Meconium androgens are correlated with ASD-related phenotypic traits in early childhood

Prenatal exposure to increased androgens has been suggested as a risk factor for autism spectrum disorder (ASD). This hypothesis has been examined by measurement of steroids in amniotic fluid and cord blood, with mixed results. To provide an orthogonal measure of fetal exposure, this study used meconium, the first stool of a newborn, to measure prenatal androgen exposure from infants in the Early Autism Risk Longitudinal Investigation (EARLI). EARLI is a familial-enriched risk cohort that enrolled pregnant mothers who already had a child with an ASD diagnosis. In the younger child, we investigated the association between meconium unconjugated (u) and total (t) concentrations of major androgens testosterone (T), dehydroepiandrosterone (DHEA) and androstenedione (A4), and ASD-related traits at 12 and 36 months of age. Autism traits were measured at 12 months with Autism Observation Scale for Infants (AOSI) and at 36 months with total score on the Social Responsiveness Scale (SRS). 137 children (61 males, 76 females) had data on both outcomes and meconium androgen measurements. Separate robust linear regressions between each of the log-transformed androgens and log-transformed AOSI and SRS scores revealed three-way interaction between sex of the child, sex of the proband, and androgen concentration. In the adjusted analyses; t-T, u-A4, and u-DHEA (P ≤ 0.01) were positively associated with AOSI scores while u-T (P=0.004) and u-DHEA (P=0.007) were positively associated with SRS total score among females with female probands. Additionally, higher concentrations of u-T (P=0.01) and t-T (P=0.01) predicted higher SRS total score in males with male probands.