Introduction
The thyroid undergoes slight “physiological” changes with aging, either as a result of its participation in the senescence process or as an effect of other systems’ changes. Thus, thyroid alterations in the elderly may be either the expression of aging evolution by itself or the result of the senescence process. Moreover, some thyroid diseases are more frequently encountered in the elderly and others, which start as subtle dysfunctions in younger people, may appear as clinically overt disorders in older individuals. Hypothyroidism from atrophic chronic autoimmune thyroiditis is known to be much more common in the elderly. Non-toxic goiter, usually starting as a diffuse thyroid enlargement in younger people, frequently acquires nodularity and autonomous function with age and eventually, but not necessarily, evolves in toxic nodular goiter. In their evolution from normal to abnormal thyroid function, both chronic autoimmune thyroiditis and goiter often show a phase of subclinical thyroid dysfunction (subclinical hypothyroidism and subclinical hyperthyroidism, respectively). Frequently, the interpretation of thyroid function tests is also difficult in aging individuals, because true “physiological” age-associated changes cannot be easily distinguished from the alterations secondary to subclinical thyroid disease, to acute or chronic nonthyroidal illness and / or drugs often taken by elderly patients. The definition of “physiological” age-associated thyroid changes with aging and the correct understanding of these modifications is not merely speculative, since it greatly helps in differentiating “physiological” alterations from subclinical thyroid disease and in resolving the question of treatment of uncertain clinical situations. Moreover, the effects of thyroid dysfunctions may be to a variable extent different in the elderly. The consequences of hyperthyroidism on cardiovascular function and of hypothyroidism on atherosclerosis, that are more marked in older than in younger individuals, are two examples of this situation.
We will present some topics on thyroid function and diseases in the elderly, with particular accent on the latest and more controversial issues.

Physiological changes of thyroid with aging
The results of studies on thyroid function in the elderly have been long controversial because of the inaccurate selection of study samples (¹). Indeed, it is now apparent that subclinical thyroid disease and nonthyroidal illness are common in the elderly. Thus an accurate selection of population samples, which must include only healthy subjects, is crucial to define the normal thyroid function in the elderly. In a population of healthy elders, carefully selected to exclude subjects with nonthyroidal illness, we showed that TSH decreased, FT4 remained steady, FT3 decreased and rT3 increased with aging (²). These changes of TSH and thyroid hormones are consistent with the presence in the elderly of a partial central hypothyroidism, associated with an impaired activity of type I deiodinase (5’D-I). A similar reduction of TSH levels had been previously reported in carefully selected healthy old subjects (³) and in a small proportion of unselected ambulatory patients, unrelated to subclinical hyperthyroidism (⁴). However, in the Whickham survey, after people with circulating thyroid autoantibodies and a family or personal history of thyroid disease had been excluded, the levels of TSH in aging people resulted unchanged (⁵). Moreover, a recent survey on a large population in the USA (⁶) reported that in the reference population, which included people without circulating thyroid autoantibodies and no other risk factors for thyroid function, TSH increased with aging. These dissimilar results may be explained, to some extent, with a diverse iodine intake and a different prevalence of subclinical thyroid disease in the populations evaluated in these studies. In conclusion, a number of complex, albeit minor, changes in functional parameters occur with aging: the question of whether these changes are beneficial or detrimental to the aging process per se remains to be established.

Aging and thyroid autoimmunity
An age-dependent increase of the prevalence of anti-thyroperoxidase (TPOAb) and anti-thyroglobulin autoautoantibodies (TgAb), particularly in females over 60 years of age has been long documented (reviewed in ^{1; 6}). Nevertheless, with the exception of atrophic lymphocytic thyroiditis, the prevalence of clinically overt autoimmune thyroid disease is not increased in the elderly. We investigated the prevalence of TPOAb and TgAb in very old subjects (⁷). The prevalence of thyroid autoantibodies resulted significantly greater in subjects aged 70-85 compared with people aged less than 50. By contrast, the prevalence in centenarians was not different from subjects aged less than 50. Therefore, the age-dependent increase in the prevalence of thyroid autoantibodies in the elderly is not seen after the ninth decade of life. In another study, we reported that the prevalence of thyroid autoantibodies was higher in unselected or hospitalized elderly patients and lower in centenarians and in “younger” healthy elderly as compared with an apparently normal population (⁸). Therefore, the appearance of thyroid autoantibodies might be related to age-associated disease, rather than to the aging process per se and healthy older subjects might represent a selected population with an unusually efficient immune system (¹). Moreover, it is well established that thyroid autoimmunity is often associated with other autoimmune diseases.
On these grounds, it is appealing to speculate a link between circulating thyroid autoantibodies and other diseases. Indeed, some authors have proposed a correlation between thyroid autoimmunity and some age-associated diseases, in which (auto)immune mechanisms are believed to be involved (reviewed in⁹) The increased prevalence of circulating thyroid autoantibodies observed in the elderly might reflect an activation of the (auto)immune process involved in the atherosclerotic process. However, the slight increase of coronary heart disease associated with positive serum thyroid autoantibodies, reported by some epidemiological studies (reviewed in¹) has not been confirmed by other reports (reviewed in ^{1; 10; 11; 12}). In 1966 Brain et al. described a patient that was on levo-thyroxine treatment for hypothyroidism, due to Hashimoto’s disease, who developed several episodes of cerebral disorders (¹³). After this first report, few papers describing an association of thyroid autoantibodies with encephalopathy (“Hashimoto’s encephalopathy”) have been published (reviewed in ¹⁴). However, chronic lymphocytic thyroiditis is seldom associated with serious neurological manifestations, with the exception of myxedema coma, and the reports on “Hashimoto’s encephalopathy” are very few compared with the large prevalence of thyroid autoimmunity in general population. A recent paper, describing an elevated numbers of perfusion defects in euthyroid patients with autoimmune thyroiditis, recalls the hypothesis that cerebral vasculitis might be implicated in this situation (¹⁵). Therefore, it is conceivable that not Hashimoto’s thyroiditis by itself but some correlated autoimmune disorder might be responsible for Hashimoto’s encephalopathy. However, a definitive role for thyroid autoantibodies in neurological disorders has not been proved yet.

Hypothyroidism and hyperthyroidism in the elderly
Definition of overt and subclinical hypothyroidism and hyperthyroidism
Overt hypothyroidism is characterized by low levels of thyroid hormones and increased levels of TSH, whereas overt hyperthyroidism is characterized by high levels of thyroid hormones and a low TSH. The term subclinical thyroid disease currently include subclinical hypothyroidism and subclinical hyperthyroidism, which are defined as states with serum TSH levels either above or below the reference range, respectively, and normal thyroid hormone levels in the absence of clinical symptoms and signs. The definition of subclinical thyroid disease is widely recognized as somewhat arbitrary and uncertainties in the definition and the terminology of subclinical hyperthyroidism and subclinical hypothyroidism should also be considered (¹⁶).
Prevalence of hypothyroidism and hyperthyroidism in the elderly
The prevalence of hypothyroidism and hyperthyroidism varies widely according to gender, age and the environment. As reported in iodine sufficient areas, hypothyroidism occurs in 4 to 9.5% of the general population, hyperthyroidism in 0.4 to 3.2% (^{5; 6; 17; 18}). Both subclinical hypothyroidism and subclinical hyperthyroidism are 2-4 fold more common than the corresponding overt conditions and are more prevalent in women and in the elderly. Hypothyroidism is more common in iodine sufficient areas, hyperthyroidism in iodine deficient areas (¹⁹). In the latter, nodular goiter with hyperthyroidism or various degrees of thyroid autonomy, including complete TSH suppression, is relatively frequent (^{20; 21; 22; 23}). The prevalence of subclinical hyperthyroidism resulted 15% in elderly subjects (>75 years) living in Pescopagano, an iodine deficient area (²¹), and 6.5% in people over 85 years living in the United States, an iodine sufficient area (⁶). In the Pescopagano survey the prevalence of subclinical hyperthyroidism resulted equal in the two sexes (²¹), whereas in the United States was more frequent in women (⁶). The prevalence of hypothyroidism increases with aging up to 17-20% of females and 3-16% of men over 75 years of age in iodine sufficient areas (^{5; 17}).

Etiology of hypothyroidism
Several conditions, transient or permanent, may be associated with an elevated serum TSH level and normal or low serum thyroid hormone concentrations (subclinical or overt hypothyroidism, respectively). Half of cases of permanent hypothyroidism of the elderly are due to chronic autoimmune thyroiditis (Hashimoto’s thyroiditis), a condition that is more common in women. Indeed, a higher prevalence of thyroiditis in patients over age 60 than in younger women has been reported by an autoptic study (²⁴). Moreover, in the Whickham survey the incidence of antithyroglobulin and antimicrosomial autoantibodies was 7% and 9% in females over 75 years of age compared with 2% and 5% in total population, respectively (⁵). A permanently elevated TSH may be also caused by a previous treatment of hyperthyroidism, by congenital hypothyroidism (including inactivating mutations of the TSH receptor) and by external radiotherapy. Circulating bioinactive TSH and heterophilic antibodies against mouse protein can induce falsely elevated levels of TSH. A transient increase of TSH levels is typical of the recovery phase from postpartum thyroiditis and subacute thyroiditis, of nonthyroidal illness and of drug (lithium, amiodarone) (reviewed in ²⁵) treatment.
Clinical presentation and diagnosis of hypothyroidism
The clinical appearance of hypothyroidism may be subtle (²⁶), particularly in elderly patients, in whom the classical symptoms of overt hypothyroidism may be confused with normal aging. Furthermore, clinical presentation of hypothyroidism in the elderly is quite different than in younger individuals (reviewed in ²⁷). “Subclinical” hypothyroidism, particularly in presence of TSH levels >10 mIU/L, has been associated with neuromuscular symptoms (²⁸), increased cholesterol levels and other lipid abnormalities (^{17; 29; 30}), cardiac alterations (^{31; 32; 33; 34}) and vascular impairment (^{35; 36}). Subclinical hypothyroidism has recently been reported to be an independent risk factor for atherosclerosis and myocardial infarction in post-menopausal women (¹¹) and in men (¹²). However, in another paper mortality was unchanged in presence of subclinical hypothyroidism (³⁷). The validity of most of the above mentioned studies has been questioned by a recent review (¹⁸). In our view, identify the etiology of subclinical hypothyroidism is mandatory, to differentiate chronic autoimmune thyroiditis, previous treatment of thyroid disease and other causes of permanent thyroid damage from rare case of isolated hyperthyrotropinemia caused by a TSH increase due to partially inactivating mutations of the TSH-receptor or other causes (¹⁶).
Treatment of hypothyroidism
Treatment of overt hypothyroidism in elderly patients should be started and monitored carefully in order to maintain TSH and FT4 within their normal ranges. The initial dose of levo-thyroxine should be very low (even 12.5 μg/day) and increased every 4-6 weeks, in order to reach the replacement dose in 3-4 months. We have recently shown that individual L-T4 requirements are dependent on lean body mass (³⁸). The replacement dose of L-T4 in the elderly is usually lower than 1.6 μg/kg/day, the dose usually employed in younger patients. This reduction appears dependent on a relative decrease in lean body mass with aging. Replacement treatment must be carefully monitored in elder patients, with TSH measurement every three months, in order to avoid overtreatment. In some patients a lesser than substitutive dose of levo-thyroxine is required to avoid ischemic heart symptoms.
While need of treatment of clinical hypothyroidism is unanimously accepted, that of subclinical hypothyroidism is under discussion. A major matter of debate concerns the role of the “evidence-based medicine” approach in managing subclinical hypothyroidism. Recently, a consensus panel recommended against treatment of patients with subclinical hypothyroidism with TSH level 4.5-10 mIU/L, because there were no large randomized control trials to support benefit (¹⁸). This and other panel’s suggestions have later been criticized by a Consensus Statement from the American Association of Clinical Endocrinologists, the American Thyroid Association and the Endocrine Society that stated that these recommendations were based on “ the lack of evidence for benefit rather than on evidence for a lack of benefit” and that there is no evidence that treatment of subclinical hypothyroidism is harmful (^39;40). We share the latter view, and recommend that a practical approach should be used, taking into account that, even if not definitively, many reports have shown that treatment may normalize some abnormalities induced by subclinical hypothyroidism. Indeed, a direct correlation between TSH values and total and LDL cholesterol levels has been demonstrated (^17;41). The optimal ranges for total and LDL cholesterol are still uncertain and the new concept is that “lower is better” (⁴²). In this setting an aggressive control of LDL cholesterol levels is required, particularly in presence of coronary heart disease (CHD) or CHD risk equivalents (⁴³), a common occurrence in the elderly. Basically, in presence of an increased cholesterol level, determination of TSH value is already a widely accepted practice (⁴³). Moreover, effectiveness of treatment of subclinical hypothyroidism in normalizing total and/or LDL cholesterol levels and other lipid abnormalities, particularly in patients with greater basal TSH and cholesterol levels, have been reported in many (^{44; 45; 46; 47}), though not all studies (^48;49). Therefore, when subclinical hypothyroidism is associated with elevated cholesterol values, clinical judgment would suggest that treatment with levo-thyroxine would be beneficial for CHD, despite the fact that no definitive data are available. Similar considerations can be drawn for the other abnormalities reported in subclinical hypothyroidism, with papers showing that treatment has beneficial effect on non-specific symptoms (^{28; 46}), cardiovascular dysfunctions (^{28;33; 46;50}) and vascular impairments(^{35; 36}). A different conclusion has been reached by other authors (⁴⁹). The bulk of available data are sufficient to support treatment of subclinical hypothyroidism, whereas management of isolated hyperthyrotropinemia is still uncertain (^{16;51; 40;52}).

Etiology of hyperthyroidism
In the elderly, the most common causes of permanent hyperthyroidism are Graves’ disease in iodine sufficient areas, nodular goiter and functioning thyroid adenoma in iodine deficient areas, respectively (^{20; 21}). A transient thyrotoxicosis may occur in silent and subacute thyroiditis and during treatment with drugs including amiodarone (particularly in iodine deficient areas) (reviewed in ²⁵), levo-thyroxine and interferon. With the exception of amiodarone-induced thyhrotoxicosis, which may represent a major therapeutic challenge (²⁵), the other transient drug-induced thyrotoxicosis are self-limited and remit with drugs withdrawal.
Clinical presentation and diagnosis of hyperthyroidism
The hyperkinetic symptoms, typical of clinical hyperthyroidism in younger patients, are generally less evident or absent in older patients, that usually exhibit the so called “apathetic” hyperthyroidism (reviewed in ⁵³). Atrial fibrillation is more common in elderly hyperthyroid patients than in younger patients (⁵⁴). Age, male gender, ischemic heart disease, congestive heart failure, aortic and/or mitralic valve disease have been identified as risk factors for atrial fibrillation in patients with hyperthyroidism (⁵⁵). Angina pectoris may occasionally occur in hyperthyroid patients with normal coronary arteries. Hyperthyroidism in the elderly is associated with greater bone density reduction and fracture risk (reviewed in⁵⁶). In clinically euthyroid patients over 60 years a TSH value <0.1 mIU/L was associated with a relative risk of atrial fibrillation of 3.1 (⁵⁷) and, more recently, a TSH level <0.5 mIU/L in people aged 60 years or older, not taking levo-thyroxine, was related to increased mortality (³⁷). Surprisingly, in the latter study no difference was noted between patients with low but detectable TSH and patients with undetectable TSH. Further studies are needed to confirm these data. Moreover, subclinical thyrotoxicosis due to suppressive levo-thyroxine treatment has been associated with cardiac abnormalities (^{58; 59}). A decreased bone density has been reported in post-menopausal women taking levo-thyroxine suppressive treatment (^{60; 61; 62}), whereas we did not find any significant effect on bone mass and bone metabolism in men in whom levo-thyroxine had been carefully titrated in order to use the minimal dose able to suppress TSH (⁶³). Since studies investigating the effects of subclinical hyperthyroidism often included patient on suppressive levo-thyroxine treatment, the opportunity to extend these conclusions to spontaneous subclinical hyperthyroidism has been questioned (⁶⁴).
FT4 is elevated in most but not all elderly hyperthyroid patients. The concomitance of T3-thyrotoxicosis, that is more common in the elderly, confirm the diagnosis. However, FT3 may be normal in many elderly patients and if the normal decrease of T3 with aging is not considered, an elevated FT3 may be mistakenly considered as normal. Furthermore, nonthyroidal illness and drugs (propanolol and iodinate contrast material) may lower T3 levels.
TSH levels are suppressed in hyperthyroidism, with the exception of central hyperthyroidism. As reported above, diagnosis of subclinical hyperthyroidism may be cumbersome and the presence of an isolated low TSH value, particularly if only partially suppressed, may not necessary indicate subclinical hyperthyroidism. Indeed, a low serum TSH value (<0.1 mUI/L) alone showed a low predictive value for hyperthyroidism in person older than 60 years, while a concomitant elevated T4 raised its predictive value (⁴). Administration of drugs that reduce TSH levels, as dopamine and glucocorticoid, must be ruled out. Positive circulating anti-thyroid autoantibodies and a goiter at palpation or ultrasound examination are useful tools to confirm the presence of thyroid abnormalities and to distinguish between Graves’ disease and nodular goiter.

Treatment of hyperthyroidism
Although clinically hyperthyroid elderly patients may be maintained on anti-thyroid drugs, definitive treatment with 131-I or surgery is advisable. 131-I is the treatment of choice, especially when concomitant cardiovascular diseases are present, while surgery may be considered in presence of obstructive symptoms caused by large goiters, nevertheless radioiodine may be successfully used in large compressive goiters (⁷⁶ reviewed in ⁶⁵). Administration of beta blockers may prevent symptoms of thyrotoxicosis following 131-I treatment. 60% of patients with atrial fibrillation reverted to sinus rhythm within four months after restoration of euthyroidism (⁶⁶). A long duration of hyperthyroidism and old age are negative determinants for reversion to sinus rhythm.
At the moment, most authors advice treatment of elderly patients with subclinical hyperthyroidism and a clearly suppressed TSH level (<0.1 mUI/L) and follow up for patients with TSH levels between 0.1 and 0.4 mUI/L (^{16; 18; 40}). Beta-blockers have been proposed to reduce the cardiac effects of levo-thyroxine suppressive treatment (⁶⁷).

Screening for subclinical hypothyroidism and subclinical hyperthyroidism
As a consequence of the uncertainty about effectiveness of treatment of subclinical thyroid disease, and particularly of subclinical hypothyroidism, the need of screening for thyroid disease is under discussion as well. Benefits of screening have been considered inadequate by some authors, who rather advocated aggressive case-finding in subjects at risk (^{18; 68}). In our view, the relatively high frequency of subclinical thyroid disease in the elderly population, the vague clinical presentation of hypothyroidism in this age group and the need for treatment of many patients are strong arguments favoring the screening for thyroid disease at this age. This is in keeping with the recommendations of the American Thyroid Association and the American Association of Clinical Endocrinologists (^{40; 69}).

Non toxic goiter in the elderly
With age, goiter size increases, thyroid nodularity develops and TSH decreases. The largest goiters are observed in the oldest age groups, living in iodine deficient areas. At this age goiters are commonly multinodular and are frequently associated with a suppressed TSH, expression of thyroid autonomy, that can progress to overt thyrotoxicosis. The prevalence of diffuse and nodular goiter in young adults participating in the Pescopagano (an iodine deficient area) survey was 30% in young adult and increased up to 75% in the age group 55-65 years, with nodular goiter accounting for about one third of the total (²¹). Goiter is more common in women than in men (⁷⁰). As reported above, in the Pescopagano survey, prevalence of thyroid functional autonomy was 15% in subjects over 75 years, with multinodular goiter accounting for the majority of cases. Large goiters may cause obstructive symptoms. Follow-up is the choice when symptoms are absent, whereas treatment is warranted in goiters presenting with hyperthyroidism (and thyroid autonomy) or compressive symptoms. Since thyroid autonomy is very common in the elderly, suppressive levo-thyroxine treatment is not recommended at this age. 131-I is the standard treatment for thyroid autonomy and hyperthyroidism, whereas surgery see above is advised for large non-toxic goiters causing significant compressive symptoms. 131-I therapy has been proposed in order to reduce thyroid volume in non-toxic goiters, with satisfactory results In multinodular goiters, the nodules are structurally and functionally heterogeneous, and are separated by varying amounts of extracellular matrix. In addition, large nodular goiters often present with cysts, hemorrhage, fibrosis and calcifications. Shrinkage of goiter in the 6-12 months after 131-I treatment has been described in several studies, but there was a large variation in the radioiodine activity administered and the degree of shrinkage. Certainly, the reduction of goiter size is related to the dose of 131-I administered per gram of thyroid tissue and very large nodular goiter, with irregular distribution of radioiodine uptake due to cold nodules and areas of involution, may require bigger doses of 131-I.
Recently, administration of recombinant human TSH (rhTSH) before 131-I has been proposed in non-toxic and autonomously functioning goiter. In the earlier studies, rhTSH increased 131-I uptake in nontoxic nodular goiter (⁷¹), particularly in cold areas (⁷²). More recent reports have indicated that pretreatment with rhTSH may increase the efficacy of 131-I therapy, supporting the concept that it may reduce the required activity of radioiodine (^{73; 74}). Acute side effects have been reported to be absent in one study (⁷³), and relatively common in another (⁷⁴). Since these studies altogether included a relatively small number of subjects with different thyroid size and function (non toxic, autonomous, toxic), definitive conclusions can not be drawn. In particular, the amount of rhTSH to be administered, the timing of its administration and the dose of 131-I have not been established yet. Thus, appropriate treatment protocols should be defined in order to avoid severe acute adverse effects caused by radiation-induced thyroiditis and oesophagitis after rhTSH pretreatment (⁷⁵).

Conclusions
The aging process is associated with a number of thyroid function changes. The question of whether and to what extent these changes are expression of the aging process per se or of an age-associated thyroidal and nonthyroidal illness is a matter of debate. Studies on healthy elderly subjects suggest that several age-related changes in thyroid function are indeed independent of nonthyroidal illness. Human aging is often associated with an increased prevalence of thyroid autoantibodies, that may be not the consequence of the aging process by itself, but rather an expression of age-associated disease. Thyroid diseases in older patients differ from those observed in younger patients in their prevalence and clinical expression and their treatment often deserves special attention because of the increased risk of complications. For all these reasons clinical evaluation of thyroid status is, to some extent, different in elderly patients and deserves a particular consideration.