An unexpectedly large number of hepatitis A virus (HAV) infections were notified among homosexual men in Rotterdam in the first five months of 1998. A case control study was conducted to investigate the hypothesis that this outbreak was associated with sexual practices and to collect information with which to focus preventive activities. Notified cases and controls selected from male members of a gay sports club completed anonymous questionnaires about known risk factors for HAV infection and sexual behaviour. Single variable analysis showed that HAV infection was associated with sexual contact with anonymous sex partners (odds ratio (OR) 4.6; 95% confidence interval (CI) 1.0-23.2) and with visits to 'darkrooms' in gay bars (OR 6.2; 95% CI 1.5-26.8). A negative association with travel abroad to western countries was observed (OR 0.4; 95% CI 0.1-1.3). The associations with visits to darkrooms (OR = 9.2; 95% CI 1.6-52.4) and travel abroad to western countries (OR 0.1; 95% CI 0.02-0.9) remained significant in a multivariable logistic regression analysis. This risk group was targeted for health education and vaccination and darkroom owners were advised to provide hygiene facilities.