Multivariate analysis of factors influencing the results of restorative proctocolectomy.

Univariate and multivariate analyses have been used to assess the influence of 14 variables on the results of 65 consecutive ileoanal pouch procedures over 5 years. There were nine failures requiring intubation, ileostomy or pouch excision. There was a significant association between failure and pelvic sepsis (P less than 0.05, n = 8), endoanal mucosectomy (P less than 0.05, n = 7), preservation of a long rectal cuff (P less than 0.05, n = 5) and lack of experience with the operation (P less than 0.05, n = 8). Of 49 patients with preoperative evidence of ulcerative colitis, three are now known to have Crohn's disease. Functional outcome was significantly impaired in patients who developed pelvic sepsis (P less than 0.01) or a postoperative fistula (P less than 0.05), and who had an endoanal mucosectomy (P less than 0.05). Success with ileoanal pouch reconstruction increases with experience. Avoidance of sepsis is associated with a lower failure rate, improved functional results and reduced hospital stay. Preliminary colectomy is also advised to exclude Crohn's disease if the diagnosis is in question.