A path analytic model for the analysis of nuclear family data is described and used to analyze the results of two major studies of cholesterol (CH) and triglyceride (TG), the Honolulu Heart Study (HHS) of Japanese–Americans and the Cincinnati Lipid Research Clinic (LRC) study of Caucasians. The studies were first analyzed separately to assess evidence for genetic and cultural transmission, marital resemblance, and maternal environmental effects for the two plasma lipids, and then simultaneously to identify the nature and sources of any between‐study‐heterogeneity. There were significant sources of heterogeneity between the two studies for CH (only marital environmental resemblance and nontransmitted sibling environmental resemblance) and for TG (only non‐transmitted sibling environmental resemblance). The two studies were homogeneous with respect to the magnitude of genetic and cultural effects; for CH genetic heritability was estimated as h2 = .594 ± .041 while cultural heritability was estimated as c2 = .035 ± .008, and for TG the two heritabilities were estimated as h2 = .259 ± .034 and c2 = .108 ± .014. An additional bivariate analysis of the association between the two lipids revealed that all phenotypic resemblance could be explained in terms of an association of non‐transmitted residual environments with little evidence for a genetic association. The relevance of these results for an understanding of the genetic epidemiology of plasma lipids is discussed.
All Science Journal Classification (ASJC) codes
- Bivariate analysis
- Familial resemblance
- Path analysis
- Plasma Lipids