Hyperthyroidism in the elderly is a serious clinical problem both for the individual and for health services which must fund associated costs. Hyperthyroidism is known to be a common disorder, a population-based survey reported by Tunbridge et al (1)conducted 30 years ago revealed a prevalence in the general population in the UK of around 2.7% in females (10-fold less in males) and of undiagnosed disease in around 0.5% of women. There are few data on prevalence specifically in elderly populations. In one UK community survey by Parle et al (2) of 1210 subjects aged over 60 years only one person was found to have undiagnosed overt hyperthyroidism. Another larger and more recent prevalence study from the same group in the UK has indicated a prevalence of undiagnosed hyperthyroidism of 0.3% in a large group aged over 65 years (3) with an additional prevalence of previously diagnosed overt hyperthyroidism of 0.3%. Another survey of 1442 subjects in Sweden indicated that 2% of over 60's had thyrotoxicosis on screening (4).
The prevalence of "subclinical" hyperthyroidism (a biochemical diagnosis characterized by a low serum TSH with normal serum thyroid hormone concentrations) (see below) is much higher than that of overt disease. The advent of sensitive assays for TSH that distinguish low from normal values led to the realization that low serum TSH was present in about 5% of elderly subjects (2), a value significantly higher than in younger age groups. There is significant variation in the reported prevalence of subclinical hyperthyroidism in the elderly, estimates in elderly populations ranging from 0.8% reported by Pirich et al (5) to 5.8% reported by Parle et al (2) although typically quoted prevalences are 1.5% in women and 1% in men over the age of 60 years (e.g. Helfand et al (6)). More recently we have reported a prevalence of subclinical hyperthyroidism of 2.1% in a large group of community dwelling subjects aged over 65 years (3).
Classical symptoms and signs of thyrotoxicosis are shown in Table 1. While some or all of these may be present in elderly subjects with thyrotoxicosis, the clinical picture is often significantly different in this age group. Problems such as weight loss and depression or agitation may predominate - so-called "apathetic" thyrotoxicosis, a condition in which more typical symptoms and signs reflecting sympathetic activation such as tremor and hyperactivity are absent as described by Nordyke et al (7). However, cardiovascular symptoms and signs often predominate. Particularly in this age group, the diagnosis of thyrotoxicosis should also be considered in the presence of other symptoms and signs considered "non-specific" in nature, such as muscle weakness, persistent vomiting, hypercalcemia, and worsening osteoporosis.
Table 1. SYMPTOMS AND SIGNS IN HYPERTHYROIDISM
|
*specific to Graves’ disease |
|
|---|---|
|
|
Cardiovascular complications of thyrotoxicosis are especially common in the elderly and may be a cause of significant morbidity and mortality. A study by Franklyn et al (8) reporting increased vascular mortality in those with a past history of thyrotoxicosis treated with radioiodine revealed a striking excess mortality from both cardiovascular and cerebrovascular causes in those aged over 60 years at the time of treatment. Much of this excess mortality was observed early after therapy. Significant increases in deaths from ischemic heart disease, hypertensive heart disease, rheumatic heart disease and deaths ascribed to dysrhythmias were found in this age group. A more contemporary study from the same group investigating those treated for hyperthyroidism in the last 20-25 years has indicated similar findings in terms of risk of vascular mortality (9).
In addition to classical findings of sinus tachycardia and systolic hypertension, it is well recognised that atrial fibrillation complicates thyrotoxicosis in about 15% of cases (10); however, the incidence of this complication rises with age so it is observed more frequently in the elderly, as Osman et al have reported recently (11). It has been estimated that atrial fibrillation occurs at least three times more commonly in those with thyrotoxicosis than those without. Development of atrial fibrillation may itself lead to deteriorating cardiac status, especially in the presence of pre-existing heart disease, and it may also be associated with embolic complications, especially cerebral embolism (12). These influences of atrial fibrillation probably contribute significantly to the increased cardiovascular and cerebrovascular mortality described above. Furthermore, the likelihood of spontaneous restoration of sinus rhythm in those with atrial fibrillation complicating thyrotoxicosis lessens with age, probably reflecting the presence of underlying ischemic, hypertensive or valvular heart disease (13). We recently confirmed an adverse influence of increasing age, as well as underlying heart disease, on the likelihood of restoration of sinus rhythm in those with AF complicating overt hyperthyroidism, while interestingly the likelihood of restoration of normal cardiac rhythm was increased in those initially hypothyroid following treatment of hyperthyroidism (11).
In view of these cardiovascular manifestations/complications, the diagnosis of thyrotoxicosis should be suspected in all subjects presenting with atrial fibrillation, as well as those with worsening heart failure, systolic hypertension and deteriorating ischemic heart disease. Nonetheless case-finding studies have shown that thyrotoxicosis accounts for less than 5% of newly diagnosed cases of atrial fibrillation (10).
The other particularly significant consequence of thyrotoxicosis in terms of morbidity and mortality is its effect on bone metabolism. It is well recognized that overt hyperthyroidism is associated with increased bone turnover and reduction in bone mineral density as reviewed by Mosekilde et al (14). Meta-analysis of available data (15) has shown that this influence is especially marked in estrogen deficient postmenopausal women. While effective antithyroid treatment results in an improvement in bone mineral density, recovery is incomplete so risks of osteoporosis associated with ageing, especially in women, are exacerbated (16). Indeed, our study of mortality in those treated with radioiodine for thyrotoxicosis revealed doubling of risk of death from fractured femur (8). Several other large-scale epidemiological studies such as that reported by Cummings et al (17) have also revealed independent associations between a history of thyrotoxicosis and risk of fracture of the femur.
There is no specific effect of ageing upon standard tests of thyroid function. Our own studies of healthy elderly subjects have shown that serum concentrations of thyroxine (T4) and tri-iodothyronine (T3) are unchanged compared with younger age groups (18). There may be a slight reduction in thyrotropin (TSH) secretion in older persons but serum TSH measurements are generally within the normal range. Nonetheless, "non-thyroidal" illnesses and drug therapies that alter tests of thyroid function (see below) are more common with increasing age and typically lead to reduced peripheral conversion of T4 to T3 and therefore reduction in serum T3 concentrations. Serum TSH may be unaffected by illness, although reduction in TSH is commonly seen, as is modest elevation in TSH particularly during the recovery phase of illness (19).
It is essential that a clinical suspicion of thyrotoxicosis is confirmed or refuted by biochemical testing before further investigation or treatment is contemplated. The single most important biochemical test is measurement of serum TSH. If the serum TSH concentration is within the normal range, then a diagnosis of thyrotoxicosis is effectively ruled out. Exceptions to this rule are rare TSH-dependent causes of hyperthyroidism, such as TSH-secreting tumors of the pituitary and syndromes of thyroid hormone resistance, although these diagnoses are more typically associated with a modest rise in TSH (with raised serum thyroid hormones, as opposed to the usual pattern of raised TSH in conjunction with low thyroid hormone levels).
The finding of a low serum TSH (through use of a sensitive TSH assay) is not, however, specific for a diagnosis of thyrotoxicosis. Low serum TSH, especially if below the normal range but nonetheless detectable, often reflects a "non-thyroidal" illness or therapy with a wide variety of drugs (Table 2). A diagnosis of thyrotoxicosis should therefore be confirmed biochemically by measurement of serum free thyroxine (T4) (and in some cases T3 if free T4 is in the high/normal range and T3-toxicosis is therefore suspected.
Table 2. EFFECT OF DRUGS ON TESTS OF THYROID FUNCTION
|
Drug |
Serum T4 |
Serum T3 |
Serum TSH |
|---|---|---|---|
|
Dopamine |
↓, → |
↓, → |
↓ |
|
Glucocorticoids |
↓, → |
↓, → |
↓ |
|
Estrogens |
↑ total T4 |
↑ total T3 |
→ |
|
Anticonvulsants |
↓, → |
↓, → |
→ |
|
Acetylsalicylic acid |
↑, → |
↑, → |
↓ → |
|
Amiodarone |
↑ |
↓ |
variable |
|
Heparin |
↑, → |
↑, → |
↓, → |
|
Fenclofenac |
↓, → |
↓, → |
→ |
|
Anabolic steroids |
↓total T4 |
↓total T3 |
→ |
In the majority of cases of thyrotoxicosis, a typical biochemical picture of elevated free T4 and T3 with associated undetectable TSH will be observed. In some cases, a biochemical diagnosis of "T3-toxicosis" is evident; this is characterized by elevation of serum T3 in the absence of a rise in T4. This biochemistry is typically observed in mild cases of toxic nodular hyperthyroidism and early in the course of relapse of Graves' hyperthyroidism. In some instances, the converse is true in that a rise in T3 is absent despite elevation in free T4 and suppression of TSH in a patient thought clinically to have thyrotoxicosis. This lack of rise in T3 may reflect the presence of another "non-thyroidal" illness, the explanation becoming evident upon re-testing once the other morbidity is eliminated.
In iodine replete parts of the world, Graves' disease is the commonest cause of hyperthyroidism. In the elderly, however, toxic nodular hyperthyroidism becomes an important cause, being responsible for the majority of cases of thyrotoxicosis. In all age groups, toxic nodular hyperthyroidism is more common in those areas of the world that are relatively iodine deficient (20). The natural history of goiter is of progression from the presence of diffuse thyroid enlargement to development of one or more nodules and eventual autonomous function of one or more of these nodules resulting in thyrotoxicosis. This natural history is typically long so the elderly patient presenting with thyrotoxicosis often describes the presence of a goiter for many years. A relatively rare cause is the presence of a single toxic adenoma - a benign tumor exhibiting autonomous secretion of thyroid hormones. This diagnosis is said by Hamburger (20) to account for less than 2% of cases of thyrotoxicosis occurring in the US. Biochemically, the development of autonomous function in a nodular goiter is first evidenced by suppression of serum TSH with normal serum concentrations of thyroid hormones ("subclinical" hyperthyroidism - see below), later followed by elevation of serum T3 and free T4.
In many cases, the cause of thyrotoxicosis is obvious from the clinical picture. The diagnosis of Graves' disease may be evident because of the presence of diffuse goiter and ophthalmopathy, whereas toxic nodular hyperthyroidism is characterized by the presence of a nodular goiter on examination of the neck. It should be noted, however, that the thyroid might be impalpable in about 30% of cases of both Graves' disease and toxic nodular hyperthyroidism. If the cause of thyrotoxicosis is not obvious, further investigation may be warranted. The presence of thyroid autoantibodies (to thyroid peroxidase - TPO and/or thyroglobulin) is suggestive (but not diagnostic of) Graves' disease; TSH receptor antibodies, although not measured routinely, are more specific for the diagnosis. Such antibodies are usually negative in cases of toxic nodular hyperthyroidism. If antibodies are positive, in the presence of a nodular goiter, both conditions may co-exist. Radioisotope scanning, using technetium-99m or iodine-125, typically shows a diffuse pattern of uptake in Graves' disease, in contrast to the presence of multiple "hot" nodules with surrounding thyroid tissue not demonstrating any uptake in cases of toxic nodular hyperthyroidism. Occasionally, a single "hot" nodule, with absent uptake elsewhere in the thyroid is observed. This finding suggests the presence of a toxic nodular adenoma.
It is not always essential to make a distinction between thyrotoxicosis due to Graves' disease and that due to toxic nodular goiter, since treatment is usually the same. It is important, however, to consider other diagnoses. As in other age groups, the elderly patient may develop transient thyroid hormone excess secondary to thyroiditis, i.e. destruction of the thyroid with release of pre-formed thyroid hormones (21). Sub-acute thyroiditis should be suspected if the patient complains of sore throat or neck tenderness, typically associated with symptoms of a viral illness or an upper respiratory tract infection. The diagnosis is confirmed by the finding of a raised erythrocyte sedimentation rate (ESR) and absent or very low uptake of iodine-123 or iodine-131. This is an important diagnosis to make since antithyroid treatment with antithyroid drugs or radioiodine is inappropriate, both because it is ineffective and because the condition settles spontaneously (usually after a self-limiting period of hypothyroidism).
An iodine-induced thyroiditis should be considered in those who give a history of iodine ingestion (e.g. in the form of sea weed preparations, other over the counter iodine containing compounds, such as expectorants) or after administration of iodine containing radiographic contrast agents. The diagnosis can again be confirmed by the finding of low iodine uptake. This condition also remits spontaneously and radioiodine therapy is contraindicated.
The diagnosis of thyroid dysfunction should be considered in an elderly patient prescribed the antiarrhythmic agent amiodarone. This drug is now widely used in the older age group for control of dysrhythmias, particularly those associated with poor left ventricular function. The managing physician should be aware that amiodarone is an iodine-containing compound that affects the results of tests of thyroid function, even in those who are euthyroid, as reviewed by Bartalena et al (22). Typically, amiodarone, through its effect on the 5' deiodinase enzyme and hence upon the peripheral conversion of T4 to T3, results in modest reduction in serum concentrations of T3 (often to below the normal range) and modest elevation in serum T4 (often to above the normal range). Measurements of serum TSH may be unaffected; typically TSH is slightly elevated early after commencement of treatment and becomes low later.
Although amiodarone results in overt thyroid dysfunction in 5-10% of cases, it is important not to over-interpret mildly abnormal results of tests of thyroid function. Thyrotoxicosis should only be diagnosed in the presence of significant elevation of free T4, importantly together with elevation in serum T3 and suppression of TSH; sometimes serum T3 is at the upper range of normal rather than elevated, probably because of associated "non-thyroidal" illness in this age group (23).
The thionamides - carbimazole (or its active metabolite methimazole) and propylthiouracil - represent the mainstay of drug treatment of thyrotoxicosis. These drugs inhibit the oxidation and organification of iodide and hence block the synthesis of T4 and T3 early in their biosynthetic pathway. They represent the most effective and rapid means of reducing circulating thyroid hormone concentrations. They can be used in several ways: short-term in preparation of the patient for definitive treatment with radioiodine or surgery, medium term in the hope of inducing remission in cases of thyrotoxicosis due to Graves' disease and finally long-term for control of clinical and biochemical thyroid hormone excess (24).
In most elderly patients, thionamides are used short-term in the preparation for curative treatment. A typical starting dose of methimazole is 20-30 mg per day as a single daily dose. In contrast, propylthiouracil is typically given in divided doses, the equivalent to methimazole 20 mg being 200mg. Doses higher than this are required only rarely and in clinical trials (of subjects with Graves' disease) high doses have not been shown to be more effective in terms of restoration of euthyroidism in important prospective studies by Reinwein et al (25) and Allanic et al (26). Since compliance is better with once daily dosing, methimazole or carbimazole are usually considered the drugs of choice, in preference to propylthiouracil. Serum free T4 should be checked 4-6 weeks after beginning therapy and the thionamide dose adjusted accordingly. It is usually possible to render the patient euthyroid (or near euthyroid) after approximately 2-3 months, so that they can proceed to curative therapy.
Drug side effects are relatively uncommon, but it is essential that all subjects (in whichever age group) be warned (preferably in writing) of the potential risk of agranulocytosis so that they present urgently for a full blood count if they develop a fever or sore throat. Agranulocytosis often, but not always, occurs in the first few weeks after beginning thionamide therapy and is probably more common in those taking higher doses (25). The latter observation represents a relative contraindication to doses of methimazole/carbimazole of greater than 20-30 mg per day; doses higher than this are rarely necessary in the elderly since toxic nodular hyperthyroidism is usually mild and therefore easy to control.
Other serious side effects can occur, notably antineutrophil cytoplasmic antibody-associated-vasculitis (typically associated with prescription of propylthiouracil) and hepatitis, although these are rare. These serious complications, together with agranulocytosis, represent absolute contraindications to further use of thionamides. Less serious side effects such as pruritic rash are more common and can usually be managed conservatively, although sometimes a change in drug therapy from one thionamide to another is required as reviewed by Cooper (27).
If the patient has an established diagnosis of Graves' thyrotoxicosis, then it may be appropriate to offer a full course of thionamide therapy in the hope of inducing long-term remission. In view of the potentially serious consequences of thyrotoxicosis (notably vascular and bone) in this age group, it is generally more appropriate to advise the patient to have definitive treatment early in the course of their disease. In general, remission rates in Graves' hyperthyroidism are less than 50%, nonetheless, there is some evidence that the remission rate in Graves’ may be higher in the elderly age group, probably reflecting the presence of milder disease. If the objective is to achieve remission or "cure" of thyrotoxicosis secondary to Graves' disease, then thionamide treatment should be prescribed for a course of probably not less than 12 or 18 months, since shorter courses are associated with a lower rate of remission21. Drug doses should be titrated according to serum concentrations of free T4 (serum TSH may remain suppressed in the medium to long-term in those with Graves' disease); for most of an 18-month course the majority of subjects will require a methimazole maintenance dose of 5-10 mg daily (propylthiouracil 50-100mg daily in divided doses). Larger dose requirements are suggestive of poor compliance. Poor prognostic features for achieving long-term remission reported by Allahabadia et al (28) (established in younger age groups) include male sex, the presence of a large goiter and biochemically severe disease at diagnosis. Most relapses of Graves' thyrotoxicosis occur 3-6 months after thionamide withdrawal. If relapse does occur then the patient should be advised to proceed to definitive treatment.
It should be noted that thionamides virtually never result in remission or cure of thyrotoxicosis secondary to toxic nodular goiter, although some spontaneous fluctuation in the severity of the disease is seen. Thionamides may thus be used short-term (as above) to induce euthyroidism prior to definitive treatment but a "course" should not be prescribed in the hope of inducing cure. In subjects who decline definitive therapy or whose life expectancy is short (because of co-morbidity) it is appropriate to prescribe thionamides life-long to control clinical and biochemical disease. Typically this scenario pertains in frail, elderly subjects. Once biochemical control has been achieved, biochemical monitoring every 3-6 months, to demonstrate euthyroidism and the absence of iatrogenic hypothyroidism, is desirable.
Beta adrenergic blockers often represent useful adjuncts to thionamides in the management of thyrotoxicosis. In cases of thyroiditis or mild cases of toxic nodular hyperthyroidism proceeding to radioiodine, they may be the only additional treatment required. Beta adrenergic blockers act promptly to reduce symptoms and signs of tremor and to improve tachycardia and associated palpitation (10; 24). Such agents should be used cautiously in elderly subjects with heart failure (although if tolerated a beneficial effect often results because of amelioration of some of the cardiovascular effects of thyroid hormone excess) and in those with asthma or chronic obstructive pulmonary disease. Propranolol has been widely used in thyrotoxic subjects but it requires multiple daily dosing; longer acting beta adrenergic blockers such as nadolol (40-80 mg daily) or atenolol (50-100mg daily) may therefore be preferred.
Other adjunctive therapies include salicylates for relief of local pain and tenderness in cases of subacute thyroiditis; occasionally glucocortocoids such as prednisolone are required short-term.
Anticoagulation with coumarin derivatives such as warfarin should be considered in elderly subjects with thyrotoxicosis complicated by atrial fibrillation. This is driven by evidence for embolic complications. There have been no controlled trials of the use of anticoagulants in thyrotoxic atrial fibrillation but overwhelming evidence of their efficacy in other settings argues in favor of their use in this situation (29), unless clear contraindications exist. Specific therapy to restore sinus rhythm should be considered but not until the patient has been rendered permanently euthyroid. This therapy may comprise pharmacological cardioversion (with agents such as sotalol) or electrical cardioversion. Restoration of sinus rhythm is more likely in those whose atrial fibrillation is of short duration and in those without underlying heart disease (10), although rates of restoration of sinus rhythm may be relatively low, even with cardiological intervention, as Osman et al have recently described (11).
This condition is widely recognized as being difficult to treat and a cause of significant morbidity/mortality in those with underlying cardiac disease (22;23).
It is reported that AIT results either from induction of Graves' hyperthyroidism (secondary to ingestion of a large iodine load since amiodarone is 37% iodine by weight) or from an amiodarone-induced destructive thyroiditis. Some experts, SUCH AS Bartelena et al (30) report that these two types can be distinguished by measurement of serum interleukin-6 (raised in destructive thyroiditis) (30) and by ultrasonographic definition of thyroid vascularity. These tests are not, however, routinely available, although it is increasingly recognized that these varieties may co-exist.
In general, thionamide therapy should be considered first line treatment of AIT. We have recently reported standard doses of methimazole/carbimazole or propylthiouracil to be effective in restoration of euthyroidism in all but one of a consecutive series of 28 subjects in the UK with AIT (31). We did not find a different clinical course in those thought at presentation to have Graves' thyrotoxicosis compared with those thought to have destructive thyroiditis. A proportion of subjects with AIT developed hypothyroidism spontaneously (presumably those with destructive thyroiditis).
Withdrawal of amiodarone is often not possible because of the serious nature of underlying dysrhythmias leading to amiodarone treatment, although it should be carefully considered. In any case, the long half-life of the drug (around 50 days) determines that any effect of amiodarone withdrawal is slow. Because of the iodine content of the drug, radioiodine therapy is ineffective and indeed contraindicated because the radioisotope is not taken up into the thyroid. Radioiodine treatment is typically not feasible until at least 6 months after amiodarone withdrawal. Several groups have described surgical treatment of AIT. Restoration of euthyroidism with thionamides is preferable pre-operatively (see below). Adjunctive agents such as perchlorate and glucocorticoids have been reported to be helpful by some authors (32); such agents themselves have significant side effects in elderly subjects with underlying heart disease.
In most cases of thyrotoxicosis occurring in elderly subjects, radioiodine therapy represents the treatment of choice since thionamide treatment alone is unlikely to be curative. Iodine-131 may be administered by mouth in the outpatient setting and is associated with few side effects. Some patients notice sore throat or neck tenderness (reflecting a radiation thyroiditis) but this is usually mild and transient. Its long-term efficacy is well established, as is long-term safety in terms of cancer risk, which we and others have investigated in large cohorts (33). There are few, if any, contraindications to radioiodine therapy apart from inability to comply with local radiation protection regulations. Such compliance may be difficult to achieve in hospital or nursing home residents, those with urinary incontinence, and those with significant mental impairment. In such cases, long-term thionamide therapy is often the best practical option (see above).
A relative contraindication to the use of radioiodine in cases of Graves' thyrotoxicosis is the presence of moderate or severe ophthalmopathy. This is based on evidence for a slightly increased risk of development or worsening of pre-existing thyroid eye disease in those treated with radioiodine compared with thionamides or surgery (34). Problematic eye disease is more likely in those with pre-existing ophthalmopathy, in smokers (smoking being an independent risk factor for development of ophthalmopathy in those with Graves’ disease) and those with severe biochemical disease. In view of evidence originally from Bartalena et al (35) that giving a course of glucocorticoid abolishes any increase in risk of ophthalmopathy in those receiving radioiodine, many experts prescribe a short course of prednisone/ prednisolone around the time of therapy.
In those with severe clinical and biochemical thyrotoxicosis it is desirable to restore euthyroidism before proceeding to radioiodine therapy. This is because of the theoretical risk of inducing "thyroid storm" due to thyroid destruction and release of pre-formed thyroid hormones following radioiodine administration, together with the need to stop thionamide therapy for up to two weeks at the time of treatment. In mild cases (judged both clinically and biochemically), such pre-treatment with thionamides may be unnecessary and radioiodine may be given as initial therapy or after short-term preparation with beta-adrenergic blockers.
Many studies have attempted to define optimal radioiodine doses in the hope of inducing euthyroidism and avoiding iatrogenic hypothyroidism in all (36-38). Such studies have examined attempts to titrate doses of radioiodine according to factors such as thyroid size (judged clinically or by imaging), isotope uptake or isotope turnover in the thyroid. Older literature suggested that cases of toxic nodular hyperthyroidism require larger doses of radioiodine to induce euthyroidism than cases of Graves' disease. It is clear, however, from review of recent literature that measures of thyroid size or isotope uptake/turnover generally do not allow effective "dose titration". Furthermore, the dose of radioiodine required to cure toxic nodular hyperthyroidism is not different from that required in Graves' disease in the majority of cases (39). In some subjects with large goiter, higher initial doses or multiple treatments are required.
Many large thyroid centers thus avoid attempts at radioiodine "dose titration" and administer empirical doses. Such an approach avoids the necessity for extra hospital visits to document isotope uptake into the gland or the need for other imaging. The dose of radioiodine administered varies between centers, and is determined in part by radiation protection restrictions that vary considerably around the world. Typically, a dose of radioiodine is chosen which can be administered in the outpatient setting and which results in cure of thyrotoxicosis in the majority after a single dose, while not inducing hypothyroidism in all. In iodine-replete parts of the world such as the US and UK, a standard dose of radioiodine is 10-15 mCi or 400-600 MBq. In our own series (40) a dose of this size resulted in cure of thyrotoxicosis in more than two thirds, at a cost of early hypothyroidism in 50%. Some centers administer larger doses to those with large goiter or to men, in view of some evidence of relative radioresistance in these groups. There is also evidence (some of it conflicting) that prescription of thionamides, especially use of propylthiouracil, before and/or after radioiodine treatment also induces relative radioresistance and hence the need for repeat dosing or a larger initial dose (40;41). It has been suggested that large doses should be administered routinely to elderly subjects, particularly those with cardiovascular disease or complications, to be certain of rapid restoration of euthyroidism. This view is tentatively reinforced by our own evidence that effective cure as indicated by the development of hypothyroidism requiring thyroxine replacement therapy is associated with a lessening of vascular mortality (compared with those not rendered hypothyroid) (9) and more likely conversion to sinus rhythm in those with AF associated with hyperthyroidism (11).
Thionamide therapy should be withdrawn 3-7 days before radioiodine (to allow iodine uptake into the thyroid) and should be recommenced after a similar period post-treatment if the elderly subject has severe disease, incomplete biochemical control, significant complications e.g. atrial fibrillation, or has experienced return of symptoms in the short period of thionamide withdrawal before radioiodine therapy. After therapy, clinical and biochemical assessment should be carried out every 4-6 weeks for the first few months so that thionamide doses may be adjusted (according to free T4) and hypothyroidism identified. A transient rise in serum TSH may be seen in the first few months after radioiodine and does not necessarily indicate permanent hypothyroidism but more marked biochemical or symptomatic hypothyroidism usually indicates the need for life-long T4 therapy. Persistence of biochemical hyperthyroidism 6 months after radioiodine therapy usually indicates the need for re-dosing. Unless small empirical doses are administered, the vast majority of patients with either toxic nodular hyperthyroidism or Graves' disease are rendered euthyroid (off all treatment) or hypothyroid (on T4) with one, two or (uncommonly) three doses (40). Occasional cases of apparent resistance to radioiodine treatment are seen.
Long-term, all patients treated with radioiodine require biochemical follow-up for detection of hypothyroidism (36). This is most efficiently achieved by using a computerized recall system to ensure annual testing of thyroid function. Such follow-up is essential since the incidence of hypothyroidism is significant even many years after radioiodine and eventually up to 90% of those treated in this way become hypothyroid. Hypothyroidism rates may be slightly lower in those with toxic nodular hyperthyroidism (40) because of relative sparing of normal thyroid tissue through concentration of isotope in "hot" autonomous nodules.
Most experts agree (as reviewed in 24) that surgery has little part to play in routine management of thyrotoxicosis. The relatively higher risk of complications of anesthetic and surgery in elderly subjects strengthens this view.
If surgery is contemplated, it is essential that clinical and biochemical euthyroidism are restored beforehand. This requires therapy with thionamides, typically for 2-3 months after diagnosis. Pre-operative preparation with beta-adrenergic blockers alone or alternatively Lugol's iodine is considered inadequate (42). Thorough preparation is essential in order to avoid thyroid storm post-operatively, as well as other significant complications of thyroid hormone excess, especially cardiovascular complications.
There is on-going debate regarding the most appropriate surgical approach for treatment of thyrotoxicosis. Many large centers now advocate the use of total thyroidectomy for Graves' hyperthyroidism since partial thyroidectomy is associated with significant rates of short - and long-term recurrence (43), while in expert hands surgical complication rates should be similar (44). Such complications include bleeding into the neck, hypoparathyroidism and damage to recurrent laryngeal nerves. Hypothyroidism is inevitable after total thyroidectomy (the patient leaves hospital on T4 therapy) but is also common after partial thyroidectomy. Life-long follow-up (as with cases treated with radioiodine) is essential for detection of hypothyroidism (and recurrence of hyperthyroidism) after partial thyroidectomy.
Cases of toxic nodular hyperthyroidism may be treated by thyroid lobectomy or excision of a single hot nodule. Such an approach has the theoretical advantage of avoidance of hypothyroidism, as well as improvement in cosmetic appearance in those with large goiter. It should be noted, however, that reduction in nodule/goiter size is also evident after radioiodine therapy, albeit after several months. Surgery may be considered appropriate if toxic nodular goiter is associated with obstructive symptoms or if there is any specific concern about the presence of co-existent malignancy in the goiter/nodules.
This is essentially a biochemical diagnosis as described above, the finding of an undetectable, as opposed to low but detectable, serum TSH concentration being of more pathophysiological significance. The most common cause of suppression of TSH in the general population is exogenous thyroid hormone therapy, typically T4. Population surveys such as that reported by canaries et al (45) have shown that approximately one quarter of those prescribed T4 long-term display reduction in TSH suggestive of mild over-treatment; (this is deliberate in the relatively small number of patients with a past history of thyroid cancer). Since T4 is prescribed to about 5% of the over 60's, this medication is a common cause of subclinical hyperthyroidism (46). Once "non-thyroidal illness" and relevant drug therapies have been excluded (see above), then nodular goiter is the next most common cause of low serum TSH in this age group. In subjects with a nodular goiter, either detectable clinically or evident on isotope imaging, suppression of serum TSH represents the earliest biochemical marker of thyroid autonomy and onset of hyperthyroidism.
There is only limited evidence to suggest that subclinical hyperthyroidism is associated with significant symptoms but there is a growing body of evidence that low serum TSH is associated with adverse effects, particularly on heart and bone. "Endogenous" subclinical hyperthyroidism, for example secondary to nodular goiter, is probably of greater significance than "exogenous" due to T4 therapy since the former is associated with high/normal serum T3 concentrations.
An important study of the Framingham population of the US reported by Sawin et al (47) revealed a 3-fold increased incidence of atrial fibrillation in those subjects aged over 60 with serum TSH of less than 0.1 mU/L compared with those with normal serum TSH. The likelihood of developing atrial fibrillation was also increased, but less markedly, in those with low but detectable TSH. The group in this survey with low TSH was heterogeneous and included some subjects taking exogenous T4 therapy. Similar findings have been reported by Cappola et al (48) who again described an association between subclinical hyperthyroidism and risk of development of AF. Furthermore, we have recently shown in a large cross-sectional study that not only is subclinical hyperthyroidism independently associated with the finding of AF on 12-lead ECG in elderly subjects, but that even variation of serum free T4 concentration within the normal reference range is associated with the finding of AF, consistent with the view that even the mildest degree of thyroid hormone excess can have cardiovascular consequences in the elderly age group (49). We have also examined vascular mortality in a group of subjects aged over 60 (excluding those taking T4) and found significant increases in both cardiovascular and cerebrovascular deaths in those with serum TSH below the normal range compared with those with normal TSH (50). These findings suggest that data showing effects of mild hyperthyroidism on indices of cardiac function and on cardiac rhythm do indeed translate into significant adverse influences, especially in elderly subjects.
Similar adverse effects of subclinical hyperthyroidism on bone may occur. Meta-analyses of studies examining effects of subclinical hyperthyroidism on bone mineral density have concluded that there are significant reductions in post-menopausal women (15), although these studies are mostly confined to subjects taking T4. There is also some evidence for improvement in bone metabolism or BMD after treatment of endogenous subclinical hyperthyroidism (51). Whether any effect of endogenous subclinical hyperthyroidism on bone metabolism translates into increased risk of fracture is remains to be established, although Bauer et al (52) have reported that subclinical hyperthyroidism is associated with increased risk of new hip and vertebral fractures after adjustment for confounding factors including previous overt hyperthyroidism. So far, evidence from epidemiological studies suggests that T4 therapy alone is not a risk factor for hip fracture, except perhaps in men, although a previous history of overt hyperthyroidism is an independent risk factor (17;53).
Concerns about effects of mild thyroid hormone excess upon heart and bone have led to a trend towards treatment of this condition. In those taking exogenous thyroid hormone, management is relatively straightforward, namely reduction in prescribed dose and re-checking of serum TSH 6-8 weeks later. For those not taking T4, many experts, especially in the US, now administer either antithyroid drugs or radioiodine to those with persistent subclinical hyperthyroidism secondary to nodular goiter or Graves' disease, especially in subjects with atrial fibrillation or other underlying cardiac disease. Prospective trials confirming benefit of such therapy have yet to be performed but consensus guidelines suggest that elderly subjects, those in AF and those with other vascular risk factors should be considered for treatment (54).
Several factors need to be considered before a decision should be made to institute either population or targeted screening in groups such as the elderly. Firstly screening programs should be instituted only for those conditions in which the benefits of screening outweigh the costs. Whether benefits outweigh the costs depends on accurate quantification of these issues, then a judgment as to whether the costs of screening are justified. Although it is clear that hyperthyroidism is common, there are no data that demonstrate that such subjects when identified by screening benefit from being so diagnosed; it is not sufficient to demonstrate only that such subjects exist. Such benefits and costs should ideally be based upon the results of a randomized controlled trial in an appropriate sample of the relevant population. In considering costs, those incurred by those who do not themselves gain from the screening program should be considered. If, for example, the screening process uses a test such as serum TSH with a high false positive rate, then a considerable number of patients may be exposed to investigations which are unnecessary, with accompanying risk and potential morbidity.
While both overt and subclinical hyperthyroidism are common in older subjects, and while there is growing evidence for adverse consequences of both of these diagnoses, present lack of evidence that treatment in a screened population improves morbidity/mortality, and that the risks of such treatment outweigh the costs, means that screening is presently unjustified (54). There should, nonetheless, be a high index of suspicion for hyperthyroidism in this age group and a low threshold for biochemical testing, especially in those with a previous personal or family history of thyroid disease or those with conditions such as atrial fibrillation that may reflect hyperthyroidism.