Ileal Pouch Anal Anastomosis Without Ileal Diversion

Abstract

Objective

To evaluate continued experience with a one-stage stapled ileoanal pouch procedure without temporary ileostomy diversion.

Summary Background Data

Most centers perform colectomy, proctectomy, and ileal pouch anal anastomoses (IPAA) with a protective ileostomy. Following a previous report, the authors performed 126 additional stapled IPAA procedures for ulcerative colitis and familial adenomatous polyposis, of which all but 2 were without an ileostomy. Outcomes in these patients question the need for temporary ileal diversion, with its complications and need for subsequent surgical closure.

Methods

Two hundred one patients underwent a stapled IPAA since May 1989, 192 as a one-stage procedure without ileostomy, and 1 with a concurrent Whipple procedure for duodenal adenocarcinoma. Patient charts were reviewed or patients were contacted by phone to evaluate their clinical status at least 1 year after their surgery.

Results

Among the patients who underwent the one-stage procedure, 178 had ulcerative colitis (38 fulminant), 5 had Crohn’s disease (diagnosed after IPAA), 1 had indeterminate colitis, and 8 had familial adenomatous polyposis. The mean age was 38 ± 7 (range 7–-70) years; there were 98 male patients and 94 female patients. The average amount of diseased tissue between the dentate line and the anastomosis was 0.9 ± 0.1 cm, with 35% of the anastomoses at the dentate line. With 89% follow-up at 1 year or more (mean 5.1 ± 2.4 years) after surgery, the average 24-hour stool frequency was 7.1 ± 3.3, of which 0.9 ± 1.4 were at night. Daytime stool control was 95% and night-time control was 90%. Only 2.3% needed to wear a perineal pad. Average length of hospital stay was 10 ± 0.3 days, with 1.5 ± 0.5 days readmission for complications. Abscesses or enteric leaks occurred in 23 patients; IPAA function was excellent in 19 of these patients (2 have permanent ileostomies). In patients taking steroids, there was no significant difference in leak rate with duration of use (29 ± 8 with vs. 22 ± 2 months without leak) or dose (32 ± 13 mg with vs. 35 ± 3 mg without leak). Two (1%) patients died (myocardial infarction, mesenteric infarction).

Conclusions

The triple-stapled IPAA without temporary ileal diversion has a relatively low complication rate and a low rate of small bowel obstruction, provides excellent fecal control, permits an early return to a functional life, and can be performed in morbidly obese and older patients.

Almost all centers perform most colectomy, proctectomy, and ileal pouch anal anastomoses (IPAA) with a protective ileostomy. ^1–3 Several centers have noted a significantly higher anastomotic leak rate in steroid-dependent ulcerative colitis (UC) patients and in male patients. ^4–7 We previously described 75 patients with a triple-stapled IPAA, of whom 68 had the operation as a one-stage procedure without temporary ileal diversion. ^8,9 In those studies, we did not find the leak rate to be higher in steroid users or male patients. Since these reports, we have performed 126 additional stapled IPAA procedures for UC and familial adenomatous polyposis (FAP), of which all but 2 were without an ileostomy. Outcomes in these patients question the need for temporary ileal diversion, with its complications and need for subsequent surgical closure.

METHODS

Preoperative Management

The initial consultation included clinical assessment of disease activity and medications, including quantity and duration of steroid therapy. The risks and benefits of a one-stage triple-stapled IPAA were discussed in detail with the patient and family. There was a detailed discussion of the potential risks of an anastomotic leak and the possible need for an emergency ileostomy should signs or symptoms of peritonitis be present, or the need for prolonged nonsurgical management until the leak resolved. There was also a discussion of the potential risks of routine ileal diversion, including anastomotic leak, dehydration from excessive ileostomy output, need for another operation to close the ileostomy, and presumably increased risk of subsequent small bowel obstruction. Severely obese patients (body mass index ≥ 30 kg/m²) were told that should a leak occur with peritonitis, it might be necessary to remove the ileoanal pouch because of the difficulty in bringing out an ileostomy when the ileum is tethered to the anus.

Three weeks before surgery, patients were asked to donate 2 units of autologous blood, provided they were not anemic. During this time, patients were asked to take three iron tablets per day (one with each meal). Patients underwent an outpatient mechanical and antibiotic bowel preparation with 4 L of a polyethylene glycol solution or a bottle of Fleet Phospho-soda and oral metronidazole the day before surgery and were admitted on the morning of surgery unless they had fulminant colitis, which had necessitated earlier admission. Patients taking prednisone were given 100 mg intravenous hydrocortisone before the induction of anesthesia and then two additional doses on the first day.

Surgical Technique

The surgical technique has been described in detail in our previous publications. ^8,9

Postoperative Management

The nasogastric tube was removed by the anesthesiologist. The patient was asked to get out of bed on the night of surgery. The urinary bladder catheter was usually removed on the second postoperative day. The Jackson-Pratt drains were left in place until the patient was having normal bowel movements and there was no fecal material in the drain. The patient was given clear liquids once bowel movements and the passage of flatus had begun and was advanced rapidly to a general, high-fiber diet as tolerated. If the patient felt bloated, developed hiccups, or vomited, a nasogastric tube was inserted and left in place until normal stool function returned. Intravenous steroids were weaned as appropriate, and the patient was converted to oral prednisone when tolerating a general diet. Because the initial stools would be frequent and watery, an effort was made to reduce perianal excoriation: patients were instructed to bring soft toilet paper with them to the hospital; they were told to blot and avoid rubbing their perianal skin; and they were given creams to protect their skin.

The greatest concern was the potential for an anastomotic leak. This required continued vigilance by the surgeon and team. If the patient developed feculent drainage in the Jackson-Pratt drains, reported rectal or pelvic pain, or developed a fever after removal of the drains, an emergent Gastrografin, water-soluble contrast radiographic enema was obtained. If a leak was present and drained by the Jackson-Pratt drains or was confined to the pelvis if the drains had been removed and the patient had no clinical signs of peritonitis, the patient was managed without ileal diversion. However, if there were signs or symptoms of peritonitis, the patient underwent emergency laparotomy, abdominal and pelvic saline lavage, and construction of a diverting ileostomy. Leaks confined to the pelvis were managed with thrice-daily 200-mL saline enemas beginning 2 weeks after the leak was diagnosed, and the patient was allowed to eat a low-fiber diet. Repeat water-soluble contrast enemas were obtained until there was documented evidence of complete healing, at which time the patient was permitted to eat a general diet. Antibiotic coverage was not given for a leak after discharge.

Follow-Up

Patients were initially seen 2 weeks after discharge, at 3 months, and yearly thereafter whenever possible. Patients who lived a great distance from Richmond were asked to have follow-up examinations with their referring gastroenterologist. Yearly pouch endoscopy was suggested. The most significant long-term problem with the IPAA procedure is the development of recurrent attacks of pouchitis. The diagnosis of pouchitis was made clinically with the development of frequent, watery diarrhea, urgency, accidents, blood, or fever that responded to treatment with oral antibiotics. It did not require endoscopic biopsy confirmation of inflammation. Treatment was usually initiated with oral metronidazole 250 mg twice daily for 2 weeks, and patients were given a PRN refill prescription. They were instructed to reinstitute treatment on their own for recurrent symptoms of pouchitis. If the patient failed to respond to metronidazole or became refractory to its effects, treatment was with generic tetracycline, 500 mg twice daily, cotrimoxazole (Bactrim-DS), one capsule twice daily, or ciprofloxacin, 250 mg twice daily. Some patients required daily antibiotic treatment to keep their pouchitis in check. Patients who failed to respond to antibiotic therapy were given 5-aminosalicylic acid (mesalamine, 5-ASA) suppositories, mesalamine or steroid enemas, or oral 5-ASA medications.

Patient Assessment

The IPAA database was queried and patient charts were reviewed from March 1989, when the triple-stapled IPAA procedure was initiated, until March 31, 1999. Patients who had not been seen in the office in the past 2 years were called by one of the ileoanal nurse coordinators (E.L.S., J.G.M.). Data retrieved included daytime and nocturnal stool frequency, presence and frequency of daytime or nocturnal accidents, need to wear a perineal pad, presence and frequency of pouchitis attacks, and need for and type of antibiotic treatment or other medications. Patients were considered to have accidents only if they occurred more than once per month. Additional data recorded included need for rectal eversion for stapling, presence of a leak, time of diagnosis of a leak (before or after drain removal, before or after hospital discharge) and its association with subsequent IPAA function, need for emergent ileostomy, rectal stricture requiring dilation under anesthesia with its possible relation to anastomotic leak, presence of a perianal or pouch–vaginal fistula, small bowel obstruction necessitating surgical adhesiolysis, male impotence or dry ejaculation and female dyspareunia, and need and reason for removal of the ileoanal pouch.

Data are presented as mean ± standard deviation. Comparative data were analyzed by analysis of variance or paired t test when appropriate; differences were considered significant at P < .05.

RESULTS

Of the 201 triple-stapled IPAA procedures performed since May 1989, all but 12 were performed by one surgeon (H.J.S.). Since our last report in 1994, ⁹ 126 additional triple-stapled IPAA procedures were performed, of which all but 2 were without temporary ileostomy as a one-stage procedure, leaving a total of 192 one-stage procedures. The two patients who underwent diverting ileostomies since our last report both had anastomotic leaks identified with methylene blue immediately after IPAA construction, and both underwent repair at the time of surgery; no leak was found when tested again with methylene blue. One of these patients developed a postoperative leak that delayed takedown of the ileostomy.

Of the one-stage group, 178 had UC, 5 subsequently were found to have Crohn’s disease (unsuspected at the time of the IPAA procedure in 3), 1 had indeterminate colitis, and 8 had familial adenomatous polyposis (FAP). One patient underwent a concurrent one-stage IPAA and standard radical pancreatoduodenectomy (Whipple procedure) for duodenal adenocarcinoma associated with large duodenal villous adenomas and FAP. There were 98 male and 94 female patients. A total colectomy and ileostomy had been previously performed in 17 patients for fulminant UC. All but four of these patients had their procedure performed elsewhere and were referred for the IPAA procedure. Acute UC was present in 38 patients and an emergent colectomy, proctectomy, and one-stage IPAA was performed in 28 patients for acute UC; of these, 10 had toxic colitis and 18 had refractory hemorrhagic colitis. The PI-30 stapler could not be placed at the distal limit of the dissection in eight male patients because of a very narrow bony pelvis; the rectum was everted and the stapler placed externally from below. The average amount of diseased tissue between the dentate line and the anastomosis was 0.9 ± 0.1 cm, with 35% of the anastomoses at the dentate line.

Pelvic abscesses occurred in 3 patients, an intraabdominal abscess in 1 patient, and enteric leaks in 19 patients. Fourteen of the leaks were at the IPAA, three were at the pouch staple line, one was in the proximal ileum, and one was in the proximal jejunum (Table 1). The ileal leak was presumed to be secondary to adhesiolysis in a patient with a previous emergent colectomy during pregnancy for fulminant colitis, and the jejunal leak was presumably secondary to a cautery injury during the colonic dissection. Nine of the leaks were diagnosed when fecal material was noted from the drain, 10 after the drain had been removed, and 8 after hospital discharge. The leak was diagnosed an average of 7.3 ± 3.8 (range 1–14) days after surgery, and it took an average of 2.5 ± 2.8 months (median 1 month, range 2 weeks to 1 year) to have documentation of healing. Examples of leaks that were managed successfully without a diverting ileostomy are shown in Figures 1 through 5.

Table 1. ANASTOMOTIC LEAKS

Figure 1. Ileal pouch-anal anastomotic leak in patient 113, demonstrated with a water-soluble contrast (Gastrografin) enema 10 days after surgery. The leak is confined to the pelvis, and the Jackson-Pratt drains have been removed.

Figure 2. Follow-up Gastrografin enema study at 1 week after surgery shows that the leak in Fig. 1 has healed.

Figure 3. Follow-up Gastrografin enema study at 1 month after surgery shows that the leak in Fig. 2 has healed.

Figure 4. Ileal pouch-anal anastomotic leak in patient 176, demonstrated with a Gastrografin enema 12 days after surgery. The leak is confined to the pelvis but does not communicate with the Jackson-Pratt drain.

Figure 5. Leak shown in Fig. 4 healed with follow-up Gastrograffin enema study at 1 month after IPAA.

One patient (a 55-year-old man with chronic UC) had previously undergone coronary angioplasty and died as a consequence of an intraoperative myocardial infarction when undergoing exploration for a leak. IPAA function was excellent in 19 of the 23 patients with IPAA leaks; 2 had permanent ileostomies and 1 had problems with excessive stool frequency and leakage.

Of the 178 UC patients, 145 (81%) were taking 34.3 ± 2.5 mg prednisone, 4 were taking azathioprine, 4 were taking 6-mercaptopurine, and 2 were taking cyclosporine before surgery. There were no significant differences (Table 2) in the leak rate in patients taking steroids versus not taking steroids, sex, prior colectomy, ileostomy or new colectomy, or acute versus chronic UC. In steroid-dependent patients, there were no significant differences (see Table 2) in the leak rate with duration of steroid use or prednisone dose. Eight severely obese patients (seven with UC, one with FAP; body mass index ≥30 kg/m²) underwent a one-stage IPAA; there was one (12.5%) leak. Several older patients underwent a one-stage triple-stapled IPAA (16 were older than 60, 8 were older than 66, 3 were older than 70 years); there was one (6%) leak, and the patient underwent diversion but subsequently had excellent pouch function.

Table 2. ANASTOMOTIC LEAKS

There were no significant differences.

There was one postoperative death (67-year-old man with UC), secondary to a massive small bowel infarction from a superior mesenteric artery thrombosis on postoperative day 9. The autopsy revealed atherosclerotic tissue in the aorta and the origin of the superior mesenteric artery.

One patient died 2 years after surgery of metastatic colon cancer. There were three other late deaths: a 75-year-old man died of metastatic lung cancer 4 years after surgery, one patient committed suicide after the diagnosis of unresectable esophageal carcinoma 5 years after surgery, and a 29-year-old man died secondary to sepsis after intensive immunosuppressive therapy for pyoderma gangrenosum that failed to respond to either colectomy/IPAA or ileoanal pouch excision 4 years later for refractory pouchitis.

The average hospital stay was 10 ± 0.3 days, with 1.5 ± 0.5 days readmission for complications.

Colonic dysplasia was present in 17 patients and adenocarcinoma in 8 before the colectomy and IPAA procedure. Of the patients with dysplasia, an unsuspected adenocarcinoma of the appendix was found in one patient at the time of his IPAA operation; he was disease-free 7 years after surgery.

Additional complications are listed in Table 3. Follow-up in the clinic or by telephone contact in these patients was 89% at 1 year or more (mean 5.1 ± 2.4 years) after surgery. Fecal frequency and continence are listed in Table 4.

Table 3. OTHER COMPLICATIONS

IPAA, ileal pouch-anal anastomosis.

* In absence of re-exploration for peritonitis.

Table 4. FECAL FUNCTION AFTER TRIPLE-STAPLED IPAA

IPAA, ileal pouch-anal anastomosis.

89% follow-up at ≥1 year (mean 4.1 ± 2.8 years) after IPAA; patients without clinical evidence of pouchitis.

Pouchitis occurred in 42% of patients to date. Of these, 10% had only one or two attacks per year; 80% had recurrent episodes necessitating treatment for 2 weeks, usually with antibiotics, at approximately 2-week intervals. Although metronidazole is usually the first antibiotic chosen because it is generic and inexpensive (used in 54%), the most effective antibiotic has been ciprofloxacin (used by 48%). Other antibiotics that have been used include Bactrim-DS (11%) and tetracycline (4%). Approximately 10% of patients had chronic pouchitis and require continuous antibiotic therapy, mesalamine suppositories (8.5%), mesalamine enemas (2.8%), steroid enemas (8.5%), or oral mesalamine (9%). Two patients had their pouch removed because of refractory pouchitis.

During the past 10 years, nine (4.5%) patients underwent pouch excision or permanent ileal diversion. Of these procedures, two were related to pouch–vaginal fistulas secondary to Crohn’s disease that was not diagnosed before the colectomy/ileoanal procedure, two were a consequence of anastomotic leaks, three were secondary to refractory pouchitis, and one was due to excessive stool frequency. One patient developed dysplasia in the residual colitic tissue above the dentate line. He underwent a completion mucosectomy and reconstruction of his ileoanal anastomosis but subsequently developed adenocarcinoma at the anastomosis and underwent pouch excision and permanent ileostomy; despite margins and lymph nodes that were free of disease, he was dying of metastatic colon cancer as of this writing.

DISCUSSION

This series expands our previous reports on a one-stage triple-stapled ileoanal pouch procedure for UC and FAP. ^8,9 Our patients had excellent control of fecal function, both day and night, with the triple-stapled IPAA. Similar results have been noted by other surgeons. 2,4,10–15 The randomized, prospective ^16–18 or retrospective ¹⁹ IPAA trials comparing the stapled to the hand-sewn mucosectomy procedure found no significant differences in fecal control. During the day, 95% of our patients report almost perfect fecal continence (90% at night); 98% state they do not need to wear a perineal pad. This is in contrast to the results of randomized studies ^16–18 or single series reports. ^19–22 The Mayo Clinic’s randomized, prospective study found a significantly better resting anal sphincter pressure in their stapled patients, ¹⁸ similar to the results of our earlier study ⁸ and those of others, ^12–15 but it failed to find significantly better control of fecal function with the stapled procedure. ¹⁸ In the most recent Mayo Clinic report, 5% of their patients had frequent nocturnal incontinence, 5% to 25% had occasional daytime incontinence, 42% to 58% had occasional nocturnal incontinence, and 16% to 39% had to wear perineal pads, depending on the patient’s sex and the length of time since surgery. ²²

We are concerned that the reason for these differences may be related to the patient assessment methods. Patients with active pouchitis may be having difficulties with fecal control secondary to this problem rather than to the function of their anal sphincter mechanism. When patients begin to report increased stool frequency and problems with continence after a stapled IPAA, it is usually a sign of recurrent pouchitis and almost always responds to antibiotic treatment. In our series, patients were either seen in the office or queried by telephone if they had not been seen in the past 2 years and were asked to assess their fecal control when they were not having episodes of pouchitis. In the other series, they were queried by a mailed questionnaire, and no attempt was apparently made to differentiate fecal control with or without attacks of pouchitis.

Our data also imply that it is not necessary to leave any tissue above the dentate line to achieve excellent fecal control, as others have suggested. ^23–25 We believe that one should get as close to the dentate line as possible with the stapled approach.

The second major advantage, in our opinion, to the triple-stapled ileoanal pouch procedure is the ability in the vast majority of patients to perform this operation as a one-stage procedure without temporary ileal diversion. This potential advantage has not been included in any of the randomized, prospective trials comparing the mucosectomy, hand-sewn to the stapled IPAA procedures. This approach, however, produces a marked increase in the surgeon’s postoperative anxiety factor because it requires vigilance on the part of the operating team, looking for signs and symptoms of an anastomotic leak after surgery. Should a leak occur, the surgeon then needs to decide whether the patient requires emergent fecal diversion. In our series, we had a 12% frequency of anastomotic leaks. Of the leaks in the most recent 126 patients since our last publication, ⁹ two thirds of the patients did not develop peritonitis and were managed successfully without a laparotomy and ileostomy. Some authors have stated that pelvic sepsis and an IPAA leak are associated with an unacceptable frequency of ileoanal pouch dysfunction. ^26,27 This has not been our experience. Of the 23 patients with leaks or pelvic abscesses, 19 have excellent pouch function. One third of our pouch leaks occurred after the patient had been discharged from the hospital. We believe it is extremely important for these patients to be able to return promptly to their surgeon should they develop rectal pain, bleeding, or fever. This mandates an urgent physical examination to determine whether there are signs of peritonitis and a water-soluble contrast enema to verify that the leak is confined to the pelvis. If the leak is not confined to the pelvis or there are signs or symptoms of peritonitis, urgent laparotomy, abdominal lavage, and ileal diversion is necessary.

Several reports have noted a significantly higher leak rate in patients taking steroids versus those who are not. 4–6,28–30 One study suggested that patients taking 30 mg prednisone per day or more for 3 years or more had a significantly higher frequency of anastomotic leaks. ⁷ This has not been our experience. We found no significant difference in the leak rate in patients who took steroids versus those who did not. The dose or duration of prednisone treatment had no relation to the leak rate. Therefore, we do not use absence of steroid therapy as a criterion for performing a one-stage pouch procedure. Other groups have also supported routine one-stage IPAA procedures without a temporary ileostomy. ^29–31 However, several studies have supported selective use 4,5,32,33 or urged caution in the routine use of a one-stage IPAA procedure. ^4,5,34

There were no additional cases of male sexual dysfunction since our previous report. ⁹ However, there were six cases of pouch–vaginal fistulas; three of these patients were found to have Crohn’s disease after their colectomy and IPAA procedure. This problem has been noted by others. ³⁵ A recent report noted healing of pouch–vaginal fistulas with infliximab. ³⁶ We have not used this medication. Eleven patients had strictures of the IPAA despite use of a 31-mm EEA stapler; only two of these were associated with an anastomotic leak, and all responded to digital dilatation under anesthesia, with subsequent self-dilatations at home. Others have also noted this complication but found most to be secondary to an anastomotic leak and pelvic sepsis. ³⁷

One of the concerns with the triple-stapled ileoanal pouch procedure is the small amount of remnant diseased mucosa between the dentate line and the anastomosis. Potential problems include persistent disease activity and the risk of cancer. Only 5% of our patients had problems with persistent anal burning and discomfort. This almost always responded to hydrocortisone suppositories or cream. No patients lost their pouch as a result of this problem. These results are similar to those reported by others. ^38,39 However, one study noted that several patients required pouch revision because of persistent problems with the residual colitic tissue from stapled ileoanal pouch procedures performed elsewhere. ⁴⁰ We had hypothesized that the risk of developing adenocarcinoma in the residual colitic tissue would be approximately 1% over 30 years. ^8,9

One of our patients, age 51, underwent the triple-stapled one-stage procedure for UC associated with dysplasia and developed sigmoid colon dysplasia in the residual colitic tissue 3 years after surgery. Subsequent mucosectomy and anal advancement was performed, but an invasive adenocarcinoma developed at the IPAA 1 year later, leading to a radical pouch excision. Despite the presence of clear margins and the absence of nodal metastases, this patient was dying of metastatic adenocarcinoma as of this writing.

Others have had success with mucosectomy of the residual colitic tissue after the stapled IPAA in one patient for severe dysplasia. ⁴¹ We have had 8 patients with colon adenocarcinoma and 16 others with dysplasia who have undergone the triple-stapled ileoanal pouch procedure. Of these, none of the 7 surviving patients with adenocarcinoma or 13 other patients with dysplasia has developed dysplasia or adenocarcinoma in this residual tissue more than 5 years after surgery. Two other patients have been described who developed adenocarcinoma at the pouch-anal anastomosis after a mucosectomy and hand-sewn IPAA. ^42,43

There is a concern about the risk of dysplasia in the ileoanal pouch in patients with UC. ⁴⁴ However, a study of 56 children who had at least three biopsies of their pelvic pouches constructed a mean of 5 years previously found no evidence of dysplasia in them. ⁴⁵ There is a report of one patient who developed carcinoma in the ileoanal pouch. ⁴⁶ Patients with pelvic pouches performed for FAP are also at risk for developing polyps in their ileoanal pouch, and carcinoma. ⁴⁷ Thus, we strongly encourage these patients to undergo annual endoscopic evaluations of their pouch; however, we have not performed routine biopsies of the pouch. Two of our patients developed polypoid nodules in the pouch that proved on microscopic examination to be inflammatory tissue associated with chronic pouchitis.

The Achilles’ heel of the ileoanal pouch procedure is the development of chronic pouchitis. ^48,49 Forty-two percent of our patients developed pouchitis; it was refractory to antibiotic treatment in 10%, necessitating the use of mesalamine suppositories or enemas, steroid enemas, or oral mesalamine medications. Three of our patients lost their pouches because of intractable pouchitis.

We had a comparatively low frequency (5%) of small bowel obstructions necessitating laparotomy. We hypothesize that many of the small bowel obstructions reported in the literature occurred at the ileostomy closure anastomosis; this is another major advantage of performing the operation as a one-stage procedure, although a recent study suggested that this problem is reduced with a stapled anastomosis. ⁵⁰

If an ileostomy were a completely benign procedure, we would support its routine use. However, in addition to small bowel obstruction, there are several other ileostomy-related complications. These include a 6% to 10% risk of anastomotic leak, dehydration and electrolyte imbalance secondary to excessive stool output, herniation at the ileostomy closure site, as well as the need for a second surgical procedure, with its associated disability. ^4,51–55

In addition, the one-stage procedure makes this operation suitable for severely obese patients with UC or FAP. It may be technically impossible to bring out an ileostomy in a patient with a large amount of subcutaneous tissue when the ileal pouch is tethered to the anus.

Emergent proctocolectomy with an ileoanal pouch procedure may also be performed in selected patients, provided the surgical procedure proceeds smoothly, the patient is not severely malnourished, and the tissues do not seem to be excessively friable. Our leak rate in these patients was similar to the rate in patients with chronic UC. Others have performed, with excellent results, a similar procedure with ileal diversion for patients with fulminant colitis. ⁵⁶ However, if surgeons do not have a large experience with the ileoanal pouch procedure, or if the patient is not an appropriate candidate, the patient should undergo total colectomy and ileostomy without any dissection of the rectum so that a subsequent proctectomy, ileostomy takedown, and ileoanal pouch construction can be performed without another diverting ileostomy.

We believe this is an appropriate procedure for older patients who have adequate sphincter function that can be adequately assessed by a preoperative digital rectal examination. Our patients older than 65 or 70 have almost all had excellent fecal control after the IPAA procedure. Others have also had good results with the stapled ileoanal pouch procedure in older patients. ^57–59

In conclusion, the one-stage triple-stapled ileoanal pouch is a challenging operation that carries a significant but acceptable rate of anastomotic leak; most of these leaks heal without the need for emergent ileal diversion. The procedure provides excellent stool control and an earlier return to a functional life. It requires a vigilant surgeon and the willingness to treat patients who develop a leak without ileal diversion or emergent laparotomy and ileostomy when there is evidence of peritonitis. The operation may be performed in severely obese and older patients with excellent results. There is a small risk of cancer or occasional problems with discomfort and minor bleeding in the small amount of residual colitic tissue left between the dentate line and the anastomosis. Although the randomized studies have not found improved fecal control with the stapled procedure, our patients who are not having attacks of pouchitis have much better fecal control than has been reported for the mucosectomy and hand-sewn IPAA procedures in any of the published randomized or nonrandomized studies.

Discussion

Dr. John H. Pemberton (Rochester, Minnesota): Dr. Sugerman reports on essentially a single surgeon’s experience with ileoanal anastomosis in over 200 patients. It is indeed an impressive series. But it is important to remember that reports of a single surgeon’s experience, which often achieve very impressive results, may not be generalizable beyond that surgeon’s practice. So my first question to you, Dr. Sugerman, is, can you expect your results to be reproduced by the larger group of surgeons operating on patients with inflammatory bowel disease and familial polyposis?

Speaking of familial polyposis, we learned early in our experience with ileal anal anastomosis that FAP patients had much better functional outcomes, as a group, than did the IBD patients, and thus have reported the results of ileoanal anastomosis for patients with FAP and inflammatory bowel disease separately. If you remove the non-IBD patients from your denominator, do your results suffer?

Although the controversy surrounding hand-sewn and stapled ileoanal anastomosis has been less acute of late, there continue to be passionate advocates of one or the other of these approaches. Your functional results are really exemplary, and you report commendably low rates of fecal incontinence compared to other reports in the literature. I think you attribute this largely to the stapling anastomosis.

We performed a randomized prospective trial comparing hand-sewn to stapled ileoanal anastomoses in about 30 patients and found no difference in functional outcomes. Because of these small numbers, however, we combined our series with the three other randomized series in the literature, and metaanalysis confirmed little functional differences between these operations. Please comment on your results in light of this prospectively acquired data.

Finally, a diverting ileostomy performed at the time of ileoanal anastomosis is practiced widely here and abroad at centers doing a lot of ileoanal surgery, as you mentioned at the beginning of your presentation. Each group has reported on ileoanal anastomosis performed without a defunctioning stoma and patients will often do well. But most surgeons at these institutions continue to believe they would rather deal with the mostly tiresome complications associated with ileostomy and its closure rather than the potentially catastrophic pelvic sepsis and pouch failure that can occur in patients without a diverting stoma. Can you tell us why you believe no stoma is better than a stoma, Dr. Sugerman? And is there ever a situation in which you would perform a defunctioning stoma at the time of the ileoanal anastomosis?

Presenter Dr. Harvey J. Sugerman (Richmond, Virginia): We believe the data are reproducible if other surgeons follow the guidelines in this manuscript and in our previous publications. If the patient is severely malnourished (albumin <2.5 g/dL), if the tissues are friable and will not hold sutures well, or if there are technical difficulties with the procedure, then I believe the patient should have a diverting ileostomy. Our leak rate was the same in the polyposis patients as in the ulcerative colitis patients, although our number of polyposis patients was small.

I believe our data show that the fecal control is much better with the stapled ileoanal pouch procedure. I have great difficulty understanding why the randomized studies haven’t shown the improved fecal control that many of us who do the stapled procedure, and who used to do the mucosectomy operation, see. Fecal continence in this and our previous series is much better than that of patients in the control arms of the published randomized studies. We wonder whether patients who have diarrhea, urgency, and accidents with pouchitis are included in the analysis of fecal continence in the stapled patients when they are assessed by mailed questionnaires. Your randomized study at the Mayo Clinic showed a better fecal continence with the stapled than with the mucosectomy and hand-sewn anastomosis, but the difference wasn’t statistically significant. This study may have had a type 2 beta error: had you entered more patients in that study, you might have found a significantly improved outcome in the stapled group. The anal sphincter pressures in your study, as in our previous study and those of other investigators, were significantly better in the stapled patients.

We believe that the other major advantage of the stapled approach is the ability to perform the operation in one stage without temporary ileal diversion. As 96% of our patients were able to undergo this operation without need for ileal diversion, as 19 of 23 of the patients with anastomotic leaks or pelvic abscesses have excellent pouch function, and as the ileostomy operation is not a free ride but carries a number of significant complications, including leaks and small bowel obstructions, we believe that the data show that a one-stage procedure permits the patient to have an earlier return to a functional life.

Dr. Zane Cohen (Toronto, Ontario, Canada): Dr. Sugerman, the results that you presented are excellent. They are almost unbelievable, and I have several questions I want to ask.

Firstly, you reported a leak rate of about 12%. You separated out your pouch–vaginal fistulas as well as your pouch–perineal fistulas. If you include both, your total leak rate is well over 17%, which, I would say, is unacceptably high. To me, leaks are the major cause of pouch failure, and therefore I would like you to explain why you have actually separated those out.

I would also like to know exactly how you measure the distance to the anastomosis. If, in fact, the dissection in 35% is taken down to the dentate line, you are actually taking out half, the top half, of the internal sphincter. It therefore doesn’t make physiologic sense that your results are the same for those patients whose dissection is taken down to the dentate line and those that have 1 or 2 cm of a muscular cuff left.

I would also like to know how you define acute colitis. To do a one-stage procedure in acute colitis and toxic megacolon I think is dangerous and unwise.

Then the question of not doing a mucosectomy in patients who have dysplasia. You described 17 patients who have either dysplasia or cancer, and you have left mucosa at risk behind. I think that that is very much against what most individuals with large recorded series would do. And in fact, one of your patients did develop dysplasia in the residual anorectal remnant and is, as you say in the manuscript, dying from cancer. I would like you to explain a little bit more about that.

You also state in your paper that it is not really necessary to do biopsies in patients who have pouchitis. I think it is very important to do biopsies because there is now a known association between the inflammatory response, chronic pouchitis, severe villus atrophy, and the development of dysplasia. I believe that there is a place for a one-stage procedure. There is a small place for it. We have to know when to do it, but not as a routine.

Dr. Sugerman: I would have to say that the pouch–anal and pouch–vaginal fistulas are probably leaks. The reason we separated them out is that these fistulas didn’t occur until between 2 and 10 years after their ileoanal pouch procedure, so they weren’t in any way clinically apparent for 2 to 10 years after the surgery. In several instances they occurred when patients developed very frequent stools associated with the development of pouchitis. I think in a way it is a matter of semantics. We are talking about clinically apparent leaks that occurred requiring subsequent follow-up and management. I believe most other series do not include the pouch–vaginal fistulas and pouch–perianal fistulas as true leaks in the way we are discussing them. When you state that our leak rate of 17%, including these fistulas, is unacceptably high, your own prior published series had an even higher leak rate than that. But you did say our results were excellent, perhaps unbelievable!

We measure the distance between the anastomosis and the dentate line with a rigid sigmoidoscope at the completion of the procedure, using the centimeter ruler on the sigmoidoscope. I believe there is variability as to the location of the dentate line from patient to patient and I certainly don’t think that we are taking out half of the internal sphincter, as you suggested, when we are able to get the anastomosis at the dentate line with the dissection from above and the PI-30 horizontal stapler.

In terms of acute colitis, we are talking about patients who developed severe colitis and are not able to be managed medically and come to surgery within a year of the diagnosis. Of these, the patients who were diagnosed with toxic colitis did not, except in two instances, have toxic megacolon. They were toxic with high fevers, high white counts, but did not have a true megacolon. I think you have to be careful that you do not restrict the diagnosis of toxic colitis to patients only with megacolon. I remember as a resident at Penn having a patient die from toxic colitis without megacolon. I think it is safe to do this operation in some acute colitis patients when the patient is not severely malnourished and when the operation proceeds smoothly. If the patient is severely malnourished with an albumin less than 2.5 and is critically ill, it is more reasonable to do a colectomy with an ileostomy and come back at a later date to do the proctectomy and pouch procedure.

With regard to the issue of doing the operation without a mucosectomy in patients with dysplasia or colon cancer, we have had a number of patients with these problems and, yes, we have had one patient with disastrous results. Dr. Fazio reported a similar patient who underwent a subsequent mucosectomy of residual tissue that had developed dysplasia. I gather from the publication that his patient has done well. None of our other patients who had colectomies for dysplasia or cancer have developed dysplasia or tumors in their residual tissue.

In terms of biopsying the pouch, where are you going to biopsy it? If the patient develops a lesion in the pouch, a lump, it should be biopsied. This has occurred in three of our patients and the tissue in each of these patients was diagnosed as inflammation. We have never seen dysplasia in the pouch. To my knowledge, there is only one case in the literature of a patient developing carcinoma in their pelvic pouch and, yes, there have been a couple of cases of dysplasia reported from Sweden.

Dr. Victor W. Fazio (Cleveland, Ohio): Dr. Sugerman, you and your colleagues have reprised your previous studies by adding the 126 patients to this report. Well, what is new, what is different, what is true, and what is important are the issues I want to discuss right now.

The “new” is the larger series. The “different” is the lowering of your leak rate from around 16% to 7% as you define it and the pouch failure rate from 16% to 4%, giving an overall leak rate of 12% or an unplanned ileostomy return rate of 5%. In our department’s experience, my colleague Dr. Remzi noted that in 175 unprotected pouches—about 8% of our ileal pouch series—anastomotic leaks occurred in 2.8% and there were zero pouch failures. But we use specific criteria and are very selective in using an unprotected pouch. What is “true” is that like most reports with one-stage pouches, bowel obstruction is certainly less common in the undiverted pouch, and your data reflects that here today. This was derived using historical case-match controls of protected ileal pouches. And what is “important”? In this unit’s experience, anastomotic and septic complications could be reduced over the course of time. The issue is, is the current rate of complications acceptable, and how can they be improved? I would offer the recommendation that a greater selective use of stoma avoidance could be used. I would like to raise several questions.

First, is the improvement in your complication rate due to the increasing individual surgeon experience, or the use of a circular stapler, or both? Given the fact that there is less tension with a stapled anastomosis, that would probably be a major factor.

Second, have you used any survey instrument to assess patient satisfaction or quality of life?

Third, how often, if at all, have you had to abandon the attempt to do an ileal pouch procedure, given abandonment rates in the literature of 4% to 6%?

Fourth, are you not pushing the envelope in attempting a pouch procedure in the morbidly obese, given your 13% leak rate with that group and need for salvage ileostomy in about the same rate? Why not simply do a colectomy and withdraw the steroids, so weight loss can be achieved before a second-stage pouch construction can be done?

Do you have any contraindications to stoma-avoiding pelvic pouches? You mentioned hypoalbuminemia, but what happens if you have incomplete donuts? What happens if you have patients on cyclosporine or immunosuppressants or if you have a leak on pouch testing?

Finally, although the results of this series are commendable, the mortality rate, pouch failure rates, and fistula rates are considerations that, given better results from similar high-volume units using temporary diversion, some selection process is warranted to use temporary diversion to give further outcomes improvement.

Dr. Sugerman: Our leak rate is really not much different than in our previous publications and our pouch failure rate is stable. The higher rate was in the original series of 64 patients in which we did a mucosectomy and hand-sewn anastomosis. We have not performed any survey instruments to assess patient satisfaction or quality of life. However, your group has done that with the stapled ileoanal procedure and your patients, with a very similar operation, have excellent results. We have had to abandon the stapled operation in one severely obese patient (0.5%) and convert to a mucosectomy and hand-sewn procedure. The anastomosis was under great tension and failed due to the development of a long, ischemic stricture in the pouch. The stapled approach appears to provide less tension than the mucosectomy and hand-sewn procedure and is less likely to be abandoned.

If the ileostomy didn’t have any risks, then I think everybody should have one. But the ileostomy is not a free ride. There are complications associated with it in every published series, including leaks, excessive fluid loss, electrolyte imbalance, and, I think, a major increase in the risk of small bowel obstruction. I believe that the risks and benefits favor the one-stage approach in the vast majority of patients.

Yes, we are pushing the envelope in the morbidly obese patient. The proctectomy and ileoanal procedure can be very, very difficult in these individuals. Our leak rate in the morbidly obese is no different than in the other patients. In terms of getting them off steroids to have them lose weight, these patients are morbidly obese and—trust me, I know a little about morbid obesity—they are not going to lose weight when you get them off prednisone. In fact, all of these patients have remained morbidly obese after their the colectomy and pouch procedure. I agree that if they don’t have complete donuts after the EEA anastomosis or if a leak is noted at the completion of the procedure with insertion of methylene-blue–colored saline under pressure into their pouch, they should be diverted. Immunosuppression has not been a contraindication in our series to a one-stage procedure and does not appear to have adverse effects, although our numbers of immunosuppressed patients are small.

I think one of the major findings in this report—and actually it came from some work that Zane Cohen’s group reported to this association a number of years ago—is that most of these leaks will heal on their own with saline enemas and a low-fiber diet, without antibiotics, with the patient at home. Prior to Dr. Cohen’s study, when I saw a leak like the ones I showed you, there was a knee-jerk reaction that the patient needed immediate ileal diversion. Our data confirm that most of these leaks don’t need fecal diversion to heal.

Dr. Theodore R. Schrock (San Francisco, California): I think that Dr. Sugerman is in need of a friend, so I rise to support his program. We too adopted a policy a few years ago of performing ileostomy only on specific indications.

In 143 patients since adopting that policy, I found it necessary, based on a preoperative or intraoperative decision, to do an ileotomy in 23 patients, or about 16%. This is a far greater ileostomy rate than you report. Judging by your criteria, our decision-making is much too cautious.

On the other hand, of the 120 patients where we did not perform an ileostomy, leaks leading to ileostomy occurred in seven patients, or 5.8%, and there were three additional patients who had an ileostomy that was unplanned for other reasons, for a total of 10, or about 8% overall. I agree with your observation that most of these leaks are contained and not catastrophes that endanger the patient.

In my opinion, this approach is only possible if one does the triple-stapled technique, as you term it. If one does a mucosectomy and a hand-sewn anastomosis or a mucosectomy and a distal hand-applied purse-string and then stapling, avoidance of ileostomy is extremely dangerous. Would you agree with that?

Dr. Sugerman: I would agree that patients who have a mucosectomy and hand-sewn anastomosis require ileal diversion. If the patient has tissues that are falling apart like wet toilet tissue and are unlikely to hold sutures or staples, or the patient is hypoalbuminemic, that patient needs a colectomy/ileostomy and a chance to come back another day when it is safer to do the proctectomy and ileoanal pouch procedure. That is why in some instances these patients had a previous colectomy, and then their ileoanal procedure after they were no longer nutritionally crippled.

Dr. David A. Rothenberger (St. Paul, Minnesota): Just a quick technical question on the management of these leaks. When you sent them home with their saline enemas, had you done anything such as digitally opening up the fistula tract at the ileoanal anastomosis to make sure it was draining through the anus? Or did they just go home without any such opening?

Dr. Sugerman: We have not digitally dilated the fistula tract and have only dilated the anastomosis when a stricture developed. Of the 11 strictures we have had, only 2 were associated with a leak. That is one of the interesting findings from these data: most of these leaks have healed without external drainage when they were contained within the pelvis.

Dr. James M. Becker (Boston, Massachusetts): You have a fairly high incidence of late anal canal complications, including strictures, fistulae, and so forth. Would you speculate about the reason for that? Is it the stapling itself, the retained disease-bearing mucosa, or the lack of the ileostomy?

Secondly, your length of stay, as you said in the abstract, is about 10 days, which is about 4 days longer than our current average length of stay. Do you think that is related to the fact that they do not have a diverting ileostomy?

Finally, you talked about steroids and the relationship to leak, no leak, or other complications, and the fact that the majority of your patients were on steroids. Most of our patients are on steroids, but they are also almost all on immunotherapy. Could you comment on how that influences your algorithm regarding the need for a diverting ileostomy?

Dr. Sugerman: I think in terms of the apparent high frequency of stricture and fistulas, I don’t think they are any higher than most series in the literature. As I mentioned, the pouch–vaginal fistulas, three of the patients had Crohn’s disease that we hadn’t appreciated prior to the procedure. One patient had had radiation therapy for a cervical cancer. It could be related to the residual tissue left behind, but I don’t think so.

In terms of the length of stay of 10 days, it may be that after the one-stage stapled procedure, when they begin to have bowel movements, the stools are often watery and frequent. It takes a little time for them to become thickened and controlled. I thought that was a rather reasonable length of stay for a patient who underwent this operation, especially since 96% of them avoided a second operation with its pain, suffering, and delay in return to a functional life. Half of the patients never required an NG tube. In terms of the risks of a one-stage procedure with the addition of immunotherapy plus steroids, our numbers are too small to make a conclusive statement, but I don’t think there is an increased risk of complications if they are on cyclosporine or 6-MP. But it will take more data to answer that question.