A prospective algorithm to reduce anastomotic leaks after rectosigmoid resection for gynecologic malignancies

Objective Determine whether a standardized protocol for temporary bowel diversion after rectosigmoid resection (RSR) for cytoreduction can reduce the rate of anastomotic leak (AL). Methods A prospective quality improvement project for patients undergoing RSR during debulking surgery from 07/2013 to 01/2016 was conducted. Patients with any of the following underwent temporary diversion: preoperative albumin ≤ 3.0 g/dL, prior pelvic radiation, RSR plus additional large bowel resection (LBR), anastomosis (AS) ≤ 6 cm from the anal verge, failed leak test or contamination of the pelvis with stool. The AL rate was compared to the historic AL rate from 01/04–06/11. Results Seventy-seven patients underwent RSR, with 27 (35.1%) receiving diverting stomas vs. 25/309 (8.1%) in the historic cohort. Additional LBR (33.3%) and AS at ≤ 6 cm from anal verge (26.3%) were the most common indications for diversion. No AL was observed among diverted patients. If one AL which occurred following protocol violation (failed leak test but not diverted) is excluded, the theoretical AL rate is 1.3% (1/77) vs. 7.8% (24/309; P = 0.039) in the historic cohort. Not excluding this case, the AL rate was 2.6% (2/77) vs. 7.8% (P = 0.11). Short-term outcomes following primary surgery were not different between diverted and non-diverted patients. Stoma-related complications were observed in 7/27 (25.9%) patients, primarily related to dehydration. Reversal surgery was successfully performed in 24/75 (88.9%) patients. Conclusions Criteria-based temporary bowel diversion for patients undergoing RSR for gynecologic cancer reduced the AL rate. Diversion was associated with acceptable morbidity and high reversal rate.