Screening for dysplasia in ulcerative colitis is expected to improve cancer-related survival by advancing the diagnosis of cancer to a more favorable pathologic stage or by identifying patients for prophylactic colectomy who are at at high risk of developing cancer. To evaluate the effectiveness of a surveillance program of annual colonoscopy and mucosal sampling, 91 screened and 95 control ulcerative colitis patients with extensive disease for at least 8 yr were compared for differences in survival, cancer detection, and colectomy rates. The two groups were of similar age at symptom onset, duration of disease, and sex. In the surveillance group, there were eight fewer deaths (p less than 0.05 by survival-curve analysis), but two more cancer deaths (not significant). Colectomy was less common (p less than 0.05) and was performed 4 yr later in the surveillance group. The effect of surveillance was not confounded by age at symptom onset, sex, or duration of disease. In our program, screening for cancer in ulcerative colitis was associated with improved survival and delayed colectomy. However, improvement was not related to the anticipated benefits of improved cancer-related survival.