Abstract
Introduction:
Over the years, several drugs used in the treatment of nonthyroidal conditions have been shown to affect thyroid function. As novel drugs are introduced, novel interactions are described. The aim of this review is to summarize clinically relevant thyroidal side effects of drugs used for nonthyroidal conditions. Special focus is given to recent developments and to drugs with the largest clinical relevance.
Summary:
Thyrosine kinase inhibitors are novel drugs used in the treatment of several neoplasias, including thyroid cancer. Thyroidal side effects are being increasingly detected with these drugs. Some drugs in this category affect thyroid hormone metabolism and therefore only affect patients on thyroid replacement. Others affect the thyroid directly profoundly, causing primary hypothyroidism. Immune modulators used in infectious, inflammatory, and neoplastic conditions also cause hyper- and hypothyroidism, through poorly understood immune or nonimmune mechanisms. The effects of amiodarone on the thyroid have been long recognized. However, given the complexity of these effects, several areas in this field remain problematic, such as the identification of subtypes of hyperthyroidism and the best treatment strategies. Lithium also has important antithyroid effects and it is a commonly prescribed medication. Its antithyroid effects may have clinical utility in selected clinical situations. Other drugs known to affect thyroid hormone absorption, metabolism, and transport are also briefly reviewed.
Conclusions:
Several drugs are known to alter thyroid function as a side effect of their primary pharmacological action. Some of these effects have been recognized for decades, but novel thyroid–drug interactions are being recognized as new drugs are developed. It is important for the clinician to be familiar with thyroid–drug interactions, as enhanced surveillance may be necessary in patients undergoing therapies known to affect thyroid function.
Introduction
Tyrosine Kinase Inhibitors and Thyroid Function
Tyrosine kinase inhibitors (TKIs) have recently emerged as promising chemotherapeutic agents in several types of malignant neoplasms, including thyroid cancer. The main pharmacologic effect of TKIs is the inhibition of several receptors with tyrosine kinase activity involved in carcinogenesis, cell proliferation, and neoangiogenesis (1). Several TKIs exert distinctive effects on thyroid function. While these effects have not been observed with all TKIs, the frequency with some of these agents is high enough to recommend thyroid function test monitoring in all patients during and after treatment. TKIs affect thyroid function through two major mechanisms: by increasing thyroid hormone requirements in patients on thyroid hormone replacement and by causing primary hypothyroidism in patients with previously normal thyroid function. The first mechanism is particularly important in patients undergoing TKI treatment for thyroid cancer after total thyroidectomy, and it probably reflects increased metabolism of thyroid hormone. The second mechanism (or mechanisms) is poorly understood, but it is likely to involve direct transient or irreversible toxicity to the thyroid gland.
Effects of TKIs on patients on thyroid replacement
In a small study, Imatinib caused a fourfold increase in TSH level in all eight patients taking levothyroxine (L-T4) after total thyroidectomy for thyroid cancer, while none of the three patients with a normal thyroid had this effect (2). The effect was reversible after discontinuation of treatment, and it could be corrected by a doubling in the daily L-T4 dosage in some patients. In one patient, changing the time of the intake of Imatinib to 4 hours after the intake of L-T4 did not correct the hypothyroidism. There were no changes in total T4 to suggest increased serum protein binding of thyroid hormone as a possible explanation. Hence, these effects of Imatinib are most likely related to increase liver metabolism of thyroid hormone.
Effect of TKIs on euthyroid patients
Sunitinib is a multitargeted TKI approved for use in renal cell carcinoma and in Imatinib-resistant gastrointestinal stromal tumor. While initial studies did not observe (or sought) effects of this drug on thyroid function, several case reports of hypothyroidism have emerged since 2005 (3 –5). In one study, 36% of patients developed primary hypothyroidism (4), while another study showed persistent hypothyroidism in 46% of cases (6). In both studies fluctuating elevations of TSH were observed in several additional cases. In all cases correction of hypothyroidism was achieved with standard L-T4 doses. In the few cases that managed to come off the medication, spontaneous return to euthyroidism was observed (6), but there have been cases of thyroid atrophy on ultrasonography, suggesting permanent hypothyroidism as a possible outcome (2). The mechanism of Sunitinib-induced hypothyroidism is unclear. In some cases, destructive thyroiditis has been directly observed (7,8). An effect on iodine uptake has been postulated (6), but not confirmed by in vitro studies (9). One study suggested an antiperoxidase effect (10). However, this proposed mechanism would not explain cases with initial destructive thyrotoxicosis, or with thyroid atrophy. The evidence available suggests that Sunitinib exerts a direct toxic effect on the thyroid gland of variable intensity. In its most severe form, significant thyrotoxicosis may precede the onset of hypothyroidism, while more commonly hypothyroidism alone ensues. More recently, further insight has been offered by careful case reports (11,12). In these patients reduced thyroid blood flow, striking thyroid shrinkage, and hypothyroidism were observed during cycles of Sunitinib. Partial regression of these changes was observed off Sunitinib cycles and after discontinuation of the drug. These findings were interpreted as a direct effect of vascular endothelial growth factor receptor inhibition in the thyroid vascular bed explaining low uptake hypothyroidism and thyroid volume shrinkage (13). An interesting question is whether the presence of these effects correlates to the antitumoral efficacy of Sunitinib in individual patients.
Sorafenib is another multikinase inhibitor used in several solid tumors. In the available studies, Sorafenib has been associated with hypothyroidism as well. In the study from Tamaskar, from the United States, 21% of patients developed hypothyroidism, mostly in the subclinical range (14). In a more recent study from Japan, in which thyroid function tests were drawn at fixed and frequent interval, 68% of patients developed hypothyroidism, and in 21% TSH was >10 mU/L. Interestingly, in the latter study in 24% of patients hypothyroidism was preceded by a phase of mild thyrotoxicosis (15). The reason for the different incidence of thyroid complications in the two studies is unclear.
While risk factors for TKI-induced thyroid dysfunction are still poorly understood, it seems reasonable to recommend periodic monitoring of thyroid function tests in all patients on these medications. Future studies should be directed at understanding whether thyrotoxicosis should prompt discontinuation of the drug, a dose reduction, or a drug holiday. Correction of hypothyroidism with L-T4 seems the only necessary step in patients experiencing this complication.
Immune Modulators Affecting Thyroid Function
Alemtuzumab and Graves' disease
Alemtuzumab is a humanized monoclonal antibody directed against the CD52 cell surface antigen. Infusion of alemtuzumab induces complement-mediated lysis of CD52 cells, which includes most lymphocytes and monocytes. While alemtuzumab has been used as an immune suppressor in several autoimmune conditions, it is mostly effective in multiple sclerosis (MS) (16). The occurrence of Graves' disease after a course of alemtuzumab was first reported in 1999 (17). Graves' disease developed in 9/27 patients with MS 6–31 months after a course of alemtuzumab. There were no distinctive features in alemtuzumab-induced Graves'. Most patients developed de novo TSH receptor antibodies and in two cases Graves' ophthalmopathy also occurred. There was a high rate of relapses after a 6 months course of carbimazole, similarly with what one would expect with sporadic Graves' disease. The incidence of Graves' disease seems to vary with the administration protocol and with the subtype of MS patients being treated. A study in 39 patients with aggressive remitting-relapsing MS showed just one case of Graves' disease (16). In a more recent study, thyroid dysfunction was found in 48/216 (22%) of patients with MS after alemtuzumab treatment. Graves' disease predominated, but there were 15 cases of autoimmune hypothyroidism (18). These studies offer compelling evidence that alemtuzumab is capable of causing classical Graves' disease in a significant number of patients, but the mechanism for this phenomenon remains obscure. Graves' disease was not reported in patients in whom alemtuzumab was used to treat disorders other than MS, such as rheumatoid arthritis (19) or chronic lymphocytic leukemia (20). It therefore appears that patients with MS are peculiarly susceptible to this complication. Genetic polymorphisms known to be associated with Graves' disease such as those at the human leukocyte antigens (HLA), tumor necrosis factor-alpha, and cytotoxic T-lymphocyte associated protein (CTLA-4) loci were not observed with increased frequency in MS patients who developed Graves' disease (17). More recently, a locus for genetic susceptibility for both Graves' disease and MS has been reported on chromosome 20, containing the gene for CD40 (21,22). It will be very interesting to test whether patients with MS and the common susceptibility polymorphism are more likely to develop Graves' disease after alemtuzumab, as in such case a truly personalized treatment protocol could be envisaged.
Graves' disease has also been reported in several cases during high activity antiretroviral therapy (HAART) in patients with human immunodeficiency virus infection (23). In a retrospective series in large urban human immunodeficiency virus clinics, the incidence of Graves' was 2.8% in women and 0.2% in men (24), which is similar to the prevalence in the general population. However, the clustering of these cases after the enrollment in HAART programs suggests that HAART may precipitate Graves' disease in predisposed subjects. This situation is vaguely reminiscent of alemtuzumab treatment, in that profound immune suppression and lymphopenia are followed by reconstitution of the immune system. Unbalanced expansion of autoreactive T-cells is one proposed explanation of this phenomenon (25).
Interferons and other cytokines
Interferon alpha (IFNα) is a human recombinant cytokine mainly used in the treatment of chronic hepatitis C, but also in several other infectious and malignant conditions. IFNα is highly effective in reducing viral load and in prolonging survival in patients with hepatitis C, especially in association with ribavirin (26,27). Dose-limiting side effects are common, most frequently malaise, depression, and hematologic side effects. Thyroid dysfunction during IFNα therapy is also quite common. Pegylated IFNα is more effective in inducing a viral response, but shows a similar risk of causing thyroid dysfunction (28). Thyroid effects of IFNα have been classified as autoimmune and nonautoimmune, mostly based on the presence or absence of markers of thyroid autoimmunity such as serum thyroid peroxidase and/or thyroglobulin antibodies (29). The most common event is the development of serum thyroid antibody in previously negative patients. The frequency of this phenomenon is highly variable in the many available studies [reviewed in detail in Mandac et al. (29)]. In most large series, the incidence is ∼10%–20%. A minority of patients go on to develop clinical thyroid dysfunction, most often classical Hashimoto's thyroiditis, with goiter and hypothyroidism. Classical Graves' disease and sometimes Graves' ophthalmopathy also develop, although less commonly. Approximately one half of patients present with a form of transient, destructive thyrotoxicosis. This is very reminiscent in its clinical course of painless or postpartum thyroiditis and similarly characterized by low radioiodine uptake in the setting of a suppressed TSH. However, in the majority of patients with destructive thyrotoxicosis in the setting of IFN therapy, thyroid antibody tests remain negative (30). This observation has suggested the classification of this disease as nonautoimmune (29). Indeed some studies have shown a direct toxic effect on thyrocytes in vitro (31). However, in the absence of pathology data, ruling out a seronegative, cell-mediated immune mechanism is currently impossible. Indeed patients with destructive thyrotoxicosis from IFNα develop a hypoechoic pattern on thyroid ultrasonography independently of whether thyroid antibody is present (32). This finding is suggestive of a lymphocytic infiltrate. Independently of the etiology, thyroid dysfunction during IFNα therapy has a significant impact. While hypothyroidism can be easily treated with L-T4, specific treatment for destructive thyrotoxicosis is not available. Symptoms of thyrotoxicosis may add to the discomfort of patients with severe malaise or flu-like symptoms from IFNα itself. In patients undergoing prolonged courses of IFNα and experiencing relapsing flares of thyroiditis, ablation with I-131 during a period of remission may be offered to prevent further episodes of the condition.
Interleukin-2 (IL-2) is used in the treatment of melanoma and renal cell carcinoma. IL-2 is used either alone or in conjunction with IFNα 2a or lymphokine-activated killer cells. Earlier studies showed a significant incidence of hypothyroidism after treatment, in the range of 20%–50% either with IL-2 alone (33) or in combination with lymphokine-activated killer cells (34). Most patients who developed hypothyroidism had positive thyroglobulin or thyroid peroxidase antibodies, suggesting autoimmune thyroiditis. Initial studies suggested that the occurrence of hypothyroidism was associated with a favorable response to treatment, but larger studies did not confirm this phenomenon (35). The opposite seems to be true: thyroid dysfunction develops more often in the responders because they receive longer courses of the treatment (36). An early phase of presumably destructive thyrotoxicosis is common, with variable degrees of hyperthyroidism (37). Interestingly, in some patients with destructive thyrotoxicosis, thyroid cytology showed a lymphocytic infiltrate in spite of negative thyroid antibody tests, suggesting a purely cell-mediated autoimmune mechanism (37). Similarly to IFNα-induced thyroid disorders, hypothyroid patients are easily treated with thyroid hormone, while destructive thyrotoxicosis only requires symptom control with beta-blockers.
Lithium
Lithium is highly effective in the long-term management of bipolar disorder. The effects of lithium on thyroid physiology have been studied extensively, but their underlying molecular mechanisms remain elusive. Lithium appears to be actively concentrated in the thyroid follicular cell (38). In rats, thyroid radioiodine uptake does not appear to be significantly affected by chronic lithium treatment, but its release is diminished, resulting in a net positive intrathyroidal iodine balance (38). Part of this effect may be the result of a block in the release of thyroid hormone (39). In spite of this effect, serum L-T4 is not low in experimental conditions, possibly due to a prolonged half-life of the hormone (40). Indeed effects of lithium on thyroid deiodinases have been described, but these could be an adaptive response to initial hypothyroxinemia, rather than being directly an effect of lithium therapy. Overall, the effects of lithium on thyroid physiology are reminiscent the Wolff–Chaikoff effect, obtained through the administration of high-dose iodine. One can speculate that in both situations, high levels on intracellular iodine are responsible for the downregulation of thyroid hormone secretion. Goiter is the most common clinical finding in lithium-treated patients (41). The reported incidence of goiter is highly variable, depending on the geographical setting, detection method, and other factors, and it ranges from 0% to ∼60% [reviewed in Lazarus (42)]. In most reports and in the experience of the author, goiter is most often diffuse, moderate in size, with normal echogenicity on thyroid ultrasonography. The cause for the goiter likely lies in chronic subtle thyroid dysfunction, but direct proliferative effects of lithium on thyroid function have also been proposed (43). There is no evidence of increased nodular thyroid disease (44), but this has not been extensively studied. The incidence of hypothyroidism is also highly variable, probably also depending on the populations studied. Emerson observed an incidence of clinical and subclinical hypothyroidism of 14% in 1973, employing a sensitive TSH assay (45). Lazarus estimated the prevalence at 3.8% in 1986 (46), but higher rates have been reported lately (42). The etiology of hypothyroidism during lithium treatment is likely due to the changes in thyroid physiology described above. Thyroid autoimmunity in general seems not to be increased in lithium-treated patients (46), but when present, it may increase the risk of hypothyroidism (45). Lithium therapy has been associated with an increased incidence of self-limited destructive thyrotoxicosis (47). The clinical features of this disorder are similar to typical painless thyroiditis, but thyroid antibodies are positive in 50% or less of patients. Several mechanisms have been called to explain this phenomenon: thyroid autoimmunity, a direct toxic effect of lithium on the thyroid gland, and a toxic effect of increased intrathyroidal iodine consequent to lithium treatment. Currently, there is insufficient data to favor one hypothetical mechanism over the other.
Treatment of lithium-induced hypothyroidism with L-T4 is no different from treatment of primary hypothyroidism. In these cases there is often a concern that even mild hypothyroidism may contribute to depressive symptoms and that increased TSH levels may favor further growth of goiter. Treatment of goiter in euthyroid patients is more controversial, but in patients with enlarging goiter a trial of L-T4 directed at lowering without suppressing TSH is reasonable. Destructive thyrotoxicosis is treated conservatively, with beta-blockers and/or calcium channel blockers, as it is usually mild and self-remitting. There is no data on treatment with corticosteroid, but these drugs would increase significantly the risk of a manic episode and therefore should not be used.
The presence of thyroid abnormalities alone is almost never a reason for lithium discontinuation, as this medication is often crucial in maintaining these patients free of the most severe manifestations of their psychiatric illness.
Because of its properties, lithium has been used as an antithyroid drug. In a small randomized trial, lithium was shown to have efficacy comparable but not superior to thionamides in terms of thyroid hormone level reduction (48). In another study, the combination lithium–carbimazole was superior to carbimazole alone (49). One can conclude that lithium has a role in the treatment of thyrotoxicosis in cases of intolerance to thionamides and as an adjunct to thionamides when a more rapid achievement of euthyroidism is necessary.
Finally, lithium has been proposed as a method for increasing the efficacy of I-131 treatment in patients with hyperthyroidism. While a large randomized study has shown no advantage of this procedure (50), a very recent large retrospective study has shown that patients pretreated with lithium had a blunted post I-131 thyroid hormone peak, a shorter time to euthyroidism, and a larger overall cure rate (51). Differences among these studies likely reflect differences in treatment protocols and in patient selection. In particular, in the Bogazzi study (51) the administration of lithium was started 5 days before the administration of I-131, whereas in the study from Bal (50), lithium treatment was began on the day of I-131 administration.
Amiodarone
The effects of amiodarone on the thyroid have been the subject of an in-depth scholarly review recently (52). Amiodarone is an important class III antiarrhythmic drug used very often in the management of atrial fibrillation and other cardiac conditions. One amiodarone molecule contains two atoms of iodine. A daily dose of 300 mg amiodarone provides 111 mg of iodine, of which about 10% (11 mg) are made available daily as inorganic iodine through deiodination of the amiodarone molecule (53). Thus, average doses of amiodarone result in a daily release of some 30–100 times the required daily allowances of inorganic iodine, estimated at 150 mcg daily in the United States. Amiodarone is a highly lipophilic drug, with a large distribution volume and an very long half-life of 40–60 days. Amiodarone is structurally similar to thyroid hormone and it may exert some antagonist effects at the thyroid hormone receptor, possibly reflected in its antiarrhythmic properties. In addition, amiodarone inhibits type I deiodinase to a clinically significant extent. Changes in thyroid function tests are common in patients taking amiodarone. The inhibition of type I deiodinase results in low triiodothyronine (T3) levels, high T4 levels, and high reverse T3 levels (54). Transient elevation of TSH is common during the first few months of treatment, possibly as a result of high iodine exposure resulting in the Wolff–Chaikoff effect (55). These thyroid function changes are to be expected in most patients. Clinicians need to be familiar with them, but active treatment is not necessary.
Clinically significant thyroid dysfunction occurs in a significant number of patients taking amiodarone. Amiodarone-induced hypothyroidism (AIH) is observed in ∼5%–15% of patients on the medication. Significant risk factors for AIH include female gender, positive thyroid antibody, and residence in an iodine-repleted region (52). The etiology of AIH is probably multifactorial, but it is deemed to involve an acquired or preexistent inability to escape from the Wolff–Chaikoff effect caused by high intracellular thyroidal iodine levels. Discontinuation of amiodarone results in resolution of AIH in 2–3 months in more than half of the patients. The rapidity of recovery is surprising given the long half-life of the drug. However, in a significant number of patients hypothyroidism is prolonged or permanent, especially if thyroid autoimmunity is present. In most cases this approach is undesired due to the efficacy of amiodarone in treating the underlying cardiac disorder and the lack of suitable alternative drugs. Treatment with L-T4 allows continuation of amiodarone treatment and adequate correction of hypothyroidism when clinically necessary.
Amiodarone-induced thyrotoxicosis (AIT) is observed in 2%–12% patients on amiodarone. Risk factors for AIT include male gender and residence in an iodine deficient region (52). Thyrotoxicosis can have deleterious effects in cardiac patients and therefore represents a serious problem. Two types of AIT are recognized. In type I AIT, hyperthyroidism is likely the direct consequence of the large iodine load on a preexisting state of thyroid autonomy, such as a pretoxic multinodular goiter or Graves' disease. In type I AIT, there is ongoing thyroidal iodine organification and thyroid hormone synthesis. Therefore, thionamides are effective in controlling the disorder. Type II AIT is a form of destructive thyrotoxicosis, likely the result of a direct toxic effect of either iodine or amiodarone itself to the follicular cell. As a consequence, thyroid hormone synthesis is nil or negligible, and the syndrome is caused by the release of preformed hormone. Antithyroid drugs are therefore ineffective in type II AIT, but corticosteroids are effective in controlling this condition. Distinguishing type I from type II AIT is therefore crucial in selecting the most appropriate treatment. Radioactive uptake and scanning is the standard test used to distinguish forms of thyrotoxicosis responsive to thionamide drugs (with normal or high uptake) from destructive forms (with very low or undetectable uptake). However, in type I AIT the large amount of circulating inorganic iodine dramatically reduces the usefulness of this test by diluting the radioactive iodine tracer dose, resulting in low uptake in the majority of cases. As a result, the clinical distinction between the two forms has relied mostly on a number of indirect findings suggesting type I AIT, such as the presence of thyroid nodules, or positive TSH receptor antibody, while specific and robust markers of type II AIT have not been found. Two imaging techniques have shown promise. Bogazzi et al. have shown that low vascular flow to the thyroid as measured by color flow Doppler sonography (CFDS) is a reliable indicator of type I AIT (56). Unfortunately, the assessment of CFDS remains highly dependent on the operator and on the equipment despite all efforts (57). More recently, 99mTc Sestamibi scanning of the thyroid has been studied. The rationale for using this technique is the concept that metabolically active thyroid cells of AIT type I would likely accumulate the tracer, while apoptotic or necrotic or stunned cells of type II AIT would not. Indeed in this small study the absence of 99mTc Sestamibi was only seen in type II AIT (58), providing, if confirmed in larger series, an excellent tool for guidance in treatment.
Treatment of type I AIT relies on antithyroid drugs. Unfortunately, the efficacy of these drugs is blunted due to the large iodine burden to the thyroid; therefore, even with maximum doses time to euthyroidism is longer than in simple Graves' disease and toxic multinodular goiter. Potassium perchlorate is a potent inhibitor of the sodium-iodide symporter (59) and is widely used in Europe and other regions to help reducing the thyroid iodine content and therefore sensitizing the drug to thionamides. Several studies have clearly shown the superiority of this regimen above the use of methimazole alone (60). The drug is not commonly available in the United States due to concerns for a cluster of cases of aplastic anemia in the 1960s. However, no reports of this adverse event have surfaced since, in spite of its resurgent use. Sodium ipodate, a potent inhibitor of type I deiodinase, has been used in adjunct to thionamides and perchlorate to rapidly drop T3 levels (61). Ipodate production has been discontinued worldwide, and therefore the drug will become unavailable once the present limited stores are used. Discontinuation of amiodarone is typically recommended with AIT type I, but there is no strong evidence in favor of this step (52). Type II AIT is effectively treated with oral corticosteroids, which lead to a resolution of hyperthyroidism in weeks or months. Discontinuation of amiodarone is probably not necessary with type II AIT.
Drugs Affecting Thyroid Hormone Replacement Therapy
The availability of oral preparations of L-T4 has made correction of hypothyroidism highly effective with peripheral thyroid hormone balance that is almost indistinguishable from normal subjects. However, the efficacy of treatment can be altered by impaired absorption of oral thyroid hormone or by changes in the metabolic clearance rate of absorbed thyroid hormone. Several drugs have been shown to affect these aspects.
Drugs affecting thyroid hormone absorption
Only 60%–80% of ingested L-T4 is normally absorbed. Absorption occurs in the duodenum and jejunum (62) and it requires stomach acidity, as suggested by impaired absorption in patients with achlorhydria (63). Several drugs have been shown to interfere with L-T4 absorption. Antacids such as proton-pump inhibitors, H2 receptor antagonists, calcium carbonate, sucralfate, and aluminum hydroxide have all been shown to significantly reduce absorption of L-T4 to a variable extent. Part of the effect is likely due to the reduction in gastric acidity effected by these drugs, but calcium, aluminum hydroxide, and sucralfate also bind the hormone, reducing its availability to the intestinal mucosa (64). Ferrous sulfate is not known to change gastric acidity but clearly reduces L-T4 absorption, probably through direct binding of the drug (65). Separating the times of intake of L-T4 of these drugs and increasing the dose of thyroxine by 20%–30% is generally sufficient to overcome these problems. Bile acid sequestrants act more profoundly on L-T4 absorption. Both cholestyramine and colesevelam markedly reduce the absorption of L-T4 when given in full doses. In addition, they appear to interfere with the enterohepatic circulation of thyroid hormone, a property that has been used to treat endogenous hyperthyroidism and thyroid hormone intoxication (66 –68). Management of hypothyroid patients on these drugs is complicated by the fact that these drugs are generally administered on a three times a day regimen. Doubling of ordinary L-T4 dose is often required with these drugs and frequent thyroid function tests are recommended.
Drugs altering thyroid hormone metabolism
Besides classical activation and inactivation reactions through the deiodinase systems, thyroid hormone is actively oxidized and conjugated in the liver by the cytochrome P450 system. It is therefore not surprising that drugs activating this system such as rifampin, phenytoin, carbamazepine, and barbiturates result in increased metabolic clearance rate of thyroid hormone. Because of the marked adaptability of the hypothalamic–pituitary–thyroid axis resulting in increased thyroid hormone output, these drugs cause only mild thyroid function test abnormalities in normal subject (69). However, in hypothyroid subjects receiving these drugs, significant increases in the L-T4 dose are required, up to 100% in patients taking rifampin, the most potent in the group. Imatinib is likely to affect thyroid hormone metabolism through a similar mechanism (see above). Because part of this effect is likely mediated through a first-pass effect during absorption form the GI tract, L-T4 doses may need to be decreased if IV administration becomes necessary, but there are no studies on this point.
Estradiol
Treatment with oral estrogen results in a marked increase in thyroxine-binding globulin. The total serum capacity for thyroid hormone is therefore increased. In normal subjects, this results in an increase in total T4 but only in a transient drop in free T4. However, in hypothyroid patients receiving thyroxine, a persistent elevation of TSH is observed and dose adjustments are necessary. This effect is not observed with transdermal estrogen therapy, as the absence of liver passage of the drug results in a much weaker effect on thyroxine binding globulin synthesis (70).
Conclusions
Several drugs are known to alter thyroid function as a side effect of their primary pharmacological action. Some of these effects have been recognized for decades, but novel thyroid–drug interactions are being recognized as new drugs are developed. Risk factors for thyroid–drug interactions have been in some cases identified. For example, the presence of preexisting thyroid autoimmunity increases the risk of lithium-, amiodarone-, and immune-modulator-induced thyroid dysfunction, suggesting the future availability of personalized monitoring screening protocols, at the very least, if not of tailored treatment modifications. Further insight may come from the understanding of the genetic background leading to thyroid–drug interaction, as described in the case of alemtuzumab-induce Graves' disease in MS patients. While many of these developments have not yet lead to changes in the daily practice of clinical thyroidology yet, it is important for the clinician to be familiar with thyroid–drug interaction and their risk factors. Enhanced surveillance may be necessary in patients undergoing therapies known to affect thyroid function.
Footnotes
Disclosure Statement
The author has no financial relationship with any entity that may have financial interests in the topics discussed in this article.
Portions of this review were presented at the Spring 2010 Meeting of the American Thyroid Association, “Thyroid Disorders in the Era of Personalized Medicine,” Minneapolis, MN, May 13–16, 2010.