Patterns of sexual and erectile dysfunction and response to treatment in patients receiving androgen deprivation therapy for prostate cancer


Ithaar H. Derweesh, Department of Urology, University of Tennessee Health Science Center, Court Avenue Rm H210, Memphis, TN 38163, USA.



To investigate the incidence of patient-reported erectile (ED) and sexual dysfunction and response to treatment in men after the induction of androgen deprivation therapy (ADT) for prostate cancer, as ADT-induced changes in serum testosterone can result in changes in libido and sexual function.


We retrospectively reviewed patients receiving ADT for prostate cancer at our institution between January 1989 and July 2005; those receiving only neoadjuvant ADT were excluded. Variables included age, race, body mass index, prostate-specific antigen level before ADT, Gleason sum, clinical stage, ADT type (medical vs surgical) and schedule (continuous vs intermittent), previous treatment for prostate cancer, presence of pre-existing or new-onset diabetes mellitus (DM), and presence of ED before ADT. After ADT induction, charts were reviewed for reporting of ED, changes in libido, and initiation of ED therapy (medical or surgical).


In all, 395 patients (mean age of 71.7 years; 59.0% African-American, 41.0% Caucasian/other, at initiation ADT) were analysed. At mean follow-up of 87.4 months, 57 (14.4%) patients reported ED; 40 of these (70%) reported new-onset ED, while 17 (30%) reported ED before ADT. Response rates were 33–80% with medical therapy, including 44% receiving phosphodiesterase-5 inhibitor monotherapy. On multivariate analysis, age <70 years (P < 0.001) and the absence of DM (P = 0.024) were associated with reporting ED after ADT.


Patients receiving ADT for prostate cancer have variable degrees of ED. Successful outcomes are possible, particularly when implementing multimodal therapy. Younger patients and those with no DM are more likely to report ED after ADT induction.


androgen-deprivation therapy


body mass index


external-beam radiotherapy


radical retropubic prostatectomy


targeted cryoablation of the prostate


(new-onset) diabetes mellitus


erectile dysfunction


phosphodiesterase type 5 inhibitor


prostaglandin E1 analogue


vacuum erection device


inflatable penile prosthesis.


With an expected 218 890 new cases and 27 050 deaths estimated in 2007, prostate cancer is the most common malignancy in men in the USA [1]. Since the discovery of the androgen dependence of prostate cancer, androgen-deprivation therapy (ADT) has become widely used for treating men with clinically localized or advanced-stage disease, and has served as both primary and salvage therapy in selected patient populations [2].

Adverse effects of ADT, including decreased libido, osteoporosis, vasomotor flushing, fatigue, anaemia, diabetes mellitus (DM), metabolic syndrome, and altered body composition are widely reported [2–9] but despite the induction of castrate testosterone levels with ADT and the potential for loss of libido and resulting erectile dysfunction (ED), we recognized a subset of patients in clinical practice that maintain erectile and sexual function, prompting further investigation into this subset. Furthermore, while these adverse effects of ADT are widely referenced [9–12], there is indeed a paucity of data analysing these endpoints, particularly for the response to ED-targeted therapy [13,14]. Thus we report our investigation into the patterns of reporting of sexual and ED and the response to treatment in men receiving ADT for prostate cancer.


After obtaining institutional review board approval, we retrospectively reviewed all patients receiving ADT for prostate cancer treatment at one centre at our institution (Veterans Affairs Medical Center, Memphis, Tennessee, USA) between January 1989 and June 2005. Patients receiving only neoadjuvant ADT were excluded. Clinicopathological variables examined included age at ADT initiation, race, body mass index (BMI), serum PSA level before ADT, Gleason grade, American Joint Committee on Cancer 1992 clinical stage, type of ADT (medical vs surgical), receipt of ADT as primary or salvage therapy, ADT schedule (continuous vs intermittent), and presence of DM before ADT, and development of new-onset DM (NODM) after ADT induction. Patients undergoing ADT were diagnosed with NODM according to WHO/American Diabetes Association recommendations with two separate fasting blood glucose measures of ≥126 mg/dL [15]. Patients with a diagnosis of DM and undergoing targeted therapy (diet, insulin therapy, oral therapy) before starting ADT were designated as having DM before ADT.

For sexual/ED, patient charts were reviewed for the presence of pre-existing ED and targeted therapy, and reports of libido status, complaints of ED and initiation of ED therapy after ADT. ED therapy included: phosphodiesterase type-5 inhibition (PDE5i), prostaglandin E1 analogues (PGE1), vacuum erection device (VED), inflatable penile prosthesis (IPP), and combined therapy. The success of ED therapy was defined as an erection sufficient for penetration and successful completion of sexual intercourse. ADT was defined as the receipt of a GnRH agonist (goserelin acetate depot, AstraZeneca PLC, London, UK), combined androgen blockade (GnRH agonist and antiandrogen), or bilateral orchidectomy.

Data were analysed by chi-square and Kruskal–Wallis anova (where appropriate), and univariate and multivariate logistic regression, with all potential explanatory covariates incorporated into models. Variables showing an association with ED were considered for multivariate analysis. Independent variables were modelled as both continuous and categorical variables, as follows: age ≥70 vs <70 years, Gleason grade sum ≥7 vs <7, PSA level ≥10 vs <10 ng/mL, and BMI ≥30 vs <30 kg/m2. All P values were based on two-sided tests of significance, with P < 0.05 considered to indicate statistical significance. The Hosmer-Lemeshow test eliminated models that fitted poorly.


Demographic data and disease characteristics for all patients are outlined in Table 1; after exclusions, 395 patients were analysed (mean age at start of ADT 71.7 years, range 46.7–89.3) and with a mean (range) follow-up of 87.4 (4.8–445.0) months. Of these, 233 (59.0%) were African-American while 162 (41.0%) were Caucasian/other. In all, 375 (94.9%) men had ADT by medical castration while 20 (5.1%) had bilateral orchidectomy. In all, 264 (66.8%) men had salvage ADT for an increasing PSA level after primary treatment, while 131 (33.2%) received primary and 359 (90.9%) received continuous ADT (surgical or medical), with 36 (9.1%) receiving intermittent ADT. In all, 113 patients (29.2%) developed NODM or had DM before ADT. Only 10 of the 395 patients (2.5%) reported a ‘normal’ libido after ADT.

Table 1.  Clinicopathological and demographic data for 395 patients with prostate cancer treated with ADT
VariableMean, median (range) or n (%)
Age at ADT initiation, years 71.7, 73.1 (46.7–89.3)
Serum PSA level, ng/mL before ADT130.8, 15.4 (0.4–6031.0)
Gleason grade sum  6.9, 7.0 (3.0–10.0)
Primary Gleason grade  3.4, 3.0 (2.0–5.0)
Secondary Gleason grade  3.5, 3.0 (1.0–5.0)
BMI, kg/m2 26.9, 26.7 (12.5–52.4)
 African-American233 (59.0)
 Caucasian/other162 (41.0)
Castration type
 Medical375 (94.9)
 Surgical 20 (5.1)
ADT type
 Primary131 (33.2)
 Salvage264 (66.8)
EBRT/brachytherapy199 (50.3)
RRP 32 (8.1)
TCAP 13 (3.3)
EBRT/TCAP  2 (0.5)
RRP/EBRT 18 (4.6)
ADT administration schedule
 Continuous359 (90.9)
 Intermittent 36 (9.1)
Clinical stage
 T1a/b 15 (3.8)
 T1c163 (41.1)
 T2a 42 (10.6)
 T2b/c 22 (5.6)
 T3a/b/c  7 (1.8)
 N+  8 (2.0)
 M+ 14 (3.5)
 Unknown125 (31.6)
 Before ADT 77 (19.4)
 NODM 36 (9.8)
 No DM282 (71.4)
Follow-up, months 87.4, 80.5 (4.8–445.0)

The clinicopathological and demographic data for the 57 patients reporting ED after ADT are shown in Table 2; of these 57, 40 (70%) reported new-onset ED, while 17 (30%) had reported ED before ADT. Fifty-one (90%) received continuous ADT and six (11%) intermittent ADT. Of the 57 men, 25 (44%) received primary ADT and 32 (56%) salvage ADT. In this subset, rates of reported ED before ADT were significantly higher in men receiving salvage ADT (13 of 32) than in those receiving primary ADT (three of 25, P = 0.04). The primary and salvage ADT groups were similar for the other clinicopathological variables. In particular, age at ADT initiation (P = 0.07) and rates of DM or NODM before ADT were similar between these groups (P = 0.62).

Table 2.  Clinicopathological and demographic data for patients with ED after ADT
Mean, median (range) or n (%) variableOverall ADT-EDPrimary ADTSalvage ADTP
No. of men 57 25 32
Age at ADT initiation, years 67.4, 67.4 (47.8–83.5) 69.7, 69.6 (55.1–72.9) 65.5, 65.0 (47.8–80.0)0.07
PSA level before ADT, ng/mL146.5, 12.5 (0.5–3600)123.4, 14.9 (4.7–1688)165.8, 9.9 (0.55–3600)0.82
Gleason grade sum  7.3, 7.0 (5.0–10.0)  7.5, 7.5 (5–9)  7.1, 7.0 (5–10)0.38
BMI, kg/m2 27.8, 27.6 (15.3–38.9) 26.3, 26.1 (15.3–35.7) 27.3, 26.9 (16.8–38.9)0.91
 African-American 37 (65) 16 (64) 21 (66)0.9
 Caucasian/other 20 (35)  9 (36)  11 (34) 
ADT Schedule
 Continuous 51 (90) 22 (88) 29 (91)0.75
 Intermittent  6 (11)  3 (12)  3 (9) 
Before ADT  9 (16)  5 (20)  4 (13)0.62
NODM 10 (18)  3 (12)  7 (22) 
Non-DM 38 (67) 17 (68) 21 (66) 
ED before or after ADT
 De novo ED 40 (70) 22 (88) 19 (59)0.016
 Before ADT 17 (30)  3 (12) 13 (41) 

Table 3 outlines the patient response to ED-directed treatment; 41 (72%) received PDE5i only, six (11%) received PGE1 analogues only (urethral suppository or intracavernosal injection), four (7%) received VED only, one (2%) received IPP, and five (9%) received combined therapy. The patient receiving IPP had failed all medical therapy after retropubic radical prostatectomy (RP) and salvage external beam radiotherapy (EBRT) for biochemical recurrence. For the overall treatment response, success was reported in 27 (of 57, 47%) patients; 18 of 41 (44%) receiving PDE5i only, two of six receiving PGE1 analogues only, two of four receiving VED only, one man receiving IPP, and four of five receiving combined medical therapy. When analysing only non-surgical treatment responses, 26 patients reported success overall, i.e. 13 (50%) patients receiving primary ADT and 13 (50%) receiving salvage ADT. Efficacy rates for non-surgical therapy were similar between primary ADT and salvage ADT groups (P = 0.8). Also, there were no differences in efficacy between primary and salvage groups for the type of non-surgical therapy for ED (Table 3).

Table 3.  Targeted therapy and success in men with ED treated with ADT
Treatment for ED after ADTN (%) receiving ED treatmentn (%) or n reporting treatment successP
OverallPrimary ADTSalvage ADT
ED after ADT57 (14)27 (47)13 (48)14 (52)0.54
PDE5i41 (72)18 (44) 8 (44)10 (56)0.95
PGE-1 6 (11) 2 1 10.64
VED 4 (7) 2 2 00.26
IPP 1 (2) 1 0 10.91
Combination 5 (9) 4 2 20.60
 PDE5i/PGE-1 1 (2) 1 0 1
 PDE5i/PGE-1/VED 2 (4) 2 1 1
 PDE5i/VED 1 (2) 1 1 0
 PGE-1/VED 1 (2) 0 0 0

Logistic regression analysis was used to identify factors predictive for or protective against reporting ED after ADT. Variables incorporated into the models included age, BMI, serum PSA before ADT, race, castration schedule, and presence of DM after ADT (data not shown). When adjusting for other variables, multivariate analysis showed that age <70 years (odds ratio 3.75, P < 0.001) and the absence of DM (3.89, P = 0.024) were associated with a significantly greater probability of reporting ED after ADT when adjusting for other variables.


ADT remains a widely used treatment for selected patients with localized or advanced prostate cancer. Distinct survival benefits have been recognized with ADT when used in patients in the form of neoadjuvant or adjuvant therapy together with EBRT and as early adjuvant therapy in patients found to have lymph node metastases at the time of RP [2]. In addition to its use in locally advanced and metastatic prostate cancer, there have been significant increases in the use of primary ADT [16]. The role of ADT in the treatment options for contemporary prostate cancer has led to much interest in recognizing and preventing ADT-associated morbidity.

GnRH agonists induce a reduction in circulating serum testosterone and oestradiol levels through down-regulation of the LH receptors in the anterior pituitary gland. Subsequently, there is a resultant decrease in the release of circulating sex hormones by the gonads [2]. Castrate levels of serum testosterone, due to either age-related changes or ADT, have been associated with altered body composition, typically resulting in the loss of lean body mass with increases in the percentage body fat. These changes have shown increases in the susceptibility for developing metabolic syndrome, loss of libido and ED [4–14,17].

In a survey-based study using Medicare Provider and Analysis and Review files, Fowler et al.[14] compared health-related quality-of-life outcomes in androgen-deprived (298) and non-androgen-deprived men (1095) after RP. In all, 166 men in the androgen-deprived group and 886 men in the other group responded to the survey questions on ED. Androgen-deprived patients had higher rates of ED after RP overall (72% vs 55%), but similar rates of ED in the month before the survey (23% vs. 22%). When asked about the quality of erections, 3% (vs 11%) of androgen-deprived men reported erections not firm enough for intercourse, with only 2% (vs 12%; P < 0.001) reporting erections firm enough for intercourse. For libido, 170 androgen-deprived and 888 non-androgen-deprived men responded to targeted survey questions. Of these, 69% of androgen-deprived men reported no sexual drive over the 30 days before the survey, compared to 29% in the other group (P < 0.001).

While we did not compare patients receiving ADT to non-androgen-deprived men, the 57 men reporting ED after ADT had similar percentages of primary (44%) and salvage (56%) patients. Overall, the series showed high rates of altered libido, with only 10 (of 395, 2.5%) patients reporting a normal libido. Despite this, patients undergoing either primary or salvage ADT had similar overall response rates to ED-directed treatments (48% vs 52%, P = 0.54, Table 3). Further, both primary and salvage ADT had similar response rates to PDE5i monotherapy (44% vs 56%, P = 0.95, Table 3).

Potosky et al.[17] analysed sexual and erectile outcomes in men receiving only either orchidectomy (132) or GnRH agonists (299) enrolled in the Prostate Cancer Outcomes Study. In that study, while pretreatment function and libido were similar in both groups, patients treated with orchidectomy reported slightly worse overall sexual function, although the differences were not statistically significant. Furthermore, changes in sexual outcomes were similar between both ADT groups. Reports of no erections increased from 35% to 78.6% in the orchidectomy group and from 37.9% to 73.3% in the GnRH group. However, 38.4% of men receiving GnRH agonist therapy reported a big/moderate problem in overall sexual function, compared to 25.6% in the orchidectomy group (P = 0.04). However, this difference was attributed to a greater percentage of patients on GnRH reporting baseline ‘big/moderate’ problems with sexual function before ADT.

While altered sexual and ED are reported there have been no reports to our knowledge examining the response to treatments in this selected patient population. In the present series of 395 men receiving ADT for prostate cancer, 57 (14.4%) reported ED after ADT and received targeted therapy. Of these, 40 (70%) had new-onset ED, while 17 (30%) had a previous history of ED before ADT induction (Table 2). Despite ADT, an overall success rate of 47% was reported for all forms of ED therapy, with a rate of 44% reported for PDE5i monotherapy (Table 3). Only one patient had an IPP placed for failed medical therapy. Therefore, we have insufficient data to report surgical outcomes in this patient population. However, the success rates with medical management were similar between primary and salvage ADT groups (Table 3).

We attempted to identify factors that would predict the reporting of ED after ADT in this population. Being younger (<70 years) and the absence of DM had a significant association with reporting ED after ADT. This is consistent with several reports showing both age and DM to be risk factors for ED. Aversa et al.[18] reported an epidemiological study evaluating the awareness of ED among Italian men. Of 16 007 men responding to the study, 41% of those with severe ED had delayed investigation and treatment for ≥3 years and only 19% had requested and tried ED-targeted medications. Thus the authors concluded that despite the widespread access and advertisement of PDE5i therapy, ED is still widely under-reported and under-diagnosed.

In the present series, age and rates of DM were similar between men having primary or salvage ADT and reporting successful medical treatment (P = 0.07 and 0.62, respectively, Table 2). Furthermore, rates of overall (P = 0.8) and non-surgical treatment success (P = 0.54) were similar between these groups (Table 3). While this group was small, our findings, as they relate to DM, are consistent with previous reports of the efficacy of ED-directed medical therapy in patients with DM [19].

There are several limitations to the present study. First, we report a retrospective review of our findings at one centre, and as such our findings are subject to the inherent biases of this type of analysis. Furthermore, we did not attempt to compare these patients with a control group (i.e. patients undergoing surveillance protocols) and therefore report on one group of patients. Also, while comprising 395 patients overall, few patients reported ED after ADT (57), but this is consistent with previous studies documenting under-reporting of ED in the general population [20]. As such, our analysis might be limited in its ability to detect all potential relationships between variables and the study endpoints. Furthermore, patients were not evaluated in a uniform and prospective manner, and were not routinely evaluated using validated instruments related to ED and sexual health. Therefore, we suspect that it is possible that more patients developed ED and went undetected. We also did not attempt to compare surgical vs medical ADT and their respective impact on sexual and erectile function, because of the disparity between the numbers in the two groups. Furthermore, we did not analyse serum testosterone levels in patients who reported or responded to ED-targeted therapy.

Thus, it is uncertain whether these study limitations potentially underestimate the effects of long-term ADT on erectile and sexual function. Nonetheless, we feel the results of this investigation are strengthened by the size of the study cohort and the duration of follow-up. To our knowledge, this is the largest single-centre series reported to date of patients evaluated for ED and response to targeted therapy in those receiving ADT for prostate cancer. The mean follow-up of 87.4 months at an equal-access healthcare centre is considerable [21,22]. Age and prostate cancer disease characteristics in the study cohort appear to be similar to the larger USA prostate cancer population, while race, lifestyle, and comorbidities might be related to regional and patient subpopulation demographics [23]. Ideally, a prospective analysis using validated ED-directed instruments with a direct comparison to matched controls would be ideal to detect all potential relationships.

Despite these limitations, we feel that these findings should at least prompt physicians who are administering ADT to patients with prostate cancer to question patients (particularly those aged <70 years), and to adopt a regimen of close, serial monitoring of ED and response (or lack thereof) to targeted therapy. These findings have prompted us to investigate changes in erectile and sexual function using validated questionnaires, implemented in a prospective fashion, as part of a surveillance plan for patients treated at our institution.

In conclusion, men receiving ADT for prostate cancer report variable degrees of sexual and/or ED. This study showed ED after ADT with initiation of targeted therapy in 14.4% of men receiving ADT. Patients were more likely to report ED if they were younger (<70 years) and had no evidence of DM. Treatment can be successful in this population, particularly with multimodal therapy, depending on the type of intervention, and does not appear to be affected by previous therapy for prostate cancer. While further investigation is required, these data support serial monitoring of sexual and ED, and physician-directed attempts at restoring erectile function in these patients.


None declared.