Thyroid hormones and mood: are population data applicable to clinical cohorts?


  • Lauren B. Marangell

  • Acta Psychiatrica Scandinavica Lauren B. Marangell Invited Guest Editor

A link between thyroid hormones and psychiatric disorders, especially depression, has been noted in a myriad of contexts. A large body of literature demonstrates that mood disorders, and their treatments, are often associated with alterations in thyroid hormones. Triiodothryonine (T3) supplementation is an effective augmenting strategy for treatment-resistant depression, as demonstrated in a meta-analysis of 292 euthyroid depressed patients (1). Both pharmacologic and non-pharmacological treatments for mood disorders are associated with alterations in serum thyroid hormone levels (2[3]–4), and in some instances enhanced intracellular conversion of thyroxin to triiodothryonine in the central nervous system (CNS) (5[6]–7).

Against this backdrop, Engum and colleagues report no significant relationship between serum thyroid hormone assays and self-ratings of depression and anxiety in a non-clinical cohort of 30 589 individuals aged 40–89 years (8). Analyses of large population databases are a valuable contribution to the psychiatric literature. However, the clinical interpretation that `psychiatric disorders and thyroid dysfunction should be diagnosed and treated separately' may be premature. As with virtually all research, the current data have limitations that warrant a more cautious interpretation. Static assessment of serum TSH levels, with or without serum T4, is not necessarily reflective of the complexities of the hypothalamic–pituitary–thyroid axis, and peripheral measures are not necessarily reflective of the CNS. This point is particularly germane to patients with mood disorders, who may be sensitive to changes in thyroid status, even when peripheral thyroid hormone assays in the normal range. The current study defines normal thyroid function as a serum TSH between 0.2 and 4.0 mU/l. We have previously reported an inverse correlation between serum TSH and cerebral blood flow and metabolism in a small group of medication-free, medically healthy patients with affective disorders (9). Further, indicating that central abnormalities in thyroid economy are often not manifest in static peripheral hormone assays, Sullivan and colleagues (10) report reduced CNS transthyretin, which might prevent adequate delivery of thyroid hormone to the brain, despite normal peripheral thyroid hormone assays. It has also been suggested that autoimmune thyroid disease may be associated with behavior disturbances even when there are no apparent aberrations in peripheral thyroid hormones. While impossible to address in the current study, it is noteworthy that patients with known thyroid disease did report higher symptoms of depression and anxiety in the Engum study.

In addition, just as small homogenous clinical populations may not be applicable to the population at large, population data may not be germane to patients with a specific disease. In other words, even if future research confirms that there is not an association between thyroid hormones and mood or anxiety symptoms in the general population, there may well be individuals who are more sensitive to fluctuations of thyroid hormones, even within the normal range. As a clinician, consideration of thyroid status remains important in psychiatric patients, both to ensure proper diagnosis and treatment. As a researcher in the area of affective disorders, the authors' hypothesis that common disorders may coexist by chance deserves consideration, but the current data must be evaluated in the context of the numerous clinical and preclinical studies that do suggest an interaction between thyroid hormones and mood. However, there is not a perfect integration of this complex body of literature, and population data, such as that presented by Engum and colleagues, is a welcome addition.


  1. This issue of Acta Psychiatrica Scandinavica also includes a short biography of Lauren B. Marangell.