Abstract
Background: Bowel surgery is a key component of advanced deep endometriosis management, with anastomotic leakage representing the most serious postoperative complication. This study aimed to identify risk factors for dehiscence after lower colorectal anastomosis and to determine effective preventive measures. Methods: This retrospective/prospective study included 425 consecutive patients aged 37.7 ± 6.0 years with laparoscopical bowel resection due to multiorgan complex deep endometriosis. All bowel surgeries were performed with use of indocyanine green (ICG). Many technical aspects of surgery and preventive procedures were analyzed which could impact leakage risk of surgery. Results: Endometriotic nodules were resected with segmental bowel resection (n = 294; 69.8%), discoid bowel resection (n = 84; 20.0%), and shaving procedure (n = 43; 10.2%). A total of 12 dehiscence events occurred (2.8%), including intraperitoneal leakage (n = 1; 0.2%), rectovaginal fistula (RVF) (n = 10; 2.3%), and rectoureteral fistula (n = 1; 0.2%); no rectovesical fistulas were observed. RVF developed only following segmental resections. Protective measures used during lower bowel procedures included fibrin glue (n = 375; 88.2%), omental flaps (n = 86; 20.2%), reinforcing sutures (n = 33; 7.8%), protective stomas (n = 25; 5.9%), and ghost stomas (n = 14; 3.3%). Among patients who developed RVFs, 90% had no protective stoma, and these cases were predominantly associated with low (from 6 to <8 cm; n = 4/77; 5.2%) and very low (from 5 to <6 cm; n = 4/10; 40%) anastomoses. In very low anastomoses (n = 4), 1 RVF occurred despite a protective stoma but there existed other strong risk factors, such as levator ani infiltration and vagina opening, whereas 3 others RVF developed in patients without a protective stoma. Notably, in ultra-low anastomoses (<5 cm), protective stomas prevented the anastomosis in 100%, and no fistula was observed (n = 3). The following factors were associated with the increased rate of RVF: segmental resection (p = 0.0355), low and very low anastomosis (p = 0.0010), lateral infiltration of the levator (p < 0.0001), concomitant hysterectomy or vaginal opening (p = 0.051), and prolonged operative time (p = 0.0010), Clostridioides difficile infection (p = 0.0001). Conclusions: RVFs occurred mainly after segmental resection (no other type of bowel resection), with very low anastomosis (5-6 cm from anal verge), in patients with levator ani infiltration and concomitant vaginal or uterine surgery; in such situations, discoid resection is the safer option. Despite the complexity of procedures, preventive strategies maintained a low overall RVF rate; no RVFs occurred in ultra low anastomoses (<5), indicating effective prevention with protective stomas.