Introduction
Rectovaginal fistula is a major complication following deep endometriosis surgery, particularly when it requires disc excision or segmental resection of the rectum along with excision of adjacent vagina1. When compared to bowel leakage, rectovaginal fistula appears more difficult to repair and may require several additional surgical procedures and a longer stools deviation time using a stoma1,2. Reducing the risk of rectovaginal fistula can be achieved by various means such as avoiding vaginal opening, placement of omentum between vaginal and rectal sutures, stitches fixing vagina to rectum and separating the two sutures, or the routine use of a transitory diverting stoma.
Guidelines for management of endometriosis specifically recommend that “clinicians refer women with suspected or diagnosed deep endometriosis management to a centre of expertise that offers all available treatments in a multidisciplinary context3. More specifically, surgical management of deep endometriosis infiltrating the colon and the rectum requires multidisciplinary teams that include colorectal surgeons. Worldwide surgical steps involving removal of endometriosis nodules infiltrating the digestive tract are performed by or in collaboration with general surgeons, who bring their experience and convictions. In the literature, results of several randomised trials concerning rectal cancer surgery lend support to routine use of stoma in the prevention of postoperative rectal fistula following removal of rectal cancers4-6. However, an automatic extrapolation of their conclusions to endometriosis surgery may be ill-advised due to the considerable differences between patients managed for rectal cancer and deep endometriosis. Theoretically some of these differences may protect against postoperative rectal fistula following endometriosis surgery (patients are young women, free of preoperative radiotherapy, in good health), while others may not (e.g concomitant excision of the vagina). As juxtaposition of bowel and vaginal sutures is considered a strong risk factor for rectovaginal fistula formation, a diverting stoma is routinely used by some teams7,8 in accordance with recommendations by various working groups9. These indicate that while a stoma is unlikely to ensure primary healing, it reduces the risk of fistula related complications, such as fecal peritonitis.
The benefits of stoma in deep endometriosis surgery remain however questionable, due to the lack of comparative studies in women managed for rectal endometriosis10. Colleagues who do not favor performing preventive stoma assert that stoma systematically requires a second surgical intervention to restore the digestive tract, leading to possible aesthetic harm, residual pain, incisional hernias or subcutaneous infections or stenosis of the colorectal anastomosis11, and complications requiring secondary surgery in 8.6% of cases12.
Our multidisciplinary team’s approach to surgical management of deep rectovaginal endometriosis has been modified. In the first period from 2005 to June 2018, for women requiring concomitant excision of rectum and vagina resulting in juxtaposition of the rectal stapled line and of posterior vagina suture, we advocated maximum prevention of rectovaginal fistula by large employ of preventive stoma13. This was applied by one of the authors’ (H.R.) at the Expert Center of Diagnosis and Multidisciplinary Management of Endometriosis, Rouen University Hospital until subsequent publication of French Guidelines for the Management of Endometriosis raised questions about scientific support for use of preventive stoma in endometriosis surgery10. From September 2018 onwards, the recommended approach favored a more restrictive use of stoma, reserved for stapled lines on low rectum, large vaginal excisions and other risk factors such as obesity, unsatisfactory rectal stapled line air test, presumed tension on stapled line, etc. This approach was used by the author during the second period at the Clinic Tivoli-Ducos Endometriosis Centre in Bordeaux. These circumstances resulted in two populations of women managed for similar endometriosis lesions, by comparable surgery but differing preventive stoma policy.
The aim of our study was to assess rectovaginal fistula prevalence, depending on the policy for preventive stoma use, in women managed for rectovaginal endometriosis and involving juxtaposition of rectal and vaginal sutures.

Patients and methods

Patients included in this series were managed at Rouen University Hospital, France, from February 2009 to June 2018, and at the Clinic Tivoli-Ducos in Bordeaux, France, from September 2018 to February 2020. They were prospectively enrolled in the CIRENDO database, which is financed by the G4 Group (The University Hospitals of Rouen, Lille, Amiens and Caen) and coordinated by one of the authors (H.R.). Inclusion criteria were: i) women managed for rectovaginal endometriosis infiltrating both the rectum and the vagina; ii) vaginal excision followed by posterior vaginal suture, with or without hysterectomy; iii) rectal nodule removal requiring rectal lumen opening, by either disc excision or segmental resection, followed by rectal suture or colorectal anastomosis juxtaposing the vaginal suture; iv) accurate recording of the height of the rectal suture, measured in cm above the anal verge. Exclusion criteria were: i) rectal nodule removal by shaving without opening of rectal lumen; ii) non-French and non-English speaking patients unable to answer the questionnaire required for inclusion in the database.
All patients were preoperatively examined by experienced gynaecological surgeons (H.R., B.R. and B.M.), who identified vaginal infiltration during clinical examination. Preoperative assessment was performed by radiologists with considerable experience in deep endometriosis and included pelvic MRI, endorectal/transvaginal ultrasound and when required, computed tomography based virtual colonoscopy. This allowed assessment of rectal nodule characteristics and identification of associated localisations involving USL, ovaries, fallopian tubes, diaphragm, urinary tract, pelvic nerves etc.
To remove rectal nodules, disc excision or colorectal segmental resection were proposed. Disc excisions were carried out using a combined laparoscopic-transanal approach, by employing either a circular stapler (a technique used by numerous teams worldwide) or a semi-circular stapler (the Rouen technique, currently used by a small number of teams in Europe) 14. For multiple bowel nodules, the afore-mentioned techniques may be associated with sparing healthy bowel located between consecutive nodules15. The choice of surgical approach was made preoperatively, and patients were fully informed of the aims, risks and expected benefits of our approach. Patients then had a preoperative visit with a colorectal surgeon, and were informed of the possibility of performing diverting stoma at the end of the procedure, specifically when vaginal and rectal sutures were juxtaposed, and to reduce the risk of complications related to rectovaginal fistula.
Surgical procedure on the bowel involved one gynaecological surgeon (H.R., B.M or B.R.), one experienced colorectal surgeon, and surgeons in training. The gynaecological surgeon removed all endometriosis localisations, including the vaginal infiltration and the colorectal surgeon performed rectal suture using transanal staplers. Omental flap was systematically placed between rectal and vaginal repair sutures in patients managed in Rouen prior to 2018, but not in Bordeaux from 2018 to 2019. The decision to create a primary stoma by ileostomy or colostomy was made by both surgeons and based on intraoperative findings, such as the close proximity of vaginal and rectal sutures following vaginal and rectal excision, unsatisfactory colorectal anastomosis bubble test results, appearance of tension on the rectal stapled line, excessive intraoperative bleeding, patient obesity, etc 12. However, the decision to not perform stoma was more frequent in patients managed in Bordeaux from September 2020 onwards and has since become routine in accordance with French guidelines for the management of endometriosis which highlight the lack of evidence in support of preventive stoma and omentoplasty10. Consequently, the use of both stoma and omental flap has progressively become limited to the management of rectovaginal fistulae. With the exception of the use of stoma and omentoplasty, all other surgical procedures were similar between the two centres, one author (H.R.) having practiced in Rouen until June 2018and in Bordeaux from September 2018.
Postoperative hospitalisation varied from 4 to 6 days. Clinical symptoms and body temperature were recorded 3 times/day, and assessment of blood values of C-reactive protein (CRP) and white blood cells (WBC) was routinely performed at day 4, 5 and 616. When patients presented intrarectal temperature >38.2°C, or a progressive increase in either CRP or WBC for two consecutive days, emergency clinical examination and computed tomography with barium enema were performed to rule out rectovaginal fistula, pelvic abscess or infected pelvic hematoma. Patients with rectovaginal fistula and without primary diverting stoma underwent emergency secondary surgery with confection of diverting stoma. In patients with hematoma or abscess but without obvious rectovaginal fistula, emergency laparoscopy was performed to drain the liquid, followed by a rectal bubble air test. Where test results were abnormal or equivocal, a secondary stoma was created prophylactically1,12.
All patients managed in Rouen and Bordeaux agreed to the prospective recording of data concerning antecedents, clinical symptoms, findings of clinical and imagery examinations, surgical procedures and postoperative outcomes through the CIRENDO (North-West Inter Regional Female Cohort for Patients with Endometriosis) database (NCT02294825). Information was obtained using self-questionnaires, surgical and histological records, while data recording, contact and follow-up were carried out by 2 clinical research technicians. Standardised gastrointestinal questionnaires were routinely used to assess pre- and post-operative digestive function: the Gastrointestinal Quality of Life Index (GIQLI)17, the Knowles-Eccersley-Scott-Symptom Questionnaire (KESS) 18 and the WEXNER scale19, the Urinary Symptom Profile (USP) 20. Prospective recording of data was approved by the French authority CCTIRS (Advisory Committee on information processing in healthcare research).
Statistical analysis was performed using Stata 11.0 software (StatCorp). Patient characteristics, surgical procedures, postoperative outcomes and score values were presented as numbers and percentages (qualitative variables) or mean and SD (continuous variables). Women managed in Rouen and Bordeaux were compared using either the Kruskal-Wallis test (continuous variables) or the Fischer exact test (qualitative variables). A logistic regression model was used to identify factors independently related to the risk of rectovaginal fistula. A P value of <.05 was considered statistically significant. The study was approved by the Rouen University Hospital Institutional Ethics Committee for Non-Interventional Research (E2020-53, June 30, 2020).