Multiple studies have shown a protective effect for physical activity on the incidence of colorectal cancer, but few have addressed its association with incident adenomas and none have investigated its association with metachronous adenomas. We interviewed 2,001 patients undergoing colonoscopy in 3 New York City practices between 1986-1988. Of these, 298 had a first diagnosis of adenomas, 506 had normal colonoscopies, 197 had metachronous adenomas and 345 had normal colonoscopies with a past history of adenomas. Subjects were queried regarding use of moderate amounts of recreational exercise (swimming, jogging, biking, racket sports and other sports) on a regular basis, and were also asked to estimate their level of physical activity. Occupational physical activity was assessed through a blinded coding of occupational titles. After adjusting for age, years of education, body mass index, total caloric intake, dietary fiber intake, dietary fat intake and years of cigarette smoking, protective effects of borderline significance were observed among males for leisure physical activity for metachronous and incident cases. Occupational physical activity was found to be significantly protective in male incident cases. Men with increased levels of either occupational or leisure activity showed a reduction for incident and metachronous cases. No effects were observed for females. Thus, in this study, both leisure and occupational physical activity were associated with a reduced risk of incident and metachronous adenomas among men but not women.