Long-term therapy with sulphasalazine protects against colorectal cancer in ulcerative colitis: a retrospective study of colorectal cancer risk and compliance with treatment in Leicestershire

Abstract

The aims of this study were to (i) estimate the prognosis of a 10-year cohort as expressed by risk of colectomy and risk of development of colorectal cancer, and (ii) assess the impact of long-term sulphasalazine on the natural course of ulcerative colitis. One hundred and seventy-five patients diagnosed between 1972 and 1981 with either total colitis (n = 143) or with limited ulcerative colitis but deceased (n = 32) were identified. Overall there was 98% case ascertainment and verification. A total of 49 patients underwent a colectomy, 6 as emergency laparotomies, 36 for failed medical management and 7 for known colorectal cancer, giving a crude colectomy rate of 23.2%. The colectomy rate was 7.2% in the year of diagnosis, decreasing in frequency over the next 4 years, then reaching a steady state of approximately 1.7% per year. In the total cohort, colorectal cancer occurred in 10 patients within the study period. The cumulative incidence of colorectal cancer 10 years after diagnosis was 2.1% and at 20 years 7.4% for the total group of patients excluding those with a colectomy. The mean duration of ulcerative colitis before diagnosis was 7.9 years (range 5-12). The crude proportions developing cancer were 5/152 (3%) in the group who took long-term sulphasalazine but 5/16 (31%) in the those who had had their treatment stopped or who did not comply with therapy. This is highly significant using a simple chi 2 test (chi 2 = 20.2, df = 1, P < 0.001). Two methods were used for survival analyses, the log-rank and the generalized Wilcoxon methods. Both give highly significant values for the crude effect of compliance (P < 0.001). Patients with ulcerative colitis who were not on long-term sulphasalazine or 5-aminosalicylic acid therapy (either because a doctor stopped it or they did not comply with treatment) were significantly more likely to develop colorectal cancer than their compliant counterparts.