The association between coffee consumption and bladder cancer was investigated in a multi-centre case-control study conducted in Spain from 1983 to 1986. A total of 497 cases (438 male and 59 female) with histopathologically confirmed bladder cancer were used in the analysis along with 566 hospital controls and 547 population controls. Odds ratios (OR), adjusted for age, province of residence, occupations at risk, consumption of artificial sweeteners and cigarette smoking, did not show any association between coffee consumption and bladder cancer for either sex. However, in non-smokers and mainly in males, for current coffee drinkers the OR was 2.78 (95% [Cl]: 0.78-9.87), while for drinkers of 2-7, 8-14 and > or = 15 cups/week the respective OR were 2.22, 3.11 and 1.87 with a dose-response relationship for lifelong consumption and years of exposure to regular coffee consumption. The OR in male non-smokers and current coffee drinkers were 2.36 (95% Cl: 0.62-9.05) with population controls only and 1.94, 2.58 and 1.48 for the corresponding levels of intensity of consumption (cups/week). The associations observed in non-smokers suggest the existence of a possible association between coffee consumption and bladder cancer, but are based on small numbers and need to be confirmed in larger studies.