Introduction. Macclesfield, Cheshire, UK; and 4 AstraZeneca, Wilmington, Deleware, USA

Transcription

1 Prevention and management of bicalutamide-induced gynecomastia and breast pain: randomized endocrinologic and clinical studies with tamoxifen and anastrozole D Saltzstein 1 *, P Sieber 2, T Morris 3 & J Gallo 4 1 Urology San Antonio Research PA, Pasteur Medical Plaza, San Antonio, Texas, USA; 2 Urological Associates of Lancaster, Lancaster, Pennsylvania, USA; 3 AstraZeneca, Macclesfield, Cheshire, UK; and 4 AstraZeneca, Wilmington, Deleware, USA (2005) 8, & 2005 Nature Publishing Group All rights reserved /05 $ A randomized, double-blind, placebo-controlled multicenter trial involving 107 men receiving bicalutamide ( Casodex ) 150 mg/day therapy following radical therapy for prostate cancer assessed tamoxifen ( Nolvadex ) 20 mg/day and anastrozole ( Arimidex ) 1 mg/day for the prophylaxis and treatment of gynecomastia/breast pain. Tamoxifen, but not anastrozole, significantly reduced the incidence of gynecomastia/breast pain when used prophylactically and therapeutically. Serum testosterone levels increased with tamoxifen relative to placebo but prostate-specific antigen levels declined in all treatment groups. Further studies are needed to define the optimum tamoxifen dose and to assess any impact on cancer control. The use of tamoxifen in this setting remains to be investigated. (2005) 8, doi: /sj.pcan Published online 1 February 2005 Keywords: tamoxifen; anastrozole; bicalutamide; gynecomastia; breast pain Introduction The nonsteroidal antiandrogen bicalutamide ( Casodex ) is being evaluated in the ongoing bicalutamide Early Prostate Cancer (EPC) program as immediate monotherapy or as an adjuvant to radical therapy, and has been found to provide benefits in patients with locally advanced prostate cancer. 1 Gynecomastia and breast pain are the main side effects of bicalutamide monotherapy, 1,2 as predicted from the pharmacologic profile of nonsteroidal antiandrogens (Figure 1). Gynecomastia is caused by an increase in estrogenic activity relative to androgenic activity at the breast tissue 3 *Correspondence: D Saltzstein, Urology San Antonio Research PA, Pasteur Medical Plaza, 7909 Fredericksburg Drive, Suite 115, San Antonio, TX 78229, USA. dsaltzstei@aol.com Arimidex, Casodex, and Nolvadex are trademarks of the AstraZeneca group of companies. Received 23 July 2004; revised 1 November 2004; accepted 8 December 2004; published online 1 February 2005 and after nonsteroidal antiandrogen administration, increased availability of testosterone leads to increased aromatization and hence an increase in circulating estrogen levels. 4 Additionally, the action of estrogen in the breast is unopposed due to the blockade of the androgen receptor by the nonsteroidal antiandrogen. Although bicalutamideinduced gynecomastia and breast pain are rarely severe, 1 evidence from the EPC program suggests that the condition may be more troublesome to men who receive bicalutamide as an adjuvant to radical therapy (radical prostatectomy or radiotherapy) for early disease than those treated with immediate hormone monotherapy. 1,2 Thus, for patients with early disease to benefit fully from bicalutamide 150 mg, management options (both prophylactic and therapeutic) are needed for bicalutamide-induced gynecomastia and breast pain. Three forms of therapy have been used to prevent or alleviate gynecomastia and/or breast pain of other etiologies: low-dose irradiation, surgery, and hormonal therapies. 3 A recent randomized study has demonstrated that a single dose of prophylactic irradiation (10 Gy) is

2 76 Negative feedback Hypothalamus Bicalutamide LHRH LH FSH Testes Pituitary ACTH Adrenal glands and tamoxifen or anastrozole in the pilot study. Therefore, we were able to proceed to the clinical trial. Here we report the design, methodology and findings of the clinical trial, as well as summarizing the hormonal changes seen in the pilot study. Breast tissue + Bicalutamide Estrogens Testosterone + DHT Aromatization Figure 1 Rationale for the clinical trial showing where the nonsteroidal antiandrogen bicalutamide acts on the hypothalamic pituitary gonadal hormonal pathways in males (adapted with kind permission 3 ). In untreated men, there is a balance between the stimulatory effects of estrogens ( þ ) and the inhibitory effects of androgens ( ) at the breast. Under nonsteroidal antiandrogen treatment, the action of androgens at the breast is blocked, while estrogenic activity rises due to increased aromatization of circulating testosterone caused by the inhibition of the negative feedback mechanism, thus the stimulatory effects of estrogens predominate. Antiestrogens block the activity of estrogens at the breast, while aromatase inhibitors will inhibit the conversion of testosterone to estrogen by aromatase. Both types of hormonal therapy therefore decrease estrogenic activity at the breast. effective in reducing the incidence of gynecomastia in bicalutamide-treated patients. 5 Similar results were reported from a recent Scandinavian study in which 69% of patients received prophyalctic irradiation for gynecomastia induced by the nonsteroidal antiandrogen flutamide. 6 Hormonal therapies that reduce estrogen activity at the breast (Figure 1), such as the aromatase inhibitors, which block the peripheral aromatization of circulating androgens to estrogens, and the antiestrogens, which are antagonists at the estrogen receptors, should also reduce the tendency for gynecomastia to occur in the presence of androgen blockade. Aromatase inhibitors and antiestrogens have both been used to treat established gynecomastia of other etiologies, with some or good efficacy Based on the above rationale, a clinical trial assessing the efficacy and tolerability of the antiestrogen tamoxifen ( Nolvadex ) 20 mg and the aromatase inhibitor anastrozole ( Arimidex ) 1 mg both as prophylaxis and treatment for gynecomastia and breast pain in non-castrate patients receiving after radical therapy for early prostate cancer was planned. Prior to initiating the clinical trial, a pilot study was undertaken to assess the effects of plus tamoxifen 20 mg and plus anastrozole 1 mg on free testosterone concentrations in healthy elderly men. This pilot study was carried out as published data 13,14 suggested that, by influencing the negative feedback effects of estradiol on the hypothalamic pituitary-axis, the addition of aromatase inhibitors and antiestrogens to therapy could increase serum testosterone to levels above those induced by alone. These higher serum testosterone levels might potentially have reduced the bicalutamide-mediated androgen-receptor blockade and, consequently, the antitumor activity of bicalutamide. However, there was no evidence of such a pharmacodynamic interaction between bicalutamide Methods Pilot study Healthy elderly male volunteers (n ¼ 14) were randomized 1:1 to either once daily for 6 weeks plus anastrozole 1 mg once daily for the final 2 weeks, or once daily plus tamoxifen 20 mg once daily for 6 weeks. Sex hormone (testosterone, dihydrotestosterone, luteinizing hormone (LH), sex hormone binding globulin (SHBG), prolactin, folliclestimulating hormone (FSH) and estradiol) serum levels were assessed periodically during the dosing period and at a follow-up visit at days. Clinical trial Patients Eligible patients had histologically confirmed localized or locally advanced (T1 4, M0, any N) prostate cancer, had undergone radical therapy (radical prostatectomy, radiotherapy) and were noncastrate. Absence of metastatic disease was confirmed by bone scan within 30 weeks of randomization. All patients had a life expectancy of at least 1 y. Patients were excluded from the trial if they had gynecomastia X2 cm in diameter or breast discomfort, had previously received hormonal therapy (except neoadjuvant hormonal therapy X6 months previously), or were currently receiving or had received in the past 6 months any medication known to cause gynecomastia. Further exclusion criteria were evidence of alcohol abuse, abnormal liver function tests, elevated estradiol levels, or decreased age-adjusted testosterone levels. All eligible patients gave written informed consent before trial entry. Trial design This was a randomized, three-arm, doubleblind, placebo-controlled trial conducted at nine centers in the USA. The trial was designed and monitored in accordance with the ethical principles of Good Clinical Practice and in accordance with the Declaration of Helsinki. The protocol was approved by the Institutional Review Board at each participating center. Randomization was performed centrally according to a schedule prepared for each center. Patient numbers were allocated in a strict sequential order and were not reused in the event of patient withdrawal. Eligible patients were randomized 1 : 1 : 1 to receive tamoxifen 20 mg, anastrozole 1 mg, or placebo in addition to for 3 months. All study drugs were taken orally once daily and treatment commenced within 3 days of randomization. After the 3-month prophylactic phase of the trial, patients received alone for 9 months. However, any patient who developed gynecomastia and/or breast pain during this period was re-treated with tamoxifen 20 mg or anastrozole 1 mg for a maximum duration of 3 months. Patients in the

3 tamoxifen and anastrozole groups received their original randomized therapy, while those in the placebo group received tamoxifen 20 mg. With the exception of randomized therapy, no other treatment for prostate cancer, including experimental therapy, was permitted during the trial. Patients were also not permitted to receive any concurrent medication known to cause gynecomastia or any alternative therapies. Physical examinations for the presence of gynecomastia were performed at screening, on initiation of randomized therapy, and every 3 months thereafter. Patients also completed a questionnaire, recording the occurrence of gynecomastia and breast pain at these time points. For both methods of assessment, the criterion for a response to randomized therapy was the complete absence of gynecomastia and/or breast pain. Blood samples were taken at each visit and sent to a central laboratory for the determination of serum prostatespectific antigen (PSA), testosterone, estradiol, LH, FSH, and SHBG levels. Plasma levels of the bicalutamide R-enantiomer at 3 and 6 months were measured by a central laboratory using enantioselective high-pressure liquid chromatography and ultraviolet detection. Details of adverse events occurring during the trial were elicited by open questioning. Adverse events were categorized using the FDA Coding Symbols for Thesaurus of Adverse Reaction Terms. Liver biochemistry (bilirubin, alanine aminotransferase (ALT), aspartate aminotransferase (AST)) was assessed at baseline and at each 3-monthly visit. Statistical methods The primary end point of the clinical trial was the development of gynecomastia and/or breast pain (assessed by both physical examination and questionnaire) at 3 months. Previously published data indicated that 70% of patients in the plus placebo group could be expected to develop gynecomastia and/ or breast pain. 15 The clinical trial was designed to have 80% power (5% two-sided significance) to detect a 35% reduction in the incidence of gynecomastia and breast pain in the tamoxifen and anastrozole groups relative to the placebo group. Secondary end points included the time to onset and the time to resolution of gynecomastia and breast pain (assessed by both physical examination and questionnaire) following re-treatment. Other secondary endpoints were the changes in serum levels of PSA, testosterone, estradiol, LH, FSH, and SHBG at 3 months relative to baseline (trial entry). Only patients who received randomized therapy were included in the efficacy and safety analyses. Data on the incidence of gynecomastia and breast pain were analyzed using the Cochran Mantel Haenszel test and the results adjusted for the potentially confounding center effect. Odds ratios and their 95% confidence intervals (CI) were also generated. The time-to-onset endpoint was analyzed using a Cox s proportional hazards model with the results again adjusted for the potentially confounding center effect. A statistical analysis of the time-to-resolution end point was not done, due to the smaller sample size used in the re-treatment phase of the trial. Results Pilot study The changes in serum levels of sex hormones in the 14 healthy volunteers (aged y) who received bicalutamide 150 mg plus either anastrozole 1 mg (n ¼ 7) or tamoxifen 20 mg (n ¼ 7) are summarized in Tables 1 and 2, respectively. None of the volunteers reported gynecomastia during the study; three reported breast tenderness, which resolved spontaneously in each case. Only one serious adverse event (abdominal pain) was reported; there were no unexpected tolerability findings. Clinical trial Patient demography and exposure to randomized therapy A total of 107 male patients aged y were enrolled into the trial and randomized to tamoxifen 20 mg (n ¼ 35), anastrozole 1 mg (n ¼ 36), or placebo (n ¼ 36) in addition to (Figure 2). The majority of patients had T2 3 disease (92.5%) at diagnosis and were node-negative (93.5%). Mean PSA levels at study entry were 1.49 ng/ml. The three treatment groups were well balanced with respect to baseline demographics and tumor characteristics (Table 3). Three patients had protocol violations, one in the tamoxifen group (use of medication known to cause gynecomastia within the 6 months prior to study entry) and two in the anastrozole group (prior hormonal therapy, one patient; hormonal level outside the normal range, one patient). None of these protocol violations were considered likely to affect the trial results and these patients were therefore included in the statistical analyses. All but one patient (in the tamoxifen group) received randomized therapy (Figure 2). Of the 106 patients who commenced randomized therapy, six patients (17.7%) in the tamoxifen group, seven (19.4%) in the anastrozole group, and 4 (11.1%) in the placebo group were subsequently withdrawn from therapy (Figure 2). All but three withdrawals were due to the occurrence of an adverse event or concurrent illness. The other reasons for study withdrawal were withdrawal of consent (n ¼ 2) and loss to follow-up (n ¼ 1). Six patients (5.6%) withdrew during the 3-month prophylactic phase of the study, three (2.8%) during the observation phase while receiving only and eight (7.5%) during re-treatment (Figure 2). Efficacy of randomized therapy as prophylaxis All patients who received randomized therapy were included in the analysis of gynecomastia and breast pain at 3 months, with the exception of one patient in the tamoxifen group who withdrew prematurely due to an adverse event (asthenia) and was lost to follow-up. At 3 months, the incidence of gynecomastia and/or breast pain assessed by physical examination was identical to that obtained by questionnaire in each treatment group (tamoxifen 11.8%, anastrozole 63.9%, placebo 69.4%; Figure 3). In the tamoxifen group, only three patients (8.8%) developed breast pain and one more patient (2.9%) developed both gynecomastia and breast pain (Figure 3). The difference between the tamoxifen group 77

4 78 Table 1 Changes in serum levels of sex hormones in healthy volunteers (n ¼ 7) receiving (from days 1 to 42) plus anastrozole 1 mg (from days 29 to 42) in the pilot study Time (days) (follow-up) Free testosterone Geometric mean (CV) (nmol/l) (30.415) (26.219) (45.199) (46.883) (36.081) % change from baseline Total testosterone Mean (s.d.) (nmol/l) (4.75) (6.31) (7.95) (7.77) (5.77) % change from baseline Dihydrotestosterone Mean (s.d.) (pg/ml) (267.3) (619.0) (548.7) (589.4) (439.1) % change from baseline Luteinizing hormone Mean (s.d.) (IU/l) 6.17 (7.37) (10.79) (12.29) (16.98) 7.37 (10.47) % change from baseline Sex hormone binding globulin Mean (s.d.) (pmol/ml) (0.016) (0.016) (0.014) (0.022) (0.017) % change from baseline Prolactin Mean (s.d.) (IU/l) 6.27 (3.64) 9.73 (3.71) 9.19 (3.60) 8.74 (3.31) 6.14 (1.96) % change from baseline Follicle-stimulating hormone Mean (s.d.) (IU/l) (29.72) (25.91) (27.2) (28.0) (28.02) % change from baseline Estradiol Mean (s.d.) (pmol/l) (6.07) (26.95) (16.46) (14.54) (15.12) % change from baseline CV, coefficient of variation; s.d., standard deviation. and the placebo group with respect to the development of gynecomastia and/or breast pain was statistically significant (Po0.0001). However, the difference between the anastrozole group and the placebo group was not significant (P ¼ 0.749). Patients in the tamoxifen group were 3.22 times less likely (relative risk estimate, 95% CI 1.28, 7.69) to develop gynecomastia and/or breast pain during the 3 months after randomization than those in the placebo group. Patients in the anastrozole group were 1.14 times less likely (relative risk estimate, 95% CI 0.91, 1.43) to develop the condition within 3 months of randomization than those in the placebo group. The median time to onset of gynecomastia and/or breast pain was significantly longer in the tamoxifen group than in the placebo group (P ¼ and as assessed by physical examination and questionnaire, respectively) (Figure 4). Interpretation of the results at 6, 9, and 12 months after randomization is confounded by the re-treatment of some patients with tamoxifen 20 mg or anastrozole 1 mg (Figure 2) for gynecomastia/breast pain developing before these timepoints, and these results are therefore not presented. Efficacy of randomized therapy as treatment The number of patients actively re-treated for gynecomastia/breast pain in the tamoxifen, anastrozole, and placebo groups was 26 (76.5%), 32 (88.8%), and 32 (88.8%), respectively (Figure 2). Resolution of gynecomastia and breast pain occurred in 65.4% of patients in the tamoxifen group, 71.8% of patients in the placebo group (who received tamoxifen) and 18.8% of patients in the anastrozole group. The median time to resolution in the tamoxifen group was 106 days as assessed by physical examination and 102 days as assessed by questionnaire; corresponding times to resolution in the placebo group were 108 and 91 days, respectively. PSA and hormonal end points The mean absolute and percentage changes in serum PSA and sex hormone levels at 3 months are presented in Table 4. Mean serum PSA levels declined in each of the three treatment groups. The mean percentage change in serum PSA level was greater in the placebo group ( 30.1%) than in the tamoxifen and anastrozole groups ( 24.4 and 16.7%, respectively). Mean serum testosterone levels increased by 162.8% in the tamoxifen group, 124.3% in the anastrozole group, and 63.7% in the placebo group. Mean percentage changes in serum estradiol, LH, and FSH levels were also greater in the tamoxifen and anastrozole groups than in the placebo group. Serum SHBG levels increased across all three groups, with the greatest percentage increase seen in the tamoxifen group.

5 Table 2 Changes in serum levels of sex hormones in healthy volunteers (n ¼ 7) receiving plus tamoxifen 20 mg (both from days 1 to 42) in the pilot study 79 Time (days) (follow-up) Free testosterone Geometric mean (CV) (nmol/l) (11.970) (18.628) (34.582) (34.303) (22.686) % change from baseline Total testosterone Mean (s.d.) (nmol/l) (2.11) (3.90) (4.42) (5.19) (6.49) % change from baseline Dihydrotestosterone Mean (s.d.) (pg/ml) (177.2) (169.8) (207.5) (224.3) (308.7) % change from baseline Luteinizing hormone Mean (s.d.) (IU/l) 3.34 (1.61) 8.21 (3.38) (6.84) (8.04) 6.56 (2.52) % change from baseline Sex hormone binding globulin Mean (s.d.) (pmol/ml) (0.011) (0.020) (0.017) (0.016) (0.019) % change from baseline Prolactin Mean (s.d.) (IU/l) 5.04 (2.30) 5.63 (2.84) 4.80 (3.46) 5.53 (3.11) 6.11 (2.86) % change from baseline Follicle-stimulating hormone Mean (s.d.) (IU/l) 5.87 (4.23) 8.71 (6.2) 8.34 (6.63) 9.33 (7.38) 7.04 (5.23) % change from baseline Estradiol Mean (s.d.) (pmol/l) (24.30) (25.36) (26.39) (40.32) (34.45) % change from baseline CV, coefficient of variation; s.d., standard deviation. 107 patients randomized 35 patients randomized to tamoxifen 20 mg plus - 1 patient did not receive randomized therapy 36 patients randomized to anastrozole 1 mg plus 36 patients randomized to placebo plus bicalutamide 150 mg 34 patients received randomized therapy - 2 patients withdrew due to AEs 36 patients received randomized therapy - 3 patients withdrew due to AEs 36 patients received randomized therapy - 1 patient withdrew due to AE 32 patients entered the observation phase - 1 patient withdrew informed consent - 1 patient withdrew due to AEs 33 patients entered the observation phase - 1 patient withdrew informed consent 35 patients entered the observation phase 26 patients had symptoms and were re-treated with tamoxifen 20 mg - 2 patients withdrew due to AEs 32 patients had symptoms and were re-treated with anastrozole 1 mg - 3 patients withdrew due to AEs 32 patients had symptoms and were re-treated with tamoxifen 20 mg - 1 patient lost to follow-up - 2 patients withdrew due to AEs Figure 2 Flow diagram of clinical trial patients.

6 80 Table 3 Baseline demographic and tumor characteristics of patients in the clinical trial Tamoxifen 20 mg plus (n ¼ 35) Anastrozole 1 mg plus (n ¼ 36) Placebo plus bicalutamide 150 mg (n ¼ 36) Mean age (range) (y) 64.6 (52 79) 64.4 (49 75) 67.1 (49 81) Mean weight (range) (kg) 82.4 ( ) 86.7 ( ) 82.7 ( ) Mean height (range) (cm) ( ) ( ) ( ) Race n (%) Caucasian 27 (77.1) 33 (91.7) 33 (91.7) Other 8 (22.9) 3 (8.4) 3 (8.4) Tumor stage, n (%) T1 1 (2.9) 1 (2.8) 2 (5.6) T2 19 (54.4) 24 (66.6) 21 (58.3) T3 15 (42.9) 9 (25.0) 11 (30.6) T4 0 (0) 2 (5.6) 2 (5.6) Lymph node status, n (%) N0 34 (97.1) 34 (94.4) 32 (88.9) N 1 (2.9) 2 (5.6) 1 (2.8) N+ 0 (0) 0 (0) 3 (8.3) Mean prostate-specific antigen (ng/ml) Patients (%) Gynecomastia Breast pain Gynecomastia plus breast pain P < P = PE Q PE Q PE Q Bicalutamide 150 mg Bicalutamide 150 mg Bicalutamide 150 mg plus tamoxifen 20 mg plus anastrozole 1 mg plus placebo (n=34) (n=36) (n=36) PE, physical examination; Q, questionnaire Figure 3 Incidence of gynecomastia and/or breast pain in the clinical trial as assessed by physical examination (PE) and questionnaire (Q) at 3 months after randomization. Pharmacokinetics The geometric mean plasma level of R-bicalutamide in the trial population (n ¼ 95) at 3 months was 16.4 mg/ml (range ), and the means and ranges were similar across all three treatment groups. The geometric mean plasma level of R-bicalutamide at 6 months (n ¼ 86) was 16.9 mg/ml (range ) and, again, the means and ranges were similar across all treatment groups. Safety All patients experienced at least one adverse event during the trial. Details of adverse events occurring in X5% of patients in any treatment group, regardless of causality, are shown in Table 5. The most common events were gynecomastia and breast pain, but Bicalutamide 150 mg + tamoxifen 20 mg Bicalutamide 150 mg + anastrozole 1 mg Bicalutamide 150 mg + placebo Physical examination Questionnaire Days p tamoxifen 20 mg vs + placebo Figure 4 Median time (as estimated by Kaplan Meier analysis) to onset of gynecomastia/breast pain in the clinical trial as assessed by physical examination and questionnaire. these were less frequent in the tamoxifen group than in the anastrozole or placebo groups. No other significant safety issues were identified. Three serious events (adenoma, aortic stenosis, urinary tract disorder) were reported during the trial, one event in each treatment group. The demographic characteristics of these patients were similar to those of the rest of the trial population. However, none of these events were considered related to trial therapy or led to withdrawal of trial therapy. There were no deaths during the trial. Fourteen patients (13.2%) withdrew from trial therapy (tamoxifen, five patients; anastrozole, six patients; placebo, three patients; Figure 2) as a result of a total of 23 adverse events, which included gynecomastia in two withdrawals and breast pain in five. One patient (0.9%) was withdrawn due to moderate rises in serum AST and ALT levels, which later resolved after treatment withdrawal

7 Table 4 Summary of changes in prostate-specific antigen and hormone levels at 3 months after randomization in the clinical trial 81 Tamoxifen 20 mg plus Anastrozole 1 mg plus Prostate-specific antigen (ng/ml) Mean absolute change (s.d.) 1.67 (5.22) 0.59 (1.98) Mean % change Testosterone (nmol/l) Mean absolute change (s.d.) (12.86) (7.26) Mean % change Estradiol, pmol/l Mean absolute change (s.d.) (281.84) (39.64) Mean % change Luteinizing hormone (IU/l) Mean absolute change (s.d.) (7.54) 7.10 (4.01) Mean % change Follicle-stimulating hormone (IU/l) Mean absolute change (s.d.) 5.13 (5.60) 0.82 (13.95) Mean % change Sex hormone binding globulin (lmol/l) Mean absolute change (s.d.) 0.03 (0.03) 0.01 (0.01) Mean % change s.d., standard deviation. Discussion Prophylactic and treatment options for the gynecomastia and breast pain associated with nonsteroidal antiandrogen monotherapy would enable patients with early prostate cancer to derive maximal benefit from their treatment. Hormonal therapies that reduce estrogenic activity in the breast tissue are one potential management option. Here, we have explored the efficacy, tolerability, and safety of the antiestrogen, tamoxifen, and the aromatase inhibitor, anastrozole, as both prophylaxis and treatment for gynecomastia and breast pain in noncastrate patients receiving bicalutamide 150 mg/day for early prostate cancer. Although the use of these hormonal agents in combination with bicalutamide might be expected to induce a rise in serum testosterone levels above that seen with bicalutamide alone, 13,14 the initial pilot study showed no evidence of any such pharmacodynamic interaction in healthy elderly men. The clinical trial was therefore able to proceed. At 3 months, the incidence of gynecomastia and/or breast pain as assessed by physical examination or questionnaire was lower in the tamoxifen 20 mg group than in the anastrozole 1 mg or placebo groups (11.8 vs 63.9 and 69.4%, respectively). Although, as expected based on the pharmacologic mechanisms involved, the benefits of prophylactic tamoxifen 20 mg did not persist once therapy was discontinued. Re-treatment with tamoxifen 20 mg was effective, with approximately two-thirds of patients experiencing complete resolution of bicalutamide-induced gynecomastia/breast pain within 3 months. Patients initially randomized to placebo in the prophylactic stage of the trial also responded similarly to tamoxifen treatment. The latter finding is in accordance with the results of previous studies on the therapeutic use of antiestrogens in established gynecomastia/breast pain of other etiologies The optimal dose of tamoxifen, either as prophylaxis or treatment for gynecomastia/breast pain (of any etiology) has not been established. Doses ranging from 10 to 40 mg/day have been used with good effect. 3 As most studies of tamoxifen for gynecomastia have been relatively small, uncontrolled, and with variable response criteria, no conclusions can be drawn relating to the dose-response relationship of tamoxifen in this setting. In the current study, a dose of 20 mg/day (that licensed for early breast cancer) was chosen and, using the complete absence of gynecomastia and/or breast pain as the criterion of response, achieved an approximate response rate of two-thirds in both the prophylactic and re-treatment phases of the clinical trial. Given that other trials have reported that a dose of 10 mg/day reduces gynecomastia and/or alleviates breast pain, 8,11 there is clearly a case for evaluating lower doses in patients receiving. In contrast to tamoxifen, anastrozole 1 mg/day exhibited only minimal efficacy in both the prophylactic and re-treatment phases of the trial. This is the first study in which a third-generation, selective, aromatase inhibitor has been evaluated for the management of gynecomastia. The literature on the use of the firstgeneration compound, testolactone, in the management of gynecomastia is sparse and the results are inconclusive. 3,7 The reasons for the disappointing efficacy of the more selective compound anastrozole, which was given at the dosage used in its licensed indications (1 mg/day), in patients receiving nonsteroidal antiandrogens are unclear. As expected, serum testosterone, estradiol, LH, and FSH levels increased during nonsteroidal antiandrogen therapy, both in the pilot study and the clinical trial. In the clinical trial, the increase in serum testosterone levels was greater in the tamoxifen and anastrozole groups than in the placebo group, despite the absence of any

8 82 Table 5 Adverse events occurring in X5% of patients in any treatment group (regardless of causality) in the clinical trial % patients Tamoxifen 20 mg plus (n ¼ 35) Anastrozole 1 mg plus (n ¼ 36) Placebo plus bicalutamide 150 mg (n ¼ 36) Breast pain Gynecomastia Pharyngitis Vasodilation Rash Accidental injury Asthenia Back pain Dizziness Sinusitis Constipation Decreased libido Diarrhea Dyspnea Gastritis Nausea Pain Pelvic pain Periodontal abscess Pruritus Somnolence Urinary frequency Urinary retention Urinary tract disorder Urinary urgency Vomiting Abdominal pain Paresthesia Urinary tract infection Dry skin Hypercholesterolemia Hypertension Increased cough Peripheral edema Urticaria pharmacodynamic interaction in either treatment group in the pilot study. The reasons for the difference between the pilot study and the clinical trial are not clear. The concentrations of free and total testosterone and dihydrotestosterone observed before and after the addition of anastrozole or tamoxifen to bicalutamide in the healthy elderly men in the pilot study were similar to those reported previously with monotherapy in prostate cancer patients. 4 However, the increases in mean serum LH levels in both treatment groups in the pilot study were greater than previously reported with bicalutamide alone. 4 Although serum testosterone levels in the tamoxifen group were higher than those in the placebo group during the prophylactic phase of the clinical trial, we did not observe any adverse impact of combination therapy on cancer control as assessed by PSA levels. However, the baseline serum PSA levels in our patients were low and we are therefore unable to rule out any long-term implications of tamoxifen 20 mg in combination with on patient outcome. The use of tamoxifen in the management of bicalutamide-induced gynecomastia and breast pain should therefore remain investigational until long-term trials have excluded an impact on prostate cancer control. Again, contrary to the pilot study, serum SHBG levels increased in all three groups in the clinical trial, most notably in the tamoxifen group. Serum SHBG levels tend to increase with age, thereby reducing free testosterone levels, which may result in sexual dysfunction. 16 However, given that the observed percentage increase in total testosterone levels following tamoxifen therapy was several-fold higher than the increase in serum SHBG levels, a clinically relevant impact on sexual function is unlikely in this trial. This is supported by our adverse event data which show no evidence of increased sexual dysfunction in these patients, with fewer than 5% of men in all treatment groups reporting either loss of libido or erectile difficulties. Future clinical studies of tamoxifen in bicalutamide 150-mg-treated patients should provide confirmation that the combination has no adverse implications for sexual interest and function. No other significant safety issues were identified during the clinical trial and there was no evidence of a pharmacokinetic interaction between bicalutamide and either tamoxifen or anastrozole, although further studies are needed to fully assess this issue. Although the mean R-bicalutamide levels in our patients were slightly lower than reported in an earlier monotherapy study (approximately 17 vs 22 mg/ml), 17 they are

9 consistent with the findings of other recent studies (unpublished data). Conclusion The incidence of gynecomastia and breast pain was significantly reduced by tamoxifen 20 mg/day initiated at the same time as monotherapy or following the development of bicalutamide-related breast changes. However, further studies are needed to determine the optimum dose of tamoxifen for both prophylaxis and treatment, and to assess any impact on prostate cancer control. Anastrozole 1 mg/day does not appear to be a viable management option for bicalutamide-induced gynecomastia and breast pain. Acknowledgements We gratefully acknowledge the contribution to the trial of the following investigators: A Cantwell (Atlantic Urological Associates, Daytona Beach, FL, USA); J Ross (Mississippi Center for Clinical Research, Jackson, MI, USA); RJ Ross (Hattiesburg Clinic, Hattiesburg, MS, USA); K Tomera (Alaska Clinical Research Center, Anchorage, Alaska); N Nemoy (Cedars-Sinai Medical Center, Los Angeles, CA, USA); D Keiller (4033 3rd Avenue, San Diego, CA); and PM Knapp (Urology of Indiana; Indianapolis, IN, USA). References 1 Wirth MP et al. Bicalutamide 150 mg in addition to standard care in patients with localized or locally advanced prostate cancer: results from the second analysis of the early prostate cancer program at median followup of 5.4 years. J Urol 2004; 172: Iversen P et al. Bicalutamide monotherapy compared with castration in patients with nonmetastatic locally advanced prostate cancer: 6.3 years of followup. JUrol2000; 164: McLeod DG, Iversen P. Gynecomastia in patients with prostate cancer: a review of treatment options. Urology 2000; 56: Verhelst J et al. Endocrine profiles during administration of the new non-steroidal anti-androgen Casodex in prostate cancer. Clin Endocrinol (Oxf) 1994; 41: Tyrrell CJ et al. Prophylactic breast irradiation with a single dose of electron beam radiotherapy (10 Gy) significantly reduces the incidence of bicalutamide-induced gynecomastia. Int J Radiat Oncol Phys Biol 2004; 60: Widmark A et al. Does prophylactic breast irradiation prevent antiandrogen-induced gynecomastia? Evaluation of 253 patients in the randomized Scandinavian trial SPCG-7/SFUO-3. Urology 2003; 61: Braunstein GD. Aromatase and gynecomastia. Endocr Relat Cancer 1999; 6: Parker LN, Gray DR, Lai MK, Levin ER. Treatment of gynecomastia with tamoxifen: a double-blind crossover study. Metabolism 1986; 35: Alagaratnam TT. Idiopathic gynecomastia treated with tamoxifen: a preliminary report. Clin Ther 1987; 9: McDermott MT, Hofeldt FD, Kidd GS. Tamoxifen therapy for painful idiopathic gynecomastia. South Med J 1990; 83: Staiman VR, Lowe FC. Tamoxifen for flutamide/finasterideinduced gynecomastia. Urology 1997; 50: Serels S, Melman A. Tamoxifen as treatment for gynecomastia and mastodynia resulting from hormonal deprivation. J Urol 1998; 59: Goss PE, Tye LM. Anastrozole: a new selective nonsteroidal aromatase inhibitor. Oncology (Huntingt) 1997; 11: Krause W, Holland-Moritz H, Schramm P. Treatment of idiopathic oligozoospermia with tamoxifen a randomized controlled study. Int J Androl 1992; 15: Soloway MS et al. Bicalutamide in the treatment of advanced prostatic carcinoma: a phase II noncomparative multicenter trial evaluating safety, efficacy and long-term endocrine effects of monotherapy. JUrol1995; 154: Ahn HS, Park CM, Lee SW. The clinical relevance of sex hormone levels and sexual activity in the ageing male. BJU Int 2002; 89: Tyrrell CJ et al. Casodext mg daily, used as monotherapy for the treatment of patients with advanced prostate cancer. An overview of the efficacy, tolerability and pharmacokinetics from three Phase II dose-ranging studies. Eur Urol 1998; 33: