A few studies have investigated whether the risk of laryngeal cancer depends on the types of alcoholic beverage consumed, providing conflicting results. We investigated this issue using the data from two case-control studies conducted in Italy between 1986 and 2000. These included 672 cases of laryngeal cancer and 3454 hospital controls, admitted for acute, non-neoplastic conditions, unrelated to smoking and alcohol consumption. Significant trends in risk were found for total alcohol intake, with multivariate odds ratios (ORs) of 1.12 for drinkers of 3-4 drinks/day, 2.43 for 5-7, 3.65 for 8-11, and 4.83 for > 12 drinks/day, as compared to abstainers or light drinkers. Corresponding ORs for wine drinkers were 1.12, 2.45, 3.29 and 5.91. After allowance was made for wine intake, the ORs for beer drinkers were 1.65 for 1-2 drinks/day, and 1.36 for > or = 3 drinks/day, as compared to non-beer drinkers; corresponding values for spirits drinkers were 0.88 and 1.15. This study thus indicates that in the Italian population characterized by frequent wine consumption, wine is the beverage most strongly related to the risk of laryngeal cancer.