Ulcerative Colitis - NCBI

4 downloads 307 Views 702KB Size Report
Apr 5, 1988 - though prophylactic surgery does not seem to be generally war- ranted. ... than 21 years of age at onset of the disease (Table 1). Twenty-one per cent of the ... Eighty-six per cent ofthe patients were hospitalized at least once.
Ulcerative

Colitis

Colorectal Cancer Risk in an Unselected Population

JORGEN N. RUTEGARD, M.D.,*t LARS R. AHSGREN, M.D.,* KARL G. JANUNGER, M.D., PH.D.* One hundred twenty-seven patients were studied, representing the unselected population of patients with ulcerative colitis from a defined catchment area of about 70,000 inhabitants during 1961-1983. Seventy-seven patients had total colitis, and 50 patients bad left-sided colitis. No prophylactic colectomy was performed. Despite this and a comprehensive follow-up of patients with long-standing, extensive colitis, primary colorectal carcinomas were diagosed in only three patients, all of whom had total colitis. This was significantly (p < 0.001) more than the expected number of patients with colorectal cancer in this subgroup, namely, 0.13. There were no deaths in colorectal carcinoma. Factors determining completeness of inclusion are discussed, and it is concluded that, in an unselected series of patients with ulcerative colitis, close cancer surveillance is indicated, although prophylactic surgery does not seem to be generally warranted.

T n HE INCREASED RISK of developing colorectal

cancer in chronic ulcerative colitis (UC) has been widely recognized.''4 However, the reported incidence of colitis-associated carcinoma varies considerably. Reported studies are primarily from referral centers, and the extent of patient selection is therefore difficult to estimate.9"'5 The reported cancer risk is biased because of imperfect inclusion of patients with milder disease, the referral of more severe cases to specialized centers, and surgical intervention, (e.g., cancer-prophylactic proctocolectomy).5""1'13,16 The aim ofthis study has been to assess the cancer risk and the cancer mortality during a long-term follow-up of patients from a defined catchment area, without the bias inherent in patient referral and prophylactic surgery. Furthermore, an attempt is made to give an objective description of the completeness of patient inclusion.

Material and Methods The series consisted of 127 patients-all of those persons known to have UC who were treated either as in- or Reprint requests and correspondence: Dr. J. Rutegard, Department of Surgery, Ornskoldsvik Hospital, S-891 89 Ornskoldsvik, Sweden. Submitted for publication: April 5, 1988.

From the Department of Surgery, Ornsk6ldsvik Hospital, Ornskdldsvik,* and the Department of Surgery, University Hospital, Ume', Sweden.t

out-patients at the community hospital of Ornskoldsvik during the years 1961-1983. The series included 75 men and 52 women, and 30% of the patients were younger than 21 years of age at onset of the disease (Table 1). Twenty-one per cent of the patients had been diagnosed before 1961, the year that they were included in the study. The remaining patients were successively included at onset of the disease during the study period (Table 2). The diagnosis and extent of disease were established by large bowel barium enema, rigid sigmoidoscopy and, in later years, colonoscopy with biopsies. During the entirety of the study period, the well-defined catchment area included about 70,000 inhabitants. All surviving patients not operated on who remained in the area were followed in a colonoscopic surveillance program that began in 1977. Only one patient was referred to another hospital for treatment. Nine patients moved out of the area and have been followed-up by personal contact and through hospital records. Complete clinical history was not available for two patients (one of whom had total colitis), but at end of the study, had not carcinoma developed in these patients, as was verified in the National Registry of Cancer. Patients with proctitis alone were not included. To check whether any patient with the diagnosis of colorectal cancer and a previously unknown UC had been overlooked, all histopathologic records (n = 387) of patients with colorectal cancer were re-examined for the years 1961-1983. No such case was found. According to accepted criteria,""7 77 of the patients were suffering from total (extensive) colitis; the remaining 50 suffered from left-sided colitis. All available radiologic films have been re-examined to secure conform interpretation.

721

722

RUTEGARD, AHSGREN, AND JANUNGER

TABLE 1. Age at Onset of Ulcerative Colitis (Number of Patients)

Ann.

Surg

* December 1988

TABLE 3. Patient Years of Observation and Cancer Diagnoses Related to Duration of Disease for the Subgroup of Patients with Total Colitis

Age

(years)

Men

Women

Total

%

0-10 11-20 21-40 41+ Total

5 17 36 17

2 14 27 9

7 31 63 26

5.5 24.4 49.6 20.5

75

52

127

100.0

Eighty-six per cent of the patients were hospitalized at least once. The cumulative operation frequency was 26% (32 of 125 patients), and in the subgroup with total colitis, the frequency was 37% (28 of 76 patients). The main indications for surgery were acute or fulminant stages of the disease or chronic disabling symptoms. Fifteen patients died during the study period, six of UC, and nine of nonrelated disease. Endpoints for the study were 31.12 1983, time of complete operation (when entire colon and rectum were removed), or death.

Statistical Methods To determine the expected number of carcinomas, a calculation was made from the official cancer statistics,'8 and a regional correction was made for the lower incidence of colorectal carcinoma in Northern Sweden. For the calculations of statistical significance, Mantel's test and a test with a Poisson distributed test variable were used. In Table 3, the number of patient years of observation in the subgroup of patients with total colitis were calculated for the disease-duration groups of 5 years. Thus, a patient with UC for 13 years contributed 5 years to the first group, 5 to the second, and 3 years to the third. Results

Colorectal Cancer Mortality No death could be directly attributed to colitis-associated cancer. However, one man refusing medical care died with fulminant colitis. At autopsy, two carcinomas (one Dukes' A, one Duke's B) were found. Another man, TABLE 2. Duration of the Disease When Included in the Study (1961 or later)(Number of Patients)

Duration (years) 0 1-10 11+

Total

Men

Women

Total

%

58 9 8 75

42 9 1 52

100 18 9 127

78.7 14.2 7.1 100.0

Duration of Disease (years)

Observation Years

Patients with Colorectal Cancer

1-5 6-10 11-15 16-20 21-25 26-30 31-35 36+ Total

326 255 184 126 65 42 16 5

1 1 1

3

who died abroad, where no autopsy was performed, had been operated on 14 years before death for colonic carcinoma. Because there was such a long interval between the operation and death, it is unlikely that he died of recurrent malignancy. Cancer Incidence During the study period, three patients, all of whom had total colitis, had their first large bowel carcinomas diagnosed after 2, 9, and 27 years' duration ofthe disease, respectively (Table 3). The expected number of cancer patients in the whole series was 0.36 (p < 0.05, test with Poisson distributed test variable). In the subgroup of patients with total colitis, the expected number was 0.13, a difference that is highly significant (p < 0.001, test with Poisson distributed test variable). Two carcinomas were found postmortem in one man, mentioned above, who succumbed to his colitis after having refused surgical treatment. He had been proposed operation because of severe symptoms and findings of dysplasia at colonoscopy. The other two patients had single carcinomas, and underwent successful operations. These two later patients were treated before 1977 (the year that-our colonoscopic surveillance program began).'9

Discussion It is apparent that every surgical department that does not perform prophylactic proctocolectomy will face a situation in which a large and growing group of patients with long-standing UC will be at risk of developing colorectal cancer. This was our situation, and it incited this investigation. When estimating increased cancer risk in a patient population, it is of the utmost importance to assess the completeness of inclusion and the selective factors.9"5 Unfortunately, published UC patient materials often differ in these respects. Several published series regard their patient materials as representing the whole population of patients with UC.3'7"' However, the inclusion rate has

Vol. 208 * No. 6

ULCERATIVE COLITIS AND CANCER RISK

not been objectively demonstrated. In an attempt to give an objective measure of the inclusion rate, we have plotted

the patient years of observation in 5-year intervals against disease-duration years. These figures, together with corresponding data from a large, often cited investigation by Kewenter et al. " are given in Table 4 and are graphically demonstrated in Figure 1. No other published patient series has enabled us to make a direct comparison. Theoretically, the decline of the interrupted curve will be influenced by the inclusion of new diagnosed cases, the death rate, the operation frequency, population movements and referral patterns. Accordingly, a rapid decline in patient years of observation depends primarily on one or more of the following factors: a high rate of inclusion of new cases, imperfection in inclusion or follow-up of patients with long-standing disease, a high operation frequency, and a high mortality. Correspondingly, a flat curve indicates a high rate of inclusion of patients with disease of long duration, a low operation frequency, and a low mortality. Thus, we regard our analyzed patient series to represent, as far as is practically possible, an unselected material from a defined catchment area. One of several possible explanations for the rapid decline in observation years in the series by Kewenter et al." is the high operation frequency of 65% for patients with total colitis, compared with ours of 37%. A high proportion of operated patients indicates a selection, which is also postulated by Prior et al.'4 We propose that the slow decline in the patient years of observation in our series is an objective description of the completeness of inclusion.

The cancer incidence in our investigation was significantly higher than in the normal population, stressing the need for careful supervision (or prophylactic proctocolectomy) of these patients. Our series of patients with total colitis had a lower number of carcinomas, compared with the study cited above," where 15 of 234 patients with extensive colitis developed 17 colonic carcinomas, although the difference was not significant (Mantel's twosided test, p = 0.095). Among the patients observed who had more than 10 years' duration of extensive disease, the difference in cancer incidence was remarkable (Table 4). This tendency to lower cancer incidence in our unselected population suggests that extrapolation of cancer determinations from large centers to unselected populations of colitis patients must be made with care. The low cancer incidence in the patient group with longer duration of disease in the present series (Table 4) is difficult to explain without postulating biases, as discussed above, or regional differences in severity of the disease that have thus far gone unrecognized. The possibility of regional differences in cancer risk has been discussed by Gilat et al.7'8 Their own patient material,7 which is of comparable size to ours for patients observed after longer duration of

723

TABLE 4. Patient Years of Observation Related to Duration ofDisease Number of Patients Developing Colorectal Cancer

Observation Years

Duration of Disease (Years) 1-5 6-10

G

0

875

326 255 184 126 65 42 16 5

523 282

11-15 16-20 21-25 26-30

147

73 * * *

31-35 36+ Total

G

0

3

1 1

4 5

2

1

1

3

15

* Not indicated. Comparison between the Gothenburg (G) study,"' and the present study from the Community Hospital, ornskoldsvik (0). Only patients with total colitis were included.

the disease, also has a low cancer risk, with only three carcinomas diagnosed in 504 patients, 259 of whom had substantial or universal colitis. Our patient with a colonic cancer, developing after only 2 years of UC with distinct onset, may be a case of nonrelated colitis and carcinoma. Yet even so, the cancer incidence was still lower than that calculated above. The risk of early cancer development has recently been demonstrated in a series in which seven of 15 patients with colitis developed adenocarcinomas before 10 years' duration after onset of symptoms.20 Our series as well as

Patient years of observation 1000

---

Kewenter et al 1 978 "Present study"

I-

500 _

;~~~~~~~~~~~~~~~~~~~~~~~ 5

10

15 15

20

25

30

35

40

Disease duration years

FIG. 1. Patient years of observation plotted against disease duration in patients with long-standing extensive ulcerative colitis: comparison between a University hospital (present study) in Sweden.

(Gothenburg)"

and

Ornskoldsvik Hospital

RUTEGARD, AHSGREN, AND JANUNGER

724

those of other investigations6'7 " also demonstrate the possibility of cancer development before 10 years' duration of the disease, indicating the need for colonoscopic surveillance earlier in the course ofthe disease, at least in cases ofactive disease, rather than relying on prophylactic proctocolectomy after 10 years' duration. In our study, we have, for tw? reasons, included patients with left-sided colitis that was radiologically determined. First, a certain overrisk for cancer, although after longer duration of the disease, has been reported.2' Secondly, a radiologic left-sided colitis is found to be a histologic total colitis in a high proportion of the cases,'9 and this may accordingly imply an increased cancer risk. We have not been able to demonstrate any further increasing cancer risk beyond 10 years of duration of the disease. Instead, the low incidence of colorectal cancer in our patients with long-standing UC lends some support to the view of Prior et al.,'4 who found a relative risk to increase during the first 15 years after diagnosis, and thereafter remain constant. In the present study, there was no mortality in colitisassociated cancer during 23 years. This finding, together with the low cancer risk in our patient population, increases our hesitation to perform preventive surgery in symptom-free patients, and supports the current trend of selecting other alternatives. If prophylactic proctocolectomy after 10 years' duration of total colitis had been our routine, more than 30 such operations would have been performed.22 Instead, we have so far used colonoscopic surveillance.19 The modern operations with proctocolectomy, ileal pouch, and ileo-anal anastomosis, may become another possibility, at least in selected patients, who prefer the operation with its drawbacks before a lifelong surveillance, as was recently discussed by Lennard-Jones.23 Conclusion

In an unselected population of 127 patients with ulcerative colitis who were treated at a community hospital over a period of23 years, the risk of developing colorectal cancer was signifcantly increased for patients with total colitis, although there was no evidence of increased risk with time. There was also no incidence of death in colitisassociated cancer during the study period, and close cancer surveillance was found to be indicated for patients with total colitis. Finally, in this patient population, because of a low cancer risk, prophylactic proctocolectomy did not seem to be warranted; however, this question should be further investigated.

Ann. Surg * December 1988

Acknowledgments We thank Dr. Anders Oden for statistical evaluation and assistance.

References 1. Bargen JA. Chronic ulcerative colitis associated with malignant disease. Arch Surg 1928; 17:561-576. 2. Brostrom 0, Lofberg R, )st A, Reichard H. Cancer surveillance of patients with long-standing ulcerative colitis: a clinical, endoscopical and histological study. Gut 1986; 27:1408-1413. 3. Carleson R, Fristedt B, Philipson J. Ulcerative colitis: a follow-up investigation of a 20-year primary material. Acta Med Scand 1962; 172:647-655. 4. Dawson IMP, Pryse-Davies J. The development of carcinoma of the large intestine in ulcerative colitis. Br J Surg 1959; 47:113128. 5. Devroede GJ, Taylor WF, Sauer WG, et al. Cancer risk and life expectancy ofchildren with ulcerative colitis. N Engl J Med 197 1;

285:17-21. 6. Edwards FC, Truelove SC. The course and prognosis of ulcerative colitis. IV. Carcinoma of the colon. Gut 1964; 5:15-22. 7. Gilat T, Zemishlany Z, Ribak J, et al. Ulcerative colitis in the Jewish population of Tel-Aviv Yafo. II. The rarity of malignant degeneration. Gastroenterology 1974; 67:933-938. 8. Gilat T, Rozen P. Risk of colon cancer in ulcerative colitis in low incidence areas: a review. In Winawer S, Schottenfeld D, Sherlock P, eds. Colorectal Cancer: Prevention, Epidemiology, and Screening. New York: Raven Press, 1980; 335-339. 9. Gyde SN. Cancer in ulcerative colitis. In Bouchier IAD, Allan RN, Hodgson HJF, Keighley MRB, eds. Textbook of Gastroenterology. London: Baillere Tindall, 1984; 927-940. 10. Katzka I, Brody RS, Morris E, Katz S. Assessment of colorectal cancer risk in patients with ulcerative colitis: experience from a private practice. Gastroenterology 1983; 85:22-29. 11. Kewenter J, Ahlman H, Hult6n L. Cancer risk in extensive ulcerative colitis. Ann Surg 1978; 188:824-828. 12. Lennard-Jones JE, Morson BC, Ritchie JK, Williams CB. Cancer surveillance in ulcerative colitis. Experience over 15 years. Lancet 1983; ii: 149-152. 13. MacDougall IPM. The cancer risk in ulcerative colitis. Lancet 1964; ii:655-658. 14. Prior P, Gyde SN, Macartney JC, et al. Cancer morbidity in ulcerative colitis. Gut 1982; 23:490-497. 15. Whelan G. Cancer risk in ulcerative colitis: why are results in the literature so varied? Clin Gastroenterol 1980; 9:469-476. 16. Coran AG. New surgical approaches to ulcerative colitis in children and adults. World J Surg 1985; 9:203-213. 17. Bartram CI, Walmsley K. A radiological and pathological correlation of the mucosal changes in ulcerative colitis. Clin Radiol 1978;

29:323-328.

18. Cancer incidence in Sweden 1978. National Board of Health and Welfare, Cancer Registry. Stockholm 1978. 19. Rutegard J, Ahsgren L, Stenling R, Janunger KG. Ulcerative colitis. Cancer surveillance in an unselected population. Scand J Gastroenterol 1988; 23:139-145. 20. Turunen MJ, Jarvinen HJ. Cancer in ulcerative colitis: what failed in follow-up? Acta Chir Scand 1985; 151:669-673. 21. Greenstein AJ, Sachar DB, Smith H, et al. Cancer in universal and left-sided ulcerative colitis: factors determining risk. Gastroenterology 1979; 77:290-294. 22. Rutegard J, Ahsgren L, Janunger KG. Ulcerative Colitis. Mortality and surgery in an unselected population. Acta Chir Scand 1988;

154:215-220.

23. Lennard-Jones JE. Compliance, cost, and common sense limit cancer control in colitis. Gut 1986; 27:1403-1407.