Abiraterone in Metastatic Prostate Cancer without Previous Chemotherapy

Study Oversight and Conduct

This study was designed by academic and sponsor-employed investigators. The lead academic author initially drafted the manuscript with sponsor input, and all coauthors subsequently provided input and approval. The sponsor provided funding for editorial assistance with an early draft of the manuscript. All authors made the decision to submit the manuscript for publication. The database was held at a third-party contract clinical research organization (CRO), and queries were issued by both the sponsor and the CRO staff. The independent CRO statistician provided the results of analysis to an independent data and safety monitoring committee, whose members were invited by the sponsor. The committee monitored safety at regular intervals and evaluated efficacy and safety at prespecified interim analyses. At the time of unblinding, analyses were performed by statisticians who were employees of the sponsor. The authors assume responsibility for the completeness and integrity of the data and the fidelity of the study to the protocol and statistical analysis plan (available at NEJM.org).

The review boards at all participating institutions approved the study, which was conducted according to the principles of the Declaration of Helsinki, the International Conference on Harmonisation, and the Guidelines for Good Clinical Practice. All patients provided written informed consent.

Patients

Eligibility criteria were an age of 18 years or older; metastatic, histologically or cytologically confirmed adenocarcinoma of the prostate; prostate-specific antigen (PSA) progression according to Prostate Cancer Clinical Trials Working Group 2 (PCWG2) criteria² or radiographic progression in soft tissue or bone with or without PSA progression; ongoing androgen deprivation with a serum testosterone level of less than 50 ng per deciliter (1.7 nmol per liter); an Eastern Cooperative Oncology Group (ECOG) performance status grade of 0 or 1 (asymptomatic or restricted in strenuous activity but ambulatory, respectively); no symptoms or mild symptoms, as defined according to the Brief Pain Inventory–Short Form (BPI-SF) (scores of 0 to 1 [asymptomatic] or 2 to 3 [mildly symptomatic], respectively); and hematologic and chemical laboratory values that met predefined criteria. Previous therapy with an antiandrogen was required. Patients with visceral metastases or patients who had received previous therapy with ketoconazole lasting more than 7 days were excluded.

Study Design and Treatment

In this multinational, double-blind, placebo-controlled study, patients were randomly assigned in a 1:1 ratio to receive abiraterone acetate plus prednisone or placebo plus prednisone. Patients were stratified according to the baseline ECOG performance status grade (0 vs. 1). Patients in the abiraterone–prednisone group received abiraterone at a dose of 1 g (administered as four 250-mg tablets), and patients in the prednisone-alone group received four placebo tablets once daily at least 1 hour before and 2 hours after a meal. All patients received prednisone at a dose of 5 mg orally twice daily. Safety and dosing compliance were evaluated during each study visit, at treatment discontinuation if applicable, and at the end-of-study visit.

End Points

The coprimary efficacy end points were radiographic progression-free survival and overall survival, defined as the time from randomization to death from any cause. Radiographic progression-free survival was determined by an independent radiologist who was unaware of study-group assignments, and dates of death were confirmed. Radiographic progression-free survival was defined as freedom from death from any cause; freedom from progression in soft-tissue lesions as measured with the use of computed tomography (CT) or magnetic resonance imaging (MRI), defined as “progressive disease” according to modified Response Evaluation Criteria in Solid Tumors (RECIST) criteria; or progression on bone scanning according to criteria adapted from the PCWG2 (Table S1 in the Supplementary Appendix).² Changes in PSA level were not included in the definition of radiographic progression-free survival.

The prespecified secondary end points were times to opiate use for cancer-related pain, to initiation of cytotoxic chemotherapy, to a decline in ECOG performance status, and to PSA progression (on the basis of PCWG2 criteria).² Other end points included radiographic progression-free survival as measured by investigators (rather than a blinded review), PSA response rate (≥50% decline in PSA level from baseline), rate of objective response according to RECIST criteria, and health-related quality of life, as measured by means of patients' reports of pain and functional status. An increase in pain was defined as an increase in the baseline pain score at two consecutive visits by 30% or more, as measured by the average of the pain scores on the BPI-SF (range, 0 to 10, with higher scores indicating worse average pain), without a decrease in analgesic use. A decline in functional status was defined as a decline of 10 or more points in the Functional Assessment of Cancer Therapy–Prostate (FACT-P) total score at any visit (range, 0 to 156, with higher scores indicating better overall quality of life).

Assessments

Efficacy assessments included sequential radiographic imaging to assess radiographic progression-free survival (CT or MRI and bone scanning) and measurement of PSA levels.² CT or MRI and bone scanning were performed every 8 weeks during the first 24 weeks and every 12 weeks thereafter. All patients underwent serial monitoring of blood chemical levels, hematologic values, coagulation studies, serum lipids, and kidney function. Cardiac safety was monitored by means of serial electrocardiography. The left ventricular ejection fraction was measured at baseline. Patient-reported outcomes were assessed at baseline and at every visit with the use of the BPI-SF. FACT-P questionnaires were completed every third visit.

Statistical Analysis

The overall level of significance for the study was 0.05, allocated between the coprimary end points of radiographic progression-free survival (0.01) and overall survival (0.04). A single analysis was planned for the coprimary end point of radiographic progression-free survival on the basis of a blinded review by the central radiologist after 378 progression-free events, which would provide a statistical power of 91% to detect a hazard ratio of 0.67 at a two-tailed level of significance of 0.01. The results of subsequent analyses of this end point based on investigator assessment are also reported. For the coprimary end point of overall survival, 773 events were required to detect a hazard ratio of 0.80 at a two-tailed significance level of 0.04 with a statistical power of 85%.

Three interim analyses were planned for overall survival, with the first analysis planned after the observation of approximately 116 of the required 773 events (15%) (in conjunction with the independent review of radiographic progression-free survival), the second analysis planned after 311 events (40%), and the third analysis planned after 425 events (55%); a final analysis was planned for after 773 events had occurred (Table S2 in the Supplementary Appendix). The group-sequential design was used for the overall survival end point with the use of the O'Brien–Fleming boundaries as implemented by the Lan–DeMets alpha spending method (Table S3 in the Supplementary Appendix).

We planned to enroll approximately 1000 patients in the study. The primary statistical method of comparison for the time-to-event end points was the stratified log-rank test with stratification according to the baseline ECOG score. The Cox proportional-hazards model was used to estimate the hazard ratio and its associated confidence interval. The Hochberg procedure was used to adjust for multiplicity testing of the secondary efficacy end points.²³ The strength of association between radiographic progression-free survival and overall survival was evaluated by means of Spearman's correlation coefficient estimated with the use of the Clayton copula.²⁴

Patients and Treatment

From April 2009 through June 2010, we randomly assigned 1088 patients to receive study treatment: abiraterone plus prednisone in 546 patients and placebo plus prednisone in 542 patients (Figure S2 in the Supplementary Appendix). The clinical cutoff date for the blinded central radiologic review of radiographic progression-free survival and the first overall survival interim analysis was December 20, 2010 (at which time 13% of deaths had occurred), and the clinical cutoff date for the second interim analysis of overall survival was December 20, 2011 (at which time 43% of deaths had occurred). The median follow-up duration for all patients was 22.2 months. Baseline demographic characteristics were well balanced between the two study groups (Table S4 in the Supplementary Appendix).

Primary End Points

Radiographic Progression-free Survival

Figure 1. Figure 1. Kaplan–Meier Estimates of Radiographic Progression-free Survival, Overall Survival, and Subgroup Analyses at the Second Interim Analysis.

Panels A and C show data for radiographic progression-free survival on the basis of investigator review, and Panels B and D show data for overall survival. The dashed line in Panels A and B indicates the median. In Panels C and D, the size of the circle reflects the number of patients affected. All analyses were performed with the use of a stratified log-rank test according to the baseline score on the Eastern Cooperative Oncology Group (ECOG) scale (a performance status grade of 0 indicates asymptomatic, and 1 restricted in strenuous activity but ambulatory). Scores on the Brief Pain Inventory–Short Form (BPI-SF) range from 0 to 10, with higher scores indicating worse average pain. ALK-P denotes alkaline phosphatase, LDH lactate dehydrogenase, and PSA prostate-specific antigen.

On the basis of the blinded central radiologic review, at the time of the first interim analysis, treatment with abiraterone plus prednisone, as compared with placebo plus prednisone, resulted in a 57% reduction in the risk of radiographic progression or death (median not reached vs. median of 8.3 months; hazard ratio for abiraterone–prednisone vs. prednisone alone, 0.43; 95% confidence interval [CI], 0.35 to 0.52; P<0.001). At the time of the second interim analysis, the median time to radiographic progression-free survival on the basis of investigator assessment was 16.5 months in the abiraterone–prednisone group and 8.3 months in the prednisone-alone group (hazard ratio, 0.53; 95% CI, 0.45 to 0.62; P<0.001) (Figure 1A). The treatment effect of abiraterone on radiographic progression-free survival was consistently favorable (all hazard ratios, <1.0) across all prespecified subgroups (Figure 1C).

Overall Survival

The planned interim analysis of overall survival was performed after 333 deaths (43% of 773 events) were observed. More deaths were observed in the prednisone-alone group than in the abiraterone–prednisone group (186 of 542 patients [34%] vs. 147 of 546 patients [27%]). Median overall survival was not reached for the abiraterone–prednisone group and was 27.2 months (95% CI, 26.0 to not reached) in the prednisone-alone group. There was a 25% decrease in the risk of death in the abiraterone–prednisone group (hazard ratio, 0.75; 95% CI, 0.61 to 0.93; P=0.01) (Figure 1B), indicating a strong trend toward improved survival with abiraterone–prednisone; however, the prespecified boundary for significance (P≤0.001) was not reached at the observed number of events. The treatment effect of abiraterone on overall survival was consistently favorable (all hazard ratios, <1.0) across all prespecified subgroups (Figure 1D). Radiographic progression-free survival was positively correlated with overall survival, with an estimated correlation coefficient of 0.72.

Secondary End Points

Table 1. Table 1. Prespecified Secondary and Exploratory Efficacy End Points.

Figure 2. Figure 2. Secondary Efficacy End Points.

Shown are the time until a decline in the Eastern Cooperative Oncology Group (ECOG) score by one point or more (Panel A), the time until the initiation of cytotoxic chemotherapy (Panel B), the time until the use of opiates for pain from prostate cancer (Panel C), and the time until prostate-specific antigen (PSA) progression according to Prostate Cancer Clinical Trials Working Group 2 criteria² (Panel D). The dashed line indicates the median. All analyses were performed with the use of a stratified log-rank test according to the baseline ECOG score.

Prespecified secondary and exploratory efficacy end points are summarized in Table 1. Abiraterone–prednisone decreased the risk of decline (by ≥1 point) in ECOG performance-status score by 18%, as compared with prednisone alone (time to decline, 12.3 vs. 10.9 months; hazard ratio for decline, 0.82; 95% CI, 0.71 to 0.94; P=0.005) (Figure 2A). The median time to the initiation of cytotoxic chemotherapy was 25.2 months in the abiraterone–prednisone group and 16.8 months in the prednisone-alone group (hazard ratio, 0.58; 95% CI, 0.49 to 0.69; P<0.001) (Figure 2B). A significant delay in the time to opiate use for cancer-related pain was observed with abiraterone (not reached vs. 23.7 months; hazard ratio, 0.69; 95% CI, 0.57 to 0.83; P<0.001) (Figure 2C). The median time to PSA progression was 11.1 months in the abiraterone–prednisone group and 5.6 months in the prednisone-alone group, a 51% reduction in risk (hazard ratio, 0.49; 95% CI, 0.42 to 0.57; P<0.001) (Figure 2D). On the basis of aggregate efficacy and safety data from the second interim analysis, the data and safety monitoring committee unanimously recommended unblinding the study in February 2012.

Other End Points

The median time to increase in pain was 26.7 months among patients receiving abiraterone–prednisone and 18.4 months among those receiving prednisone alone (hazard ratio, 0.82; 95% CI, 0.67 to 1.00; P=0.049) (Table 1). The median time to a decline in the FACT-P total score was 12.7 months in the abiraterone–prednisone group and 8.3 months in the prednisone-alone group (hazard ratio, 0.78; 95% CI, 0.66 to 0.92; P=0.003). The rates of PSA response and objective response to therapy were significantly higher in the abiraterone–prednisone group than in the prednisone-alone group (Table 1).

Safety

Table 2. Table 2. Adverse Events.

Table 3. Table 3. Adverse Events of Special Interest.

Adverse events are summarized in Table 2 and Table 3. Grade 3 or 4 adverse events were reported in 48% of patients in the abiraterone–prednisone group and 42% of patients in the prednisone-alone group; serious adverse events were reported in 33% and 26% of patients, and adverse events resulting in death were reported in 4% and 2% of patients, respectively. Fatigue, arthralgia, and peripheral edema were among the adverse events reported more frequently in the abiraterone–prednisone group than in the prednisone-alone group. Grade 3 or 4 adverse events classified as hepatotoxicity, consisting primarily of a reversible elevation in aminotransferase levels, were reported in 8% of patients in the abiraterone–prednisone group and 3% of patients in the prednisone-alone group. No patient in either study group died from hepatotoxicity-related adverse events.

The frequency of adverse events resulting in treatment discontinuation was similar in the two study groups. A total of 19% of patients in the abiraterone–prednisone group and 12% of patients in the prednisone-alone group had adverse events leading to dose modification or interruption of study treatment. In the two study groups, the most frequently occurring adverse events resulting in death were those related to disease progression (0.6% of patients in each group). The proportions of patients with grade 3 or 4 serious adverse events were similar in the two groups. Adverse events that were classified as cardiac disorders were reported in 19% of patients in the abiraterone–prednisone group and 16% of those in the prednisone-alone group. Mineralocorticoid-related toxic effects were more common in the abiraterone–prednisone group than in the prednisone-alone group, including hypertension (22% vs. 13%), hypokalemia (17% vs. 13%), and fluid retention or edema (28% vs. 24%), and were mostly grade 1 or 2 adverse events.

In our study involving men with metastatic castration-resistant prostate cancer, abiraterone plus low-dose prednisone resulted in prolonged radiographic progression-free survival (median time to event, 16.5 months vs. 8.3 months; hazard ratio, 0.53), as compared with placebo plus prednisone. Patients receiving abiraterone also had an extended time until the initiation of opiate analgesia, treatment with cytotoxic chemotherapy, or a decline in performance status, as well as delays in PSA progression, onset of pain, and decline in health-related quality of life. The proportion of men in the abiraterone–prednisone group with a PSA response and the time to PSA progression are consistent with outcomes reported in earlier phase 1 or 2 studies of abiraterone.^20-22 In addition, a strong trend toward improved survival (hazard ratio, 0.75) was evident at the time at which 43% of the prespecified total number of events required for the final analysis had occurred. This consistent pattern of benefit resulted in the unanimous decision of the data and safety monitoring committee to recommend unblinding of the study and crossover of patients in the prednisone-alone group to abiraterone treatment.

Despite the various therapies available for men with metastatic castration-resistant prostate cancer, a need remains for effective nontoxic agents that can improve and maintain the quality and duration of life while preventing the morbidity associated with disease progression.²⁵ Second-line hormonal manipulation with antiandrogens, diethylstilbestrol, and ketoconazole has long been used on the sole basis of symptom relief and PSA-level response data.⁸ This pattern of use has persisted despite the availability of two new agents with a survival benefit: docetaxel, the use of which is limited by toxic effects, and sipuleucel-T, the use of which is limited by a lack of demonstrable antitumor activity. The durable antitumor effect and safety profile of abiraterone confirms earlier experience that it can be used long term without concern for life-threatening toxic effects.^21,22

Glucocorticoids have beneficial effects in patients with metastatic castration-resistant prostate cancer, and prednisone has been an active comparison agent in randomized trials for decades.^4,7,19 Our data show that targeting persistent extragonadal androgen synthesis²⁶ leads to benefits that exceed those with standard prednisone therapy used in current clinical trials. An additional notable finding is that the median overall survival of 27.2 months with prednisone alone is the longest survival prospectively observed in this patient population, possibly a consequence of antitumor activity of the prednisone control and the activity of subsequent effective therapies.

In addition to the marked improvement in radiographic progression-free survival, treatment with abiraterone was associated with a trend toward improved overall survival. Evidence of the magnitude of the survival benefit of abiraterone–prednisone, as compared with prednisone alone, was that treatment effects were consistently favorable across all prespecified patient subgroups, including older men and those with a decreased performance status, increased pain, and increased disease burden (Figure 1D). The use of abiraterone after crossover among patients originally assigned to the prednisone-alone group may affect the ability to show statistical significance in subsequent analyses of overall survival. Despite the high disease burden and the proportion of patients with high-grade tumors (Gleason score, ≥8) who were enrolled, the survival curves did not separate until after approximately 12 months. This finding can be ascribed to the use of an active prednisone control and the low rate of early death in asymptomatic or mildly symptomatic patients with metastatic castration-resistant cancer.

Early deaths related to cancer may occur in patients with a tumor phenotype against which androgen modulation may have little effect. Although we do not know the effectiveness of therapies (including abiraterone) that were used after termination of the study treatment, the prevalence of subsequent therapy was higher in the prednisone-alone group than in the abiraterone–prednisone group (60% and 44%, respectively) (Table S5 in the Supplementary Appendix). The most common subsequent therapy in the two groups was docetaxel. Between-group disparities in subsequent therapies may be attributable to the greater number of patients in the abiraterone–prednisone group who continued to receive the drug, as compared with the prednisone-alone group: 166 of 542 patients (31%) in the abiraterone–prednisone group vs. 86 of 540 patients (16%) in the prednisone-alone group.

The safety of abiraterone in this study was similar to that previously reported in men with metastatic castration-resistant prostate cancer and disease progression after docetaxel chemotherapy.¹⁹ No toxic effects unique to this patient population were identified (a finding that was consistent with previous studies), despite a longer duration of abiraterone–prednisone treatment. Liver-function abnormalities (typically seen in the first 3 months of therapy) and cardiac toxic effects were more common in the abiraterone-treated patients than in the prednisone-alone group. Cardiac abnormalities tended to appear later. Discontinuation of therapy because of toxicity occurred in 10% of patients in the abiraterone–prednisone group and in 9% of patients in the prednisone-alone group.

In summary, the results show benefit from the use of abiraterone in patients with asymptomatic or mildly symptomatic metastatic castration-resistant prostate cancer who have not received previous chemotherapy. These findings include increased rates of radiographic progression-free survival and overall survival, as well as clinically meaningful secondary end points, such as delays in the use of opiates for pain and chemotherapy and patient-reported outcomes related to health-related quality of life.

The authors' affiliations are as follows: the Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco (C.J.R., E.J.S.); Massachusetts General Hospital Cancer Center (M.R.S.) and Dana–Farber Cancer Institute (M.-E.T.) — both in Boston; the Institute of Cancer Research and the Royal Marsden Hospital, Sutton, United Kingdom (J.S.B.); Janssen Research and Development (A.M., Y.P., T.K., T.G.) and David Geffen School of Medicine, University of California (A.P.) — both in Los Angeles; M.D. Anderson Cancer Center, Houston (C.J.L., E.E.); St. George Private Hospital, Kogarah (P.S.), Haematology and Oncology Clinics of Australia, Brisbane (P.M.), and St. John of God Hospital, Subiaco (S.N.) — all in Australia; Institut Gustave Roussy, University of Paris Sud, Villejuif, France (K.F.); Institut Català d'Oncologia de l'Hospitalet (J.M.P.) and Universitari Vall d'Hebron (J.C.) — both in Barcelona; Radboud University Medical Center, Nijmegen (P.F.A.M.), and VU University Medical Center, Amsterdam (W.R.G.) — both in the Netherlands; Memorial Sloan-Kettering Cancer Center, New York (E.B., H.I.S., D.E.R.); University of Montreal, Montreal (F.S.), Juravinski Cancer Centre Hamilton Health Services, Hamilton, ON (S.D.M.), and London Health Sciences Centre, London, ON (E.W.) — all in Canada; Hospital Network Antwerp (ZNA)–Middelheim, Antwerp (D.S.), and University Hospital Leuven, Leuven (H.V.P.) — both in Belgium; Urologikum Hamburg, Hamburg, Germany (H.S.); University of Colorado Cancer Center, Aurora (T.W.F.); Duke University Hospital Medical Center, Durham, NC (D.J.G.); University of Washington, Seattle (E.Y.Y., C.S.H.); and Department of Clinical Therapeutics, University of Athens Medical School, Alexandra Hospital, Athens (E.E.).