A multicentre hospital-based incident case-control study with 520 male gastric cancer cases aged <75 years and an equal number of age- and sex-matched controls without cancer has been carried out in Poland to assess potential risks due to smoking and alcohol consumption. It was shown that after adjusting for socio-demographic and dietary confounders and vodka drinking, smoking cigarettes had no significant effect on risk. The estimated relative risk (RR) increased to 2.27 (95% confidence interval [Cl] : 0.97-5.28) for intestinal cardia cancer for those who smoked cigarettes without filters. The RR for stomach cancer grew as the frequency and amount of vodka drunk increased. People drinking vodka at least once a week had about a threefold higher risk compared to non-drinkers (RR = 3.06, 95% Cl : 1.90-4.95). The effect of vodka drinking on risk was particularly strong for non-cardia cancers of the intestinal type. Those who usually drank vodka before breakfast had an elevated risk (RR = 2.98, 95% Cl : 1.60-5.53) which was also present in all the subgroups investigated. Cardia and non-cardia cancer showed differences with respect to the interaction between tobacco smoking and vodka drinking. For cancers of the cardia region the risk was low for non-smokers or those who drank small amounts of vodka. The risk for cardia cancer increased considerably for smokers of cigarettes without filters and vodka drinkers who consumed large amounts (RR = 3.70, 95% Cl : 1.13-12.06). For the non-cardia region a uniform increase could be observed for vodka drinking regardless of cigarette smoking status. The findings of this study support the hypothesis that the effect of tobacco smoking and vodka drinking may be different for cardia cancer compared to the more distal cancers.