Diverted versus undiverted restorative proctocolectomy for chronic ulcerative colitis: an analysis of long-term outcomes after pouch leak short title: outcomes after pouch leak.

BACKGROUND: The safety of undiverted restorative proctocolectomy (RPC) is debated. This study compares long-term outcomes after pouch leak in diverted and undiverted RPC patients.

METHODS: Data were obtained from a prospectively maintained registry from a single surgical practice. One-stage and staged procedures with an undiverted pouch were considered undiverted pouches; all others were considered diverted pouches. The outcomes measured were pouch excision and long-term diversion defined as the need for loop ileostomy at 200 weeks after pouch creation. Regression models were used to compare outcomes.

RESULTS: There were 317 diverted and 670 undiverted pouches, of which 378 were one-stage procedures. Pouch leaks occurred in 135 patients, 92 (13.7%) after undiverted, and 43 (13.6%) after diverted pouches. Eighty-six (64%) leaks were diagnosed within 6 months of pouch creation. Undiverted patients underwent more emergent procedures within 30 days of pouch creation (p < 0.01). Pouch excision occurred in 14 (33%) diverted patients and 13 (14%) undiverted patients (p = 0.01). Thirteen (32%) diverted patients and 18 (21%) undiverted patients (p = 0.17) had ileostomies at 200 weeks after surgery. In multivariable analyses, diverted patients had a higher risk of pouch excision (HR 3.67 p < 0.01), but similar rates of ileostomy at 200 weeks (HR 1.8, p = 0.19) compared to undiverted patients.

CONCLUSIONS: Despite a likely selection bias in which "healthier" patients undergo an undiverted pouch, our data suggest that diversion does not prevent pouch excision and the need for long-term diversion after pouch leak. These findings suggest that undiverted RPC is a safe procedure in appropriately selected patients.