Recent data indicate that the familial aggregation of congenital cardiovascular malformations (CCVM) varies with the type of defect in the index case. Using a logistic regression model that allows for dependence among family members, we calculated the risk of any CCVM to case relatives compared with relatives of controls. Data from 3,908 first-degree relatives of 570 matched cases and controls identified from 1981 through 1985 by the Baltimore-Washington Infant Study were used in the analyses. Overall risk for any CCVM in case relatives was increased four-fold over that of control relatives. While relatives of cases with flow lesions (including right and left heart defects, as well as perimembranous ventricular septal defect [VSD]) had a five-fold increase in risk, the risk to relatives of nonflow lesion cases did not differ significantly from the risk to relatives of controls. Sex, maternal age, miscarriage history in the mother, and birth order had no apparent effect on risk among siblings. However, there was an indication of increased risk in relatives of nonwhite cases with VSD compared to relatives of matched controls. However, with these data it was not possible to distinguish between environmental and genetic sources of this familial aggregation.