Multivariate models for human genetic analysis: aggregation, coaggregation, and tracking of systolic blood pressure and weight.

A multivariate path model parameterizing the sources of familial aggregation and coaggregation of systolic blood pressure and weight, as well as their tracking across time, is applied to longitudinal data collected in Muscatine, Iowa. Genetic, common household, and individual environmental effects, pleiotropy, and a direct regression effect of blood pressure on weight are parameterized. The sample consisted of 998 individuals distributed in 261 families of whom 601 were measured on four successive occasions. The data were divided with times 1 and 2 forming group 1, and times 3 and 4, group 2. Model fitting and estimation was performed using group 1, followed by testing the model and estimates using the data in group 2. Heritability estimates for systolic blood pressure and weight were .15 and .54, respectively. The genetic correlation between these traits was nonsignificant, but there was a significant direct regression effect. The results indicate that 30% of the full-sib correlation for systolic blood pressure is attributable to the aggregation of weight. In terms of tracking, 59% and 60% of the predicted systolic blood pressure and weight correlations, respectively, were attributable to genetic effects. Testing the model from group 1 in group 2 indicates that the qualitative relationships between blood pressure and weight are stable with time.