Maternal Vitamin D Levels during Pregnancy and Autism Spectrum Disorders (ASD)

Background: Vitamin D deficiency has increased concurrently with the rise in prevalence of Autism Spectrum Disorder (ASD), and evidence for vitamin D’s role in brain and immune function has led to interest in potential associations with ASD. Prior studies of ASD have examined surrogates of vitamin D or levels after diagnosis, not during critical fetal developmental periods. Published studies with Vitamin D measured during pregnancy have not been specific to ASD, but rather for broader aspects of neurodevelopment. We previously reported no association of newborn Vitamin D levels and ASD and here examine another susceptibility period.

Objectives: To examine maternal Vitamin D levels in mid-pregnancy in association with ASD in offspring, and to explore potential differences by phenotype (ASD with and without comorbid intellectual disability (ID)) or demographic sub-groups.

Methods: Data are drawn from a case-control study among children born 2000-2003 in Southern California who had banked biospecimens available from routine genetic screening. Autism cases were identified through the CA Department of Developmental Services and controls were selected from general population births in the same area and birth years. 25-Hydroxyvitamin D (OHD) was measured in serum (N=534 cases, 421 controls) collected at 14-20 weeks gestation using a sensitive assay (as OHD2 and OHD3, then summed). OHD levels were categorized as deficient (<50nmol/L), insufficient (50-74nmol/L) and sufficient (≥75nmol/L, referent category), and also examined continuously. Linearity was assessed through generalized additive models and addition of a quadratic term in logistic models. Crude and adjusted odds ratios (AOR) and 95% confidence intervals (95%CI) were calculated.

Results: OHD was deficient in 9.5% and insufficient in 25.6% of mothers, with AORs (95%CI) for ASD of 0.79 (0.49-1.3) and 0.94 (0.68-1.3) respectively. There was also no association of linear continuous OHD levels. Results were consistent for ASD with or without ID. However, when stratified by maternal race/ethnicity, non-Hispanic Whites had the expected protective association with higher OHD levels (AOR=0.92, 95%CI=0.85-0.99 per 10nmol/L OHD), but other race/ethnic groups did not. Upon further exploration, a non-linear pattern, inverted j-shape, was observed overall between OHD and ASD (adjusted beta for quadratic term = -0.0001, p=0.02), with the peak around 100 nmol/L. This pattern was similar when examining OHD3 separately or for ASD by comorbid ID. By maternal race/ethnicity, this pattern was attenuated in the non-Hispanic Whites and more apparent in “other” race/ethnic groups.

Conclusions: In one of the first large studies to examine prenatal Vitamin D levels in relation to clinical ASD, we saw no association overall with Vitamin D deficiency/insufficiency or continuous levels, in contrast to hypotheses. However a significant non-linear association was observed and high levels of Vitamin D appear to be protective, particularly among White mothers. Race differences may in part be due to variation in OHD levels, with White women having a distribution shifted to higher levels. The suggestive finding of protective effects as well at lower Vitamin D levels requires further investigation to explain. Future analyses will also examine co-exposures to organo-halogenated compounds and genetic interactions.