Prostate-specific antigen changes and prostate cancer in hypogonadal men treated with testosterone replacement therapy

Authors


Matt Coward, 2113 Physicians Office Bldg CB♯7235, 170 Manning Dr, Chapel Hill, NC 27599-7235, USA.
e-mail: rcoward@unch.unc.edu

Abstract

OBJECTIVES

To retrospectively review hypogonadal men receiving testosterone replacement therapy (TRT), and evaluate the changes in prostate-specific antigen (PSA) levels over an extended period, and thus evaluate the occurrence of prostate cancer, as a primary concern in treating late-onset hypogonadism (LOH) is the potential increased risk of prostate cancer; we also recorded the cardiovascular effects of TRT.

PATIENTS AND METHODS

In all, 81 hypogonadal men (mean age 56.8 years) were followed for a mean (range) of 33.8 (6–144) months after starting TRT. All men had a normal baseline PSA level before TRT and had routine laboratory investigations, including measurements of body mass index (BMI), haematocrit, lipid profile, and liver function tests (LFTs). Testosterone and PSA levels were assessed every 6–12 months. Patients with a biopsy-confirmed or recent history of prostatitis before treatment were excluded. TRT was discontinued in men who developed prostate cancer.

RESULTS

Before and 36 months after treatment the total testosterone levels were 241.1 and 379.8 ng/dL (P < 0.05), respectively. Four men (4.9%) developed prostate cancer at a mean (range) of 32.5 (22–41) months after starting TRT. In men without prostate cancer (95.1%), PSA levels did not increase significantly at 1-year intervals for 5 years. There was no statistical difference in PSA level change from baseline to 36 months when patients without prostate cancer were stratified into groups according to age (≤50, 55–65 and ≥70 years). In men with prostate cancer there was an increase in mean PSA level from baseline to 18 months of 1.8 ng/mL, and to 36 months of 3.2 ng/mL (P < 0.05). Total cholesterol improved from 203.8 to 166.6 mg/dL (P < 0.05) after 36 months of TRT; the BMI, haematocrit and LFTs did not change significantly.

CONCLUSIONS

LOH is an increasingly prevalent disease characterized by a symptomatically low testosterone level, and TRT is effective in normalizing serum testosterone levels, providing a beneficial cardiovascular effect, and improving sexual function and overall quality of life. PSA levels remain stable after normalization of testosterone for ≥5 years, prostate cancer can be effectively diagnosed and treated in men taking TRT, and the incidence of prostate cancer among men with LOH on TRT is no greater than that in the general population.

Abbreviations
LOH

late-onset hypogonadism

TRT

testosterone-replacement therapy

(L)(H)DL

(low-) (high-) density lipoprotein

BMI

body mass index

LFTs

liver function tests.

INTRODUCTION

Late-onset hypogonadism (LOH), or androgen decline in the ageing male, is a syndrome caused by the well-known decline in gonadal production of androgens in men that occurs with ageing. It is characterized by clinical symptoms that accompany low serum androgen levels. LOH affects a wide range of organ systems and has a broad severity of clinical presentations. Sexual dysfunction, including erectile dysfunction, decreased libido, difficulty achieving orgasm, and decreased penile sensation, is usually the presenting complaint, although fatigue, depressed mood, impaired cognition and decreased muscle mass are other common symptoms [1]. If these symptoms are found in conjunction with low serum testosterone levels, LOH may be diagnosed [2]. Treatment is appropriate when the testosterone level is below the lower limit of normal, generally accepted to be 300 ng/dL [3].

The incidence of LOH is increasing; serum testosterone levels decrease with age [4], and according to USA Census data from 2008, the population of men age ≥65 years is projected to double by the year 2035 [5]. An extrapolation from the Massachusetts Male Aging Study found a prevalence of LOH of almost half a million new cases per year in men in their fifth, sixth and seventh decades of life [6]. With the increasing incidence of LOH and better understanding of treatment benefits, the use of testosterone-replacement therapy (TRT) has been steadily increasing in recent years, as shown by a 500% increase in prescription sales of testosterone since 1993 [7].

The benefits of TRT are easily recognized, primarily because as levels of serum testosterone normalize, desired improvements in sexual function, libido and mood, as well as in overall quality of life, are reached. TRT also has a positive effect on cardiovascular health; one well-documented effect of TRT is a decrease in total cholesterol and low-density lipoprotein (LDL) levels [8]. Recent studies have shown a link between LOH and a high-risk, increasingly prevalent condition termed ‘metabolic syndrome’, also known as insulin-resistance syndrome, or syndrome X [9]. Metabolic syndrome is characterized by at least three of the following: abdominal obesity, hypertriglyceridaemia, low high-density lipoprotein (HDL) levels, hyperglycaemia and hypertension [10]. Laaksonen et al.[11] suggested that hypogonadism is an element of metabolic syndrome, based on their large population-based study. Treatment of LOH with TRT might improve the symptoms of metabolic syndrome, and vice versa, treatment of metabolic syndrome might improve the symptoms of LOH [12].

Despite its benefits TRT has several contraindications and questions remain about its effects on prostate health [13]. BPH is an androgen-sensitive disease that might be accelerated by endogenous and exogenous testosterone. Prostate volume is lower in hypogonadal men than in men with normal serum testosterone levels [14], and while prostate volumes of men with LOH undergoing TRT increase up to the levels of age-matched controls, TRT has not been shown to increase prostate volume beyond that of eugonadal men of the same age [15]. An increase in LUTS has not been documented in men undergoing TRT. Similar to BPH, prostate cancer is exquisitely androgen-sensitive. The first-line treatment for metastatic prostate cancer is androgen deprivation, either through bilateral orchidectomy or medical castration with GnRH agonists. While castration causes regression of prostate cancer, there is no proof that testosterone causes progression. Despite the lack of evidence, controversy persists as to whether testosterone can ‘awaken’ a clinically insignificant prostate cancer.

Short-term changes in PSA levels during TRT have been reported. Because a definitive, double-blind, randomized, placebo-controlled study on the effect of TRT on PSA levels has yet to be published, much of the available data are derived from relatively small, often retrospective studies. A systematic review of 18 short-term TRT trials found an average initial increase in PSA of 0.3 ng/mL [16]. A review of six studies of short-term TRT in slightly older men with LOH found a higher average increase in PSA of 0.43 ng/mL [16]. Rhoden and Mortgentaler [17] reported that 579 hypogonadal men on TRT from nine different studies, with a mean (range) duration of treatment of 11.8 (3–36) months, had a mean PSA level increase of 0.46 ng/mL. Among the longest follow-up data available on the subject, Wang et al.[18] followed 163 hypogonadal men on TRT for 3 years; they reported that after a small increase in mean (sd) PSA level from a baseline of 0.85 (0.06) to 1.11 (0.08) ng/mL at 6 months, there were no further significant increases with continued TRT over the following 3 years (P < 0.150). The effect that TRT has on PSA levels after 3 years has not been reported.

The mean rate of prostate cancer in published TRT trials is ≈1%[13]; in the review by Rhoden and Mortgentaler [17] of nine studies of PSA changes in men with LOH receiving TRT, they reported seven cases of prostate cancer in a total of 579 men (1.2%), although the mean duration of treatment was only 11.8 months. Three patients (1.8%) within the study by Wang et al.[18] of 163 hypogonadal men on TRT, were diagnosed with prostate cancer over a 3-year period. The incidence of prostate cancer with long-term TRT is unknown.

Thus the aim of the present study was to retrospectively review patients receiving TRT for LOH, to evaluate the long-term changes in PSA and any cardiovascular effects, as well as to review men diagnosed with prostate cancer during TRT.

PATIENTS AND METHODS

We reviewed the medical records of 267 men with a diagnosis of hypogonadism (International Center for Diseases-9, code 257.2) from January 2000 to June 2006 at the urology clinic at our institution. To be diagnosed with hypogonadism the patient must have reported symptoms in addition to having had a serum testosterone level of <300 ng/dL. Men must have been seen in our clinic at least twice or more with the diagnosis, yielding a total of 156 men with LOH for review.

For inclusion in the study, each patient must have had an initial testosterone measurement, PSA level and routine laboratory investigations, including body mass index (BMI), haematocrit, lipid profile and liver function tests (LFTs). The patients must have been started on TRT and then had at least one subsequent testosterone and PSA level measured within a year of starting TRT. The form of TRT was selected by patient preference. All men included in the study had a normal baseline PSA level before starting TRT, and had routine DREs during treatment. Patients with a biopsy-confirmed diagnosis or recent history of prostatitis before TRT were excluded from the study. If a man developed prostate cancer during treatment, TRT was discontinued and treatment for cancer was initiated. After applying these criteria, 81 men with LOH were selected for a complete chart review.

Each patient’s age at the time of diagnosis, and race, were identified as demographic data. The patient’s symptomatic response to treatment was determined subjectively by the patient in follow-up clinic visits as either a positive response or no response. Comorbid conditions, including BPH, coronary artery disease, diabetes mellitus type 2, hypertension, hyperlipidaemia, and obstructive sleep apnoea were also assessed. The treatment dates were rounded to the nearest 6-month interval.

Basic statistical data were obtained at each 6-month interval for all variables. The data were compared using Student’s t-test or one-way anova, with P < 0.05 considered to indicate statistical significance.

RESULTS

In all, 81 hypogonadal men (mean age 56.8 years, sd 11.5, range 25–82) were included in the study; Table 1 shows the characteristics of the study population. The patients were followed for a mean (range) of 33.8 (6–144) months after starting TRT, administered as a gel in 75 (93%), by injection in four (5%) and by a transdermal patch in one (1%). There was a positive symptomatic response in 90% of patients. The total and free testosterone levels before and at 3 years were significantly different (P < 0.05), as shown in Table 1.

Table 1.  The characteristics of the 81 patients
CharacteristicMean (sd, median, range) or n (%)
  • *

    P < 0.05.

Age, years56.83 (11.47, 55, 25–82)
Race
 White  62 (77)
 Black  17 (21)
 Other   2 (2.5)
Follow-up, months33.78 (29.7, 24, 6–144)
Symptomatic response  73 (90)
Comorbidities
 BPH  27 (33)
 Hypertension  44 (54)
 Coronary artery disease  20 (25)
 Diabetes mellitus  26 (32)
 Hyperlipidaemia  33 (41)
 Obstructive sleep apnoea  7 (9)
Baseline PSA level, ng/mL  1.32 (0.98)
Total testosterone, ng/dL
 Baseline 241.1 (103.5)
 3 years 379.8 (203.3)*
Free testosterone, ng/dL
 Baseline 7.93 (5.30)
 3 years13.34 (7.24)*

Total cholesterol improved from a mean (sd) of 203.8 (42.9) to 166.6 (23.5) mg/dL (P < 0.05) after 3 years of TRT, although the BMI, HDL, triglycerides and LDL did not change significantly. LDL decreased from 125 at baseline to 97 mg/dL, and triglycerides decreased from 230 at baseline to 165 mg/dL, but these differences were not statistically significant. The haematocrit increased from 41.7% to 47.3% (P = 0.07), and LFTs did not change significantly. The cardiovascular effects on the study population are shown in Table 2.

Table 2.  The cardiovascular effects of TRT
Mean (sd)Baseline3 years
  • *

    P < 0.05.

Total cholesterol, mg/dL203.8 (42.86)166.6 (23.45)*
Triglycerides, mg/dL230.3 (119.9)165.8 (68.56)
LDL, mg/dL125.4 (40.12)97.44 (47.37)
HDL, mg/dL47.47 (13.15)43.22 (7.66)
BMI, kg/m2 31.04 (6.45)32.10 (8.55)
Haematocrit 41.68 (5.46)47.33 (4.17)

The mean (sd) baseline PSA level for all men was 1.32 (0.98) ng/mL. In men without prostate cancer (95%) the PSA level did not increase significantly at 12-month intervals for 5 years. For these 77 men, the mean PSA level at 1, 2, 3, 4 and 5 years were 1.49 (P = 0.42), 1.35 (P = 0.89), 1.31 (P = 0.97), 1.10 (P = 0.58) and 1.43 ng/mL (P = 0.82), respectively (Fig. 1). There was no statistically significant difference from baseline at any of the yearly intervals. There was also no statistical difference in change in PSA level from baseline to 3 years when patients without prostate cancer were stratified into groups according to age (≤50, 55–65 and ≥70 years).

Figure 1.

The change in mean PSA over time in hypogonadal men on TRT with or without prostate cancer.

Four men (5%) developed prostate cancer at a mean (range) of 32.5 (22–41) months after starting TRT (Table 3). These men were slightly younger than men without cancer (mean age 54.5 vs 56.9 years) although this difference was not statistically significant (P = 0.7). In men with prostate cancer there was an increase in mean PSA level from baseline at 18 months of 1.8 ng/mL (P = 0.15), while the mean PSA level increased from baseline by 3.2 ng/mL (P < 0.05) at 3 years. Figure 1 shows the PSA dynamics in men with and without prostate cancer over a 5-year period. The sharp increase in PSA level in the group with cancer was first apparent at 18 months. The baseline PSA level among the men with cancer was slightly higher than the cancer-free group, at 2.23 vs 1.27 ng/mL (P = 0.05), respectively.

Table 3.  The characteristics of the four men with prostate cancer
Patientno.StageGleasonsumTreatmentMarginsPSA nadir, ng/mL
  1. EBRT, external beam radiotherapy; RRP, radical retropubic prostatectomy; RALP, robotic-assisted laparoscopic prostatectomy; NA, not applicable; *treated elsewhere.

31T1c3 + 3EBRTNA0.3
40T1c3 + 3RRPNegative<0.1
61T1c3 + 4RALPNegative<0.1
67T1c3 + 3NA*NANA

All four men who developed cancer were found to have low-grade, stage T1c disease. Three of these men had Gleason 3 + 3 disease, while one had Gleason 3 + 4 disease. One of the men with Gleason 3 + 3 disease chose external beam radiotherapy, and his PSA nadir was 0.3 ng/mL. The second patient with Gleason 3 + 3 disease had a radical retropubic prostatectomy; his margins were negative and his PSA remains undetectable. The patient with Gleason 3 + 4 disease chose robot-assisted laparoscopic radical prostatectomy; the margins were negative and his PSA level remains undetectable. The third patient who had Gleason 3 + 3 disease while on TRT pursued treatment for his cancer elsewhere.

DISCUSSION

TRT was effective in normalizing serum testosterone levels, with total testosterone increasing from 241.1 (103.5) at baseline to 379.8 (203.3) ng/dL (P = 0.003) at 3 years. This normalization resulted in a positive symptomatic response in 90% of the present patients. Improvements in sexual function, primarily erectile dysfunction, and in overall quality of life of the patients continue to support TRT as an effective treatment for LOH.

An added benefit of TRT is the favourable cardiovascular effect. The mean total cholesterol levels in the present patients improved from 203 mg/dL at baseline to 167 mg/dL at 3 years (P < 0.05). Although the changes were not statistically significant, the trends in LDL levels from 125 mg/dL at baseline to 97 mg/dL at 3 years (P = 0.19), and triglyceride levels from 230 to 166 mg/dL (P = 0.15) were concordant with other data reported [8]. A systematic review by Shabsigh et al.[19] found that TRT improves insulin sensitivity, central obesity, and lowers total cholesterol and LDL.

In men without prostate cancer (95%) the PSA level remained stable after normalization of testosterone for 5 years. Previously published compilation data from nine different studies with a mean (range) treatment duration of 11.8 (3–36) months report a range of PSA changes in men with LOH on TRT, from −0.01 to +1.0 ng/mL, with a mean PSA increase of 0.46 ng/mL [17]. The present study is one of the first to report changes in PSA in hypogonadal men on TRT for >3 years; although not statistically significant, the mean increase in PSA level from baseline to 5 years was 0.11 ng/mL.

Prostate cancer can be effectively diagnosed and treated in men taking TRT, and the incidence of cancer among men with LOH on TRT in the present study was no higher than that in the general population. In the review by Rhoden and Mortgentaler [17] of nine studies of PSA changes in men with LOH receiving TRT, they reported seven cases of prostate cancer in 579 men (1.2%), although the mean duration of treatment was only 11.8 months. In the present study, 4.9% of men developed prostate cancer, but the duration of treatment was significantly longer at 5 years, and the mean time to a diagnosis of prostate cancer was 32.5 months.

The four men who developed cancer had a mean increase in PSA level from baseline to 18 months of 1.8 ng/mL, and a mean increase in PSA level from baseline of 3.2 ng/mL at 3 years, leading to a prostate biopsy and diagnosis of cancer. An increase in PSA level after starting TRT was recently shown to be predictive of the diagnosis of prostate cancer [20], and this was the case in the present patients. None of the 77 men who remained cancer-free had an increase in PSA level prompting a biopsy during the study period, and all four men with a significant increase in PSA level and subsequent biopsy were found to have cancer. The baseline PSA level in the group with cancer was slightly higher than in the cancer-free group, but the difference was <1.0 ng/mL, and the mean baseline PSA level in the cancer group was still normal at 2.23 ng/mL; thus, the clinical significance of this observation is uncertain.

All four men with prostate cancer had low-grade disease; they were treated using standard protocols, by presenting treatment options and allowing a plan to be decided upon mutually between the provider and the patient. Two men chose prostatectomy, and both had negative margins and currently undetectable PSA levels. One man was referred for external beam radiotherapy and ultimately had a PSA nadir of 0.3 ng/mL. Therefore, these cases of low-grade disease were effectively diagnosed and treated despite the men taking TRT in the months preceding their diagnosis. Our results support the recent review by Morgentaler [21], asserting that there has been neither historical nor current data supporting a scientific basis for the belief that testosterone causes prostate cancer to grow.

The primary limitations to the present study were its retrospective design and lack of a control group. The study followed 81 patients, more than in similar studies, but still too small to have adequate power for statistical significance in many of the results. The present data from patients followed for >5 years is long-term compared with previously published reports, but a 10- or 20-year follow-up is necessary to thoroughly characterize the long-term effects of TRT on PSA and the development of prostate cancer.

In conclusion, LOH is an increasingly prevalent disease characterized by a symptomatically low testosterone level. TRT is effective in normalizing serum testosterone levels and providing a positive symptomatic response. In addition to its improvement in sexual function and overall quality of life, it has a beneficial cardiovascular effect, particularly by decreasing total cholesterol levels.

We found that PSA levels remain stable after testosterone normalization for ≥5 years. In addition, prostate cancer can be effectively diagnosed and treated in men taking TRT. The incidence of prostate cancer among men with LOH on TRT was no higher than that in the general population. Thus, TRT effectively treats LOH and does not increase the overall risk of developing prostate cancer.

CONFLICT OF INTEREST

None declared.

Ancillary