An approach to the multivariate analysis of high-density-lipoprotein cholesterol in a large kindred: the Bogalusa Heart Study

The genetic determination of high-density-lipoprotein cholesterol (HDL-C) levels was evaluated using segregation analysis techniques in a large multigenerational kindred with a high prevalence of coronary heart disease and myocardial infarction. Univariate segregation analysis of HDL-C levels with the effects of age and sex removed by regression provided evidence of a Mendelian mode of inheritance for a portion of the variability in HDL-C levels. Subsequent analyses included low-density-lipoprotein cholesterol (LDL-C) levels and several behavioral and anthropometric variables that affect HDL-C levels. Pedigree discriminant analysis was used to find the linear functions of the variables that maximized the likelihood given the pedigree structure and assuming monogenic segregation. The best linear function was found to be approximately equivalent to the log of the HDL-C to LDL-C ratio, with concomitant and environmental effects removed by regression. Genetic hypotheses were tested by cross-validation; linear functions derived from data on each side of the kindred were used to test hypotheses on the other side of the kindred. On one side of the kindred, all hypotheses were accepted. On the other side of the kindred, only Mendelian inheritance of the linear function was indicated. Segregation of the age- and sex-adjusted HDL-C values, and of the linear function, was evaluated using a regressive model that allows for intrafamilial correlations in addition to a monogenic effect. All analyses provide evidence for levels of HDL-C being controlled by a major locus with neither dominant nor recessive expression.