Final safety and efficacy analysis of the specific endothelin A receptor antagonist zibotentan (ZD4054) in patients with metastatic castration-resistant prostate cancer and bone metastases who were pain-free or mildly symptomatic for pain: a double-blind, placebo-controlled, randomized Phase II trial

Authors


Nicholas D. James, Division of Cancer Studies, University of Birmingham, Birmingham B15 2TT, UK.
e-mail: n.d.james@bham.ac.uk

Abstract

Study Type – Therapy (RCT)
Level of Evidence 1b

OBJECTIVES

To report the final analysis of a Phase II trial, which investigated the safety and efficacy of the specific endothelin A receptor antagonist zibotentan (AstraZeneca, Macclesfield, UK) in patients with metastatic castration-resistant prostate cancer (CRPC).

PATIENTS AND METHODS

Patients with CRPC and bone metastases who were pain free or mildly symptomatic for pain were randomized to receive once-daily oral tablets of zibotentan 10 mg, 15 mg or placebo. The primary endpoint was the time to progression and secondary endpoints included overall survival, change in the number of bone metastases, and safety.

RESULTS

In total, 312 patients were randomized (placebo, n= 107; zibotentan 10 mg, n= 107; zibotentan 15 mg, n= 98). The median duration of study treatment and median follow-up time were 4 and 22 months, respectively. At the final analysis, there were no statistical differences of the primary outcome of time to progression between treatment groups, although an improvement in overall survival was observed in the zibotentan groups compared to placebo. Consistent with the previous analyses for overall survival, hazard ratios (HRs) of less than one were sustained for both zibotentan 15 mg (HR, 0.76; 80% CI, 0.61–0.94; P= 0.103) and 10 mg (HR, 0.83; 80% CI, 0.67–1.02; P= 0.254). The most commonly reported adverse events considered to be related to zibotentan treatment were peripheral oedema, headache and nasal congestion.

CONCLUSIONS

The results obtained in the present study support endothelin A receptor antagonism as an approach for treating patients with CRPC. To confirm the survival signal observed in the present study, zibotentan is being investigated further in the ENdoTHelin A USE (ENTHUSE) Phase III clinical trial programme.

Abbreviations
AE

adverse event

CRPC

castration-resistant prostate cancer

CTCAE

Common Terminology Criteria for Adverse Events

ET

endothelin

HR

hazard ratio

PSA

prostate specific antigen

RECIST

Response Evaluation Criteria In Solid Tumours

SAE

serious adverse event

INTRODUCTION

Prostate cancer is the second leading cause of cancer-related death in men in the Western world and this disease accounted for an estimated 25% of newly-diagnosed cancers in the USA in 2009 [1,2]. The current mainstay of treatment for patients with advanced prostate cancer is hormone-based therapy, such as androgen deprivation therapy, which has been shown to be effective for prolonged periods [3]. However, disease progression occurs in most patients whose prostate cancer eventually becomes castration resistant (defined as a rising prostate specific antigen (PSA) level, despite serum testosterone controlled below a castrate level), which is associated with a poor prognosis [3].

In recent years, there has been interest in how best to describe patients with a rising PSA despite castration status. The terms androgen independent, hormone refractory or hormone resistant have been used, often interchangeably in clinical practice. However, a new descriptor, ‘castration-resistant prostate cancer’ (CRPC) has emerged, and provides the most accurate description of this patient population.

Unfortunately, treatment options are limited after progression to CRPC. Chemotherapy is the only treatment with a proven survival benefit for patients with metastatic CRPC, with secondary hormone manipulation being the only option for non-metastatic disease [3]. Therefore, there is a major unmet need for new therapies that can improve survival at the same time as maintaining or improving the health-related quality of life of patients with CRPC.

The growth factor endothelin (ET)-1 is considered to play an important role in regulating the growth and proliferation of a variety of tumours, including prostate cancer [4]. Activation of the ETA receptor by ET-1 promotes tumour cell growth and survival, mediating processes involved in tumour invasion and metastasis, angiogenesis and the inhibition of apoptosis [5]. By contrast, activation of the ETB receptor induces cell death through apoptosis, which contributes to the inhibition of tumour progression [6]. The ETB receptor also mediates clearance of ET-1 [7]. These findings support the rationale for specific inhibition of the ETA receptor as a viable treatment strategy for patients with CRPC.

Zibotentan (ZD4054, AstraZeneca, Macclesfield, UK) is an oral, specific ETA receptor antagonist with no detectable activity at the ETB receptor [8]. Preclinical studies in a number of cancer models have shown that zibotentan inhibits cell invasion [9], proliferation [10] and metastases [11]. At single doses up to 240 mg in healthy volunteers, zibotentan produced no increase in plasma levels of ET-1, confirming its specificity for the ETA receptor in the clinical setting [12]. Initial analyses (April 2006 and February 2007) of data from this Phase II clinical trial of zibotentan in patients with CRPC and bone metastases who were pain free or mildly symptomatic for pain, showed that zibotentan was well tolerated and associated with a promising improvement in overall survival compared to placebo [13]. Exploratory findings suggest that zibotentan has no detrimental effect on health-related quality of life. However, the absence of post-progression follow-up meant that it was not possible to determine any differential impact on health-related quality of life between treatment groups [14].

This present study reports the final analysis (data cut-off December 2008) of this Phase II clinical trial investigating the safety and efficacy of zibotentan in patients with CRPC and bone metastases who were pain free or mildly symptomatic for pain.

MATERIALS AND METHODS

STUDY DESIGN AND PATIENTS

This was a Phase II, double-blind, placebo-controlled, parallel-group, multicentre, randomized trial, which was performed in 65 centres in 14 countries across Europe, North America, Australasia and South East Asia. The study was performed in accordance with the Declaration of Helsinki and Good Clinical Practice. All patients provided their written informed consent, and appropriate local institutional review board approval was obtained.

The methods for this study have previously been described in detail [13]. In brief, the study recruited men aged ≥18 years with CRPC and bone metastases who were pain free or mildly symptomatic for pain (defined as no requirement for opiate analgesia) and who had not received previous chemotherapy. Patients had documented evidence of bone metastases, were surgically or continuously medically castrated with serum testosterone ≤2.4 mmol/L with rising PSA levels, and life expectancy of at least 6 months. Exclusion criteria included: radiotherapy to bone metastases within 4 weeks of randomization; previous cytotoxic chemotherapy (estramustine was considered oestrogen therapy, and therefore allowed); antiandrogens within 4 weeks (flutamide) or 6 weeks (bicalutamide, nilutamide) of randomization; and radionuclide treatment of bone metastases within 12 weeks of randomization.

Patients were randomly assigned once-daily oral tablets of zibotentan 10 mg, 15 mg or placebo in a 1:1:1 ratio under a block randomization scheme stratified by centre (Fig. 1). Dose selection was based on the maximum tolerated dose in a multiple-ascending dose study in men with metastatic prostate cancer (15 mg), and the lower of two doses shown to reduce ET-1-induced forearm blood flow in healthy volunteers (10 mg) [15]. All patients received zibotentan or placebo on a background of best supportive care according to local practice. This trial was registered with ClinicalTrials.gov, Unique ID NCT00090363.

Figure 1.

Study population flowchart.

STUDY OBJECTIVES

The primary objective of the present study was to assess the effect of zibotentan on time to disease progression, a composite endpoint that was assessed from the date of randomization to the date of any of the following events: clinical progression (defined as symptoms requiring chemotherapy, radiotherapy, surgery excluding transurethral resection with negative histology, or initiation of a new hormone-based therapy); cancer pain requiring opiates; objective progression of soft tissue metastasis according to Response Evaluation Criteria In Solid Tumours (RECIST); or death in the absence of progression. PSA progression or change in the number or appearance of bone lesions assessed by scintigraphic imaging did not count as progression events. Secondary objectives included overall survival, time to PSA progression, tumour response rate according to RECIST [16], change in number of bone metastases and safety and tolerability.

ASSESSMENTS

Time to progression was assessed from the date of randomization to the date of any defined progression event. Patients were required by the protocol design to discontinue study treatment after disease progression but continue to be followed up in the trial. Patients who had not progressed at the time of the analysis were censored using the last available assessment date.

Overall survival was assessed from the date of randomization to the date of a patient’s death. Assessments for time to PSA progression and bone metastasis have been described previously [13]. In patients with soft tissue metastasis at baseline, objective tumour response and progression were assessed according to RECIST [16]. The appearance of new soft tissue lesions constituted a progression event. Adverse events (AEs) were monitored and coded according to the Medical Dictionary for Regulatory Activities terminology, and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 3 [17].

STATISTICAL ANALYSIS

The primary endpoint for this study was time to progression. To detect a 50% difference in median time to progression, equivalent to detecting a hazard ratio (HR) of 0.67 at the 20% significance level (two-sided) with 80% power, a total of 165 progression events were required.

The primary statistical analysis was for time to progression and was planned after 165 progression events. In response to the results for the secondary endpoint of overall survival obtained at the primary analysis (data cut-off April 2006), two additional analyses were planned for the times at which a minimum of 110 [13] (data cut-off February 2007) and 200 (data cut-off December 2008) deaths had occurred. Time to progression, overall survival and time to PSA progression were analyzed using the Cox regression model. Overall significance levels were adjusted using an O’Brien–Fleming approach to take account of multiple analyses that were performed during the present study [18].

The analysis reported for the present study was performed after the final cut-off in December 2008.

RESULTS

STUDY POPULATION

A total of 312 patients were randomized to the three treatment cohorts (zibotentan 10 mg, n= 107; zibotentan 15 mg, n= 98; placebo, n= 107). The present study includes analyses of data from the final cut-off (December 2008). All 312 patients were included in the intention-to-treat and safety populations, and all analyses were based on this population of patients. Demographics and baseline characteristics were similar in the three treatment groups, except for the number of bone metastases, which was greater in the zibotentan 15 mg group than the other groups. Detailed demographic and baseline characteristics have been described previously [13].

EFFICACY

Exposure and time to progression

The median duration of follow-up was 653 days (21.8 months), 682 days (22.7 months) and 520 days (17.3 months), for patients in the zibotentan 15 mg, zibotentan 10 mg and placebo groups, respectively. In comparison, the median time of total exposure to treatment was 113 days (3.8 months), 136 days (4.5 months) and 112 days (3.7 months) for zibotentan 15 mg, zibotentan 10 mg and placebo groups. Therefore, the median duration of treatment was 120 days, whereas the median duration of overall follow-up was 618 days. These differences occurred because the study was designed for patients to stop treatment when the primary endpoint of time to progression was reached. Consistent with the previous analyses [13], no statistically significant differences were observed in time to progression for either zibotentan 15 mg compared with placebo (HR, 0.86; 80% CI, 0.72–1.04; P= 0.309) or zibotentan 10 mg compared with placebo (HR, 1.06; 80% CI, 0.89–1.27; P= 0.673) at the final analysis (Fig. 2 and Table 1). A high proportion of progression events occurred early in the study, as shown by the significant number of patients who progressed in the first 3 months.

Figure 2.

Kaplan–Meier curve for time to disease progression at the final analysis.

Table 1.  Efficacy results at previous and final analyses
VariableFirst analysis (April 2006)Second analysis (February 2007)Final analysis (December 2008)
Placebo (n= 107)Zibotentan 10 mg (n= 107)Zibotentan 15 mg (n= 98)Placebo (n= 107)Zibotentan 10 mg (n= 107)Zibotentan 15 mg (n= 98)Placebo (n= 107)Zibotentan 10 mg (n= 107)Zibotentan 15 mg (n= 98)
  • *

    Symptoms requiring chemotherapy, radiotherapy, surgery (excluding transurethral resection with negative histology), or a new dose of hormone-based therapy;

  • †Progression of soft tissue metastases according to Response Evaluation Criteria In Solid Tumours.

Median exposure (months) 3.7 3.9 3.7 3.7  4.3 3.8  3.7  4.5 3.8
Number of progression events737561991008310310290
 Clinical progression*26232236 2930 37 2930
 Opiate analgesia18181423 2821 23 2922
 Objective progression25332332 4027 34 4029
 Death 4 1 2 8  3 5  9  4 9
Time to progression         
 Median (months) 3.6 4.0 3.8 3.7  4.6 3.8  3.7  4.6 3.8
 Hazard ratio vs placebo 0.88 0.83  1.09 0.94  1.06 0.86
 80% CI 0.71–1.09 0.66–1.03  0.91–1.31 0.78–1.14  0.89–1.27 0.72–1.04
 P vs placebo 0.448 0.278  0.553 0.702  0.673 0.309
Number of deaths18 91351 3334 75 7462
 Reason: prostate cancer13 8 941 2827 62 6453
 Reason: other or unknown 5 1 410  5 7  6 1010
Overall survival         
 Median (months)14.416.515.117.3 24.523.5 19.9 23.523.9
 Hazard ratio vs placebo 0.38 0.61  0.55 0.65  0.83 0.76
 80% CI 0.22–0.64 0.38–0.99  0.41–0.73 0.49–0.86  0.67–1.02 0.61–0.94
 P vs placebo 0.019 0.190  0.008 0.052  0.254 0.103

Overall survival

At the final analysis, a total of 211 (68%) deaths had occurred. A promising improvement in overall survival was seen in both zibotentan groups compared to placebo (Fig. 3 and Table 1). Compared to placebo, a reduction in the risk of death of 24% (HR, 0.76; 80% CI, 0.61–0.94; P= 0.103) and 17% (HR, 0.83; 80% CI, 0.67–1.02; P= 0.254), was observed in the zibotentan 15 mg and zibotentan 10 mg groups, respectively (Table 1). Figure 4 illustrates the findings of the three overall survival analyses, which are based upon 40 (13%), 118 (38%) and 211 (68%) death events at the initial, secondary and final analyses, respectively. The HR was closer to one at the final analysis than for the previous analyses and, as a consequence, was no longer statistically significant. Initially, Kaplan–Meier overall survival curves diverged, indicating a benefit with zibotentan treatment, but then became parallel when most patients had stopped study treatment (Fig. 3). Interpretation of the tails of the Kaplan–Meier overall survival curves is limited by the small number of patients remaining in the trial at the final analysis.

Figure 3.

Kaplan–Meier curve for overall survival at the final analysis.

Figure 4.

Comparison of hazard ratio vs placebo and 80% CI for overall survival at the first, second and final analyses. ITT, intention-to-treat.

PSA progression

There was no significant difference in serum PSA levels between the zibotentan treatment and placebo groups, as described previously [13], with serum PSA levels increasing over time in all treatment groups. PSA data collection stopped subsequent to clinical progression.

Tumour response rate

As shown in the previous analyses [13], there were no responders according to RECIST criteria in any treatment group, at the final analysis, and the proportion of non-responders (i.e. those with stable disease, progressive disease and non-evaluable disease) were comparable across all treatment groups.

Systemic anticancer therapy received after discontinuation

Data on systemic anticancer therapy usage following discontinuation were not available for a considerable proportion of patients (n= 37, 37.8%; n= 38, 35.5%; and n= 43, 40.2% of patients in the zibotentan 15 mg, zibotentan 10 mg and placebo groups, respectively) because consent was not obtained from patients and ethics committees to collect this information retrospectively. For those patients where data were available, systemic anticancer therapy (excluding radiotherapy) usage after discontinuation of the study treatment was similar across treatment groups (n= 48, 49%; n= 56, 52.3%; and n= 48, 44.9% of patients in the zibotentan 15 mg, zibotentan 10 mg and placebo groups, respectively). The most commonly used therapies were taxanes (96 patients), anthracyclines and related substances (27 patients), antiandrogens (40 patients), glucocorticoids (32 patients) and bisphosphonates (35 patients). Overall, the known taxane (i.e. docetaxel; one patient also took paclitaxel) usage was well balanced across treatment groups, with 33 (34%), 37 (35%) and 35 (33%) patients receiving treatment in the zibotentan 15 mg, zibotentan 10 mg and placebo groups, respectively.

Bone metastases

The number of bone metastases was slightly higher at baseline in the zibotentan 15 mg group than in the zibotentan 10 mg or placebo groups. Zibotentan 10 mg treatment led to a significant decrease in the number of bone metastases at study discontinuation compared to placebo (treatment ratio, 0.83; 80% CI, 0.70–0.98; P= 0.155). By contrast, zibotentan 15 mg had no effect on the number of bone metastases compared to placebo (treatment ratio, 1.02; 80% CI, 0.85–1.23; P= 0.874). Analysis of the median percentage change from baseline in the number of bone metastases at study discontinuation showed that, for a typical patient, bone metastases increased following discontinuation of study treatment in all groups (zibotentan 15 mg, 206%; zibotentan 10 mg, 164%; placebo, 202%) with overlapping 80% CIs showing little difference between groups.

SAFETY

Approximately 90% of patients experienced at least one AE (Table 2). The most commonly reported AEs considered to be related to zibotentan treatment were peripheral oedema, headache and nasal congestion.

Table 2.  Adverse events experienced by ≥10% patients in any treatment arm
VariableNumber (%) of patients*
Placebo (n= 107)Zibotentan 10 mg (n= 107)Zibotentan 15 mg (n= 98)Total (n= 312)
  • *

    Number (%) of patients who reported at least one adverse event for a preferred term.

Patients with any adverse event93 (86.9)99 (92.5)89 (90.8)281 (90.1)
Peripheral oedema11 (10.3)42 (39.3)47 (48.0)100 (32.1)
Headache13 (12.1)39 (36.4)43 (43.9)95 (30.4)
Nasal congestion4 (3.7)30 (28.0)33 (33.7)67 (21.5)
Back pain17 (15.9)18 (16.8)19 (19.4)54 (17.3)
Nausea17 (15.9)18 (16.8)14 (14.3)49 (15.7)
Rhinitis allergic2 (1.9)16 (15.0)15 (15.3)33 (10.6)
Dyspnoea6 (5.6)17 (15.9)8 (8.2)31 (9.9)
Fatigue10 (9.3)18 (16.8)6 (6.1)34 (10.9)
Arthralgia7 (6.5)10 (9.3)12 (12.2)29 (9.3)
Constipation18 (16.8)12 (11.2)9 (9.2)39 (12.5)
Anorexia7 (6.5)12 (11.2)7 (7.1)26 (8.3)
Insomnia10 (9.3)6 (5.6)10 (10.2)26 (8.3)
Pain in extremity12 (11.2)10 (9.3)5 (5.1)27 (8.7)

The overall frequency of grade 3 and 4 CTCAEs was low and similar across the treatment groups. Differences between placebo and zibotentan treatment groups were difficult to evaluate because of low patient numbers; anaemia, peripheral oedema, dyspnoea, headache, cardiac failure, nausea and pulmonary embolism of CTCAE grade 3 or higher occurred in at least three patients who received zibotentan but not in any patients receiving placebo. Few patients discontinued treatment as a result of AEs; in the zibotentan 15 mg, zibotentan 10 mg and placebo groups, respectively, 12 (12.2%), 12 (11.2%) and six (5.6%) patients discontinued from the study as a result of AEs, most of which were considered to be treatment related (Table 3).

Table 3.  Adverse events leading to discontinuation and serious adverse events that occur in two or more patients
VariableNumber (%) patients
Placebo (n= 107)Zibotentan 10 mg (n= 107)Zibotentan 15 mg (n= 98)
Patients with an adverse event leading to discontinuation6 (5.6)12 (11.2)12 (12.2)
 Oedema peripheral0 (0.0)2 (1.9)6 (6.1)
 Dyspnoea0 (0.0)1 (0.9)2 (2.0)
 Nasal congestion0 (0.0)0 (0.0)2 (2.0)
 Headache0 (0.0)2 (1.9)0 (0.0)
 Back pain2 (1.9)1 (0.9)0 (0.0)
 Dysuria1 (0.9)1 (0.9)0 (0.0)
 Fatigue1 (0.9)1 (0.9)0 (0.0)
 Myocardial infarction1 (0.9)1 (0.9)0 (0.0)
Patients with any serious adverse event14 (13.1)26 (24.3)18 (18.4)
 Anaemia1 (0.9)3 (2.8)2 (2.0)
 Cardiac failure0 (0.0)3 (2.8)0 (0.0)
 Cardiopulmonary failure0 (0.0)2 (1.9)0 (0.0)
 Dyspnoea0 (0.0)0 (0.0)2 (2.0)
 Fatigue0 (0.0)1 (0.9)1 (1.0)
 Left ventricular dysfunction0 (0.0)1 (0.9)1 (1.0)
 Non-cardiac chest pain0 (0.0)1 (0.9)1 (1.0)
 Pneumonia0 (0.0)1 (0.9)1 (1.0)
 Pulmonary embolism0 (0.0)1 (0.9)1 (1.0)
 Haematuria2 (1.9)1 (0.9)0 (0.0)
 Myocardial infarction2 (1.9)1 (0.9)0 (0.0)
 Vomiting1 (0.9)0 (0.0)1 (1.0)

Serious adverse events (SAEs) were infrequent in the present study, with similar numbers being reported by patients in all treatment groups (Table 3). There were nine deaths, six of which were in the placebo group, two were in the zibotentan 15 mg group and one was in the zibotentan 10 mg group. One patient in the zibotentan 15 mg group died as a result of pneumonia, which was considered to be treatment related.

Eleven patients experienced events relating to cardiac failure (ventricular dysfunction, cardiac failure, congestive cardiac failure and cardiopulmonary failure), all of whom were receiving zibotentan treatment. Ten of these patients had a history of cardiac abnormalities at baseline. Anaemia was reported as an AE by 21 patients, of which six were considered a SAE. These SAEs occurred only in patients with extensive bony involvement or evidence of rapidly progressive disease.

DISCUSSION

Consistent with previous analyses [13], the final analysis of this Phase II trial in patients with CRPC and bone metastases who were pain free or mildly symptomatic for pain showed no statistically significant difference between zibotentan and placebo for the primary endpoint of time to progression. However, as in the previous analyses, zibotentan continued to suggest a survival advantage compared to placebo.

The lack of a statistically significant difference between placebo and zibotentan treatment on time to progression, in the context of a positive effect on overall survival, may reflect the definition of disease progression used in the present study. A previous study of the immunological therapy sipuleucel-T for the treatment of patients with CRPC also reported an increased survival benefit with no difference in time to progression compared to placebo treatment [19]. In the present study, progression was a composite endpoint comprising clinical progression, requirement of opiate analgesia, objective progression and death in the absence of disease progression. These criteria have historically been used to assess the efficacy of chemotherapy, and whether such endpoints are useful in discriminating treatment effects of novel non-cytotoxic targeted therapies is now being questioned [20]. The Prostate Cancer Clinical Trials Working Group 2 recently recommended the development of outcome measures tailored to the specific mechanism of action of the drug being investigated. Cytotoxic drug trials should therefore include assessments of both control/relieve/eliminate or prevent/delay endpoints, and trials for noncytotoxic agents should be designed with prevent/delay endpoints [21]. However, components of composite endpoints such as those described above may be subjective in nature, and variations in interpretation by physicians and patients at multiple centres may occur. In the present study, a high proportion of progression events occurred within the first 3 months of study treatment. Preclinical models indicate that zibotentan has antiproliferative effects [12], which may have a less immediate impact on the disease than chemotherapy. Thus, the definition of progression in the present study, which is sensitive to relatively minor changes in prostate cancer symptoms, may have reduced the ability to differentiate treatment effects between study arms.

The overall survival advantage following zibotentan treatment appears to be robust: the survival benefit at the initial analysis following 40 (13%) deaths was sustained at the second analysis after 118 (38%) deaths, and is shown again at this final analysis after 211 (68%) deaths had occurred. Consistent with the results of the two earlier analyses, HRs <1 were observed at the final analysis for both doses of zibotentan treatment, indicating an improved overall survival relative to placebo. The early separation of the zibotentan and placebo Kaplan–Meier overall survival curves was followed by a period where the deaths occurred at similar rates (parallel curves), suggestive of a benefit of zibotentan that diminishes once treatment has been stopped. This contributes to the movement of the HR towards one another over time and reflects the long follow-up period during which patients were no longer receiving study drug, and were instead receiving similar post-study therapies in all three treatment groups. The treatment benefit of zibotentan does still appear to persist long after treatment discontinuation as shown by the continued separation of the Kaplan–Meier overall survival curves. Consequently, the median time to death is longer in the zibotentan 15 mg and 10 mg groups than in the placebo group. However, the long period of follow-up (21 months) relative to the short (4 months) median duration of study treatment may also explain why the overall survival results no longer achieve statistical significance for zibotentan 10 mg at the final analysis. The results obtained in the placebo group are consistent with those previously reported for patients in this disease setting [22,23], indicating that the overall survival benefit observed in the present study was unlikely to be the result of a poorly performing placebo group. Furthermore, treatment groups were well balanced in their use of systemic anticancer therapies after discontinuation of study treatment, which suggests that these therapies did not affect overall survival. These data indicate that there may be further benefit from extending the duration of treatment with zibotentan.

Zibotentan treatment had little effect on PSA levels compared to placebo. This is consistent with preclinical data, which have shown that ETA receptor antagonism with zibotentan does not alter PSA secretion [24]. Indeed, the data obtained in the present study are not unexpected because zibotentan is an agent specifically targeted at the ETA receptor, without cytotoxic activity. Zibotentan therefore has no direct interaction with the androgen receptor (the principal regulator of PSA gene synthesis and secretion) or overt cell killing capacity (which would indirectly reduce PSA levels). PSA sampling in the present study was discontinued after patients stopped study treatment because of disease progression. Therefore, the follow-up period may have been insufficient to observe any long-term effects that the study treatment may have had on PSA levels.

The overall number of bone metastases was highly variable and increased over time for all treatment groups. Treatment with zibotentan 10 mg led to a significant reduction in the number of new bone metastases compared to placebo; however, treatment with zibotentan 15 mg did not significantly reduce the appearance of new bone metastases. Bone metastatic involvement increased in all treatment groups after the discontinuation of study treatment, but least so in patients receiving zibotentan 10 mg. However, intergroup overlapping of the 80% CIs suggests differences between treatment groups are unlikely.

Zibotentan was well tolerated, with the most commonly reported AEs being either transient or manageable, reflecting the vasodilatory pharmacology of ETA receptor antagonism. SAEs were infrequent, with anaemia and cardiac failure being the most commonly reported following treatment with zibotentan. Of 11 patients who reported cardiac failure as a SAE, 10 had a history of cardiac abnormalities, which may have increased their predisposition for this condition. The effects of ETA receptor antagonism on patients with cardiac failure was investigated recently but showed neither beneficial, nor detrimental effects on cardiac failure [25]. However, the small number of patients in the present study preclude ruling out any causal relationship entirely at this stage and the condition will be monitored carefully in the ongoing Phase III trial programme. Of those patients who reported anaemia as a SAE, all had extensive bony involvement or evidence of rapidly progressive disease and most patients who experienced anaemia had reduced haemoglobin levels at baseline. On the basis of the more benign tolerability profile of zibotentan at 10 mg and the lack of difference in the efficacy between the zibotentan 10 mg and 15 mg groups, the 10 mg dose has been selected for use in the Phase III trial programme.

In conclusion, the primary endpoint of time to progression was not met in the present study, although a promising signal in improvement of overall survival in patients with CRPC and bone metastases who were pain free or mildly symptomatic has persisted across three analyses. Further benefit may be achieved by continuing zibotentan treatment beyond progression because the endothelin axis can stimulate tumour growth through all stages of prostate cancer. The results obtained support the strategy of targeting the ETA receptor in patients with prostate cancer. To confirm the overall survival signal observed in the present study, zibotentan is being investigated further in the ENTHUSE (ENdoTHelin A USE) Phase III clinical trial programme.

ACKNOWLEDGEMENTS

We thank Dr Claire Routley, from Mudskipper, who provided editorial assistance funded by AstraZeneca. This research was sponsored by AstraZeneca.

CONFLICT OF INTEREST

Nicholas D. James received research funding and consultancy fees from AstraZeneca; Heather Payne received fees for lectures, advisory boards and consultancy from AstraZeneca; Bernard A. Zonnenberg received research funding from AstraZeneca. Stuart McIntosh, De Phung and Thomas Morris are employees of AstraZeneca. Nancy A. Dawson received research funding from AstraZeneca.

Ancillary