The results of the pelvic pouch procedure were reviewed to assess the surgical complication rate and outcome of patients who had had the procedure performed with a stapled ileo-anal anastomosis with and without a defunctioning ileostomy. Between December 1982 and March 1992, 483 patients underwent a pelvic pouch (PP) procedure. Patients were divided into three groups: group I consisted of 325 patients (178 men and boys and 147 women and girls) who underwent a PP procedure with a handsewn ileoanal anastomosis (IAA) with a defunctioning loop ileostomy. In group II, there were 87 patients (47 men and boys and 40 women and girls) who had a stapled IAA with a defunctioning ileostomy. Group III patients consisted of 71 patients (43 men and boys and 28 women and girls) who had a stapled IAA with no covering ileostomy. Assessment was made of the IAA leak rate, the surgical complications, the reoperation rate, and functional outcome. Early surgical complications included 40 (12%) IAA leaks in group I patients compared with only six (7%) leaks in group II patients who had a stapled IAA (p < 0.05). In group III patients, who had a stapled IAA but no covering ileostomy, there were 13 leaks (18%). Eleven of these 13 leaks healed spontaneously with tube drainage; one patient remains with a rectal tube in place 6 weeks after operation, and only one patient has required a reoperation (defunctioning ileostomy). Functionally, all patients with a healed IAA after a leak have had an excellent result comparable to those without a leak. Patients who were male, older than age 40, on steroids, and had had a true one-stage PP procedure, had a greater risk of developing an IAA leak. In two patients, there was intraoperative difficulty, and one of these patients had an IAA leak after operation. Disease activity at the resection margin and patient weight did not affect the leak rate. Our results suggest that the IAA leak rate is significantly reduced in patients with a stapled IAA with an ileostomy compared with those with a handsewn IAA. Omission of the defunctioning ileostomy is associated with a higher IAA leak rate, but spontaneous healing occurs in almost all patients without impairment of functional results. In patients in whom the ileostomy is omitted, the IAA leak rate is greatest in male patients who have undergone a true one-stage PP procedure, are on steroids, and are older than age 40.

Persistent rectovaginal fistula presents a difficult management problem. Choice of operation must be tailored to the underlying pathology and type of repair previously done. Advancement flap repair is generally not recommended for persistent complex fistulas or for simple fistulas that have failed a previous advancement flap repair.

The functional success rate of the ileoanal reservoir procedure for ulcerative colitis is quite high. Despite the few early and late complications described there is now widespread acceptance of this procedure in the management of ulcerative colitis. We report a patient who developed an adenocarcinoma in the rectal cuff four years after having a pelvic pouch procedure. This new late complication brings to light several points including the importance of a radical total mucosectomy. The purpose of this paper is to discuss concern as to whether or not this procedure is indicated in colitis patients in whom severe dysplasia is the primary indication for surgery. Abdominal colectomy, rectal mucosectomy and ileoanal pouch anastomosis has gained increased acceptance as an alternative to total proctocolectomy and end ileostomy in the management of ulcerative colitis. A concern about this procedure is the fate of any islets of rectal mucosa left behind or regenerated in the cuff of rectal muscle stripped of its mucosa. The literature suggests that residual islets of mucosa appear in up to 20% of these procedures.' Regeneration of islets, however, does not seem to occur.2 This concern is particularly relevant in patients in whom the indication for surgery is severe dysplasia.3The increased risk of developing colonic carcinoma in patients with ulcerative colitis has been known for many years, especially in younger age groups who have had more extensive colitis for longer than 20 years.45 Cancer in colitis is frequently a fatal illness whch is entirely preventable by a total colectomy.3 A recent report suggests that a pelvic pouch is indicated even in the presence of colonic cancer providing that the cancer does not invade the pelvic floor muscles and the resection margins do not damage the sphincter mechanism.6Case report We report the first patient to have undergone a total colectomy, rectal mucosectomy and ileoanal pelvic pouch anastomosis, who subsequently developed a cancer in the residual rectal cuff. This 59 year old man was diagnosed as having ulcerative colitis at age 21. He was well for 28 years until 1977 when he was found to have

Ras mutations are an important early event in a number of carcinogen-induced rodent tumors. Colon carcinogenesis induced in rats by azoxymethane is a useful model as it mimics the adenoma-carcinoma sequence observed in humans. In addition, aberrant crypt foci develop in the rat and these lesions appear to be potentially important precursors to adenomas in colorectal cancer. Recent studies have shown that specific K-ras codon 12 and 13 mutations are present in up to 66% of carcinogen-induced rat colon adenocarcinomas. We studied the frequency of these mutations during the aberrant crypt focus-adenoma-carcinoma sequence in azoxymethane-induced Fisher F344 rats. K-ras codon 12 GAT and codon 13 GAC mutations were detected with a sensitive assay based on the amplification of DNA using the polymerase chain reaction. No mutations were present in normal mucosa. Of 27 aberrant crypt foci, K-ras mutations were identified in 2 lesions containing 5 and 10 aberrant crypts, respectively. Mutations were present in 1 of 23 and 10 of 27 adenomas and adenocarcinomas, respectively. These data suggest that K-ras mutations play a role during the stages of carcinogenesis in azoxymethane-induced rat colon cancer. The demonstration of a genetic mutation in aberrant crypt foci provides further evidence for the significance of these lesions as precursor markers of malignant potential during colorectal tumorigenesis.