Testosterone Treatment of Older Men—Why Are Controversies Created?

Cunningham and Toma have published an article entitled “Why Is Androgen Replacement in Males Controversial?” (1) in this issue of JCEM. This represents a new look at an old controversy. The article reviews the status of testosterone treatment in older men and serves as an extension to previously published guidelines on hypogonadism by Bhasin et al. (2) and recommendations published by Wang et al. (3) on late-onset hypogonadism. There are many causes for controversy in the practice of medicine, not the least of which is insufficiency of high-grade, evidence-based data. This is likely to occur when a medication was established in clinical use before the modern era of pharmaceutical regulation. This new review describes the insufficiency of evidence-based data to be certain that the benefit-to-risk ratio justifies testosterone treatment in older men, whereas taking the position that androgen replacement therapy in younger and middle-aged men is well established. The authors do acknowledge, as do the above guidelines (2), that relatively few large-scale, double-blind, placebo-controlled, multiple end point studies had been performed on testosterone treatment in men of any age. This position was also taken by the Institute of Medicine (4). Most of the data come from relatively small-size studies and meta-analyses of small- or medium-size databases (5,6,7,8).

Testosterone has been available as a pharmaceutical medication since the 1930s and has been used for many decades to treat symptoms of impaired libido and failure of secondary sexual development. With time and expanded understanding of androgen biology, it became apparent that testosterone had effects on a wide range of target organs. Deficiency of testosterone could produce a decrease in lean body mass, an increase in fat mass, a decrease in bone mineral density, impaired mood (preclinical depression), and decreased erythropoiesis, as well as a decrease in facial and body hair. Thus, the male sex hormone became a hormone of metabolic significance. Subsequent epidemiological studies also suggested that testosterone deficiency was associated with obesity and predisposes men to develop metabolic syndrome, type 2 diabetes mellitus (9,10,11), and increased mortality due to cardiovascular disease (12,13).

In the past two decades, it became apparent that testosterone levels peaked in men in the decade between ages 20 and 30 and decline progressively and slowly over many decades. The rate of decline has varied in different studies and populations from 1 to 2% per year (14,15) to less than 1% in others (15,16). This progressive decline results in increasing numbers of men in the older age groups with serum testosterone levels below the lower limits of normal for a reference population of younger men. Many men with low serum testosterone are relatively asymptomatic or their symptoms cannot be easily distinguished from other causes; thus, the prevalence of symptomatic testosterone deficiency in the older age groups is lower than that based on serum testosterone levels (16,17). This raises the question of whether or not serum testosterone concentration below the reference range for young men is a physiological phenomenon associated with aging or a pathological state. Regulatory agencies have questioned whether therapeutic intervention is a lifestyle drug or a treatment for a real medical disorder. This concern may have been encouraged by the development of many niche clinics and practices using multiple medications to reverse aging in men without clear-cut evidence of benefit.

One of the arguments presented by Cunningham and Toma is that the pathogenesis of low testosterone in older men is not well defined. This is certainly true because many men in the middle to older age group do not fit the simple definition of either primary or secondary hypogonadism but have a mixed type of testosterone deficiency with impairment of both testicular and hypothalamic pituitary signals (low serum testosterone with elevated, normal, or subnormal concentration of gonadotropins) (18). There are many acute and chronic diseases that will produce low serum testosterone levels that may normalize when the underlying comorbid condition is improved. Although the pathogenesis for the decline of testosterone with age is unknown, this is also true for the majority of men with hypogonadism in younger age groups.

Another important issue is that we are not certain what threshold should be used for the definition of testosterone deficiency in men of any age. The most frequent practice is to accept the reference range for a young, healthy male population to define the lower limits of normality. This is rational in that men in the young and middle-aged groups that fall below this level had previously been shown to benefit from testosterone interventional treatment. The assumption is that older men who fall below this reference range for younger men will also benefit from replacement testosterone treatment. This argument ignores the fact that we have limited data to assess relative benefit at specific serum testosterone concentrations in the borderline serum testosterone windows. This dilemma has been mitigated by recent data (17) and may be greatly resolved by data from studies in progress (http://clinicaltrials.gov/ct2/show/NCT00799617). Wu et al. (17) in the European Male Aging Study have attempted to define the levels of testosterone in an ambulatory population that are associated with symptoms including those seen in younger men with androgen deficiency. The study showed that different serum testosterone thresholds exist for individual clinical symptoms (17). This is consistent with previous reports by Bhasin et al. (19) suggesting that sexual symptoms may have a lower threshold for symptoms than some of the other manifestations such as reduction in lean body mass and increase in fat mass. Furthermore, the reference range for serum testosterone in a general population is wide (e.g. 280–1000 ng/dl), and the clinical manifestations of testosterone action are not directly correlated with serum testosterone concentrations within the normal range. Thus, for instance, men in a normal population at the higher end of the serum testosterone reference range have no greater libido than men with lower normal range testosterone concentrations.

The review explains that biologically active testosterone may not correlate with total testosterone due to differences in SHBG resulting in divergent levels of free or biologically active testosterone concentrations. The review and the guidelines recommend that total testosterone be used to define the threshold of normality and that free or biologically active testosterone measurements be reserved for patients whose total testosterone levels fall in the borderline range. This is a pragmatic decision that may be revised by use of calculated free testosterone concentrations or development of simplified but highly specific and sensitive new assays for free and biologically active testosterone with reliable reference ranges provided.

Lastly, with regard to the controversies about the importance of disease and diagnosis, we must ask how much of the controversy about the diagnosis and treatment of older men with symptomatic low testosterone state represents age bias. Should our considerations of testosterone treatment in symptomatic older men be different than our considerations of treatment for younger men? Do we think that older men who might benefit from testosterone treatment should be excluded just on the basis that they are older and that this is a normal phenomenon of aging, or do we think that they deserve therapeutic intervention to attempt to prevent the decrease in vitality, muscle strength, and mass that could be associated with frailty? Osteoporosis is clearly associated with low testosterone; the combination of frailty and increased susceptibility to fracture are important medical risks to older men and lead to increased morbidity and mortality. The benefits of testosterone treatment on increasing muscle mass and strength as well as increased bone mineral density seem to be clear, but whether these benefits are associated with prevention of fractures and reduction in fall rate is not known (4). It is less clear that interventional treatment with testosterone will reverse the susceptibility to these disorders or improve the conditions when they exist before treatment.

What is clear from the review of Cunningham and Toma (1) is that we need more data to determine whether testosterone treatment of older men with symptomatic low testosterone states will benefit such men. The authors have argued that we must await the results of ongoing large-scale interventional trials such as the National Institute of Aging, National Institutes of Health-supported Testosterone Trials (http://clinicaltrials.gov/ct2/show/NCT00799617).

The authors also ask whether older men are more susceptible to risks from testosterone intervention? Here the authors emphasize two areas of concern: 1) prostate (prostate cancer and benign prostatic hyperplasia), and 2) cardiovascular health. They point out that most of the epidemiological data suggest that low serum testosterone is a risk factor for cardiovascular disease (12,13), and that the interventional data are inadequate to determine risk because available studies are too short in duration and underpowered to determine whether there will be a benefit, a neutral effect, or an increased risk from testosterone interventional treatment. They do point out that available meta-analyses suggest that testosterone intervention is not a risk factor for cardiovascular complications (20), but a recent report from a study (21) in older frail men treated with relatively high doses of testosterone suggested that there may be reasons for concern about cardiovascular risk, whereas another study (22) showed no increase when the doses of testosterone were lower. Thus, cardiovascular disease risk and testosterone treatment in older subjects require a longer, large-scale, adequately powered study (23). The authors also reviewed the issue of prostate cancer. The data at present do not support the concept that testosterone is a risk factor (24) for development of prostate cancer, but Cunningham and Toma correctly argue that inadequate data are available to determine whether long-term testosterone treatment will increase the risk of either benign or malignant prostate disease. It should also be noted that there is no particular reason to believe that long-term treatment with testosterone in younger men with chemical testosterone deficiency with or without symptoms represents a greater or lesser long-term risk in terms of adverse effects on the prostate or the cardiovascular system than treatment beginning in older men.

Cunningham and Toma stated in their review that transdermal testosterone preparations are more difficult than injectables to titrate the dose and to ensure compliance. The issues of adherence are certainly true if one assumes that the injections are administered and documented in a doctor’s office. The ability to titrate dose, however, is quite possible with transdermal preparations, particularly when a metered-dose dispenser is used.

Regarding adolescent treatment, the authors indicate that treatment with human chorionic gonadotropin (hCG) may be of psychological benefit in that it increases testis size and could be a benefit to future fertility. There are very limited data that teenagers check the size of their testes and have psychological benefit from a larger size testis rather than a smaller size testis. Furthermore, there is limited evidence that early treatment of hCG in adolescence will benefit future fertility to a greater degree than reserving treatment with gonadotropins to the time when fertility is desired. They also state that testosterone treatment is not appropriate for men with hypogonadotropic hypogonadism and infertility; this is certainly true if fertility is an immediate need at that time. On the other hand, the majority of men with hypogonadotropic hypogonadism benefit by testosterone replacement therapy at the time of their diagnosis and only require the much more expensive and less friendly hCG and FSH injections at the time when fertility is required. It is our experience that testosterone interventional replacement therapy does not preclude future benefit from hCG/FSH treatment for fertility.

In the end, testosterone treatment in men of all age groups should be used judiciously. More data are needed to determine which patients will benefit from treatment and for which symptoms. Long-term studies are necessary to determine whether treatment with testosterone in young, middle-aged, or older men will enhance, decrease, or have no effect on cardiovascular and prostate disorders.