Maternal Thyroxine and Fetal Brain Development: The Latest Chapter, a Look Back, and Considerations for the Future

Abstract

Defining the relationship between maternal hypothyroxinemia during early pregnancy (free T4 [FT4] 10th centile) and offspring neurocognitive development is an ongoing challenge to those exploring the biology of maternal-fetal interactions in search of opportunities for improving neurocognitive function. Although not yielding a definitive characterization of this relationship, the report by Finken et al (1) in this issue of the JCEM represents an important step forward. Studies of this type are labor-intensive and require a considerable investment of time and organizational skill. The authors enrolled a large cohort of women during early pregnancy, collected baseline demographics, performed thyroid testing, and then monitored outcome as part of the Amsterdam Born Children and their Development (ABCD) study. When offspring of those women reached the age of 5 years, the investigators measured various aspects of neurocognitive function via a computerized assessment program. Application of a computerized measurement tool in this context allowed all responses to be determined objectively. Several features of this study add confidence to its conclusions: 1) the study was community-based; 2) TSH, FT4, and thyroperoxidase (TPO) antibody measurements were performed early in pregnancy; 3) children were tested at an age when results are reliable; 4) numbers of children tested were sufficient to detect important differences; 5) testing was blinded and carried out by trained investigators in a school setting; 6) the computerized assessment tool provided objective performance measurements; and 7) the ABCD study collected extensive demographics and biophysical data that were available for use in regression analyses. Computerized assessment consisted of 4 tasks. The 175 children of hypothyroxinemic mothers performed significantly less well only on the first task (baseline speed, a simple reaction time test that required minimal cognitive effort). The other 3 tasks measured visuomotor coordination, inherent response inhibition, and response flexibility. On these latter 3 tasks, children of hypothyroxinemic mothers did not perform differently from the remaining children. For the first task, however, their mean response speed was 41.3 milliseconds slower than that of 1584 children whose mothers’ FT4 levels were above the 10th centile. Each child performed this task 32 times with each hand, allowing the standard deviation of response speed to be measured as well. This was also significantly greater, indicating more instability in performing the task repeatedly. The difference in response speed remained significant after adjustment for multiple variables. Furthermore, the slowing of response speed was more pronounced when limited to children of mothers with FT4 levels fifth centile. Finally, mean reaction time was even longer among 44 children of mothers whose FT4 measurements were 10th centile and whose TSH measurements were 90th centile. Given that children of hypothyroxinemic mothers performed the other 3 tasks satisfactorily, the longer term developmental implications are unclear, although documentation of the slower response speed was convincing. In addition to reporting neurocognitive data, this study highlights maternal characteristics that segregate at the population level with low FT4. The higher mean body mass index among hypothyroxinemic women, for example, has been documented in several other studies (2–6),