Ensuring Appropriate Thyroid Hormone Therapy
	The “correct” maintenance dose of thyroxine The development of assays for TSH with a functional sensitivity of 0.1 mU/l or less demonstrated that the dose of thyroxine necessary to restore serum TSH to normal in most patients with primary hypothyroidism was 100–150 µg daily and not the much higher dose of 200 to 400 µg previously employed and recommended in most major textbooks in the 1960s. A significant proportion of patients taking thyroxine have undetectable serum TSH concentrations 1 , but are clinically euthyroid and asymptomatic. This has been considered unimportant as it was felt that the TSH-producing thyrotropes of the anterior pituitary were not only uniquely sensitive to minor changes in thyroid hormone concentration within their respective reference ranges 2 but also, unlike other tissues and organs, derived proportionately more T 3 from local monodeiodination of T 4 , and less from circulating T 3 3 . In other words, a “hyperthyroid” thyrotrope did not reflect the metabolic activities of the rest of the body. This view was challenged, however, by a series of reports, showing that if thyroxine were given in doses sufficient to suppress serum TSH without raising serum T 3 and T 4 into the thyrotoxic range, there were changes in the function of other target organs, such as heart, liver, kidney 4, 5 and bone 6 that were similar to but much less marked than those present in patients with hyperthyroidism. There is some evidence of tissue adaptation to thyroid hormone excess 7 and the changes may to some extent depend upon the cause of the hyperthyroidism 8 . It was mainly the anxiety that a low serum TSH concentration might eventually lead to reduced bone mineral density that caused the American Thyroid Association (ATA) to state that “the goal of therapy is to restore most patients to the euthyroid state and to normalize T 4 and TSH concentrations”9 . This policy has inevitably led the pharmaceutical industry, principally in North America, to produce a bewildering variety of strengths of thyroxine tablets in an attempt to allow the physician to fine-tune the replacement dose. It soon became clear that a significant minority of patients satisfying the criteria of the ATA recommendation did not feel as well as they would wish and only achieved the desired sense of well-being when taking thyroxine in a dose of 50 µg in excess of that necessary to restore serum TSH to normal 10 . Many physicians now take the view that a suppressed serum TSH concentration is acceptable in those taking thyroxine replacement therapy as long as serum T 3 is unequivocally normal, a sentiment supported by a UK consensus statement 11 . Serum TSH concentrations are of no value in assessing the correct dose of thyroxine in patients with central hypothyroidism, and measurements of both serum T 3 and T 4 combined with a biochemical index of thyroid hormone action, such as soluble interleukin 2 receptors, have been suggested 12 . In practice, however, the physician is likely to be influenced more by the well-being of the patient. Is “a little too much” thyroxine dangerous? There is evidence that endogenous subclinical hyperthyroidism due to early Graves’ disease or nodular goitre is associated with reduced bone mineral density,13–16 atrial fibrillation,17 increased left ventricular mass and enhanced systolic but impaired diastolic function.18 Increasingly, the term exogenous subclinical hyperthyroidism is used to describe clinically euthyroid patients taking thyroxine in whom serum TSH is suppressed and there is often an assumption that the exogenous form carries the same risk of morbidity as the endogenous variety. This may be the case in those patients who have developed hypothyroidism following iodine-131 or surgery for Graves’ disease or nodular goitre in whom an autonomous remnant may remain 19 . If a standard dose of thyroxine is prescribed to those patients, serum T3 may lie in the upper part of the reference range as is usually the case in endogenous subclinical hyperthyroidism. However, the majority of patients with a suppressed serum TSH concentration, resulting from thyroxine therapy, have a high normal or elevated serum T 4 and a serum T 3 in the middle of the reference range. Indeed the T 4 :T 3 ratio is greater than would be anticipated in endogenous subclinical hyperthyroidism 20 . Patients with exogenous subclinical hyperthyroidism are, therefore, a heterogeneous group which explains why there are as many publications denying an association with reduced bone mineral density as there are those which claim a link. Even if the generic criticisms of the validity of meta-analysis as a method of combining the results of multiple small studies are disregarded, it is difficult to accept the findings of the most often quoted study as not only did it find increased bone loss in postmenopausal women receiving thyroxine suppressive therapy but also inexplicably in premenopausal women taking replacement therapy in whom serum TSH was normal 21. Various reports have either failed to show an increased rate of fracture in patients with endogenous subclinical hyperthyroidism 1,22 or the risk has disappeared when those with a previous history of hyperthyroidism were excluded 23 . There must be similar reservations about one of the Framingham studies which revealed that a low serum TSH concentration was a risk factor for atrial fibrillation in older persons;24 the number taking thyroxine was small (41) and, as in the bone mineral density studies, inevitably a heterogenous group. It would, therefore, be quite wrong to generalize about the wisdom of avoiding a suppressed serum TSH concentration in those receiving thyroxine replacement. Apragmatic approach should be adopted for each patient, depending upon his or her unique circumstances (Table 1). Suppression of serum TSH concentrations is part of the management of patients with differentiated thyroid carcinoma and may be valuable in achieving maximum goitre shrinkage in patients with Hashimoto’s thyroiditis and significant thyroid enlargement. Although unlikely, any evidence of overtreatment should prompt a reduction in dose of thyroxine of 25–50 µg daily in order to achieve a serum TSH concentration which is low, but detectable, as there is no evidence that this lesser degree of suppression is deleterious in the longterm. Starting dose of thyroxine There is no evidence base for determining how thyroxine therapy should be initiated, but it is customary to prescribe 50 µg daily, increasing to 100 µg daily after 3–4 weeks. Measurement of serum T 4 and TSH at two months after starting will dictate any further adjustment of dosage. In the elderly, and particularly in those with symptomatic ischaemic heart disease, a starting dose of 25 µg daily is advisable with increments of 25 µg every 3–4 weeks, recognising that full replacement dosage may not be achieved without coronary artery bypass grafting or angioplasty. In those in whom hypothyroidism has developed rapidly and has been detected early, such as following total thyroidectomy for thyroid carcinoma, a full replacement dose of 100–150 µg daily may be given from the outset. Variation in dosage of thyroxine Once thyroxine therapy is established it is good practice to review patients annually and measure serum TSH not only to ensure compliance but also to determine whether an adjustment of dose is required. The most common reason for a raised serum TSH in a patient taking 150 µg or more of thyroxine daily is poor compliance; and the apparent paradox of raised serum TSH and raised T 4 is most likely due to overzealous tablet taking for the few days earlier by a patient who had previously been taking thyroxine sporadically. There is, however, an increasing list of circumstances in which the dosage of thyroxine may need to be adjusted, usually upwards (Table 2). The most recent observation is that women taking oestrogen either for contraception or for hormone replacement therapy may require an increase in the dose of thyroxine, although this is not as common an occurrence as during pregnancy 25 . Thyroxine alone or in combination with triiodothyronine? For the first half of the 20 th century the only treatment available for hypothyroidism was animal thyroid extract which, although usually effective, was of variable potency and the rapid serum peak of its triiodothyronine component gave rise to troublesome palpitations in some patients 1 to 2 hours after ingestion. It was largely abandoned in the 1960s when readily standardised synthetic thyroxine became widely available 26 and, as triiodothyronine was derived only from peripheral monodeiodination of the ingested thyroxine, peak concentrations were avoided. Some patients, however, swore by the increased efficacy of thyroid extract and would go to considerable lengths to obtain supplies, but they were considered eccentric as indeed were the practitioners who supplied them. Recently there has been a renewal of interest in treating hypothyroidism with both thyroxine and triiodothyronine based on the results of studies in animals and in patients. In rats which had been thyroidectomized, and therefore made hypothyroid, it was only possible to restore normal tissue levels of T 3 and T 4 and normal serum concentrations of T 3 , T 4 and TSH by administering a combination of both hormones. It required supraphysiological doses of thyroxine given alone to achieve normal tissue concentrations of T 3 27,28 . In patients a combination of thyroxine and triiodothyronine approximating the ratio normally secreted by the thyroid gland resulted in significant improvements in mood and neuropsychological function, as compared with a higher dose of thyroxine alone – and, of most interest, this improvement occurred without suppression of serum TSH 29 . Further evidence that monotherapy with thyroxine may be inadequate derives from a study of weight gain following treatment of hyperthyroidism. Those patients who developed hypothyroidism and were given thyroxine in a dose which restored serum TSH to normal gained significantly more weight than those who remained euthyroid with normal serum TSH as a result of endogenous secretion of T 3 and T 4 . 30 If such findings are replicated it will be important for synthetic combinations of thyroxine and triiodothyronine to be available in a ratio of 10:1, with the latter in slow release form to avoid cardiac adverse effects. The average daily replacement therapy is likely to be thyroxine 100 µg and triiodothyronine 10 µg. Although most patients are satisfactorily controlled with thyroxine alone, an appropriate combined preparation would perhaps improve well-being in those dissatisfied with replacement rather than suppressive doses of monotherapy while at the same time removing the possible risks associated with exogenous subclinical hyperthyroidism.