Endothelin-1 and the endothelin A (ETA) receptor have been implicated in prostate cancer progression in bone. This study aimed to determine whether the specific ETA receptor antagonist, zibotentan, prolonged overall survival (OS) in patients with castration-resistant prostate cancer and bone metastases who were pain-free or mildly symptomatic for pain.
Patients were randomized 1:1 to zibotentan 10 mg/day or placebo, plus standard prostate cancer treatment. The primary endpoint was OS. Secondary endpoints included times to pain progression, chemotherapy use, new bone metastases, and safety. Efficacy endpoints were analyzed using a log-rank test.
A total of 594 patients were randomized (zibotentan, n = 299; placebo, n = 295). Median OS was 24.5 months in zibotentan-treated patients versus 22.5 months for placebo, but the difference did not reach statistical significance (hazard ratio, 0.87; 95.2% confidence interval, 0.69-1.10; P = .240). No statistically significant differences were observed for any secondary efficacy endpoints. Peripheral edema (44%) and headache (31%) were the most commonly reported adverse events in the zibotentan group. Cardiac failure events were higher in the zibotentan group than placebo (any grade, 5.7% and 1.7%; Common Terminology Criteria for Adverse Events grade ≥3, 3.0% and 1.0%, respectively); these were manageable and reversible.
Prostate cancer is the second most frequently diagnosed cancer worldwide and accounts for almost 10% of all male cancer deaths in the developed world.1 Although death rates for patients with prostate cancer have been decreasing in many developed countries, in part because of improved early detection and treatment options,1 patients with castration-resistant prostate cancer (CRPC) have a poor prognosis with limited therapeutic options.2 The disease state is defined by evidence of prostate cancer progression, despite serum testosterone being controlled below a castrate level.2
The potent vasoconstrictor, endothelin-1 (ET-1) appears to contribute to prostate cancer progression, particularly in bone.3 ET-1 has been detected in nearly every primary and metastatic prostate cancer, independent of the hormonal milieu.4 Acting through the endothelin A (ETA) receptor, ET-1 is causal in the osteoblastic response of bone, contributing to the vicious cycle of cancer progression.5 Neutral endopeptidase, the enzyme responsible for local clearance of bioactive peptides such as ET-1, is significantly decreased with androgen deprivation,6 adding to the pathologic effects of ectopic ET-1 secretion in CRPC. Acting through the ETA receptor, ET-1 is an inhibitor of apoptosis in a variety of cell types, including prostate cancer.7 Resistance to androgen deprivation–induced apoptosis is the distinguishing feature of CRPC. Collectively, these observations support the rationale for specific inhibition of the ETA receptor for the treatment of patients with CRPC.
Zibotentan (ZD4054), an oral-specific ETA receptor antagonist, was studied in a phase 2 study of patients with CRPC and bone metastases who were pain-free or mildly symptomatic for pain.8, 9 The primary endpoint, time to disease progression, was not met, but a signal for prolonged overall survival (OS) was observed in patients receiving zibotentan compared with placebo. The aim of this phase 3 study was to determine the efficacy and safety of zibotentan in a similar but larger population of patients with metastatic CRPC (clinicaltrials.gov, identifier NCT00554229). This study was part of the large international phase 3 ENTHUSE (EndoTHelin A USE) program, which also included studies of zibotentan in combination with docetaxel in metastatic CRPC and as monotherapy in patients with nonmetastatic CRPC.
MATERIALS AND METHODS
Eligible patients had pathologically-confirmed prostate cancer who had undergone surgical or continuous medical castration (gonadotropin-releasing hormone analogues, luteinizing hormone-releasing hormone analogues, or estramustine) with serum testosterone ≤2.4 nmol/L (70 ng/dL; lower limit of quantification in many centers taking part in this trial) and had metastatic disease with documented biochemical progression. Biochemical progression was defined as ≥2 stepwise increases in prostate-specific antigen (PSA) levels over a period of ≥1 month with ≥14 days between each measurement; the last PSA measurement had to be an increase of ≥50% or an absolute increase of ≥10 ng/mL over the initial PSA value. The final PSA value had to be ≥1.2 ng/mL in patients who had undergone a radical prostatectomy and ≥5 ng/mL in all other patients. Other inclusion criteria were evidence of bone metastases on radionuclide bone scan (disease involvement <75% of the spine, pelvis, and ribs; confirmation by computed tomography, magnetic resonance imaging scan, or x-ray was required for ≤3 lesions), asymptomatic/mild pain from prostate cancer (defined as a score of ≤2 in the “worst pain” item of the Brief Pain Inventory [BPI]), World Health Organization performance status of 0/1, and adequate hematologic, hepatic, and renal function. Patients were ineligible if they had received any of the following therapies: cytotoxic chemotherapy for recurrent prostate cancer; targeted anticancer therapies; endothelin receptor antagonists; systemic retinoids or potent cytochrome P450 inducers within 2 weeks of starting study treatment; radiotherapy to bone lesions or prostatic bed, or any anticancer investigational drug within 4 weeks of starting study treatment; radionuclide therapy within 12 weeks of starting study treatment. Additional exclusion criteria were evidence of impending spinal cord compression, known/suspected central nervous system metastases, cardiac failure (New York Heart Association class ≥II) or myocardial infarction within the past 6 months, hemoglobin <9 g/dL, or creatinine clearance <50 mL/minute.
All patients provided written informed consent. The institutional review boards or independent ethics committees of all investigational sites approved the protocol.
Patients were stratified by center and randomized 1:1, using an interactive voice response system, to receive zibotentan 10 mg/day po (by mouth) or placebo (Fig. 1). Study treatment was blinded using the unique code generated during randomization. Patients could receive the following standard supportive/palliative care: regular follow-up; chemotherapy; symptomatic therapy for pain control; steroids; bisphosphonates and second-line hormonal manipulations including anti-androgens, estramustine, ketoconazole, megestrol, and estradiol. Other experimental agents were not permitted. Patients continued randomized treatment until a withdrawal criterion was met.
Study Endpoints and Evaluations
The primary endpoint was OS, defined as time to death from any cause. Key secondary efficacy endpoints included time to pain progression (time to change in pain due to metastases [defined as an increase from baseline to a minimum worst pain BPI score of 5 with no decrease in analgesic use] or death); time to opiate use (opiate use for ≥1 week for pain due to prostate cancer metastasis); skeletal-related events (pathological fracture; vertebral compression fracture not related to trauma; prophylactic surgery or radiation for impending fracture or spinal cord compression; spinal cord compression); time to chemotherapy use, or death; time to new bone metastases (time to appearance of ≥4 new bony lesions confirmed by bone scan or death); progression-free survival (PFS; time from randomization until clinical progression of disease [defined as any of the following: increased pain, as defined above; skeletal-related event; objective progression of visceral or nodal disease (modified Response Evaluation Criteria In Solid Tumors [RECIST] version 1.0); ≥4 new bony lesions confirmed by bone scan; or death from any cause in the absence of progression]); and time to PSA progression (time to the first PSA value ≥50% from baseline, seen in ≥2 consecutive PSA values). Health-related quality of life (HRQoL) was assessed according to the functional well-being domain of Functional Assessment of Cancer Therapy for Prostate cancer (FACT-P) and total FACT-P score. Symptoms were assessed using the 8-item FACT-P symptom index (FAPSI-8). Adverse events (AEs) were monitored throughout the study and were graded according to Common Terminology Criteria for Adverse Events (CTCAE), version 3.0. Safety data were reviewed regularly by an independent data monitoring committee (IDMC).
Recruitment of 580 patients was planned; ≥263 deaths were required for formal analysis. Based on a recruitment period of 18 months and a predicted median OS for the placebo group of approximately 19 months, 263 deaths were estimated to occur approximately 30 months after the first patient had entered the study. If the true hazard ratio (HR) for zibotentan versus placebo was 0.67, the analysis would have 90% power to demonstrate a statistically significant effect in OS at the 5% level. For the primary endpoint of OS, the nominal significance level of 5% was adjusted to 4.8% to allow for a single interim analysis.
The efficacy and HRQoL analyses included all patients randomized and the safety analyses included all patients who received ≥1 dose of study medication. Efficacy endpoints were analyzed using a log-rank test.
Between November 20, 2007, and February 13, 2009, 594 patients from 182 sites in 27 countries were randomized to zibotentan 10 mg/day (n = 299) or placebo (n = 295) (Fig. 2). Demographics and baseline characteristics were similar between the treatment groups (Table 1).
Table 1. Patient Demographics and Baseline Characteristics
Placebo (n = 295)
Zibotentan 10 mg (n = 299)
Abbreviations: PSA, prostate-specific antigen; RECIST, Response Evaluation Criteria In Solid Tumors; WHO, World Health Organization.
Median age, y (range)
Race, n (%)
WHO performance status, n (%)
Previous cancer treatment, n (%)
Median PSA, ng/mL (range)
Bone metastases, n (%)
RECIST soft tissue lesions, n (%)
No evaluable lesion
Target and nontarget
Brief Pain Inventory score – pain at its worst, n (%)
No statistically significant improvement in OS was observed (HR, 0.87; 95.2% confidence interval [CI], 0.69-1.10; P = .240; median OS, 24.5 [zibotentan] versus 22.5 [placebo] months) (Fig. 3). The OS results were consistent across the prognostic and geographic subgroups evaluated, including age (<75, ≥75 years), PSA levels (<median value, ≥median value), bone metastases (≤4, 5-20, ≥21), and bisphosphonate use (Fig. 4).
Secondary Efficacy Endpoints
No statistically significant differences were observed between treatments for time to pain progression, time to chemotherapy use, time to new bone metastases, and PFS (Table 2). At the time of data cutoff, the proportion of patients who had been treated with chemotherapy subsequent to randomization was identical in both arms (72.9%; zibotentan, n = 218/299; placebo, n = 215/295).
Table 2. Summary of Secondary Efficacy Endpoint Analyses
Hazard Ratio (95% CI)
Zibotentan 10 mg
Abbreviation: CI, confidence interval.
Time to pain progression
P = .521
Time to chemotherapy use
P = .421
Time to new bone metastases
P = .201
P = .894
There were no statistically significant differences in time to deterioration of either HRQoL (FACT-P; HR, 1.01; 95% CI, 0.85-1.21; P = .890) or symptoms (FAPSI-8; HR, 1.03; 95% CI, 0.86-1.24; P = .707) between treatment groups.
Zibotentan treatment was generally well tolerated; the median duration of treatment exposure was 9.5 (range, 0.1-30.6) months for zibotentan and 11.0 (range, 0.1-29.0) months for placebo.
Common AEs that occurred more frequently in the zibotentan arm included peripheral edema (44.3% versus 19.0%), headache (31.2% versus 12.5%) and nasal congestion (16.1% versus 3.1%; Table 3). Both arms showed a similar incidence of grade ≥3 events (42.3% zibotentan, 40.7% placebo; Table 4). The most common grade ≥3 events were anemia (5.4% zibotentan, 2.7% placebo) and neutropenia (3.4% zibotentan, 4.4% placebo). Neutropenia events were generally reported in patients receiving chemotherapy. Overall, 32.9% and 35.6% of patients in the zibotentan and placebo groups, respectively, had serious AEs (SAEs); the most frequently reported in the zibotentan group were pneumonia, cardiac failure, and anemia (2.0% of patients each), and pulmonary embolism (1.7%). There were 14 deaths due to AEs in the zibotentan group versus 26 with placebo. AEs leading to treatment discontinuation occurred in 23.2% of patients receiving zibotentan and 12.9% receiving placebo (Table 5); this difference primarily reflects the higher rates of peripheral edema and nasal congestion leading to discontinuation in the zibotentan arm.
Table 3. Adverse Events (Any CTCAE Grade) Experienced by ≥10% Patients in Either Treatment Arm
Number of Patients (%)
Placebo (n = 295)
Zibotentan 10 mg (n = 299)
Abbreviations: CTCAE, Common Terminology Criteria for Adverse Events.
Any adverse event
Table 4. Grade ≥3 Events Experienced by ≥1% Patients in the Zibotentan Group
Number of Patients (%)
Placebo (n = 295)
Zibotentan 10 mg (n = 299)
Any grade ≥3 event
Table 5. Adverse Events Resulting in Discontinuation in >1% of Patients in Either Treatment Arm
Number of Patients (%)
Zibotentan 10 mg
(n = 295)
(n = 299)
Any adverse event leading to discontinuation
Cardiac failure events were higher in the zibotentan group (any grade, 5.7%; grade ≥3, 3.0%) than placebo (any grade, 1.7%; grade ≥3, 1.0%). In the zibotentan group, of the 17 patients with cardiac failure events, all but 1 patient received treatment and most events (14) resolved either without interrupting treatment (n = 4), or after temporary or permanent discontinuation (n = 4 and 6, respectively); 8 of 17 events were considered to be SAEs. There was a sustained 5 mm Hg greater reduction in mean systolic and diastolic blood pressure in the zibotentan group compared with placebo, but no increase in pulse or weight. Generally, there were no differences in hematology laboratory parameters between the treatment arms. However, a sustained decrease in mean hemoglobin levels of 1.0 to 1.5 g/dL was reported in the zibotentan group.
Overall, there were 283 deaths: zibotentan, n = 136 (108 due to prostate cancer); placebo, n = 147 (113 due to prostate cancer).
The addition of zibotentan 10 mg/day to best supportive care did not result in a significant improvement in OS in this study of patients with CRPC and bone metastases who were pain-free or mildly symptomatic for pain (HR, 0.87; CI, 0.69-1.10; P = .24; median OS, 24.5 months vs 22.5 months). These data are disappointing, and confound the promising OS signal observed in a randomized phase 2 study of zibotentan in a similar patient population.9 There are, however, notable differences between the phase 2 and phase 3 trials, which should be considered when designing and conducting future clinical trials in CRPC.
First, the primary endpoint in the phase 2 study was time to progression, not OS.8 After observing an increased OS at the primary analysis, two additional analyses were planned for the times at which a minimum of 110 (data cutoff February 2007) and 200 (data cutoff December 2008) deaths had occurred.8, 9 This phase 3 study was designed and initiated, based on the outcome of the second analysis (110 deaths), where median OS was 17.3 months for placebo and 24.5 months for zibotentan 10 mg. At the final analysis (200 deaths), the difference in median OS had narrowed: 19.9 months for placebo, 23.5 months for zibotentan 10 mg.9 In hindsight, the phase 2 study would have benefited from an OS primary endpoint and a more complete follow-up; the phase 3 study may also provide a good example of regression to the mean. Cautious and conservative statistical assumptions are appropriate when designing a pivotal trial with a previously untested primary endpoint. A much larger study would have been required to demonstrate significance and, even in that setting, it is reasonable to question whether a 2-month OS difference is clinically meaningful in CRPC.
Second, the placebo group in the phase 3 study survived longer than expected. The trial was designed based on the published median OS for patients with metastatic CRPC of approximately 19 months.10 The better survival is consistent with other, more recent, data from phase 3 trials in CRPC: in a sipuleucel-T study, median OS for placebo was 21.7 months.11 As defined—CRPC patients with bone metastases who were pain-free or mildly symptomatic for pain—the inclusion criteria may have been too broad to select a homogeneous population for study. Future trial designs will benefit from an improved understanding of CRPC disease progression.
Third, median drug exposure in the phase 2 study was shorter (3.7 months for placebo, 4.3 months for zibotentan 10 mg) and data about subsequent systemic anticancer therapy usage were available for less than half of patients.8, 9 Subsequent use of anticancer therapy following start of study treatment in the phase 2 trial was balanced between arms (placebo, 44.9%; zibotentan 10 mg, 52.3%) and was lower than in this phase 3 trial (72.9%, both arms). In our study, similar proportions of patients received subsequent anticancer therapy, including chemotherapy (placebo, 26.8%; zibotentan, 24.4%), radiotherapy (22.0%; 21.4%), hormonal therapy (85.8%; 87.6%), and immunotherapy (1.0%; 0.3%), with no apparent effect on the overall outcome of this study. It is not possible to determine the effects of exposure, nature and timing of chemotherapy between the 2 studies, but emphasizes the challenge in designing clinical trials with primary OS endpoints in CRPC. This is particularly true now that there are 6 agents (docetaxel,12 sipuleucel-T,11 cabazitaxel,13 abiraterone,14 MDV3100,15 and radium-223 chloride16) approved and/or reporting improvements in OS in CRPC.
Fourth, there were demographic differences between the studies. The phase 3 trial included East Asian patients (from China, Hong Kong, Japan, Singapore, South Korea, and Taiwan) and the prespecified subgroup analysis of overall survival showed the treatment effect hazard ratio was numerically worse in the overall Asian population compared with patients of other ethnic (predominantly Caucasian) backgrounds. However, this observation should be interpreted with caution, given the wide and overlapping confidence intervals, and the absence of data on the substratified Asian population. Interestingly, epidemiological studies have shown Asian populations with prostate cancer, whether residing in Asia or in the United States, to have better OS rates than Caucasian patient populations located in the same geographical region, which could indicate a genetic rather than geographical basis for these differences.17, 18 The influence of genetic background on CRPC progression and response to therapy remains an unquantifiable variable in clinical trial design and holds the greatest promise for understanding the disease. Collectively, these differences highlight the need for new and class-specific endpoints, more reflective of mechanisms of action, to better facilitate the design of clinical trials and inform treatment decisions.19
Despite dramatic effects on cancer-bone interactions in preclinical models,5 ETA receptor antagonists have had disappointing results in patients with CRPC in the phase 3 setting. Every study of an ETA receptor antagonist in CRPC has been consistent in demonstrating a modest but nonsignificant effect favoring the agent. The selective ETA receptor antagonist atrasentan, which also reported positive results in a number of phase 2 studies,20-23 failed to significantly impact disease progression in phase 3 studies in patients with either metastatic or nonmetastatic CRPC.24, 25 The atrasentan studies also demonstrated the difficulty of conducting trials in this disease state. For example, both trials were underpowered to demonstrate a difference. Asymptomatic radiographic progression in metastatic CRPC was unexpectedly rapid and median progression was reached at the first 3-month assessment.24 In nonmetastatic CRPC, progression to metastases is protracted, and premature discontinuation, particularly in the United States, was unexpectedly high; this appeared to be largely driven by anxiety about PSA-only progression.25 The failure of the metastatic atrasentan study informed the recommendations developed by the Prostate Cancer Clinical Trials Working Group (PCWG2), where confirmatory scans are recommended for new bone lesions.26
Zibotentan had an acceptable safety profile and no new concerns were identified. The most commonly reported AEs were peripheral edema and headache, which are consistent with the vasodilator properties of ETA receptor antagonists. Congestive heart failure is a common comorbidity in elderly patients with prostate cancer, present in approximately 10% of patients 12 months before diagnosis of prostate cancer.27 Cardiac failure events, actively solicited due to a signal in the phase 2 study,9 although more commonly reported in zibotentan-treated patients, were manageable and nonfatal, with patients making full recoveries in most cases. The incidence of cardiac failure in the current study was similar to that observed in both the earlier phase 2 study of zibotentan and the atrasentan phase 3 study in a similar patient population.9, 24
Several emerging treatment approaches, including novel targeted agents, hormonal therapies, bone-targeted agents, and immune modulators, are currently in advanced development or have received regulatory approval for patients with metastatic CRPC.28 The autologous vaccine sipuleucel-T has been approved to treat asymptomatic or minimally symptomatic metastatic CRPC.11 The CYP17 inhibitor abiraterone acetate14 and the tubulin-binding taxane cabazitaxel13 have been approved for the treatment of patients with metastatic CRPC in the postdocetaxel setting. Denosumab, a RANK ligand inhibitor, has been approved for the prevention of skeletal-related events in metastatic CRPC.29 Radium-223 chloride and MDV3100 have not yet received regulatory approval, but the ALSYMPCA (radium 223)16 and AFFIRM (MDV3100)15 phase 3 trials in the advanced disease setting were stopped early based on the results of preplanned interim analyses, demonstrating significant survival benefit over the comparator arm in each case. Numerous other phase 3 studies are ongoing with a variety of treatment modalities, including MDV3100 and abiraterone in chemotherapy-naive patients and orteronel (endocrine therapies), aflibercept and tasquinimod (antiangiogenic), dasatinib (Src inhibitor), ipilimumab (CTLA-4 inhibitor), and OGX-011 (clusterin inhibitor).28, 30-35 The development of all these new agents is the result of an increased understanding of molecular mechanisms associated with disease progression. To ensure the success of future phase 3 trials in CRPC, efforts should be made to optimize individual assessment of predictive biomarkers of drug efficacy and to better select patients most likely to benefit from treatment.
In conclusion, zibotentan 10 mg/day plus standard of care did not result in a significant improvement in OS in patients with CRPC and bone metastases who were pain-free or mildly symptomatic for pain and had not received chemotherapy for metastatic disease. More recently, the 2 additional phase 3 studies in the zibotentan ENTHUSE program have also ended. The study investigating zibotentan as a monotherapy in patients with nonmetastatic CRPC (NCT00626548) was stopped following an interim efficacy review by the IDMC, and the study of zibotentan in combination with docetaxel in patients with metastatic CRPC (NCT00617669) failed to reach its primary endpoint of OS. As a result, zibotentan is no longer being investigated as a potential treatment for patients with prostate cancer.
We thank the ENTHUSE Program Institutions for participating in the study and Dr Claire Routley from Mudskipper Bioscience who provided medical writing assistance funded by AstraZeneca.
This study was funded by AstraZeneca.
CONFLICT OF INTEREST DISCLOSUREs
J.B. Nelson, K. Fizazi, K. Miller, C. Higano, J.W. Moul, and H. Akaza have attended advisory boards for AstraZeneca. J.B. Nelson, K. Fizazi, K. Miller, and J.W. Moul have received honoraria from AstraZeneca. J.W. Moul owns stock in AstraZeneca. M. Gleave is chief scientific officer for and owns stock in OncoGenex Technologies, has attended advisory boards for, and received research funding from OncoGenex Technologies, Sanofi-Aventis, AstraZeneca, and Pfizer (research funding only). T. Morris, S. McIntosh, and K. Pemberton are employees of, and own stock in AstraZeneca.