Factors associated with failure in managing pelvic sepsis after ileal pouch-anal anastomosis (IPAA)--a multivariate analysis.

Background: Pelvic sepsis is known to cause a detrimental outcome after ileal pouch-anal anastomosis (IPAA). The aim of this study was to examine potential factors associated with failure in managing pelvic sepsis after IPAA.

Methods: We performed univariate and multivariate logistic regression analysis on 2518 IPAA patients between 1983 and 2005. Failure was defined as pouch failure, the need for a permanent ileostomy, or mortality as a result of sepsis. There were 157 patients (6.2%) with pelvic sepsis after IPAA. These involved anastomotic leak 34% (54/157) and fistula 25% (40/157). There were 5 mortalities related to sepsis. Mean age at surgery was 38.1 +/- 14.4 years and mean follow-up was 5.5 +/- 4.7 years.

Results: Pouches were saved in 75.8% patients. Univariate analysis identified early sepsis (P = .040), preoperative steroid use (P = .007), and need for percutaneous drainage (P = .004) as significant factors associated with treatment success. Factors associated with failure were hypertension (P = .026), hand-sewn anastomosis (P = .038), associated fistula (P = .0003), need for transanal drainage (P = .0002), need for laparotomy to control septic complications (P < .0001), delayed ileostomy closure (P = .0003), and need for a new diverting ileostomy (P < .0001). By using multivariate analysis with selected covariates, significant factors associated with failure were associated fistula (P = .0013), need for transanal drainage (P = .003), delayed ileostomy closure (P = .022), need for a new ileostomy diversion (P = .004), and hypertension (P = .039). We developed a predictive scoring system for failure to use in management plans and decision-making for the treatment of septic complications of IPAA.

Conclusions: Pelvic sepsis after IPAA has a significant impact on pouch failure. This predictive model for failure may play an important role in providing risk estimates for successful outcomes.