Address correspondence to:
Marco Centanni, M.D.
Dept of Experimental Medicine,
Policlinico Umberto I, Viale Regina Elena 324, 00161 Roma, ITALY
Pho. and FAX +3906 4997-2604 E-mail: marco.centanni@uniroma1.it

Background

Levothyroxine is an effective and reliable drug, widely prescribed for replacement of defective iodothyronines synthesis and secretion (1-3), or to lower TSH in patients with non toxic multinodular goitre (1, 4, 5) or with thyroid cancer (1, 2). Serum TSH represents the best marker for assessing the appropriateness of thyroxine dose (1,2,6-8) both in replacement and in interventional mode. However, due to the lack of sensitive TSH assays, large doses of thyroxine were prescribed for many years (2), which were liable for iatrogenic clinical or subclinical hyperthyroidism. Chronic hyperthyroidism has detrimental effects on some tissues or functional pathways (9,10) such as cardiovascular system and bone remodeling process (1,10-12). Thus, even the appropriateness of thyroxine treatment, when given to lower the TSH below the normal range, has been questioned (1-5). On this ground, it is worth to note that a significant reduction of the described harmful effects has been observed by individually tailoring the dose (13,14). Anyway, these concerns and the availability of more sensitive TSH assays (8), led to a progressive general reduction of thyroxine dose on both interventional and replacement treatment (1). On the other side, growing evidence highlighted that even undertreatreatment and/or mild hypothyroidism have detrimental effects on several tissues and functions (15-20). Indeed, undertreatment during pregnancy has been described as harmful to the fetus as well as to the successful pregnancy (15-17). In adult patients, mild or subclinical hypothyroidism has been implicated in impaired lipid metabolism and atherosclerosis as well as in cardiovascular disease (18-20). Therefore, the need for an individually tailored dose potently emerges from these findings. The continuous search for an optimal daily dose has led to consensus that 1.5-1.6mg/Kg body weight/day is the dose able to restore TSH into the normal range in most hypothyroid patients (1,2) as well as 2.0-2.2mg/Kg body weight/day is the one that lower TSH under the normal range (7,21). Despite these efforts, a number of patients fail to show a clinical and chemical response to the expected dose of T4, leading to uncertainty about the optimal dose in every single patient (1-3). The consequence of that is the need for an increased dose and of care and monitoring as well as repetition of needless and costly diagnostic workup. While the lack of complete clinical effect of replacement T4 treatment deal with the appropriateness of serum TSH as marker of euthyroidism in all tissues (22-24), an impaired absorption of levothyroxine may explain most of the incomplete biochemical effects of T4 in both replacement (1-3) and interventional therapy (25).

The absorption of thyroxine

The daily dose required to obtain the therapeutic effect is not a linear function of the ingested dose of thyroxine which is the main, but not the only decisive event. The amount of absorption of oral thyroxine is, in fact, also key to obtain successful treatment. The absorption of oral thyroxine (70 to 80% of the administered dose) takes place at the intestinal level, is incomplete (26,27), and does not differ in euthyroid and hypothyroid patients (1). Noteworthy, conjugated iodothyronines are secreted, partially deconjugated and reabsorbed in different intestinal sites (26, 28). So far, the direct measurement of thyroxine malabsorption is difficult because several factors must be taken into account (26). The tissue specific metabolism of T4 is a further problem for large investigations in humans (29,30). Hays MT (26), in elegant studies using double isotope equilibrium method, has shown that about 20% of T4 is absorbed at the level of duodenum, about 40% is absorbed in the upper and the remaining fraction in the lower jejunoileum. Studies in euthyroid subjects from Benvenga S et al. (31), showed that peak values of T4 absorption (ΔT4) occurs between 30 and 60 minutes following drug administration and that most of the hormone absorption occurs within the first 90 minutes. In the same study, the interference of food has been proven to delay the absorption of oral thyroxine too (31). In fact, serum TSH was not lowered despite the high doses administered until ingestion of food was postponed from 15 minutes to at least one hour following T4 ingestion (31).

Increased need for thyroxine

The complexity of a direct measurement of thyroxine absorption makes serum TSH levels the best diagnostic tool to evaluate thyroxine treatment effectiveness (1-3,6,7). There are some circumstances in which patients fail to show a chemical response to thyroxine treatment and larger doses of thyroxine are required to attain the desired serum TSH concentrations (1,3,26). Psychological, nutritional, and pharmacological circumstances in which the T4 dosage should be adjusted are summarized in Table I. Interference appear to be mainly related to a) patient characteristics, nutritional habits and compliance with the drug; b) the pharmaceutical characteristics of thyroxine and c) the interference of other drugs. So far, mechanisms other than malabsorption of T4 may be responsible for the increased need for thyroxine in these patients (see 1,3 for review) and some of them deserve mention. The effect of pregnancy and of estrogen treatment is very well known and reviewed elsewhere (1-3). Certain patients, due to poor compliance with the prescribed regimen, do not assume thyroxine regularly, a condition known as pseudomalabsorption (32). Similarly, some patients assume T4 while having breakfast or with a minimal delay (15-20 min). According to the abovementioned studies (31), a lag time of one hour between oral T4 and food ingestion may improve the efficiency of T4 absorption. These observations have been recently confirmed in patients with non-toxic multinodular goitre (25). Using that treatment design, a mild suppression of TSH was obtained at a median dose (1.53mg/Kg body weight/day) which usually normalized serum TSH in most hypothyroid patients (21) and was lower than the one required to suppress thyrotropin (7). This suggests that more efficient thyroxine absorption occurred at the intestinal level in real fasting conditions. So far, delaying breakfast up to one hour following levothyroxine ingestion may reduce the amount of the required T4 dose in most patients (25). Increased need for thyroxine may also ensue from the different T4 bioavailability in areas where a number of generic and brand name thyroxine preparations are available (33). The occurrence of these differences in the bioequivalence of thyroxine preparations is probably infrequent, but switching from one preparation to another should be avoided anyway (33). Some drugs have also been shown to interfere with thyroxine dose (1,3) (Table I). Most of them seem to interfere by sequestrating oral thyroxine into the intestinal lumen but some also interacting with the acid environment. In particular, calcium carbonate, aluminum hydroxide, sucralfate and proton pump inhibitors are all drugs known to interfere with gastric acid pH and/or secretion (25,34-36). In few circumstances the need for thyroxine may be also decreased, namely in elderly patients or when body weight decreases and finally in patients treated with androgens (1,3).

Malabsorption of thyroxine

Increased need for thyroxine not necessarily indicates T4 malabsorption, which also requires the presence of gastrointestinal disorders. Studies dealing with T4 malabsorption were in the form of case reports and mainly focused on intestinal disorders. Erratic but reliable reports indicated that celiac disease (37,38), lactose intolerance (39), short bowel syndrome (3), chronic giardiasis (40) are all associated with an increased need for thyroxine (table II). The recent report of an increased need for thyroxine in patients with impaired gastric acid secretion (25) highlighted also a novel role for the stomach in the subsequent intestinal T4 absorption (Table II). Indeed, the gastric acid producing machinery (i.e. the oxintic glands) is destroyed in patients with atrophic body gastritis (41), is blocked in those treated with proton pump inhibitors (PPI) (42) as well as partially destroyed and counteracted by NH3 production in those with Helicobacter pylori infection (43,44). Daily T4 requirement was found to be almost 1/3 higher in patients with H.pylori-related gastritis, atrophic gastritis and maximal in those with both these conditions, than that observed in control patients (25).  The prospective approach was consistent with these results and showed a reversible TSH increase in thyroxine-treated patients de novo-infected by H.pylori or simultaneously treated with PPI (25). The mechanism by which intestinal absorption of thyroxine may be impaired in hypo-achlorhydric patients remains, however, unclear but hypothesis can be made. The efficiency of intestinal T4 absorption may be altered by the ionization status and conformational characteristics of T4 molecule. The native lipophylic thyroxine enters target cells through both passive diffusion and in a carrier-mediated, inhibitable way (45). Instead pharmaceutical T4 preparation is a hydrophilic sodium salt that may remain undissociated under hypochlorhydric gastric environment and thus less efficiently absorbed through the intestinal lipid bilayer. Whatever the mechanism would be, the meaning of these findings is not restricted to the addition of further causes of increased need for thyroxine (Table II). In fact, while atrophic gastritis is a rare disease (41), H. pylori infection is widespread all over the world (43) and proton pump inhibitors are among the most prescribed drugs in the world (42). Therefore, the estimated optimal daily dose of T4 may be biased in a relevant number of patients taking oral T4, due to the large number of unaware infected or treated patients. Also, that occult variable may perhaps explain the large range of T4 dose reported by several authors (2,6,7). Conversely, malabsorption of T4 may also represent a tool to diagnose occult gastrointestinal diseases. In fact, once established that malabsorption represents a major disrupter of the relationship between the oral T4 dose and the expected TSH, patients treated with thyroxine may be divided in responders and non-responders. While the responders show a positive proof of not having the described gastrointestinal problems, the non-responders represent a group at risk for having these problems. In these latter, careful exclusion of pseudomalabsorption as well as of food and drug interference is needed, but then the presence of gastrointestinal diseases must be investigated. This is particularly true when an iron-deficient anemia, often associated to celiac disease (46), H. Pylori infection (47) and atrophic gastritis without pernicious anemia (48), is concurrently present. Based on these evidence and considerations, a diagnostic flow chart of T4 malabsorption, shown in Fig 1, may be proposed. Active H. pylori infection must be screened as a first line as it is the most frequent (43). The assay of fasting gastrinemia and pepsinogen I should be assayed following a negative response from ¹⁴C-Urea breath test (41,42). Increased gastrinemia and/or reduced pepsinogen I in serum should lead to EGDS with multiple biopsies of gastric body and antrum and duodenal mucosa (48). A third line of investigations may be proposed when previous steps would have been negative and should include lactose breath test to detect lactose intolerance (39) and parasitic search in the stool with special attention to the presence of giardia lamblia (40). A successful diagnostic workup may uncover the vast majority of gastrointestinal cause of poor response to thyroxine treatment, but would also be useful to diagnose and treat subclinical or silent, but not harmless, gastrointestinal disorders. A further issue to be taken into account is the pathogenesis of some of these gastrointestinal diseases: in fact both the atrophic body gastritis (48) and celiac disease (46) are of autoimmune origin and frequently associated to autoimmune thyroiditis, in turn the main cause of hypothyroidism in adult patients (11). Characteristic of these associations is the putative common pathogenic mechanism and the significant risk to have further occult immunoendocrinopaties in the patient and in its relatives (49). The complex of these diseases has been classified as autoimmune polyglandular syndromes, a definition, however, which is insufficient to describe all these associations and still in waiting for updated classification. The disclosure of these often silent diseases by using the screening described in fig. 1, may help to better characterize the immunological status of  patients and of their relatives.

Concluding Remarks

Gastrointestinal malabsorption of oral thyroxine is quite common and may be a major determinant of the T4 daily requirement for both replacement and interventional therapy. Conversely, an unexplained increased need for thyroxine should be trigger for gastrointestinal diseases diagnostic workup.

LEGEND TO FIGURES
Fig.1- Diagnostic flow chart of T4 malabsorption
EMA = serum anti–endomisium antibodies; tTg = serum anti-transglutaminase antibodies;
EGD = esophagogastroduodenoscopy