Br. J. Surg. Vol. 69 (1982) 414-416 Printed in Great Britain

Neoplasia and u reterosigmoidostomy: a col onoscopy su rvey

Patients who have undergone implantation of ureters into the sigmoid colon (ureterosigmoidostomy) are known to be at high risk of developing cancer of the colon many years later. The operation is often performed in infancy for congenital abnormalities of the bladder, thus creating a long term surveillance problem.

Six of 34 patients (1 7.6 per cent) who had undergone uretero- sigmoidostomy were found at a screening jibresigmoidoscopy to have adenomas of the left colon or severe dysplasia of the stoma, a mean interval of 22 years after their urinary diversion. Four other patients had previously had a sigmoid adenoma or cancer; thus, 29 per cent overall had developed colonic neoplasms, almost all closely related to their stomas. Another patient who, by error, was excluded from the surveillance, died of metastatic cancer of the sigmoid colon within the study period.

Regular surveillance by jibresigmoidoscopy of patients who have had urinarv diversion affords the omortunitv to detect and remove

M. STEWART, F. A. MACRAE AND C. B. WILLIAMS

.r.r 1 L

~ h " ~ ~ ~ ~ ~ , ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ n s of England. 35/43 Lincolns Inn Field. London

Mark's Hospital, City Road, London ECIV ZPS.

po ten tiall; malignant sigmoid lesions. In addition to its clin ical importance, it allows a closer study of this unique model of colon carcinogenesis. WC2A 3PN. F. A. Maude and C. B. Williams. St

Estimates of the risk of malignancy at the ureterocolic anasto- mosis in patients who have had ureterosigmoidostomies (Table I ) have varied between 18 1 and 7000 times that of the normal population (1-3). Some of these statistics are misleading in that a prevalence of carcinoma in the groups studied has been compared with an annual incidence in the normal population. In a review of the St Peter's Hospitals' cases (unpublished), we compared the incidence of colon cancer in our patients with a cumulative incidence in the general population. There was an excess risk of 80- 100-fold.

Updating our own experience (6) and review of the literature reveals that most affected patients underwent diversion for benign disease, predominantly ectopia vesicae. Adenomas (22.5 per cent) and adenocarcinornas (54 per cent) are by far the commonest tumours described. The mean age of tumour diagnosis is 34 years (with a range of 13-71 years) and the mean latent interval between diversion and diagnosis of tumour is 24 years (ranging from 7 to 50 years). The tumour is

site specific, occurring at, or close to, the ureteric implant. Rediversion confers no immunity: urine can be in contact with contiguous colon for only a matter of months, but nevertheless malignancy develops many years later at the site of the unexcised ureteric bud. Tumour has been described with various forms of ureteric anastomoses, including Maydl's trigonosigmoidostomy (7), Coffey (8) and Leadbetter (9) im- plantation techniques. The common presenting symptoms are rectal bleeding (51 per cent), loin pain (31 per cent) and abdominal pain (20 per cent).

In this study we have attempted to establish the prevalence of malignant and premalignant changes at the ureteric anasto- mosis in patients with longstanding rectal bladders.

Patients and methods Thirty-four consecutive patients attending for routine long term follow-up after ureterocolic diversion were included in the study. A single exception was a patient who, in error, was not referred for the

Table I: ESTIMATED RISK OF COLON CANCER IN PATIENTS WITH URETEROCOLIC ANASTOMOSIS IN PATIENTS DIVERTED FOR BENIGN DISEASE

Incidence No. ofpaiients No . of cases per 100000

Author followed up of carcinoma population Risk

Urdaneta et al. (1)

Parsons et al. (2)

Eraklis and Folkman (3)

St Peter's cases (unpublished)

23 3 13 000 550 x normal population

29 2 6 900 280 x normal population

82 4 4900 181 xwhole population

7000 x population aged < 25 yr

population * 200 7 3 500 80-100 x normal

The number of cases of colorectal cancer in the first three series gives a cumulative incidence of between 4.9 and 13 per cent. When all are added together the cumulative, incidence is 9 out of 134 or 6.7 per cent. The mean age at tumour diagnosis in the St Peter's cases is 46 years. The cumulative incidence of colon cancer in Birmingham is 0.1 I per cent in persons living up to 46 (4). of which approximately one-third occur in the rectosigmoid ( 5 ) . Thus, the cumulative incidence of rectosigmoid cancer is 0.04 per cent, giving an excess risk, in the St Peter's cases, of 80-100-fold.

Neoplasia and ureterosigmoidostomy

study. In most cases the patients had originally undergone diversions for benign disease. commonly ectopia vesicae or epispadias (62 per cent). Tuberculosis, incontinence and urethral stricture were other antecedents for diversion. In 4 cases ( I 1.8 per cent) diversion was carried out for malignancy. Conversion to ileal or colonic loop had, at some stage, been performed in 12 patients (35 per cent), and 10 patients (29 per cent) had undergone nephro-ureterectomy.

Fibresigmoidoscopy was arranged as a single outpatient visit. No prior bowel preparation was used: but two phosphate enemas were administered on arrival. The patients were not sedated and no recovery period was necessary following the procedure. Methylene blue (20 ml I per cent solution), 5 mg of frusemide and 40 mg of the antispasmodic hyoscine N-butyl bromide (Buscopan) were injected intravenously at the beginning of the procedure to help identify the ureteric stomas. The more flexible paediatric colonoscope was used as the sigmoid colon was frequently fixed by postoperative adhesions which made manipulation more difficult.

Stomas were usually situated 20--40cm proximal to the anal verge and 5-15’cm apart. The appearances varied from flat slit-like orifices 1-2 mm in size to polypoid lesions up to 1 cm in diameter. Where possible, biopsies were taken from each stoma and the surrounding mucosa. Snare polypectomy of adjacent lesions was carried out only if the ureteric orifices could be positively identified by the passage of blue dye. In our earlier experience, failure to take this precaution led to ureteric stomatectomy’ and subsequent removal of a poorly function- ing kidney (10).

Results Among the 34 patients, there were 55 available ureteric orifices. Thirty-five were visualized, a success rate of 64 per cent in terms of identification of individual ureteric orifices.

On the basis of histological information derived from biopsy, three subgroups of patients were defined: (a) no pathology group (19 patients), (b) minimal pathology group (mild inflam- matory changes, metaplastic polyps) (9 patients) and ( c ) adenoma dysplasia group (6 patients). The overall mean age at the time of investigation was 41 years, although patients in the adenoma group tended to be older (Table II). The mean number of years since diversion (and thus of follow-up) was 5 years longer in the adenoma/dysplasia group than in the group with no pathology.

41 5

Table 11: LATENT INTERVAL SINCE DIVERSION

AGE A T TIME OF INVESTIGATION AND

Interval since Age ( y r ) diversion (yr )

Mean Range Mean Range ~~~

No pathology group 36 19-77 22 9-38 Minimal uatholoev erouu 40 13-78 20 7-31 Adenomi dysplaza i r o i p 56 4 3 4 6 21 6-46 All patients 41 13-17 22 6-46

i --

Fig. 1. pective colonoscopic study. ( a , Same patient.)

The distribution of the six adenomas detected in the pros-

Only two patients had symptoms pertaining to the gastro- intestinal tract. One patient, with abdominal pain, was found to have a duodenal ulcer but no colonic pathology. The other patient, who admitted to intermittent rectal bleeding, was found to have a tubulovillous adenoma.

Fig. 1 shows the distribution of adenomas in the colon. The six adenomas occurred in 5 patients and were all macro- scopically visible. Two were at the anastomotic site. One was a stalked polyp of the sigmoid colon, visible from the rectum, and another was at 25cm, with an uncertain relationship to the ureteric stomas. The fifth was a small lesion located in the proximal descending colon, and the sixth a 1-cm polyp near the sigmoid/descending junction. The sixth patient was found to have an irregular stoma from which a generous snare loop biopsy was taken. She is currently awaiting further follow-up.

The histology of these tumours, together with their surgical treatment is outlined in Table III. None of the biopsies of mucosa surrounding the ureteric stomas showed adenomatous change or epithelial dysplasia.

Although this was essentially a prospective study, it is important to note that 4 of the 34 patients had, in the past, undergone surgical excisions of symptomatic tumours at the ureterocolic anastomosis, 3 within the past 10 years, and one 20 years previously (Table III). One patient excluded from this study was, by error, not referred for endoscopy. Later that

Table III: HISTOLOGY AND TREATMENT OF TUMOURS A T URETEROCOLIC ANASTOMOSIS

Tumour Histology Size Treatment

Diagnosed at 1. Tubular adenoma with colonoscopy mild dysplasia (present study)

2. Tubular adenoma with mild dysplasia

3. Two tubular adenomas with mild dysplasia

4. Tubular adenoma with mild dysplasia

5 . Stoma with severe epithelial dysplasia

6. Villous adenoma with moderate epithelial dysplasia

Tumour previously 7. Villous adenoma

8. Villous adenoma 9. Adenocarcinoma

resected

(well-differentiated)

(well-differentiated) 10. Adenocarcinoma

1 mm

1 cm

1 cm+ 1 cm

1.5cm

1.5cm

6 cm

22cm

>2cm

Hot biopsy

Snare polypectomy

Snare polypectomy

Snare polypectomy

Snare loop biopsy

Colotomy and polypectomy

Local excision

Local excision Local resection and

reimplantation of ureter into colon

and ileal condult Anterior resection

41 6

Table IV: ADENOCARCINOMAS AND POLYPS ARISING A T THE URETEROCOLIC ANASTOMOSIS (6 )

MEAN AGE AND LATENCY OF

M . Stewart et al.

Surveillance of patients after urinary diversion is important and should be in centres with the necessary experience and skills, including fibresigmoidoscopy. In addition to the clinical importance, this surgically created meeting place of the urinary stream and the intestinal mucosa provides a unique model for the study of colon carcinogenesis.

Tumours at the Mean age at ureterocolic diagnosis anastomosis lvr l Mean latency

Polyps* 33.5 19.8 Adenocarcinoma 34.7 25.8

*Ninety per cent of these polyps were adenomas.

year, she underwent resection of a sigmoid carcinoma but subsequently died of metastatic disease at the age of 29 years.

Excluding this last case, 10 patients, or 29 per cent of the patients studied, had at some stage, developed a neoplastic lesion in the sigmoid or descending colon.

Discussion A largely asymptomatic group of patients with ureterosig- moidostomies has been followed up for an average of 22 years. Fibresigmoidoscopy and biopsy in 6 patients revealed lesions with the potential for malignant transformation at or close to the ureteric orifice. No rectal lesions were encountered, and biopsy of surrounding colonic mucosa showed no abnormality. In this respect, patients acted as their own controls.

Data concerning the incidence of adenomas of the sigmoid colon in asymptomatic persons aged 40 years are unavailable. Estimates of their prevalence based on autopsy reports show wide variation (7-51 per cent). However, the published results of major autopsy series distinguishing adenomas from other polyps suggest a prevalence of about 10 per cent for recto- sigmoid adenomas in this age group (1 1 - 13).

The prevalence of adenoma/dysplasia in this study (17.6 per cent) seems surprisingly low, considering the high estimated risk of carcinoma in retrospective series. However, all except one of the neoplasms in our series were 1 cm or over in size, whereas the majority of adenomas in autopsy series are under 1 cm. Detection of larger adenomas is potentially more im- portant, as the incidence of invasive malignancy increases progressively with size (14). These findings would support the adenomaxarcinoma sequence hypothesis ( 1 5) if the effect of urinary diversion was to increase the size of existing adenomas or the rate of progression of adenoma to carcinoma rather than encourage the development of adenoma per se. Our findings also complement other data (Table ZV) which indicate that adenomas occurring at the ureterocolic anastomosis occur, on average, 6 years earlier than carcinomas (6). This may suggest that malignant transformation takes 5- 10 years to develop and occurs more frequently and more rapidly in these patients.

Acknowledgements F. A. M. is the recipient of an Applied Health Sciences Fellowship from the Australian National Health and Medical Research Council. We are grateful for the review of pathology slides by Dr B. C. Morson and for the secretarial assistance of Mrs Pam Primavesi and Mrs Denise Ellison.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

Urdaneta L. F., Duffell D., Creevy C. D. et al. Late development of primary carcinoma of the colon following ureterosigmoid- ostomy: a report of three cases and literature review. Ann. Surg.

Parsons C. D., Thomas M. D. and Garrett R. A,: Colonic adenocarcinoma: a delayed complication of ureterosigmoid- ostomy. J. Urol. 1977; 118: 31-4. Eraklis A. J. and Folkman M. J.: Adenocarcinoma at the site of the ureterosigmoidostomies for exstrophy of the bladder. J . Pediatr. Surg. 1978; 13: 730-4. Doll R., Muir C. and Waterhouse J. (ed.) UICC: Cancer in Five Continents, Volume 11. Berlin: Springer-Verlag, 1970. De Jong V. W., Dai N. E., Muir C. S. et al.: The distribution of cancer within the large bowel. Int. J . Cancer 1972; 10 463-77. Stewart M., Hill M. J.. Pugh R. C. B. et al.: The role of N-nitrosamine in carcinogenesis at the ureterocolic anastomosis. Br. J . Urol. 1981; 53: 115-18. Maydl K.: Ueber die Radikaltherapie der Ectopia Vesicae Urinariae. Wien. Med. Wochenschr. 1894; 25: 11 15-16. Coffey R. C.: Physiologic implantation of the severed ureter or common bile duct into the intestine. J A M A 191 1; 56: 397-403. Leadbetter W. F.: Consideration of problems incident to perfor- mance of ureterosigmoidostomy-report of technique. J . Urol.

Williams C. B. and Gillespie P. E.: Accidental removal or ureteral stoma at colonoscopy. Gasfroinfest. Endosc. 1979; 2 5 109- 10. Arminski T. C. and McLean D. W.: Incidence and distribution of adenomatous polyps of the colon and rectum based on 1000 autopsy examinations. Dis. Colon Rectum 1964; 7 249-61. Chapman I.: Adenomatous polypi of large intestine: incidence and distribution. Ann. Surg. 1963; 157: 223-6. Blatt L. J.: Polyps of the colon and rectum: incidence and distribution. Dis. Colon Recfum 1961; 4: 277-82. Muto T., Bussey H. J. R. and Morson B. C.: The evolution of cancer of the colon and rectum. Cancer 1975; 36: 225 1-70. Hill M. J., Morson B. C. and Bussey M. J. R.: Aetiology of adenoma carcinoma sequence in large bowel. Lancet 1978; 1:

1966; 164: 503-13.

1951; 65: 818-30.

245-7.

Paper accepted 19 January 1982.