The levels of different lipid, lipoprotein and haemostatic variables were assessed in 615 control subjects of the ECTIM Study, defined by five groups of alcohol consumption: non-drinkers, 0 < or = 15 g/day, 15 < . < or = 36 g/day, 36 < . < or = 66 g/day and > 66 g/day. After adjustment for age, body mass index, cigarette consumption and country, alcohol consumption was associated with an increase in HDL-cholesterol (0.47 +/- 0.02 to 0.59 +/- 0.01 g/l in non-drinkers and > 66 g/day consumers, mean +/- S.E.M., P < 0.0001), apolipoproteins A-I and A-II (1.37 +/- 0.03 to 1.60 +/- 0.03 g/l and 0.32 +/- 0.01 to 0.41 +/- 0.01 g/l, respectively, P < 0.0001), LpA-I, LpA-I:A-II (0.46 +/- 0.01 to 0.50 +/- 0.01 g/l and 0.75 +/- 0.02 to 0.91 +/- 0.02 g/l, respectively, P < 0.001) and PAi-1 activity (134 +/- 11 to 177 +/- 11 U/ml, P < 0.001). Conversely, no increases were found for total and LDL-cholesterol, triglycerides, apolipoproteins B and C-III, LpE:B, LpC-III:B, fibrinogen and factor VII. Hence, among the lipid and haemostatic variables studied, only HDL parameters and PAi-1 activity were increased by alcohol consumption.