The value of lipoprotein (a) [Lp(a)] in the prediction of coronary artery disease risk very early in life remains to be established in different racial groups. Serum Lp(a) distribution and its relation to parental histories of myocardial infarction were examined in 2,438 children (8-17 years old) from a biracial community. Parental myocardial infarction was used as a surrogate measure of future risk of disease in the offspring. Lp(a) levels averaged 1.7-fold higher in blacks than in whites (p less than 0.0001). A small but significant sex difference (females greater than males, p less than 0.05) was seen in both races. Race was the only independent variable that contributed appreciably (9%) to the variability of Lp(a) in serum. White children with parental myocardial infarction (n = 90) had increased levels of Lp(a) compared with those without parental myocardial infarction (22.4 versus 17.1 mg/dl, p less than 0.01). Furthermore, among white children, the prevalence of parental myocardial infarction was higher in those with Lp(a) levels of more than 25 mg/dl than in those with values of 25 mg/dl or less (9.5% versus 5.4%, p less than 0.01). In contrast, the relation of Lp(a) to parental myocardial infarction was not seen in black children. No associations were observed between parental myocardial infarction and serum levels of any of the lipids or lipoprotein cholesterol classes in children of either race. Serum Lp(a) levels may prove valuable in the assessment of coronary artery disease risk early in life among white populations. These findings also emphasize the need to evaluate the atherogenic potential of Lp(a) in different racial groups.