L-T4 Therapy in the Presence of Pharmacological Interferents

Pharmacological interference on L-thyroxine (L-T4) therapy can be exerted at several levels, namely from the hypothalamus/pituitary through the intestine, where the absorption of exogenous L-T4 takes place. A number of medications interfere with L-T4 therapy, some of them also being the cause of hypothyroidism. The clinician should be aware that some medications simply affect thyroid function tests with no need of modifying the dose of L-T4 that the patient was taking prior to their prescription. Usually, the topic of pharmacological interference on L-T4 therapy addresses the patient with primary hypothyroidism, in whom periodic measurement of serum thyrotropin (TSH) is the biochemical target. However, this minireview also addresses the patient with central hypothyroidism, in whom the biochemical target is serum free thyroxine (FT4). This minireview also addresses two additional topics. One is the costs associated with frequent monitoring of the biochemical target when L-T4 is taken simultaneously with the interfering drug. The second topic is the issue of metabolic/cardiovascular complications associated with undertreated hypothyroidism.

As mentioned below, there may be consequences on health from undertreated hypothyroidism (UTH).
Another study evaluated changes in number of L-T4 prescriptions and dose of L-T4 before and during exposure to potential drug-drug interactions (DDIs) (24). In 5,426 L-T4 users aged ≥18 years (7.5% of persons under care), prescriptions and doses of L-T4 increased during exposure by 6 and 5%, respectively, suggesting that clinicians increase the number of L-T4 prescriptions to achieve TSH levels as low as those before DDIs (24).
A retrospective study evaluated TSH changes in 10,999 Scottish residents who were prescribed L-T4 before starting IM (25). Iron, calcium, PPI, and estrogens increased TSH significantly, with an increase >5 mU/L in 7.5, 4.4, 5.6, and 4.3% patients, respectively. TSH decreased significantly (0.17 mU/L) in patients on statins and changed insignificantly in patients on H2 receptor antagonists or glucocorticoids (25).

THYROID DYSFUNCTION AND SUBOPTIMAL TREATMENT CAUSED BY INTERFERING MEDICATIONS OCCUR IN A PROPORTION OF PATIENTS
The Scottish study (25) is important because it reminds that not all L-T4-replaced hypothyroid patients taking IM become undertreated. For instance, a 3-month supplementation of 1,200 mg/d calcium carbonate reversibly increased TSH in 13/ 20 patients (65%), with TSH >4.0 mU/L in four (20%) (41). Overall, TSH increased from 1.60 ± 0.22 to 2.71 ± 0.43 mU/L (+69%). Calcium carbonate adsorbed T4 dose-dependently at pH 2 but not pH 7.4, thus explaining L-T4 malabsorption (41), and reduced T4 pharmacokinetics (42). Lower L-T4 bioavailability applies to other calcium salts (43). Those data (41) match data of SB (44). Fifty postmenopausal women with L-T4-treated PH started taking 600-1000 mg/d calcium carbonate ≤2 h after L-T4. UTH (TSH >4.12 mU/L) occurred in 9/50 women (18%). Overall, TSH increased from 1.93 ± 0.51 mU/L to 3.33 ± 1.93 28. Physicians who are not endocrinologists, but who are familiar with the diagnosis and treatment of hypothyroidism should be able to care for most patients with primary hypothyroidism. However, patients with hypothyroidism who fall into the following categories should be seen in consultation with an endocrinologist. These categories are (i) children and infants, […] and (ix) unusual causes of hypothyroidism such as those induced by agents that interfere with absorption of L-thyroxine, impact thyroid gland hormone production or secretion, affect the hypothalamic-pituitary-thyroid axis (directly or indirectly), increase clearance, or peripherally impact metabolism.Grade C, BEL 3

(5)
Are there medications and supplements that should not be co-administered with levothyroxine in order to avoid impaired absorption? 3b. We recommend that where feasible, levothyroxine should be separated from other potentially interfering medications and supplements (e.g., calcium carbonate and ferrous sulfate). A 4-h separation is traditional but untested. Other medications (e.g., aluminum hydroxide and sucralfate) may have similar effects but have been insufficiently studied.Weak recommendation. Weak quality evidence. What medications may alter a patient's levothyroxine requirement by affecting either metabolism or binding to transport proteins? 3e. Initiation or discontinuation of estrogen and androgens should be followed by reassessment of serum thyrotropin at steady state, since such medications may alter the levothyroxine requirement. Serum thyrotropin should also be reassessed in patients who are started on agents such as tyrosine kinase inhibitors that affect thyroxine metabolism and thyroxine or triiodothyronine deiodination. Serum thyrotropin monitoring is also advisable when medications such as phenobarbital, phenytoin, carbamazepine, rifampin, and sertraline are started. Strong recommendation. Low-quality evidence.  (10), drugs that "may alter the levothyroxine dose required by a patient" are indicated by one asterisk (*) if the mechanism is "by affecting T4 metabolism or transport", and by the section sign ( §) if the mechanism is "by affecting levothyroxine absorption". The dagger ( †) identifies medications that may trigger hypothyroidism, which is of the central type in the case of bexarotene, and may also be of the central type in the case of medications that act by lowering TSH secretion and causing hypophysitis.
In the literature, interfering medications are categorized variably, depending on mechanism, with some medications having multiple mechanisms. For instance, the 2012 American Thyroid Association (ATA) guidelines (4) considered four mechanistic categories, with information provided when mechanisms are multiple, namely: (i) direct and indirect effects on the hypothalamic-pituitary-thyroid axis; (ii) thyroid gland hormone production and secretion; (iii) increased clearance; (iv) interference with absorption. In the 2014 ATA guidelines (5), categories were two, viz. (i) impaired absorption, and (ii) altered metabolism or binding to transport proteins, with emphasis on the first category. In a book (20) medications are listed under five categories: (i) central TSH inhibition; (ii) absorption; (iii) synthesis and secretion; (iv) transport; (v) metabolism. In another chapter of this book (10), emphasis is given to drugs that may alter the L-T4 dose by affecting (i) T4 metabolism or transport, and (ii) L-T4 absorption. In another book (22), drugs are listed under two major categories, viz. interfering with (i) hypothalamic-pituitary-thyroid function, (ii) thyroid function. Whenever "thyroiditis" appears in the second column of the Table (Mechanism), the clinical implication is that such silent thyroiditis may translate into monophasic thyrotoxicosis, monophasic hypohyroidism, or biphasic dysfunction (thyrotoxicosis followed by hypothyroidism). However, hypothyroidism may sometimes be permanent. In the case of lithium, a few cases of thyrotoxicosis have been reported. Interferon alpha, alemtuzumab, ipilimumab, tremelimumab, nivolumab and pembrolizumab may even trigger true Graves' disease. Clearly, whenever a L-T4-treated hypothyroid patient experiences increased discharge of thyroid hormones (thyrotoxicosis or hyperthyroidism), because he/she is simultaneously taking a medication that causes such side-effect, L-T4 therapy has to be withdrawn. Upon rechecking thyroid function tests during the follow-up period in such patient, L-T4 replacement is started again. mU/L (+73%), but when all women took calcium 6-8 h after L-T4, all had TSH <4.12 mU/L (2.16 ± 0.54 mU/L) (44).
In 71 L-T4-treated PH patients who started the antituberculosis drug rifampin, an increased L-T4 dose was required for 50% of 46 patients (TSH-suppressive group) and 26% of 25 patients (replacement group) (47). Lack of thyroid remnant, time interval between starting rifampin and TSH measurement, and baseline L-T4 dose/kg body weight were significant risks for UTH (47). Ethionamide and paraaminosalycilic acid (PAS) are used in multidrug-resistant tuberculosis (MDR-TB). After initial reports (48)(49)(50), a metaanalysis on 6,241 MDR-TB patients (31) showed that PH prevalence in MDR-TB patients averaged 17.0%, with ethionamide and PAS were the most frequently reported drugs associated with hypothyroidism. Tubercolosis is a common opportunistic infection in HIV-seropositive persons, and antiretroviral therapy (ART) may induce PH. Of 69 HIVinfected MDR-TB patients under anti-TB and antiretroviral therapies, 37 (54%) had PH (51). Co-administration of PAS and ethionamide doubled the risk of hypothyroidism (RR = 1.93) (51). In MDR-TB patients receiving anti-TB, one-fourth developed PH: 32% in patients who received a regimen containing ethionamide, 35% in patients who received a regimen containing PAS, and 44% in HIV-positive patients on ART (52).
OH or SCH develop in up to 15 or 34% of lithum-treated patients (20), the annual rate of developing PH being 1.5%, (6.4% in thyroid antibody-positive and 0.8% in thyroid antibodynegative individuals) (70). Women <60-year-old are at greatest risk to develop thyroid disease (71). Indeed, women develop OH or SCH three times more frequently than men (25.8 vs. 8.7%), with prevalence among women exceeding 50% by the age of 65 years (35). During tricyclic antidepressant therapy, TSH levels are unchanged, and T4 and FT4 levels decrease within the euthyroid range, though other studies reported unchanged thyroid function tests (TFTs) (72). Selective serotonin reuptake inhibitors (SSRIs) affect TFTs variably, usually with no or downward changes of FT4 and FT3, and no or upward change of TSH within the corresponding reference range (72). L-T4 requirements increased in nine L-T4-treated hypothyroid patients under sertraline (73). Confirming the sertraline-T4 interaction, in two patients under TSH-suppressive L-T4 therapy, TSH levels rose into the normal range (73). A 3month duration study was conducted in 57 patients with major depression and 10 control patients (72). The study patients were hypothyroid on adequate L-T4 therapy (n = 28) who were randomized to fluoxetine (n = 13) or sertraline (n = 15), and euthyroid (n = 29) who were treated with fluoxetine (n = 15) or sertraline (n = 14). Controls had hypothyroidism on adequate L-T4 therapy without depression (72). No changes occurred in the L-T4-replaced hypothyroids under either SSRI. In response to a letter (74) commenting that difference with the early study (73) could be that" many … patients … were athyreotic", the authors admited that "in most [patients] the cause of hypothyroidism was autoimmune" (75). Noteworthy, a 41-yearold man with a history of bipolar disorder and schizophrenia had myxedema coma after therapy with sertraline and ariprazole (76). After discharge, TSH remained high (34 IU/ml) on sertraline, ariprazole and 200 µg/d L-T4.
The anti-obesity drug orlistat inhibits gastro-intestinal lipases and is minimally absorbed. Based on data of the UK Medicines Information pharmacists (79), only two cases of interaction between orlistat and L-T4 were reported (80,81), most likely via T4 malabsorption. It was recommended that L-T4 and orlistat "should be separated by 4 hours, and increased monitoring of […] thyroid hormone levels may be prudent" (79).
Metformin lowers serum TSH in L-T4-treated OH and L-T4untreated SCH patients, but not in euthyroid patients (33). After its chronic administration, FT4 was unchanged and TSH decreased in L-T4-treated or L-T4-untreated hypothyroid diabetics, but not in euthyroid subjects (86).
A single oral dose of clofibrate decreased significantly hyperthyrotropinemia of PH patients (87). Clofibrate did not change discernibly basal and TRH-induced TSH secretion in euthyroids. Similar results were given by meclofenoxate hydrochloride (87), suggesting that both drugs inhibit TSH secretion in PH patients possibly acting on the hypothalamus/ pituitary (87). Clofibrate also increased serum T4-binding capacity of TBG, lowering serum FT4 in 11/12 hyperlipoproteinemic patients (92%) (88).
The adrenocytolytic drug mitotane is used to treat adrenocortical carcinoma (ACC) (89). In 17 patients with radically resected ACC, mitotane was administered associated with glucocorticoid replacement therapy (90). Excluding three patients under L-T4 treatment and one with SCH, during the first year, FT4 became subnormal in 12/13 evaluable patients (92%), with L-T4 replacement started after 9 months in four. At last follow-up, FT4 levels were unchanged compared with the 12 month-evaluation, but another three patients needed L-T4 replacement. Five women with mitotane-treated ACC showed features of CH, namely low FT4, normal FT3 and TSH, with impaired TSH response to TRH (91). Mitotane increased serum FT3/FT4 ratio, suggesting enhanced T4 to T3 conversion, a compensatory mechanism of hypothyroidism (91). CH was reported in a girl with mitotane-treated ACC (92); full restoration of FT4 required increasing the L-T4 dose. After completing chemotherapy, TFTs remained normal, and L-T4 replacement was discontinued.
High-normal TSH levels impact unfavorably on mortality. In 9,020 adults, SCH (TSH >5.60 mIU/L) and high-normal TSH (1.96-5.60 mIU/L) were associated with increased all-cause mortality (HR = 1.90 and 1.36) vs. the middle-normal TSH group (1.20-1.95 mIU/L) (104), with CVD mediating 14.3 and 5.9% of the association, respectively (104). A study on 611 hospitalized elderly patients evaluated all-cause mortality up to 66 months after discharge (105) and concluded that (i) in treated hypothyroid patients, median TSH levels of 5-10 IU/L associated with increased mortality, (ii) treatment should aim at achieving euthyroidism to improve survival (105).

POTENTIAL FUTURE DEVELOPMENTS
In the "Areas for future research" heading of the 2012 ATA guidelines (4), a section was devoted to "Agents and conditions having an impact on L-thyroxine therapy and interpretation of thyroid tests". Except for reminding that the residual functioning thyroid tissue is a major factor for a given IM to cause thyroid dysfuntion and L-T4 dose adjustments, no ideas were presented (4).
Future developments have already occurred considering the availability of L-T4 formulations (liquid, softgel) that are refractory or much more resistant to IM than tablet L-T4 (9,14,98,100,(109)(110)(111)(112)(113)(114). In the author's opinion, their use seems preferable to the strategies of (i) increasing stepwise the dose of L-T4 (with associated frequent monitoring of TFTs and risk of iatrogenic thyrotoxicosis if the IM is decreased in dose or withdrawn); (ii) adding supplementation with either 1 g/d (115) or 0.5 g/d (116) vitamin C to acidify the intragastric pH. In the Argentinian study, the 28 patients had no known cause for UTH (115), while in the Colombian study, the 31 patients had endoscopy/gastritis-proven gastritis (116). Both 2-month-long trials with vitamin C (115,116) lack formal pharmacokinetics studies and challenge of patients with IM-associated UTH. Indeed, coadministration of acidic beverages is one way of solving the problem of decreased bioavailability of drugs whose bioavailability under conditions of increased intragastric pH (117). Further to T4, there are a number of other drugs with decreased absorption at high intragastric pH, such as ketoconazole, itraconazole, atazanavir, cefpodoxime, enoxacin, dipyridamole, raltegravir, alendronate, digoxin, and nifedipine, to name a few (117). The other way is the development of "formulations that can minimize or mitigate the effects of increased gastric pH on the bioavailability" (117).
Considering the magnitude of polypharmacy, particularly in the elderly (118), and the aforesaid unfavorable impact of UTH on metabolic and CVD outcomes, more research is needed to substantiate those outcomes in the setting of IM-associated UTH. Once pejorative outcomes are confimed, monitoring of hypothyroid patients under IM should be tightened, with the biochemical monitoring not restricted to TSH solely.

AUTHOR CONTRIBUTIONS
SB made the work, drafted the article, revised it, and gave the final approval for the publication.