Examining Sex Differences in Adaptive Behavioral Development in High Risk Infants with ASD, Social Communicative Delay, and Typical Development

Background:  There is notable phenotypic heterogeneity among high-risk (HR) siblings of children with autism who do not meet ASD diagnostic criteria. A substantial proportion of non-ASD HR siblings manifest a profile of subthreshold ASD symptoms and/or low developmental level compared to low-risk (LR) children (Messinger et al., 2013). More work is needed to understand heterogeneity in HR siblings and to potentially elucidate sex-specific differences, if present. Under-representation of females in research may obfuscate female protective effects or sex-specific ASD and subclinical social-communicative development.

Objectives: To determine whether sex differentially influences behavioral trajectories in ASD, social-communicative concern, and typical outcomes in a sample of HR infant siblings while controlling for genetic loading in families.

Methods:  We examined 334(137F) HR and 123(55F) LR infant siblings, assessed at 6, 12, and 24-months. ASD classification at 24 months was based on CBE using DSM-IV criteria and by meeting ADOS criteria (HR-ASD, N=72(15F)). HR siblings with 24-month receptive or expressive language Mullen Scales of Early Learning scores one SD below the mean and/or ADOS social affect subscale score >5 were classified as HR-SocCom (N=87(36F)) to capture subthreshold, yet concerning social-communicative development. The remaining HR infants (N=175(86F)) are referenced as HR-TD. No LR infants had an ASD diagnosis. Experimentally controlling for genetic loading, the HR sample was restricted to participants with male probands, except HR-ASD females (as a female containing family may reflect different inheritance pattern). Five LME models were fit to each Vineland Adaptive Behavioral Scale (VABS) subscale and summary score. Predictors of interest include age, group, sex, and interactions. Benjamini-Hochberg correction was applied to pairwise group comparisons.

Results:  Infants are more likely to be male in the HR-ASD group as compared to the HR-TD (RR=1.56, p=.00005) and the HR-SocCom (RR=1.35, p=.006). We observe equal sex proportion in the HR-SocCom and HR-TD groups (RR=1.15, p=0.241). All best-fit LME models of VABS subscales and composite score, except motor skills, include age, group, sex, and group-by-age effects. The motor skills best-fit LME model included age and group. Group-by-sex interactions did not improve fit in any model. 6-month ABC scores revealed the following pattern of group differences (LR-TD>HR-TD>HR-ASD, whereas the HR-SocCom group did not differ from the HR-TD or the HR-ASD group). 6-month motor scores revealed the following pattern of group differences (LR-TD>HR-TD=HR-SocCom>HR-ASD). Trajectories across the interval differentiated the groups in a consistent manner in the ABC, communication, and socialization domain, such that LR-TD=HR-TD>HR-SocCom>HR-ASD.

Conclusions: Experimentally controlling for proband sex, we did not observe specific sex effects on group trajectories of adaptive behavior from 6-24 months of age. Pairwise group differences revealed a gradient of attenuated adaptive behavioral development across groups. Relative risk discrepancy between HR males and females is maintained in ASD but not social-communicative concern, indicating sex may protect from diagnostic, but not subthreshold levels of concern. These data are consistent with previous reports (Constantino & Todd, 2003; 2005) suggesting that the M:F sex ratio differs by degree of affectation.