Lowering the risk of autism spectrum disorder with folic acid: can there be too much of a good thing?

See corresponding article on page 1432.

Autism spectrum disorder (ASD) affects 1% to 2% of American children (1). Its incidence may be increasing, stimulating investigations of possible causes. ASD has a strong genetic component, estimated to be 50% or greater (2); epigenetic alterations and nutrition have been suggested as contributing factors. Micronutrients such as B vitamins are of particular interest because of the ways in which they could interact with genetic or epigenetic factors. Folate is especially relevant because of its role in pre- and postnatal brain development.

Many studies have explored the influence of maternal folic acid (FA) intake on the risk of ASD in their offspring. There are 2 reasons to think that folate could be important in preventing ASD. First, maternal FA supplementation in the periconceptional period protects against the occurrence of neural tube defects, which are also neurodevelopmental disorders. Second, 1-carbon units derived from folate pathways are the main source of methyl groups for essential DNA and histone methylation reactions. Inadequate availability of such methyl groups could potentially affect gene expression, as suggested in studies exploring folate-related risk factors for ASD (3). In fact, most reports find that FA protects against ASD (4).

Paradoxically, given the known importance of folate for normal development, there is some concern that there could be adverse effects of high exposure. These could result from high FA intake or unmetabolized folic acid (UMFA) in serum. The latter occurs when FA is ingested in larger doses than the intestinal or hepatic enzyme dihydrofolate reductase can reduce to the natural, biologically active form (5). A large Spanish prospective study found that periconceptional intake of FA supplements > 1000 µG/day was negatively associated with some markers of cognitive development at age 4 (6). Most of the markers showed no significant association, however, and the positive findings could be due to chance given the large number of statistical comparisons reported. The Boston Medical Center cohort study of 1257 mother-child pairs found that moderate self-reported maternal prenatal multivitamin use was protective against ASD, but both low (≤2 times per week) and high (>5 times per week) supplement use was associated with an increased risk of ASD. It is noteworthy that these findings were not specific to folate, as vitamin B12 concentrations were also associated with increased risks of ASD. The authors described this as a hypothesis-generating study. (7). A follow-up study from the same cohort examined plasma UMFA in cord blood from the infants and showed a significant association between higher UMFA and ASD in Black children but not in the group overall. The authors appropriately concluded that the study “raises more questions than it could answer” (8).

In the current issue of the American Journal of Clinical Nutrition, Husebye et al. (9) contribute to this debate by presenting data on UMFA in pregnant women receiving antiseizure medications (ASM) from a large Norwegian Mother, Father and Child Cohort Study (MoBa). The authors report that there were no adverse neurodevelopmental effects of UMFA exposure in utero in children of mothers receiving ASM for epilepsy. Such women are often told to take high doses of FA, either because some ASMs interfere with folate metabolism or in hopes of decreasing the risk for neural tube defects. MoBa has been used previously to examine several related topics, including a report that FA was not associated with ASD in the full MoBa cohort (10).

This study has several strengths. MoBa has excellent data on FA exposure, measured UMFA, maternal use of ASM, and reported autism scores and language outcomes in the offspring (N = 227) of study pregnancies. The children were followed up for 8 years, although there was a substantial dropout rate. Many women were taking high doses of FA (4 mg or more), making it more likely that an adverse effect would have been uncovered if there were one.

It is important to note that several factors influence how much UMFA is present in the serum, particularly the dose ingested and the time since ingestion. Also, genetic variation may affect the efficiency of dihydrofolate reductase. Unfortunately, information on the time since last FA exposure was not available. Thus, it is difficult to determine whether a high concentration of UMFA reflects a low recent exposure or a high exposure at an earlier time.

Does either high FA intake or UMFA have an adverse effect on neurodevelopment? Few studies found FA intake above the upper level to be beneficial, but there is little evidence that it is harmful either. UMFA represents no more than 1%–2% of the total folate concentration in serum. Animal studies suggest adverse effects of high FA intakes on offspring developmental processes, but no consistent adverse effects have been reported in humans. Moreover, UMFA concentrations are likely to be higher in those ingesting larger quantities of FA, making it difficult to separate the effects of the 2 (5).

What can we then conclude? Normal doses of FA have been shown to have a beneficial effect on development. A biological mechanism for an adverse effect has not been persuasively demonstrated, but the effect of UMFA inside the cell is largely unknown. Overall, the 2 studies on UMFA to date are contradictory, which speaks against a serious harmful effect, but the studies have serious limitations. Future investigations should focus on women exposed to high doses of FA during pregnancy and on the basic science of UMFA at the cellular level.

Notes

This work was funded by the Intramural Research Program, Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH.

Abbreviations used: ASD, autism spectrum disorder; ASM, anti-seizure medications; FA, folic acid; UMFA, unmetabolized folic acid.