Quantitative Autistic Trait Aggregation in Siblings of Autistic Probands in 1,246 IAN Families: Further Support for Differential Genetic Transmission of Simplex and Multiplex Autism

Background: Previous research has suggested that simplex and multiplex autism may involve divergent mechanisms of inheritance.  In a previously published study involving 210 multiplex autism families and 80 simplex autism families, we observed differential patterns of familial aggregation of quantitative autistic traits (QAT) for males in simplex versus multiplex autism. 

Objectives: To explore the aggregation of QAT in an independent sample of 1,101 self-identified simplex and 145 self-identified multiplex autism families, encompassing a total of 2,772 children in autism-affected families.

Methods: All data was acquired through the Interactive Autism Network (IAN), an internet-based volunteer register for U.S. families, for which eligibility includes:  a) having at least one child diagnosed with an autism spectrum disorder (ASD) by a professional; and b) a full biological parent or legal guardian willing to participate.  For this analysis we included families in which there existed: 1) a 4-18 year old ASD proband; 2) at least one full biological sibling in the same age range; and 3) QAT measurement for the proband and sibling(s) completed by parent-report using the Social Responsiveness Scale (SRS).  2,760 of these children were also rated by their parents using the Social Communication Questionnaire (SCQ).  

Results: SRS and SCQ total scores were moderately correlated (ICC 0.62 for affected children), supporting substantial correspondence between quantitative severity and DSM-IV criterion endorsement.  Principal components factor analysis of SRS data revealed a unitary factor structure, supporting use of singular (total) SRS scores in the data analyses.  The QAT distribution for children in self-identified simplex autism families was distinctly bimodal, with a nadir in the distribution at an SRS T-score of 60, and a cluster representing unaffected siblings tightly grouped about a mean of 20.0.  Of note is that 10 per cent of the children deemed unaffected by their families fell in the “pathological” cluster of the bimodal distribution.  Although a bimodal distribution was also observed for females in multiplex families, the distribution for males in multiplex families was unimodal, i.e. without an appreciable cluster of unaffected children.  The respective means for all boys whose SRS scores fell below a uniform quantitative severity threshold (60T) differed by 0.5 SD between simplex and multiplex families (t=-3.85; df=633; p <0.001).  Affected children in both groups exhibited a wide, continuously-distributed range of severity. 

Conclusions: We observed clear evidence of familial aggregation of continuously-distributed QAT in males in multiplex families, however most unaffected individuals in self-identified simplex families appeared to be devoid of aggregation of such traits—in essence they appeared categorically unaffected.  A remarkable feature of the distribution of QAT in this large sample of simplex ASD families was the location of the notch in the bimodal distribution of parents’ reported scores (the threshold at which 90% of families differentiated affected from unaffected children), which fell only 1.5 SD above the general population mean. These results support a hypothesis of differential genetic transmission of ASD in simplex versus multiplex families, and also indicate that traditional methods for designation of affected status may result in substantial underestimation of sibling recurrence risk.