A recent report from the U.S. Center for Disease Control’s (CDC) Autism and Developmental Disabilities Monitoring (ADDM) Network revealed that the prevalence of childhood autism spectrum disorder (ASD) is continuing to increase across all population groups, with the most recent data (2012 birth cohort) estimating a prevalence of 1 in 36 among 8-year-olds in the U.S. [1]. An important trend in the surveillance data over the past decade [2] is that the prevalence of ASD among minority children has not only “caught up” with that of non-Hispanic whites (NHW), but is now starting to outpace NHW prevalence. Among 8-year-olds in the 2012 birth cohort, the prevalence of ASD was 31.6, 24.3, and 29.3 for Hispanic, White-Non-Hispanic, and Black-Non-Hispanic populations, respectively [1]. This trend was particularly pronounced in Maryland, where a Black child was 2.0 × as likely as a white child to be diagnosed with ASD [1]. See Supplemental Table 2 for prevalence by race and ethnicity across the ADDM birth cohort.

There are a number of factors that can increase diagnostic prevalence of a predominantly-inherited condition over time and across generations. One that may differentially impact populations across disparate ancestral origins or cultures is the phenomenon of assortative (preferential) mating. Assortative mating has been repeatedly documented on the basis of greater-than-expected spousal correlations for highly-heritable autistic traits in the general population [3,4,5], and among the parents of children affected by autism [4,5]. Excess sharing of polygenic risk specific to autism has been confirmed among spousal pairs in families affected by autism in prior molecular genetic analyses [3, 6], and the degree of phenotypic correlation for autistic traits among parents of children with autism was observed to be particularly pronounced among Hispanic spousal pairs (ICC = 0.60) in a Miami, FL study [7]. Across generations, mate selection for likeness in a given trait increases diversity-of-outcomes of offspring in the next generation and in particular can increase the prevalence of offspring expressing phenotypic traits at the higher and lower extremes of the distribution for their generation.

In this analysis, we leverage a contemporary and simultaneously-ascertained epidemiologic cohort of parents in Missouri and California to provide updated information on the degree to which spousal pairs exhibit correlations in autism-related trait variation in Hispanic (California) and non-Hispanic (Missouri) US populations, to provide a contemporary context for the new information reported by ADDM, and to propose consideration of the tracking of this phenomenon as a potential influence on ASD prevalence in the US.