No abstract

Aim This study aimed to assess outcomes of Hartmann's reversal (HR) after failure of previous colorectal anastomosis (CRA) or coloanal anastomosis (CAA).Methods All patients planned for HR from 1997 to 2018 following the failure of previous CRA or CAA were included.Results From 1997 to 2018, 45 HRs were planned following failed CRA or CAA performed for rectal cancer (n = 19, 42%), diverticulitis (n = 16, 36%), colon cancer (n = 4, 9%), inflammatory bowel disease (n = 2, 4%) or other aetiologies (n = 4, 9%). In two (4%) patients, HR could not be performed. HR was performed in 43/ 45 (96%) patients with stapled CRA (n = 24, 53%), delayed handsewn CAA with colonic pull-through (n = 11, 24%), standard handsewn CAA (n = 6, 14%) or stapled ileal pouch-anal anastomosis (n = 2, 4%). One (2%) patient died postoperatively. Overall postoperative morbidity rate was 44%, including 27% of patients with severe postoperative complication (Clavien-Dindo ≥ 3). After a mean follow-up of 38 AE 30 months (range 1-109), 35/45 (78%) patients presented without stoma. Multivariate analysis identified a remnant rectal stump < 7.5 cm in length as the only independent risk factor for long-term persistent stoma. Among stoma-free patients, low anterior resection syndrome (LARS) score was ≤ 20 (normal) in 43%, between 21 and 29 (minor LARS) in 33% and ≥ 30 (major LARS) in 24% of the patients.Conclusion HR can be recommended in patients following a failed CRA or CAA. It permits 78% of patients to be free of stoma. A short length of the remnant rectal stump is the only predictive factor of persistent stoma in these patients.Keywords Hartmann's reversal, redo surgery, anastomotic leakage, definitive stoma, functional resultsWhat does this paper add to the literature? This study suggests that reversal of Hartmann's procedure in patients following a failed colorectal or coloanal anastomosis permits 78% of the patients to be stoma free. A short remnant rectal stump (< 7.5 cm) was the only predictive factor of permanent stoma in these patients.