Stable Intraprostatic Dihydrotestosterone in Healthy Medically Castrate Men Treated With Exogenous Testosterone

Context:

Concern exists that T replacement therapy (TRT) might increase the risk of prostate disease. There are limited data regarding the impact of TRT on prostate androgen concentrations.

Objective:

Determine the dose-dependent effects of exogenous T administration on intraprostatic androgen concentrations.

Design:

Twelve-week, double-blinded, randomized, placebo-controlled trial.

Setting:

Academic medical center.

Participants:

Sixty-two healthy eugonadal men, aged 25–55 years.

Interventions:

Subjects were randomly assigned to receive injections of acyline, a GnRH antagonist (used to achieve medical castration), every 2 weeks plus transdermal T gel (1.25 g, 2.5 g, 5.0 g, 10 g, or 15 g daily), or placebo injections and transdermal gel for 12 weeks.

Main Outcomes:

Serum T and dihydrotestosterone (DHT) were measured at baseline and every 2 weeks during treatment. Intraprostatic T and DHT concentrations were assessed from tissue obtained through ultrasound-guided prostate needle biopsies at week 12. Androgens were quantified by liquid chromatography-tandem mass spectrometry.

Results:

51 men completed the study and were included in the analysis. There were no significant adverse events. Exogenous T resulted in a dose-dependent increase in serum T and DHT concentrations (190–770 and 60–180 ng/dL, respectively). Although intraprostatic T differed among dose groups (P = .01), intraprostatic DHT was comparable regardless of T dose (P = .11) and was 10- to 20-fold greater than intraprostatic T.

Conclusions:

In healthy, medically castrate men receiving exogenous T, the total intraprostatic androgen concentration (predominantly DHT) remained stable across serum T concentrations within the physiological range. These findings further our knowledge of the relationship between serum and intraprostatic androgens and suggest that physiological serum T achieved by TRT is unlikely to alter the prostate hormonal milieu.

Healthy men received medical castration and varying doses of testosterone gel replacement. Despite dose-dependent increase in serum androgens, intraprostatic androgen milieu remained unchanged.

The use of T replacement therapy (TRT) is increasing (1), with global sales of T now exceeding $2 billion annually. Chronically low T exposure in men leads to adverse effects on body composition, bone mass, and sexual function (2) and may increase the risk of diabetes, cardiovascular disease, and mortality (3–5). Aging in men is characterized by a gradual and progressive decline in T production; 20% of men over 60 years of age exhibit serum T levels below the reference range for young, healthy men (6, 7). TRT, therefore, has the theoretical potential to broadly improve health outcomes in hypogonadal men (8–10), a tenet supported by initial results of the Testosterone Trials (11). Despite escalating use, however, the relative risks and benefits of TRT remain poorly defined, and concerns persist, particularly regarding the risk of prostate disease in men using TRT (2, 12).

Androgens are critical regulators of prostate development and have been implicated in the pathogenesis of prostate disease. Androgen deprivation therapy substantially lowers intraprostatic androgen concentrations and remains a cornerstone of prostate cancer therapy (13). Furthermore, men with untreated hypogonadism exhibit reduced prostate size and lower serum levels of prostate-specific antigen (PSA) (14). Such observations have contributed to the notion that lower systemic androgen exposure confers protection from prostate disease, but substantial clinical data do not support this conclusion (15). The prevalence of prostate disease rises among older men coincident with a decline in T production, and low serum T levels have been associated with higher-grade prostate malignancies (15, 16). Moreover, observational longitudinal studies have failed to demonstrate an association between higher endogenous serum T levels and the incidence of prostate cancer (17) or an increased risk of prostate cancer with long-term TRT (18). Although limited in size and duration, randomized clinical trials to date have not reported a higher incidence of prostate disease among men receiving TRT (19), although none have been powered for such clinical endpoints.

Androgen-responsive tissues appear to exhibit differential sensitivity to changes in circulating T levels. Exogenous T exerts linear dose-dependent effects on tissues such as muscle, and organ-specific symptoms of androgen deficiency occur at variable thresholds of serum T (20, 21). However, intraprostatic androgen levels remained readily detectable, even in the setting of medical castration (22), and low-dose androgen administration in healthy men yielded no effect on intraprostatic androgens (23, 24). These data imply dissociation between intraprostatic and serum androgen concentrations and could provide a physiological explanation for differential tissue sensitivities to TRT. Intraprostatic androgen concentrations are subject to local regulation by type 2 5α-reductase, resulting in high intraprostatic concentrations of dihydrotestosterone (DHT) compared to T. Understanding the dose-dependent effects of exogenous T on intraprostatic androgen concentrations could contribute to optimizing dosing strategies for TRT by targeting circulating T concentrations that maximize clinical benefit while minimizing potential risks.

To understand the relationship between serum and intraprostatic DHT and T concentrations across the physiological range, we performed a double-blinded, randomized, placebo-controlled intervention study in healthy men who were administered a range of doses of daily T gel while the hypothalamic-pituitary-testicular axis was simultaneously suppressed with a GnRH antagonist. We hypothesized that within the normal range of circulating T concentrations for healthy young men, intraprostatic DHT and T concentrations would remain stable.

Subjects and Methods

Participants

Healthy male volunteers, 25–55 years old, were recruited via advertisement and gave written informed consent before screening (Supplemental Table 1 contains inclusion and exclusion criteria). All study procedures took place at the University of Washington Medical Center, Seattle, Washington, and were approved by the Institutional Review Board.

Study design and interventions

After screening, subjects were randomized (using random block allocation, block size 4) in a double-blinded fashion to one of six treatment groups for 12 weeks. Group 1 received placebo injection (every 2 weeks) plus daily placebo transdermal gel. Groups 2–6 received injections of the GnRH antagonist, acyline (PolyPeptide Laboratories) 300 μg/kg every 2 weeks, plus the following doses of daily 1% transdermal T gel (Abbvie Inc.): group 2, 1.25 g; group 3, 2.5 g; group 4, 5 g; group 5, 10 g; or group 6, 15 g.

Subjects were seen biweekly during the intervention and at recovery (week 18); vital signs and blood for serum chemistries, hormone, and complete blood counts were collected at all visits. Androgens are known to stimulate erythropoiesis (25); therefore, we assessed hematocrit at baseline and throughout treatment as a measure of peripheral androgen effect. International prostate symptom scores (IPSS) (26) and sexual function questionnaires (27) were collected every 4 weeks. Prostate volume and PSA were quantified at baseline, week 12, and recovery. Ten prostate core biopsies were obtained at week 12 using an 18-gauge needle via transrectal ultrasound guidance with local anesthetic (1% lidocaine), snap-frozen in liquid nitrogen, and stored at −80°C (22). Androgen concentrations were individually quantified in three biopsies per subject and averaged.

Outcome assessments

The primary outcome measures were serum and intraprostatic androgen concentrations. These were measured by liquid chromatography-tandem mass spectrometry in a blinded manner as described (28). This assay is validated for T quantification by the Centers for Disease Control and Prevention using certified standards. The lower limit of quantification for each androgen was ≤0.01 ng/mL in all cases. The intra-assay coefficients of variation were: T = 4.9%, DHT = 4.4%, dehydroepiandrosterone = 7.6%, and androstenedione = 3.5%. Normal reference ranges for serum T and DHT using these methods were established from 118 morning samples from healthy men ages 18–55 years. Serum LH, FSH, and SHBG were quantified by immunofluorometric assay (29), and serum chemistries, blood counts, and PSA were measured in the clinical laboratory at the University of Washington, Seattle, Washington. Free T was calculated as described by Södergård (30). For each subject, three separate prostate core biopsies (average core weight, 7.1 ± 1.4 mg) were individually extracted and quantified; androgen concentrations presented are the average of these three values.

Power and statistical analyses

Based on previous data (22), the inclusion of 12 subjects per group afforded an 80% power to detect a 30% difference in intraprostatic DHT concentration between the lowest dose T group and the other groups receiving T gel (α 0.05, assuming 30% variation). Because the data were not normally distributed, nonparametric statistical analyses were performed, and the results are presented as medians with interquartile ranges. Comparisons between groups at a given time point were made using a Kruskal-Wallis ANOVA (P < .05 considered significant) with a Wilcoxon rank-sum post hoc test for pairwise comparisons, with a Bonferroni adjusted P < .008 being considered significant. Within-group changes from baseline were analyzed using a Wilcoxon signed rank test, with an unadjusted P < .05 being considered significant. Spearman's correlation was used to evaluate the associations between serum and intraprostatic hormone concentrations and hematocrit. All analyses were performed using either STATA 10 (StataCorp) or GraphPad Prism 6.0 (GraphPad Software, Inc).

Results

Subjects

Ninety-eight subjects were screened; 53 completed the study, and 51 were included in the analyses (Supplemental Figure 1). Two subjects were excluded from the entire analyses due to noncompliance. We excluded one subject, with persistent anemia throughout the study, from the hematocrit analysis to quantify the effects of exogenous T on normal erythropoiesis. There were no significant differences at baseline between subjects across the groups (Table 1 and Supplemental Table 2). Baseline prostate parameters including PSA, prostate volume, and IPSS were also similar (Table 2).There were no serious adverse events (AEs) during the study, including no cardiovascular events, significant increases in blood pressure, or prostate-related AEs (detailed in Supplemental Table 3). Treatment-related AEs possibly related to treatment-related hypogonadism included hot flashes, decreased libido, erectile dysfunction, and fatigue. Some subjects experienced skin irritation and dryness with the gel, acne, and pain at the injection site. None of these required intervention.

Table 1.

Baseline Characteristics of Study Subjects by Daily T Dose Group

Variable	Placebo	1.25 g T	2.5 g T	5 g T	10 g T	15 g T	All Groups	P Value
No. of subjects	8	9	8	9	8	9
Age, y	45 (34–48)	45 (41–46)	46 (39–51)	34 (30–44)	42 (33–47)	51 (50–54)	44 (35–50)	.07
BMI, kg/m2	28 (26–33)	28 (25–33)	26 (24–28)	25 (25–27)	25 (22–28)	26 (25–28)	26 (25–29)	.39
T, ng/dL	450 (315–516)	448 (345–509)	554 (462–730)	419 (387–505)	482 (429- 591)	452 (396–536)	459 (390–530)	.35
DHT, ng/dL	36 (27–47)	37 (24–44)	43 (39–59)	39 (18–49)	49 (44–55)	36 (28–53)	40 (28–52)	.24
LH, mIU/mL	3.5 (3–4.5)	5 (4–6)	4 (2–5.5)	3 (3–4)	4 (4–5)	3 (3–5)	4 (3–5)	.61
FSH, mIU/mL	5.5 (4–8.5)	4 (3–6)	5 (3–6)	5 (3–6)	5.5 (5–7)	7 (4–8)	5 (4–7)	.61
Hematocrit, %	43 (42–45)	43 (41–44)	43 (42–44)	44 (41–45)	45 (42–45)	43 (41–44)	43 (41–45)	.91

Abbreviation: BMI, body mass index. Values are expressed as median (interquartile range). Treatment groups received different doses of 1% T gel; for instance, the 1.25-g T group applied 1.25 g of 1% T gel daily for 12 weeks, and so on.

Table 2.

Prostate Parameters by Group at Baseline and End of Treatment

Variable	Baseline	End of Treatment (wk 12)	P Value of Change Within Group
PSA, ng/dL
Placebo	0.80 (0.61–0.91)	0.82 (0.57–0.97)	.40
1.25 g T	0.71 (0.66–0.88)	0.48 (0.44–0.62)	.02a
2.5 g T	1.09 (0.75–1.25)	0.61 (0.39–0.81)	.08
5 g T	0.74 (0.34–0.86)	0.58 (0.41–0.78)	.91
10 g T	0.67 (0.41–0.91)	0.52 (0.43–0.67)	1.0
15 g T	0.69 (0.35–0.91)	0.76 (0.50–1.2)	.16
All groups	0.75 (0.42–0.96)	0.58 (0.43–0.80)
P value of ANOVA across groups	.56	.54
Prostate volume, cm3
Placebo	20 (16–22)	19 (18–22)	.64
1.25 g T	19 (13–23)	18 (14–22)	.55
2.5 g T	17 (17–21)	19 (15–21)	.97
5 g T	17 (14–18)	15 (14–17)	.36
10 g T	17 (14–20)	16 (15–20)	.74
15 g T	19 (15–19)	20 (16–20)	.01a
All groups	18 (14–21)	18 (15–20)
P value of ANOVA across groups	.75	.29
IPSS
Placebo	2.5 (1.5–3)	1 (0–2)	.13
1.25 g T	1 (1–3)	2 (1–3)	1.0
2.5 g T	2 (0–4.25)	2.5 (1.5–5.25)	.41
5 g T	0 (0–1)	0 (0–0)	.50
10 g T	1.5 (0–3)	2.5 (1.5–7.75)	.13
15 g T	2 (1–3)	4 (1–10)	.08
All groups	2 (0–3)	2 (0–4)
P value of ANOVA across groups	.63	.14

Values are expressed as median (interquartile range). Treatment groups received different doses of 1% T gel; for instance, the 1.25-g T group applied 1.25 g of 1% T gel daily for 12 weeks, and so on.

^a

statistically significant.

Serum hormones

We calculated average on-treatment serum T and DHT concentrations in all study groups using biweekly levels measured from weeks 2–12, in part to account for intraindividual variability in T concentrations observed with transdermal T gel (31). As expected, serum T (Figure 1A) and DHT (Figure 1B) concentrations increased with increasing doses of exogenous T to produce serum T levels ranging from hypogonadal to the high end of the normal range. Similar increases in free T concentrations were observed (P < .0001) (Supplemental Figure 2). Serum DHT concentrations were supraphysiological in most of the dose groups. Serum levels of other steroids were not different across the groups (Supplemental Figure 3, A and B).

Figure 1.

Serum T (A) and serum DHT (B) concentrations during treatment (weeks 2–12) and intraprostatic T (C) and intraprostatic DHT (D) concentrations at the end of treatment (week 12).

White solid lines represent medians. Black hashed lines represent normal reference ranges. Significant differences are denoted as relative to 15 g T (*), 10 g T (#), 5 g T ($), and 2.5 g T (Ω) groups, respectively. ng/g, nanograms per gram of prostate tissue.

Intraprostatic androgens

Despite significant dose-proportionate increases in serum T and DHT concentrations, after 12 weeks of treatment, there were no significant differences across the groups in intraprostatic DHT concentrations (Figure 1D). Intraprostatic DHT levels were roughly 10–20 times intraprostatic T (median intraprostatic DHT, 4.03 ng/g; and median intraprostatic T, 0.22 ng/g). Intraprostatic T differed across the groups at week 12, with significantly lower levels in the 2.5-g group compared to both the placebo and 15-g groups (Figure 1C). Despite higher levels of serum T in the 15-g group, intraprostatic T was not significantly higher in this group than placebo. All other intraprostatic steroids were similar across the groups (Table 3).

Table 3.

Additional Intraprostatic Sex Steroid Concentrations at End of Treatment (Week 12)

Variable, ng/g	Placebo	1.25 g T	2.5 g T	5 g T	10 g T	15 g T	P Value
17-OHPreg	1.19 (1.06–1.44)	1.25 (0.93–2.07)	1.23 (0.92–2.04)	1.05 (1.01–1.89)	1.02 (0.94–1.45)	1.36 (1.29–1.47)	.94
17-OHP	0.05 (0.04–0.06)	0.05 (0.05–0.06)	0.05 (0.04–0.05)	0.05 (0.04–0.05)	0.05 (0.04–0.05)	0.05 (0.04–0.07)	.65
Androstenedione	0.16 (0.10–0.33)	0.11 (0.10–0.18)	0.12 (0.09–0.15)	0.13 (0.12–0.15)	0.18 (0.15–0.28)	0.14 (0.12–0.22)	.44
Androsterone	0.17 (0.14–0.21)	0.15 (0.12–0.22)	0.12 (0.08–0.14)	0.17 (0.11–0.24)	0.17 (0.12–0.43)	0.21 (0.17–0.23)	.16
DHEA	29.3 (19.1–42.0)	26.2 (13.5–35.8)	22.1 (16.2–30.1)	26.8 (26.0–29.2)	24.4 (23.0–28.2)	18.6 (14.7–22.0)	.29
Pregnenolone	28.7 (23.6–33.8)	29.5 (21.3–36.4)	32.7 (23.3–51.2)	32.4 (28.6–45.5)	27.9 (17.4–37.6)	23.9 (18.1–25.0)	.14
Progesterone	0.09 (0.06–0.20)	0.08 (0.06–0.09)	0.06 (0.04–0.09)	0.08 (0.07–0.13)	0.09 (0.08–0.12)	0.07 (0.07–0.08)	.69

Abbreviations: 17-OHPreg, 17-hydroxypregnenolone; 17-OHP, 17-hydroxyprogesterone; DHEA, dehydroepiandrosterone. Values are expressed as median (interquartile range). Treatment groups received different doses of 1% T gel; the 1.25-g T group applied 1.25 g of 1% T gel daily for 12 weeks and so on.

Correlation of serum and intraprostatic androgens

The average serum T concentrations during treatment were positively correlated to both intraprostatic DHT (Spearman's r = 0.41; P = .003) and intraprostatic T (Spearman's r = 0.51; P < .001) at week 12. Notably, the correlation with intraprostatic DHT disappeared if subjects (n = 12) with very low (<220 ng/dL) or very high (>900 ng/dL) serum T were excluded. Similarly, average serum DHT concentrations during treatment were positively correlated with both intraprostatic DHT (Spearman's r = 0.34; P = .01) and intraprostatic T (Spearman's r = 0.29; P = .04), but these correlations were lost when only subjects with serum T levels of 220–900 ng/dL were included (n = 39).

Other prostate measures

PSA did not increase during treatment in any group; however, there was a statistically significant decrease in the 1.25-g group (Table 2). Prostate volume did not change across the groups (P = .29), except for the 15-g group, in which there was a small increase from baseline to end of treatment (P = .01). There were no significant changes in IPSS across the groups throughout the study (Table 2).

Hematocrit

Hematocrit at the end of treatment differed significantly across the groups (Figure 2); the two lowest dose groups (1.5 g and 2.5 g) had lower hematocrits than the 15-g group. The 15-g group exhibited a significant increase in hematocrit over the treatment period (P < .01). The end-of-treatment hematocrit showed a moderate correlation with average serum T concentrations during treatment (r = 0.45; P = .0009).

Figure 2.

Hematocrit (%) at the end of treatment (week 12).

*, Significant differences relative to 15-g T group.

Discussion

Our findings demonstrate that in healthy men with GnRH antagonist-induced androgen deficiency, intraprostatic androgen concentrations do not parallel increasing serum androgen concentrations achieved with escalating sub- to supraphysiological doses of exogenous T. DHT, a more potent androgen than T, is the predominant intraprostatic androgen. Across the physiological range of serum T concentrations, intraprostatic T and DHT levels are remarkably stable. Concentrations of intraprostatic DHT were similar across all treatment groups, despite serum DHT levels that rose well above the physiological range in men receiving 2.5–15 g of daily transdermal T. Intraprostatic T concentrations were stable for all dose groups compared to placebo treatment and only varied when comparing the lowest (1.25 g) and highest (15 g) dose groups. Consistent with these minimal differences in intraprostatic androgen levels, no increases in PSA were evident among subjects in any treatment group. Similarly, prostate volume was stable across the 12-week treatment period for all but the highest dose group, in which there was a very small increase in prostate volume. Consistent with these quantitative analyses, no changes in prostate-related symptoms (IPSS) were observed. In contrast to effects on the prostate, hematocrit rose in a dose-dependent manner with increasing serum T concentrations. These results demonstrate that circulating sex steroids have distinct and tissue-specific effects on individual androgen-responsive end organs.

Limited data available regarding the relationship between circulating and intraprostatic androgen concentrations are consistent with those presented here. Marks et al (32) observed no change in intraprostatic T and DHT concentrations with TRT in older hypogonadal men treated with low-dose T injections for 6 months, despite increases in serum T concentrations compared to baseline. Our study builds on these findings. The use of transdermal T gel applied daily over the treatment period allowed us to examine the impact of stable circulating androgen levels across a broad range of serum concentrations, particularly at the higher end of the normal range for T, where many clinicians target therapies for hypogonadal men. Moreover, we found that supraphysiological levels of circulating DHT induced by transdermal T treatment had no impact on intraprostatic DHT, consistent with a previous trial of DHT gel administration for a much shorter duration (24). Finally, we measured serum and intraprostatic androgens by liquid chromatography-tandem mass spectrometry, allowing for quantification of additional androgens and avoiding potentially confounding cross-reactivity manifest in older immunoassays.

Androgen deprivation clearly lowers intraprostatic androgens both in healthy men (22) and men with prostate cancer (33), demonstrating that without adequate circulating substrate (T), the intraprostatic androgen milieu is compromised. In our study, we noted a moderate correlation between serum T and intraprostatic androgens, but this correlation was lost if we only considered serum T concentrations in the normal physiological range. Therefore, our data support the hypothesis that the prostate maintains a buffering capacity within the physiological range of serum T, and that this buffer is lost only with extremes of low or high serum T. It is worth noting that the variance in intraprostatic T was highest in the highest dose group (15 g). It is possible that contamination with blood (in this case, containing high concentrations of T) may have contributed to higher levels of intraprostatic T in some subjects in this group.

Previous data indicate that whereas PSA and prostate volume decrease when serum T levels fall well below the lower end of the normal range, PSA and prostate volume remain stable across a wide range of serum T levels once a threshold serum T is exceeded (20, 24, 34, 35). Our data give a biological basis for these observations, given the stability of the intraprostatic androgen milieu across a broad range of serum T. Our data are consistent with (without proving) a “saturation model” of prostate response to androgens (36). This model proposes that androgen receptor (AR) signaling within the prostate is maximally stimulated at serum T concentrations of approximately 200 ng/dL. At serum T levels below this value, intraprostatic androgen signaling becomes attenuated, but limited incremental increases in androgen signaling occur at higher circulating T levels (15, 36). Consistent with these findings, we observed a small but significant decrease in PSA exclusively among subjects receiving the lowest T dose (serum T average = 190 ng/dL), but PSA did not increase in any of the high-dose treatment arms. However, in the highest-dose group, we noted a slightly higher intraprostatic T and prostate volume and trends toward increases in PSA and IPSS that might argue against AR saturation. Whether or not higher doses of exogenous androgens (such as those used in doping) alter intraprostatic androgens and the effects of such changes on AR signaling in the prostate microenvironment cannot be determined from our data. Notably, and of clinical relevance, our data suggest that within the recommended therapeutic doses of TRT (5–10 g T gel daily), there is a buffering of intraprostatic androgens, particularly DHT, and no differences in relevant AR-regulated endpoints including PSA, prostate volume, or IPSS are seen. Similarly, across the physiological range of serum T, no correlation was observed with intraprostatic DHT.

Clinical data suggest that baseline T status may influence prostate response to exogenous T exposure (37). Our study design utilized healthy men rendered acutely castrate, and it is therefore possible that TRT in men with long-standing hypogonadism may not mimic our findings. Moreover, our results should not be extrapolated to patients with prostate cancer or benign prostatic hyperplasia, who may respond differently to exogenous androgens. In men with existing prostate cancer, those who had undergone prior castration exhibited a higher risk of disease progression than eugonadal patients after T administration (37). Nonetheless, administration of exogenous T to men with hypogonadism consistently confers only nominal changes in lower urinary tract symptoms, PSA, and prostate volume (14). Furthermore, TRT was not associated with prostate growth or disease progression in a small cohort of hypogonadal men undergoing active surveillance for known prostate cancer (38), and limited data from randomized controlled trials have not suggested an increase in prostate-related morbidity or mortality with TRT (39). Thus, further clinical studies are needed to delineate the dose-dependent effects of exogenous T on prostate in men with underlying T deficiency of varying severity and duration.

Additional limitations of the current study include exclusion of older subjects, small sample size, and the relatively short treatment duration. Although the concern for adverse prostate effects from TRT has mainly focused on men >60 years old, we enrolled middle-aged men in this physiology study to minimize risk to subjects of incidental prostate cancer detection and the potential inclusion of men with underlying prostate disease that might confound the dose-response relationship between T treatment and intraprostatic androgen concentrations. Identifying healthy men willing to undergo prostate biopsies proved challenging. As a result, it is possible that we were unable to identify a small difference in intraprostatic DHT and T between groups. However, given the similarity in intraprostatic DHT across groups, the possibility of a type II error seems relatively small. In terms of study design, it would have been ideal to also perform a baseline prostate biopsy in subjects to allow for robust paired analysis of intraprostatic DHT using group as an interaction term. However, given the invasiveness and discomfort of this procedure, paired prostate biopsies in healthy men were not thought to be feasible. Our current analysis approach may have also introduced some error when comparing serum and intraprostatic hormone measurements. We investigated the impact of serum T levels across a dose range for 12 weeks; it is possible that changes in intraprostatic androgens, PSA, or prostate volume could become evident with more sustained exposure to exogenous T, although results from Marks et al (32) over a 6-month treatment period are consistent with those presented here.

The strengths of our study include randomized, double-blinded, controlled design; stable serum T levels produced by T gel compared to T injections; and the measurement of androgens by a validated mass spectrometry assay, which is more accurate and sensitive than immunoassay and thus particularly important for measuring low levels of T found in the prostate. In addition, we were able to quantify additional intraprostatic androgens and therefore examine compensatory changes in the intraprostatic androgen milieu, which we did not observe. Our results are consistent with previous work demonstrating that healthy older men undergoing medical castration followed by im T administration exhibit dose-dependent effects on hemoglobin but not PSA, suggesting that the nonlinear response of intraprostatic androgens to serum T we report here using transdermal gel may be similar in men using injections for TRT (40). However, this remains to be tested experimentally because the peaks in serum T concentrations associated with T injections could cause transient increases in intraprostatic T. Finally, these data are of particular relevance in the design of future large clinical trials assessing the risks of TRT following the recently published Testosterone Trial (11) and suggest that targeting serum T levels in the mid-physiological range will appropriately minimize, while at the same time assessing, risks to prostate health of study participants.

Overall, our findings indicate that intraprostatic androgen concentrations remain stable and do not show a dose-response across a broad physiological range of serum T levels. Our data also underscore the tissue-specific regulation of androgen concentrations, with the prostate exhibiting a high degree of local regulation and capacity to maintain stable androgen levels across a range of serum T and DHT concentrations. Notably, intraprostatic T levels and prostate volume increased only in the treatment group that received the highest dose of exogenous T, one well above that commonly used in clinical practice and recommended for TRT. Long-term, randomized intervention studies are clearly needed to address the possible risks of TRT and the effects of age, antecedent androgen deprivation, and duration of treatment on prostate health. Nevertheless, our findings demonstrate that TRT at treatment doses that are sufficient for clinical benefit and peripheral androgen effects do not increase intraprostatic androgen concentrations in healthy men.

Acknowledgments

We are grateful to the Eunice Kennedy Shriver National Institute of Child Health and Human Development for providing acyline at no cost to the study (IND no. 53539). We thank our research study coordinator, Ms. Kathryn Torrez Duncan, and study nurse Marilyn Buscher, RN, as well as the research subjects, without whom this work would not be possible.

This work was supported by the National Institute of Aging, National Institutes of Health (NIH) through Grant RO1AG037603 (to S.T.P.). Additional support was received as follows: NIH Grant 5T32DK007247-37 (to A.T.), the Robert B. McMillen Professorship (to S.T.P.), University of Washington Nutrition Obesity Research Center Pilot and Feasibility Award P30 DK035816, the American Heart Association Clinical Research Program, and the Eunice Kennedy Shriver National Institute of Child Health and Development Grant 6K12 HD053984 (all to K.B.R.). The work was also supported by a VA Advanced Fellowship in Geriatrics (to L.A.C.) and resource and facilities use at VA Puget Sound Health Care System in Seattle, Washington (to B.T.M. and A.M.M.).

This study is registered at ClinicalTrials.gov, trial identifier: NCT 01327495.

Disclosure Summary: Transdermal T gel and placebo gel were provided to the study at no cost by Besins Healthcare (Bangkok, Thailand) as an investigator-initiated grant. Besins Healthcare had no input into the design, conduct, data analyses, or manuscript preparation and provided no other financial support for the study. J.K.A. received grant support from Clarus. A.M.M. received research support from Abbott Laboratories and GlaxoSmithKline and is a consultant for Abbott Laboratories, Lilly USA, LLC, Endo Pharmaceuticals, and Clarus. S.T.P. received drug supplies for an investigator-initiated study from Besins Healthcare. A.T. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

A.T. and L.A.C. are co-first authors who have contributed equally to this work.

Abbreviations

 
• AE

  adverse event

 
• AR

  androgen receptor

 
• DHT

  dihydrotestosterone

 
• IPSS

  International Prostate Symptom Score

 
• PSA

  prostate-specific antigen

 
• TRT

  T replacement therapy.