Alcohol abuse is a major predictor of premature death, and also an independent risk factor for coronary death. Alcoholics are often heavy smokers. In a large primary prevention trial in middle-aged men with 7,495 participants, registration data were used to identify subjects with alcohol problems. Smokers had slightly lower blood pressure and were somewhat leaner than non-smokers, but had slightly higher serum cholesterol levels. Alcohol-registered subjects also tended to have lower blood pressure levels, and higher serum cholesterol. The relative risk of non-fatal myocardial infarction during the follow-up period of 11.8 years was essentially doubled in smokers compared to non-smokers, regardless of registration for alcohol problems. Among the non-alcoholic subjects, the relative risk of coronary death in smokers was double that of non-smokers, whereas the risk in non-smoking alcoholic subjects was not significantly increased. In smoking alcoholics the relative risk was substantially raised to 4.2 (3.0-7.0; 95% c.i.). In multivariate analysis both smoking and alcohol abuse were independently associated with coronary death. A possible mechanism might be through a combination of tobacco-induced coronary arteriosclerosis and the cardiotoxic effects of alcohol. As to total mortality, a smoking non-alcoholic man had a relative risk of dying almost double that of a non-alcoholic non-smoker. Among non-smoking alcoholics the risk was three times and, in smoking alcoholics over four times that of the non-alcoholic non-smokers.