Since the publication of the Clinical Society of London’s “Report on Myxoedema” in 1888, it has been recognized that thyroid disease may give rise to psychiatric disorders that can be corrected by reestablishment of normal thyroid hormone levels. Later, Asher reemphasized that patients with profound hypothyroidism may present with depressive psychosis. As outlined in Table 41-1, the symptoms of hypothyroidism often mimic those of depression, whereas those of hyperthyroidism include anxiety, dysphoria, emotional lability, and intellectual dysfunction, as well as mania or depression, the latter especially characteristic among elderly patients presenting with apathetic thyrotoxicosis.

Because patients with thyroid disease may manifest frank psychiatric disorders that are reversible with endocrine therapy, the thyroid axis has been extensively studied in patients presenting with a wide variety of behavioral disturbances. Various abnormalities of thyroid function have been identified, particularly in depression. In most depressed subjects, the basal serum thyroid-stimulating hormone (TSH), thyroxine (T₄), and triiodothyronine (T₃) levels are within the normal range, although in one report, a third of such patients were observed to have suppressed TSH levels.

A “blunted” TSH response to TRH administration (defined as a TSH rise < 5 mU/L) is seen in approximately 25% of depressed subjects. The blunted TSH response is said to be observed more often in unipolar than bipolar depression, but differentiating these disorders with TRH testing has been disappointing. The blunted TSH response is a “state” marker that normalizes on recovery from depression.

The mechanism for the blunted TSH response in affective disorders is not known; however, glucocorticoids, known to inhibit the hypothalamic-pituitary-thyroid axis, are elevated in depression and could be responsible. This suppressed TSH response is not specific to depression and may be observed in alcohol withdrawal, starvation, normal male aging, renal failure, acromegaly, Cushing’s syndrome, and hypopituitarism. The blunting may also result from medications such as T₄, glucocorticoids, growth hormone, somatostatin, dopamine, and phenytoin, all of which have been reported to diminish this response.

In normal subjects, TSH begins to rise in the evening before the onset of sleep and reaches a peak between 11:00 pm and 4 am. In depression, the nocturnal TSH surge is frequently absent, resulting in reduced thyroid hormone secretion. This finding supports the view that functional central hypothyroidism may occur in some depressed subjects. Sleep deprivation, which has an antidepressant effect, returns the TSH circadian rhythm to normal. The mechanism for the impaired nocturnal TSH rise is unknown.

Although the blunted TSH response is well recognized in depression, it is less clearly appreciated that an enhanced TSH response may occur in up to 15% of depressed subjects with normal baseline thyroid function tests. Most such patients have antithyroid antibodies, a finding suggesting that the TSH hyperresponse may indicate latent hypothyroidism caused by autoimmune thyroiditis. When autoimmunity is tested using the antithyroid peroxidase antibody (anti-TPO) rather than the less specific antimicrosomal antibody, the prevalence of autoimmune thyroid disease is even higher. Not all studies, however, have found an increased prevalence of antithyroid antibodies in depressed subjects when compared with matched control groups.

Approximately 20% of patients admitted to a hospital with acute psychiatric presentations, including schizophrenia and major affective disorders, but rarely dementia or alcoholism, may demonstrate mild elevations in their serum T₄ levels and, less often, T₃ levels. The basal TSH is usually normal but may demonstrate blunted TRH responsiveness in up to 90% of such patients. These findings do not appear to represent thyrotoxicosis, and the abnormalities spontaneously resolve within 2 weeks without specific therapy. Such phenomena may be secondary to central activation of the hypothalamic-pituitary-thyroid axis resulting in enhanced TSH secretion with consequent elevation in circulating T₄ levels.

In depressed patients, the most consistent abnormality of the thyroid axis may be an increase in serum total or free T₄ levels, although usually within the conventional reference range. This elevation generally regresses following successful treatment of the depression.

Thyroid function test abnormalities are common in older individuals. In otherwise normal female patients who are more than 60 years old, the prevalence of elevated TSH values and/or positive antithyroid antibodies is 10% or more. Subjecting apparently asymptomatic individuals with slight elevations of serum TSH but normal T₄ and T₃ levels to a battery of psychological tests has revealed significant differences from control subjects on scales measuring memory, anxiety, somatic complaints, and depression in many but not all studies reported. It is becoming increasingly recognized that depression is much more common in elderly individuals. Whether borderline hypothyroidism plays a role in these behavioral disturbances requires clinical attention. Further investigation should also be directed at studying the outcomes of intervention with levothyroxine.

Among alcoholic patients and those suffering from anorexia nervosa, suppressed T₃ levels with elevations in reverse T₃ and normal TSH values are consistent with the euthyroid sick syndrome. These findings likely result from caloric deprivation.

Medications commonly used to treat psychiatric illness have been shown to affect thyroid function tests (Table 41-2).

Lithium carbonate, used to treat bipolar disorders, interferes with both the release and organification of thyroid hormone. Therapeutic lithium levels diminish both T₃ and T₄ release from the thyroid gland, whereas at higher (probably toxic) levels, iodine uptake and organification may also be inhibited. Following a 3-week therapeutic course of lithium carbonate, suppression of serum T₄ and T₃ levels and associated elevations of basal serum TSH values and exaggerated TSH responses to TRH administration may be noted; these abnormalities generally return to normal within 3 to 12 months, even if the medication is continued.

Goiter is the most common thyroid disorder in lithium-treated patients. Hypothyroidism can also occasionally develop, particularly in patients who have thyroid glands that have been compromised by disorders such as Hashimoto’s thyroiditis and Graves’ disease previously treated with 131-iodine therapy. However, it is uncommon for hypothyroidism to occur if pretreatment thyroid function was completely normal and patients are thyroid antibody negative. If considered clinically necessary, lithium may be continued and T₄ added to treat patients who develop goiter or hypothyroidism.

The effects of phenytoin (Dilantin), occasionally used for bipolar disorder, on thyroid function are quite complex. Suppressed values of total T₄ and, occasionally, free T₄ are observed in a significant minority of patients who are treated on a long-term basis with phenytoin alone and in more than 75% of patients in whom the drug is combined with carbamazepine (Tegretol). The lower total T₄ levels are likely secondary to displacement of T₄ from thyroxine-binding globulin (TBG), whereas the reduced free T₄ levels result from enhanced clearance of T₄ through phenytoin-induced hepatic microsomal oxidative enzyme activity. Generally, the suppressed T₄ levels are accompanied by normal T₃ and free T₃ levels and normal TSH concentrations. Normal basal TSH values with diminished TSH responses to TRH have been attributed to potential phenytoin agonism at the T₃ receptor. However, other studies have suggested that this may be an assay artifact because free T₄ values have been found to be normal or mildly elevated in analyses using undiluted serum.

Carbamazepine (Tegretol) is used increasingly in bipolar disorder. Long-term use with maintenance of therapeutic serum levels suppresses serum T₄ values in more than 50% of patients. This may be the result of enhanced hepatic metabolism of T₄. TRH stimulation testing before and after initiation of carbamazepine therapy reveals that TSH responsiveness is reduced by the addition of this drug; this finding has led to speculation that carbamazepine may inhibit thyroid function through effects on the pituitary gland. Displacement of T₄ from TBG, similar to that seen with phenytoin, has additionally been cited as a potential effect.

Both phenobarbital and valproic acid are reported to lower serum levels of T₄ in patients treated on a long-term basis, the former via enhanced hepatic T₄ clearance and the latter likely from protein binding changes. Heroin, methadone, and perphenazine commonly increase serum TBG levels and therefore may elevate serum total T₄ levels, although TSH and free T₄ values remain normal. Amphetamines induce hyperthyroxinemia through enhanced secretion of TSH, an effect that appears to be centrally mediated.

Antidepressants do not generally cause abnormal peripheral thyroid hormone levels but may affect thyroid hormone metabolism in the central nervous system (CNS). Selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs) appear to promote activity of type 2 deiodinase (D2), which increases conversion of T₄ into T₃ in the brain. However, peripheral circulating total T₄ and free T₄ levels, often show a modest decline, though still within the normal range after treatment with various pharmacologic classes of antidepressants, or electroconvulsive therapy (ECT). There are case reports of sertraline-, paroxetine-, and escitalopram-related asymptomatic hypothyroidism, but more recent studies evaluating fluoxetine and sertraline have shown no clinically significant change in thyroid function test results.

The use of TCAs in thyrotoxic patients should be pursued with caution because cardiac dysrhythmias may be exacerbated or precipitated. Further, the monoamine oxidase (MAO) inhibitors may cause hypertension in thyrotoxic patients, although these drugs generally do not affect thyroid function or serum thyroid hormone levels.

Asher’s report on “myxoedematous madness” demonstrated that thyroid hormone deficiency resulted in depression that reversed with thyroid hormone administration. This led to studies of the role of thyroid hormone therapy alone in the treatment of depression and other psychiatric diseases and open studies of high-dose T₄ for refractory bipolar and unipolar depression. Euthyroid individuals with typical hypothyroid symptoms who are considered depressed on psychological testing do not improve when treated with T₄. In fact, patients presenting with symptoms of hypothyroidism with normal thyroid function test results respond more positively to placebo. Although initial reports of T₃ as single therapy were promising, these studies were methodologically flawed, so the role of thyroid hormone by itself in the treatment of depression in the absence of abnormal thyroid function has not been established.

Studies have shown that symptomatic patients with subclinical hypothyroidism (elevated serum TSH but normal T₄ and T₃ levels) can have significant impairment of memory-related abilities, health status, and mood, as well as anxiety, somatic complaints, and depressive features, when compared with euthyroid controls. Greater than 60% of patients with subclinical hypothyroidism report some degree of depressive symptoms. Effects of treatment with levothyroxine have been mixed. Some studies suggested that normalization of thyroid function with levothyroxine therapy, as determined by the serum TSH, may completely reverse these neuropsychiatric features. Conversely, other studies suggested that hypothyroid patients receiving levothyroxine replacement may continue to experience neuropsychiatric abnormalities despite a normal TSH when compared with controls of similar age and sex. Further, when thyroid hormone is withdrawn from subjects with underlying hypothyroidism, gradually increasing cognitive deficits, sadness, and anxiety symptoms are observed over the ensuing few weeks. These findings indicate that the patient presenting with depression must be assessed for thyroid dysfunction because the presence of even subclinical hypothyroidism may provide an opportunity for resolution of the depression with thyroid hormone treatment.

Murray introduced thyroid hormone therapy to the clinical world in 1891. From the beginning, this was a treatment based on a combination of T₄ and T₃ derived from animal thyroid extracts. Variability of the animal extracts in regard to T₄ and T₃ content and ratios from batch to batch and brand to brand led to the replacement of “natural” combination therapy using extracts with pharmaceutically more precise synthetic T₄ and T₃. Eventually, the simplicity of T₄ monotherapy was adopted as usual therapy. The T₄ dosage was titrated based on symptoms until more sensitive TSH assays became available. During the 1980s, sensitive TSH assays allowed titration of thyroid hormone therapy to “normal,” and this resulted in significant dose reductions of as much as 100 μg per day. Once euthyroid, however, some patients continued to complain of symptoms consistent with some components of hypothyroidism. Since that time, multiple reports have appeared evaluating the effectiveness of combining T₃ with T₄ to improve neuropsychological outcomes. The 1999 report of Bunevicius et al, for example, seemed to indicate that substituting 12.5 μg of T₃ for 50 μg of the individual’s usual T₄ dose resulted in improvement in mood and neuropsychological function. Several double-blind randomized controlled trials designed to correct design flaws of previous trials subsequently failed to reproduce the positive effects reported by Bunevicius et al and did not demonstrate improvement in self-rated mood, well-being, or depression scales with the addition of T₃ to T₄ therapy. Moreover, these studies failed to demonstrate differences in cognitive function, quality of life, or subjective satisfaction with treatment, but they did report that anxiety scores were significantly worse in patients treated with the T₄-T₃ combination. However, a subset of patients may benefit from T₃ supplementation. Preliminary evidence suggests that patients with certain mutations in the D2 gene may have a positive response to T₃ potentiation, but prospective trials have not yet been conducted. At this time, it does not appear justified to use combined T₄ and T₃ treatment in hypothyroid patients who complain of depressive symptoms after biochemical euthyroidism is restored.

Adjuvant therapy has been said to be logical when depression fails to resolve after 6 weeks of adequate antidepressant medication. Such resistance occurs in about 30% to 45% of cases. The role of adjuvant thyroid hormone with TCAs has been investigated in euthyroid patients with depression for more than 25 years. T₃ doses of 25 to 50 μg daily increase serum T₃ levels and cause suppression of serum TSH and T₄ values. Two separate therapeutic effects of T₃ therapy have been studied: first, its ability to accelerate the onset of the antidepressant response; and second, its ability to augment antidepressant responses among those considered pharmacologically resistant.

Given that the antidepressant effect of TCAs is known to be delayed, the role of T₃ in accelerating the therapeutic onset of these drugs has been investigated. Several reports detailing the clinical outcomes of starting T₃ (5-40 μg daily) along with varying doses of TCAs, as well as SSRIs, at the outset of therapy have appeared in the literature. The study populations were inhomogeneous, consisting of patients with various types of depression. Furthermore, the studies had important methodologic limitations, including small sample sizes, inadequate medication doses, lack of serum medication level monitoring, and variable outcome measures. Because two relatively large, prospective, randomized placebo-controlled studies came to opposite conclusions, it still has not been clearly established that T₃ accelerates the antidepressant effect of TCAs. A meta-analysis of double-blind clinical trials comparing SSRI-T₃ treatment with SSRIs alone showed no significant difference in remission rates.

An additional hypothesis is that adding small doses of T₃ to the antidepressant therapy of patients who have little or no initial response will enhance the clinical effectiveness of the antidepressant. Resistance to antidepressants is defined as inadequate remission after 2 successive trials of monotherapy with different antidepressants in adequate doses, each for 4 to 6 weeks before changing to alternative therapies. However, a course of 8 to 12 weeks of ineffective antidepressant therapy is commonly deemed unacceptable, and strategies designed to augment the response are being sought. Early studies assessing T₃ effectiveness in augmenting the antidepressant response were neither placebo controlled nor focused on patient populations that could be directly compared. The first placebo-controlled, double-blind, randomized study reported results in 16 unipolar depressed outpatients who had experienced no improvement in their clinical outcomes with TCAs alone. The intervention consisted of adding 25 μg of T₃ or placebo daily for 2 weeks before the patients were crossed over to the opposite treatment for an additional 2 weeks. No beneficial effect of T₃ was apparent. The only other placebo-controlled, randomized, double-blind trial investigating this question involved 33 patients with unipolar depression treated with either desipramine or imipramine for 5 weeks before random assignment to placebo or 37.5 μg of T₃ daily. After 2 weeks of observation on T₃, during which TCA levels were monitored, significantly more patients treated with T₃ (10 of 17; 59%) had a positive response than did placebo-treated patients (3 of 16; 19%). A subsequent open clinical trial of imipramine-resistant depression, using a prolonged TCA treatment period preceding the addition of T₃, showed no demonstrable T₃ effect.

The SSRI group of substances (including fluoxetine and sertraline) is the preferred antidepressant treatment in the United States today. A large, double-blind, placebo-controlled study to determine the role of T₃ as augmentation therapy did not demonstrate an effect of T₃ in augmenting the response of paroxetine (an SSRI) therapy in patients with major depressive disorder, but a similar study using sertraline and T₃ showed a conflicting positive response. Responders in the Cooper-Kazaz report seemed to have had lower circulating thyroid hormone levels before treatment and to have experienced greater decreases in TSH levels as a result of the intervention. This finding may indicate that patients benefiting from the addition of T₃ may have been subtly hypothyroid, and the addition of T₃ compensated for this deficiency. A metaanalysis of the available data suggests that coadministration of T₃ and SSRIs has no significant clinical effect in depressed patients when compared with SSRIs alone. More research is necessary to determine whether those patients with a functional type 1 deiodinase (D1) gene polymorphism may be more responsive to T₃ cotreatment.

T₃ has been reported to augment the antidepressant effect of ECT. However, there is little to no evidence to guide the duration of treatment with T₃, and there are few studies on the side effects of long-term administration.

For the 10% to 15% of patients with bipolar disorder with four or more episodes of manic-depressive psychosis yearly (rapid cyclers), the prevalence of autoimmune thyroid disease may reach 50% or higher. Therapeutic intervention with standard therapy such as lithium is frequently disappointing. Open-label studies treating such patients with T₄ in pharmacologic doses sufficient to suppress serum TSH and elevate T₄ levels to approximately 150% of normal showed that treatment may decrease the manic and depressive phases in both amplitude and frequency, and it led to remission in some patients. Given these encouraging results, controlled studies on the efficacy of T₄ or T₃ seem warranted.

Thyroid hormones play a critical role in the development and function of the CNS. T₃ receptors are widely distributed throughout the brain, and there is much evidence that thyroid hormone regulates brain function through interaction with the catecholaminergic system. Thyroid hormone action in brain tissue is accomplished through the binding of T₃ to its nuclear receptor. The T₃ is derived from T₄ by the action of D2, which is located throughout the CNS.

Most studies using thyroid hormone as adjuvant therapy have used T₃ rather than T₄. In those reports assessing the advantages of one over the other, T₃ was considered superior. In a randomized trial combining T₄ or T₃ with antidepressants, only 4 of 21 patients (19%) treated with 150 μg/day of T₄ for 3 weeks responded, whereas 9 of 17 (53%) responded with 37.5 μg/day of T₃. Further studies of open T₄ treatment in antidepressant-resistant patients have appeared, but the lack of controls makes outcome interpretation difficult. One of these studies indicated that responders to T₄ had significantly lower pretreatment serum T₄ and reverse T₃ levels, a finding leading the investigators to believe that the responders may have been subclinically hypothyroid. Adjuvant therapy with T₄ rather than T₃ may be indicated when subclinical hypothyroidism or rapid cycling bipolar disease is present. Because T₄ equilibrates in tissues more slowly than T₃, treatment with T₄ for at least 6 to 8 weeks, and preferably longer, would be necessary to determine its efficacy in this situation.

As personalized medicine evolves, therapies will inevitably become more directed. Research directed at type 1 deiodinase (D1), which is important for serum conversion of T₄ to T₃, suggests that certain polymorphisms of D1 may be associated with a positive response to T₃ potentiation of SSRIs. These patients with certain alleles have inherently lower D1 activity and therefore have naturally lower serum T₃ levels. When compared with placebo, these patients have decreased depression scores at 8 weeks with T₃ supplementation in combination with sertraline.

Conversely, there is evidence that specific subsets of patients may not respond to T₃ potentiation. Genetic researchers have identified an organic anion transporting polypeptide (OATP) that is thought to be key in delivery of T₄ to the brain. Polymorphisms in the OATP1C1 gene appear to be linked to increased depressive symptoms among patients with hypothyroidism. These patients do not appear to have any decrease in depressive scores with T₃ supplementation when compared with controls. The clinical significance of these findings is yet to be determined, but it may have a meaningful impact on the future of depressive treatment.

It has been postulated that D2 activity in the CNS is deficient in depression, thereby giving rise to a state of brain hypothyroidism coexisting with systemic euthyroidism. Alternatively, D2 activity may be depressed by the elevated cortisol levels seen in depression and stress, thus resulting in T₄ instead being converted to reverse T₃ (rT₃) by “inner ring” brain 5 deiodinase (type III deiodinase [5D-III]) activity, with decreased T₃ and increased rT₃ levels. Notably, T₃ treatment is feasible because T₃ is not dependent on transport by transthyretin, which is low in depression, and therefore would ensure adequate T₃ delivery across the blood-brain barrier to the brain.

It has been shown that desipramine, a TCA, and fluoxetine, an SSRI, both enhance D2 activity in the CNS, thus presumably increasing the availability of T₃ in the brain. This could conceivably account for the clinical efficacy of these classes of drugs.

It seems prudent to check thyroid function test results in those psychiatric patients who are at increased risk for developing thyroid disease. Women more than 45 years of age, patients with known autoimmune diseases, individuals with a family history of thyroid disease, and those receiving lithium or suffering from dementia should be screened for underlying thyroid abnormalities. Patients receiving medications known to influence the interpretation of thyroid function tests should have these considered when interpreting the results of testing.

It is recommended that T₄ therapy be offered to any depressed patient with an elevated serum TSH, especially if it is accompanied by increased antithyroid antibody titers or a low free T₄. Thyroid hormone replacement may alleviate the depression in these individuals. Conversely, antidepressant therapy, if required, may be ineffective before normalization of thyroid axis parameters. One observation suggests that there is a subset of patients with hypothyroidism who are taking T₄ and who experience higher depression and anxiety scores than their euthyroid counterparts. These patients in particular, who may have D2 gene polymorphisms, may benefit from the addition of T₃, but prospective trials have not yet been conducted. In patients with refractory depression but normal systemic thyroid function, adjuvant T₃ therapy may not be worth considering.