Familiality of Social Responsiveness Scale Scores in the Nurses' Health Study II

Background:  

The Social Responsiveness Scale (SRS) is a widely used measure of autistic traits and social functioning. Familiality of SRS scores has been suggested, but how scores associate with factors related to autism spectrum disorders (ASD) has not been fully examined in large population-based studies.

Objectives:  

To examine familiality of SRS scores, assess associations with child comorbidities, and explore whether adjustment for parent SRS scores, as a measure of underlying genetic susceptibility/ broader autism phenotype (BAP), altered associations with certain risk factors for ASD. We also examined whether these factors were related to ASD severity as measured by child SRS scores.

Methods:  

Participants were part of a nested case-control study within the Nurses’ Health Study II. SRS forms were mailed to participants; mothers (nurse participants) completed child forms and spouses completed parent forms. Familiality was assessed through Pearson correlation coefficients for parent-child scores and by t-tests comparing child scores according to elevated parent scores (defined as the top 20% of the distribution). Child scores according to maternal report of comorbid neuropsychiatric conditions were compared by t-test. T-tests stratified by case status were used to compare scores based on: demographic factors (education, income, parental age), maternal obesity, gestational diabetes, and depression. Logistic regression was used to obtain estimates of the association between these factors and offspring ASD adjusted for parental SRS scores and other potential confounders.

Results:  

470 cases and 1,647 controls returned SRS forms (representing over 70% of the follow-up study). Case fathers had significantly higher SRS scores than control fathers (p<.0001). Parent scores were correlated in both cases and controls (r in total study group =.32), suggesting assortative mating. Child scores were higher for those with parent-reported neuropsychiatric conditions in controls, though still within the normative range; in case children, only maternally-reported comorbid obsessive-compulsive disorder was associated with higher child scores. Parent elevated scores in controls were associated with significantly higher child scores (p <.0001) relative to those with parent scores in the lower 80% of the distribution, corresponding to a shift in child raw score of approximately 15 points; the effect was stronger when the father had the elevated score, when the child was male, or when both parent scores were elevated. Elevated scores in case parents were not associated with concordant increases in child scores, though case parents were significantly more likely to have concordantly elevated scores (both parents in top 20% of distribution) than were control parents (p=0.0008). This association persisted in adjusted analyses (odds ratio for parent concordantly elevated scores comparing cases to controls=2.09, 95% CI 1.22, 3.59). Maternal depression was associated with increased mother and child SRS scores; other factors were not associated with scores. In models adjusting the previously identified risk factors for parental SRS scores associations were somewhat attenuated.

Conclusions:  

Our results demonstrate familiality of autistic traits, as measured by the SRS, and suggest associations in this cohort with previously identified ASD risk factors are not largely impacted by parental BAP.