cost, prolonged hospital stay, patient discomfort, and the potential adverse oncological and survival effects of both superficial and deep SSI, it is important that surgeons and those involved in the primary care of colorectal cancer and colorectal benign patients implement aspects of care bundles that are proven.
Wound protectors are commonly used in colorectal surgery and are recommended in open abdominal surgery in the ACS and SIS Guidelines [60]. However, there are some conflicting results on this in the literature [60,[77][78][79]. The combination of MBP and antibiotic (PO) preparation is recommended for all elective colectomies according to ACS and SIS guidelines [77]. Other surgical techniques such as quilting or killing the dead space to reduce seroma and the use of subcuticular suturing should be looked at with increasing evidence that these may reduce wound infection rates [80,81]. This paper did not specifically look at laparoscopic versus open colorectal surgery and this is something that will need to be done into the future, stratifying cohorts or having separate or comparative studies [82]. Although we found that colorectal wound bundles significantly reduce the risk of SSI and length of hospital stay our study has several limitations. Firstly the vast majority of the included studies were retrospective cohort studies with heterogeneous interventions; no assessment of risk of bias was carried out. Secondly the primary outcome measure of SSI does not have a specified length of follow-up. Thirdly only four studies provided sufficient raw data to carry out a metaanalysis on risk factors for SSI; a small number of patients were included in each analysis. In addition, the effect on wound bindle efficacy in patients with immune compromise, or ongoing Covid infection has not been widely studied.

Conclusion
This meta-analysis has identified significant reductions in wound infections with implementation of wound bundles.
As Surgeons we have the responsibility to ensure we routinely use wound bundles which should become routine in colorectal surgery. Future work encompasses the need for standardisation of wound complications, standardised follow-up of patients and internationally agreed research definitions.