Abstract
The characterizing features of autism spectrum disorder (ASD) are continuously distributed in nature; however, prior twin studies have not systematically incorporated this knowledge into estimations of concordance and discordance. We conducted a quantitative analysis of twin–twin similarity for autistic trait severity in three existing data sets involving 366 pairs of uniformly-phenotyped monozygotic (MZ) twins with and without ASD. Probandwise concordance for ASD was 96%; however, MZ trait correlations differed markedly for pairs with ASD trait burden below versus above the threshold for clinical diagnosis, with R2s on the order of 0.6 versus 0.1, respectively. Categorical MZ twin discordance for ASD diagnosis is rare and more appropriately operationalized by standardized quantification of twin–twin differences. Here we provide new evidence that although ASD itself is highly heritable, variation-in-severity of symptomatology above the diagnostic threshold is substantially influenced, in contrast, by non-shared environmental factors which may identify novel targets of early ASD amelioration.
Highlights
In any disease, the concordance rate for identical twins reared together serves as a foundational parameter for estimating heritability
Among the monozygotic twin pairs affected by autism spectrum disorder (ASD) in this sample, the categorical probandwise concordance was 0.91 for the Autism Genetic Resource Exchange (AGRE) sample, 0.61 for the Interactive Autism Network (IAN) sample, and 0.82 for the combined clinical monozygotic twin sample
The equivalence in the range of differences identified in the IAN and AGRE samples makes it likely that the MZ twin concordance rate reported by parents for community diagnosis within IAN is an underestimate, since AGRE pairs with comparable levels of difference presumed discordant by their parents were more commonly confirmed to be concordant for ASD
Summary
The concordance rate for identical twins reared together serves as a foundational parameter for estimating heritability. The discovery, that autistic symptomatology exhibits a fully continuous distribution in the general population (Constantino 2011)—and that the distribution reflects highly overlapping causal influence across the range from sub-clinical to clinical affectation (Constantino et al 2010; Robinson et al 2011)— raises critical questions about the extent to which calculations of MZ concordance may be underestimated, and whether previously-reported heritability estimates for sub-clinical severity (Constantino and Todd 2003) extend to measured variation in severity in the clinical range of affectation This was preliminarily addressed in a population study that implemented diagnostic observations among screen positive MZ twin pairs involving a small number of affected pairs (n < 20) and a substantial number of unaffected pairs (Colvert et al 2015), among whom severity correlations were observed to be strong in both groups. This observation contrasted with an earlier report of dissimilarity in symptom burden between clinically-affected identical
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
