No major effect of twinning on autistic traits.

BACKGROUND It has been questioned whether the process of twinning might be a risk factor for autism spectrum conditions (ASC) and autistic traits. AIM We sought to determine whether autistic traits and probable disorder, as measured by the Childhood Autism Spectrum Test (CAST), were more pronounced in twins compared to singletons. SAMPLES Data were analyzed from two large population-based samples of UK children, twins (n = 5,142 twin pairs, aged 8 years) and singletons (n = 2,805, aged 5-9 years). RESULTS Distributions of CAST scores in both groups were negatively skewed and scores for twins were more variable than singletons. Mean CAST total scores and standard errors (SE) were not significantly different for twins (5.1; SE 0.04) compared to singletons (4.9; SE 0.08). Moreover, contrary to expectations, the likelihood of scoring above the threshold for possible ASC was significantly lower in the twins than the singletons (OR = 0.69; P = 0.002). Subsidiary analyses of CAST scores according to sex, twin type, and subscale scores representing the subdomains of autism found a few significant differences (P<0.01), but the effect sizes for these differences were small and none exceeded η(2) = 0.005. The explanation for these small differences remains obscure, but the very small effect sizes mean they are of little importance. CONCLUSIONS Our results do not provide evidence to support twinning as a risk factor in the development of autistic traits.