Hormone Replacement in the Elderly: Facts and Fantasy

Ashok Balasubramanyam, MD


August 09, 2005

Editorial Collaboration

Medscape &


The rapid increase in the number of older people is a marked trend all over the world. In the United States, this demographic shift has been dramatic: In 1900 there were 3 million persons over the age of 65 years, comprising 4% of the population; in 2000 there were 35 million persons over 65 years, comprising 13% of the population.

Organ dysfunction as well as disease and disability increase with older age, and are often associated with changes in the circulating levels of a number of hormones, including sex hormones, growth hormone (GH), and thyroid hormones. But are these hormonal changes a "natural" consequence of aging, or do they affect only a portion of the aging population? How can true deficiency be measured in the face of an overall age-related decline in hormone levels and actions? What are the mechanisms of the decline? Finally, when and how should these hormones be replaced in the elderly, and what are the beneficial and deleterious effects of such replacement? These questions were addressed in a symposium at The Endocrine Society's 87th Annual Meeting.

Testosterone Replacement in Elderly Men

Alvin Matsumoto, MD,[1] of the University of Washington, Seattle, discussed important issues surrounding the use of testosterone replacement therapy in elderly men. Some of these issues are quite basic: How should we interpret low testosterone levels in the elderly, and, because the levels of bioavailable testosterone are regulated in a number of ways in both the circulation and at target tissues, how do we account for the possibility that blood testosterone levels may not reflect levels in target tissues, or alterations in the sensitivity of those tissues to the hormone?

The Baltimore Longitudinal Study of Aging showed that there is a gradual decrease in blood testosterone levels with age, likely due to a greater decline in testosterone production than in testosterone clearance. The same is true of the weaker adrenal androgens. In contrast, the Massachusetts Male Aging Study showed that blood levels of the potent androgen dihydrotestosterone (which is derived from testosterone by the action of the enzyme 5-alpha reductase) don't decline with age, or actually increase slightly. Somehow, these alterations have to be correlated with their putative physiologic consequences as they occur with aging, ie, decreases in muscle mass, bone mass, and sexual function, and an increase in fat mass.

A good example of the complex effects of these multiple changes in the regulation of androgens is the aging prostate gland. Why does the prostate enlarge contemporaneously with a decline in circulating testosterone levels? Could there be a discrepancy between blood levels and tissue levels of androgens? An abstract presented by Dr. Stephanie Page and colleagues,[2] in Dr. Matsumoto's group, carried out a bold experiment to address this question. They measured levels of testosterone and dihydrotestosterone in both blood and simultaneously obtained prostate biopsy specimens in groups of men treated with placebo, acycline (an inhibitor of 5-alpha reductase), or acycline plus testosterone.

The evidence is fairly clear that testosterone replacement has numerous benefits for young men with hypogonadism. However, there are no long-term, controlled studies that establish guidelines for treatment, or clearly demonstrate beneficial outcomes of such treatment in elderly men. The clinician faced with an elderly man with low testosterone levels and a number of physiologic changes that may be related to reduced androgenic effects must decide whether it is beneficial to treat this person with testosterone replacement. On what should the clinician base the decision? The operational definition of male hypogonadism (≥ 3 signs or symptoms; total serum testosterone < 200 ng/dL, or free testosterone < 8.9 ng/dL) is based on a large cohort of men in their 20s.[3] Can this apply to the elderly? In the elderly, how does one distinguish a pathologically low testosterone level or a low androgen state that would benefit from therapy from an age-related low androgen state in which androgen replacement may not be of any help or may even result in morbidity?

The challenges of defining androgen deficiency in the elderly are numerous. First, there are numerous pitfalls in regard to interpreting blood androgen levels, because (1) normal ranges of total testosterone are generally established for young populations; (2) there may be transient decreases in blood testosterone levels in association with intercurrent illnesses or medications, which generally affect the total testosterone level by altering synthesis of sex hormone-binding globulin (SHBG); and (3) measurement of luteinizing hormone and follicle-stimulating hormone is frequently unhelpful, because hypogonadism in the elderly male often has both primary and secondary components. Second, the signs and symptoms of androgen deficiency in the elderly are often nonspecific and could easily be confused with those associated with depression or a variety of comorbid illnesses or medications. Finally, there is the thorny, hard-to-quantify factor of differences in tissue sensitivity to circulating androgens. These issues leave open the key question: What is the threshold blood hormone level, and what are the specific signs and symptoms that are useful to define treatable hypogonadism in the elderly? Treatment options are further affected by whether one can expect linear dose-response effects in the key tissues. This is by no means a simple problem because excellent data from the group of Shalender Bhasin, MD, at the University of California, Los Angeles, show that the physiologic responses extend to doses well beyond the normal range of blood testosterone levels.

All this raises more questions than answers, and carefully conducted studies are needed to address these questions rigorously and practically. In the meantime, the physician needs to measure blood testosterone levels with reliable assays (or measure free testosterone levels when changes in SHBG are suspected), check testosterone levels in the basal "healthy" state, and exercise good clinical judgment to decide whether to treat an elderly man for possible hypogonadism.

Thryoid Dysfunction in the Elderly

Mark Vanderpump, MD, FRCP,[4] of the Royal Free Hospital, London, United Kingdom, discussed thyroid dysfunction in the elderly. Diagnosing and treating hypothyroidism is fraught with a different set of considerations than diagnosing and treating hypogonadism. Signs and symptoms of hypothyroidism overlap with those of the physiologic and behavioral changes associated with aging, but there is no evidence that hypothyroidism is an inevitable consequence of aging.

The actual prevalence of hypothyroidism in the elderly is controversial. Dr. Vanderpump reviewed data from a large English cohort, the Whickham study, which in 1977 reported that the frequencies of goiter, elevated thyroid-stimulating hormone (TSH) levels, and antithyroid antibodies increased with age. However, in the United States, the Third National Health and Nutrition Examination Survey (NHANES III) data indicate that TSH levels do not change much with age in healthy adults, although there is an increase in the prevalence of subclinical hypo- or hyperthyroidism in the later decades of life. More detailed testing reveals that advancing age is generally not associated with changes in total thyroxine levels, either basally or in response to TSH. However, the rates of both secretion and clearance of thyroxine decline, and this could affect the dosage of exogenous thyroxine therapy in hypothyroid patients as they advance in age.

The prevalence of true, overt thyroid disease, as revealed in a number of studies, does tend to increase with age, reaching a frequency of 1% to 2% for hypothyroidism and .2% to -.7% for hyperthyroidism. The manifestations of thyroid dysfunction in the elderly vary from those in younger adults. Hypothyroidism presents often with psychological changes or dementia, and hyperthyroidism with fewer sympathetic nervous system signs and with more neurologic signs, weight loss, and anorexia than in younger adults. The rates of hyperthyroidism rise more sharply in the elderly who live in regions of relative iodine deficiency. It is also important to remember that approximately 15% of new-onset atrial fibrillation in the elderly is due to thyrotoxicosis.

An important issue in the elderly is the high rate of subclinical thyroid disease, defined as the presence of suppressed TSH or elevated TSH with thyroxine levels in the normal range. The prevalence is 4% to 10% for subclinical hyperthyroidism and 1% to 2% for subclinical hypothyroidism. Are these conditions associated with end-organ effects or morbidity; do they presage an inevitable slide into overt disease; and should they be treated? In regard to subclinical hypothyroidism, the Whickham study revealed that elderly patients with this condition had no significant morbidity beyond a mild dyslipidemia and some echocardiographic evidence of a mild decrease in left ventricular systolic function. The data leave open the question of whether low doses of thyroxine to restore TSH levels to normal would improve these mild changes.

A meta-analysis of interventions in subclinical thyroid disease evaluated the quality of the evidence for treatment benefit, and found little overall advantage.[5] However, the rate of progression of subclinical hypothyroidism to overt hypothyroidism in the elderly is high, and correlates directly with the initial level of TSH and the presence of thyroid antibodies, so clinicians should at least maintain close follow-up of elderly patients with subclinical hypothyroidism.

In contrast, subclinical hyperthyroidism is associated with inimical effects on end-organ function. For instance, it is associated with an increase in the risk of developing atrial fibrillation, left ventricular systolic changes and mortality from cardiovascular disease, and increased rates of bone loss and dementia. There is also evidence that the cardiovascular effects can be reversed by treatment. However, the meta-analysis concluded that there is insufficient evidence for routine antithyroid treatment of subclinical hyperthyroidism.

GH Treatment in the Elderly

Perhaps the most controversial of all issues in hormone replacement in elderly people is GH treatment. Dr. Andrew Toogood,[6] of the University of Birmingham, Birmingham, United Kingdom, discussed this contentious issue. Numerous studies show that GH secretion is markedly impaired with advancing age, declining at a steady rate during each decade of adult life. Concomitant physiologic changes include body composition changes (increased fat mass and decreased lean mass, bone mass, and exercise tolerance) and an elevated cardiovascular risk profile. But can we say that these changes are due to a true pathologic state of GH deficiency?

Dr. Toogood's group attempted to answer this question by investigating the hypothesis that if elderly persons have a state of true GH deficiency, then overt diseases of the hypothalamus or pituitary that affect the GH axis should not affect the functional status of their GH axis very significantly. They compared 24 elderly patients with known hypothalamic/pituitary disease to 24 age- and body mass index-matched controls for GH profile, body composition, and biochemical parameters. Those with hypothalamic/pituitary disease had significantly lower mean basal and stimulated GH levels and insulin-like growth factor 1 (IGF-1) levels. However, there was a great deal of overlap in GH/IGF-1 levels between patients with disease and controls. There was an increase in fat mass in the disease group, but no difference compared with controls in lean mass or bone mineral density. These and other data suggest that there is some rationale for GH replacement in elderly patients with definable hypothalamic-pituitary pathology and GH deficiency, but not as yet in the elderly who have signs or symptoms of GH deficiency but do not have biochemical evidence of GH deficiency. Thus, the Growth Hormone Research Society, based on a consensus conference in 1988, has indicated that 2 criteria should be met to define GH deficiency: (1) a history of hypothalamic/pituitary disease and (2) peak serum GH level < 3 mcg/L during an insulin tolerance test.

These sober and thoughtful guidelines fly in the face of an enormous marketing drive to place before the public the notion that GH treatment is a universal panacea for an assortment of aging ills. Is there any evidence of benefit to justify treating apparently "normal" elderly persons with GH? Dr. Toogood reviewed a large number of studies that have addressed this question, and concluded that although GH does reduce total fat mass and increase total lean mass in these persons, the results in relation to bone mass and muscle mass are disappointing. Furthermore, there is a high rate of adverse events, such as carpal tunnel syndrome, gynecomastia, arthralgia, and peripheral edema, and increased levels of IGF-1 could aggravate prostate cancer or breast cancer. An important caveat is that these studies were conducted in healthy elderly subjects, so one may ask whether GH therapy may be of benefit to frail elderly persons. Further studies are needed to address this issue.

On a final note, the comments of an ethicist, Ann Danoff, MD, of New York University, New York, NY, offered an important social perspective on hormone replacement in the healthy elderly. There is a great proliferation in this age of direct, Internet-based (mis)information dissemination of hormonal "youth elixirs" with greatly exaggerated claims. Dr. Danoff urged the medical community to keep well informed about good science in the field, and to always remember the principles of respect and justice in the treatment of the vulnerable elderly with incompletely proven or dubious hormonal treatments. It will be important to use both emerging evidence and sound ethical sense to consider always whether we are treating a true deficiency, with reasonable hopes of improving health, or just "medicalizing" a natural process of senescence.

  1. Matsumoto AM. Testosterone (T) levels in the aging male: challenges in the diagnosis of androgen deficiency and actions of T on target tissues in older men. Program and abstracts of The Endocrine Society's 87th Annual Meeting; June 4-7, 2005; San, Diego, California. Symposium 6-1.

  2. Page ST, Lin DL, Amory JK, et al. Prostate tissue dihydrotestosterone but not testosterone levels are decreased by medical castration in normal, middle-aged men. Program and abstracts of The Endocrine Society's 87th Annual Meeting; June 4-7, 2005; San, Diego, California. OR57-6.

  3. Araujo AB, O'Donnell AB, Brambilla DJ, et al. Prevalence and incidence of androgen deficiency in middle-aged and older men: estimates from the Massachusetts Male Aging Study. J Clin Endocrinol Metab. 2004;89:5920-5926.

  4. Vanderpump M. Thyroid dysfunction in the elderly. Program and abstracts of The Endocrine Society's 87th Annual Meeting; June 4-7, 2005; San, Diego, California. Symposium 6-2.

  5. Surks MI, Ortiz E, Daniels GH, et al. Subclinical thyroid disease: scientific review and guidelines for diagnosis and management. JAMA. 2004;291:228-238.

  6. Toogood AA. Growth hormone: to treat or not? Program and abstracts of The Endocrine Society's 87th Annual Meeting; June 4-7, 2005; San, Diego, California. Symposium 6-3.