Fax: (202) 626-4833
Article first published online: 11 JUL 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 2, pages 361–376, 15 July 2002
How to Cite
Samson, D. J., Seidenfeld, J., Schmitt, B., Hasselblad, V., Albertsen, P. C., Bennett, C. L., Wilt, T. J. and Aronson, N. (2002), Systematic review and meta-analysis of monotherapy compared with combined androgen blockade for patients with advanced prostate carcinoma . Cancer, 95: 361–376. doi: 10.1002/cncr.10647
This work was developed under contract with the Agency for Healthcare Research and Quality (AHRQ contract number 290-97-0015). The Blue Cross and Blue Shield Association Technology Evaluation Center is an Evidence-Based Practice Center of the AHRQ.
Charles L. Bennett has received grants from Schering-Plough and Astra-Zeneca and has acted as a consultant for Schering-Plough.
- Issue published online: 11 JUL 2002
- Article first published online: 11 JUL 2002
- Manuscript Accepted: 15 JAN 2002
- Manuscript Revised: 7 JAN 2002
- Manuscript Received: 9 MAY 2001
- androgen antagonists;
- prostatic neoplasms
The current systematic review and meta-analysis compared monotherapy and combined androgen blockade in the treatment of men with advanced prostate carcinoma. Outcomes of interest included overall, cancer specific, and progression-free survival; time to treatment failure; adverse events; and quality of life.
The literature search identified randomized trials comparing monotherapy (orchiectomy and luteinizing hormone-releasing hormone [LHRH] agonists) with combination therapy using orchiectomy or a LHRH agonist plus a nonsteroidal or steroidal antiandrogen. Dual independent review occurred. The meta-analysis used a random effects model.
Twenty-one trials compared survival after monotherapy with survival after combined androgen blockade (n = 6871 patients). The meta-analysis found no statistically significant difference in survival at 2 years between patients treated with combined androgen blockade and those treated with monotherapy (20 trials; hazard ratio [HR] = 0.970; 95% confidence interval [95% CI], 0.866–1.087). The authors determined a statistically significant difference in survival at 5 years that favored combined androgen blockade (10 trials; HR = 0.871; 95% CI, 0.805–0.942). For the subgroup of patients with a good prognosis, there was no statistically significant difference in survival. Adverse effects leading to withdrawal from therapy occurred more often with combined androgen blockade. To the authors' knowledge there is little evidence published to date comparing the effects of combined androgen blockade and monotherapy on quality of life, but the single randomized trial that adequately addressed this outcome reported an advantage for monotherapy over combined androgen blockade.
A thorough examination of the usefulness of combined androgen blockade must balance the modest increase in expected survival observed at 5 years against the increased risk of adverse effects and the potential for adversely affecting the patient's overall quality of life. [See editorial on pages 209–10, this issue.] Cancer 2002;95:361–76. © 2002 American Cancer Society.
Initial reports from nonrandomized studies generated enthusiasm for combined androgen blockade coupled with either surgery or an luteinizing hormone-releasing hormone (LHRH) agonist to prevent testicular production of testosterone with an antiandrogen to block the action of adrenal androgens.1, 2 Early studies used nonsteroidal antiandrogens (NSAAs) with purely antiandrogenic activity. Cyproterone acetate, a steroidal antiandrogen with mixed antiandrogenic and androgenic properties, also has been investigated. Subsequent pilot studies reported that men who were treated with combined androgen blockade survived longer than historic controls. These observations led to randomized controlled trials to test the hypothesis that men with advanced prostate carcinoma survived longer after combined androgen blockade than after monotherapy. Nevertheless, debate has continued with regard to whether combined androgen blockade has an advantage over monotherapy in the treatment of advanced prostate carcinoma.
Reviews and overviews that addressed the controversy regarding combined androgen blockade have reached differing conclusions. Critics note that the preponderance of trials found no statistically significant difference that favored combined androgen blockade, and are concerned with the additional costs and adverse effects.3–5 Advocates of combined therapy find support in the larger size of the positive trials and issues related to the design of the negative trials.6–10
The Prostate Cancer Trialists' Collaborative Group (PCTCG) addressed this controversy using individual patient meta-analysis that combined results from 27 randomized trials with a total of 8275patients.11 Potential limitations in the approach taken in the PCTCG meta-analysis have been widely discussed in the literature.7, 8, 10, 12 A threshold issue was whether the consistently negative trials using a steroidal antiandrogen, which has progestational and other hormonal effects, should be combined with those trials in which an NSAA, which selectively blocks the androgen receptor, was used. Because the flare reaction has been reported to occur with LHRH agonists but not with orchiectomy, a second issue is whether the results from trials in which patients in the control arm did not receive short-term antiandrogens to block flare should be combined with those trials that controlled flare or those in which orchiectomy was used. In 2000 the PCTCG reported that the typical duration of follow-up was nearly 5 years and that 72% of patients were no longer alive at the time of last follow-up. Using an intention-to-treat analysis, actuarial estimates of survival at 5 years were 23.6% for monotherapy and 25.4% for combined androgen blockade. The difference in survival of 1.8% (95% confidence interval [95% CI], 0–4%) was not statistically significant at 5 years (P > 0.1).
A second meta-analysis that was restricted to randomized trials that used NSAAs in the combination regimen was reported by Caubet et a.13 This meta-analysis excluded studies published only as abstracts, those that did not publish data regarding survival, and one trial that used short-term cyproterone acetate for flare in control patients receiving an LHRH agonist alone. Only nine trials included the necessary statistical summaries in the published reports to permit estimation of the relative risk by at least one of the authors' two methods. The meta-analyses used a random effects model to generate pooled estimates of the relative risk for survival from each of the two methods used to estimate the log hazard ratios. Both methods of analysis yielded statistically significant relative risks that favored combined androgen blockade. A third recent meta-analysis of trials using flutamide14 was to our knowledge the first to incorporate the results of the largest randomized trial completed to date, the Intergroup trial (INT 0105).15 This meta-analysis also found a statistically significant increase in survival with the use of combined androgen blockade. The purpose of the current systematic review and meta-analysis was to investigate the relative effectiveness of combined androgen blockage versus monotherapy in men with advanced prostate carcinoma. Previous meta-analyses have focused exclusively on survival, whereas the current review sought data concerning survival as well as other outcomes.
This review was designed prospectively to define study objectives, search strategy, study selection criteria and methods for determining study eligibility, data elements and methods for abstraction, and methods for study quality assessment. Two independent reviewers completed each step in this protocol and resolved disagreements by consensus. Disagreements were infrequent and generally were resolved by reconciliation of an oversight. When survival rates were estimated from figures in publications, disagreements always were < 5%, and the consensus estimate was the midpoint.
All efficacy studies were randomized controlled trials. Reviewers assessed the study quality dimensions recognized in the literature as a source of bias16: adequacy of randomization method, adequacy of concealment of allocation and use of blinding, and documentation of withdrawals and whether results were analyzed in an intent-to-treat fashion. With the exception of blinding and intent-to-treat analysis, published reports commonly provided insufficient information to permit valid assessments of these quality dimensions. Therefore, studies that blinded patients and investigators to group assignment and that used an intent-to-treat analysis of overall survival and/or progression-related outcomes were classified as higher quality studies for sensitivity analysis. Blinding was considered not applicable when orchiectomy was one of the study arms, although patients undergoing orchiectomy could be blinded with regard to whether they received an antiandrogen.
Literature Search and Study Selection
MEDLINE, Cancerlit, Embase, and Cochrane Library databases from 1966 until March 1998 and Current Contents on Diskette through August 24, 1998 were searched for all articles that included at least one of the following terms in their titles, abstracts, or keyword lists: leuprolide, goserelin, buserelin, flutamide, nilutamide, bicalutamide, cyproterone acetate, diethylstilbestrol (DES), or orchiectomy (castration, orchidectomy).
Search results were limited to studies concerning human subjects indexed under Medical Subject Heading (MeSH) “prostatic neoplasms.” Randomized controlled trials were identified using the U.K. Cochrane Center search strategy.17 The total retrieval was 1477 references, which were checked against the Cochrane Controlled Trials Register, the CENTRAL register, and trials cited in 2 recent meta-analyses. No additional trials were identified.
The study selection criteria for the current review limited reports of the efficacy outcomes of interest to those randomized controlled trials that made any one of the following four comparisons: 1) orchiectomy compared with orchiectomy plus an antiandrogen; 2) an LHRH agonist compared with an LHRH agonist plus an antiandrogen; 3) orchiectomy compared with an LHRH agonist plus an antiandrogen; or 4) either orchiectomy or an LHRH agonist alone compared with either orchiectomy or an LHRH agonist plus an antiandrogen. Randomized controlled trials that compared different doses of the same agent only were excluded. For adverse events, Phase II studies that reported on withdrawals from therapy were included. In addition, all studies reporting on quality of life also were included.
The patient population of interest for the current review was comprised of men with advanced prostate carcinoma, including regional or disseminated metastases (D1 or D2 disease; N+/M0 or M1) and minimally advanced disease (C disease; T3-4/N0 or Nx/M0). Studies that selected patients with minimally advanced disease were included due to the recent trend in clinical practice of initiating androgen suppression therapy earlier in the disease course. We also examined outcomes that were analyzed by patient prognostic factors such as tumor grade, extent of disease, and performance status. Outcomes of interest were overall, cancer specific, and progression-free survival; time-to-treatment failure; adverse effects; and quality of life. When available, data regarding patient preferences were included.
We encountered well described difficulties18, 19 in capturing infrequent events from small trials and inconsistencies among trials in measuring and reporting adverse events. Summarized herein is the most reliable index of serious adverse events, the rate of withdrawal from therapy. A summary of adverse events by category (e.g., cardiovascular, endocrine, etc.) is in the full evidence report.20
Overall survival was the only outcome for which there were adequate data for meta-analysis. We used the general approach to meta-analysis of trials in prostate carcinoma described by Caubet et al.,13 with additional guidance from Whitehead and Whitehead.21 The problem of combining evidence from studies with several different treatment arms requires going beyond standard meta-analysis techniques.22 The solution to the problem entails defining parameters that describe the possible interventions. The poor survival rates reported for metastatic prostate carcinoma have implied a large value for the hazard rate (rate of death across time). We made the same assumption that is used in standard meta-analysis, namely, we assume that the effect measure (hazard ratio in this case) remains constant across studies. Because there currently are several different treatments, we assumed that all the hazard ratios between the various treatments remain constant. The model is a generalization of the random effects model described by DerSimonian and Laird.23 It essentially is the same model used by Egret24 except that it is applied to continuous outcomes instead of dichotomous outcomes. The model is a generalization that includes both fixed and random effects terms. The fixed effects terms are the individual study intercepts. The random effects terms are the slopes for the treatment effects. Estimates of all parameters, including the extra variation, are obtained by maximum likelihood.
Based on the above assumptions, the objective was to estimate the hazard rate for each arm of each study, or to estimate the proportional hazards term and its standard error. We obtained estimates from other statistics for studies that did not provide this directly. Caubet et al.13 suggested estimating the log hazard ratio β from the chi-square value of the log-rank test:
in which e is the number of events. When Kaplan–Meier curves were given, an estimate of individual hazards generally was possible. Finally, if survival at a given point in time, S, was as S = X/N (in which X is the number of subjects surviving and N is the number of subjects followed), the hazard was estimated as: β=−Log(S)/t, in which t is the time at which survival was measured. To use this meta-analysis methodology, we constructed a table of hazard rates for each arm of each study. The meta-analysis was performed with software developed at the Duke Clinical Research Institute.
Sensitivity analyses tested for heterogeneity of methods (including the effect of including studies of lower methodologic quality), participants, and interventions. An initial analysis determined whether the results of castration and DES are comparable and thus whether it was valid to pool studies in which the control arms used either of these monotherapies. Separate analyses also compared the available monotherapies and categories of monotherapies. All meta-analysis results were reported as the hazard ratios relative to orchiectomy.
Overview of the Evidence Base
The literature search and study selection process identified 27 randomized controlled trials (including 7987 patients) that compared the outcomes of monotherapy with the outcomes of combined androgen blockade. A nonsteroidal antiandrogen was used in the combined androgen blockade arm of 20 trials with 6085 patients. Twelve trials (n = 4583 patients) used flutamide and 8 trials (n = 1502 patients) used nilutamide. None of these trials used bicalutamide as the antiandrogen. Seven trials (n = 1902 patients) used the steroidal antiandrogen cyproterone acetate. In addition, a 28th trial (n = 813 patients) compared 4 regimens for combined androgen blockade. Drug treatments were continued until death, disease progression, or patient withdrawal due to adverse effects in all studies.
Only 4 of the 12 trials that used an LHRH agonist for monotherapy also used an initial brief treatment with an antiandrogen to control the tumor flare reaction.25–28 Therefore the other eight trials may be biased against the monotherapy arm because they did not use an antiandrogen to control flare. The flare reaction is not reported to occur when orchiectomy is used as monotherapy. Therefore, the 14 trials in which the control arm was orchiectomy alone may provide a better comparison of combined androgen blockade with monotherapy.
All trials were restricted to patients undergoing primary hormonal therapy. The mean or median age of these patients was between 65–75 years. The overwhelming majority of patients in trials of combined androgen blockade had metastatic disease, largely Stage D2 (93% of all patients randomized). Enrollment was limited to patients with metastatic disease in 20 trials. Of the remaining trials, seven included patients with no evidence of metastases (Stages C disease or T3-T4/N0/M0 disease). The final trial29 did not provide information regarding the disease stage distribution of the randomized patients.
For the most part, these studies were well balanced with regard to stage of disease and, when reported, the distribution of other prognostic factors. However, only four trials provided data that compared treatment arms for one or more primary outcomes after stratifying by prognostic group.15, 30–32 Two other author groups stated that they completed comparative analyses on subgroups, but did not publish the relevant data.25, 33 The lack of reporting of results stratified by prognostic groups is a notable deficiency of this body of evidence.
Efficacy outcomes summarized herein include overall survival, cancer specific survival, progression-free survival and/or time to disease progression, and time to treatment failure. Overall survival after monotherapy was compared with survival after combined androgen blockade in 21 studies (Table 1), including 23 separate comparisons of a monotherapy with a combination regimen and a total of 6871 patients. Cancer specific survival was reported in five trials. Time to treatment failure was reported in six trials. Some measure of disease progression was reported in 21 trials. Increase in the serum prostate specific antigen (PSA) level rarely was used as evidence of disease progression.
|Systematic review||Meta-analysis, 2 yrs||Meta-analysis, 5 yrs|
|No. of trials||No. of patients||No. of trials||No. of patients||No. of trials||No. of patients|
|Orchiectomy + NSAA||6||2480||6||1261||2||922|
|LHRHa + NSAA||9||3156||8||1519||5||1003|
|Orchiectomy + cyproterone||4||578||4||292||2||135|
|LHRHa + cyproterone||2||657||2||270||1||111|
Nearly all the trials provided some data regarding the adverse effects of treatment. However, the specific adverse outcomes that were reported varied markedly among the trials. In addition, one randomized trial, a feasibility study of a second trial, and a cross-sectional survey reported on quality of life.
Among the 21 studies (a total of 6871 patients) reporting comparisons of overall survival after monotherapy with survival after combined androgen blockade (Table 2), data were obtained at the following intervals: 1 year (19 trials), 2 years (20 trials), and 5 years (10 trials). One trial was excluded from Table 2 because it compared two regimens of combined androgen blockade but did not include a monotherapy arm.48 Fifteen of the 21 trials reporting on overall survival (including a total of 5636 patients) used nonsteroidal antiandrogens for combined androgen blockade; 6 trials (including a total of 1235 patients) used cyproterone acetate. Overall, 18 of 21 trials (including a total of 5485 patients) found no statistically significant difference with regard to survival between monotherapy and combined androgen blockade.
|Trial||Control intervention||Treatment intervention||No. at risk, entry Cx||No. at risk, entry Tx||Median OS (mos) Cx||Median OS (mos) Tx||% alive 1-yr Cx||% alive 1-yr Tx||% alive 2-yr Cx||% alive 2-yr Tx||% alive 5-yr Cx||% alive 5-yr Tx||2-yr HR||95% CI||Reporteda|
|HR (95% CI)||P value|
|Zalcberg et al., Australian Multicenter Randomized Trial34||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus flutamide, 750 mg||110||112||31.0||23.0||88%||74%||60%||48%||1.435||0.964–2.137||0.21|
|Eisenberger et al., NCI SWOG/ECOG INT-101515||Bilateral orchiectomy||Bilateral orchiectomy plus flutamide, 750 mg||685||697||29.9||33.5||85%||86%||62%||63%||28%||32%||0.938||0.629–1.398||0.91 (0.81–1.01)||0.14|
|Beland et al., Canadian Anandron Study Group35||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus nilutamide, 300 mg||96||98||18.9||24.3||72%||77%||38%||49%||0.737||0.500–1.086||0.137|
|Namer et al., French Multicenter Single-Dose Trial36||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus nilutamide, 300 mg||53||45||82%||82%||1.000||0.439–2.276|
|Dijkman et al., International Anandron CAB Trial37||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus nilutamide, 300 mg 1 mo then 150 mg||232||225||23.6||27.3||77%||83%||51%||60%||18%||27%||0.759||0.381–1.509||0.0326|
|Brisset et al., French Multicenter Two-Dose Anandron Trial38||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus nilutamide, 150 mg||43||46||24.0||22.0||83%||82%||1.265||0.463–3.458|
|Brisset et al., French Multicenter Two-Dose Anandron Trial38||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus nilutamide, 300 mg||43||38||24.0||22.0||83%||79%||1.000||1.000–1.000|
|Williams et al., Hammersmith/R PMS Trial (UK)39||Bilateral orchiectomy||Bilateral orchiectomy plus cyproterone, 300 mg||24||20||92%||80%||91%||75%||3.050||0.617–15.079|
|Robinson et al., EORTC Protocol 3080533||Bilateral orchiectomy||Bilateral orchiectomy plus cyproterone acetate, 50 mg||110||111||22.9||21.8||76%||70%||48%||44%||15%||13%||1.119||0.774–1.616||0.751|
|Jorgensen et al., SPCG Study No. 240||Bilateral orchiectomy plus placebo||Bilateral orchiectomy plus cyproterone acetate, 150 mg||136||137||33.0||33.0||84%||82%||64%||66%||0.931||0.621–1.395|
|Klosterhalfen and Becker, Hamburg Trial (Germany)41||Bilateral orchiectomy||Bilateral orchiectomy plus cyproterone acetate, 50 mg||16||24||31.0||29.0||81%||85%||69%||62%||6%||23%||1.288||0.428–3.875|
|Crawford et al., NCI Intergroup Trial30||Leuprolide, 1 mg sc daily plus placebo||Leuprolide, 1 mg sc plus flutamide, 750 mg||300||303||29.0||35.0||84%||90%||59%||66%||23%||26%||0.788||0.605–1.025||0.035|
|Bono et al., Italian Leuporelin Group Trial28||Leuprolide, 3.75 mg sc 28 days||Leuprolide, 3.75 mg sc 28 days plus flutamide, 750 mg||120||121||31.6||32.1||95%||95%||79%||78%||10%||15%||1.054||0.610–1.820||0.83|
|Crawford et al., North American Multicenter Trial42||Leuprolide plus placebo||Leuprolide plus nilutamide||202||209||27.0||28.6||56%||58%||0.944||0.714–1.248|
|Periti et al., Italian Multicenter Trial43||Leuprolide, 3.75 mg sc 28 days||Leuprolide, 3.75 mg sc 28 days plus nilutamide, 300 mg 1st mo, then 150 mg||126||1.000||1.000–1.000|
|Boccardo et al., PONCAP Trial44||Goserelin, 3.6 mg sc 4 wk||Goserelin, 3.6 mg sc 4 wk plus flutamide, 750 mg||186||187||32.0||34.0||85%||85%||60%||67%||0.784||0.520–1.182||0.4|
|Fourcade et al., French Multicenter Trial45||Goserelin, 3.6 mg sc 28 days plus placebo||Goserelin, 3.6 mg sc 28 days plus flutamide, 750 mg||125||120||80%||78%||1.113||0.643–1.927||0.78|
|Tyrrell et al., International Prostate Cancer Study Group Trial46||Goserelin, 3.6 mg sc 28 days||Goserelin, 3.6 mg sc 28 days plus flutamide, 750 mg||282||287||37.7||42.4||90%||87%||65%||72%||34%||42%||0.763||0.567–1.026||0.14|
|DeVoogt et al., EORTC Protocol 3084325||Buserelin, 1.5 mg sc, then 1.2 mg IN; plus cyproterone 150 mg 2 wk||Buserelin, 1.5 mg sc, then 1.2 mg IN; plus cyproterone acetate, 150 mg contin.||113||111||25.0||25.0||78%||88%||53%||52%||17%||14%||1.030||0.700–1.516||0.98|
|DiSilverio et al., Italian Multicenter Trial47||Goserelin 3.6 mg sc 28 days||Goserelin, 3.6 mg sc 28 days plus cyproterone acetate, 200 mg||156||159||30.1||23.8||83%||81%||63%||49%||1.544||1.096–2.174||0.26|
|Iversen et al., DAPROCA Trial # 8631||Total or subcapsular bilateral orchiectomy||Goserelin, 3.6 mg sc 28 days plus flutamide, 750 mg||133||129||27.6||22.7||77%||80%||58%||47%||20%||20%||1.386||0.968–1.985||0.49|
|Denis et al., EORTC Protocol 3085332||Bilateral orchiectomy||Goserelin, 3.6 mg sc 28 days plus flutamide, 750 mg||163||163||27.0||34.0||83%||85%||54%||61%||20%||28%||0.802||0.572–1.126||0.77;8q(0.60–0.99)||0.04|
|DeVoogt et al., EORTC Protocol 3084325||Bilateral orchiectomy||Buserelin, 1.5 mg sc, then 1.2 mg IN; plus cyproterone acetate, 150 mg contin.||118||111||25.0||25.0||78%||88%||53%||52%||12%||14%||1.030||0.703–1.509||0.98|
Table 3 summarizes the six trials that reported significant differences or trends. Three of the six trials reported a statistically significant difference30, 32, 37 and all favored combined androgen blockade. All three trials reporting statistically significant results used a nonsteroidal antiandrogen in the combination arms (two trials used flutamide and one used nilutamide). Three trials reported a trend,15, 35, 46 all favoring combined androgen blockade. We define a trend as Kaplan–Meier curves that differed with a significance level of 0.05 < P < 0.20.15, 35, 46 There was no apparent survival advantage from combined androgen blockade reported in any of the six trials that used cyproterone acetate in the combination arm.
|Study||Treatment arms||No.||Median (mos)||2 yrs (%)||5 yrs (%)||P value|
|Dijkman et al., 199737||Orchiectomy orch + nilut||232||23.6||51||18||0.0326|
|Δ = 3.7||Δ=9||Δ=9|
|Crawford et al., 198930||Leuprolide leup + flut||300||29.0||59||23||0.035|
|Δ = 6||Δ=7||Δ=3|
|Denis et al. 199332||Orchiectomy goser + flut||163||27||54||20||0.004|
|Δ = 7||Δ=7||Δ=8|
|Eisenberger et al. 199815||Orchiectomy orch + flut||685||29.9||62||28||0.14|
|Δ = 3.9||Δ=1||Δ=4|
|Beland et al. 199035||Orchiectomy orch + nilut||96||18.9||38||0.137|
|Δ = 5.4||Δ=9|
|Tyrrell et al. 199146||Goserelin goser + flut||282||37.7||65||34||0.14|
|Δ = 4.7||Δ=7||Δ=8|
Cancer Specific Survival
Data regarding cancer specific survival were reported in 6 trials; 5 trials (n = 1272 patients) used an NSAA in the combination arm and 1 trial (n = 342 patients) used cyproterone acetate. Three of the 6 trials found a statistically significant difference in favor of combined androgen blockade (n = 972 patients). All three trials utilized orchiectomy in the control arm and an NSAA in the combination arm. Two studies have compared orchiectomy plus placebo with orchiectomy plus nilutamide; Dijkman et al.37 included 457 patients and Beland et al.35 randomized 189 patients. Denis et al.32 compared orchiectomy alone with goserelin plus flutamide (n = 326 patients). The three trials that reported similar cancer specific survival in each arm (P > 0.20) assigned patients to the following treatments: buserelin or buserelin plus nilutamide,49 choice of buserelin or orchiectomy alone or buserelin plus cyproterone acetate,25 and orchiectomy alone or goserelin plus flutamide.30
Progression-Free Survival and/or Time to Disease Progression
Five trials that used an NSAA (n = 2867 patients) and 1 trial that used cyproterone acetate (n = 221 patients) reported progression-free survival. Eleven trials that used an NSAA (n = 2534 patients) and 5 trials that used cyproterone acetate (n = 1324 patients) reported time to disease progression. There may be similarities between outcomes labeled as either progression-free survival or time to disease progression and time to treatment failure; however, reports varied with regard to the degree of detail used to describe these outcomes and the criteria used to assess them. Absent sufficient detail concerning definitions to distinguish between outcomes, information is presented herein according to how outcomes were labeled in the selected articles.
Only two studies were found to report a statistically significant difference in progression-free survival between arms, favoring combined androgen blockade, and each utilized an NSAA. Dijkman et al.37 compared orchiectomy plus placebo with orchiectomy plus nilutamide. Crawford et al.30 compared leuprolide plus placebo with leuprolide plus flutamide. A third trial44 found a trend toward longer progression-free survival after combined androgen blockade in the comparison of goserelin alone with goserelin plus flutamide. Progression-free survival was similar in both arms of the remaining three trials that compared orchiectomy with orchiectomy plus cyproterone acetate,33 orchiectomy or goserelin alone with either orchiectomy or goserelin plus flutamide,29 or orchiectomy alone with orchiectomy plus flutamide.15
The only study that found a statistically significant difference between treatment arms with regard to time to disease progression compared orchiectomy with goserelin plus flutamide.32 One additional trial reported a trend in favor of combined androgen blockade.36 There were no apparent strong trends in favor of either arm in the remaining 13 trials.
Time to Treatment Failure
Five trials that used NSAAs (n = 1685 patients) and 1 trial that used cyproterone acetate (n = 315 patients) reported time to treatment failure. None of these studies reported a statistically significant difference between monotherapy and combined androgen blockade. There was a trend favoring the monotherapy arm in one trial that compared goserelin with goserelin plus flutamide (n = 569 patients).46 There were no trends determined in the remaining four trials.
Adverse Events Leading to Withdrawal of Therapy
To our knowledge the evidence comparing adverse effects is limited but favors monotherapy over combined androgen blockade (Table 4).20 Withdrawal from therapy occurred more often among patients treated with combined androgen blockade (10%) than among patients treated with monotherapy (2%).
|Treatment||No. of studies||No. of patients||No. of withdrawals||% withdrawals|
|Leuprolide, 1 daily||1||268||0||0.0%|
|Goserelin, 3.6 1 mo||11||1679||33||2.0%|
|Goserelin, 10.8 3 mo||2||77||1||1.3%|
|Orchiectomy + nilutamide, 150||2||271||38||14.0%|
|Orchiectomy + nilutamide, 300||3||209||24||11.5%|
|Orchiectomy + cyproterone acetate, 150||1||102||3||2.9%|
|Orchiectomy + cyproterone acetate, 300||1||20||2||10.0%|
|Goserelin 3.6 1 mo + flutamide, 750||5||846||94||11.1%|
|Orchiectomy/LHRHa + flutamide, 750||9||2804||233||8.3%|
|Orchiectomy/LHRHa + bicalutamide, 50||1||401||41||10.2%|
|Orchiectomy/LHRHa + nilutamide, 150/300||5||480||62||12.9%|
|Orchiectomy/LHRHa + cyproterone acetate, 150/300||2||122||5||4.1%|
|Combined androgen blockade, pooled||29||5255||502||9.6%|
Quality of Life
To our knowledge little evidence exists comparing the effects of combined androgen blockade and monotherapy on quality of life, but the best available evidence favors monotherapy over combined androgen blockade. A substudy of the recently completed Southwestern Oncology Group (SWOG)/Eastern Cooperative Oncology Group (ECOG) trial (INT 0105) is to our knowledge the only source of data from a randomized controlled trial that compares patients treated with a monotherapy with patients treated with combined androgen blockade with respect to formal measures for quality of life.50, 51 Data were collected at randomization and at 1 month, 3 months, and 6 months after the initiation of treatment on 3 treatment-related symptoms (diarrhea, gas pain, and body image), on physical functioning, and on emotional functioning. Patients in this substudy who were treated with orchiectomy plus flutamide reported significantly more diarrhea at 3 months (P < 0.001) and worse emotional functioning at 3 months and 6 months (P < 0.003) compared with those patients treated with orchiectomy plus placebo. There also were nonsignificant trends toward more side effects among those patients in the combination arm with the following physical functioning: fatigue, abdominal gas, overall pain, and body image. The lack of long-term follow-up data is a limitation of the current study.
A cross-sectional survey used three separate instruments to measure quality of life in patients receiving an LHRH agonist plus flutamide.52 However, this study compared only those patients who developed disease progression while receiving combined androgen blockade with those who did not develop disease progression. No patients treated with a monotherapy were included. A feasibility study conducted by the European Organization for Research and Treatment of Cancer (EORTC) (as part of Protocol 30853) included an optional assessment of quality of life at entry and at each follow-up visit using a validated questionnaire.53, 54 However, only 63 of the 327 patients randomized to treatment with either orchiectomy or goserelin plus flutamide completed the questionnaire at 1 or more clinic visits
Twenty studies (n = 6745 patients) presented data regarding survival at 2 years. Table 5 presents the combined estimates for the hazard ratios at 2 years after the initiation of treatment for 4 classes of combined androgen blockade and for monotherapy using any LHRH agonist relative to treatment with orchiectomy alone. Table 6 presents combined estimates for the hazard ratio at 2 years after the initiation of treatment for all methods of combined androgen blockade relative to monotherapy, with sensitivity analyses for heterogeneity of participants and for quality of study methods. Table 7 presents combined estimates of hazard ratios at 2 years after the start of treatment for combined androgen blockade by specific antiandrogen.
|Treatment||HR||95% CI, lower||95% CI, upper|
|Orchiectomy + NSAA||0.977||0.781||1.222|
|LHRH agonist + NSAA||0.945||0.779||1.147|
|Orchiectomy + cyproterone acetate||1.073||0.595||1.935|
|LHRH agonist + cyproterone acetate||1.335||0.988||1.803|
|Group||HR||95% CI, lower||95% CI, upper|
|Restricted to patients with D2 disease||0.965||0.852||1.093|
|Restricted high-quality studies||0.957||0.804||1.139|
|Group||HR||95% CI, lower||95% CI, upper|
|Antiandrogen = nilutamide||0.878||0.564||1.368|
|Antiandrogen = flutamide||0.945||0.779||1.147|
|Antiandrogen = cyproterone acetate||1.168||0.919||1.484|
|Antiandrogen = NSAA||0.926||0.812||1.056|
These results (Table 5) generally demonstrated no difference between the combination therapies and orchiectomy itself. The exception was LHRH agonist plus cyproterone acetate, which demonstrated poorer 2-year survival, nearly statistically significant, compared with orchiectomy. Furthermore, the different combination therapies do not appear to be different from each other (chi-square value = 7.68 for 4 degrees of freedom; P = 0.104). The results are shown in Figure 1.
One sensitivity analysis was performed for those studies that restricted the subjects to those with D2 disease. This analysis eliminated a total of seven studies.31, 36, 38, 39, 44–46 The result of this analysis also is shown in Table 6 and is shown graphically in Figure 1. A second sensitivity analysis was conducted for those studies that were determined to be of high quality based on whether they were double-blinded and whether an intent-to-treat analysis was performed. This eliminated eight studies.36, 38–40, 42, 44, 46, 47 The result of this analysis also is shown in Table 6 and Figure 1. It is clear that the subset of studies had little effect on the estimate of the hazard ratio. There was a trend toward an overall reduction in mortality, but the trend was not close to being statistically significant.
To compare the results of the current analysis with those of other meta-analyses, the studies were split into groups based on the treatment. The results are shown in Table 7 and in Figure 2. Results are presented by the specific antiandrogen used as part of combined androgen blockade. These results suggest that, excluding cyproterone acetate, the specific antiandrogen does not alter the comparison between monotherapy with combined androgen blockade.
In addition to the analyses performed on 2-year survival rates, the subset of 10 studies (n = 4443 patients) reporting 5-year survival also was included in a meta-analysis (Table 8) (Fig. 3). The number of patients included in the meta-analysis of 5-year survival represented 66% of the patients included in the meta-analysis of 2-year survival. This analysis demonstrated a statistically significant reduction in mortality at 5 years. The combined hazard ratio was 0.871, and the upper limit of the 95% CI was 0.942.
|Study, CAB regimen||HR||95% CI, lower||95% CI, upper|
|Klosterhalfen and Becker,41 orchiectomy + cyproterone acetate||0.522||0.2232||1.223|
|Iversen et al.,31 goserelin + flutamide||1.000||0.7401||1.351|
|Denis et al.,32 goserelin + flutamide||0.791||0.6028||1.038|
|Robinson et al.,33 orchiectomy + cyproterone acetate||1.075||0.7712||1.500|
|de Voogt et al.,25 buserelin + cyproterone acetate||0.927||0.6675||1.288|
|Crawford et al.,30 leuprolide + flutamide||0.917||0.7509||1.119|
|Djikman et al.,37 orchiectomy + nilutamide37||0.763||0.6062||0.962|
|Tyrell et al.,46 goserelin + flutamide||0.804||0.6470||0.999|
|Bono et al.,28 leuprolide + flutamide||0.824||0.5965||1.138|
|Eisenberger et al.,15 orchiectomy + flutamide||0.896||0.784||1.024|
A note of caution is necessary because the previous analysis was incomplete in the following respect. Although 10 studies were used to report 5-year survival, 20 studies were included in the analysis of 2-year survival. It is possible that this subset is subject to publication bias, in that trials that continue to demonstrate no effect at later follow-up may be unlikely to publish updates. Two sensitivity analyses for publication bias were performed to determine whether the studies that reported 2-year survival were somehow different from those reporting 5-year survival.
The first sensitivity analysis compared the 2-year hazard ratios of the studies that reported survival at 5 years with the 2-year hazard ratios of the studies that reported only survival at 2 years. The results of this analysis are provided in Table 9. The estimate of the ratio of the two hazard ratios also was tabulated. Although the point estimates for the two groups of trials were somewhat different, the difference was not found to be statistically significant.
|Group||HR||95% CI, lower||95% CI, upper|
|Trials with 5-year data||0.9272||0.8310||1.032|
|Trials without 5-year data||1.047||0.8668||1.264|
|Ratio of ratios||0.8857||0.7129||1.100|
The second sensitivity analysis for publication bias combined those studies reporting survival at 5 years with those studies reporting only survival at 2 years. For this analysis we assumed that the studies with data only at 2 years would continue to have the same hazard ratio through 5 years. The combined hazard ratio for all studies and, for comparison, the hazard ratio for only those studies reporting survival at 5 years are provided in Table 10. When the studies reporting survival only at 2 years were combined with the studies that reported survival at 5 years, the estimated hazard ratio moved closer to 1, but the 95% CI still excluded 1. This suggests that even if the studies that only reported survival at 2 years were continued to 5 years, the combined results still would demonstrate a statistically significant hazard ratio at 5 years. However, the magnitude of the difference attributed to combined androgen blockade would not be greater and actually might be less.
|Group||HR||95% CI, lower||95% CI, upper|
|Trials with 5-year data||0.8710||0.8055||0.9418|
|All trials, substituting 2-year HRs for trials without 5-year data||0.9146||0.8461||0.9887|
The meta-analysis found no statistically significant difference in overall survival between monotherapy and combined androgen blockade at 2 years (total of 20 trials; hazard ratio = 0.970; 95% CI, 0.866–1.087). Based on more limited data, the meta-analysis found a statistically significant increase in overall survival for combined androgen blockade at 5 years (total of 10 trials, 66% of patients in 2-year meta-analysis; hazard ratio = 0.871; 95% CI, 0.805–0.942). Sensitivity analyses suggest that if complete 5-year data were available, the magnitude of the difference attributed to combined androgen blockade would not be greater. The meta-analysis found that combined androgen blockade using flutamide or nilutamide appeared to be equivalent. The hazard ratio was 0.878 (95% CI, 0.564–1.368) in trials using nilutamide and 0.945 (95% CI, 0.779–1.147) in trials using flutamide. No trials comparing combined androgen blockade using bicalutamide with monotherapy met the study selection criteria for the current study. The meta-analysis found a hazard ratio of 1.168 (95% CI, 0.919–1.484) when combined androgen blockade used cyproterone acetate compared with a hazard ratio of 0.926 (95% CI, 0.812–1.056) when combined androgen blockade used an NSAA.
To our knowledge findings favoring combined androgen blockade over monotherapy are few and are of a lesser magnitude than expected based on early nonrandomized studies. The hypothesis that combined androgen blockade provides a greater benefit than monotherapy for a subgroup of advanced prostate carcinoma patients with good prognostic factors is not supported by the available evidence. To date, the main subgroup of interest has been patients with minimal disease and a good performance status. However, the SWOG/ECOG trial (INT 0105),15 which to our knowledge is the largest single study published to date and the only study that stratified randomization by extent of disease, reported no significant difference in survival between combined androgen blockade and monotherapy in these good-risk patients.
The preponderance of evidence, as documented in the current systematic review, found no statistically significant difference in survival between men treated with combined androgen blockade and those treated with monotherapy. This evidence is comprised of 27 randomized controlled trials, including a total of 7987 patients. Among these is the most recent and to our knowledge largest trial by the SWOG/ECOG (INT 0105) (n =1382 patients), which at a short follow-up duration found no significant difference between combined androgen blockade and monotherapy. Additional trials that address the aggregate population of men with advanced prostate carcinoma are unlikely to alter the balance of the evidence.
The evidence favoring combined androgen blockade is comprised of 3 trials, which report increases in median survival ranging from 3.7–7 months: the current meta-analysis of 5-year survival (hazard ratio = 0.871; 95% CI, 0.805–0.942) and prior meta-analyses restricted to studies using NSAAs for combined androgen blockade (hazard ratio = 0.84; 95% CI, 0.76–0.93)13 or those studies using flutamide for combined androgen blockade (hazard ratio = 0.90; 95% CI, 0.79–1.00).14 Although the meta-analyses demonstrate a statistically significant difference in survival at 5 years, the magnitude of the difference appears to be modest.
The current systematic review and meta-analysis demonstrate that, compared with monotherapy, combined androgen blockade does not improve survival at 2 years in men with advanced prostate carcinoma. Results did not appear to differ by the type of therapy, stage of the disease, or quality of the study. The frequency of adverse events appears to be greater with combined androgen blockade than with monotherapy and results more frequently in the withdrawal of therapy. The modest overall survival difference attributed to combined androgen blockade at 5 years must be weighed against the increased risk of adverse events and the potential for adversely affecting overall quality of life. Differences in costs between combined androgen blockade and monotherapy have been addressed by Bayoumi et al.55
Using different statistical techniques, the recently published PCTCG meta-analysis11 concluded that the difference between combined androgen blockade and monotherapy in 5-year survival ranges from 0–5%. When this meta-analysis addressed all trials, the difference in 5-year survival between treatments was 1.8%, but this difference was not statistically significant. Separate analysis of studies of survival when cyproterone acetate was used in combined androgen blockade found survival to be significantly worse than with monotherapy. When studies using cyproterone acetate were excluded, the statistically significant difference in survival between combined androgen blockade (with flutamide or nilutamide) and monotherapy at 5 years was 2.9%. It must be noted that although the current systematic review sought evidence on a variety of outcomes (including survival, adverse events, and quality of life) the PCTCG article addressed only survival.
A limitation of the current systematic review is that the patients currently being considered for androgen suppression differ from those in the majority of trials we reviewed. This is due to the widespread use of PSA measurement to detect and monitor prostate carcinoma patients. Consequently, men now are considered for androgen suppression earlier in the natural history of the disease than was previously the case. To our knowledge, the relative effects of monotherapy and combined androgen blockade remain unknown for contemporary patients.
Evidence collected from patients regarding the effects of various androgen suppression therapies on the patient's quality of life is needed urgently. Such information would increase the value of the patient educational materials used in shared decision-making. Collecting quality of life evidence and comparing results from separate trials will be facilitated when investigators reach agreement with regard to which disease-specific quality of life parameters should be measured and then determine how best to measure them. The reports from two studies54, 56 suggesting that patients and their physicians have markedly different perceptions of the effects of therapy on the quality of life serve to emphasize the urgency of such research. It also is important to evaluate how access to such quality of life information affects patients' choices regarding their treatments.
- 5Androgen withdrawal therapy: new perspectives in the treatment of prostate cancer. In: RhagavanD, ScherRI, LeibelSA, et al., editors. Principles and practice of genitourinary oncology. Philadelphia: Lippincott-Raven, 1997: 583–591., .
- 7Combined androgen blockade. In: VogelzangNJ, ScardinoPT, ShipleyWU et al., editors. Comprehensive textbook of genitourinary oncology. Baltimore: Williams & Wilkins, 1996: 875–884., .
- 12Androgen deprivation strategies for metastatic prostate cancer. In: RhagavanD, ScherRI, LeibelSA, et al., editors. Principles and practice of genitourinary oncology. Philadelphia: Lippincott-Raven, 1997: 591–597., .
- 16Assessing the quality of reports of randomised trials: implications for the conduct of meta-analyses. Health Technol Assess. 1999; 3: 1–98., , , et al.
- 17MulrowCD, OxmanAD, editors. Cochrane Collaboration Handbook [updated 9 December 1996]. Available in: The Cochrane Library [database on disk and CD-ROM]. The Cochrane Collaboration. Issue 1. Oxford: Update Software, 1997.
- 201999 AHCPR Pub. No. 99-E0012. Available from URL: http:// www.ahcpr.gov/clinic/prossumm.html [accessed date], , , et al. Relative effectiveness and cost-effectiveness of methods of androgen suppression in the treatment of advanced prostate cancer. Evidence Report/Technology Assessment: Number 4. Rockville (MD): Agency for Health Care Policy and Research;
- 24Egret for Windows. Cambridge, MA: Cytel Software Corp., 1999.
- 33The final analysis of the EORTC Genito-Urinary Tract Cancer Co-Operative Group phase III clinical trial (protocol 30805) comparing orchidectomy, orchidectomy plus cyproterone acetate and low dose stilboestrol in the management of metastatic carcinoma of the prostate. Eur Urol. 1995; 28: 273–283., , , , , .
- 42A randomized, controlled clinical trial of leuprolide and anandron (LA) vs leuprolide and placebo (LP) for advanced prostate cancer (D2cap) [abstract]. Proc Annu Meet Am Soc Clin Oncol. 1990; 9: A523., , , et al.
- 43Depot leuprorelin acetate alone or with nilutamide in the treatment of metastatic prostate carcinoma: interim report of a multicenter, double-blind, placebo-controlled study [abstract]. Can J Infect. 1995; 6 (Suppl C): 292C., , , .
- 48Clinical benefits of bicalutamide compared with flutamide in combined androgen blockade for patients with advanced prostatic carcinoma: final report of a double-blind, randomized, multicenter trial. Casodex Combination Study Group. Urology. 1997; 50: 330–336., , , et al.
- 50Quality of life (QOL) endpoints in advanced stage prostate cancer: a randomized, double blind study comparing flutamide to placebo in orchiectomized stage D2 prostate patients (PTS) [abstract]. Proc Annu Meet Am Soc Clin Oncol. 1997; 16: A183., , , et al.
- 53Quality of life assessment in prostatic cancer. Prog Clin Biol Res. 1988; 260: 1191–21., .