OBJECTIVE To determine, among persons who have had a large colon polyp, the risk of subsequent colon cancer at a site distant from that polyp. METHODS Follow-up was done for 226 persons at the Mayo Clinic who had had a > or = 1-cm polyp demonstrated on barium enema between 1965 and 1970 and for whom yearly colon surveillance examination was recommended. Information was collected from Mayo Clinic records and from contact with patients, physicians, and other hospitals regarding the results of surveillance examinations and the development of colon cancer. Colon surveillance was routinely done at the Mayo Clinic using the technique of single contrast barium enema with vigorous manual fluoroscopic examination and proctoscopy. The expected rate of colorectal cancer (CRC) was calculated based on previously published rates for this community. RESULTS Patients received, on average, four colon examinations in addition to the examination that discovered the index polyp. During 2126 person-years of follow-up, 16 persons developed a colon cancer at a location other than the site of the index polyp, in comparison with 4.0 expected cases, for a standardized incidence ratio of 4.0 (95% CI,2.3, 6.4). The cancers were large (mean 4.5cm) at presentation, and eight of the 16 cancers had been preceded within 3 yr by at least one negative barium enema. CONCLUSIONS The rate to develop colon cancer in persons who have had a large colon polyp es about 4 times the expected rate, suggesting that such persons should be considered for aggressive colonoscopic surveillance. The failure to detect early cancer or its precursors by surveillance barium enema is probably explained by inherent insensitivity of single contrast barium enema.