Inherited ASD Susceptibility In Never-Diagnosed Females: Implications for Intergenerational Transmission, Gender Ratio, and the Diagnosis of Autism

Background:  Traditional diagnostic paradigms for social-communicative impairment result in the most severe 1.4% of the male population distribution being categorized as having an Autism Spectrum Disorder (ASD) in childhood.  The same paradigms result in only the most severe 0.5% of the female population distribution being categorized as ASD. 

Objectives:  To ascertain the range of phenotypic severity that encompasses inherited ASD symptomatology among females, and to explore whether latent ASD susceptibilities are as likely to be transmitted by undiagnosed sisters of individuals with ASD as by their undiagnosed brothers. 

Methods:  We analyzed parent-report data on the autistic symptomatology of siblings of children with ASD enrolled in a national volunteer register (the Interactive Autism Network, IAN, http://ianproject.org) to determine the range of standardized cutoff scores over which prevalence among females (n=854) represents a significant departure from general population prevalence.  Next we explored intergenerational transmission in 209 pedigrees of ASD-affected children in a longitudinal sibling study at Washington University in St. Louis, specifically addressing the question of whether ASD susceptibility represented by clinical-level affectation status of uncles (full biological male sibs of the parents) occurred more commonly in maternal versus paternal lines.  ASD status of uncles was reported by parents of probands, operationalized by DSM-IV diagnostic criterion endorsement, and confirmed by standardized symptom severity ratings. 

Results:  When considering a series of standardized severity cutoffs for designation of affectation status, prevalence among female siblings of ASD probands significantly exceeded general population prevalence through the 5^th percentile cutoff.  Relative risk (RR) of sisters of ASD probands exceeding a 1^st percentile cutoff was 8.0 (eight times the general population risk for girls); RR for exceeding the 5^th percentile cutoff was 2.0 (p <.001); these figures are conservative given rater contrast effects which tend to reduce severity scores of presumed-unaffected siblings in ASD-affected families.  With respect to the pedigree data, 3.1 per cent of uncles of ASD probands were reported to have clinical-level ASD symptomatology (representing at least triple the general population prevalence for ASD in adult males; p<.01)—this proportion was not significantly different between paternal and maternal lineages.

Conclusions:  Use of arbitrary, non-standardized cutoffs for case designation (as occurs when using traditional diagnostic assessments for ASD) results in marked underestimation of ASD-related-susceptibilities in females and inflates estimation of the male:female gender ratio for autism, which may, in part, be explainable by a subtle shift (0.25 SD) in the population distribution for social communicative impairments between boys and girls.  Undiagnosed mothers are at least as likely as undiagnosed fathers to transmit to their offspring that aspect of familial ASD susceptibility represented by affectation status of their own siblings. These data warrant revisions to the manner in which ASD-related susceptibility and symptomatology are ascertained and characterized among females.