The Modern Era of Testosterone Therapy

Before the early 1990s, the use of testosterone therapy was rare, and was limited almost exclusively to younger men with severe cases of testosterone deficiency (TD) owing to pituitary tumors, anorchia, or genetic abnormalities such as Klinefelter syndrome. Over the past 20 years there has been a remarkable and steady growth in the use of testosterone therapy. This has occurred as a result of increased physician awareness of TD and the benefits of treatment, together with increased convenience of testosterone formulations. Notable benefits include improved sexual desire and performance, improved energy, increased muscle and bone density, and improved metabolic status, similar to benefits reported at the advent of testosterone use in the 1940s. The reinvigorated interest in testosterone therapy has led to a reexamination of traditional assumptions concerning prostate cancer and testosterone. Despite important advances in our understanding of this topic, the use of testosterone therapy continues to be controversial because of prostate cancer fears, and this remains the greatest concern among physicians around the world with regard to the use of testosterone therapy.

The Androgen Hypothesis

Stemming from reports in the 1940s, the androgen hypothesis has come to include the following features: prostate cancer is an androgen-dependent cancer, high testosterone levels contribute to the development of prostate cancer, high testosterone causes rapid growth of prostate cancer, and low testosterone is protective against development of prostate cancer and causes prostate cancer to regress. Ever since, medical students and physicians have been taught that high testosterone promotes prostate cancer development and there seemed to be no reason to doubt this axiomatic concept. The relationship between testosterone and prostate cancer was classified as “fuel for a fire” or “food for a hungry tumor”. In an international survey published in 2007, as many as 70% of health care providers were concerned about the association of testosterone therapy and prostate cancer. It was not until recently that this conventional wisdom was challenged.

The breakdown of the androgen hypothesis evolved throughout the years, beginning in the early 1990s. Morgentaler and colleagues published a study in which testosterone deficient men with normal prostate-specific antigen (PSA; <4.0 ng/mL) and a normal digital rectal examination underwent a sextant prostate biopsy before initiating testosterone therapy. Interestingly, 11 of the first 77 men had prostate cancer. Compared with the 15.2% prostate cancer rate noted by Thompson and colleagues in the placebo arm of the Prostate Cancer Prevention Trial, this 14.3% rate was shockingly similar. This was the first indication that low testosterone may be a risk factor for prostate cancer, and not protective against prostate cancer and its progression.

Since this revelation, more than 20 population-based longitudinal studies have shown no relationship between prostate cancer and serum testosterone or other androgens. The Endogenous Hormones and Prostate Cancer Collaborative Group published high-level evidence from a metaanalysis consisting of 18 studies that included 3886 men with incident prostate cancer and 6438 controls. The results demonstrated no direct association between endogenous serum androgens and the development of prostate cancer. Additionally, Muller and colleagues analyzed 3255 men in the placebo arm of the reduction by Dutasteride of Prostate Cancer Events trial. Men underwent prostate biopsies at 2 years and 4 years and there was no relationship found between testosterone or dihydrotestosterone levels and prostate cancer risk.

Although high testosterone levels were thought to contribute to the development of prostate cancer, there is a complete lack of compelling evidence in the literature. An extensive review found that men with higher endogenous testosterone or who had undergone testosterone therapy were not at increased risk of prostate cancer. Supraphysiologic doses of testosterone for up to 9 months in healthy men failed to demonstrate a significant increase in PSA or prostate volume. The notion that “more testosterone is bad, less testosterone is good” was not necessarily true.

The Saturation Model

However, physicians still recognize that initiation of androgen deprivation causes rapid declines in PSA and that cessation of androgen deprivation causes rapid increases in PSA. Revisiting the landmark work of Huggins and Hodges, the traditional view suggests a continuous relationship between serum testosterone and prostate cancer growth. Studies from Prout and Brewer and Fowler and Whitmore present an alternative possibility. Both papers noted no evidence of progression in men with prostate cancer not treated by androgen deprivation or castration. The evidence presents a paradox: how can prostate cancer be so sensitive to androgen deprivation, yet seem to be indifferent to variations in serum androgens under other circumstances?

The resolution of this paradox is the saturation model. The saturation model accounts for the key observation that prostate tissue is exquisitely sensitive to changes in serum testosterone at low concentrations, but becomes indifferent to these changes at higher concentrations. Exposure to increasing concentrations of androgen causes prostate tissue growth, but this growth rate plateaus when the concentration reaches a limit. There exists a threshold (saturation point) beyond which there is no further ability to induce androgen-driven changes in prostate tissue growth. Similar relationships are found throughout biology. This explains why dramatic PSA changes are noted when serum testosterone is manipulated into or out of the castration range, but minimal PSA changes occur when supraphysiologic testosterone doses are administered to normal men.

At least 2 mechanisms underlie the model. The first is the finite ability of the androgen receptor to bind androgen. Maximal binding of the androgen to the androgen receptor (saturation) in human prostate is achieved in vitro at approximately 4 nmol-1. In vivo, the saturation point is approximately double this value, at 8nmol/L or approximately 250 ng/dl. Because the primary actions of androgen on prostate tissue occur via binding to androgen receptor, once the androgen receptors are saturated the presence of higher androgen concentrations should not elicit further biochemical response. A second mechanism is that intraprostatic androgen concentrations seem to be somewhat independent of serum concentrations. Marks and colleagues demonstrated that intraprostatic concentrations of testosterone and dihydrotestosterone were unchanged after 6 months of testosterone injections, despite large increases in serum testosterone concentrations. This raises the possibility that the prostate maintains a relatively homeostatic microenvironment with regards to androgens, relatively unaffected by changes in serum androgens.

This saturation model derails the notion that testosterone is “fuel for a fire” or “food for a hungry tumor.” An analogy that better fits the available evidence is “testosterone is like water for a thirsty tumor.” Once that thirst has been quenched by adequate testosterone concentrations, additional androgens serve as nothing more than excess.

The presentation of the saturation model and the shift in concepts regarding testosterone and prostate cancer have important clinical implications. Approximately 20% to 30% of elderly men over the age of 60 experience TD. Symptoms of TD vary, but include fatigue, weakness, decreased libido and energy, erectile dysfunction, reduced muscle and bone mass, and increased abdominal fat. Testosterone therapy is an effective, commonly used treatment shown to be effective in mitigating the bothersome symptoms of TD.

As physicians gain a better understanding of TD and its consequences, there has been a reevaluation of the risks of testosterone therapy, particularly with regard to prostate cancer. Some clinicians still fear that testosterone therapy may unmask occult prostate cancer in otherwise healthy men with TD, but more evidence is mounting in favor of the benefits of testosterone therapy. Whereas prior history of prostate cancer was considered an absolute contraindication to the use of testosterone therapy, physicians are recognizing the benefits of testosterone treatment in certain populations.

Although no randomized, controlled trials have been performed to assess testosterone therapy and prostate cancer risk, evidence to date fails to suggest increased risk. Calof and colleagues conducted a metaanalysis of 19 placebo-controlled testosterone therapy trials and found no significant increase in prostate cancer or development of PSA levels greater than 4.0 ng/ml in men treated with testosterone therapy versus placebo. A systematic review of 11 placebo-controlled studies by Shabsigh and colleagues showed that men who received testosterone therapy had neither increased risk of prostate cancer nor greater Gleason grade among those who developed prostate cancer. In the UK Androgen study, Feneley and Carruthers followed 1365 British men 28 to 87 years of age and found the risk of prostate cancer diagnosis to be similar to age-matched controls. Only 1 of 20 patients with high-grade prostatic intraepithelial neoplasia developed prostate cancer during 12 months of testosterone therapy. Combined with the studies cited previously reporting no increased risk of prostate cancer based on endogenous androgen concentrations, these data can reliably ease the fear that higher testosterone concentrations increase the risk of prostate cancer.

Association of Low Serum Testosterone with Prostate Cancer

Clouded by the long-held fear that high serum testosterone is a risk for prostate cancer, there has been little appreciation for a literature that strongly shows a relationship between low serum testosterone concentrations and high-risk prostate cancer. Men with the lowest tertile of serum testosterone had nearly double the risk of being diagnosed with prostate cancer on biopsy compared with men with less severe TD, in a study of 345 men with TD and PSA levels of less than 4.0 ng/mL. Among men with prostate cancer, a high Gleason score has been reported with lower serum testosterone concentrations. In an open clinical study conducted by Mearini and colleagues, 37% of the 65 patients with prostate cancer had testosterone levels of less than 2.5 ng/mL. One study from García-Cruz and colleagues including 137 men undergoing biopsy for suspicion of prostate cancer found an inverse relationship between serum testosterone and prostate cancer. Salonia and colleagues measured serum testosterone on the day before radical prostatectomy (RP), and found that the risk of seminal vesicle invasion, a markedly poor prognostic indicator, was increased significantly in men with low testosterone levels, including a 3-fold increased risk in men with severely reduced serum testosterone.

Low levels of serum testosterone have also been reported in association with poor prognosis from prostate cancer. Yamamoto and colleagues reported increased rates of biochemical recurrence after RP in patients with low testosterone, and low levels of free testosterone were reported as an independent prognostic factor for prostate cancer progression in a study of 154 men undergoing active surveillance for prostate cancer.

These studies are consistent with the epidemiology of prostate cancer, in that there is increased prevalence of high-grade prostate cancer as men age (Thompson and colleagues ) and testosterone levels decline.

Testosterone Therapy in Men with a History of Prostate Cancer

Testosterone therapy for men with a history of prostate cancer remains controversial, particularly because prostate cancer has been considered a contraindication for testosterone therapy for several decades, and testosterone product labels carry this warning as well. However, with the emergence of the saturation model and evidence mounting against the traditional androgen hypothesis, the paradigm has shifted. Multiple investigators have now reported on the use of testosterone therapy in men with prostate cancer, with published case series beginning in the mid-2000s. Overall, these studies have provided reassuring results regarding the risk of testosterone therapy in men with prostate cancer. In a study by Kaplan and Hu, Surveillance, Epidemiology, and End Results data were linked to Medicare data to identify 149,354 men diagnosed with prostate cancer between 1992 and 2007. Of these, 1181 (0.79%) received testosterone therapy after diagnosis. No differences were reported in overall survival, cancer-specific survival, or in the use of salvage androgen deprivation therapy in men with or without use of testosterone after diagnosis.

In 2004, Kaufman and Graydon reported no prostate cancer recurrences in 7 men treated with testosterone therapy after RP. Median follow-up in the study was 2 years, with the longest follow-up of 12 years. Agarwal and Oefelein reported no recurrences in 10 testosterone deficient men with history of RP who received testosterone therapy for up to 19 months. In a larger case series of 57 men with predominantly low and intermediate risk prostate cancer, Khera and colleagues reported no biochemical recurrences after a median follow-up time of 13 months. Men were treated with testosterone therapy for an average of 36 months after RP. In the largest series to date, Pastuszak and colleagues examined the records of 103 hypogonadal men after RP. In this study there were also 49 eugonadal men who underwent RP but did not receive testosterone therapy, included as a comparison group. Approximately one-quarter of these men in each group were considered high risk based on Gleason score 8 to 10, positive margins, or the presence of positive lymph nodes at surgery. With a median follow-up of 27 months, biochemical recurrence rates were 4% in the testosterone-treated group and 16% in the untreated, eugonadal group.

In 2007, Sarosdy published a case series evaluating 31 men with TD receiving testosterone therapy after brachytherapy treatment for prostate cancer. The median time for initiation of testosterone therapy was 2 years after treatment and median follow-up was 5 years (range, 1.5–9). There was no evidence of biochemical recurrence and none of the men halted testosterone therapy owing to prostate cancer recurrence. Morales and colleagues reported a prospective case series consisting of 5 men that had been previously treated with external beam radiation therapy for prostate cancer. After reaching nadir PSA levels, these men received testosterone therapy. Median follow-up was 14.5 months (range, 6–27). The PSA level did increase transiently in 1 patient, but none exceeded 1.5 ng/mL to raise any concern of biochemical recurrence.

A small number of reports also describe testosterone therapy in men who underwent radiation therapy for prostate cancer. In 2013, Pastuszak and colleagues evaluated 13 men with low and intermediate risk prostate cancer treated with testosterone therapy after radiation (brachytherapy or external beam radiation therapy). After median follow-up of 29.7 months, no biochemical recurrences were reported. More recently, Balbontin and colleagues reported no biochemical recurrence in a case series of 20 men with low-risk prostate cancer treated with brachytherapy. Testosterone treatment lasted for a median of 14 months and median follow-up was 31 months (range, 12–48). Notably, PSA level declined from 0.7 to 0.1 ng/mL. More recently, Pastuszak and colleagues also present a multiinstitutional cohort of 98 men who received testosterone therapy after radiation treatment for prostate cancer. Of these men, 76.6% had low or intermediate risk prostate cancer and the median follow-up was 40.8 months. A small but statistically significant increase in PSA levels was noted with testosterone therapy, from 0.8 to 0.9 mg/mL, and 6 men (6.1%) experienced biochemical recurrence. This recurrence rate is lower than previously reported rates of biochemical recurrence after radiation therapy. However, one must be cautious in drawing conclusions owing to the limited sample size of this and other studies, their retrospective study design, and single-arm methodology.

In addition, there are now several reports of testosterone therapy in men undergoing active surveillance for prostate cancer. Morgentaler and colleagues performed a retrospective study of 13 hypogonadal men receiving testosterone therapy for at least 12 months while undergoing active surveillance for prostate cancer. Median follow-up was 2.5 years (range, 1–8.1). At initial biopsy, 12 men had low-risk prostate cancer and 1 man had intermediate risk prostate cancer (Gleason 3 + 4). No prostate cancer progression was noted, and no cancer was found in 54% of biopsies. Recently, Kacker and colleagues retrospectively reviewed a larger series of men on active surveillance, comparing progression rates in 28 men with TD who underwent testosterone therapy with 96 men with TD who did not receive testosterone therapy. Median follow-up was 38.9 months and 42.4 months for the testosterone group and the no-testosterone groups, respectively. Progression rates were similar between the 2 groups.

A cautionary note was struck by Morales, who reported erratic PSA responses to testosterone therapy in 6 men with untreated prostate cancer and one man without documented prostate cancer. However, no follow-up biopsy results were reported, making it difficult to interpret these results.

Discussion

Given the proven benefits of testosterone therapy for testosterone-deficient men, clinicians are faced with a dilemma. Large numbers of men around the world, including younger men aged 40 to 50 years, have been treated and cured of prostate cancer. With some of these men having long life expectancies, is it reasonable to deprive these men of a treatment that may provide important benefits and enhanced quality of life based on a historical concern that does not seem to be supported by current scientific evidence?

When evidence contradicts theory, it is useful to try to understand how the theory came into being. From the early 1940s when Huggins first reported that castration lowered serum acid phosphatase in men and that testosterone injections increased it, there was extremely limited clinical experience with testosterone therapy. This situation changed with introduction of the first branded topical testosterone products in the late 1990s and early 2000s. Before that time, the only experience physicians had with manipulation of testosterone levels, particularly urologists, was to lower serum testosterone into the castrate range as treatment for prostate cancer. This effectively lowered PSA. There was no reason, therefore, to question the axiom that high testosterone contributed to prostate cancer development and growth, and that low testosterone protected against it. It was therefore stunning to discover in 2006 that the original conclusion by Huggins—that testosterone injections “activated” prostate cancer—was based on only a single hormonally intact patient. The reliable, albeit temporary, results observed by all urologists with lowering androgen concentrations, and the lack of experience with raising testosterone contributed to the belief that higher testosterone concentrations were risky.

Today, it must be recognized that there exists no evidence that testosterone therapy increases prostate cancer risk in testosterone-deficient men. Although the fear of aggravating prostate cancer with testosterone therapy among physicians is understandable given our training, it should be clear that the theoretic underpinning of this concept has been shown to be unsound, and there are now numerous clinical experiences and publications that demonstrate that the risk of worsening prostate cancer outcomes with testosterone therapy seems to be small, if present at all.

The leading controversy today is whether testosterone therapy can be safely offered to men with a history of prostate cancer. Recommendations for this have been provided in a recent review by a group of experts in the field. Our own recommendations are similar. Candidates should be those with symptomatic TD who stand to benefit from treatment. Informed consent should be provided to patients, specifically including information that safety data are limited, no controlled prospective studies have yet been performed, and there is thus an unknown degree of risk of cancer progression or recurrence with testosterone therapy. We require a signed consent form stating these points in our practice. In addition, patients should be informed of the standard risks associated with testosterone therapy, including acne, erythrocytosis, fluid retention, peripheral edema, infertility, testicular atrophy, and gynecomastia.

It is useful to remember that prostate cancer recurrence occurs at a significant rate in men treated for prostate cancer, even if no testosterone therapy is offered. This baseline risk of approximately 15% must be considered when offering testosterone therapy. This means that some men will experience prostate cancer recurrence or progression whether or not they receive testosterone therapy. However, there will be a reflex presumption on the part of patients and other physicians that it was the increase in serum androgens that triggered the recurrence or progression. Adequate counseling of patients before beginning testosterone therapy is of great help in countering this eventuality, and we repeat this conversation regularly during the course of treatment.

Among men with a history of prostate cancer, the safest candidates for testosterone therapy are those with low-risk disease who have an undetectable PSA after RP. A relatively low-risk group also includes those with excellent PSA responses to radiation therapy. Until more data are available, testosterone therapy in men on active surveillance should be undertaken only with great caution. And testosterone therapy should be avoided in men with advanced or metastatic disease, and in men currently treated with androgen deprivation.

Another point to consider is that any man, with or without a history of prostate cancer, is likely to experience an increase in PSA if their baseline serum testosterone is below the saturation point. This varies from one individual to another, but the saturation point seems to be approximately 250 ng/dL. Thus, if a man begins treatment with a serum testosterone of 150 ng/dL it can be expected that his PSA will increase for the first 3 to 6 months, whereas this is less likely in a man with a baseline serum testosterone of 290 ng/dL.

For all men receiving testosterone therapy with treated or untreated prostate cancer, follow-up should be rigorous, especially in the first year of treatment. Physicians must be prepared for prostate cancer recurrence and the possibility that it may be interpreted as a result of testosterone therapy. Hematocrit and hemoglobin should be measured 2 to 4 times in the first year, then annually. PSA levels should be measured every 3 to 4 months in the first year, then biannually. Digital rectal examination should be performed 1 to 2 times in the first year, then annually. It should also be noted that for those men undergoing active surveillance, an annual prostate biopsy should be performed to ensure cancer stability. After stability has been determined, longer intervals may be suitable.