Differential Enrollment in Twin Registries: Its Effect on Prevalence and Concordance Rates and Estimates of Genetic Parameters

Abstract

In the NAS-NRC Registry, all major diseases are more common in DZ than in MZ twins. Furthermore, concordance rates for most disorders are lower in the registry than would be expected. In this article we propose a general model which seeks to explain these phenomena. The model explores the impact of traits which increase or decrease the probability of enrollment of individuals given that the registry, like the NAS-NRC, includes only pairs where both members are enrolled. If the trait decreases the probability of selection into the registry, both the prevalence of and concordance for the trait in the registry will be lower than that found in the population. A trait which increases the probability of selection has the opposite effects. However, the magnitude of these effects are a function of the population concordance. If population concordance differs in MZ and DZ twins, the effect of differential enrollment will not be the same for the two zygosity groups. The article examines the impact of differential enrollment on estimates of heritability and common environment and explores ways in which estimates of prevalence and concordance rates can be obtained which are free of the bias introduced by selection.