The thyroid axis in ageing

Holger Leitolf, Jens Behrends and Georg Brabant1

Department of Clinical Endocrinology, Center for Internal Medicine and Dermatology, Medizinische Hochschule Hannover (MHH), 30625 Hannover, Germany

Abstract. The hypothalmo—pituitary—thyroid axis, among various endocrine systems, undergoes physiological alterations associated with the ageing process. Directly age-related changes have to be distinguished from indirect modifications which are caused by simultaneous thyroidal or non-thyroidal illness or other physiological or pathophysiological states whose incidence increases with age. In summary, direct changes of the hypothalmo—pituitary—thyroid axis seem to be subtle and suggestive of a decreased hypothalamic stimulation of thyroid function. In parallel, disease-specific alterations such as the development of thyroid autonomy or changes in energy intake or sleep lead to pronounced alterations ofthyroid function with age which may dominate the underlying ageing of the hypothalmo—pituitary—thyroid axis itself. The following article attempts to delineate some aspects of the interplay of the regulation of thyroid function and the ageing process.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242) p 193-204

Ageing is associated with physiological alterations in various endocrine systems including the hypothalamo—pituitary—thyroid axis. A number of studies have attempted to delineate alterations in thyroid physiology due to the ageing process (Felicetta 1988, Levy 1991, Mooradian & Wong 1994). When focusing on age-dependent changes, alterations of the hypothalamo—pituitary—thyroid axis have to be considered which are not directly age-related but rather indirectly caused by simultaneous thyroidal or non-thyroidal illness or other physiological and pathophysiological modifications whose incidence increases with age. Nevertheless, these events have important influences on thyroid physiology and function.

Physiological regulation of the hypothalamo—pituitary—thyroid axis is characterized by pulsatile stimulation of pituitary thyrotropin (TSH) secretion resulting from an interplay of stimulatory hypothalamic influences mediated by

1This chapter was presented at the symposium by Georg Brabant, to whom correspondence should be addressed.

thyrotropin-releasing-hormone (TRH) pulses as well as inhibitory hypothalamic influences mediated by dopamine and somatostatin release. After being converted from the prohormone tetraiodothyronine (or thyroxine, T4) by specific deiodinases, triiodothyronine (T3) inhibits synthesis and secretion of TRH (Segerson et al 1987) and stimulates the release of inhibitory factors such as dopamine and somatostatin from the respective hypothalamic centres. Within the thyrotropic cells of the pituitary gland, T3 directly inhibits the synthesis of the TSH b-subunit and thereby TSH secretion by interaction with specific T3 receptors (Gurr & Kourides 1985). Thyrotropin is released from the pituitary gland in a pulsatile pattern with approximately one pulse every two hours. However, these pulses are not uniformly distributed but rather cluster within the evening hours. Fusion of TSH pulses with increasing pulse amplitude, leads to the nightly increase in mean TSH serum levels with a maximum between 0200 h and 0400 h (Brabant et al 1986, Greenspan et al 1986, Clark et al 1987, Table 1).

Many factors have been described which influence the secretory pattern of TSH. Pulse amplitude is rapidly dampened by increasing circulating T4 levels in healthy subjects. A similarly rapid effect can be observed after bolus application of glucocorticoids but, in contrast to thyroid hormones, this effect appears to be specifically exerted on the hypothalamic level whereas the responsiveness of the pituitary to TRH remains unaltered (Re et al 1976, Brabant et al 1987, 1989). Physiological alterations such as sleep and energy supply also exert profound effects on the hypothalamo—pituitary—thyroid system. In healthy volunteers, 36 hours of fasting significantly decreases TSH levels and almost completely suppresses the normal TSH increase during the night. Despite this profound effect, the pattern of TSH pulses and their frequency remains constant, indicating a selective effect on TSH pulse amplitude (Romijn et al 1990). As it is known that leptin plasma levels decrease with fasting, these may also serve as potential mediators. When starving rodents received intraperitoneal leptin injections twice daily, the fasting-dependent decrease in hypothalamic—pituitary—thyroid axis was reversed (Seoane et al 2000). Experimental sleep alterations have been shown to modify TSH secretion and change hypothalamic and pituitary responsiveness. Acute sleep withdrawal in healthy young subjects induces an activation of the nightly increase in TSH secretion whereas a longer lasting sleep deprivation period and sleep fragmentation induces a dramatic decrease in the mean 24 h TSH secretion (Brabant et al 1990, Behrends et al 1998, Spiegel et al 1999).

The examples described above are important as they all may affect the hypothalamo—pituitary—thyroid axis in ageing and may thus obscure physiological alterations. Moreover, in iodine-deficient regions such as Germany, the prevalence of goitre is estimated to range between 30—50% with an increasing frequency with age (Berghout et al 1990, Hintze et al 1991a, Hampel et al 1995). Alterations in iodine supply are known to change thyroid function. Most likely due

TABLE 1 Age-related changes of the thyroidal axis and the peripheral responsiveness to thyroid hormones



TSH response to TRH

= or decreased

Nocturnal TSH peak


Thyroid response to TSH

= or decreased

Radioactive iodine uptake


T4 production

= or decreased

5' deiodinase activity


T3 production

= or decreased

T4 / T3 degradation

= or decreased

Serum thyroid hormone binding


Total or free T4 serum concentration

= or decreased

Total or free T3 serum concentration


Reverse T3 serum concentration

= or increased

Metabolic rate


Lipid peroxidation

= or decreased

Malic enzyme, Na/K-ATPase, S14


to a higher sensitivity of the thyroid gland to TSH, at least in short-term experiments, circulating TSH levels drop by approximately 50% without any effect on the frequency of TSH pulses when iodine supply is experimentally decreased (Brabant et al 1992). Prolonged iodine deficiency results in a high number of patients with nodular goitre estimated to account for more than 15% of the population in Germany beyond 60 years of age. This percentage is even further increased if only hospitalized subjects are considered (Felicetta 1988). Overt or occult thyroid diseases have to be considered when assessing the physiological alterations of the hypothalamo—pituitary—thyroid axis associated with ageing, and their clinical presentation, at least in part, may differ substantially from the symptoms and signs found in younger patients (Kobberling et al 1981, Hennemann & Krenning 1987, Nordyke et al 1988, Trivalle et al 1996). In many studies, thyroid autonomy and autoimmune thyroid disease have been identified as the leading causes for hyperthyroidism in ageing (Trzepacz et al 1989, van Coevorden et al 1989, Brabant et al 1991, Greenspan et al 1991, Levy 1991, Mariotti et al 1993, Mooradian & Wong 1994, Samuels 1998). A large study of 583 healthy subjects including 34 centenarians revealed an age-dependent increase in thyroid-specific antibodies with a significant peak between 70 and 85 years, accompanied by a decrease in total and CD5+ B cells (Mariotti et al1992).

The predominant cause of hyperthyroidism in old age, however, is related to the widespread therapeutic use of thyroid hormones. Data from the Framingham study demonstrated in an unselected population of 2575 adults beyond the age of 60 (mean age 68.6 years) that 6.9% (2.3% of men and 10.0% of women) were receiving medical treatment with thyroid hormones. It is interesting to note that apart from the therapeutic use in goitre patients and as substitution therapy for hypothyroidism, approximately 20% of the subjects studied were using thyroid hormones without appropriate identifiable cause. Insufficient treatment could be demonstrated in 37% of the hypothyroid patients using thyroid hormones by elevated TSH serum levels (Sawin et al 1989). Approximately 6% of the group with lowered TSH serum levels in the Framingham study were hyperthyroid when all laboratory and clinical indicators were combined. This suggests that lowered TSH serum levels in the elderly may be under a dominant influence of other factors than the thyroid status alone (Sawin et al 1991). The ratio of 'truly' hyperthyroid subjects as compared to other causes of lowered thyrotropin serum levels changes even more dramatically when the study is focused on hospitalized patients. Our group recently investigated patients of a rehabilitation clinic after treatment of acute illness (Brabant et al 1996). Restricting this prospective analysis to those 619 patients with no previous suspicion of active thyroid disease during the primary hospital stay, 5 subjects (0.8%) were identified to have elevated TSH serum levels indicating hypothyroidism. As expected for a region with deficient dietary iodine supply, the prevalence of suppressed TSH serum concentrations was much higher (22.6%). In 19% of the subjects from this group (i.e. 4% of the total group), overt hyperthyroidism was diagnosed supporting the insidious clinical signs of thyroid dysfunction in elderly subjects (Stolte et al 1998, Sawin et al 1994). This fits to recent data in hospitalized patients where a frequency of 0.8% overt clinical (4.2% subclinical) hypothyroidism was found in an iodine-deficient region, whereas 1.5% of the patients (10.4%) in a region with sufficient dietary iodine supply and 7.6% of the patients (23.9%) with high dietary iodine supply could be shown to be clinically (subclinically) hypothyroid. In contrast, clinical and subclinical hyperthyroidism was observed in 3.4% of the patients in a region with iodine deficiency, 3.0% in an area with sufficient dietary iodine and 0% in a region with high dietary iodine supply (Szabolcs et al 1997).

In large population studies, the incidence of hypothyroidism in elderly patients varies from less than 1% to 17% depending on the iodine supply. Women are more commonly affected than men, and subclinical hypothyroidism is more frequent than overt hypothyroidism. Virtually all the cases of hypothyroidism found are related to autoimmune thyroid disease (Hintze et al 1991b).

Apart from the iodine supply, thyroid function may be altered by the availability of other micronutrients such as selenium in the thyroid gland itself as well as in major target tissues for thyroid hormones such as the liver. Ravaglia et al (2000) demonstrated lower serum selenium levels in a group of healthy subjects beyond 90 years of age as compared to younger subjects. This was accompanied by other changes such as lower zinc serum concentrations. Many studies underline the importance of selenium for normal bioactivity of the type II deiodinase (Beckett et al 1993, Meinhold et al 1993, Mitchell et al 1997, Hotz et al 1997) at least in serious iodine deficiency.

Total energy intake exerts a profound influence on thyroid function, mainly on the hypothalamic level but also on the pituitary and thyroid level. The importance of inadequate energy intake in ageing has been discussed in large population studies. In the Euronut—Seneca study (de Groot et al 1992) on 2600 subjects born between 1913 and 1918 in 18 European communities, body mass index was below 20 kg/m2 in up to 15% of elderly men and 17% of elderly women, in addition, the daily energy intake in these subjects was found to be lower than their respective requirements.

Similarly, in severe non-thyroidal illness (NTI) the decrease in mean TSH secretion (Custro et al 1994) follows a comparable pattern to fasting with significantly reduced TSH pulse amplitude but an unchanged pulse frequency (Romijn et al 1990). Malnutrition may at least play a partial role in the explanation of disturbed thyroid function in NTI. The prevalence of chronic diseases in the Euronut—Seneca study varied between 59% and 92% (Danforth & Burger 1989, de Groot et al 1992), thereby indicating that NTI may introduce a significant bias when assessing age-dependent changes of the thyroid axis. Under these pathophysiological conditions, the glucocorticoid axis may be activated. This may suppress, to a variable degree, TSH secretion, an effect potentially important in physiological ageing where an activation of glucocorticoid secretion has been previously reported (van Cauter et al 1996, Harper et al 1999, Bergendahl et al 2000).

Finally, sleep may play an important role. It is well known that the sleep pattern changes with age. With increasing age, total sleep duration decreases and sleep fragmentation increases. In addition the relative time intervals in slow wave sleep are decreased as less deep sleep stages and REM sleep dominate (van Cauter et al 2000). When allowed to sleep only four hours every 24 h over one week, thereby mimicking the sleep pattern of elderly subjects, a profound effect on TSH secretion with very low circulating TSH levels at the end of the experiment could be demonstrated in young volunteers (Spiegel et al 1999). This indicates that sleep-dependent processes not directly related to ageing may change TSH secretion and the functional activity of the thyroid axis.

Most of the epidemiological studies on thyroid function in ageing are based on population studies including subjects where the above mentioned mechanisms may very well result in a significant study bias due to disease-related alterations or independent age-related problems (Mariotti et al 1995). Reliable studies focusing on healthy subjects are scarce. A careful evaluation of healthy centenarians selected according to criteria of the Eurage Senieur protocol (Mariotti et al 1993) revealed subtle changes of the thyroid axis. Free triiodothyronine and TSH serum concentrations were slightly decreased in parallel to a small but significant increase in reverse triiodothyronine serum concentrations. Free thyroxine serum levels were unchanged when comparing a group of 41 healthy centenarians with 98 healthy subjects aged 20-64 years or with 33 subjects aged 65-80 years, respectively. The changes observed were so subtle that they would not have been evident if only subjects up to 80 years of age had been included. These results are compatible with a decline in hypothalamic-pituitary activity during ageing resulting in a decreased TSH secretion. Studies on the 24 h rhythm of TSH in young and elderly subjects showed that the nocturnal TSH peak was blunted in the elderly (Barreca et al 1985, van Coevorden et al 1989, Greenspan et al 1991). Analysis of the pulsatile release pattern of thyrotropin revealed that this change selectively rests on a decreased TSH pulse amplitude whereas the temporal structure of the circadian TSH secretion with a nadir in the late afternoon and a peak around midnight in conjunction with an increased frequency of TSH pulses appears to be preserved (Brabant et al 1991). Thyrotropin response to a defined TRH challenge in the study group, male subjects from 67-84 years of age, was attenuated, pointing to a decreased pituitary TSH secretory reserve despite comparable triiodothyronine and thyroxine serum (van Coevorden et al 1989), which would be compatible with a hypothalamic or pituitary mechanism.

Animal studies in male Fisher rats suggest an age-related decrease in hypothalamic TRH synthesis followed by a decreased formation of thyrotropin b subunit within the pituitary gland (Cizza et al 1992). Studies in humans using the dopamine antagonist metoclopramide indicated a decreased responsiveness of pituitary TSH secretion to TRH (Targum et al 1989). This may be explained by a change in the nocturnal increase of TSH secretion due to an increased nightly dopamine tone (Greenspan et al 1991). As dopamine has direct inhibitory effects on pituitary TSH synthesis (Shupnik et al 1986), the attenuated response to TRH may fit into this concept. As an alternative or additional mechanism, an altered set point for the feedback action of thyroxine in elderly subjects was proposed, due to a putatively increased pituitary T4 to T3 conversion (Lewis et al 1991) and a selective decrease of free serum T3 levels (Mariotti et al 1993). Such age-related alterations in the activity of T4 to T3 conversion have not only been described for the activity of the 5' deiodinase in the pituitary, but also in the liver (Donda & Lemarchand-

Beraud 1989). On the contrary, TSH appears to stimulate deiodinase activity and may thus counteract these effects (Kohrle 1990).

In summary, changes of the hypothalamic—pituitary—thyroid axis with age seem to be subtle and suggest a decreased hypothalamic stimulation of thyrotropin release. In parallel, disease-specific alterations such as the development of thyroid autonomy, or changes in energy intake or sleep may dominate these small physiological changes and lead to pronounced alterations of thyroid function with age, which are only indirectly related to ageing of the hypothalamo— pituitary—thyroid axis itself.

0 0

Post a comment