In 45 subjects after a cerebrovascular attack (age less than 55 years) and their first degree relatives -45 siblings, 60 children, 9 parents and 31 partners--lipoprotein (a) levels by the ELISA method as well as some other lipids, lipoproteins and apoproteins were assessed and compared with two control groups (I and II) of healthy subjects with a negative family history of early atherosclerosis. The lipoprotein (a) levels were significantly higher not only in probands, as compared with the control group I (27.4 +/- 30.7 x 16.0 +/- 18.9; p < 0.05, Mann Whitney U-test) but also in their children when compared with control group II (28.1 +/- 31.9 x 12.5 +/- 16.3; p < 0.05). The lipoprotein (a) levels did not correlate with age, there was, however, a significant positive correlation with total cholesterol (r = 0.2244, p < 0.01), LDL-cholesterol (r = 0.2834, p < 0.0001) and apolipoprotein B (r = 0.1759, p < 0.05) when Spearman's correlation coefficient was used. Familial dyslipidaemia Lp (a) defined as lipoprotein (a) higher than 25 mg/dl (or 30 mg/dl resp.) at least in two close relatives in one family was found in 27% (and 21.6%) evaluated families (10/37 and 8/37 resp.). "Risk" levels of lipoprotein (a) > 25-30 mg/dl were found most frequently in combination with other types of hyperlipoproteinaemia--in subjects with types IIa and IIb, in families in combination with familial combined hyperlipidaemia. The results indicate that "risk" levels of lipoprotein (a) are a frequent genetically conditioned metabolic disorder in patients with premature ischaemic cerebrovascular attacks.