Although most patients with ulcerative colitis escape surgery, 2 per cent of those with distal colitis and up to 33 per cent of those with extensive disease will require an operation.

Severe acute colitis is characterized by the passage of more than six loose, bloody motions per day, together with systemic signs, including tachycardia, fever, and hypoalbuminaemia. The crucial points in management are early recognition, aggressive medical therapy, and regular review by a joint medical and surgical team. Stool is cultured to rule out a specific bacterial cause or the rare case of amoebic colitis. Corticosteroids are administered intravenously (400 mg hydrocortisone or 64 mg methylprednisolone daily) and rectally (100 mg hydrocortisone in 100 ml normal saline, infused into the rectum over 15–20 min). The patient is allowed only sips by mouth; fluids are given intravenously.

This treatment is successful in 75 per cent of patients, whose illness resolves within 5 to 7 days. They will feel well, passing two to three semiformed, bloodless motions each day. The remaining 25 per cent either deteriorate despite treatment or fail to recover completely; they may suffer a relapse when a normal diet is resumed. In both of these instances urgent surgery is indicated, since these patients will all require surgery in the subsequent few weeks and there seems little point in delay. A surgical procedure that does not result in a permanent ileostomy is also more acceptable to anxious and ill patients. The administration of antibiotics or steroids at a higher dosage has no therapeutic advantage and is not recommended. Early studies suggested a possible use for cyclosporin in the treatment of acute colitis, but the results of controlled trials are awaited.

If within the first 24 h of admission the patient has a pulse rate greater than 100 beats/min, a temperature over 38°C, or passes more than nine stools per day, or if mucosal islands are visible on a plain abdominal radiograph (Fig. 1) 1043, surgery will probably be required. Perforation and colonic dilatation (the upper normal limit of colon diameter is 5 cm) that fails to resolve rapidly when medical treatment is started or that develops during treatment is an indication for emergency surgery. Massive colorectal bleeding may also require emergency surgical intervention. Intraoperative colonic lavage and colonoscopy can help locate the source of bleeding and may obviate the need for proctocolectomy.

The main indication for elective surgery is chronic illness that responds poorly to medical treatment or that is punctuated by frequent episodes of severe acute colitis. The threshold for referral by physicians and the degree of debility which many patients are prepared to tolerate has been reduced with the advent of restorative proctocolectomy. Physical, social, and employment factors should all be considered in the evaluation of a patient, as these can be affected by symptoms of the colitis or by the medication used in its treatment. The patient's age and the duration of illness are also important considerations when deciding on the appropriateness of surgical intervention.

Benign colorectal strictures produced by ulcerative colitis, and some extraintestinal manifestations such as extensive pyoderma gangrenosum (Fig. 2) 1044 and iritis (but not sclerosing cholangitis or ankylosing sponylitis), are rare indications for surgery.

Patients with ulcerative colitis have an increased risk of developing colorectal carcinoma (Fig. 3) 1045. This situation is particularly true for those who have chronic total colitis which started as a severe first attack in childhood. A precancerous stage is indicated by the finding of dysplasia in colorectal biopsy specimens. Surveillance programmes have been established to examine the large bowel at colonoscopy every 2 years when the disease has been present for 10 years. Biopsy specimens are taken either of specific lesions or from random sites of the mucosal field, as dysplasia in one area can indicate the presence of an unsuspected malignancy elsewhere.

Results from these programmes suggest that the probability of developing a carcinoma is 3 per cent at 17 years, 5 to 7 per cent at 20 years and 9 to 1 per cent at 25 years. Very few, if any, cancers develop before a 10-year-history of colitis.

Unfortunately, dysplasia is not an ideal marker of malignancy. Its detection is subject to observer variation, and cancer can develop in the absence of changes at a distant site. Identifying dysplastic changes, however, is the only currently available method for detection of a precancerous state, and is most reliable when severe or high-grade changes are found in a biopsy from a villous or polypoidal lesion (Fig. 4) 1046. A similar degree of dysplasia from an area of flat mucosa is less likely to be associated with synchronous carcinoma. The finding should be verified by a second, independent, pathologist and colectomy should be advised only if dysplastic change is present at two sites in the colon. Surgery is also recommended when low-grade dysplasia is found in a raised lesion; its presence in flat mucosa is an indication for increased vigilance only.

Until a more sensitive and specific marker of premalignancy is found, the true value of screening for colorectal cancer in ulcerative colitis remains uncertain.

Until quite recently, complete proctocolectomy (or panproctocolectomy) was the best operation for patients with ulcerative colitis who required surgery. This position has, however, been challenged by the advent of restorative proctocolectomy which, with few exceptions, is now considered the first-choice elective surgical treatment for this condition. Restorative proctocolectomy has also forced a re-evaluation of the recommended management for those undergoing urgent or emergency operations. The various options available and their indications are shown in Table 1 329.

Until the development of the Brooke ileostomy and advanced materials for stoma appliances, surgery was a last resort. In the 1950s, however, proctocolectomy became the established procedure.

This operation has several advantages: all diseased tissue is excised at one operation, it removes the risk of cancer, it is a well tried and usually uncomplicated procedure, and the patient can return to normal activities expeditiously. Unfortunately, the patient is left with a permanent ileostomy, which may have serious social and psychological consequences. In the long-term the ileostomy may require revision and small bowel obstruction may develop due to adhesions. Chronic perineal sinus occurs in 5 to 10 per cent of cases and delayed healing of the perineal wound is not uncommon. Sexual dysfunction arising from damage to pelvic autonomic nerves occurs in 0.5 to 1 per cent of male patients. Dyspareunia may be a result of adhesions in female patients.

With the advent of restorative surgery, this procedure is currently indicated only in elderly patients, particularly those with weak anal sphincters, and in those unwilling to subject themselves to a more complicated operation with the possibility of additional morbidity and a longer hospital stay. In Oxford, one-quarter of our patients elect to have a complete proctocolectomy; the majority choose the restorative procedure.

With the patient in the modified lithotomy Trendelenburg position, the abdomen and perineum are prepared. The rectum is washed out with an antiseptic and a close perianal purse string is placed. The previously marked ileostomy site is trephined. The abdomen is opened through a long mid-line incision, the colon mobilized and the vessels divided comfortably near the bowel, from the caecum to sigmoid. The pelvic dissection can be carried out in the close rectal or mesorectal plane (Fig. 5) 1047. The perineal operator meanwhile makes a close perianal incision and develops the intersphincteric plane. Infiltration with dilute adrenaline may aid this dissection. The specimen is removed through the perineal wound. The perineum is closed in layers with absorbable sutures. Suction drains are placed from above.

The terminal ileum is delivered through the stoma trephine and everted to make a 2- to 3-cm spouted (Brooke) ileostomy. The lateral space is closed to prevent small bowel herniation.

Complications include pelvic bleeding and intra-abdominal sepsis. Small bowl adhesion obstruction occasionally requires laparotomy. Stoma protraction or retraction requiring revisional surgery occurs in 5 to 10 per cent of patients.

Specialist nurse support in stomatherapy is essential. The stomatherapist will mark the appropriate stoma site, and provide information, counselling and postoperative care in stoma management.

This procedure retains the advantages of a complete proctocolectomy and replaces the spouted Brooke ileostomy with a flush, continent stoma and an intra-abdominal reservoir constructed of ileum (Fig. 6) 1048. Early technical problems, particularly with the nipple valve, have been resolved, but the advent of restorative proctocolectomy has virtually abolished the indications for this procedure. It could be considered in a patient who has already had a complete proctocolectomy and who expresses a strong wish to be rid of the ileostomy.

This procedure is attractive because it is relatively quick and straightforward, and it can be associated with a good functional outcome. However, its use remains controversial because, by leaving behind the rectum it is associated with a continuing risk of carcinoma. It might be considered in a patient with a compliant rectum which is relatively spared from disease. The presence of a functioning sphincter is essential, and the patient must be reliable as life-long monitoring is required. Coexisting sclerosing cholangitis or portal hypertension, which would make rectal dissection hazardous, are also indications for this operation. It is also an option, in children, allowing them to pass through adolescence without a stoma or until a pouch is indicated.

This is the procedure of choice for the elective treatment of ulcerative colitis. It is not indicated in the presence of Crohn's disease, a lower-third rectal cancer (a high early cancer does not exclude restorative proctocolectomy if local tumour clearance can be achieved without sphincter excision), or poor anal sphincter function. Current operations involve the creation of an ileal reservoir of various designs, followed by an anastomosis to the anal canal.

Both complete proctocolectomy and subtotal colectomy, with formation of an ileostomy and mucous fistula, have been used in the treatment of acute colitis. However, with the possibility of later restorative surgery, subtotal colectomy has become the procedure of choice. Timing of surgery is obviously critical, and with careful management the ‘late case’ should be avoidable. Turnbull introduced the principle of ‘blow hole’ colostomy, in which the thin friable perforating colon was managed by multiple ostomies to the abdominal wall. This avoided the hazardous resection of a difficult colon, but it is rarely required nowadays. The splenic fixture is usually the most dangerous area in an emergency colectomy, and this part of the procedure must be undertaken with great care, packing off the area in case of an inadvertent perforation. If omentum is adherent to the colon it should be left in place and should not be preserved.

The best way to manage the mucous fistula remains undecided. It can be sewn to the skin at the lower end of the laparotomy wound should the rectum break down or, less conventionally, closed and placed subcutaneously. This will avoid creation of a second stoma but allow for discharge of blood and pus through the wound should the rectum break down. In either case, subsequent identification of the rectum is straightforward. Closure of the rectum at the peritoneal reflection (as in Hartmann's procedure) carries the risk of possible perforation of the inflamed stump and abscess formation. Conservative proctocolectomy which includes removal of the rectum to the level of the pelvic floor is not advised. Proctocolectomy is associated with a high degree of morbidity and may make any subsequent pelvic dissection technically difficult. It is only indicated in the rare patient with colitis and rectal bleeding. The various options for managing the rectum after subtotal colectomy are shown in Fig. 7 1049.

In 1947, Ravitch and Sabiston demonstrated that in both animals and humans it was possible to remove the rectal mucosa and leave denuded but functional rectal muscle. This mucosal proctectomy could be followed by creation of an ileoanal anastomosis, pulling the ileum through the rectal muscle cuff and sewing it to the modified skin of the anal canal. Ravitch performed this operation in two young adults with ulcerative colitis, who were subsequently continent. Although the technique was tried by a number of other surgeons over the following decade, the procedure never gained widespread acceptance because of operative complications, incontinence, intolerable frequency, and perianal dermatitis. However, the considerable success of mucosal proctectomy and pull-through procedures for children with Hirschsprung's disease tempted some paediatric surgeons to try a similar approach in young patients suffering from ulcerative colitis, with encouraging results.

A better understanding of the mechanisms of normal continence and defecation, together with the introduction of the Kock pouch, were the stimuli for the development of the pelvic ileal reservoir. The use of a single narrow tube of ileum as a replacement for rectum resulted in very poor reservoir function. Parks and colleagues introduced a pelvic triple-loop ileal reservoir, which was sewn to the anus after mucosal proctectomy. Utsunomiya in Japan and Fonkalsrud in the United States developed similar procedures shortly afterwards.

Parks' original operation has undergone considerable modifications since it was first described. Several of these, together with other important issues surrounding restorative proctocolectomy, are discussed below.

Many patients suffering from ulcerative colitis are young and sexually active. Preservation of the pelvic nerves, and in particular the nervi erigentes, is of great importance. Early proctocolectomy was associated with a 10 per cent incidence of impotence, but adoption of a perimuscular plane for rectal dissection reduced the incidence of this complication to below 0.5 per cent. Similar results have been achieved by dissection in the avascular plane around the mesorectum posteriorly and laterally but close to the bowel wall anteriorly (Fig. 5) 1047. This technique is reported to be quicker and associated with less blood loss; it is widely used in North America. Formal comparison of the two methods and information on the relative incidences of sexual dysfunction are not available.

The pouch originally described by Parks and Nicholls was a triple-loop S pouch with a long efferent limb which emptied poorly and frequently required intubation to effect complete evacuation. This problem has been overcome by shortening the length of the efferent segment. A two-loop J pouch, four-loop W pouch, and lateral isoperistaltic or H pouch have also been described. These pouches can be hand-sewn or stapled and use approximately 40 to 50 cm of terminal ileum. The various designs are illustrated in Fig. 8 1050; their function is compared below. Figure 9 1051 shows a triplicated S pouch in situ in the pelvis.

When restorative proctocolectomy was first introduced, the rectum was transected 10 cm above the pelvic floor and the mucosa striped off the underlying muscle by a tedious endoanal dissection up from the dentate line. This achieved the aim of excising all potentially diseased mucosa, but also removed the anal transitional zone, an area of cuboidal rather than columnar mucosa which extends for approximately 1.5 cm above the dentate line and has a rich sensory nerve supply.

The rectal muscle cuff has not been shown to be essential for the appreciation of pouch fullness; indeed it may increase the incidence of pelvic sepsis if left in place. On the other hand, the anal transitional zone is thought to play an important role in sensory discrimination as well as incontinence. The postoperative functional results of restorative proctocolectomy are better if this mucosa is preserved. However, there is a long-term risk from leaving behind potentially diseased mucosa and retention of the anal transition zone is not universally accepted. Pathological changes suggestive of ulcerative colitis have been found in anal canal mucosa from patients undergoing proctocolectomy. Even more worrying are reports of dysplasia and adenocarcinoma in this zone. In Oxford, no dysplasia has been found in a review of over 50 consecutive specimens, and clinical carcinoma arising in this area in patients with long-standing ulcerative colitis is virtually unheard of.

Based on the current balance of information, O'Connell and Williams advise against mucosectomy in the absence of rectal mucosal dysplasia. However, mucosectomy should be carried out if rectal dysplasia is diagnosed preoperatively or is found on subsequent histological examination of the surgical specimen which must, for this reason, be meticulous. Figure 10 1052 illustrates the possible levels of ileoanal anastomosis.

The question of whether the ileoanal anastomosis should be hand-sutured or stapled is less controversial. Minor defects in continence are not uncommon after manual endoanal anastomosis. Such deficiencies may be caused by retraction and dilatation injury to the anal sphincters at the time of operation. Certainly, impaired internal sphincter function has been noted. It was hoped that circular stapling devices would improve continence, as they produce minimal mechanical sphincter damage. However, objective evidence suggests that a significant fall in internal anal sphincter activity still occurs after stapled anastomosis, although the cause of this problem remains obscure. The only demonstrable improvement in pouch–anal function after stapling the anastomosis is a reduction in nocturnal evacuation. However, stapling devices do have the advantage of speed and can produce a great saving in time if used throughout the operation (Fig. 11) 1053. Four-loop W reservoirs cannot be constructed easily with staplers.

Until quite recently, all authors agreed that the reservoir and ileoanal anastomosis should be temporarily defunctioned with a loop ileostomy. This accepted practice has now been challenged. If the surgeon is familiar with the technique of pouch surgery, the operation is technically straightforward, and if the patient was not maintained on high-dose steroids preoperatively, the ileostomy can be omitted as, based on present evidence, this does not seem to be associated with additional morbidity. This avoids a subsequent surgical procedure which has its own attendant complications and which requires hospital admission.

If a defunctioning ileostomy is formed, a period of 6 to 8 weeks is allowed to elapse before closure. If an ileostomy is omitted, the pouch is intubated with a wide-bore balloon catheter for at least 7 days to divert the faecal stream. The perianal region should be protected by barrier cream and a careful record must be kept of fluid balance. Patients can be allowed home 10 to 14 days postoperatively if they are well and confident. Early review is sensible in order to monitor and advise on pouch function. Constipating medication will reduce stool frequency in the first few months.

Very few deaths have been reported after restorative proctocolectomy, no doubt as a result of careful patient selection. However, there is appreciable morbidity which declines as the surgeon becomes increasingly familiar with the techniques involved. Pelvic sepsis, with or without anastomotic breakdown, develops in 8 to 25 per cent of cases and can be difficult to manage as it cannot drain freely. Healing by fibrosis can impair the eventual functional result. Adhesion obstruction also occurs frequently and further laparotomy is needed in up to 15 per cent of patients. Strictures are seen in 10 per cent of patients, but usually respond to simple dilatation. Fluid and electrolyte loss from the defunctioning ileostomy can produce a dehydration syndrome which may be exacerbated by steroid withdrawal.

Following restorative proctocolectomy, most patients defecate spontaneously about five to seven times every 24 h, and will be able to defer defecation without urgency. Few patients suffer frank incontinence, but minor imperfections such as spotting or soiling occur in up to 33 per cent during the day and 56 per cent at night. Some 20 per cent of patients use codeine phosphate or loperamide hydrochloride.

The determinants of ileoanal pouch function have recently been reviewed. The best clinical results are associated with a large capacity, compliant pouch which empties completely and which sits above normally active anal sphincters. Total daily stool volume, postoperative pelvic sepsis, and pouchitis are also important. Some evidence suggests that three and four-loop pouches produce the lowest stool frequency because they are more capacious than those formed from two loops of ileum. However, this is disputed. It is recognized that the pouch functions in part by modifying terminal ileum motility, but the relationship between this, capacity, and compliance remains unexplained.

Outright failure which requires pouch excision occurs in only 5 per cent of patients and is caused by persistent pelvic sepsis, undiagnosed Crohn's disease, or unacceptable stool frequency.

The long-term health of the pouch mucosa is of more concern and has recently been the subject of extensive review. Early studies on Kock pouches showed chronic inflammation as well as villous atrophy, and similar changes are seen in most pelvic reservoirs. Colonic metaplasia has been noted on histological examination and is supported by studies using histochemical and immunological markers. The cause of these changes is not known, but probably involves bacteriological and immunological factors. The risk of future dysplasia and even frank malignancy cannot be ignored but, to date, no such occurrence has been reported in either a pelvic or a Kock pouch.

More acute inflammatory changes also occur and can produce pouchitis—a condition which is characterized by diarrhoea in the presence of endoscopic and histological features of acute inflammation (Fig. 12) 1054,1055 It has only been diagnosed unequivocally in patients who had ulcerative colitis, affecting between 10 and 20 per cent of those undergoing restorative proctocolectomy. Pouchitis is more common in patients who had extensive disease than in those whose colitis was left-sided or distal. The pathogenesis of pouchitis is not known and is almost certainly multifactorial. Contact between the pouch mucosa and ileal contents with stasis and bacterial overgrowth are essential features. The importance of genetic predisposition, the presence of specific bacterial strains, epithelial defects, and immunological abnormalities is uncertain, but these have all been implicated as causative agents. Outlet obstruction, pouch ischaemia, and Crohn's disease are important differential diagnoses which should be excluded.

In the absence of a controlled therapeutic trial, the treatment of pouchitis is empirical. Metronidazole is probably the most commonly used agent and may function as an immunosuppressive agent as well as an antibiotic. Enemas containing steroid or 5-aminosalicylic acid derivatives, oral sulphasalazine, and short courses of oral steroids can also be used. Most episodes of pouchitis will resolve when treated with one or a combination of these agents. Very rarely, pouchitis is unresponsive and must be treated by the formation of a defunctioning ileostomy or even pouch excision.

Almost all patients undergoing restorative proctocolectomy are satisfied with the outcome of their operation and over 90 per cent prefer the pouch to a permanent ileostomy because of increased self-confidence, cleanliness, sexual self-image, lack of interference with social and sports activities, and ease at work. Impaired sexual function is related to the proctocolectomy rather than the pouch.

No serious long-term nutritional sequelae have emerged after restorative proctocolectomy, but a mild microcytic anaemia associated with a low serum iron concentration is seen in up to 30 per cent of patients. Serum vitamin B&sub1;&sub2; levels may be marginally lowered but serum folate concentration is normal, and there is no abnormality of fat and fat-soluble vitamin absorption or of liver function.

Berry AR, De Campos R, Lee ECG. Perineal and pelvic morbidity following perimuscular excision of the rectum for inflammatory bowel disease. Br J Surg 1986; 73: 675–7.

de Silva HJ, Kettlewell M, Mortensen NJ, Jewell DP. Acute inflammation in ileal pouches (pouchitis). Eur J Gastroenterol Hepatol 1991; 3: 343–9.

Everett WG, Pollard SG. Restorative proctocolectomy without temporary ileostomy. Br J Surg 1990; 77: 621–2.

Fazio V. W. The role of anastomosis for ulcerative colitis. In: Allan RN, et al. eds. Inflammatory Bowel Disease. Edinburgh: Churchill Livingstone, 1990; 433.

Fonkalsrud EW. Endorectal ileoanal anastomosis with isoperistaltic ileal reservoir after colectomy and mucosal proctectomy. Ann Surg 1984; 199: 151–7.

Hulten L, Svaninger G. Facts about the continent iloesotomy. Dis Colon Rect 1984; 27: 553–7.

Jewell DP. Medical management of severe ulcerative colitis. Int J Colorectal Dis 1988; 3:186–9.

King DW, Lubowski DZ, Cook TA. Anal canal mucosa in restorative proctocolectomy for ulcerative colitis. Br J Surg 1989; 76: 970–2.

Lennard-Jones JE, Ritchie JK, Hilder W, Spicer CC. Assessment of severity of colitis: a preliminary study. Gut 1975; 16: 579–84.

Lennard-Jones JE, et al. Precancer and cancer in extensive ulcerative colitis: findings among 401 patients over 22 years. Gut 1990; 31: 800–6.

Nicholls RJ, Lubowski DZ. Restorative proctocolectomy: the four loop (W) reservoir. Br J Surg 1987; 74: 564–6.

O'Connell PR, Williams NS. Mucosectomy in restorative proctocolectomy. Br J Surg 1991; 78: 129–30.

Parks AG, Nicholls RJ. Proctocolectomy without ileostomy for ulcerative colitis. Br Med J 1978; ii: 85–8.

Shepherd NA. The pelvic ileal reservoir: apocalypse later? Br Med J 1990; ii: 886–7.

Utsonomiya T, et al. Total colectomy, mucosal proctocolectomy and ileo anal anastomosis. Dis Colon Rect 1980; 23: 459–66.

Watts JMcK, de Dombal FT, Goligher JC. Longterm complications and prognosis following major surgery for ulcerative colitis. Br J Surg 1966; 53: 1014–23.

Williams NS. Restorative proctocolectomy is the first choice elective surgical treatment for ulcerative colitis. Br J Surg 1989; 76: 1109–10.