In a multi-center case-control study, we evaluated the risk of esophageal cancer in the Japanese population. All patients and controls were inpatients in the surgical departments of seven hospitals nationwide. Patients eligible for the study were those newly diagnosed as having primary esophageal cancer. One control per case was selected from among patients admitted to the same hospital, and 141 male pairs were analyzed using logistic regression analysis. The results showed dose-response relation between the risk of esophageal cancer and both the quantity (g/week) and frequency (times/week) of alcohol drinking (P value for trend = 0.0001). Although a statistically significant risk increase was shown among moderate to heavy smokers (15 < or = cigarette/day < 25) (odds ratio, 4.35:95% confidence interval, 1.81-10.49), the dose-response for cigarette smoking was unclear (P value for trend = 0.07). No combined effect of alcohol drinking and cigarette smoking was found. A frequent intake of fruit was associated with a decreased risk (P value for trend = 0.02). After adjustment for alcohol consumption, cigarette smoking and fruit intake were found not to be associated with the risk, whereas a preference for high-temperature food and drink showed a statistically significant positive association (P value for trend = 0.02). Drinkers who consumed shochu most frequently showed a three-fold increased risk over that for beer consumers, although the association disappeared after adjusting for the amount of alcohol consumed. The present results confirm alcohol intake and a preference for high-temperature food to be associated with an increased risk of esophageal cancer and show the amount of alcohol consumed, rather than the type of alcoholic beverage, to be the main risk determinant.