How do you approach the problem of TSH elevation in a patient on high-dose thyroid hormone replacement?


John C. Morris, Division of Endocrinology, Mayo Clinic, Rochester, MN 55905, USA. E-mail:


Persistent elevation of TSH levels in patients under treatment for hypothyroidism is a relatively common clinical problem in endocrinology practice. The most common cause for this phenomenon is poor patient compliance with their thyroid hormone tablets. In the compliant patient, however, multiple aetiologies are possible and a methodological and stepwise approach to the patient's problem will uniformly identify a cause, or at least a resolution.

Thyroid hormone therapy is generally one of the most gratifying and effective hormone replacements within the clinician's armamentarium. Most patients are satisfactorily managed with a single tablet of synthetically derived levothyroxine daily. To make that possible, a wide variety of tablet sizes are available from a number of manufacturers and feature both brand-name products and several generic preparations. Although the average dose for effective and optimal replacement varies somewhat from patient to patient, most hypothyroid patients are managed within a fairly narrow dose window that varies according to body weight, the average being near 1·6–1·8 µg/kg.1 Although total body weight is a convenient measurement upon which to base initial dosing, correct replacement dosing correlates better with lean body mass than total body weight.2 In most patients the circulating concentration of TSH serves as a reflection of thyroid hormone effect upon the pituitary and thereby as an effective marker of the adequacy of the replacement dose. However, one of the common clinical problems that I am asked to review is that of the patient who requires higher doses of levothyroxine for normalization of their TSH or whose TSH level remains persistently elevated despite these high doses.

For example, a recent patient I was referred was a 40-year-old woman from a neighbouring community who was diagnosed with hypothyroidism 1 year prior to her visit. In brief, at that time she expressed symptoms compatible with hypothyroidism, such as severe constipation, fatigue and a modest amount of weight gain, and her thyroid gland was firm, moderately diffusely enlarged but without nodularity, suggestive of autoimmune thyroid disease. Laboratory investigation demonstrated marked TSH elevation (145 mIU/l, normal 0·3–5·0 mIU/l) and low free thyroxine (3·9 pmol/l, normal 10·3–23·2 pmol/l). Her primary care physician started thyroxine replacement at standard doses and adjusted the dosage upwards on three occasions because of persistently elevated TSH levels. After taking 300 µg of levothyroxine for 6 weeks, she was referred for further evaluation when her TSH remained elevated (77 mIU/l) and free thyroxine remained low at 7·7 pmol/l.

My approach to this patient and those with persistently elevated TSH levels despite doses of thyroid hormone that should be adequate is outlined below. I recommend following these steps, although the order of them and the number that may need further pursuit may vary from one patient to the next, based upon the individual patient's circumstances.

  • 1Confirm the diagnosis and laboratory results. Frank primary hypothyroidism by laboratory definition requires low levels of thyroid hormones (total and free T4 and T3) and elevated TSH. The finding of a persistently elevated TSH level is not enough to confirm the diagnosis here, it is also crucial to measure thyroid hormone levels (T4 and T3). Markedly elevated TSH levels without low or at least low-normal thyroid hormones suggests other diagnoses or reasons for the discrepant dose requirements such as heterophilic antibody interference with TSH measurements, TSH secreting pituitary tumours, or thyroid hormone resistance syndromes. If the thyroid hormone levels are not low, more investigation and establishing the correct diagnosis are essential. Elevated thyroid hormone levels that indicate the patient is taking and absorbing the thyroxine tablets appropriately suggest that investigations including more careful quenching of the serum samples for heterophilic antibody interference by the laboratory, serial dilution of the TSH sample, screening of the family members or genetic testing for thyroid hormone resistance, and pituitary imaging for evidence of pituitary tumours may be helpful. Measurement of free T4 and/or free T3 by equilibrium dialysis may at times be helpful as these more direct methods of assay are less susceptible to the effects of thyroxine binding proteins.
  • 2Ask about compliance. The most common reason for unusually high thyroid hormone dose requirements in my practice is poor compliance with the daily dosing of levothyroxine. One day's tablet accounts for 14% of the total weekly dose and because of the long half-life of levothyroxine, missing a day will have an influence on thyroid hormone and TSH levels that is manifest over several days. Thus, having a discussion with the patient about compliance should always be the first step in the process. Many patients will acknowledge forgetting their tablets occasionally if asked in a nonaccusatory, nonjudgemental manner. What is sometimes difficult here is determining how often ‘occasionally’ occurs.
  • 3Check the patient's medication bottles and tablets. At times the patient's reported dose may differ from that prescribed and of course errors by the pharmacist also occur, resulting in tablets inside the bottle that differ from those reported on the label. The colour coding of thyroxine tablets is helpful for determining the dose the patient is taking, but many patients, and occasionally their physicians, have colour vision deficiencies making actual inspection of the tablets an important step in understanding exactly what the patient is currently ingesting. Confirming records of prescriptions and refill records with the pharmacy may also be helpful in documenting compliance, or lack thereof.
  • 4Review the thyroxine ingestion history. The most efficient and reproducible way of taking levothyroxine is to ingest the tablets on an empty stomach and avoid ingesting other medications or food for 30–60 min afterwards. A fairly large and still growing number of medications, supplements and even food items can alter the fraction of an ingested dose that is absorbed.3,4 The ingestion of one or more of these items at or near the time of dosing with thyroxine can substantially change the dose requirement in an individual patient, especially when practised as routine. The most common offenders that I see are calcium and iron supplements. Some multivitamin preparations may also influence thyroxine absorption but the effect does not seem as clear or commonly problematic as that for calcium and iron supplements. A list of medications that may interfere is included in Table 1, but this list is certainly incomplete as new offending medications are reported frequently.
  • 5Investigate for malabsorption. Unfortunately, levothyroxine is not fully absorbed after oral ingestion. On average, only about 70–80% of the available tablet dose is absorbed in euthyroid individuals.5 Interindividual variability in the efficiency of gastrointestinal (GI) absorption is fairly large and this variability accounts for most of the range of requirement seen between compliant patients after adjustment for body size. Malabsorption syndromes increase the requirement for levothyroxine by further reducing the fraction of the ingested dose that is absorbed. Patients with short bowel from prior small bowel bypass or resection commonly require higher than expected T4 doses. If the patient has frequent, voluminous stools, a malabsorption disorder may be evident and measurements of stool fat can confirm this diagnosis. However, thyroxine malabsorption has been reported as the initial finding in patients with otherwise asymptomatic malabsorptive syndromes, especially coeliac disease.6 I screen patients that get this far along the diagnostic pathway with measurement of tissue transglutaminase antibodies and, if positive, send them for GI evaluation, usually including small bowel biopsy and stool fat measurements. Correction of the malabsorption will normalize or at least improve thyroxine absorption in these patients. Helicobacter pylori infection, especially when accompanied by atrophic gastritis and achlorhydria, has been reported to impair thyroxine absorption by up to 37% in patients with multinodular goitre. Antibiotic treatment of the H. pylori infection was also demonstrated to improve absorption and reduce thyroxine requirement in those patients.7
  • 6Consider increased turnover or excretion. A number of drugs or clinical conditions may increase the turnover or excretion of thyroid hormone and thereby increase considerably the requirement in individuals that are thyroid hormone dependent. Some examples are phenytoin, carbamazapine and rifampin. Several of the new kinase inhibitors, such as imatinib and sunitinib, that are entering the clinic for various malignancies appear to influence thyroxine requirements in this manner and this may be a class effect, although reports to date are few.8–10 In addition, patients with nephrotic syndrome who excrete large quantities of albumin may have increased thyroxine requirements due to binding of T4 to the excreted albumin.11,12 Women experience increased thyroxine requirements during normal pregnancy, that may reach as much as 50% at its peak.13
  • 7Perform a thyroxine absorption test. A clinical test to estimate thyroxine absorption has been proposed that may have utility in patients who have unexpectedly high T4 requirements. This test involves administration of a single large dose of levothyroxine, usually in the range of 1000 µg, then monitoring T4 levels in blood over time.4 Although I have occasionally performed this test, I do not find it to be generally helpful unless it demonstrates completely ‘normal’ results, thereby supporting patient noncompliance. Unfortunately, there is no well-established standard to which individual patient results can be compared, especially one done in hypothyroid patients with normal absorption. In the few circumstances when I have asked a patient to perform the test, the results indicate that the patient does absorb thyroxine when exposed to a large dose, but leaves me wondering if it is a normal amount or not, because of the lack of adequate normal standards. Furthermore, severe hypothyroidism itself may impair absorption, presumably due to oedema of the small bowel mucosa and this cannot be quantified by the test. Thus, I have not found that this test helps me very much in the decision-making process in individual patients.
  • 8Treat the patient. If an error by the pharmacy or patient, or a compliance-related cause, is identified, correction of these underlying factors will help to resolve the situation. In some circumstances compliance issues are best treated by increasing the tablet size, assuming that compliance will continue at a similar rate. As a ‘last resort’ in patients with continued poor compliance, evidence supports successful use of thyroxine administered once weekly. This quantity is roughly equivalent to the entire weekly total as calculated above, but given as a single oral dose,14 and administration can be monitored if needed. Removing interfering drugs or changing ingestion patterns when present can also be helpful when possible. In many patients the best course is to increase the patient's T4 dose and titrate upwards until the TSH and T4 levels normalize. Except in patients with surgical small bowel syndromes, I have not yet seen a patient who required parenteral administration of thyroid hormone for maintenance. Even relatively large doses of levothyroxine administered orally are less expensive and better tolerated that intravenous or intramuscular injections.
Table 1.  Drugs that increase thyroxine requirements
  1. ? indicates reports and/or data are suggestive but inconclusive.

Reduced absorption
 Ferrous sulphate
 Calcium carbonate (? other calcium salts)
 Aluminium hydroxide
 Sodium polystyrene sulfonate (Kayexalate)
 Sevelamer HCl
 Chromium picolinate
 Proton pump inhibitors
Increased metabolism
 (? Sunitinib and sorafenib)

The patient described at the beginning of this article was very convincing in her description of faithfully ingesting her thyroxine tablets and had for several months been doing so in the absence of other, potentially interfering medications or supplements after advice from her primary care provider. Although she still described symptoms of hypothyroidism, I could not elicit symptoms suggestive of malabsorption. Her thyroid function tests confirmed inadequately replaced primary hypothyroidism in that her TSH remained elevated at 29·4 mIU/l and her free and total thyroxine levels were at the lower end of the normal range, this after ingesting 300 µg of thyroxine for approximately 8 weeks. The screens for malabsorption and coeliac disease (transglutaminase antibodies and stool fat measurement) were negative and urinalysis was normal. I elected to increase her levothyroxine dose to 400 µg daily and 6 weeks later her TSH was below normal (0·2 mIU/l) and free thyroxine was elevated (25·7 pmol/l). We reduced the thyroxine dose to 350 µg daily and after 8 weeks found that her TSH and free thyroxine levels had returned to normal. Her fatigue and constipation also improved and she stated that she felt much better than before.

I have seen a small number of patients (unreported) that appear to have selective malabsorption of thyroxine. These patients clearly absorb levothyroxine poorly but do not have evidence of generalized malabsorption or of coeliac disease and respond well to increasing the levothyroxine to levels well above those considered usual, in the range of 400–600 µg per day. At present I am not aware of a physiological explanation for these findings but I suspect one will be forthcoming with further investigation into thyroxine transport and absorption. The patient described above may be a representative.