A Retrospective Cohort Study to Determine Predictive Factors for Abdominal Wound Disruption Following Colorectal Surgery

Abdominal wound disruption (AWD) is a postoperative complication that increases length of hospital stay, mortality, and cost. A retrospective cohort study was conducted to identify predictors of AWD, defined in the National Surgical Quality Improvement Program User Guide as a spontaneous reopening of a previously surgically closed (midline) wound that occurs within 30 days after index elective surgery in patients undergoing colorectal surgery.Data from the American College of Surgeons National Surgical Quality Improvement Program (2006–2012) were searched, supplemented by institutional review board-approved chart review. Patients were identified using Current Procedural Terminology codes for open and laparoscopic abdominal colorectal procedures. Data were collected to predefined worksheets and entered into a statistical analysis program and included demographics; comorbidities; pre- and postoperative laboratory tests including white blood count, blood glucose, and albumin levels; date and type of procedure; wound classification; postoperative complication rate; type of access; time to disruption; surgical site infection (SSI); and use of the SSI intervention/prevention bundle. The Wilcoxon rank sum test was utilized to compare independent continuous variables between the groups, and Fisher’s exact test was utilized to compare categorical variables. Variables with a P value <.2 at univariate logistic regression were included in multivariate logistic regression analysis. Time-to-event variables were compared using Cox regression analysis. Of the 690 patients included in the study, 16 (2.3%) developed an AWD. Mean age was 61.9 ± 15.3 years and 61.3 ± 15.0 years in AWD and non-AWD groups, respectively (P = .704). AWD occurred more frequently in men than women (75% vs. 50%; P = .040) and in patients who did compared to those who did not develop a deep incisional SSI (12.5% vs. 2%; P = .044). Preoperative albumin level was significantly lower in AWD (3.2 ± 0.8 vs. 3.8 ± 0.7; P = .006), as well as the proportion of post-bundle implementation (18.75% vs. 65.7%; P = .041). No significant differences were observed for any of the other variables examined. Per multivariate analysis, male gender (P = .05), absence of SSI bundle (P = .026), and hypoalbuminemia (P = .01) were independent predictors of AWD after elective colorectal resections. Time to AWD was significantly shorter in patients without SSI (P <.001). Results indicate implementation of the SSI bundle decreased AWD rates. Further prospective studies are needed to confirm these findings.

Wound disruption is a postoperative complication that contributes to increased morbidity and mortality rates and direct and indirect costs for individuals and health care providers.¹ Abdominal wound disruption (AWD) is a severe complication of colorectal surgery with reported incidence rates of 0.4% to 3.5% and mortality rates as high as 45%^2-4; it has been referred to in the literature as abdominal wound dehiscence, fascial dehiscence, or burst abdomen. 

Attempts to improve surgical outcomes led to the National Veteran’s Administration Surgical Risk study in 1980s,⁵ and the National Surgical Quality Improvement Program (NSQIP) was initiated in 1990s.⁶ Patient Safety Indicators (PSI) were developed by the Agency for Healthcare Research and Quality of the United States Department of Health and Human Services to provide information on potential postoperative complications.⁷ One of the provider level indicators (PSI-14) measures AWD. National rates of postoperative wound disruption were at 0.48 per 100 000 US residents with the Healthcare Cost and Utilization Project, a risk-adjusted rate of 1.11 per 1000 eligible patients in 2008.^8,9 

The NSQIP database was used to identify predictors of perineal wound disruption after abdominoperineal excision of the rectum¹⁰ and overall wound disruption after colorectal resections regardless of location of the resection.¹¹ In their retrospective review comparing 363 AWD cases with 1089 controls, van Ramshorst et al² reported AWD leads to evisceration requiring immediate re-operation (100%) and increases the length of hospital stay (P <.001), incisional hernia (P <.001), and mortality rates (16%). Several reports using multivariate logistic regression found heterogeneity among reported predictors of AWD in colorectal surgery.^3,10-12 In a retrospective study of 17 044 cases from the NSQIP database, Webster et al³ reported the independent risk factors for AWD were cerebrovascular accident with no residual deficit, history of chronic obstructive pulmonary disease (COPD), current pneumonia, emergency procedure, operation time >2.5 hours, postgraduate year 4 level resident as surgeon, superficial or deep wound infection, failure to wean from the ventilator, and 1 or more complications other than dehiscence. Wounds classified as clean and return to the operating room during admission were found to be protective factors. In their retrospective study of 2761 patients undergoing major abdominal surgery during a 5-year period, Riou et al¹² found age >65 years, wound infection, pulmonary disease, hemodynamic instability, and ostomies in the incision to be significant predictors of AWD. In another study using NSQIP data, Althumairi et al¹⁰ reported American Society of Anesthesiologists (ASA) classification ≥4 (P = .003), history of smoking (P <.001), history of COPD (P = .03), body mass index (BMI) ≥35 kg/m² (P = .001), and closure with a flap (P <.001) to be risk factors for wound disruption after abdominoperineal excision. In another study that utilized the NSQIP database, Moghadamyeghaneh et al¹¹ found the highest risk of wound disruption was among patients with wound infection (P <.01).

The surgical site infection (SSI) bundle was first introduced by the Institute for Healthcare Improvement¹³ (IHI) in 2001 and included 3 or more evidence-based interventions with the potential to prevent SSI to be implemented in a consistent manner for all patients. The SSI bundle was introduced in the authors’ cohort in 2009 and consisted of 4 groups of interventions: pre-hospital, pre-, intra-, and postoperative (see Figure 1). 

One disadvantage of NSQIP data is the lack of information on the use and/or effect of the SSI bundle. To the best of the authors’ knowledge, no studies exist in the literature that evaluate the impact of the SSI bundle on AWD rates. The aim of this study was to identify predictive factors for AWD in patients undergoing colorectal resections.

Study design. A retrospective cohort study was conducted using the records of colorectal patients operated on by a single surgeon at a single institution. Demographics and relevant perioperative variables of all patients in the participant use data files who underwent abdominal colorectal surgery at the authors’ institution from 2006 to 2012 were extracted from the NSQIP database (see Table 1). Institutional Review Board approval was obtained before initiating the study, which was designed according to Strengthening the Reporting of Observational Studies in Epidemiology guidelines.¹⁴ 

Data collection and definitions. Patient records were identified using Current Procedural Terminology codes for open and laparoscopic abdominal colorectal procedures. Data abstracted included demographics; comorbidities; pre- and postoperative laboratory tests including white blood count (WBC), blood glucose, and albumin levels; date and type of procedure; wound classification; postoperative complications, including surgical site infection (SSI); type of access (open vs. laparoscopic); time to disruption; and use of the SSI intervention/prevention bundle. These data were entered into predefined Excel tables. The primary endpoint of the study was wound disruption, defined in the American College of Surgeons NSQIP User Guide¹⁵ as a spontaneous reopening of a previously surgically closed wound that occurs within 30 days after the principal operative procedure. Hyperglycemia was defined as blood glucose level of >140 mg/dL. Hypoalbuminemia was defined as serum albumin level of <3.5 g/dL. 

All incisions were closed in layers according to the facility’s Division of Colon and Rectal Surgery wound closure guidelines. The fascia at laparoscopic port sites was closed using single-strand polydioxanone sutures; double-strand polydioxanone loops were employed to close midline laparotomy wounds. No preventive mesh was used. Skin closure was performed using titanium staples or interrupted monofilament sutures.¹⁶ 

Infection. Superficial SSI was defined in NSQIP as an infection that occurs within 30 days after the operation and involves only skin or subcutaneous tissue of the incision; deep SSI was defined as an infection that occurs within 30 days after the operation, appears to be related to the operation, and involves deep soft tissues (eg, fascial and muscle layers) of the incision. Organ/space SSI was defined as an infection that occurs within 30 days after the operation, appears to be related to the operation, and involves any part of the anatomy (eg, organs or spaces), other than the incision, which was opened or manipulated during an operation.¹⁶

Statistical analysis. Data were entered into SPSS software, version 18 (SPSS Inc, Chicago, IL), for statistical analysis. Mean and standard deviation were used as descriptive statistics for continuous variables; percentage and ratio were used to express categorical variables. Wilcoxon rank sum test was utilized to compare independent continuous variables between the groups. Categorical variables were compared using Fisher’s exact test, and univariate logistic regression was used to identify factors with the most significant impact on wound disruption rates. Variables with a P value <.2 were included in multivariate logistic regression analysis. Continuous variables such as preoperative albumin or blood glucose were converted to dichotomous variables for logistic regression with cut-off values as previously defined (hyperglycemia and hypoalbuminemia). Time-to-event variables were compared using Cox regression analysis. Statistical significance was defined as 2-sided, P ≤.05.

Of the 690 patients included in the database, 16 (2.3%) developed a wound disruption. Age, BMI, diagnosis, previous abdominal surgery, diabetes mellitus, chronic steroid use, current dialysis, liver cirrhosis, a history of COPD or neoadjuvant chemo- and radiotherapy use, pre- and postoperative blood glucose level, perioperative hyperglycemia rate, preoperative blood transfusion rate, type of surgery, wound class, type of access, creation of an ostomy at time of surgery, the use of an open wound management technique, and anastomotic leak rates were not significantly different between patients with and without AWD (see Table 1).  More men than women (75% vs. 50%; P = .040) and more patients with hypertension requiring medication than those who did not (81% vs. 53%; P = .039) developed AWD. Thirteen (13) of the 16 AWDs developed before implementation of the SSI bundle; the pre-SSI bundle AWD rate was 13 of 379 (3.4%) compared to 3 of 311 (0.96%) following SSI bundle implementation (P = .041). Higher ASA scores were noted more frequently in AWD patients (ASA III–V: 100% vs. 78.3%; P = .026). Preoperative WBC and albumin levels also were significantly different between AWD and non-AWD patients; leukocytosis (11.7 ± 3.7 vs. 9.2 ± 8.0; P = .033) and hypoalbuminemia (3.2 ± 0.8 vs. 3.8 ± 0.7; P = .006) were more common in persons who developed AWD. SSI rates were significantly higher in AWD patients, owing to deep incisional SSIs that were significantly more frequent in AWD patients than in those who did not experience AWD (12.5% vs. 2%) (see Table 1).

Using univariate logistic regression, the most significant factors affecting AWD rates were found to be male gender, ventilator dependency, severe COPD, hypertension requiring medication, ASA class, hypoalbuminemia, overall SSI, deep incisional SSI, and absence of SSI bundle (see Table 2). Male gender, hypoalbuminemia, and absence of SSI bundle (pre-SSI bundle implementation] were the 3 independent predictive factors of AWD (see Table 3).

The mean time-to-AWD in patients with all 3 subtypes of SSI and without SSI was 15.25 ± 12.4 days vs. 15.57 ± 7.7 days. The onset of AWD was significantly later in AWD patients with SSI (P <.001) (see Figure 2). All patients with AWD underwent emergency wound closure and had no recurrent AWD or SSI within 30 postoperative days. 

AWD is a postoperative complication leading to increased length of hospital stay, mortality, and costs. In their retrospective study including 25 636 eligible patients who had undergone abdominopelvic surgery between January 1, 2008, and December 31, 2012, Shanmugam et al⁹ reported patients with any wound disruption included in the US Nationwide Inpatient Sample had 9.6% excess mortality, 9.4 days of excess hospitalization, and $40 323 in excess hospital charges relative to matched controls. These variables are worse for AWD. A recent European retrospective cohort study¹⁷ showed patients with AWD after laparotomy had 107.5% excess in-hospital mortality, 15.6 days of excess hospitalization, and 14 327 Euros excess cost. A comparative study¹⁸ (N = 967) evaluating long-term outcomes showed among the 37 patients with AWD, 83% developed incisional hernia at a median follow-up of 40 months. These patients also reported significantly worsened body image perception and decreased physical and mental quality-of-life scores.

In the current study, deep incisional SSI rates were significantly higher in patients with AWD. This finding is in line with previous reports evaluating perineal wound disruption after closure with a myocutaneous flap.^10,19 Other NSQIP database studies^11,20 reported superficial incisional and organ/space SSI rates were also significantly higher in patients with AWD.

Independent predictors of AWD in the present cohort included male gender, hypoalbuminemia, and absence of the SSI bundle. Male gender was found to be an independent predictor of AWD in previous studies as well. In a model based on regression coefficients, van Ramshorst et al² and Kenig et al²¹ reported 75% of AWD patients were men with a risk score of 0.7; other studies^17,22 have reported approximately 70% of patients with AWD were men, including a prospective study²² with 50 patients with AWD and a gender ratio of 2.84:1. However, some studies,^10,20,23,24 including 2 retrospective cohort studies with 44 (15 patients with AWD) did not find gender to be a predictor. In their post hoc analysis of data from a recent randomized trial involving 1386 patients who underwent emergency or elective laparotomy, Dahl et al²⁵ found smoking (3.8% smokers vs. 2.4% nonsmokers; P = .04) and alcohol abuse defined as daily intake of more than 4 drinks (15% abusers vs. 2.3% nonabusers; P <.0001) may account for the increased rates of AWD among men in several studies. 

Serum albumin level <3.0 g/dL previously was shown to be a predictor of AWD following colorectal surgery.¹¹ In the current cohort, albumin level <3.5 g/dL was an independent predictor of AWD. A previous study²⁶ (N = 108 898) that also used NSQIP data to evaluate the association of hypoalbuminemia with postoperative complications among patients undergoing colorectal resection reported that even modest hypoalbuminemia (3 g/dL ≤ serum albumin level <3.5 g/dL) increased AWD rates.

In most previous studies, the role of SSI bundle implementation in AWD was not reported. The development of a SSI was a predictive but not an independently predictive factor of AWD in the current cohort; use of the SSI bundle was predictive. The SSI bundle, first introduced by the IHI in 2001, was defined as 3 or more evidence-based interventions implemented in a consistent manner for all patients with the potential to prevent SSI.¹³ In the current study, use of the SSI bundle seemed to have the potential to reduce AWD rates in patients after elective colorectal surgery. 

Previous research¹¹ reported AWD was preceded by an SSI in 86% of the patients. The pathogenesis of AWD seems to differ in patients with and without SSI. In the current study, time to AWD was significantly shorter in patients who did not develop an SSI. Hence, implementation of SSI bundle and correction of preoperative hypoalbuminemia seems to reduce the risk of wound disruption.

Overall, the use of previously acquired data can have inherent inaccuracies. In addition, an important limitation of this study may be that complication rates are underestimated because the NSQIP database only records follow-up data for 30 days postoperatively. Because the NSQIP database does not include all potential variables that may affect AWD (eg, intraoperative variables), not all risk factors could be ascertained. 

Male gender, absence of the SSI bundle, and hypoalbuminemia (albumin level <3.5 g/dL) were independent predictors of AWD after colorectal resection. Deep incisional SSI was associated with increased rates of AWD. In this study, implementation of the SSI bundle decreased AWD rates. Further prospective studies are needed to confirm these findings.

Dr. Gachabayov is a research fellow; Ms. You and Mr. Sullivan are research assistants; and Dr. Bergamaschi is Chief, Division of Colorectal Surgery, State University of New York, Stony Brook, NY.

Please address correspondence to: Roberto Bergamaschi, MD, PhD, FRCS, FASCRS, FACS, Taylor Pavilion, Suite D-365, 100 Woods Road, Valhalla, NY 10595; email: rcmbergamaschi@gmail.com.