A randomized controlled trial was performed to assess the role of loop ileostomy in totally stapled restorative proctocolectomy. Entry criteria included all patients who were not on corticosteroids in whom on-table testing revealed a watertight pouch with intact ileoanal anastomosis. Of 59 patients undergoing restorative proctocolectomy over 36 months, 45 were eligible and were randomized to loop ileostomy (n = 23) or no ileostomy (n = 22). The age and diagnosis of the groups were similar. There were no deaths; two ileoanal anastomotic leaks occurred, one in each group. Ileoanal stenosis occurred in five patients with and one without an ileostomy. The incidences of wound and pelvic sepsis, bowel obstruction and pouchitis were similar. Twelve patients (52 per cent) developed ileostomy-related complications. The median total hospital stay was 23 (range 13-75) days with ileostomy and 13 (range 7-119) days without (P < 0.001). This study indicates that there is a low risk of pelvic sepsis which is not increased by avoiding a protective ileostomy. Loop ileostomy was associated with a high incidence of complications.

Twenty of 81 patients treated by restorative proctocolectomy for presumed ulcerative colitis had some features of Crohn's disease: 10 were classified as definite Crohn's disease and 10 as indeterminate colitis. These pathological features were first apparent during synchronous colectomy and pouch construction in 10 of 11 cases. In the remainder, histological features of possible Crohn's disease were first identified during rectal excision (n=6), preliminary subtotal colectomy (n=2), and after pouch excision (=2

Twenty-seven patients developed a fistula after 168 restorative proctocolectomies. Thirteen fistulas were enterocutaneous (two with communication to the bladder); their origins were from the pouch (three patients), the ileoanal anastomosis (three), the pouch appendage (three), a previous loop ileostomy (two) and iatrogenic small bowel injury (two). Two patients had Crohn's disease. The pouch was removed in four patients, one of whom died from chronic small bowel obstruction; the remaining nine have satisfactory pouch function after fistula excision. Ten pouch-vaginal fistulas occurred, all from the ileoanal anastomosis; four were extrasphincteric. Four of these patients had underlying Crohn's disease. Only two patients, with Crohn's disease and indeterminate colitis, required pouch excision; the remainder have good pouch function after treatment of the fistula. There were three pouch-perineal fistulas, all from the ileoanal anastomosis; these were successfully managed by seton fistulotomy. There was one pouch-vesical fistula, successfully treated by excision of the fistula and pouch appendage.

(Gut 1993; 34: 680-684) Restorative proctocolectomy is a useful operation for ulcerative colitis, familial adenomatous polyposis, and some diseases where colonic function is severely disturbed.

A retrospective study compared the outcome of restorative proctocolectomy in patients who had a covering ileostomy (n = 53) with those who had no proximal stoma (n = 32). Those who had a loop ileostomy had a higher incidence of anastomotic leakage (21 per cent), pelvic abscess (32 per cent) and postoperative fistula (28 per cent) than those with no covering ileostomy (6, 12 and 12 per cent respectively). Intestinal obstruction occurred in 23 per cent of those with an ileostomy, compared with 6 per cent in those who had no stoma. The functional outcome was identical.