3.4.9 Treatment of hypothyroidism

Treatment of primary hypothyroidism is usually both gratifying and simple and, in most cases, lifelong. Thyroxine, as l-thyroxine sodium, is the therapy of choice and is available in the UK as tablets of 25, 50, and 100 μg. A greater variety of tablet strength is marketed in other parts of Europe and North America. Thyroxine has a half-life of some 7 days and should be given as a single daily dose which improves compliance. Thyroxine, taken at bedtime, is associated with higher thyroid hormone concentrations and lower thyroid-stimulating hormone (TSH) concentrations compared to the same dose taken in the morning, probably due to greater gastrointestinal uptake of thyroxine during the night (1). Omitting the occasional tablet is of no consequence and those who forget to take their medication, e.g. on vacation, will experience little in the way of symptoms for the first 2 weeks.

Before the availability of sensitive assays for TSH the recommended dose of thyroxine in most major textbooks of medicine was 200–400 μg daily. These doses were associated with high serum thyroxine (T₄) concentrations, e.g. total T₄ 180–200 nmol/l (normal 60–150 nmol/l), thought to be needed before it was recognized that thyroxine was converted to the metabolically active triiodothyronine (T₃) by widespread peripheral monodeiodination, and with serum TSH concentrations that were unresponsive to thyrotropin-releasing hormone. Subsequently, it was shown that doses of thyroxine of as little as 100–150 μg daily were adequate in restoring TSH secretion to normal. The consensus, however, was that the pituitary thyrotrophs were uniquely sensitive to changes in serum thyroid hormone concentrations within their respective reference ranges (2), and that these cells derived proportionately more of their triiodothyronine than other organs from local deiodination of thyroxine (3). Suppression of thyrotropin secretion was not, therefore, necessarily regarded as a sign of overtreatment with thyroxine.

Opinion changed, however, with the advent of thyrotropin assays with a functional limit of detection of 0.1 mU/l or less, which were capable of distinguishing normal from low concentrations. Doses of thyroxine sufficient to suppress thyrotropin secretion without necessarily increasing serum thyroid hormone concentrations into the thyrotoxic range appeared to have more widespread effects. These included changes in nocturnal heart rate, left ventricular wall thickness, systolic time intervals, urinary sodium excretion, liver and muscle enzyme activity, red cell sodium concentrations, and serum lipid concentrations, similar to, but less marked than, those present in overt thyrotoxicosis (4). These changes may not only resolve with prolonged treatment, as there is evidence of tissue adaptation to thyroid hormone excess (5), but also depend upon the cause of the hypothyroidism (6). It was the concern that a low serum TSH concentration might be associated with reduced bone mineral density (7) that prompted the American Thyroid Association to make its landmark statement (8), since reinforced (9), that ‘the goal of therapy [with thyroxine] is to restore patients to the euthyroid state and to normalize serum T₄ and TSH concentrations’. This advice was strengthened by the report that a low serum TSH was a risk factor for the development of atrial fibrillation in older people (10), even although the patients in the study were a heterogeneous group only some of whom were taking thyroxine. The pharmaceutical industry reacted by producing a variety of strengths and colours of thyroxine tablets in an attempt to ensure the recommended biochemical control in patients taking replacement therapy. There is, however, no consensus about what constitutes the most appropriate dose or form of thyroid hormone replacement (11). Most hypothyroid patients treated according to the above guidelines have no complaints and feel returned to normal health. However, a substantial minority claim only to achieve a sense of wellbeing if thyroxine is given in a dose of 50 μg greater than that needed to restore normal TSH secretion (12). There is no convincing evidence that this degree of ‘overtreatment’ is a risk factor for osteoporosis (13), or is associated with increased morbidity or mortality (14). Furthermore, studies of weight gain following destructive therapy for Graves’ disease suggest that restoration of serum TSH to normal by thyroxine alone may not constitute adequate hormone replacement (15). It makes sense, therefore, to allow hypothyroid patients who are dissatisfied with the outcome of restoring serum T₄ and TSH concentrations to normal to increase the dose of thyroxine such that serum TSH is suppressed, in which case serum free T₄ is likely to be between 20 and 25 pmol/l. In this circumstance it is essential, however, that the serum T₃ concentration is unequivocally normal.

Practice varies slightly from centre to centre and between countries, but a reasonable starting daily dose of thyroxine in a middle-aged patient with no history of cardiac disease is 50 μg, increasing to 100 μg, and then to 125–150 μg at intervals of 2–3 weeks. After 3 months or so of therapy any minor adjustment to the dose can be made such that the serum concentrations of T₄ (free or total) and TSH are at the upper and lower parts of their respective reference ranges. The reason for the stepwise increment in the dose of thyroxine is the fear that a sudden increase in metabolic rate in a patient with long-standing severe hypothyroidism may unmask previously unrecognized ischaemic heart disease and precipitate angina, myocardial infarction, dysrhythmia, or even sudden death, although the evidence for such a cautious approach is anecdotal. On the other hand, it is quite appropriate to prescribe what is thought to be a full replacement dose with immediate effect in a young patient in whom the thyroid failure is known to have been of short duration, such as following total thyroidectomy for differentiated thyroid carcinoma. In contrast, in older patients in whom thyroxine requirements are reduced, and in those with concurrent symptomatic ischaemic heart disease, it is customary to begin with a dose of thyroxine of 25 μg daily with increments of 25 μg daily every 3–4 weeks. Worsening angina is no longer a reason for suboptimal replacement therapy, as coronary artery bypass surgery or angioplasty is safe and effective before clinical and biochemical euthyroidism has been established.

Patients begin to feel better within 10–14 days of starting thyroxine, even in doses as little as 25 μg daily. Reduction in body weight, which is rarely more than 10% and largely due to fluid loss, and improvements in periorbital puffiness are among the early responses, whereas maximum improvement in hair and skin texture may take up to 3 months, and reversal of the rare feature of cerebellar ataxia, considerably longer.

Once the correct dose of thyroxine has been established it is good practice to evaluate the patient and measure serum T₄ and TSH concentrations annually to improve compliance, as long-term medication is often not taken regularly or in the recommended dose, and thyroxine is no exception. Weekly administration of 7 times the daily dose of thyroxine may be of benefit in poorly compliant patients (16), but there is little or no experience of its efficacy and safety. The most common reason for a raised serum TSH concentration in a patient taking 150 μg or more of thyroxine daily is poor compliance. The seemingly anomalous combination of raised serum T₄ and TSH concentrations is most likely due to overzealous tablet-taking for a few days before a clinic visit by a patient who was previously taking thyroxine sporadically. Whereas computerized follow-up schemes are the most effective method of ensuring that thyroid function tests are performed regularly, there is an unfortunate tendency for advice about changing dosage of thyroxine to be based solely on biochemical results without considering the clinical status of the patient.

Even in conscientious patients, regular review of dosage is advisable as requirements may change for a variety of reasons (Box 3.4.9.1). The concurrent administration of any of several drugs may necessitate an increase in thyroxine dosage to maintain a normal serum TSH concentration, the most recently recognized being omeprazole (17) and the antidepressant sertraline (18). Ingestion of dietary fibre supplements may reduce bioavailability of thyroxine by its adsorption on to wheat bran (19).

The mean dose of thyroxine required by patients developing hypothyroidism during the first 1–2 years after surgery or ¹³¹I therapy for hyperthyroidism due to Graves’ disease is lower than in those with spontaneous primary hypothyroidism, but a higher dose may be required in later years. The explanation is the continued presence of stimulating TSH-receptor antibodies in the early stages after ablative therapy for Graves’ disease, resulting in nonsuppressible secretion of thyroid hormones by the thyroid remnant. As the production of the antibodies declines, this autonomous secretion declines as well (Fig. 3.4.9.1). Rarely, patients with long-standing primary hypothyroidism develop Graves’ disease due to a switch in production from blocking to stimulating TSH-receptor antibodies.

Most patients require an increase in thyroxine dosage during pregnancy, as serum TSH concentrations rise into the hypothyroid range if the prepregnancy dose of thyroxine is maintained. The average increase in thyroxine dosage required is 50 μg daily, and this may be evident within 6 weeks of conception. The principal reason for this change in thyroxine requirement is the increase in the serum concentration of thyroxine-binding globulin in pregnancy, which results in decreased serum concentrations of free T₃ and T₄. These decreases cannot be compensated for by increased thyroidal secretion because of lack of functioning thyroid tissue (20).

Finally, there are numerous manufacturers of thyroxine preparations and, from time to time, there is divergence in bioequivalence. This possibility should be entertained in any patient in whom serum TSH concentrations rise when no new medication has been prescribed and the dose of thyroxine has been stable for many years. It is important that the same preparation of thyroxine is dispensed at each prescription refill (11).

Most patients with primary hypothyroidism require lifelong thyroxine therapy. However, hypothyroidism may be transient and even short-term treatment with thyroxine may be unnecessary. This is the case for the thyroid failure developing within the first 6 months after surgery for Graves’ disease and failure to appreciate such a phenomenon has led to spuriously high estimates of postoperative hypothyroidism. Raised serum TSH concentrations, as high as 200 mU/l, and low or undetectable serum T₄ concentrations may be recorded in 30% of patients at 3 months after surgery, with or without symptoms of mild hypothyroidism. By the sixth month, however, without thyroxine substitution, serum T₄ and usually TSH concentrations have returned to normal (Table 3.4.9.1)(21). A similar pattern of thyroid function occurs after ¹³¹I therapy for Graves’ disease, and a rise in the concentration of blocking TSH-receptor antibodies has been implicated in the thyroid failure (22). It follows that permanent hypothyroidism should not be diagnosed before 6 months have elapsed following treatment of Graves’ disease by surgery or ¹³¹I. If, because of symptoms at 2–4 months, thyroxine treatment is deemed necessary, a suboptimal dose of 50–75 μg daily should be prescribed, which allows meaningful assessment of thyroid function at 6 months. If at that stage the serum TSH concentration is elevated, the thyroid failure should be considered permanent and the dose of thyroxine increased appropriately, but if serum TSH is normal or low, the thyroxine should be stopped and the thyroid function reassessed 4–6 weeks later. An exception to this policy should be made in patients with significant orbitopathy, as raised serum TSH concentrations are a risk factor for worsening of thyroid eye disease. In such patients it is important to avoid any degree of thyroid failure following ¹³¹I or surgical treatment of Graves’ disease by early treatment with adequate doses of thyroxine, accepting that the question of permanent or temporary hypothyroidism can be resolved at some stage in the future.

Other examples of transient hypothyroidism include the recovery phase of subacute (De Quervain’s), painless, and postpartum thyroiditis; Hashimoto’s thyroiditis, particularly if excess iodine or iodine-containing drugs, such as amiodarone, have been implicated in the development of the thyroid failure; in the neonatal period in children born to a minority of mothers with autoimmune hypothyroidism due to the transplacental passage of TSH-receptor blocking antibodies; and in 5% of patients with chronic autoimmune thyroiditis as a result of the disappearance of these same antibodies from the serum. In addition, the use of iodine-containing antiseptics applied vaginally during labour and topically to the skin of the newborn infant may also result in transient hypothyroidism. Temporary thyroid failure may occur 2 months to 2 years after starting treatment with interferon-α for hepatitis C. Raised serum TSH concentrations of greater than 15 U/l are often recorded in patients with untreated or inadequately treated Addison’s disease, but usually fall to normal with glucocorticoid replacement. Similarly, raised serum TSH concentrations may be found during the recovery phase of nonthyroidal illness.

This was first used successfully by mouth in the treatment of hypothyroidism in 1892 and for the next 50 years or so was the only therapy. These extracts from oxen, sheep, and pigs contain, among other iodinated amino acids and proteins, thyroxine and triiodothyronine in variable amounts. They lost favour among endocrinologists in the 1960s due to problems with standardization, and because synthetic l-thyroxine became readily available. From time to time thyroid extract has enjoyed a renaissance, usually among practitioners on the fringes of medicine, because of its effectiveness in weight reduction, but only as a consequence of inducing mild hyperthyroidism which is not in the best interests of the patient.

These combinations, e.g. Liotrix and Novothyral, are not widely used as most, if not all, contain molar ratios of T₄ to T₃ significantly less than the molar ratio for secretion by the human thyroid gland of 14:1, thereby providing an excess of triiodothyronine. Administration of triiodothyronine as a bolus, alone or in combination with thyroxine, coupled with its rapid intestinal absorption, contributes to the appearance of peaks of elevated serum T₃. These are often associated with undesirable cardiac effects, such as palpitations, as the heart derives most of its triiodothyronine from plasma. It has been assumed that all the necessary triiodothyronine is derived from peripheral monodeiodination of orally administered thyroxine, a view for which there is recent support in respect of concentrations of T₃ in the serum (23). However, there is evidence from the thyroidectomized, and therefore hypothyroid, rat that it is only possible to restore normal concentrations of T₃ in all tissues, while maintaining normal serum concentrations of T₃, T₄, and TSH, by giving a combination of thyroxine and triiodothyronine, and not thyroxine alone, unless in supraphysiological doses (24). That a similar situation might exist in humans was suggested when a combination of thyroxine and triiodothyronine, approximating the ratio normally secreted by the thyroid gland, resulted in significant improvements in mood and neuropsychological function, when compared to a higher dose of thyroxine alone, and without suppressing serum TSH (25). This claim has not been substantiated (26) but some patients prefer the combination therapy despite the absence of any objective benefit. There is also emerging evidence that patients who do respond to the addition of triiodothyronine to thyroxine therapy possess an inherited variant of the type II deiodinase gene, present in 16% of the population (27). The pragmatic approach in patients who fail to achieve the desired sense of wellbeing, despite a dose of thyroxine which suppresses serum TSH to less than 0.05 mU/l, albeit with an unequivocally normal serum T₃ concentration, is to reduce the dose of thyroxine by 50 μg daily and add triiodothyronine in a dose of 10 μg daily. If the patient remains symptomatic it is clear that thyroid dysfunction is not responsible. This most commonly occurs in female patients and they should be encouraged to address outstanding issues at home and in the workplace which may be the cause of the nonspecific symptoms of weight gain, tiredness, and low mood. Many of these patients are menopausal and oestrogen replacement is likely to be more effective than tinkering with the dose or form of thyroid hormone replacement.

Occasionally, patients may have been started on treatment with thyroxine for nonspecific symptoms such as tiredness and weight gain, without confirmatory tests of thyroid function, on the basis of equivocal results, or in a situation when the thyroid failure may have been temporary, such as in the first year postpartum. In order to assess the continued need for thyroxine, treatment should be stopped for 4 weeks if the serum TSH concentration is normal, or for 6 weeks if it is undetectable, in order to allow recovery of the suppressed pituitary thyrotrophs. Measurement of serum T₄ and TSH concentrations at this stage will determine whether the patient is truly hypothyroid.