Evidence for overlapping genetic influences on autistic and ADHD behaviours in a community twin sample

High levels of clinical comorbidity have been reported between autistic spectrum disorders (ASD) and attention deficit hyperactivity disorder (ADHD). This study takes an individual differences approach to determine the degree of phenotypic and aetiological overlap between autistic traits and ADHD behaviours in the general population.

The Twins Early Development Study is a community sample born in England and Wales. Families with twins born in 1994-6 were invited to join; 6,771 families participated in the study when the twins were 8 years old. Parents completed the Childhood Asperger Syndrome Test and the Conners' DSM-IV subscales. Teacher data were also collected on a sub-sample. High scores on the Conners' subscales were used to identify possible ADHD cases. Potential ASD cases were interviewed using the Development and Well-Being Assessment. Multivariate structural equation model-fitting was employed, as well as DeFries Fulker extremes analysis and liability threshold model-fitting.

Significant correlations were found between autistic and ADHD traits in the general population (.54 for parent data, .51 for teacher data). In the bivariate models, all genetic correlations were >.50, indicating a moderate degree of overlap in genetic influences on autistic and ADHD traits, both throughout the general population and at the quantitative extreme. This phenotypic and genetic overlap still held when sex, IQ and conduct problems were controlled for, for both parent and teacher data. There was also substantial overlap in suspected cases (41% of children who met criteria for an ASD had suspected ADHD; 22% with suspected ADHD met criteria for an ASD).

These results suggest there are some common genetic influences operating across autistic traits and ADHD behaviours throughout normal variation and at the extreme. This is relevant for molecular genetic research, as well as for psychiatrists and psychologists, who may have assumed these two sets of behaviours are independent.