Maternal Hypothyroxinemia and Autism
Abstract & Commentary
By Sotirios Keros, MD, PhD
Instructor, Department of Pediatrics, Division of Pediatric Neurology, Weill Cornell Medical College
Dr. Keros reports no financial relationships relevant to this field of study.
Synopsis: Very low levels of early-maternal free thyroxine are associated with a 4-fold increased risk of parental reports of autism symptoms in their offspring.
Source: Román GC, et al. Association of gestational maternal hypothyroxinemia and increased autism risk. Ann Neurol 2013 Aug 13; doi: 10.1002/ana.23976. [Epub ahead of print].
Autism spectrum disorder (asd) represents a somewhat diverse set of conditions that are linked by key clinical features. Currently, more than 100 genes have been identified that cause or are otherwise potential risk factors for autism. In addition, it is clear that there are many environmental factors that affect ASD risk. It has long been known that congenital hypothyroidism and maternal hypothyroidism result in cognitive disabilities. Relatively recently, it has been suggested that maternal hypothyroxinemia during the first trimester negatively affects cognitive development in humans, and that transient maternal hypothyroxinemia in animal models leads to alterations in brain structure and function that resemble those seen in human ASD.
In this study, the authors evaluated maternal thyroid function during pregnancy and parental report of autism symptoms taken from the Netherlands-based Generation R Study. The study enrolled 8870 pregnant women who delivered between 2002 and 2006. Of these, 4039 maternal-child pairs met the study inclusion criteria of: consent for the study; maternal thyroid measurements consisting of free thyroxine (fT4), serum thyroptropin (TSH), and thyroid peroxidase antibodies before 18 weeks of gestation; and information about autistic symptoms in the child through age 6. Mild hypothyroxinemia was defined as an fT4 level less than 10th percentile (n = 295; fT4 < 11.82 pmol/L) with an otherwise normal TSH, and severe hypothyroxinemia was defined as less than the fifth percentile (n = 136; fT4 < 10.99 pmol/L). In addition, the subset of women who fell between the 5th and 10th percentile were labeled as only mild hypothyroxinemia (n = 159; 10.99 < fT4 < 11.82 pmol/L). Autistic symptoms in children were assessed by the parents as reported on the Pervasive Developmental Problems (PDP) subset of the Child Behavioral Checklist for Toddlers as well as the Social Responsiveness Scale (SRS). Children with a PDP score above the 93rd percentile were classified as having borderline PDP, while those above the 98th percentile were labeled as clinical PDP. Children who had a PDP score of greater than 98th percentile as well as an SRS score in the top 5% were classified as a probable autistic child. Statistical models were adjusted for sex, ethnicity, gestational age at birth, birth weight, education, smoking history, parental age, thyroid medication during pregnancy, and pregnancy folate and CRP.
The study found that early maternal severe hypothyroxinemia was associated with a four-fold risk (odds ratio [OR], 3.89; 95% confidence interval [CI], 1.83-8.20) of having a probable autistic child. When looking solely at PDP scores, severe hypothyroxinemia conferred an OR of 2.02 (95% CI, 1.16-3.51) and 2.6 (95% CI, 1.30-5.18) on the development of borderline and clinical PDP, respectively. Severe hypothyroxinemia was also associated with increased SRS scores. Outcomes were similar when comparing the mild hypothyroxinemia group with the only mild hypothyroxinemia group and, thus, the authors did not find a statistically significant dose-response relationship, as assessed by the risk of developing borderline PDP or clinical PDP. In addition, there was no dose-response relationship between hypothyroxinemia and scores on the PDP subtest and SRS as measured by linear regression. TSH levels and TPO antibodies during pregnancy were not correlated with autistic features.
Commentary
This study provides strong evidence for an association between low fT4 levels in pregnancy and autism. Although this study cannot prove causation, it strengthens the link to this modifiable risk factor and gives support to a potential policy of testing and treatment of otherwise euthyroid maternal hypothyroxinemia during early pregnancy. One limitation of the study is that the children did not have clinically diagnosed autism. But the parental reporting scales used are well validated and highly sensitive and specific for a clinical diagnosis. Also, more than half of the original enrollees in the study were not ultimately included due to loss to follow-up or lack of autism reports and, thus, these results are subject to selection bias. However, the large sample size and population-based birth cohort design otherwise help to provide additional evidence to the hypothesis that thyroxine is critical for normal fetal brain development early in gestation.