Br. J. Surg. Vol. 68 (1981) 874-878 Printed in Great Britain

The value of preserving the anal sphincter in operations for ulcerative colitis and polyposis: a review of 22 mucosal proctectomies D A V I D J O H N S T O N , N O R M A N S. WILLIAMS, D A V I D E. N E A L A N D A N T H O N Y T. R . A X O N *

SUMMARY Ulcerative colitis and polyposis are both diseases of the mucosa. They can be cured by colectomy combined with selective mucosal proctectomy, without sacrifice of the anal sphincters or damage to bladder or sexual function. Terminal ileum, either as a straight tube or in the form of a pouch, is drawn down through the denuded tube of anorectal muscle and anastomosed to the mid-anal canal. A temporary defunctioning ileostomy is always used. Caecum has also been used as a neorectum after mucosal proctectomy. but without long term success.

Twenty-two patients, 20 with ulcerative colitis and 2 with polyposis, have been treated by mucosal proc- tectomy in the past 4 years, with no mortality. The caeco-anal procedure proved a failure because of re- currence of colitis, although the early functional results were good. After ileo-anal anastomosis, continence was perfect by day, but 2 patients had occasional lapses at night. The disadvantage of straight ileo-anal anastomosis is frequency of bowel action (6-9 times a day) . even with codeine medication, although the patients considered the operation a success. Some form of pelvic reservoir is therefore desirable and our early experience with the triplicated ileal pouch is encouraging.

ULCERATIVE COLITIS is primarily a disease of the mucosa. The muscular wall of the bowel is involved only late in the course of the disease and even then the muscles of the lower rectum, pelvic floor and anal canal usually remain in good working order. Surgical re- moval of these muscles in the course of procto- colectomy leads, of course, to loss of continence for faeces and the patient has to have a permanent ileostomy .

Loss of the anus and rectum, and so of continence, has hitherto been accepted, albeit reluctantly, because it was believed that when the rectal mucosa had been removed, the afferent side of the reflex arc subserving continence would be lost, and so the patient would be incontinent, even if his anal sphincters were preserved. Today, thanks largely to the work of Parks and his colleagues (I), this idea is known to be erroneous, because patients who have lost all their rectal mucosa, or even their entire rectum, are found to be continent, provided that the puborectalis component of the pelvic floor and the mucosa and muscles of the anal canal remain intact. Thus, there is now firm evidence that reflexes subserving continence persist after removal of the rectal mucosa and it would appear that there are sensors outside the rectum that can influence the behaviour of the anal sphincter ( I , 2). This improved understanding of anorectal physiology, allied to certain technical advances, is currently transforming the surgery of rectal carcinoma. Applied ,to the surgery of ulcerative colitis, it undoubtedly means that many

patients will no longer have to endure the embarrass- ment of an ileostomy, nor young men run the risk of impaired sexual function.

In this paper we describe our experience with 22 selective mucosal proctectomies (3) over the past 4 years. In 5 patients, the procedure was used merely to deal with a persisting inflamed rectal stump after previous emergency colectomy, but in 17 alimentary continuity was restored by means of a caeco-anal or ileo-anal anastomosis, which was performed endo- anally without eversion or mobilization of the anal canal and anorectal ring (4).

Patients and methods There were 22 patients, 13 men and 9 women, aged from 17 to 54 years. Most were youn adults in their 20s and 30s, and only 2 were over 40 years of age. Twenty had ulcerative colitis, 1 had familial polyposis coli and 1 man aged 52 had multiple colonic polyps, one of which later proved to be malignant. Further details of the 17 patients in whom alimentary continuity was either restored or its restoration was planned are shown in Table I. In the remaining 5 patients, mucosal protectomy was used instead of complete (or intersphincteric) rectal excision to deal with the problem of a persisting inflamed rectal stump after previous emergency colectomy for ulcerative colitis: these patients were happy with their ileostomy and the objective was merely to remove the rectal disease with minimal risk to bladder and sexual function and without a perineal wound. In them, a mucosal roctectom was performed up to mid-rectal level, above wiich the fu i thickness of the rectal wall was excised by an abdominal approach.

At first, only patients who were anxious to avoid a per- manent stoma, or who already had a stoma and wished to be rid of it, were accepted for mucosal proctectomy, but for the past 2 years we have used mucosal proctectorny as the procedure of choice. Thus our indications now for mucosal proctectomy with ileo-anal anastomosis are failure of medical treatment for ulcerative colitis (in which case a colectomy is performed at the same time) or persisting active disease in the rectal stump after previous colectomy. Since the operation !akes from 6 to 1 I h to perform, its use has been confined to fit’ young people. Many of them, of course., have in fact been

somewhat debilitated and are on steroid therapy. Active inflammation in the rectum is not a contraindication to the use of mucosal proctectomy. The anal sphincters, however, must not be patulous or damaged by sepsis, and patients with Crohn’s disease are considered unsuitable for this procedure. Twelve of the 20 patients with ulcerative colitis had previously undergone subtotal colectomy; in the other 8, colectomy was performed at the same time as mucosal proctectomy.

Operative technique Before operation, the patient’s general condition is improved as far as possible, and for 4 days the colon is cleansed by means of oral aperients and gentle enemas, steroid enemas

University Department of Surgery and Gastroenterology Unit, The General Infirmary, Leeds LSI 3EX.

Ulcerative colitis and polyposis 875

Table 1: DETAILS OF PATIENTS WITH MUCOSAL PROCTECTOMY AND ANASTOMOSIS FOIIOW-UP No. of after ileo- bowel

Patient Age stomy closed Current actions no. Sex (yr) Diagnosis (months) medication Continence in 24 h Comment

Group I: Ileo-unul unustomosis I M 26 uc 40 2 M 17 uc 35

3 M 31 uc 32

4 M 26 uc 30

5 F 23 P 6

6 M 18 UC Not closed I M 9 UC Not closed 8 M 36 UC Not closed

Group I!: Ileo-anal unustomosis with pelvic ileul pouch 9* M 52 P 20

lo* F 24 uc 2 I 1 F 28 UC Not closed

Group 111: Cueco-unul anastomosis 12 M 39 uc 17

13 F 30 uc 12

14 M 54 uc 16

IS M 36 uc 15

16 F 28 uc 15

17 F 24 uc 12

Codeine Perfect Codeine Perfect by day

occasional soil- ing at night

Codeine Perfect

Codeine

Codeine Perfect by day some urgency, occasional minor incon- tinence at night

-

- Perfect

- Perfect -

Prednisone Perfect now (soiled at night at first)

cult to assess (neurotic)

urgency

- Variable: diffi-

- Perfect, but much

- Perfect

- Perfect

- Perfect

6-8 4- 10 (7)

6-8

6-8

7-10

4

3

5

10-20

3-5

3-6

3-4

Good result Good result

Good result, anal skin sore at times

Failure due to 'cuff abscess and stricture

Good result

Failure due to urgency

Excellent result so far Good progress, but

and discomfort at anus: ileostomy re-established

ileostomy not yet closed

Recurrence of colitis: well on steroids

Recurrence of colitis: re-op.

Early failure due to severe diarrhoea

Recurrence of colitis: re-op.

Recurrence of colitis: re-op.

Recurrence of colitis: re-OD.

UC, Ulcerative colitis; P, polyposis. *Patient 9 had a duplicated type of pouch and patient 10 a triplicated (Parks) pouch.

also being used in an attempt to control inflammation in the rectum. Gentamicin (80mg) and metronidazole (500mg) are given intravenously I h before the operation, but no oral antibiotics are used. These antibiotics are given again &hourly for 18 h. The details of the operative procedure have been described recently by Martin (5) and Parks et al. (6) and only an outline need be given here. Phuw I: Muco.sul procfectomy: The bladder is catheterized and the patient placed in the lithotomy position. The rectum should already be clean, but any content that remains is aspirated. Neomycin solution is instilled and aspirated, and gauze is inserted high up to prevent faeces reaching the field of operation. The anus is gently dilated, Pyrah's rectractors inserted and the rectal mucwa raised off the internal sphincter muscle by means of adrenaline-in-saline solution (1 : 200 000 Fig. I). Starting just above'lhe dentate line. the mucosa of the upper half of the anal canal and the lower 5-8cm of rectum are then removed. as a cylinder, which is gradually fitted over the operator's index finger to facilitate traction as the dissection proceeds upwards. A good light is essential. NO attempt is made to evert the anus or rectum, but once the initial difficulties in the anal canal have been surmounted. the mucosal cylinder itself prolapses through the anus, and so

access is surprisingly good. The mucosa of the upper anal canal is often friable, but once the puborectalis sling is passed, the rectal mucosa is tougher and withstands strong tra$on. Meticulous haemostasis is essential because a cuff haematoma is likely to lead to abscess formation, fibrosis, stricture and eventual failure.

Difficult and tedious as this dissection is, it is even more difficult from above, so no effort is spared to carry the dissection as high as possible from below, a point about 8-IOcm from the anal margin being the objective, which is indeed achievable in most cases. This takes 2-3 h. Then the rectum is packed with gauze, and the patient placed in the Lloyd-Davies position. Phuse It The abdomen is opened by a midline incision and the entire colon mobilized, but its arteries are not divided. The distal sigmoid and upper rectum are most carefully mobilized, the operator dissecting right on the muscular coat to avoid damage to the hypogastric nerves at the pelvic brim and to the autonomic nerves to bladder and sexual organs more distally. Numerous small blood vessels have to be divided. A circumferential cut through the muscular layers of the rectum is then made. at or just above the peritoneal reflection, to enter the submucosal plane, which is then followed downwards to

876 D. Johnston et al.

... ‘. ,’,- . >

Fig. 1. The mucow of.the upper anal canal and rectum is raised off the muscle by infiltration of adrenaline-in-saline into the submucosd.

meet the previous plane of dissection (Fig. 2). The mucosal cylinder is then extracted. The greatest patience must be exercised at this stage to secure perfect haemostasis. The entire colon and the rectal mucosa are then removed en bloc.

1. Ileo-anal anmiornosis. In 8 patients, distal ileum was then drawn down through the rectal muscular tube and sutured to the anus (Fig. 3).

2. As frequency of bowel action proved to be a problem after ‘straight’ ileo-anal anastomosis, a caeco-anal or ascend- ing colon-anal anastomosis was used in 6 patients with long standing distal colitis in whom radiological and colonoscopic assessment before operation had suggested that the proximal colon was normal. About 12-15cm of caecum and ascending colon were preserved in this procedure.

3. When the cam-anal operation in its turn failed because of recurrent ulcerative colitis, amileal pouch was fashioned in 3 patients and connected to the anus by a short length (4cm) of ileum. This was a triplicated pouch (6) in 2 patients and a duplicated pouch in 1 patient.

It is essential to mobilize the ileal mesentery fully, so that the ileum will reach the anus without tension. The severe haemorrhage that was encountered in phase 111 of the opera- tion in one patient was probably due to ‘blind’ traction on a tight ileal mesentery by the perineal operator, leading to rupture of blood vessels in the mesentery.

When the ileum has been delivered through the anus, without tension or rotational deformity, and with its mesentery taking the shortest route downwards, that is, anteriorly, the margins of the transected rectal wall are tacked lightly to the seromusclar layer of the ileum with catgut sutures (Fig. 3) and a suction drain is placed in the cuff between the rectum and ileum and brought out in the left iliac fossa.

A diverting loop or terminal ileostomy is then fashioned, and the abdomen closed (Fig. 4). To ical neomycin solution is used to wash the pelvic cavity and t!e wound edges at regular intervals during phase 11. Ampicillin powder ( I g) is placed in the subcutaneous tissues of the abdominal wound, and, if the patient is fat, a thin corrugated drain is also placed in the wound and removed 24 h later. Phase III: The patient is again in the lithotomy position. A peranal ileo-anal anastomosis is then made with interrupted Dexon 2/0 sutures, as Parks (6) describes (Fig. 3), between mucosa and muscle of the anal canal and all layers of the ileal wall. (The end of the ileum was closed with staples or sutures in the abdomen to prevent contamination and is opened again at this stage.) The ‘neorectum’ is then packed with gauze to

Fig. 2. The dissection of the rectal mucosal cylinder proceeds in the submucosal plane. Most of the dissection is done from below. Note that the dissection should start just above the dentate line in the anal canal

Fig. 3. Ileum is drawn down through the anorectal muscular tube and anastomosed to the anal mucosa and internal sphincter muscle, just above the dentate line. It is essential that blood should not collect in the cuff between the rectum and the ileum.

bring the walls of the concentric cylinders into apposition and discourage haematoma formations. No drain is brought out in the perineum and we do not decompress the pouch. The defunctioning ileostomy is closed 6-12 weeks later.

The operation takes from 6 to I I h to perform, and is thus demanding both to staff and to the patient. Particular care must therefore be taken to maintain the patient’s body tem erature and protect bony pressure points and nerves such as t i e ulnar and the lateral popliteal.

Results Operative mortality was nil. Postoperative com- plications occurred in 1 1 of the 22 patients (50 per cent,

Ulcerative colitis and polyposis 877

6-40 months. They are perfectly continent during the day, but 2 of them occasionally experience a minor degree of incontinence at night. This does not trouble them. All 4 patients feel that the operation has been a success. However, all 4 are taking regular codeine medication and their stool frequency ranges from 6 to 10 in 24 h, with mean of 7 or 8. Soon after operation, stool frequency was 8- 16 times per day, and the peri- anal skin tended to become excoriated, but some degree of adaptation has obviously taken place. These 4 patients are all working and none reports any disturbance of bladder or sexual function. Ileo-anal anastomosis and pelvic reservoir: A man aged 52 with multiple polyps had a pouch constructed from duplicated ileum. He was perfectly continent, but had to defecate 7-10 times a day, experienced a good deal of urgency and discomfort near the anus and after about a year asked to have his ileostomy restored. This was done and at the same time the pouch was removed. A 24-year-old woman had a triplicated pouch (6). She is completely continent and defecates (without having to intubate the pouch) 4 times a day. Follow-up is only 2 months, however. She is delighted with the result. The third patient also has a Parks type of pouch (6) and is well, but the ileostomy has not yet been closed.

Fig. 4. A defunctioning ileostomy is essential to guard against sepsis in the pelvis. It may be terminal, as here, or a loop.

Table II: POSTOPERATIVE COMPLICATIONS Actual Possible

Haemorrhage. sepsis and stricture 1 22 Wound infection 2 22 Pelvic sepsis and stricture 2* 22 Pressure sore on occiput I 22 Nerve palsy (lateral popliteal, I 22

temporary)

I slight)

After closure of ileostomy 3 13

spontaneously)

Stricture at anastomosist (2 severe, 9 17

Small bowel obstruction After first operation 1 22

Reoperation 2 22 Small bowel fistula (closed I 22

Caeco-anal anastomosis: There was one early failure, out of 6 patients, in the only patient in whom the cut end of the caecum would not reach the anus. The side of the caecum had to be anastomosed to the anus. He experienced persistent severe diarrhoea when the ileostomy was closed and had to have the ileostomy restored.

In the other 5 patients, the operation seemed very successful at first because they were fully continent and stool frequency was only 3-6 times (mean 4 times) in 24 h, presumably because of the presence of an intact ileo-caecal valve and the absorptive capacity of the caecum. However, after a relatively short period that ranged from 3 to 15 months, active colitis re-appeared in the caecum, leading to troublesome diarrhoea, and 4 of the 5 patients had to have the ileostomy restored.

~~~~ ~ ~

* Includes 1 haemorrhage, sepsis and stricture. ?Seven of the strictures were soft and easily dilated.

Table 11). The most serious complications were the two ‘cuff abscesses, which caused stricturing, and failure of the operation. Small bowel obstruction occurred in 4 patients. Anastomotic strictures in the anal canal were common, but usually responded readily to dilatation. Mucosal proctectomy alone: These 5 patients recovered rapidly and without complications. Follow-up ranges from 3 to 16 months. Three patients have no symptoms, but two still have a slight irritating dis- charge from the anus, for which one patient needs to use a barrier cream. Ileo-anal anastomosis without a reservoir: Two of the 8 patients developed ‘cuff abcesses between the rectal tube and the ileum and this caused the operation to fail in one of them. Though continent after the ileostomy was closed, this patient suffered severe diarrhoea, probably because of stricture formation, and he had to have his ileostomy restored. He is now in good health. The other patient has not yet had his ileostomy closed. He, too, has a stricture, which is being dilated. Two other patients have yet to have the ileostomy closed. The remaining 4 patients have been followed-up for

Discussion Colectomy combined with mucosal proctectomy is clearly a demanding procedure, both for the surgeon and for the patient. Do the results justify the extra effort? While the results reported here are in many ways unimpressive because in about half the patients the operation failed, nevertheless some important lessons were learned and the results are improving. In many patients, too, the period of follow-up is short; thus, further adaptation may take place in the small intestine and improved anorectal function may yet result. Though tedious, the operation is safe. All the inflamed mucosa has been removed, as in conventional proctocolectomy, and all the patients, even the ‘failures’ have been restored to good health and vigour.

Apart from the safety of the procedure, which is in large measure attributable to the defunctioning ileostomy and the use of prophylactic antibiotics, the most exiciting finding has been that patients who have had their entire rectal mucosa removed and replaced with ileum are indeed completely continent, at least during waking hours, provided that there has been no technical mishap such as a ‘cuff abscess in the pelvis.

878 D. Johmton et al.

Even during sleep, most patients are completely con- tinent, and only two have occasional soiling, which is not enough to be socially unacceptable. Thus, our experience is in agreement with that of Martinet al. and Parks et al. (5,6).

These, then, are important conclusions: if the results from Cincinatti, London and Leeds are combined, the mortality in 60 patients was nil, the disease was completely eradicated in all, yet all patients who did not develop pelvic sepsis were continent. Continence was complete during the day, but a few patients experienced minor leakage at night.

To this common experience we would add our finding that attempts to preserve the ileocaecal valve and caecum seemed doomed to failure, because how- ever long the colitic process has been confined to the left side of the colon when the caecum is utilized as a ‘neorectum’, an early flare-up of colitis in it must be expected. This is an intriguing and disappointing out- come which we are at a loss to explain.

Granted technical perfection, then, in the performance of the operation, we can now virtually guarantee to a colitic patient that he can be cured of the disease, yet be continent and avoid a stoma. Continence, however, is not enough, because if the patient has to defecate 10 times a day, he might well prefer to have an ileostomy. The question that remains, therefore, is how to secure for the patient as good a quality of life as possible, by the construction of a ‘neorectum’ with good reservoir capacity, so that urgency and frequency of defecation may be avoided. The solutions to this problem may vary, depending on the age of the patient. Thus, in the Children’s Hospital at Cincinatti, Martin et al. ( 5 ) noted that considerable adaptation took place for 6- 12 months after mucosal proctectomy. At first, their 13 young patients had frequent watery stools, often associated with incontinence at night, but after 6- 12 months they were completely continent and passed 2-8 semi-formed stools daily. At St Mark’s Hospital in London, Parks and his colleagues found that after construction of a triplicated ileal pouch, mean stool frequency was only 3.8 per 24 h, and all patients were continent by day (6). Thus, the Parks’ pouch seems at present to provide the best solution for adult patients. Its main disadvantage

seems to be that half the patients cannot defecate spontaneously, but have to evacuate the pouch by means of a catheter, whereas our patients and those of Martin et al. ( S ) , with straight ileo-anal anastomosis, could defecate spontaneously. However, the St Mark‘s group have recently obtained better results in this respect by making the pouch larger and drawing it down nearer to the anal canal. Our single patient with this type of pouch has a large one which she can evacuate without the need for a catheter.

In conclusion, despite the difficulties which we and our patients have experienced, we think that the results obtained here and elsewhere are very encouraging. It seems to us preferable to use a natural sphincter to control the exit to the ileum, rather than try to construct an artificial valve (7). It is also better to have no stoma at all than to have a stoma, even one that is flush with the abdominal wall. This operation follows the principle of ‘primum non nocere’, whereby only diseased tissue is removed, all normal structures and functions being preserved. It follows that until this operation has been fully evaluated, the anus and rectum of colitic patients should not be removed unless absolutely necessary.

References I . LANE R. H. s. and PARKS A. G.: Function of the anal

sphincters following colo-anal anastomosis. Br. J . Surg. 1977; 64: 596-9.

2. WILLIAMS N. s., PRICE R. and JOHNSTON D.: The long term effect of sphincter preserving operations for rectal carcinoma on function of the anal sphincter in man. Er. J . Surg. 1980; 67 203-8.

3. RAVITCH M. M. and SABISTON D. c.: Anal ileostomy with preservation of the sphincter. Surg. Gynecol. Obstet. 1947;

4. PARKS A. G.: Transanal technique in low rectal anastomosis. Proc. R. SOC. Med. 1972; 65: 975-6.

5. MARTIN L. w., LECOULTRE c . and SCHUBERT w. K.: Total colectomy and mucosal proctectomy with preservation of continence in ulcerative colitis. Ann. Surg. 1977; 186

6. PARKS A. G., NICHOLLS R. J . and BELLIVEAU P.: Proctocolectomy with ileal reservoir and anal anastomosis. Br. J . Surg. 1980; 67: 533-8.

7. KOCK N. G., DARLE N., HULTEN L. et al.: Ileostomy. Curr. Prohl. Surg. 1977; 14 1-52.

8 4 1095-9.

477-80.