Opinion on the nature of surgery for small bowel Crohn's disease has varied since the earliest operations described by Crohn and colleagues. At first wide resection was the treatment of choice, but high relapse rates led to a period when bypass procedures were popular. In due course these patients appeared to have a high incidence of septic and fistula complications from their bypassed segments and there was a return to radical surgery once more. Since Crohn's disease is a diffuse intestinal problem, and the majority of patients undergo surgery at some time during the course of their disease, there has been a move towards minimal or conservative surgery. The rate of recurrence increases with the length of follow-up, and there is a group of patients who require repeated resections at increasingly frequent intervals. It is this group who are at greatest risk of developing a surgically induced short-bowel syndrome and malnutrition. Recurrence of Crohn's disease is independent of the presence of microscopic disease at the resection margins and, as yet, there is no successful adjuvant medical therapy to decrease the incidence of recurrence following surgery.

Strictly speaking, there is no such thing as a surgical recurrence of Crohn's disease, but rather a recrudescence of a previously quiescent focus of Crohn's disease at a different site.

Since Crohn's disease cannot be cured by wide surgical excision, surgical management is reserved for the complications of Crohn's disease. The main indications are to:

(1)remove or relieve an area of stenosis causing persistent or recurrent obstructive symptoms;

(2)treat enterocutaneous or enterovesical fistulae;

(3)drain an intra-abdominal abscess—this role is now being achieved increasingly by percutaneous radiological techniques;

(4)control acute or chronic blood loss, though this is rare;

(5)treat free perforation, but this is exceedingly uncommon.

Since the risk of recrudescence of Crohn's disease is sufficient to require the treatment of approximately 6 per cent of cases a year, many patients will need a number of operations during a lifetime. It is important when operating to deal only with the specific complication which has been the indication for the procedure, and bowel should not be removed just because it is affected by Crohn's disease. Surgery should be as conservative as possible, but stenosed bowel, even in a defunctioned segment, should not be left untreated. The greatest care must be taken with surgical technique, ensuring well-vascularized, tension-free anastomoses and the avoidance of inadvertent damage to adjacent loops of the bowel.

Gastroduodenal disease is rare and almost always occurs in association with disease elsewhere. It can cause bleeding or stenosis, and disease in the duodenum most frequently produces clinical symptoms (Fig. 1) 965. At endoscopy it can be difficult to distinguish between peptic ulcer and Crohn's disease. Peptic ulcers almost always respond to H&sub2;-receptor antagonists, whereas Crohn's disease does not.

Resection of the duodenum is not feasible and therefore this is the one area of the small bowel in which bypass is the preferred treatment. At one time gastrojejunostomy was thought to be the treatment of choice. However, in some patients who have already had extensive small-bowel resection, gastric hypersecretion may result, and it has been customary to add a vagotomy to reduce the high risk of stomal ulceration. This vagotomy may, in turn, compound the patient's problem with diarrhoea.

A selective vagotomy has the theoretical advantage that diarrhoea is less likely to occur. With the advent of H&sub2;-receptor antagonists and omeprazole it is possible to control gastric secretion in Crohn's disease patients. Where possible, a strictureplasty should be carried out on pyloric or duodenal disease to save the patient the complications of a gastroenterostomy. There are preliminary reports of the use of balloon dilatation for duodenal stenosis caused by Crohn's disease, but it is too early to judge its success.

The most commonly affected site is the terminal ileum. However, there may be multiple foci throughout the small bowel (Fig. 2) 966 and here it is often difficult to tell exactly which lesion is responsible for the patient's symptoms. The most common indication for surgery in Crohn's disease of the small bowel is recurrent intestinal colic, provoked by solid food. Presentation is often insidious or subacute and may or may not be associated with diarrhoea. Weight loss is a major symptom. The patient may be anorexic, or be literally frightened to eat for fear of pain.

It is wise to give the patient full bowel preparation just in case there is any colon involvement. Routine antiembolus measures are taken, including stockings and low-dose subcutaneous heparin. Since there is good evidence that patients with Crohn's disease have organisms in lymphatics and lymph nodes outside the bowel wall, antibiotic prophylaxis with a cephalosporin and metronidazole is given for between 2 and 5 days. In this situation the antibiotics are probably being used for established infection rather than to prevent inoculation of a clean abdominal operation site. We usually give intravenous steroid cover (hydrocortisone 100 mg thrice daily) for the operation, introducing oral steroids when the patient is taking fluids by mouth. This is reduced to a dose of 10 mg, on which the patient is maintained until review. Although many surgeons would operate on a patient flat on the operating table, ileosigmoid or ileorectal fistulae in Crohn's disease are common and, in complicated cases, access to the rectum may be necessary. Therefore I usually operate on patients in the modified lithotomy Trendelenburg position used for colorectal surgery. A full laparotomy is carried out, carefully examining the whole length of the intestine for external macroscopic features of Crohn's disease.

This is best treated with a conservative ileocaecal resection to include a short margin of macroscopically normal gut on each side of the resection specimen (Fig. 3) 967. In patients with severe disease, however, the presence of minor macroscopic evidence of Crohn's disease at the anastomotic site is not important and the emphasis should be on preserving gut length. There has been some controversy over the use of various suture materials and the precise type of anastomosis used, and their possible effect on disease recrudescence. There is no definite evidence of an effect of either of these; I usually carry out a straight end-to-end ileocolic anastomosis, using absorbable suture material (3-0 Vicryl). It is helpful to mark the site of the anastomosis with metal clips so that future radiological investigations can localize new disease.

Actuarial methods are now widely used for the effective comparison of results of treatment between various centres because they correct for the varying lengths of follow-up between individual patients.

The cumulative operation rate for patients with distal ileal disease at 5 years from the time of diagnosis is 80 per cent. Cumulative reoperation rate after the first resection for distal ileal Crohn's disease at 5 years after the first operation is 25 per cent and at 10 years this rises to 35 per cent. If an ileocolic anastomosis is inspected carefully by colonoscopy, aphthous ulceration can be seen on the ileal side of the anastomosis as early as 6 months after surgery. There is no evidence that age at the time of initial diagnosis has any effect on recurrence rate.

Although recurrent disease is therefore very common, surgical treatment of distal ileal disease rapidly restores the patient with chronic persistent symptoms to good health. On average, patients have an operation once every 10 years, so they could expect to have 3 or 4 operations over a lifetime.

Lee pioneered the concept of treating Crohn's disease strictures by strictureplasty. He operated on patients with intestinal obstruction and malnutrition in whom excisional surgery was thought to be contraindicated because of diffuse small bowel involvement or a short-gut syndrome resulting from previous excisions. Lee proposed the concept of minimal surgery to overcome sites of obstruction. In some places this would be a mini-resection, and in others strictureplasty. In a subsequent report of 10 years' experience of strictureplasty for obstructive Crohn's disease, including the first patients operated on by Lee in Oxford, 24 patients had 86 strictureplasties at 30 separate operations. There were no deaths, fistulae, or wound infections related to the operation site and patients made positive weight gains in the postoperative period.

Ten further operations for obstructive symptoms were carried out in four patients 18 months after their previous strictureplasties. All but one of the previous strictureplasties in the additional operations were found to be widely patent and not the cause of the recurrent obstruction. Alexander Williams carried out 198 strictureplasties in 64 patients. In addition to using the technique in short stenoses, he has used a Finney pyloroplasty-like method for stenotic areas of more than 10 cm in length. Major complications developed in five patients who developed enterocutaneous fistulae, a clinical leak rate of 2.5 per cent. At 6 months 80 per cent had an excellent symptomatic response, being free of pain with improvement in general well-being.

Another study looked at the site-specific recurrence rate in 41 patients treated by strictureplasty, and reported that no recurrence occurred in those patients. A control group continued to be treated by small-bowel resection and there was no difference between the two groups. There is no evidence as yet that disease remaining at the strictureplasty site precipitates early recrudescence.

In multifocal disease it is essential to identify every stricture. Even though a stricture may not look very narrow from the outward appearance, this can be misleading (Fig. 4) 968. I usually choose the middle stricture and open it on the antimesenteric border, along the long axis of the small intestine. A 20 French gauge Foley balloon catheter is inflated with saline to give balloon diameters of 15, 20 and 25 mm (Fig. 5) 969. The intestine is then characterized from the duodenojejunal flexure to the caecum by passing the catheter along the gut lumen and then inflating the balloon, allowing it to impact at each stricture site as it is slowly withdrawn. This ensures that no strictures are undetected and prevents the dangerous scenario of obstruction downstream from a suture line. The diameter is recorded and its position marked with a suture. A stenosis of less than 20 mm diameter is usually treated by strictureplasty. The stricture site is held in stay sutures and a longitudinal incision made by diathermy (Fig. 6) 970. The bowel is opened out, and, if the appearance is suspicious, a biopsy should be taken to exclude the possibility of cancerous change, although we do not do this routinely. The stricture site is then closed transversely in the manner of a Heinecke-Mickulicz pyloroplasty (Fig. 7) 971, although some slightly longer strictures have been anastomosed in the manner of Finney. We use a continuous absorbable suture in a single layer. In our experience the median length of the strictureplasty has been 3 cm. Simultaneous intestinal resections have been undertaken for longer sections of disease, but we make these mini-resections as far as possible. Multiple strictureplasties can be carried out at any one sitting and the strictureplasty sites are usually marked with metal clips for future reference. A stapling technique has been described using a GIA stapler, but the resulting anastomosis looks more like a side-to-side than a strictureplasty.

There is some evidence that disease may heal, or at least not progress. Removing the obstructive element, even though the original triggering factors are likely to be present, may modify the progression of the disease. Once obstruction has occurred, a chain of events is precipitated resulting in mucosal leakiness, transmural fissuring, and, eventually, abscess and fistula formation.

Small-bowel Crohn's disease can be complicated by enteroenteric, enterovesical, or enterocutaneous fistulae. It is important to emphasize that postoperative fistulae may be due to an anastomotic breakdown, and not associated with any residual or recurrent active Crohn's disease.

The incidence of external fistulae is 17 to 22 per cent and of internal fistulae, 9 to 25 per cent. Preoperative abscess as a complication of Crohn's disease occurs in some 10 per cent of patients and, following resection, postoperative abscess is found in 14 per cent, of which a third occur in patients with a preoperative abscess. The organisms involved are usually enteric, including Escherichia coli, Bacteroides fragilis, and enterococci. Wound infection and intra-abdominal abscesses are common problems in patients with Crohn's disease.

These abscesses may be primary or secondary. They probably occur at the site of a transmural fissure and may be the first stage in the development of a fistula as the abscess tracks to another loop of gut or the abdominal wall. The common sites for an abscess are the pelvis, abdominal wall, and psoas sheath, or around the segment of affected bowel. Abscesses penetrating posteriorly into the psoas produce symptoms of fever, weight loss, and a flexion deformity of the hip. They can point at the groin or buttock (Fig. 8) 972. Drainage of the abscesses invariably results in a fistula. Secondary abscesses are more common and follow surgical procedures, often in malnourished septic patients. They may also occur slowly and present many months later.

Most external fistulae pass through the site of a previous surgical incision, and a spontaneous fistula through the abdominal wall is rare.

An external fistula following appendicectomy usually arises from a segment of ileum actively involved with Crohn's disease rather than the appendix stump. If Crohn's disease is found at a laparotomy for presumed ‘appendicitis’ the appendix should be removed. This will ensure that future attacks of pain can be managed without confusion. Where there is extensive caecal involvement, an ileocaecal resection is preferred.

This is surprisingly rare from small-bowel Crohn's disease. Sudden onset of pain in a patient with an exacerbation of Crohn's disease treated with steroids should be regarded as a possible perforation until proven otherwise.

Fistula and abscess go hand in hand, and fistula cannot be treated successfully without the elimination of any concurrent abscess.

Fever, weight loss, diarrhoea, and abdominal pain are the usual symptoms of an internal fistula, although some enterenteric fistulas can be entirely asymptomatic. A discharge sinus is the hallmark of an external fistula, although only about 50 per cent may discharge faeces, the rest producing gas or pus (Fig. 9) 973. There may be a mass or area of tenderness to suggest an abscess.

Nutritional and biochemical assessment should include haemoglobin, serum albumin, body weight, lymphocyte count, and liver function tests. Barium contrast studies, especially a small-bowel enema, are best for demonstrating areas of small-bowel disease, a fistula, and even an abscess (Fig. 10) 974. Fistulography provides important information about the complexity of a track. Ultrasound and computerized tomography (CT) scans are the best methods for demonstrating an abscess, and CT-guided percutaneous drainage of abscesses is becoming increasingly successful (Fig. 11) 975. Endoscopy will help to define the extent of mucosal disease and the presence and nature of a stricture, but rarely identifies a tiny fistula track, for example at an ileocolic anastomosis.

Many internal fistulae are asymptomatic and often a chance finding at laparotomy. Any area of active small-bowel Crohn's disease adherent to an adjacent organ should be regarded as a fistula until proven otherwise. Disease is usually on the small-bowel side of an enterocolic fistula. This can be pinched off and the hole in the colon closed. Formal resection is seldom necessary. Enterocutaneous fistulae usually arise at an anastomosis or in a segment of non-involved bowel damaged at surgery. The management of these can be very challenging and requires a team approach. Within the first 24 hours, fluid and electrolyte problems must be corrected and the skin protected with stomadhesive and a wound-management appliance. Intravenous nutrition should next be established through a dedicated central venous line. If there is an abscess it must be drained, preferably by a percutaneous technique. Broad-spectrum antibiotics are given for a defined period and are not a substitute for adequate drainage. Then the fistula anatomy can be defined. It is reasonable to try a period of conservative management, and some 60 to 70 per cent of postanastomotic fistulae will heal spontaneously.

Spontaneous enterocutaneous fistulae rarely heal. Although medical trials with azathioprine, 6-mercaptopurine, and cyclosporin have been reported, surgical exploration and resection of the fistula with anastomosis is usually necessary.

Operative treatment can be successful in experienced hands, provided there is not extensive sepsis or profound hypoalbuminaemia, and distal obstruction must be carefully excluded. If the patient is grossly hypoalbuminaemic or septic, the procedure of choice is resection and exteriorization of both bowel ends or, less optimally, a proximal bypass leaving the disease and fistula in place. When the patient is well and the sepsis drained, a second procedure to restore continuity is permissible. It may be necessary to wait for many weeks until a patient is fit enough for a further operation. Where the patient is well, resection and primary anastomosis are indicated (Fig. 12) 976.

An abscess pointing at the skin surface should be drained but is usually followed by a fistula. The introduction of ultrasound and CT have not only improved diagnosis and localization of abscesses, but percutaneous drainage techniques have largely replaced open laparotomy with all its hazards in a sick patient. The drainage of an intra-abdominal abscess is an absolute priority and undrained sepsis can be lethal. Antibiotics should be given in short effective courses, taking into account bacteriological reports on any swabs taken from an abscess.

Repeated urinary tract infections, turbid urine, and pneumaturia suggest an enterovesical fistula. This is not an absolute indication for surgery but they seldom heal and eventually come to surgery. The small bowel can usually be ‘pinched’ off the bladder and resected (Figs. 13, 14) 977,978. The bladder itself only requires curettage and careful closure. A catheter is left in the bladder for at least a week.

Alexander Williams J, Irving M. Intestinal fistulas. Bristol: Wright, 1982.

Alexander Williams J. The technique of intestinal strictureplasty. Int J Colorect Dis 1986; 1: 54–7.

Allan RN, Keighley MRB, Alexander Williams J, Hawkins C., eds. Inflammatory bowel diseases. 2nd edn. Edinburgh: Churchill Livingstone, 1990.

Coutsoftides T, Fazio VW. Small intestine cutaneous fistulas. Surg Gynecol Obstet 1979; 149: 333–6.

Dehn TCB, Kettlewell MGW, Mortensen NJMcC, Lee ECG, Jewell DP. Ten-years experience of strictureplasty for obstructive Crohns disease. Br J Surg 1989; 75: 339–41.

Fazio W, Galandiuk S. Strictureplasty in diffuse Crohns jejun ileitis. Dis Colon Rect 1985; 28: 512–18.

Greenstein AJ, et al. Perforating and non perforating indications for repeated operations in Crohn's disease. Gut 1988; 29: 588–92.

Keighley MRB, Eastwood D, Ambrose NS, Allan RN, Burdon DW. Incidence and microbiology of abdominal and pelvic abscess in Crohn's disease. Gastroenterology 1982; 83: 1271–5.

Lee ECG, Papionnu N. Minimal Surgery for chronic obstruction in patients with extensive or universal Crohn's disease. Ann R Coll Surg Eng 1982; 64: 229–33.

Sayfan J, Wilson DAL, Allan A, Andrews H, Alexander Williams J. Recurrence after strictureplasty or resection for Crohn's disease. Br J Surg 1989; 76: 335–8.

Williams JG, Wong WD, Rothenberger DA, Goldberg SM. Recurrence of Crohn's disease after resection. Br J Surg 1991; 78: 10–19.