OUP user menu

Testosterone Replacement Therapy Outcomes Among Opioid Users: The Testim Registry in the United States (TRiUS)

Gary Blick MD, Mohit Khera MD, MBA, MPH, Rajib K. Bhattacharya MD, Dat Nguyen PharmD, Harvey Kushner PhD, Martin M. Miner MD
DOI: http://dx.doi.org/10.1111/j.1526-4637.2012.01368.x 688-698 First published online: 1 May 2012


Objective. Among patients with hypogonadism-associated comorbidities, opioid users have the highest incidence of hypogonadism. Data from the Testim Registry in the United States were analyzed to determine the efficacy of testosterone replacement therapy in opioid users vs nonusers.

Design. Prospective, 12-month observational cohort registry.

Subjects. Hypogonadal men (N = 849) prescribed Testim (but not necessarily testosterone replacement) for the first time.

Interventions. Testim 1% testosterone gel (5–10 g/day).

Outcome Measures. Total and free testosterone, sex hormone-binding globulin, prostate-specific antigen, sexual function, mood/depression, and anthropometric data were assessed. Changes from baseline were analyzed using repeated measures mixed-effects analysis of variance; multiple linear regressions of changes in testosterone levels with sexual function, mood, and opioid use were computed.

Results. 90/849 patients (10.6%) reported opioid use at baseline; 75/90 (83%) used opioids for ≥30 days prior to baseline. Baseline total testosterone and prostate-specific antigen were not statistically different between opioid users and nonusers; there was a trend for higher sex hormone-binding globulin (P = 0.08) and lower free testosterone (P = 0.05) in opioid users. After 1 month, both opioid users and nonusers had significant (P < 0.001) increases in total and free testosterone, which continued through 12 months. Sexual function and mood improved significantly in both opioid users and nonusers over 12 months, and significantly correlated with change in total testosterone.

Conclusions. Testosterone replacement therapy increased serum testosterone in hypogonadal opioid users and nonusers alike. The data suggest that with testosterone replacement, hypogonadal opioid users might be expected to have similar improvements in sexual function and mood as opioid nonusers.

  • Opioid
  • Testosterone
  • Hypogonadal
  • Testosterone Replacement Therapy
  • Testosterone Gel


Both opioid addiction and long-term opioid therapy for chronic pain can cause physical symptoms of testosterone deficiency (hypogonadism), including loss of libido, infertility, fatigue, depression, anxiety, loss of muscle mass, osteoporosis, and impotence. These symptoms are associated with hypogonadism through opioid-mediated suppression of the hypothalamic–pituitary–gonadal axis, which controls the production of testosterone [1]. Up to 75% of individuals on long-term opioid therapy for chronic pain and/or with opioid addiction may be hypogonadal [2].

Endogenous and exogenous opioids bind to opioid receptors in the hypothalamus, pituitary gland, and testis. When bound in the hypothalamus, opioids decrease the release of gonadotropin-releasing hormone (GnRH) or modify its pulsatility, resulting in decreased release of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) from the pituitary, decreased gonadal steroid production, and hypogonadism [3]. In addition, opioids may have direct effects on the testis, such as decreased secretion of testosterone and decreased testicular interstitial fluid. Other effects of opioids are increased prolactin, which decreases testosterone production, and modified production by the adrenal glands of dehydroepiandrosterone (DHEA), an important precursor of testosterone [1].

Hypogonadism, whether occurring due to natural causes or secondary to opioid use, is often associated with health risks and quality-of-life issues, such as age-related declines in bone mineral density, muscle mass, strength, and physical function (thereby predisposing men to health problems, such as musculoskeletal frailty, osteoporosis, and fractures [4,5]); decreased sexual function; depressed mood; and metabolic changes involving abdominal obesity, diabetes, insulin resistance, impaired glucose tolerance, increased low-density lipoprotein (LDL) cholesterol, increased LDL/HDL cholesterol ratios, and metabolic syndrome [6–11].

There are several treatment options for men experiencing opioid-induced hypogonadism, including nonopioid pain management, opioid rotation (although this approach has not been studied), or treatment with testosterone replacement therapy (TRT) [1]. TRT in opioid users requires consideration of the risks of hypogonadism as well as the risks and benefits of the treatment [1,9]. In one TRT outcome study that focused on chronic opioid users, testosterone patch therapy over 6 months in 16 men with opioid-induced androgen deficiency (OPIAD) was reported to be effective in increasing testosterone levels and improving sexual function and mood [12]. Potential adverse events and risks associated with TRT include erythrocytosis, acne, detection of subclinical prostate cancer, growth of metastatic prostate cancer, reduced sperm production and fertility, gynecomastia, worsening of benign prostatic hyperplasia symptoms, or decreases in HDL cholesterol [1,9]. Findings from the Testosterone in Older Men with Mobility Limitations (TOM) trial suggest that TRT may be associated with cardiovascular risk [13]; however, this is somewhat controversial given the limitations of the TOM study, and further research is necessary.

In our analysis, we used data from the Testim Registry in the United States (TRiUS), which enrolled a large group of hypogonadal men on Testim® (1% topical testosterone gel; Auxilium Pharmaceuticals, Malvern, PA) for 12 months. Patients were drawn from diverse clinical settings from around the United States. Although registries do not include a placebo or other control group, these prospective, observational studies provide an opportunity to evaluate outcomes that reflect how treatments are used in actual clinical practice. Further, registries include a wide variety of patients who are representative of those that clinicians can expect to encounter in routine practice, rather than a specific sample of patients who are required to meet the restrictive inclusion and exclusion criteria used in clinical trials.

The objective of this analysis was to evaluate the effectiveness of TRT in hypogonadal men who were opioid users in comparison with opioid nonusers. Key outcomes included testosterone levels, sexual function, depression, and anthropometric measures.


Study Design and Enrollment

TRiUS was a large, prospective observational cohort registry of hypogonadal men treated with testosterone 1% gel and followed for 12 months. Details of the TRiUS methodology and baseline characteristics of the registry population, including baseline symptoms and comorbidities, have been published previously [14]. Briefly, between March and December 2008, patients were enrolled from 72 clinical settings, including endocrinology, urology, general internal medicine, and HIV specialist practices. Inclusion criteria were men with a diagnosis of hypogonadism who were being prescribed Testim for the first time, which could include patients who were nonresponders to previous TRT with different preparations. Exclusion criteria included hypersensitivity to any ingredients in the testosterone gel, including testosterone, or carcinoma of the breast or known or suspected prostate cancer. All study sites had local or central Institutional Review Board (IRB) approval, and research was carried out in compliance with the Declaration of Helsinki as currently amended. Study participants completed IRB-approved consent forms and were free to discontinue at any time for any reason.

Diagnosis and laboratory measures were performed according to each investigator's customary practice. Each subject was prescribed by their physician either 5 g (one tube) or 10 g (two tubes) of testosterone 1% gel per day, applied to clean skin of the upper back/shoulder area. Physicians were free to modify dosing as needed over the course of the study. Follow-up examinations were recommended at 1, 3, 6, and 12 months after the initiation of therapy, though actual follow-up visits were at the discretion of the physician. Visits occurring within a specified time window before or after a suggested time point (visits within 14 days for months 1 and 3, within 30 days for month 6, and visits at day 330 through 14 months for month 12) were grouped within that time point. If a patient had multiple visits within a specific time window, the visit closest to the designated central time point (30, 90, 180, and 360 days) was used.


A physical examination, which was performed at baseline and at follow-up visits (at the discretion of the physician), included anthropometric measurements, vital signs, a broad clinical laboratory assessment, including prostate-specific antigen (PSA) and testosterone levels, a digital rectal examination if deemed appropriate by the physician, and a checklist of concomitant medications for pain, depression, and erectile dysfunction. At baseline, comorbid conditions and previous medical history was also collected [14]. Instructions were given on the use of patient diary cards to record the daily use of TRT; information from the diary cards was used to determine adherence to therapy. Opioid use was recorded as part of the concomitant medications checklist. Clinicians marked a box on the checklist to indicate patient use of opioid medications and listed the specific opioid(s) being used, dose, start and stop dates, and whether use was continuing for each opioid medication listed.

The Brief Male Sexual Function Inventory (BMSFI) [15], an 11-item self-reporting questionnaire measuring sexual function with 10 items used to generate a total composite score, and the Patient Health Questionnaire (PHQ-9) [16], a 9-item self-reporting questionnaire measuring mood/depression, were administered at baseline and at follow-up assessments. The BMSFI total score ranges from 0 to 40, with higher scores denoting better sexual function. The PHQ-9 total score ranges from 0 to 27, with higher scores denoting worse depressive symptoms.

This registry was observational and not designed to collect safety data. However, investigators were instructed to report all serious adverse events, per regulations for an FDA-approved product.

Statistical Analysis

All numerically continuous values are presented as mean ± standard deviation (SD), unless otherwise specified. Observed data from each visit were analyzed. Per-subject changes in parameters from baseline were analyzed using a repeated measures mixed-effects analysis of variance (ANOVA), where the random effect was defined as “patient,” and the two fixed effects were defined as “visit” and “opioid user vs opioid nonuser.” Bivariate correlations were determined using Pearson correlation coefficients. Differences in proportions were tested using Fisher's exact test. Multiple linear regression analyses based on changes from baseline were also computed to determine differences between slopes for opioid users and nonusers. P values ≤0.05 were considered statistically significant. No Bonferroni correction was made for simultaneous multiple comparisons. All analyses were performed using SAS® Version 9.1 (SAS Institute, Cary, NC).


Baseline Characteristics

This analysis includes data from 90 participants with continuing opioid use at baseline, drawn from the total study sample of 849 participants from the TRiUS registry. Exactly one opioid was used at baseline for each opioid user, and these were: hydrocodone (N = 41, 46%), oxycodone (N = 25, 28%), morphine (N = 6, 7%), buprenorphine (N = 4, 4%), codeine (N = 4, 4%), propoxyphene (N = 4, 4%), oxymorphone (N = 2, 2%), tramadol (N = 2, 2%), fentanyl (N = 1, 1%), and methadone (N = 1, 1%). Median duration of opioid use prior to the baseline visit was 293 days, and 75/90 (83%) patients used opioids for more than 30 days.

The mean age of opioid users was 48.3 years, significantly younger (P = 0.002) than the mean age for opioid nonusers (N = 759) of 52.6 years (Table 1). Previous medical history differed between the two groups (Table 1). Opioid users had significantly lower baseline BMI (29.7 vs 31.6, P = 0.01), were less frequently obese (37% vs 54%, P = 0.001), and were less frequently diagnosed with dyslipidemia than nonopioid users (11% vs 21%, P = 0.03). More opioid nonusers were diagnosed with metabolic syndrome (National Cholesterol Education Program criteria [17]) than opioid users (38% vs 26%), although the difference was not statistically significant (P = 0.1). Other comorbid conditions markedly more prevalent in the opioid nonusers were coronary artery disease (19% vs 10%, P = 0.04), diabetes mellitus (13% vs 6%, P = 0.04), and hypertension (21% vs 13%, P = 0.1). However, HIV history was more common to opioid users (18% vs 9%, P = 0.01), as was the use of antidepressants (24% vs 15%, P = 0.02). Baseline levels of total testosterone (TT), PSA, sex hormone-binding globulin (SHBG), and previous testosterone use were not significantly different between opioid users and nonusers, and free testosterone (FT) levels were higher (P = 0.05) for opioid nonusers (Table 1). Individual FT and TT levels at baseline for opioid users are shown in Figure 1.

View this table:
Table 1

Baseline demographic and clinical characteristics

CharacteristicOpioid Users (N = 90)Opioid Nonusers (N = 759)P Value
Age, years (mean ± SD) (N)48.3 ± 12.0 (90)52.6 ± 12.2 (755)0.002
Race % (N)
  Caucasian89 (80)82 (619)0.2
  African American6 (5)9 (67)
  Hispanic2 (2)5 (41)
  Asian02 (18)
  Other/missing race3 (3)2 (14)
Total testosterone, ng/dL mean ± SD (N)280 ± 170 (82)287 ± 149 (642)0.7
Free testosterone, pg/mL mean ± SD (N)27 ± 39 (63)43 ± 65 (399)0.05
SHBG, nmol/L mean ± SD (N)35 ± 16 (15)27 ± 15 (117)0.08
BMI, kg/m2mean ± SD (N)29.7 ± 7.0 (89)31.6 ± 6.8 (757)0.01
  Obese ≥30 kg/m2 % (N)37 (33)54 (406)0.004
Total cholesterol, mg/dL mean ± SD (N)186 ± 57 (29)182 ± 45 (443)0.7
Triglycerides, mg/dL mean ± SD (N)200 ± 121 (30)174 ± 121 (444)0.3
PSA, ng/mL mean ± SD (N)0.94 ± 0.79 (34)1.13 ± 1.13 (430)0.4
Systolic BP, mm Hg mean ± SD (N)129 ± 17 (90)129 ± 15 (755)0.8
Diastolic BP, mm Hg mean ± SD (N)80 ± 9 (89)79 ± 10 (756)0.5
Fasting blood glucose, mg/dL mean ± SD (N)106 ± 48 (37)106 ± 36 (418)0.9
Prior TRT27 (24)24 (179)0.5
Medical history % (N)
  Metabolic syndrome (ATP III definition, baseline visit)26 (15)38 (198)0.11
  Hypertension13 (12)21 (156)0.1
  Dyslipidemia11 (10)21 (159)0.03
  Coronary artery disease10 (9)19 (143)0.04
  Diabetes mellitus6 (5)13 (101)0.04
  HIV18 (16)9 (66)0.01
  Depression10 (9)8 (59)0.4
  Sleep disorders (including apnea)6 (5)4 (33)0.6
Current medication use % (N)
  ED (PDE5 inhibitor use)13 (12)20 (152)0.2
  Antidepressant use24 (22)15 (112)0.02
  • ATP = Adult Treatment Panel of the US National Cholesterol Education Program; BMI = body mass index; BP = blood pressure; ED = erectile dysfunction; PDE5 = phosphodiesterase type 5; PSA = prostate-specific antigen; SD = standard deviation; SHBG = sex hormone-binding globulin; TRT = testosterone replacement therapy; TT = total testosterone.

Figure 1

Individual levels of total testosterone (dots; right-hand axis) and free testosterone (bars; left-hand axis) at baseline among 90 opioid users, grouped by type of opioid used. ANOVA = analysis of variance.

Post TRT

Initial TRT dose was either one tube (5 g gel = 50 mg testosterone) or two tubes (10 g gel = 100 mg testosterone) applied daily. Among opioid users, 59/90 (65.6%) were prescribed 1 tube and 31/90 (34.4%) were prescribed two tubes; this was significantly different than for the opioid nonusers, in which 684/759 (90.1%) were prescribed one tube and 75/759 (9.9%) were prescribed two tubes (P < 0.001). At the last recorded visit, 57/90 (63%) opioid users were prescribed one tube, and 33 (37%) were prescribed two tubes, while 608/759 (80%) opioid nonusers were prescribed one tube and 151/759 (20%) were prescribed two tubes of TRT (P < 0.001). At month 12 (N = 280), the prescribed dose was not significantly different between opioid users (84% prescribed one tube) and opioid nonusers (66% prescribed one tube) (P = 0.1). Adherence to daily TRT as self-reported in a daily diary was similar across both opioid users and nonusers. The adherence rate for opioid nonusers (N = 428) was 92 ± 15% and for opioid users (N = 44) was 93 ± 14% (P = 0.7).

Changes in testosterone levels from baseline to 12 months were similar in both the opioid users and nonusers. At 1 month, both groups had significant increases from baseline in levels of TT (Figure 2) and FT (Figure 3). Changes from baseline at 6 and 12 months were also highly significant for opioid nonusers; changes from baseline for the opioid users were similar but not statistically significant because of the small number of obtained testosterone levels available.

Figure 2

Mean (±standard error) total testosterone at baseline and after 1, 6, and 12 months of testosterone replacement therapy in opioid users and nonusers. Significant increases were observed at all postbaseline time points in opioid nonusers. The ability to detect statistically significant changes among opioid users was limited by the small number of patients for whom postbaseline data were available. P values based on repeated measures mixed-effects ANOVA models. * P ≤ 0.001 vs baseline. † P < 0.05 for difference between changes. ANOVA = analysis of variance.

Figure 3

Mean (±standard error) free testosterone at baseline and after 1, 6, and 12 months of testosterone replacement therapy in opioid users and nonusers. Significant increases were observed at all post-baseline time points in opioid nonusers. The ability to detect statistically significant changes among opioid users was limited by the small number of patients for whom post-baseline data were available. P values based on t-tests (baseline) and repeated measures mixed-effects ANOVA models (postbaseline). * P = 0.05 between groups; † P ≤ 0.001 vs baseline; ‡ P < 0.001 for difference between changes. ANOVA = analysis of variance.

Sexual function, as measured by BMSFI score, showed a similar statistically significant mean improvement for each of the groups after TRT (Figure 4). Baseline BMSFI scores were 27.7 ± 1.38 among opioid users and 28.3 ± 0.40 among nonusers. Both opioid users and nonusers showed a significant improvement from baseline at month 1 (+3.2 ± 0.6, P < 0.001 for nonusers and +3.8 ± 1.8, P = 0.04 for opioid users), month 6 (+4.6 ± 0.7, P < 0.001 for nonusers and +7.6 ± 2.0, P < 0.001 for opioid users), and at month 12 (+6.5 ± 0.6, P < 0.001 for nonusers and +6.7 ± 2.2, P = 0.003 for opioid users). Overall, there was no difference between opioid users and nonusers in BMSFI change at any visit. Multiple linear regression analysis of post-TRT data showed that the change in BMSFI (multiple R = 0.33, P < 0.001) was related to both the change in TT (P < 0.001) and visit time point (P < 0.001), but was not related to opioid use because there were no significant differences in regression slopes (difference: −0.005 ± 0.003, P = 0.1).

Figure 4

Mean (±standard error) Brief Male Sexual Function Inventory (BMSFI) score at baseline and after 1, 6, and 12 months of testosterone replacement therapy in opioid users and nonusers. Higher scores indicate better sexual functioning. The trend lines show similar rates of improvement over time for opioid users and nonusers. P values based on repeated measures mixed-effects ANOVA models. * P < 0.001 vs baseline; † P = 0.04 vs baseline; ‡ P = 0.003 vs baseline. ANOVA = analysis of variance.

Improvements in mean scores on the PHQ-9 depression index were also found (Figure 5). While opioid users had a significantly higher PHQ-9 total score (indicating greater depression symptoms) at baseline than nonusers (11.4 ± 0.78 vs 8.3 ± 0.23; P < 0.001), the PHQ-9 total score significantly decreased at each visit for both groups (Figure 5). There was no significant difference between the two groups in PHQ-9 change at any visit. Multiple linear regression analysis of post-TRT data showed that the change in PHQ-9 (multiple R = 0.35, P < 0.001) was related to both the change in TT (P < 0.001) and visit time point (P < 0.001), and that there was a statistically significant difference in the slope of the change in PHQ-9 score for opioid users compared with nonusers (difference: −0.011 ± 0.004, P = 0.003). As seen in Figure 5, this significant difference in slopes shows that the negative slope of the change in the PHQ-9 was greater for opioid users than for nonusers.

Figure 5

Mean (±standard error) Patient Health Questionnaire-9 score at baseline and after 1, 6, and 12 months of testosterone replacement therapy in opioid users and nonusers. Lower scores indicate fewer depressive symptoms. The trend lines show a greater rate of improvement over time for opioid users compared with nonusers. P values based on repeated measures mixed-effects ANOVA models. * P < 0.001 between groups; † P < 0.001 vs baseline; ‡ P = 0.02 vs baseline; § P = 0.002 vs baseline. ANOVA = analysis of variance.

Opioid use among opioid users at baseline decreased over the course of the study to 74% at month 6, and to 63% at month 12 (among patients who had data at these visits; Table 2), but this may be reflected in the relatively high dropout rate among opioid users; reasons for discontinuation of opioid treatment were not documented. However, other prescription drug use remained relatively stable (Table 2). At baseline, 24% of opioid users also used a selective serotonin reuptake inhibitor (SSRI); this decreased slightly to 21% at the end of the study. Opioid nonusers were less likely than opioid users to use an SSRI at the beginning of the study (15%) and at the end (9%). The use of phosphodiesterase type 5 (PDE5) inhibitors also seemed to remain stable throughout the study, with their use more common among opioid nonusers (20%) at baseline than opioid users (13%).

View this table:
Table 2

Opioid and concomitant medication use

Opioid Users (N = 90)Opioid Nonusers (N = 759)P Value
Opioid use, N/total (%)
  Month 617/23 (74)0/297 (0)
  Month 1212/19 (63)0/265 (0)
PDE5 inhibitor use, N/total (%)
  Baseline12/90 (13)152/759 (20)0.2
  Month 64/23 (17)54/297 (18)1.0
  Month 123/19 (16)49/265 (18)1.0
SSRI use, N/total (%)
  Baseline22/90 (24)112/759 (15)0.02
  Month 65/23 (22)33/297 (11)0.2
  Month 124/19 (21)24/265 (9)0.1
  • PDE5 = phosphodiesterase type 5; SSRI = selective serotonin reuptake inhibitor.

There were no meaningful differences in anthropometric changes between the two groups, including body mass index (BMI) and waist–hip ratio, and no differences in changes in blood pressure, glucose, or other clinical chemistry levels (data not shown). While opioid users tended to have slightly lower values for both BMI and waist–hip ratio at baseline, there was almost no change from baseline in either group over the course of the 12-month treatment period (Table 3).

View this table:
Table 3

Anthropometric outcomes

BMI kg/m2Waist-to-Hip Ratio
BaselineMonth 12BaselineMonth 12
Visit ValueVisit ValueChange from BaselineVisit ValueVisit ValueChange from Baseline
Opioid nonusersMean ± SE31.6 ± 0.2531.6 ± 0.40−0.23 ± 0.080.98 ± 0.030.98 ± 0.004−0.01 ± 0.005
P value10.0040.1
P value20.10.2
Opioid usersMean ± SE29.7 ± 0.7427.6 ± 1.640.24 ± 0.290.96 ± 0.020.97 ± 0.02−0.03 ± 0.02
P value10.40.09
  • Change from baseline analyses based on least squares mean, standard error, and P values from repeated measures mixed-effects ANOVA models.

  • ANOVA = analysis of variance; BMI = body mass index. P value1 = within-group change from baseline; P value2 = between opioid users and opioid nonusers; SE = standard error.

The registry was observational and not designed to collect safety data, although investigators were instructed to report all serious adverse events. The adverse events investigators reported included skin rash (N = 6), itching (N = 2), redness (N = 1), skin allergy (N = 1), prostate cancer (N = 1), increased lipid panel (N = 1), and elevated PSA (N = 1). One patient died following surgery for an arteriovenous malformation; the death was judged not related to TRT.


This is the first report of long-term TRT outcomes among a sample of hypogonadal men using opioids compared with hypogonadal men not using opioids. Hypogonadal opioid users and nonusers had similar levels of TT, FT, SHBG, and PSA at the baseline visit. The opioid users were 4 years younger, and there were a number of statistically significant differences at baseline between the two groups that could not be explained by the age difference. Prevalence of CAD, dyslipidemia, diabetes, obesity, hypertension, and the metabolic syndrome were close to, or even more than, double among opioid nonusers vs opioid users at baseline. This raises the question of how hypogonadal opioid users could be healthier and have less cardiovascular risk factors than hypogonadal opioid nonusers of a similar age and ethnic distribution. We suspect that the opioid users were hypogonadal because of the opioid use, where the majority (83%) of opioid users had been on opioids for ≥30 days prior to baseline. Thus, we speculate that opioid use significantly lowered testosterone levels, but did not have any opportunity to adversely affect long-term health and cardiovascular risk factors. The long-term profile of hypogonadism is strongly associated with increases in cardiovascular risk factors. In our study, low testosterone caused by opioid use was not associated with these cardiovascular risk factors over the 12 months of this study. An extensive review of animal and human research studies on the association of opioid effects on food intake, obesity, diabetes, and the metabolic syndrome is available [18].

We estimate from our registry cohort that 10.6% (95% exact CI, 8.6–12.9%) of hypogonadal men are hypogonadal because of opioid use. Therefore, clinicians should be mindful of the potential for opioid-induced hypogonadism, as it appears to be a common effect of opioid use [1]. Among a group of male patients receiving intrathecal morphine for chronic nonmalignant pain, low levels of serum testosterone and pituitary gonadotropins indicated hypogonadism in all of the patients [19,20]. In another study of 12 men treated with oral opioids for at least 1 year for chronic noncancer pain, 75% demonstrated hypogonadism, and osteopenia was found in half of the subjects [21]. Significant hypogonadism and sexual dysfunction were demonstrated in male cancer survivors using high-dose oral opioids for cancer-related chronic pain [22,23]. Methadone-treated addicts are well known to have hypogonadism and a high prevalence of erectile dysfunction associated with hypogonadism and depression [24]. These observations have resulted in the recommendation that testosterone be monitored during opioid treatment due to the effect of opioids on testosterone production [9].

Our analysis showed that for opioid users with low testosterone, both TT and FT levels can be significantly increased with the use of TRT over 12 months. This is particularly important considering recent findings that morphine administration increased 5-alpha reductase mRNA expression in the liver and aromatase mRNA expression in the brain and gonads of male rats [25], suggesting that opioids may lead to lower plasma and brain testosterone levels by increasing testosterone metabolism, as well as by their effects on gonadotropin secretion. Thus, exogenous supplementation of testosterone that increases FT throughout the body may be important in maintaining normal levels of mRNA expression and function for the hypogonadal male opioid user.

Sexual function at baseline and after TRT was measured at 1, 6, and 12 months using the BMSFI questionnaire. Baseline sexual function was not different in opioid users or nonusers (P = 0.6). However, changes from baseline for the total BMSFI score in both groups were statistically significant at months 1, 6, and 12 (Figure 4). There was no difference in changes between opioid users and nonusers at any visit and no significant interaction terms. The multiple linear regression analysis identified that the change in BMSFI was a linear function of the change in TT and visit. This demonstrates that TRT was equally effective for the improvement of sexual function in both opioid users and nonusers alike.

Opioid users had significantly higher baseline PHQ-9 scores (Figure 5) and had significantly higher use of antidepressants (Table 1). While both groups showed significant decreases in depressive symptoms at each visit, there was a significantly greater decrease in PHQ-9 scores over time for the opioid users (Figure 5). PHQ-9 scores remained slightly higher among opioids users compared with nonusers at 12 months, although there were no statistically significant differences in PHQ-9 scores at either 6 or 12 months. This suggests an improvement in mood symptoms with TRT use among opioid users, and that this improvement begins early and continues in a general linear fashion for at least 12 months. The improvement in PHQ-9 scores among opioid nonusers was also evident, but at a significantly slower rate (Figure 5). It is not known from the present study how long the consistent linear decrease would continue and if the PHQ-9 scores among the two groups would reach the exact same level.

Improvements in sexual function and depression in hypogonadal men receiving TRT are not unexpected, as they have been observed in previous short-term controlled studies [26,27] and long-term open-label extension studies [28]. Thus, the results from the opioid nonusers in this sample are consistent with the findings of earlier studies. However, this analysis further suggests that these improvements are also observed specifically in hypogonadal men who are opioid users.

Despite the understanding of the effects of opioids on testosterone, there have been few studies of TRT that specifically evaluated outcomes in the context of chronic opioid use. In one study, 16 men with OPIAD who completed 24 weeks of treatment with testosterone patch therapy experienced improvements in mood/depression and sexual function outcomes [12]. More recently, a study of 17 men with OPIAD and who were receiving epidural morphine for chronic noncancer pain found that TRT (testosterone gel) for up to 12 months was associated with improved sexual function [29]. However, these studies had a limited number of patients, and because they did not include any patients who were not opioid users, potential differences in outcomes between opioid users and nonusers could not be evaluated. Although specific measures differed among the various trials and our study did not evaluate men with OPIAD specifically, our analysis provides additional support for the effectiveness of TRT in opioid users in improving outcomes, such as sexual function and mood. Moreover, the results suggest that the degree of improvement is not adversely affected by opioid use.

There are several limitations of this study. Registry participants were not equally split between opioid users and nonusers. Thus with the low number of opioid users who volunteered for the study, statistical significance could not be reached for several of the outcomes. This was not a randomized clinical trial but was a prospective observational registry examining outcomes in patients observed in clinical settings. Due to the nature of the study, it is not known whether the use of opioids caused the hypogonadism among opioid users. In addition, long-term studies are generally associated with high rates of noncompletion. Of the 849 patients in the overall TRiUS registry, 305 (36%) completed the entire study; of the 545 that did not complete, 145 (17% of 849) had a study duration of 0–29 days, 94 (11%) had a study duration of 30–89 days, 111 (13%) had a study duration of 90–179 days and 195 (23%) had a study duration of 180 days or more. Other factors about the participants in the study, such as concurrent medications or medical conditions, duration of opioid use, or differences in the pain management protocol used for each patient, could have influenced the outcomes; observed improvements in BMSFI and PHQ-9 scores and the reduction in opioid use over the course of the study, in other words, could be due in part to these factors. Finally, due to the fact that this was a registry study and all clinical visits and laboratory assessments were at the discretion of the treating physician, not all patients contributed data for each of the month 1, 6, and 12 visits.


Even with the limitations noted above, the results from this analysis suggest that TRT has as significant an effect on testosterone levels among opioid users with hypogonadism as opioid nonusers with hypogonadism. This effectiveness does not appear to be mitigated by any safety issues. With TRT use, two key quality-of-life measures, sexual function and mood, also improved in both participant groups. Thus, this study suggests that for hypogonadal men who are opioid users, TRT may provide benefits that are at least as favorable as those seen in hypogonadal men who are not taking opioids.


The authors thank Sherri Jones, PharmD of MedVal Scientific Information Services, LLC for providing medical writing and editorial assistance.


  • Author Disclosures:

  • GB—Research funding: Auxilium, Boehringer Ingelheim, Gilead, Pfizer, Sangamo, and ViiV. Advisor/consultant: Auxilium, Bristol-Myers Squibb, Merck, Tibotec, and ViiV. Speaker: Abbott, Auxilium, Boehringer Ingelheim, Bristol-Myers Squibb, Merck, Tibotec, and ViiV.

  • MK—Research funding: Allergan. Speaker: Auxilium and Slate.

  • RKB—Advisor/consultant: Auxilium. Speaker: Abbott, Bristol-Myers Squibb, Novartis, and sanofi-aventis.

  • DN—Employee: Auxilium.

  • HK—Employee: Auxilium.

  • MMM—Board membership: Sexual Medicine Society of North America. Research funding: Auxilium and GlaxoSmithKline. Advisor/consultant: Abbott, Auxilium, and Endo.

  • Funding: Funding to support this study and the preparation of this manuscript was provided by Auxilium Pharmaceuticals. This manuscript was prepared according to the International Society for Medical Publication Professionals' “Good Publication Practice for Communicating Company-Sponsored Medical Research: The GPP2 Guidelines.”


View Abstract