The results of a prospective randomized study of 1056 patients with colorectal adenomas are presented. After initial polypectomy from 1978 to 1992, patients were allocated at random to different follow-up intervals varying from 6 to 48 months, except 53 patients who were allocated to intervals of 6 months. The examinations were mainly done by colonoscopy. Ten patients developed colorectal carcinoma, a number similar to that expected (7.96), when compared with a sex- and age-matched normal Danish population. The expected number of carcinomas was also calculated from adenoma to carcinoma conversion rates estimated in other studies and compared with that observed. If all carcinomas develop in large (> or = 10 mm) adenomas or adenomas with severe dysplasia, the expected number of carcinomas would have been 62 and 110, respectively, indicating a significant reduction of carcinomas in the present study. One patient died of colorectal carcinoma, which is significantly lower than the number expected (7.58). Two patients died of complications from therapeutic and diagnostic colonoscopy--that is, 2 deaths in 3959 colonoscopies. In conclusion, the follow-up strategy has resulted in a mortality from colorectal carcinoma which is reduced when compared with the normal population, in spite of an apparently similar incidence of carcinoma. However, previous suggested adenoma-carcinoma conversion rates indicate that a major reduction of incidence actually has taken place.