Extensions to multivariate normal models for pedigree analysis. II. Modeling the effect of shared environment in the analysis of variation in blood lead levels

Abstract

A multivariate normal model for pedigree analysis is applied to blood lead measurements from 617 individuals in 80 families in Melbourne, Australia, studied in 1977-1978. A new method is introduced for estimating time dependence of the family covariance matrix for blood lead levels; this time dependence can be interpreted as arising from the effects of common family environment on blood lead levels. Methods for the testing of assumptions and detection of outlying pedigrees and outlying individuals are applied. No correlation between blood lead levels of spouses was observed, but an effect of shared family environment was suggested by the difference between an estimated sibling correlation of about 0.5 for young sibling pairs living together and of about 0.1 for older siblings no longer living together. As there was no significant polygenic additive effect, the non-zero correlation between older siblings is more likely to be due to continuing effects of (environmental) factors shared in youth, rather than to a polygenic dominant effect. It is estimated that smoking 20 cigarettes per day is associated with an increase of about 12 per cent in blood lead level.