OBJECTIVE: This study investigated the association between green tea consumption and leukemia. METHODS: A total of 252 cases (90.3% response) and 637 controls (53.4% response) were enrolled. Controls were matched for cases on age and gender. Information was collected on participants' living habits, including tea consumption. Green tea was used as a standard to estimate the total amount of individual catechin consumption. We stratified individual consumption of catechins into four levels. Conditional logistic regression models were fit to subjects aged 0-15 and 16-29 years to evaluate separate associations between leukemia and catechin consumption. RESULTS: A significant inverse association between green tea consumption and leukemia risk was found in individuals aged 16-29 years, whereas no significant association was found in the younger age groups. For the older group with higher amounts of tea consumption (>550 units of catechins), the adjusted odds ratio (OR) compared with the group without tea consumption was 0.47 [95% confidence interval (CI) = 0.23-0.97]. After we adjusted for smoking status and medical irradiation exposure, the overall OR for all participants was 0.49 (95% CI = 0.27-0.91), indicating an inverse relation between large amounts of catechins and leukemia. CONCLUSION: Drinking sufficient amounts of tea, especially green tea, which contains more catechins than oolong tea and black tea, may reduce the risk of leukemia.