Although cigarette smoking is a well-known risk factor of lung cancer, associations of cigarette smoking with the risk of other sites have not been fully elucidated in Japan. To simultaneously evaluate the associations of cigarette smoking with the risks of cancers of the stomach, lung, colon, and rectum, which have been the leading cancer sites in recent years in Miyagi Prefecture, Japan, we conducted a hospital-based case-control study. Study subjects consisted of 614 stomach, 515 lung, 324 colon, and 164 rectal cancer cases and 2444 hospital controls admitted to a single hospital in Miyagi Prefecture from 1997 to 2001. Information on smoking habit and other lifestyle factors was collected using a self-administered questionnaire. Distributions of referral base among cases and controls were also investigated. For each site, odds ratios (ORs) and 95% confidence intervals (95% Cls) for smoking habit were estimated with adjustment for age, year of survey, history of alcohol drinking, family history of index cancer, and occupational history, respectively, using an unconditional logistic regression model. Cigarette smoking (ever vs. never) was associated with an increased risk of stomach (OR=1.62; 95% Cl 1.20–2.19) and lung (OR=3.82; 95% Cl 2.49–5.86) cancer among males and lung cancer among females (OR=2.02; 95% Cl 1.28–3.18). For female stomach cancer, the association with cigarette smoking was uncertain (OR=0.65, P=0.1533). For rectal cancer, a significant increased risk was observed in both-sex-combined analysis. There was no association between cigarette smoking and the risk of colon cancer. Detailed analysis showed that the association of cigarette smoking with cancer risk might be modified by the patient referral pattern, i.e., screened or not screened. The present results indicate that the association of cigarette smoking with cancer risk may differ among sites and sexes. In terms of the population attributable risk, a large proportion of leading cancers in males appears to be related to cigarette smoking.