In 13 hypercholesterolemic children, re-screened for serum cholesterol after a 1-year interval, hypercholesterolemia was confirmed in only 61.5% of the cases. A tentative explanation seems to be the statistical principle of regression towards the mean. The lipid--lipoprotein analysis showed that serum and LDL cholesterol concentrations in the 13 hypercholesterolemic children and their parents were significantly higher compared to controls (children and parents). At re-screening, hyper-LDL cholesterolemia was present in only 8 of the 13 children (61.5%); 4 cases exhibited hyper-HDL cholesterolemia (30.7%). The high prevalence of the parents repeating the lipoprotein abnormality and the electrophoretic pattern found in the propositi (children) confirms the familial aggregation of the hypercholesterolemic states (hyper-LDL and hyper-HDL cholesterolemia). In conclusion the results of our study stress the importance of determining the lipid--lipoprotein composition, rather then merely evaluating total serum cholesterol in order to make a correct diagnosis of the hypercholesterolemic state. It should also be emphasized that the lipoprotein disturbances and their familial aggregation may be detected early in childhood, suggesting that the familial screening for risk factors of atherosclerosis should be done at pediatric age.