To elucidate the natural history of colorectal polyps and to observe the influence of endoscopic polypectomy on the incidence of colorectal cancer, we conducted a retrospective cohort study of all patients who had undergone colonoscopic examination at the Center for Adult Diseases, Osaka, between April 1974 and December 1985. The study subjects consisted of a control group (760 non-polyp patients) and a polyp group (648 polyp patients, including 136 treated by polypectomy at the initial examination). These subjects were followed up until the end of 1987 by record linkage with the Osaka Cancer Registry's file to observe the occurrence of colorectal cancer. The O/E (observed/expected numbers derived from the general population) was 5.1 (95% confidence interval = 2.5-9.4) and 1.0 (0.1-3.6) for the polyp and control group, respectively. When subjects in the polyp group were categorized into polypectomy and non-polypectomy sub-groups, the O/E was 2.3 (0.1-12.6) and 8.0 (3.4-15.8) respectively. The relative risk of undergoing polypectomy to developing subsequent cancer was estimated at 0.3 (0.1-2.1). These results suggest an increased risk of developing cancer among polyp patients and the possibility of prophylactic effect of polypectomy against subsequent cancer. A large-scale and long-term follow-up study is required to confirm these findings.