UC Davis

Introduction: The etiology of autism spectrum disorder (ASD) is unknown, though both environmental and genetic factors likely play a role in its development. Research suggests that maternal exposure to thyroid disrupting compounds during pregnancy could increase the likelihood of ASD in the child. Additionally, changes in dietary patterns among women of reproductive age may contribute to a decreased consumption of iodine, a nutrient necessary for thyroid hormone production. Severe maternal iodine deficiency and resulting hypothyroidism are well-documented predictors of neurodevelopmental disability in children, but less is known about the relationship between maternal thyroid function and ASD development or mild to moderate maternal deficiency and ASD.Objective: 1) Summarize literature on maternal thyroid dysfunction, iodine, and ASD outcomes (diagnosis and traits) and identify gaps 2) Characterize the relationship between maternal thyroid function during pregnancy and ASD outcomes 3) Characterize the relationship between maternal iodine status during pregnancy and ASD outcomes. Methods: First, scientific publications on maternal thyroid dysfunction and child ASD outcomes in humans were reviewed (through 2023) and synthesized into narrative review. Then, two original research studies were conducted on different study populations. In the first study, maternal thyroid hormone concentrations (thyroid stimulating hormone [TSH] and free thyroxine [free T4]) were measured from pregnancy blood samples collected from participants in Markers of Autism in Babies: Learning Early Signs (MARBLES), a high likelihood autism cohort study. Statistical associations between maternal thyroid dysfunction (hypothyroid conditions or hyperthyroid conditions compared to euthyroid), thyroid hormones, and ASD diagnosis or non-typical development (non-TD) compared typical development (TD) were examined. Autism traits (social responsiveness scale standardized scores [SRS]) and child cognitive ability (Mullen Scales of Early Learning [MSEL]) were also examined as outcomes. The second study included participants from the Environmental Health Influences on Children’s Health Outcomes (ECHO) Cohorts program. Individual maternal iodine intake per day was estimated using at least two pregnancy urinary iodine concentrations per mother. Mothers were classified as deficient based on the National Academy of Sciences, Engineering, and Medicine Estimated Average Requirement (EAR) and upper tolerable limit for iodine for women who are pregnant. Continuous, quintile, and binary (deficient vs. adequate) iodine intake was examined in relation to continuous standardized SRS scores and binary SRS scores (possible ASD or no ASD indicated). Results: The literature review identified multiple population-based studies that found positive associations between maternal hypothyroidism and child ASD. Relationships between hyperthyroidism or iodine status and ASD were less consistent. Misclassification bias for outcome and exposure may have occurred in studies that used medical records. In MARBLES (n=285 pregnancies from 273 mothers), 9.5% of pregnancies were affected by hypothyroid conditions and 6.7% were affected by hyperthyroid conditions. Maternal hypothyroid conditions during pregnancy were significantly associated with an increased chance of ASD compared to TD (AOR 3.17 95% CI 1.13, 8.93; p=0.03) and non-TD compared to TD (AOR 4.27, 95% CI 1.36, 13.38; p=0.01) after adjusting for confounders. Hyperthyroid conditions during pregnancy were not significantly associated with either outcome. First trimester natural log of TSH was significantly association with some percentiles of the SRS and MSEL. In ECHO, 42.5% of mothers were categorized as having deficient iodine intake with the remaining categorized as adequate intake (n=653 mother-child pairs). In adjusted quintile analysis, the 1st and 4th quintile were significantly associated with about a quarter of a standard deviation increase in SRS scores compared to the 3rd quintile (adjusted coefficient 1st quintile 2.70, 95% CI 0.56, 4.83; p=0.013; adjusted coefficient 4th quintile 2.65, 95% CI 0.52, 4.7; p=0.015). Similar results were found in the binary SRS outcome model for the 4th quintile only (adjusted coefficient 2.64, 95% CI 1.07, 6.55; p=0.04). Neither deficient iodine intake nor continuous iodine intake was significantly associated with binary or continuous SRS scores. Discussion: This analysis of MARBLES data contributed to the literature by using a sample where maternal thyroid dysfunction was assessed based on laboratory analysis in all pregnancies and all children were assessed for ASD. To our knowledge, this analysis of ECHO participants was the first to examine maternal iodine intake and ASD outcomes using a sample from the United States. Findings from MARBLES support previous findings that maternal hypothyroidism during pregnancy is associated with an increased likelihood of ASD diagnosis. Because an association was also found for non-typical development, maternal thyroid dysfunction may not be an etiological factor that is specific to ASD development. Findings from the ECHO align with previous research that also found minimal evidence of an association between maternal iodine status during pregnancy and child ASD traits. Both studies may have been limited by a small sample size, smaller number of first trimester pregnancies, and survival bias. Conclusion: Studies that include multiple urine and blood samples per trimester per person and collect information on consumption of iodine rich foods and vitamins in relation to urine sample timing would be valuable in further exploring this relationship. Because iodine deficiency is suspected to contribute to ASD development by disrupting thyroid function, additional research is needed that examines thyroid function along with maternal iodine status and child ASD. Thyroid dysfunction (especially hypothyroidism) and iodine deficiency among pregnant women should continue to be monitored along with child developmental outcomes.