Multiple threshold models for the affective disorders: the Yale-NIMH collaborative family study

From the Yale-NIMH collaborative family study, the 1482 first degree relatives of 90 bipolar 1, and 163 major depression probands were examined to test the hypothesis that bipolar 1 and major depression are due to a single underlying genetic liability. We attempted to fit multifactorial-polygenic and single-major-locus multiple threshold models for sex and severity to the relatives in the major depression and bipolar 1 families. With relatives classified as affected only if they met criteria for major depression or bipolar 1, there was at best only marginal support for these models. Differences between these and previously reported results were examined in relation to differences in underlying assumptions. Additional analyses of these and other data from families of NIMH bipolar 2 and schizoaffective probands suggest that different methods of age adjustment, the relative placement of bipolar 2 and schizoaffective disorders in a hypothesized liability continuum and the inclusion or exclusion of sex thresholds were not primarily responsible for differences in the fit of genetic threshold models. Factors which do appear to be important are variability in rates between samples, the possibility of genetic heterogeneity, and the presence of a large secular increase in affective illness over the past three generations; the secular trend cannot be accommodated in the models of genetic transmission examined.