Correlation of Autism Traits in Families of Children with Autism Spectrum Disorders

Background: Prior research has indicated that autism spectrum disorder (ASD) is strongly heritable and that there are valuable lessons to be learned by studying family members of individuals with ASD.  The Social Responsiveness Scale (SRS), developed by Constantino and colleagues (2000) is a quantitative measure of ASD traits. Constantino and Todd (2005) found intergenerational correlations of these traits (on the order of 0.4) between children and their parents, and between the parents themselves, in a non-clinical sample. 

Objectives: It has yet to be determined if behavioral correlations between parents and between children and parents are also observed in a sample of families where at least one child has received a diagnosis of ASD.  The present investigation is designed to address this question using a unique and comprehensively characterized sample that includes children with ASD as well as their unaffected siblings, mothers and fathers. 

Methods: The three primary inclusion criteria for the current study were that families must (1) be biologically related, (2) live in the same household and (3) include at least one child who had received a diagnosis of an ASD (autism, pervasive developmental disorder-not otherwise specified or Asperger’s disorder), confirmed by the ADOS and/or ADI-R.  As of December 2008, sixty-seven families were included in the present sample, with a total of 163 participants.  This was comprised of 45 mothers, 43 fathers, 50 affected probands (37 male, 13 female) and 25 unaffected siblings (12 male, 13 female).  At the time of participation, parents ranged in age from 25 to 54 years (M=40.70 years) and children ranged in age from 3 to 24 years (M=7 years).  Data characterizing autism traits were collected on all family members (mother report on father, father report on mother, mother report on children) via the SRS (Constantino et al., 2000), a questionnaire measure including 65 items and yielding raw scores between 0 and 195.  The standardized version of the measure for children was used as well as a research version for adults (SRS-Adult Research Version; Constantino, 2005), and another for young children (SRS-Preschool Version; Constantino, 2006). 

Results: The mean raw SRS score for mothers was 36.20 (SD=23.64), fathers M=32.40 (SD=25.36); unaffected siblings M=21.44 (SD=16.63) and probands M=96.92 (SD=32.90).  Intraclass correlation coefficients were estimated on the following six pairings: mother-father, mother-proband, mother-unaffected sibling, father-proband, father-unaffected sibling and proband-unaffected sibling.  Note that not all pairings were available for all families as not all questionnaires have been returned yet.  We found significant correlations between mothers and fathers (n=39), ICC = .28 (p<.05) and between fathers and unaffected siblings (n=13), ICC = .45 (p<.05).

Conclusions: Even in this small preliminary sample, significant associations emerged between members of families of children with ASD.  These findings indicate that – as has been reported in a non-clinical population – there may be intergenerational and assortative mating effects on ASD symptoms in unaffected family members of children with ASD. As this sample collection is ongoing and expanding, these findings will be updated with the addition of many more families over the upcoming months.