The cardioprotective effect of moderate alcohol consumption, which is mediated largely by an elevation of high-density lipoprotein (HDL) levels, may be absent in patients with primary disorders of lipid metabolism. In 881 untreated, unrelated patients with primary hyperlipidemia, we estimated the effect of alcohol on fasting serum HDL cholesterol concentration adjusted for age, gender, diet, and smoking with a multiple linear regression model. The number of abstainers, moderate users (1-30 g daily), and heavy alcohol consumers (> or = 30 g daily) was not different between three types of hyperlipidemia. Among hypercholesterolemic (n = 514) and combined hyperlipidemic (n = 227) patients, consumers of alcohol had higher HDL levels compared to abstainers (P(for trend) = 0.0001 and 0.0002, respectively). Heavy drinkers with combined hyperlipidemia had higher triglyceride levels compared to moderate drinkers and abstainers (P(for trend) = 0.008). Alcohol consumption had no effect on the lipid profiles in 140 patients with hypertriglyceridemia. In conclusion, combined hyperlipidemic and hypertriglyceridemic patients may profit from an encouragement to stop consuming alcohol. On the other hand, moderate use of alcohol may delay the development of coronary artery disease in hypercholesterolemic patients by an elevation of HDL levels.