A twin study of anxiety-related behaviours in pre-school children

Background: From middle childhood onwards, substantial evidence points to phenotypic differentiation between anxiety diagnostic categories such as generalised anxiety, separation anxiety, specific phobia, and obsessive-compulsive disorders. However, little is known about the genetics of these categories and especially about the phenotypic and genetic structure of related behaviours in pre-school children. Methods: We examined the phenotypic differentiation and genetics of mother-reported anxiety-related behaviours in 4,564 four-year-old twin pairs, from a population-based sample. Results: Confirmatory factor analyses provided support for five correlated factors: General Distress, Separation Anxiety, Fears, Obsessive-Compulsive Behaviours, and Shyness/Inhibition. Genetic influences were found on all five factors, but the pattern of influences differed considerably across them, with particularly high heritability estimates for Obsessive-Compulsive Behaviours and Shyness/Inhibition, and substantial shared environmental influence on Separation Anxiety. Multivariate genetic analyses revealed moderate genetic correlations between the five factors. Genetic overlap was particularly pronounced between General Distress and the other anxiety-related behaviours, accounting for about half of their covariance. Genetic variance on Obsessive-Compulsive Behaviours was the least correlated with the other scales. The shared environmental influences correlated highly across the factors, accounting for the greatest proportion of covariation between Separation Anxiety, Fears and Obsessive-Compulsive Behaviours. The non-shared environment influences were largely variable specific. Conclusions: These data provide evidence for phenotypic and genetic overlap as well as differentiation between aspects of anxiety-related behaviours in young children. We conclude that research with young children will benefit from more specific assessments of anxiety-related behaviours in addition to less differentiated assessments of 'internalising' symptoms.