The transmission of schizophrenia under a multifactorial threshold model

Abstract

Family studies of schizophrenia have reported elevated rates of both definite and definite-plus-probable schizophrenia among the relatives of definite schizophrenics. These elevated rates imply a strong association between the two forms of diagnosis and suggest some form of familial transmission. Here we have used recently developed maximum likelihood methods to investigate this association and characterize the nature of the familial transmission. Results indicated that although the two forms of diagnosis were strongly related, they could not be considered alternative manifestations of a single liability distribution. Heritability estimates for either form of diagnosis were comparable (h2 = .668 +/- .052 and c2 = .191 +/- .038 for definite while h2 = .628 +/- .073 and c2 = .236 +/- .106 for definite-plus-probable), although cultural transmission (i.e., c2) was statistically significant only for definite-plus-probable. For either form of diagnosis, residual twin resemblance was statistically significant and could not be explained in terms of the effects of genetic dominance. These results are comparable to those of an earlier analysis based upon a similar data set. Finally, the statistical correction used to adjust for between-study heterogeneity in morbidity risk figures did not noticeably alter the parameter estimates.