Subclinical Hypothyroidism and Depression: Is There a Link?

Abstract

A number of large-scale observational studies, as well as a few small-scale intervention studies, have failed to find any link between subclinical hypothyroidism and depression.

The brain is a major target organ for thyroid hormone, and overt hypothyroidism can cause symptoms compatible with depression [reviewed in Samuels (1)]. For this reason, it is routine clinical practice to screen patients with depression for thyroid dysfunction. Patients with moderate to severe overt hypothyroidism can expect improvement in their depressive symptoms with levothyroxine (L-T4) therapy, although symptoms do not always completely resolve.

On the other hand, many patients with depression who are screened for thyroid dysfunction have mildly elevated thyrotropin (TSH) and normal free thyroxine (T4) levels (mild or subclinical hypothyroidism). The patient and care provider may attribute the depressive symptoms to the mild thyroid disorder and initiate L-T4 therapy. Many of these patients are subsequently disappointed that their symptoms do not improve, which leads to overtreatment or use of nonstandard thyroid hormone preparations, with attendant risks. Corroborating this clinical impression, a recent report from the United Kingdom showed that L-T4 is being prescribed more frequently and for more mild degrees of hypothyroidism (2). This suggests that L-T4 is often prescribed for marginal reasons, which are commonly neuropsychiatric in nature. Despite this widespread use of L-T4, ∼5% to 10% of hypothyroid patients on L-T4 replacement therapy have persistent affective symptoms despite normal TSH levels [reviewed in Wiersinga (3)].

With this clinical context, what do we know objectively about depression and mild or subclinical hypothyroidism? Some past observational studies reported an increased prevalence of depression in subclinical hypothyroid patients compared with the general population (1). However, this is not a universal finding. In fact, the largest studies found no differences in depression between euthyroid and subclinical hypothyroid patients, either at baseline or when patients were followed over time (1, 4). Strengths of these latter studies, compared with many of the earlier studies, include large sample sizes, broad age ranges, and absence of bias (population based). Complementing these observational studies, four randomized, placebo-controlled, blinded studies of L-T4 therapy in patients with subclinical hypothyroidism showed no improvement in depression or psychological distress scores (5–8). These small-scale but well-conducted studies provide additional evidence that subclinical hypothyroidism and depression are not causally linked.

The current article by Kim et al. (9) adds valuable information to this body of literature. In this study, 92,206 young to middle-aged adults without depression or known thyroid disease at baseline were examined at least twice in a 4-year time period. TSH, free T4, and free triiodothyronine (T3) were measured at each visit. Mean age was 40 years, and 33% were women. In total, 4384 (4.8%) had subclinical hypothyroidism (which they defined as a TSH >5.0 mU/L with a normal free T4 level), a similar prevalence to other large studies (10). Participants completed the Center for Epidemiologic Studies–Depression (CES-D) survey at each visit, which is a validated instrument for assessing recent depression symptoms. A cutoff score of 16 was used to define clinically significant depression, a level that has been shown to have high sensitivity and specificity (11). By this criterion, incident depression occurred in 7323 participants (8%). The authors appropriately accounted for a number of relevant covariates, including age, sex, body mass index, smoking status, alcohol use, exercise, and medical comorbidities. A unique and valuable addition was their inclusion of a validated questionnaire for perceived stress, as that is an important confounder for depressive symptoms.

The authors conclusively found that there was no association between the presence of subclinical hypothyroidism and the development of depression during a mean follow-up of 2 years by the CES-D cutoff of 16 in this large cohort. They performed additional analyses to address two common issues in this field. (1) In many patients with mildly elevated TSH levels followed without treatment, TSH levels normalize on repeated measurement (12). Therefore, it is likely that a significant number of patients in this study who had subclinical hypothyroidism at baseline had normal TSH levels on repeat measurement. To address this, the authors performed a sensitivity analysis using repeated measurements of thyroid hormones as time-varying variables, and found similar results. (2) Many studies linking subclinical hypothyroidism to adverse end points, particularly cardiovascular outcomes, have shown that TSH values >10 mU/L confer the most risk. For this reason, the American Thyroid Association recommends treating patients with persistent TSH elevations above this level (13). The authors divided their subclinical hypothyroid patients by TSH levels of ≤10 mU/L or >10 mU/L and found similar negative results in both groups. (Note that TSH levels were as high as 57 mU/L in their subclinical hypothyroid group, a level well above what most experts would classify as subclinical hypothyroidism. However, exclusion of outlier TSH levels would likely only strengthen their conclusions.)

Although they did not emphasize it, the authors also conducted an interesting subanalysis in the large euthyroid group (n = 87,822) and found no association between thyroid hormone levels and increased risk of incident depression. There has been a great deal of interest lately in the question of whether variations in thyroid function within the reference range have clinical relevance [reviewed in Taylor et al. (14)]. Previous studies investigating possible links between depression and reference-range thyroid hormone levels have been variable, although the largest and best conducted reported no correlations (1). The current study corroborates those findings.

In addition to these sensitivity analyses, there are a number of other strengths to this study, particularly the large sample size, as this is by far the largest such study to date. The authors also acknowledge some limitations, including the relatively short follow-up period. The cohort was drawn largely from employees of local companies and government agencies and their spouses and so was younger and likely healthier than the overall population. Finally, the authors used a binary definition of depression (CES-D >16), which is the cutoff in widespread clinical use as a screening tool for depression. It is possible that analyzing the data using CES-D scores as a continuous variable might have shown subtle differences between subclinical hypothyroid and euthyroid patients.

Given the evidence summarized above, it is not surprising that patients with mild hypothyroidism do not reliably experience relief of affective symptoms with L-T4 therapy. When this occurs, patients often request L-T4/levotriiodothyronine (L-T3) combination therapy. This is based on the rationale that many L-T4–treated hypothyroid patients have low serum levels of the active thyroid hormone T3 despite normal TSH levels and high-normal to mildly elevated free T4 levels (3). This could theoretically contribute to low intracellular T3 levels and “tissue hypothyroidism.” However, cross-sectional studies have not found correlations between depression and serum T3 levels in treated hypothyroid patients (15). More relevant, a number of placebo-controlled, blinded interventional studies of combined L-T4/L-T3 therapy have failed to find significant improvements in mood (3).

Other patients who are dissatisfied with L-T4 monotherapy request desiccated thyroid extract (DTE), which contains a high L-T3/L-T4 ratio, based on a belief that it is more “natural” and more effective than synthetic L-T4. However, a recent randomized, blinded, crossover study failed to find differences in symptoms, quality of life, or depression when DTE was substituted for L-T4 (16). This suggests that DTE is not superior to L-T4 for affective symptoms.

With all of these negative studies, why do 5% to 10% of L-T4–treated patients with normal TSH levels report persistent symptoms, often neuropsychiatric in nature, and often despite trying alternate thyroid hormone preparations? One possible answer is that patients with self-knowledge of a thyroid condition are more likely to report symptoms (“labeling effect”). This is supported by the literature: in a population-based study of over 25,000 patients, there was no correlation between thyroid function and self-reported depression overall, but the subgroup with known thyroid disease had an increased risk of depression, despite normal TSH levels (17). Hypothyroid patients and their physicians may be more aware of depressive symptoms and ascribe them to the thyroid disease when in fact it is a separate problem, as depression is extremely common in the general population.

Another possible explanation for persistent symptoms in some L-T4–treated patients regards polymorphisms in the deiodinase 2 or thyroid hormone transporter genes (3). These could theoretically lead to lower intracellular levels of active thyroid hormones, even with normal circulating TSH and T3 levels. In one large cross-sectional study, there was no association between the Thr92Ala polymorphism in deiodinase 2 and quality of life in euthyroid patients (18), but in another study, L-T4–treated patients with the same polymorphism had decrements in psychological well-being and mood compared with patients without this polymorphism (19). This study also reported improvements in psychological well-being (but not in depression) in patients with this polymorphism in a trial of combined L-T4/L-T3 therapy. Finally, L-T4–treated patients with polymorphisms in the brain-specific thyroid hormone transporter OATP1C1 had increased levels of fatigue and depression, although this did not correlate with responses to combined L-T4/L-T3 therapy (20). Thus, it is possible that certain subsets of L-T4 patients may in fact have depressive symptoms due to decreased intracellular thyroid hormone levels, although the data are quite preliminary. This hypothesis requires further exploration in targeted populations with these polymorphisms.

Given the findings summarized above and the valuable contribution by Kim et al. (9) in the current study, what conclusions can be drawn regarding a possible link between subclinical hypothyroidism and depression? First, major affective symptoms are not associated with subclinical hypothyroidism but are likely due to independent psychiatric diagnoses, which are common in the general population and occur with similar frequency in patients with subclinical hypothyroidism. Second, symptoms are more apparent when patients are aware of their thyroid status, suggesting that they may be related to the self-knowledge of thyroid disease. However, subtle deficits in mood may exist in subclinical hypothyroidism, at least in subsets of patients, and this requires further targeted investigation. At present, it may be reasonable to initiate treatment of mild hypothyroidism in patients with affective complaints, but realistic expectations should be set regarding the likelihood of symptom resolution. Moderate to severe symptoms are unlikely to be due to subclinical hypothyroidism and need to be evaluated and treated as separate disorders.

Glossary

Abbreviations:

CES-D

Center for Epidemiologic Studies–Depression

DTE

desiccated thyroid extract

L-T3

levotriiodothyronine

L-T4

levothyroxine

T3

triiodothyronine

T4

thyroxine

TSH

thyrotropin