Maternal Metabolic Conditions and ASD: Comparing Associations from High-Risk and General Populations

Background: The association between maternal metabolic dysfunction – in the form of obesity, hypertensive disorders, diabetes or specific physiologic biomarkers – and autism spectrum disorder (ASD) is one of the most consistent findings in ASD epidemiology.

Objectives: To compare associations between maternal metabolic conditions and ASD in children in the MARBLES (cohort of high-risk mothers and CHARGE population-based case-control studies in California.

Methods: Participants included 156 mother-child pairs (45 with ASD) from the MARBLES prospective cohort of pregnant women who have a biological child with ASD and followed until the child reaches 36 months; and 676 ASD cases and 467 controls from the CHARGE population-based case-control study. Metabolic conditions in both studies were ascertained using medical records and structured telephone interviews with the mother, and conditions included diabetes (type 2 or gestational), hypertension (chronic or gestational, preeclampsia), and prepregnancy obesity (body mass index ≥30). In both studies, ASD diagnosis was assessed with Autism Diagnostic Observation Schedule (ADOS) and Autism Diagnostic Interview–Revised (ADI-R); typical development was determined using the Mullen Scales of Early Learning (MSEL) and Vineland Adaptive Behavior Scales (VABS), with composite scores ≥70 on both assessments.

Results: Preliminary findings showed a higher prevalence of metabolic conditions in MARBLES compared with CHARGE among mothers whose child was diagnosed with ASD. Overall, 47% of MARBLES mothers were burdened with metabolic conditions (diabetes, hypertensive disorder, and/or obesity) in contrast to 35% of CHARGE mothers. Twenty-two percent of MARBLES mothers, compared to 9.8% of CHARGE mothers, had diabetes (mainly the gestational form). Metabolic conditions were also more prevalent in MARBLES mothers of typically developing children in contrast to CHARGE, with 38% compared to 25% affected. Time trends, population composition, and the choices in comparison groups will be evaluated and discussed.

Conclusions: These comparisons are unique in that the two analyses employ the same extensive exposure and outcome ascertainment techniques. This allows an analysis focused on the influence of study design, population and time frame on the association between maternal metabolic conditions and ASD. The CHARGE case-control design enrolled children with specific diagnoses who may or may not have had siblings, while the MARBLES enriched cohort design recruited families with a ASD-affected child and assessed neurodevelopmental outcomes in subsequent offspring. Compared with CHARGE, MARBLES pregnancies occurred later temporally; mothers were likely to be older; and children diagnosed with ASD became part of multiplex families, which are more likely to have shared genetic risk factors. Analyses restricted to overlapping birth years and adjusted for maternal education, and race/ethnicity are planned, as is a comparison of multiplex and simplex CHARGE families.