There is no adequate information on the carcinogenicity of coffee and, specifically, on a potential association of coffee drinking with non-Hodgkin's lymphoma. Consumption of coffee and other methylxanthine-containing beverages has been researched in a case-control study conducted in northern Italy. A total of 429 cases of incident histologically confirmed non-Hodgkin's lymphoma and 1,157 controls in hospital for acute, non-neoplastic, non-immunological, non-digestive tract diseases were interviewed during their hospital stay. Relative risk (RR) estimates and their 95% confidence intervals (CI), according to consumption of coffee and other methylxanthine-containing beverages, were derived from multiple logistic regression equations including terms for age, sex, study centre, body mass index, alcohol and smoking status. Compared with non-drinkers, the RR was 1.2 (95% CI, 0.8-1.7) for coffee drinkers. No trend in risk emerged with number of cups of coffee consumed/day (RR = 1.1 for one and three cups; RR = 1.2 for two; or RR = 0.9 for four cups/day), or duration of coffee intake (RR = 1.2 for less than 20 years; RR = 1.3 for 21-30 years; and RR = 1.1 for more than 30 years). Similarly, no significant association was observed with consumption of decaffeinated coffee (RR = 0.9) or tea (RR = 1.2). Consumption of cola was associated with a borderline risk (RR = 1.7; 95% CI 1.0-2.7). We found no association between non-Hodgkin's lymphoma and consumption of regular or decaffeinated coffee and tea.