Journal of Gastroenterology & Endoscopy

The Complexity of Colorectal Resection for Endometriosis

Maude Trepanier, Kristen Kent, Olivia Ziegler and Michael Deutsch*

Colorectal Surgery, Penn State Health Milton S, Hershey Medical Center, Hershey, USA

*Corresponding author
Michael Deutsch, Colorectal Surgery, Penn State Health Milton S, Hershey Medical Center, Hershey, USA.

Introduction
Endometriosis is the result of an inflammatory reaction from implants of endometrial-like tissue outside of the uterus and affects 10-12% of women of reproductive age [1]. Endometriosis can implant in various locations, including the sigmoid colon and rectum which is reported to be involved in up to 90% of intestinal lesions [2-4].

Bowel endometriosis can resemble typical endometriosis symptoms including dysmenorrhea and dyspareunia [2]. Alternatively, it can cause more bowel-specific issues such as diarrhea, rectal bleeding, constipation, dyschezia, and even bowel obstruction [2,5]. Surgical intervention has been shown to be effective at resolving these symptoms, however, there is scant literature on standardized surgical management of bowel endometriosis [6,7]. Conversely, resection of the diseased diverticular colon has robust data and standardized surgical guidelines. 

The objective of this study is to examine the short-term outcomes of patients undergoing colorectal resection for endometriosis and compare those outcomes with patients undergoing elective resection for diverticular disease, a similarly benign and largely left-sided colorectal condition. We employed National Surgical Quality Improvement Project data to compare outcomes for left sided colectomy for diverticular disease versus endometrial disease. We hypothesize the endometrial cohort will be associated with increased morbidity given the complexity in management of this pathology. 

Methods
Data Source and Patient Population
The American College of Surgeons National Surgical Quality Improvement Project (ACS-NSQIP) 2016 to 2020 database was queried for all patients undergoing colorectal resection for a diagnosis of endometriosis. The ACS-NSQIP is a Health Insurance Portability and Accountability Act (HIPAA)-compliant data file that contains case information from participating hospitals across the United States and Canada. Patients were identified as having a colorectal resection for the diagnosis of endometriosis (ICD9 codes 617.x and ICD10 codes N80.x) as the primary or secondary procedure based on primary procedure Current Procedural Terminology (CPT) codes. Colorectal resections were based on the following CPT codes: 44140-44160, 44202-44213, 45110-45123, and 45395-45397. Proctectomy was defined as the following CPT codes: 44145-44147, 44155-44158, 44207, 44208, 44211, 44212, 45110-45114, 45116, 45119, 45123, 45126, 45395 and 45397. Patients that did not undergo colorectal resection (i.e. colostomy or ileostomy reversal) were excluded from the analysis. This cohort contained a subset that was also captured using the colectomy-targeted participant-use file (PUF) which included additional variables such as bowel preparation, ileus, and anastomotic leak.

Comparison Cohort
A comparison cohort was created by querying the same ACS-NSQIP 2016-2020 database for female patients undergoing elective colorectal resection for diverticular disease using ICD9 codes 562.x and ICD10 codes K57.x. The same CPT procedure codes were used as in the endometriosis cohort. This comparison cohort was chosen as diverticular disease is a benign condition that mostly requires left-sided resections similarly to bowel endometriosis. In order to exclude patients who had a semi-elective operation for acute diverticulitis (i.e. those who underwent surgical intervention for failure of medical management during the same admission), those who were admitted for 1 day prior, or more, to their operation were excluded. Patients who underwent closure of colostomy or ileostomy for diverticular disease were also excluded.

Outcome Definitions
Standard NSQIP variable definitions were used [8]. Main clinical outcomes of interest were 30-day overall complications, reoperations, readmissions, and total length of stay (LOS). Postoperative complications included all events captured by the NSQIP: surgical site infections (superficial, deep, and organ space), wound disruption, pneumonia, unplanned intubation, acute renal failure requiring dialysis, urinary tract infection, cerebrovascular accident, cardiac arrest requiring cardiopulmonary resuscitation, myocardial infarction, transfusion within 72h, Clostridium difficile infection, and sepsis/septic shock. The colectomy-targeted PUF also recorded anastomotic leak and ileus.

Statistical Analysis
All data are presented as n (%) or mean (SD), unless otherwise specified. Univariate analyses were performed using chi-squared test for categorical variables, and Student’s t-test or ANOVA for continuous variables. Regression analyses were performed (using multiple logistic for dichotomous variables such as complications, readmissions, and reoperations, or Poisson for count data such as length of stay) to determine the effect of disease indication on the main outcomes, adjusting for the ACS-NSQIP morbidity risk score, rectal resection, and minimally invasive approach. The ACS-NSQIP morbidity risk score already incorporates the most important risk factors for morbidity and is an estimate of the probability of overall morbidity [9,10]. A subgroup analysis was performed for patients ≤45 years old (this represented the 75th percentile of patient age in the endometriosis cohort). All analyses were performed using STATA 15.1 (StataCorp, College Station, TX).  

Results
A total of 117, 287 women underwent colorectal resections between 2016 and 2020, of which 590 (0.5%) had a diagnosis of endometriosis. {Table 1} reports the patient and operative characteristics of these cases. The proportion of colorectal resections performed for endometriosis significantly increased over time {Figure 1}. Most cases (80%) were performed by a general surgeon as primary surgeon, with the remaining 20% by gynecologists. A concomitant hysterectomy and/or oophorectomy was performed in 30.7% (181/590). A proctectomy was performed in 51% and a stoma created in 11%. The majority (66%) underwent a minimally invasive approach in the overall cohort, and in 79% in the colectomy-targeted PUF with a conversion rate of 7%. Table 2 reports the clinical outcomes of the endometriosis cohort. The overall 30-day complication rate was 21%, including reoperation in 3% and readmission in 9%. Mortality occurred in 1 patient only. The median length of stay (LOS) was 4 days (IQR 3-5). Colectomy-targeted data was available in 349 patients, of which 9% had an ileus and 3% had an anastomotic leak.

A comparison of the patient and operative characteristics of the endometriosis and diverticular cohorts is shown in {Table 1}. Endometriosis patients were significantly younger, had lower body mass index, less likely to be white, less comorbid, and therefore had a lower predicted risk of morbidity and mortality when compared to patients with diverticular disease. Endometriosis patients also were less likely to receive a minimally invasive procedure and had longer operative length. In the colectomy-targeted PUF, endometriosis patients were less likely to receive a mechanical or oral antibiotic bowel preparation. {Table 2} reports the differences in clinical outcomes between endometriosis and diverticular cohorts. There were no differences in overall 30-day complications, median length of stay, readmissions, or unplanned reoperations, however there was a higher proportion of venous thromboembolism and pulmonary embolism in the endometriosis cohort. On multiple regression analysis, endometriosis patients had higher overall 30-day complications (adjusted OR 1.70, 95% CI 1.38, 2.09) and longer length of stay (adjusted IRR 1.17, 95% CI 1.13, 1.22).

*defined in the colectomy-targeted PUF as “leak, treated with non-interventional/non-operative means”
†p-value comparing the distribution of the different types of anastomotic leak

On subgroup analysis in patients ≤45 years of age, the endometriosis cohort was associated with longer median length of stay and a higher incidence of overall 30-day complications (Table 3). There also were differences in certain complications that were not seen in the overall cohort such as hemorrhage requiring transfusion, urinary tract infection (UTI), and ileus. On multiple regression analysis, endometriosis remained independently associated with increased overall 30-day complications (adjusted OR 2.1, 95% CI 1.56, 2.74) and longer length of stay (adjusted IRR 1.21, 95% CI 1.16, 1.28), but with larger magnitudes of effect compared with the overall cohort.