INTRODUCTION: Pouch-related fistula occurs in 5-10 per cent of patients after restorative proctocolectomy. The present study identified risk factors associated with the development of such fistulas. METHODS: Data on preoperative and postoperative risk factors were recorded from 1965 patients who underwent restorative proctocolectomy in a single tertiary centre between 1983 and 2001. Cox regression analysis was used to identify independent predictors of pouch-perineal, pouch-abdominal wall and pouch-vaginal fistula during follow-up. RESULTS: Median patient follow-up was 4.1 (range 0-19) years. By 15 years' follow-up, pouch-vaginal fistulas had occurred in 44 women (5.2 per cent). The prevalence of ileal pouch-perineal and pouch-abdominal wall fistula was 3.6 per cent (70 patients) and 1.5 per cent (30 patients) respectively. Independent predictors of pouch-related fistula identified by multivariate analysis were diagnosis of indeterminate colitis or Crohn's disease (hazard ratio (HR) 1.28 (95 per cent confidence interval (c.i.) 1.00 to 1.65) and 1.73 (95 per cent c.i. 1.07 to 3.48) respectively versus ulcerative colitis or familial adenomatous polyposis), previous anal pathology (HR 3.43 (95 per cent c.i. 2.43 to 4.84) and 4.02 (95 per cent c.i. 1.27 to 12.77) respectively for perineal abscess and fistula in ano versus no previous anal pathology), abnormal anal manometry (HR 4.29 (95 per cent c.i. 2.33 to 7.91)), patient sex (HR 0.74 (95 per cent c.i. 0.58 to 0.95) for men versus women) and pelvic sepsis (HR 3.79 (95 per cent c.i. 2.48 to 5.79)). CONCLUSION: This study suggests that Crohn's disease and the clinical signs that favour the diagnosis of Crohn's disease may contribute to the development of pouch-related fistula.
INTRODUCTION: Pouch-related fistula occurs in 5-10 per cent of patients after restorative proctocolectomy. The present study identified risk factors associated with the development of such fistulas. METHODS: Data on preoperative and postoperative risk factors were recorded from 1965 patients who underwent restorative proctocolectomy in a single tertiary centre between 1983 and 2001. Cox regression analysis was used to identify independent predictors of pouch-perineal, pouch-abdominal wall and pouch-vaginal fistula during follow-up. RESULTS: Median patient follow-up was 4.1 (range 0-19) years. By 15 years' follow-up, pouch-vaginal fistulas had occurred in 44 women (5.2 per cent). The prevalence of ileal pouch-perineal and pouch-abdominal wall fistula was 3.6 per cent (70 patients) and 1.5 per cent (30 patients) respectively. Independent predictors of pouch-related fistula identified by multivariate analysis were diagnosis of indeterminate colitis or Crohn's disease (hazard ratio (HR) 1.28 (95 per cent confidence interval (c.i.) 1.00 to 1.65) and 1.73 (95 per cent c.i. 1.07 to 3.48) respectively versus ulcerative colitis or familial adenomatous polyposis), previous anal pathology (HR 3.43 (95 per cent c.i. 2.43 to 4.84) and 4.02 (95 per cent c.i. 1.27 to 12.77) respectively for perineal abscess and fistula in ano versus no previous anal pathology), abnormal anal manometry (HR 4.29 (95 per cent c.i. 2.33 to 7.91)), patient sex (HR 0.74 (95 per cent c.i. 0.58 to 0.95) for men versus women) and pelvic sepsis (HR 3.79 (95 per cent c.i. 2.48 to 5.79)). CONCLUSION: This study suggests that Crohn's disease and the clinical signs that favour the diagnosis of Crohn's disease may contribute to the development of pouch-related fistula.