A case-control study of colon adenoma in relation to dietary habits was undertaken in Saitama, Japan. The population included 105 patients with newly diagnosed colon adenoma, 843 general population controls, and 84 hospital controls who had no pathologic lesions at colonoscopy. The adenoma patients and the hospital controls underwent total colonoscopy between November 1992 and March 1994. Information about diet, cigarette smoking and alcohol use were obtained with a self-administered questionnaire. Unconditional logistic regression analysis was used to compare adenoma patients with general population controls and with hospital controls. We found that current cigarette smokers had a higher risk (OR = 2.86, CI = 1.20-6.80) than did nonsmoking hospital controls. Current alcohol use did not affect the risk of colon adenoma. Ex-drinkers had a higher risk (OR = 2.36, CI = 1.02- 5.45) than did nondrinking general population controls. The same tendency was observed in comparison with hospital controls. The consumption of raw vegetables was inversely related to the risk of colon adenoma when adenoma patients were compared with general population controls (P for trend = 0.02) or hospital controls (trend P=0.17). Having one bowel movement per day was associated with a lower risk than having less or more frequently bowel movements when adenoma patients were compared with either general population or hospital controls. The present study suggests that the high consumption of raw vegetables is important in the primary prevention of colon adenoma.