Understanding the Brain Manifestations of Hypothyroidism in Adult Patients

With the advent of advanced MRI technology to study the impact of non-neurological diseases on the brain, thyroidologists worldwide have sought to identify the structural and functional brain manifestations of hypothyroidism in children and adults. The neuroimaging literature on adult hypothyroid patients has, however, yielded inconsistent and often confusing results. ^{1 –3} This is in part because previous research has typically relied on single imaging modalities and, with a few exceptions, ⁴ most have not mapped the findings to behavioral sequelae of hypothyroidism.

In the current issue of Thyroid, ⁵ Su et al. have taken this line of inquiry a step further. They used a combination of two MRI platforms to assess both brain structure and brain function in adults with newly diagnosed untreated hypothyroidism, based on thyrotropin measurements above the reference range and free thyroxine below. The sample included 44 hypothyroid adults attending a single hospital in Gansu China and 54 well-matched healthy community-based controls, all of Han ethnicity.

Their cognitive abilities were evaluated using a quick screening test for dementia, specifically the Montreal Cognitive Assessment (MoCA), and their emotional status through validated depression and anxiety questionnaires (all translated into Chinese). Participants underwent MRI scanning and had blood drawn for thyroid hormone and lipid measurements.

Three protocols were used to analyze the MRI scans. One, performed on structural scans, involved deformation-based morphometry (DBM), a technique that compares the gray matter content of discrete regions within a person's brain with those from a “normal” templated brain provided by the imaging software. Patient and control groups were compared for their degrees of deviation within regions. The second technique used functional MRI to identify brain regions that are active during rest, known as resting-stage functional MRI (rs-fMRI). Third, functional connectivity (FC) analysis was performed on the rs-fMRI data to ascertain regions of coactivation thought to represent discrete neural networks. For each network, the “seed” or starting node was based on DBM results.

Su et al. observed that the hypothyroid group had worse performance than controls on the MoCA indices and was also more likely to be depressed and anxious. DBM analysis demonstrated five cortical regions with significantly reduced gray-matter volumes in the patient group. Two regions located in the right hemisphere were the superior frontal gyrus and superior temporal gyrus, and three in the left hemisphere were the middle frontal gyrus, dorsolateral prefrontal gyrus, and supplementary motor area. The DBM analysis also identified three regions with enlarged volumes in the hypothyroid group, namely the right and left Crus I regions of the cerebellum and the left precentral cortical gyrus.

The FC analysis showed patients had three sets of highly coactivated regions, each connecting the Crus I part of the right cerebellum to the left inferior frontal gyrus (triangular part), left precentral gyrus, and left angular gyrus of the parietal lobe. Through a further analysis, the network showing the strongest connectivity in patients was the Crus I—angular gyrus connection. In patients, better MoCA language scale performance was positively correlated with larger left supplementary motor and precentral gyrus areas. Thyroid hormone levels were unrelated to any of the clinical or MRI indices.

There are three main take-away messages from this study. First, the DBM analysis showed that the brain regions involving reduced volumes in hypothyroid patients are those known to support abilities that are often compromised in these individuals. These abilities include emotional control, speech, attention, working memory, and graphomotor skill. Second, the three enlarged regions in patients found through DBM suggest a compensatory response to offset disease sequelae such as the sluggishness these patients commonly report. For example, their enlarged precentral cortex volumes may have arisen from the extra effort and attention they need to carry out tasks such as writing or drawing, which depend on this cortical region.

Third, the FC analysis identified three neural networks that were more strongly activated in patients than controls. Each network connected the right Crus I region of the cerebellum to left hemisphere regions that support language (pars triangularis), motor (precentral gyrus), and emotion–regulation functions (angular gyrus). Based on recent cerebellum studies showing that in addition to its traditional role in motor coordination and balance, the cerebellum serves as backup for some neuropsychological functions, particularly when a cortical region is defective. ⁶ Could patients be needing this resource to carry out these functions more than controls and could this additional neurocircuitry engagement contribute to the fatigue and brain fog ⁷ seen in some hypothyroid patients?

There are several noteworthy limitations of this study. First, the MRI tools and analytic techniques, while state-of-the art, did not cover all brain aspects known to be affected by hypothyroidism. The authors did not investigate the subcortical structures and white matter, ^{8 –10} nor use task-based fMRI, which identifies brain abnormalities when performing cognitive tasks. ¹¹ Second, the clinical test battery was limited and did not comprehensively evaluate all abilities and behaviors or symptoms (e.g., brain fog) of concern in this population. ^7,11 Third, patients most at risk of brain abnormalities were not identified, in part because factors such as disease duration and illness severity were not examined or inadequately measured in the current sample.

Fourth, as this was a cross-sectional study of untreated patients, it is not known whether present findings are transient or persist once therapy is provided. It is also not known whether patients who do not respond adequately to treatment are more likely to have permanent brain defects. Although a few neuroimaging studies conducted to date have involved longitudinal designs and reported no difference from controls after treatment (e.g., Refs. ^4,12 ), these studies typically involved small sized samples and so may have missed important persisting effects. Fifth, the current sample spanned a wide age range from 18 to 60 years, which is when a number of age-related brain changes are known to occur. ¹³ And finally, the study failed to measure both levels of iodine, a key component of thyroid hormone, and environmental thyroid disruptors.

Despite these limitations, the research by Su et al. ⁵ represents a commendable advance over previous study in identifying the structural and functional brain changes associated with untreated adult hypothyroidism. This study is carefully crafted and methodologically sound, invoking state-of-the-art analytic techniques. Findings are novel and enlightening and, importantly, correspond to the clinical sequelae of the disease. Future studies need to assess the brain more broadly, expand the clinical test battery, use longitudinal designs to evaluate treatment effects, and identify the patients most at risk of permanent brain impairments. Nevertheless, kudos to these researchers for setting us on the right track!