Introduction

Endogenous production of thyroxine ([T4] tetraiodothyronine/3,5,3′,5′ tetraiodothyronine]) and triiodothyronine ([T3] 3,5,3′ triiodothyronine) is the primary function of the human thyroid. The chemical structure of levothyroxine (T4) was first identified by Edward Kendal in 1914 [1]. Thyroxine was successfully synthesized in 1927 by Harrington [2, 3]. As the synthesized molecule was an acid, oral absorption was not efficient which limited the clinical utility of the hormone until a sodium salt of thyroxine (LT4) was developed in the 1950’s providing an orally administered preparation for clinical use [4]. Triiodothyronine (T3) structure and synthesis (LT3) occurred in 1952 [5–7]. In 1970, Braverman, Sterling and Ingbar documented the endogenous generation of T3 from LT4 in athyreotic subjects [8] This finding provided the mechanism underlying the utility of LT4 mono-therapy. Clear evidence exists that T3 is predominately (80%) produced by this peripheral conversion due to the activity of 5′-deiodination of T4 and only 20% of T3 in the circulation is secreted either after intrathyroidal deiodination of T4 or directly by the thyroid [9]. T3 is the most active thyroid hormone with affinity for the thyroid hormone nuclear receptor 10–20 fold higher than that of T4 [9]. It has been abundantly demonstrated that physiologic levels of T3 are generated in athyreotic subjects treated with replacement doses of LT4 [9–11]. Once in the circulation, synthetic levothyroxine (LT4) and Liothyronine (LT3) are indistinguishable from the two thyroid hormones produced in the human body [12].

The development of thyroid hormone as a useful clinical intervention

Thyroid treatment has been comprehensively reviewed by Slater and dates as far back as 239 BCE in China and the second century of the common era (CE) in the western literature [13]. Chinese cretins were treated with sheep thyroids in the 6th century CE [13]. The observation that thyroidectomy resulted in the onset of myxedema (cachexia strumipriva) was proposed as causative in 1883 by Semon [14] and reported by Kocker [13, 15]. This association was based on a survey of surgeons published in 1888 [16]. Transplantation of animal thyroid tissue into a myxedematous patient was reported in 1890 by Bettencourt and Serrano [17, 18]. A prompt clinical response, including an increase in temperature and positive clinical effects lead to the conclusion that the effect was due to absorption of components of the implanted tissue rather than tissue incorporation into the recipients body [13, 15]. These Portuguese investigators also published the initial report of subcutaneous injections with an extract of animal thyroid in 1890 [13, 15, 19]. George Murray of Newcastle-upon-Tyne described his initial experience with thyroid extract derived from two and one half sheep thyroids injected intermittently over a few weeks in 1891 [20]. Murray went on to report clinical outcomes of further patients illustrating the effectiveness of this approach and documenting the first fatal cardiac complications after the initiation of thyroid hormone treatment [13, 21]. Among the clinical benefits listed from these early reports were rapid weight loss, measurable increases in body temperature, urine volume and nitrogen excretion which occurred when thyroid extract was administered to clinically myxedematous patients [22]. Preparation of these injectable thyroid extracts was very demanding, time consuming and expensive, additionally reports of acute and chronic complications associated with the injections were coming to light [13, 23]. A transition to the preparation of extracts for oral ingestion improved the consistency of clinical outcomes [13, 24–26]. Within a few years, dosing guidance and cautions in regard to iatrogenic over-dosage appeared [13, 27]. Patients were “titrated” to clinical endpoints and considered adequately treated if no longer complaining of hypothyroid symptoms, and not demonstrating obvious thyrotoxic symptoms when evaluated by experienced physicians [28]. The symptom as an endpoint has subsequently been deemed unreliable due to the non-specificity of “hypothyroid symptoms” [29–31]. Measures of success including body weight, pulse rate, skin quality, sensitivity to cold and quantitative measurement of Achilles tendon relaxation time were documented [32]. The danger of over-dosage was recognized very early on [21] and resulted in recommendations that treatment start with a low dose to be gradually increased as required and that caution be used in the elderly and those with coronary artery disease [27].

Observations of physiology, clinical assessment and pharmacokinetics on extracts, LT4 and LT3

Outcomes showed that ingestion of thyroid hormones by biochemically normal but symptomatically “hypometabolic” individuals resulted in no increase in their basal metabolic rate (BMR), but BMR decreased significantly for weeks following the discontinuation of thyroid intake [33], indicating a suppression of normal thyroid function by the ingestion of exogenous thyroid hormones [34].

Thyroid function testing eventually grew beyond measurements of BMR, thyroidal 131-I uptake and protein-bound iodine (PBI) as reported in the 1960’s [35]. In the era prior to reliable TSH testing, PBI was difficult to interpret in those treated with thyroid extracts when BMR had been restored to goal. PBI varied widely dependent on the particular thyroid hormone preparation ingested. BMR normalizing doses of triiodothyronine (T3) for example were associated with sub normal PBI while BMR normalized with l-thyroxine (LT4) resulted in substantial elevation of PBI [35] compared to the PBI measured during thyroid extract use [36]. Additionally, PBI could vary substantially when using thyroid extract from different manufactures [35]. The measured PBI in the circulation of those on thyroid extracts was accounted for primarily by the thyroxine content. Due to variation in T3 content, individuals with low or normal PBI could experience thyrotoxic symptoms thought to be dependent on the T3 content of the product being ingested [37]. Such content variability lead to calls for an a replacement of animal thyroid hormone preparations with synthetic fixed dose preparations of LT4 and LT3 [38]. Others attributed the clinical variability observed during extract therapy to compliance issues, concluding that serum TSH was more reliable than PBI [39].

Radioimmune assays of circulating thyroid hormones developed in the early 1970’s, reliably measured both T4 and T3 levels in the serum [40]. The T4 and T3 profiles of euthyroid normal controls and hypothyroid patients ingesting different types of thyroid hormone preparations were documented [40]. Controls, had essentially constant T4 and T3 levels throughout the day [40]. Following LT3 ingestion, athyreotic subjects demonstrated substantial increases of T3 within the first 6 h post ingestion, returning to baseline by 24–48 h. When athyreotic individuals were stabilized on ingested LT4, minor increases of circulating T4 levels were observed up to 6 h after the ingestion while T3 levels remained normal and constant. Hypothyroid patients ingesting thyroid hormone extract, demonstrated supraphysiologic surges of T3 levels in the first 10 h after ingestion while little if any change in T4 levels occurred [40]. The conclusion of these studies was that the pattern of both circulating thyroid hormone levels was close to physiologic only after the ingestion of l-thyroxine alone [40].

The gradual transition from thyroid extracts to levothyroxine use

Advances in the pharmaceutical manufacture of thyroid hormones made synthetic LT4 and LT3 available for clinical use in the 1950s. Initially, these products were introduced into the US market without the need for United States Food and Drug Administration (FDA) approval as they were viewed as forms of thyroid hormone contained in thyroid hormone extracts that had been on the market prior to 1938. Because thyroid replacement therapy predated FDA regulation, extracts were not subject to the new regulation in the US. In 1949 the sodium salt of LT4 was synthesized and noted to be more readily absorbed [41]. The first introduction of regulated Sodium-levothyroxine products in North America occurred in 1951 [42]. Appropriate thyroid hormone replacement therapy before 1970 [8] was thought to require the administration of both LT4 and LT3 to mimic normal thyroid function. Several studies of LT3/LT4 combinations appeared in the late 1960s and early 1970s to explore this approach as an alternative to thyroid hormone extracts [28, 32, 43] During this time, a transition towards LT4 monotherapy was the next innovation. Smith et al. reported on hypothyroid patients treated with LT-4 mono-therapy for primary hypothyroidism and judged their clinical status after a minimum of 6 months on stable dosing to be clinically euthyroid [44] Although 18.4% preferred the combination of LT4/LT3, 33% preferred LT4 alone and the remaining 48.3% were apparently satisfied with both treatments [44]. These authors concluded that many subjects were unable to distinguish treatment with LT4 from LT4/LT3, that few (18%) preferred LT4/LT3 and that at these doses, there was a high incidence of side effects among those using the combination and many (33%) preferred L-T4 therapy which was well tolerated [44].

At this time, the peripheral conversion of LT4–T3 was documented in experiments performed by Braverman, Ingbar and Sterling and published in 1970. Eleven Hypothyroid patients, seven of whom had postoperative hypothyroidism and three had been treated with 131-I for Graves’ disease were all receiving clinically adjusted doses of LT4 monotherapy ranging from 50–600 mcg/day. Triiodothyronine (T3) was found in the sera of all with levels ranging from 243–680 ng/100 ml (normal 170–270) [8]. The amount of T3 detected was far in excess of any potential LT3 contamination of the LT4 they were ingesting. The conversion of exogenously administered LT4 as the source of the T3 was confirmed by applying 125I labeled LT4 both orally and intravenously by detecting 125I labeled T3 in the circulation thereafter [8]. Based upon these studies, the original premise that synthetic thyroid hormone products could achieve comparable clinical results to thyroid hormone extract products, either in combination or as LT4 alone was documented. The observations of Braverman et. al. provided a solid mechanism to explain the documented ability of LT4 monotherapy to normalize thyroid function. The subsequent work of Surks et al. confirmed the obvious endogenous generation of T3 in those treated with LT4 monotherapy and documented nearly physiologic diurnal levels of T4 and T3 after the chronic ingestion of LT4 alone [40]. This peripheral generation of physiologic circulating T3 levels in surgically hypothyroid individuals has also been confirmed more recently [10].

The migration of clinical treatment away from Thyroid hormone extracts and combinations to LT4 monotherapy

The introduction of serum TSH testing to clinical practice revealed heretofore unrecognized abnormalities in thyroid function among those on extract [45–50] and further energized the academic community to recommend abandonment of thyroid hormone extracts in favor of LT4 monotherapy [45]. This was accepted by the clinical community without major resistance once the ease of administration, the simplification of monitoring with TSH was widely used and reports equating clinical outcomes with LT4 monotherapy to those without thyroid disease confirmed both safety and efficacy. One such study reported outcomes of middle-aged women treated with LT4 for 1–28 years duration. A 12-year follow-up assessment captured the clinical status of 99.7% of participants [51]. The incidence of myocardial infarction, diabetes mellitus, cerebrovascular accidents, cancer, and death in women treated with LT4 was the same as euthyroid controls. Eighty three percent of these women were still on LT4 at follow up and 22 of these 24 (92%) had TSH and TT3 within the normal range. When the LT4 treated hypothyroid women were compared with control subjects, identical laboratory and clinical data were observed with the exception of a higher BMI, taller stature and lower cholesterol. LT4 treated subjects had the same life quality estimate as controls [51].

The evolution of thyroid hormone dosing with LT4 monotherapy

Contemporary, individually tailored [52] dosing of LT4 using regulated preparations and applying today’s assessment tools including the serum TSH to determine the patients clinical and biochemical status is substantially different than when LT4 was initially introduced. Several of the initial studies cited above used LT4 doses that were so high that they would currently raise suspicion of noncompliance or gastrointestinal malabsorption. Several developments resulted in our current thinking about LT4 mono-therapy dosing. The most obvious clinical innovation was the use of precise TSH and thyroid hormone level measurements. As symptom based clinical assessments gave way to biochemical investigation, clinicians recognized clear evidence of both overdosage (suppressed TSH or TSH response to thyrotropin releasing hormone [TRH] and/or elevated T3 or T4 levels) [8, 45] as well as under dosage (elevated TSH or exaggerated response to TRH) [53] when patients appeared to be “clinically euthyroid”. With the aid of biochemical assessment, the estimated LT4 replacement dose declined steadily from the 300+ mcg daily (greater than 3 mcg/kg) range [8, 44] to more moderate doses of 1.6–2.0 mcg/kg daily [54–61] and it was observed that adequate thyroxine dosing required less per kilogram body weight in the elderly [50, 56, 57]. Following changes in the quality control process of commercial products prior to release [62] (see below), the LT4 content of tablets being ingested became more uniform and predictably similar to the labeled amount [61–63]. As a result of this, subjects previously estimated to have a particular mcg/kg dose requirement based upon the labeled amount by USP quality standards, were noted to achieve similar control at lower doses of the newer generation of LT4 preparations [61]. As a result of these processes, the contemporary general recommendation for estimating full replacement dosage of LT4 is 1.6–1.7 mcg/kg/day [53, 61, 64, 65].

Refinement of LT4 manufacture, regulation and assessment of bio-equivalency

Prior to 1980 the iodine content of thyroid hormone tablets (and thyroid extract tablets as well) was considered the standard by which the USP assessed the potency of thyroid hormone tablets [53]. This was a direct result of the FDA assumption that thyroid hormones were an established therapeutic intervention which were in general use prior to the FDAs establishment and therefore beyond the scope of FDA regulation [53, 66]. As such, LT4 and for that matter LT3 were initially recognized as components of thyroid hormone extract, on the US market when the FDA was formed, and prior to 1997 were not subject to oversight regulation. This USP standard, based on analytic methods available at the time of introduction, did not require quantification of actual thyroid hormone content in LT4 tablets but rather considered tablets from a particular manufacturer or batch to meet the standard if the iodine content fell within the prescribed limits [63]. From their introduction on the US market, each LT4 manufacturer conducted their own business following USP standards for pharmaceutical products, making changes to their manufacturing processes and formulations to maintain this iodine standard without FDA oversight [67]. This unregulated process of independent manufacturing and individual product production resulted in significantly conflicting results when various preparations were compared with one another [63, 68]. Tablet analysis for actual LT4 content per tablet often detected substantial differences in content when several brands were evaluated [61, 63, 69, 70] and occasionally significant batch differences when the same brand was evaluated [71].

In 1980, high performance liquid chromatography (HPLC) quality control of LT4 was introduced by one manufacturer [53, 62] resulting in the distribution of one LT4 product at nearly 100% of advertised potency [63]. A second manufacturer, adhering to the prevailing USP standard, distributed a different name brand LT4 product that had consistently less LT4 content [61, 63]. Both of these products met the USP iodine content standard but HPLC analysis was necessary to detect the differences in LT4 content [63]. Both were internally consistent but when substituted for one another, differences in clinical outcomes occurred [53, 63, 72, 73]. The differences in clinical outcomes narrowed after [53] HPLC quality control was utilized by both products. Subsequently, USP adopted an HPLC standard to assure predictable LT4 content for all products distributed in the US.

Extracts from animal thyroid glands had been shown to be useful since at least 1891 so were already being utilized when the FDA was formed in 1906. Thyroid hormone extracts were the sole option in use when the Federal Food, Drug and Cosmetic act of 1938 enhanced the regulatory authority of FDA [74]. The Kefauer-Harris Amendments of 1962 were introduced in the wake of the thalidomide malformations reported after use in pregnancy. These amendments required FDA to assess the efficacy and safety of drugs introduced to the US market since 1938. As Thyroid extract use predated this cutoff, only thyroid hormone products introduced subsequently were subject to this higher level of regulation even though FDA declined to do so initially until some cause for concern was found [74].

Following receipt of 58 adverse drug experience reports, and because of levothyroxine sodium’s narrow therapeutic index, FDA concluded that it was particularly important that the amount of available active drug be consistent for a given tablet strength. Furthermore, it was observed that products had failed to maintain potency through the expiration date and that the active ingredient of the same dosage and brand varied from lot to lot [75]. A message in the Federal Register Vol. 62, No. 157:43535–8, 14 Aug 1997 alerted the thyroid community to a major change in the manufacture, distribution and use of LT4 products. Due to these concerns of stability and potency of preparations distributed in the US, FDA declared LT4 products to be new drugs, indicating they would subject to completing new drug applications (NDA) to meet FDA manufacturing, safety and consistency standards in order to be allowed to be further distributed in the US [75]. A comprehensive series of guidance to industry documents were created to illuminate the path to FDA approval of all LT4 products [76, 77]. Initially 7 LT4 preparations (Table 1) were approved under this NDA process, each of which was recognized as a unique product which met the strenuous FDA NDA standards [76, 77]. At the time of these NDA approvals, none of these products were considered interchangeable with any other as the FDA process for assessment of bio-equivalency and therapeutic equivalence had not been implemented.

Table 1 Initial NDA and ANDA approved products, current therapeutic equivalence (AB) ratings
Full size table

The process of evaluating the potential interchangeability among LT4 products followed next. Bioequivalence (BE) studies were put forward by FDA as the means of evaluating substitution of one product with another at the same dosage and became a critical component of the abbreviated new drug application (ANDA) submission process. The ANDA is the direct route onto the pharmaceutical market place for generic versions of NDA approved medications. The purpose of ANDA studies is to demonstrate BE between pharmaceutically equivalent (same active ingredient, form [i.e. tablet], dose etc.) generic drug products and the corresponding NDA listed reference drug [78]. Bioequivalence is defined as the absence of a significant difference in the rate and extent to which the active ingredient (AI) becomes available at the site of drug action when administered in the same molar dosage [78]. The expressed purpose of determining BE is to then extrapolate these results and designate the generic medication as therapeutically equivalent (TE) to the NDA product. This is to assure that bioequivalent products can be substituted for each other without concern for potential adjustment in drug dose or the need for any follow up therapeutic monitoring [78]. The FDA maintains that the most efficient method of assuring TE is to assure that the formulations perform in an equivalent manner. Therefore for two orally administered drug products to be bioequivalent, the test product should exhibit the same rate and extent of absorption of AI as the reference drug product.

These judgements are made on the basis of pharmacokinetic (PK) studies, understanding that measuring the active ingredient at the site of action is not possible and assuming some relationship exists between the efficacy/safety and concentration of active moiety in the systemic circulation [78]. Ultimately, PK measurements are viewed as a bioassay assessing the release of drug substance from the drug product into the systemic circulation and constitute the process of BE determination as no clinical verification is required for two drugs to then be designated as TE. FDA created a specific LT4 PK protocol [78]. But a study by Blakesely et al. predicts that this approach of determining bioequivalence results in absorbed dose differences of 12.5% not being detected, thus allowing products to be considered equivalent when more than 10–12% differences are present [79].

Once designated as BE, therapeutic equivalence is assumed, and a TE code is assigned. A codes are assigned to drug products that FDA considers TE, for example AA is assigned when no known or suspected BE problems have been reported or exist while the AB code would be used if actual or potential BE problems have been resolved with adequate in vivo and/or in vitro evidence supporting BE (PK testing). Once designated as AB, the medications can be freely substituted for one another at the pharmacy unless this practice is specifically restricted by the physician. On the other hand, B codes are assigned to drug products FDA considers NOT TE (non-interchangeable), in the case of LT4, products are labeled as BX when the data are insufficient to determine TE (presumed inequivalent). The BX rating is applicable when two products have not been compared pharmacokinetically or if the PK data generated fails to meet the standard of bioequivalence. Once these standards were in place, a generic manufacturer demonstrated PK data consistent with therapeutic equivalence to initially one, then 2 other name brand LT4 products eventually going through the process of BE determination with 4 name brands. As outlined in Table 1, as each reference product was challenged, the AB rating was expanded to include AB1 to indicate TE with Unithroid®, AB2 for Synthroid®, AB3 for Levoxyl® and eventually AB4 to communicate interchangeability with Levothroid® which had been NDA approved as Thyro-Tabs® but then distributed under the Levothroid® name.

As previously stated, 7 medications had originally been NDA approved. As outlined in Table 1, four NDA approved products were until recently distributed as name brand products (Thyro-Tabs [Levothroid®] has been discontinued as of a 11 September 2016 review of the FDA Orange Book), two products (Levo-T [distributed by Sandoz] and Unithroid) have been distributed as generic products and at least 2 levothyroxine products were approved through the ANDA process with LT4-Mylan currently rated interchangeable as AB1-4 and the other product is currently off the US market. By my count that amounts to 5 LT4 tablet products currently available in the US. Additionally a LT4 gel-cap product (Tirosint®) has been NDA approved by FDA and 3 powdered formulations are available for intravenous use (Table 1). Tirosint® is a unique product which is available in the US as a soft gel capsule and in Europe as a liquid LT4. In addition, a soft gel preparation of levothyroxine is now also available in Italy and likely elsewhere in Europe distributed as Tiche® (IBSA Pambio Noranco, CH) [80]. These two forms of LT4 have been reported to have enhanced absorption when ingested with meals and in the face of gastroenterological conditions associated with variation of absorption [81]. In a comparison with thyroid hormone tablets the soft gel form was demonstrated to release LT4 in a pH independent manner thus assuring LT4 in solution for absorption in the small bowel independent of gastric acidity [82]. In 2 studies, the soft gel LT4 has demonstrated more LT4 bioavailability when ingested by patients with gastric disorders [83–85] or when taken simultaneously with coffee [86]. This enhanced absorption results in lower LT4 doses of soft gel being required to maintain thyroid status compared to doses observed when the same patients ingested tablets. Levothyroxine in liquid form is available in Europe and has been demonstrated to be effective in pediatric patients and several observational studies of adults have concluded that liquid LT4 is both well absorbed and effective in clinical practice [81]. Of note, as both of these products are derived from LT4 solutions, bioavailability would be anticipated to be greater than that seen when ingesting LT4 tablets as no dissolution of the tablet excipients must occur before the LT4 is available for bowel absorption. As such, lower doses of LT4 are generally necessary than those listed above for traditional tablet based therapy [81].

Further improvement of LT4 product performance

FDA recognized the narrow range within which LT4 must be titrated. As a result, FDA approved a requirement that LT4 products meet a 95–105% potency specification throughout their shelf life to ensure that the LT4 potency not degrade more than 5% of labeled potency amount (http://www.fda.gov/cder/drug/infopage/levothyroxine/qa.htm accessed 19 September 2016). Manufacturers were to meet these new standards by October 2009. It was anticipated that formulation or manufacturing changes would be made by some manufacturers in order to meet the 95% potency standard. All involved believed that this should further improve outcomes of LT4 Rx performance in clinical practice (http://www.fda.gov/cder/drug/infopage/levothyroxine/qa.htm accessed 19 September 2016). As a consequence, the brand name Levoxyl® was reformulated and in 2009 submitted pharmacokinetic data to FDA demonstrating that the new Levoxyl® was similar to the previous formulation [87]. The branded gel-cap LT4 product Tirosint® also reformulated, providing PK verification that the new Tirosint® was similar to the NDA approved product [88]. The PK relationship of AB3 rated generics to the new Levoxyl® was not reassessed so the original ANDA pairings have remained unaltered. In summary, FDA regulation of the LT4 market has resulted in the availability of well manufactured products meeting well defined standards and providing a framework assuring ongoing surveillance of product quality. The tightening of LT4 product potency to a 95% standard rather than the 90–110% standard generally applied to other pharmaceutical products, reassures the clinician that patient outcomes should be only minimally affected from refill to refill of the same product. Several of the LT4 products do have similar pharmacokinetic performance however this is not assured with each of the AB rated pairs [89] so some variation of clinical outcomes are likely expected if subjects are freely substituted from refill to refill [90].

Optimal use of levothyroxine preparations

Patients treated with levothyroxine are assumed to be easily titrated and maintained in a euthyroid state. In actuality, most cross sectional analyses of outcomes in independent populations reveal that less than optimal outcomes are seen in 32 [91]–54% [92] of individuals. As few as 5% [93] and as many as 38% [92] of individuals have been reported to be overdosed with LT4. On the other hand, 15 [94]–27% [95] are noted to be under replaced. An obvious consideration is the potential that adverse clinical outcomes result in those with less than optimal biochemical control. A study of 17,684 patients on LT4 clinically characterized individuals based on weighted TSH outcomes [96]. Compared to those considered euthyroid, individuals with suppressed or elevated TSH had significantly higher incidence of cardiovascular, dysrhythmia or fracture related admission or death [96]. Another investigation involved participants in a longitudinal study of aging [97]. Of 1029 individuals euthyroid at baseline, 87 were being treated with LT4. Fifty eight NON Cancer patients were started on LT4 during the follow up period and were expected to be euthyroid as a therapeutic goal. New onset of thyrotoxicosis was observed in 5/58 (8.6%) of those newly treated with LT4 and in 5/87(5.7%) subjects continuing LT4 during follow up. Compared to initially euthyroid controls, those treated with LT4 demonstrated an 11.8 fold increased risk of developing thyrotoxicosis [97]. Older caucasian females were significantly more likely to be overdosed with LT4 [97].

Suboptimal outcomes are related to compliance irregularities as about 1/3 of individuals report being non-adherent often, or very often [98]. Patients do not take medications as prescribed as they may doubt the efficacy of the treatment; intend to save money or stop when symptoms are abated. About 30 % do not understand the need to continuously take LT4 to maintain euthyroidism [99, 100]. More than 60% simply forget to take the pills on a regular basis [98]. Kesselheim et al. examined factors involved in self-discontinuation of newly prescribed medications and found that differences in pill shape and color with refills were a source of confusion and diminished adherence with therapy [101]. To enhance compliance, steps may be taken such as informing the patient of the effects of not following the treatment and avoiding unnecessary changes in treatment regimens such as substitution of one source with another in the pharmacy. Regular pharmacy follow-up can improve adherence by over 30% [98, 102]. Other strategies include the use of a daily pill box with instructions that at the end of the week all 7 LT4 pills should have been ingested assuming daily dosing. Additionally putting pills next to ones toothbrush or alarm clock, setting reminders on cell phones, watches, posting notes may be of benefit. Finally, consideration may be given to supervised once weekly (7× the daily dose) LT4 administration when other interventions are ineffective [98, 102].

Interference with LT4 dissolution and absorption has substantial impact on clinical outcomes. Food is the most common factor associated with a disturbance of thyroid hormone effectiveness and consistency. Ingestion of LT4 with a meal interferes with effective TSH suppression when administered for goiter growth inhibition [103]. When co-administered with a standard test breakfast, studies document a 38–40% reduction of absorption of LT4 [104]. In a study of 65 hypothyroid subjects ingesting therapeutic LT4 doses, subjects were randomized to ingesting their LT4 in the fasting state 1 h before breakfast (BB), at bedtime at least 2 h after their last meal (HS) or along with breakfast (WB) [105]. TSH values were significantly lower and more consistent while LT4 was ingested in the fasting state (BB) compared to at bedtime and both approaches resulted in lower TSH than when LT4 was taken as part of breakfast [105]. Similar findings were reported by Silva Perez et al. [106]. Again, LT4 ingestion after an overnight fast 60 min prior to breakfast was compared to LT4 ingestion with breakfast. TSH levels were higher when LT4 was ingested with breakfast and ingestion of LT4 while fasting was preferred by a plurality [106]. Medications have been demonstrated to influence the absorption or metabolism of LT4 [64]. Table 2 lists medication interference to be considered in the differential diagnosis of suboptimal outcomes in LT4 treated patients. Estimation of the clinical impact of commonly used medications has been assessed in patients on LT4 therapy indicating that iron, calcium, proton pump inhibitors and estrogen supplements are associated with increases in TSH while statin use may result in TSH decline and no significant effect of H2 blockers nor Glucocorticoids are commonly observed in routine practice [107].

Table 2 Medications reported to impact levothyroxine outcomes
Full size table

Pregnancy changes binding proteins and thyroxine metabolism resulting in inadequate LT4 replacement raising concerns of maintaining adequate thyroxine exposure for the fetus during the first trimester [108]. Recommendations for accommodating these predictable changes in thyroid hormone requirements include independently increasing the dose of LT4 by 25–30% upon recognizing a missed menstrual cycle or positive home pregnancy test by adjusting the LT4 from once daily dosing to a total of nine doses per week (29% increase) [108]. Further, treated hypothyroid patients (receiving LT4) planning pregnancy should have their dose adjusted to optimize serum TSH values to <2.5 mIU/L preconception [108].

Gastrointestinal malabsorption of LT4 may result from celiac disease [109], lactose intolerance [110–112], pancreatic insufficiency, obstructive liver disease, cirrhosis of the liver, small bowel bacterial overgrowth [98, 113]. Additionally, conditions resulting in mal-dissolution of LT4 tablets may result in a disturbance of LT4 release from tablet excipients. Decreased HCL secretion, alters gastric pH which may have significant impact on the dissolution of some LT4 products [82]. Specific conditions such as H. Pylori infection and atrophic gastritis both clearly result in suboptimal outcomes in LT4 treated subjects [113, 114]. Several reports document abnormal thyroid function tests with low free T4 (FT4) estimates and reflexively elevated TSH levels when urinary protein losses include the binding proteins responsible for transporting T4 in the blood [115]. Typically this pattern is characterized by elevations of TSH with low T4 and FT4 associated with negative antithyroid antibodies and inadequate response to increasing doses of LT4 [116]. The need for LT4 and abnormalities in thyroid function tests may be expected to be resolved with effective treatment of the nephrosis [115].

Differences in the bioavailability among LT4 preparations have altered outcomes when switches occur from one product to another [117]. As FDA bioequivalence criteria may allow differences in bioavailability of 12.5% or more to be considered therapeutically equivalent [79], this has resulted in at least one pairing of products considered interchangeable with a documented difference of 12.5% [118]. The absolute frequency with which adverse events occur in LT4 treated patients is not clear, but a controlled survey study identified nearly 200 incidents, associated with pharmacy substitution of one product for another [90]. Consistency of refills with the same LT4 source is recommended by the three major endocrine societies based in the US [119].

The impact of intervention once specific causes of suboptimal outcomes are identified, is not always significant. In a study of 17,500 hypothyroid patients followed in an automated registry, TSH values were monitored by an endocrinologist. When elevated TSH values were observed, the reviewer communicated with the primary prescribers of LT4 and made recommendations to pursue further evaluation, treatments and/or dose adjustments of LT4 [120]. At baseline individuals who appeared to be on excessive amounts of LT4 were further characterized. Interventions for further work up or dose adjustments were made and outcomes were reassessed during a follow up analysis. Only 125/17,500 (1%) were taking >225 mcg/day at baseline. Issues identified to explain the apparent “malabsorption” were celiac disease in less than 5 % of cases, compliance issues in more than 15% of cases, medication interference in 21% and gastric parietal cell antibodies were positive in 22% of those taking excess amounts of levothyroxine [120]. More than 35% of cases had no readily identifiable condition associated with their excessive LT4 dose. After the suggested intervention, LT4 dosage was reduced by more than 30% in subjects who had had drug interferences identified. Those with positive PACs were noted to be on 22% less LT4 and patients with celiac disease were on 20% less after appropriate intervention. Those with no known etiology for their excessive LT4 dose also improved demonstrating a 13% dose reduction but those with compliance issues were taking about 16% more LT4 when reevaluated [120]. Clearly compliance is a difficult factor to effectively address.

In summary, evaluation of the patient with less than optimal biochemical outcomes on LT4 therapy is anything but straight forward or limited. Issues to be considered include; compliance, renal thyroxine loss and the potential impact of interfering medications or food intake. Additionally issues related to consistent clinical monitoring and communication [121] as well as the duration of therapy [93] may impact efforts to achieve optimal control. When dosing LT4, it should be remembered that body weight, gender and age [122] may need to be considered when estimating and titrating to biochemical endpoints. Substitution of LT4 preparations with differing bioavailability may result in subtle but measurable differences in clinical and TSH outcomes. The observation that excessive dosing may however be a reflection of malabsorption of LT4 or alternatively be related to diagnostic artifacts due to interfering antibodies producing falsely elevated TSH assay results [123] must be kept in mind. This TSH artifact will be best recognized as FT4 levels rise without the expected suppression of TSH. Examination of the patient’s thyroid function tests on another diagnostic platform or after dilution of samples may provide insight into a potential TSH reliability issue when present [123].

Recommendations for clinical use and ongoing controversies

Although recently published guidelines for thyroid hormone replacement therapy endorse the use of levothyroxine mono therapy [64, 65, 124, 125], others continue to view the use of the thyroid hormone extracts as an alternative to the use of LT4 alone or in combination with triiodothyronine [126–128] for patients with hypothyroidism. A recent prospective, appropriately controlled comparison of thyroid hormone extract versus l-thyroxine mono therapy revealed no objective evidence of differences in neurocognitive function nor symptoms among subjects who were maintained in a euthyroid state with close monitoring [128]. Subjects treated with extract however did report some weight loss and 48.6% preferred the thyroid hormone extract period of treatment. Although TSH levels were similar in the two groups, significantly higher serum T3 levels were observed in those treated with extract especially 3 h after ingesting the medication. Interestingly, those with the highest T3 levels seemed to prefer the LT4 treatment period [128]. As a reminder, combination treatments of LT4 and LT3 were initially introduced to provide a predictable dosage alternative to thyroid extracts [28, 32, 43, 44], more recently the addition of LT3–LT4 was re-introduced as potentially superior to LT4 alone among patients reporting symptoms attributed to hypothyroidism [129, 130]. There have been many experimental protocols developed to test the hypothesis that LT4 alone is inadequate substitution for hypothyroid patients. Most of these studies, using many different ratios of LT4–LT3 have been unsuccessful in demonstrating that this combination therapy provides superior outcomes to LT4 alone [131–134].

National guidelines on the treatment of hypothyroidism

National and international guidelines on the treatment of hypothyroidism in adults, clearly support the use of LT4 monotherapy as the first line approach [64, 65, 124, 125]. Most of these recommendations go on to state that there is insufficient evidence to support the use of thyroid hormone extracts as an alternative to LT4 monotherapy [64, 65, 124, 125]. Where the guidelines differ to some degree is when consideration of LT3 add on therapy would be considered. The European Thyroid Association guideline suggests the experimental use of LT4/LT3 in a 14:1 ratio [124] and the most recent American Thyroid Association (ATA) guideline also considers LT4/LT3 to be an experimental alternative [65].

In special patient populations the guidelines go further. The recent, 2011 Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease during Pregnancy and Postpartum strongly recommends that treatment of maternal hypothyroidism be accomplished with oral levothyroxine mono-therapy. This same guideline strongly discourages use of other thyroid preparations such as LT3 or thyroid extract [108]. Similarly, the American Association of Clinical Endocrinologists and ATA designate L-thyroxine monotherapy as preferred in hypothyroid pregnant women or those with hypothyroidism planning pregnancy [64]. Levothyroxine mono therapy is overwhelmingly recommended as the best proven therapeutic approach to patients with hypothyroidism.