TRAPEZE: a randomised controlled trial of the clinical effectiveness and cost-effectiveness of chemotherapy with zoledronic acid, strontium-89, or both, in men with bony metastatic castration-refractory prostate cancer

Prostate cancer

Prostate cancer is a major worldwide health problem which accounts for nearly one-fifth of all newly diagnosed male cancers. In the UK, approximately 35,000 men are diagnosed with prostate cancer each year, and in 2008 almost 10,000 men died from the disease. 1 The disease is mostly one of older age, but significant numbers of men of working age will develop the disease. Figure 1 summarises the age distribution of incident cases and deaths.

FIGURE 1.

Prostate cancer incidence and mortality in the UK. 2–4

Although adenocarcinoma of the prostate most often presents as local (stage T1 or T2) disease, in which the malignancy is confined to the prostate, a significant proportion of patients progress despite initial treatment with ablative surgery or radiotherapy, often in combination with hormonal therapy. A minority of patients present with de novo metastatic disease. Figure 2 summarises the treatment options across the disease spectrum.

FIGURE 2.

Prostate cancer treatment paradigm. Cancer has spread to the pelvic lymph nodes (N+) or to lymph nodes, organs, or bones distant from the prostate (M+). CRPC, castration-refractory prostate cancer; Ra223, Radium-223.

Hormone therapy

A hormone (from the Greek ὁρμή, meaning ‘impetus’) is a chemical released by a cell in one part of the body to affect cells in other parts of the organism. Cells respond to a hormone when they express a specific receptor for that hormone. The hormone binds to the receptor protein, resulting in the activation of a signal transduction mechanism that ultimately leads to cell type-specific responses. Hormone therapies can thus work on a number of points in this pathway and there are examples of all of these in prostate cancer, which are summarised in Table 1.

Hormone therapy has been a mainstay of prostate cancer since the seminal studies of Huggins and Hodges,5 published in 1941, demonstrating substantial and prolonged remissions from prostate cancer with the use of either surgical castration or oestrogen therapy. Diethylstilboestrol is the first example of a successful drug treatment for advanced cancer, and, while now supplanted in this role, it remains in use 70 years later. As is now well known, while responses to hormone therapy may be dramatic, with durations running into many years, they are rarely curative and typically last 18–24 months depending on disease stage. This period after failure of initial androgen deprivation therapy has been known by many terms over the years, including androgen-independent prostate cancer and castration-refractory prostate cancer (CRPC). However, with the recognition that relapsing tumours remain dependent on androgen receptor-mediated pathways and the licensing in relapsing disease of abiraterone,6–8 a steroid synthesis inhibitor, and enzalutamide,9 an androgen receptor-targeting agent, the term castration-refractory prostate cancer is increasingly used. This term is, however, unpopular with patient groups and, while accurate, may yet also be supplanted if anyone can think of a term with less pejorative overtones.

Broadly speaking, there are two routes into long-term hormone therapy: via localised disease, radical therapy and relapse, and de novo advanced disease (Figure 3).

FIGURE 3.

Pathways to advanced disease. Natural history for metastatic patients. ADT, androgen deprivation therapy; PSA, prostate-specific antigen.

Figure 3 shows disease burden expressed via the prostate-specific antigen (PSA) level on the vertical axis. For most purposes, however, PSA does equate with disease burden. In particular, in late-stage disease managed with non-hormonal therapies the relationship is not that close and PSA is not recognised as a surrogate end point for clinical trials. In early hormone-sensitive disease, the concordance between PSA changes and clinical ones is close. One consequence of the use of the PSA test is that managment tends to be PSA-driven rather than clinically-driven. In the case of patients relapsing after failed local therapy, clinicians are faced with a rising PSA but often no radiological evidence of disease for many years – termed a biochemical relapse. Patients in this situation will often be started on hormone therapy many years before any clinical consequences of relapse. Randomised trials in this setting have shown that intermittent therapy is as good as continuous therapy and probably should be regarded as the standard of care.

Management of metastatic disease

Initial management of men with locally advanced or metastatic prostate cancer is some form of androgen deprivation therapy. This will generally control disease for 1–3 years, following which progressive clinical failure will ensue – CRPC. In patients with metastatic CRPC (mCRPC), one of the most common sites of spread is bone. The development of bone metastasis, and the associated pain, results in a high level of mobility problems, leading to a loss of functional independence in men, and is a major cause of mortality [bone marrow failure, pathological fractures, spinal cord compression (SCC) and other bone-related complications]. Bone morbidity is often quantified in clinical trials via a composite end point termed the skeletal-related events (SREs). The elements that make up this end point are summarised as:

• pathological fracture

• SCC

• radiotherapy to bone

• hypercalcaemia

• change in anticancer treatment to treat bone pain.

The reduction in the frequency or severity of SREs that any particular patient experiences during the individual disease pathway may provide additional health-related quality-of-life (HRQoL) benefits. The true benefit in terms of HRQoL is not yet completely known, although recent data from clinical trials have begun to show the HRQoL benefits of bisphosphonates. 10,11 In addition to the potential quality-of-life (QoL) benefits, patients may also gain actual survival benefit from either mono or combination therapy. Although bisphosphonates therapy and/or chemotherapy may be considered as central to the treatment of patients with bone metastases, other therapies such as radioisotopes are available and are widely used for patients with mCRPC.

Chemotherapy

For many years chemotherapy was considered too toxic to be of value in men with advanced prostate cancer. There were a number of reasons for this, including later diagnosis in the pre-PSA era, difficulty in assessing responses and problems in managing toxicity, such as nausea and vomiting. The advent of PSA-driven diagnosis and management, while remaining controversial in terms of use as a screening test, has undoubtedly resulted in a strong trend to earlier diagnosis now dating back several decades. This in turn has meant that men are diagnosed younger with advanced disease. Secondly, the use of PSA monitoring post-primary treatment has meant that men relapsing after failed radical therapy are picked up early and so, when mCRPC does develop, treatment can be instigated when men remain fit enough to cope with it. Definitive proof of benefit from palliative chemotherapy came from a landmark National Cancer Institute of Canada trial led by Ian Tannock from Toronto. The trial compared prednisone alone with prednisolone plus mitoxantrone given 3-weekly for up to 10 cycles. This relatively small study of 161 patients published in 1996 set out to compare palliative end points rather than survival-based ones. 12

A palliative response was observed in 23 out of 80 patients who received mitoxantrone plus prednisone, compared with 10 out of 81 patients who received prednisone alone. In an additional seven patients in each group, analgesic medication was reduced without an increase in pain. The duration of palliation was longer in patients who received chemotherapy (with a median of 43 weeks to symptom worsening) than in those treated with prednisone alone (median of 18 weeks to symptom worsening). There was significant crossover from the prednisone arm to the chemotherapy arm and no difference in overall survival. Thus, this study clearly established the principle that chemotherapy could provide palliative benefit but did not show a survival benefit. Subsequent mitoxantrone trials produced similar results, although the crossover between the chemotherapy and no-chemotherapy arms means that, essentially, it is not known whether or not chemotherapy with this agent produces a survival benefit.

In the late 1990s, a variety of agents started to be evaluated in what was then called hormone-refractory prostate cancer (HRPC). Docetaxel emerged as the lead candidate for evaluation in large phase trials, and two landmark studies were published in the New England Journal of Medicine in 2004. 13,14 One trial, the TAX327 study,13 compared weekly or 3-weekly docetaxel with the Tannock mitoxantrone regimen. The second trial (SWOG 991614) compared a combination of docetaxel and estramustine with the same control arm. Both trials showed improved palliative outcomes compared with mitoxantrone and, very importantly, an overall survival (OS) advantage for 3-weekly docetaxel and the docetaxel–estramustine combination with hazard ratios (HRs) of 0.76 and 0.8, respectively, despite significant crossover to docetaxel in the mitoxantrone arms of both studies. All patients in both trials received prednisone as per the original Tannock paper. These trials confirmed unequivocally that chemotherapy could both prolong survival and give worthwhile palliation without undue toxicity. They also established that docetaxel is a superior agent to mitoxantrone. On the basis of these trials, a 3-weekly schedule of docetaxel plus prednisolone for up to 10 cycles has emerged as the standard of care for mCRPC and was approved by the National Institute for Health and Care Excellence (NICE) for this purpose in 2006 (Figure 4).

FIGURE 4.

Comparison of estimated probability of overall survival of mitoxantrone and docetaxel for CRPC. 15

A number of agents have been studied in the second-line chemotherapy setting. Of these, to date only cabazitaxel has shown a survival advantage and obtained a licence. The key trial, TROPIC, compared cabazitaxel with mitoxantrone given on the standard Tannock trial schedule and showed an improvement in median survival from 12.7 to 15.1 months. 16 Cabazitaxel was licensed in 2010 ahead of abiraterone, which obtained a licence in 2011 in the same post-docetaxel setting. As both drugs improve survival, although they have completely different modes of action, there is clearly an unresolved issue over choice and sequencing (or indeed combination) of agents. The position has now been further complicated by the licensing of enzalutamide post chemotherapy based on the AFFIRM trial,9 plus the extension of the abiraterone licence to chemo-naive patients. 6 The licence for enzalutamide was also expanded to cover pre-chemotherapy patients following the PREVAIL trial. 17

Additionally, the recent publication of the results of the ALSYMPCA trial of radium-22318 demonstrated both improved OS and reduced skeletal complications with six injections, one every 28 days, of radioisotope compared with placebo. 18 How chemotherapy should best be integrated with other therapeutic options for patients with bone metastasis is at present not defined.

Bisphosphonates

Bisphosphonates inhibit bone catabolism by reducing the numbers of functioning osteoclasts and have been used to manage bone metastases. Zoledronic acid (ZA), but not some older bisphosphonates, also arrest cell proliferation, induce apoptosis, and inhibit the growth factor stimulation of cultured prostate cancer cells. 14 In trials in relapsing mCRPC, ZA reduced the time to SRE as well as the frequency of subsequent SREs. 19,20 The ZA licensing trials19,20 have proved very controversial, as the fracture end point was assessed by regular skeletal survey with blinded radiological assessment. Hence, there is significant doubt as to whether many of the small fractures detected were precursors of a subsequent real ‘clinical’ SRE or radiological features of no significance. ZA is not currently recommended for use in the UK by NICE because of doubts as to its cost-effectiveness.

Radioisotopes have been used to palliate bone pain for over 20 years. A variety of radioisotopes are available; the most commonly used during the trial recruitment era were strontium-89 (Sr-89)21,22 and samarium-153. 23 Both accumulate selectively in bone metastases compared with non-involved bone. There is some evidence that Sr-89 may reduce overall health-care costs compared with standard methods of delivering radiotherapy. 24 There are a number of previous studies of combined use of chemotherapy with radioisotopes. Of particular note, Tu et al. 25 combined combination chemotherapy with Sr-89 in a small randomised trial with promising results suggesting a survival advantage in chemotherapy responders allocated to Sr-89.

Since the publication of the MRC PR05 study,26,27 more potent bisphosphonates have been evaluated in mCRPC. The most widely studied has been zoledronate, which has a 40- to 850-fold higher potency than clodronate in pre-clinical models of bone resorption. 28 It has also been shown to be more effective than pamidronate (90 mg) in controlling malignant hypercalcaemia29,30 In addition, zoledronate has demonstrated direct anticancer activity, including inhibition of proliferation of breast cancer and prostate cancer cells in vitro. 31,32

In prostate cancer trials in relapsing mCRPC, ZA reduced the time to SREs as well as the frequency of subsequent SREs. 19,20 However, it is clear from looking at the components that make up the SREs that these vary hugely in clinical significance and, in addition, are to a degree subjective. In particular, the ZA licensing trials19,20 have proved very controversial as the fracture end point was assessed by regular skeletal survey with blinded radiological assessment. As such there is significant doubt about whether many of the small fractures detected were precursors of a subsequent real ‘clinical’ SRE or radiological features of no significance. The subsequent trials comparing ZA with denusomab33 used the same methodology and so can be subject to the same criticism. As a result, neither agent is recommended for use in the UK by NICE. The impact of ZA on SREs is illustrated in Figure 5; the bisphosphonate showing decreases in skeletal complications in both lytic and blastic lesions in a comparison with pamidonate. 20

FIGURE 5.

Proportion of patients with SREs, demonstrating reduction with ZA.

In vitro evidence suggests synergistic killing of breast and prostate cancer cells when combined with chemotherapy. 32 Furthermore, ZA was licensed in the ‘pre-docetaxel’ era; hence, whatever the merits of SRE prevention, the role of zoledronate in the chemotherapy era was effectively undefined. It was therefore logical to evaluate docetaxel with ZA in men with mCRPC affecting bone. In view of the controversy over the SRE as an end point, we did not undertake routine skeletal evaluations as in the zoledronate and denosumab licensing trials but collected data only on ‘clinical’ SREs; that is, those reported by the patient or diagnosed on the basis of symptoms such as those from SCC. We combined this clinically orientated approach to the SRE with a health economic assessment of the impact of the various trial interventions, the intention being that, if the clinical utility of combination therapy were confirmed, we should be able to produce robust estimates of the cost-effectiveness at the same time.

Radioisotopes

A variety of radioisotopes are available, the most commonly used during the trial recruitment era being Sr-89 and samarium-153. Both accumulate selectively in bone metastases compared with uptake rates in non-involved bone. Sr-89, a bone-seeking radionuclide, is a pure β-emitter with a half-life of 50 days, has a high uptake in osteoblastic metastases, and remains in tumour sites for up to 100 days. Sr-89 provides pain relief in up to 80% of patients, and complete freedom from pain in approximately 10%, for periods that can exceed 3 months. 21,34 In a randomised controlled Phase III trial, the combination of Sr-89 injection and external beam radiotherapy improved pain relief, delayed disease progression and enhanced some QoL measures compared with external beam radiotherapy alone. 21 However, another Phase III randomised controlled trial has suggested that, in some patients, systemic Sr-89 may be inferior to local-field radiotherapy in terms of survival (7.2 months vs. 11.0 months; p = 0.0457). 22 The selection of patients has a significant impact on outcome, response and duration of response to radionuclide therapy, as bone pain palliation is reduced in those who have widespread metastatic disease or a short life expectancy. 35–38 Consequently, the use of radionuclides appears to be optimal at an early stage in disease management. However, their efficacy is reduced or lost with repeated use, and overtreatment can also lead to irreversible pancytopenia. As noted above (see Bisphosphonates), there is some evidence that Sr-89 may reduce overall health-care costs compared with standard methods of delivering radiotherapy. 39

There are a number of previous studies of combined use of chemotherapy with radioisotopes. Tu et al. combined combination chemotherapy with Sr-89 in a small randomised trial with promising results suggesting a survival advantage in chemotherapy responders allocated to Sr-89. 25 More recently, Fizazi et al. ,40 Tu et al. 41 and Morris et al. 42 have combined docetaxel with samarium-153 in Phase I/II trials, confirming safety for the combination. No published randomised trials have addressed the safety or efficacy of docetaxel with either Sr-89 or samarium-53.

As new treatments have appeared for CRPC, these treatments have been less frequently used. However, recent data with a new radioisotope radium-223 seem set to change this picture. Like Sr-89, radium-223 is a calcium mimetic. Recently completed placebo-controlled Phase III trials in symptomatic CRPC patients showed a prolongation of survival and also a delay and reduction in symptomatic (as opposed to radiological) SREs. 18 Levels of adverse reactions reported in the trial were low. The agent was licensed in 2013 and is an important new therapeutic option for men with CRPC, especially as the trial included men both pre and post chemotherapy, as well as those deemed unfit to ever receive chemotherapy.

Osteoporosis

Patients eligible for the study are at risk of osteoporosis in view of their previous therapy (androgen deprivation, possible steroid exposure, age) as well as from some on-study therapies (steroids, docetaxel). Osteoporosis was therefore considered in the causality of any SRE. A bone density substudy formed part of this trial.

Arm A: control – docetaxel plus prednisolone

Docetaxel 75 mg/m² (up to a maximum dose of 165 mg) was administered intravenously at 3-weekly intervals (21 days). Participants also received oral prednisolone 10 mg daily throughout trial treatment or until disease progression or associated treatment toxicity.

Trial chemotherapy ceased after cycle 6 for Phase II participants but continued for up to 10 cycles for Phase III participants, ceasing for pain or tumour disease progression, or other cause decided by the treating clinician or patient choice. As noted above (see Trial design) patients could receive further chemotherapy off-trial in keeping with NICE guidance.

Arm B: docetaxel, prednisolone plus zoledronic acid

Docetaxel and prednisolone were administered as per the control arm. ZA was administered intravenously after completion of docetaxel administration at a dose of 4 mg, subject to pre-treatment creatinine clearance being greater than 60 ml/minute; creatinine clearance of < 60 ml/minute would incrementally reduce the dose given, as detailed in section 6.1.3 of the protocol (see Appendix 1). Following the completion of chemotherapy, participants received continuing ZA at 4-weekly intervals, as clinically indicated, until pain or tumour disease progression or withdrawal. It was recommended that patients treated with ZA also receive vitamin D and calcium supplements throughout treatment.

Arm C: docetaxel, prednisolone plus strontium-89

Docetaxel and prednisolone were administered as per the control arm, for six cycles. Subject to satisfactory haematological and clinical parameters on clinical assessment 21 days after the sixth docetaxel treatment, participants received a single 150-MBq dose of Sr-89 on the 28th day after the sixth cycle.

Chemotherapy ceased after cycle 6 for Phase II participants, but for Phase III participants continued for up to 10 cycles after a period of between 28 and 56 days of Sr-89 administration, allowing for bone marrow function to be adequately recovered.

Arm D: docetaxel, prednisolone, zoledronic acid plus strontium-89

Patients in this arm received docetaxel, prednisolone and ZA for six cycles, as per arm B participants, plus clinical and haematological assessment and Sr-89 administration, as per arm C participants. Following a recovery period of between 28 and 56 days, chemotherapy, prednisolone and ZA treatment resumed until disease progression, associated treatment toxicity or patient withdrawal. As per the arm B treatment regime, following the end of chemotherapy, patients received continuing ZA administrations at 4-weekly intervals, as clinically indicated, until disease progression or until other discontinuation criteria were met. It was again recommended that patients treated with ZA also receive vitamin D and calcium supplements throughout treatment.

Further off-study treatment

All further off-study treatment, for example chemotherapy, bisphosphonate and radioisotope therapy, as well as newer drugs, such as abiraterone, enzalutamide and radium-223, received after study treatment were captured on the Concomitant Medication Running Form. The choice of further treatment was at the discretion of the participant’s clinician.

Case report forms

Data collected on each subject were recorded by the investigator or his/her designee on case report forms (CRFs). Originals of the CRF were returned to the trial management office, whereas photocopies were retained by the site.

Quality-of-life data

All eligible participants were asked to consider taking part in the QoL part of the study. QoL was assessed using patient-completed questionnaires, i.e. the European Quality of Life 5-Dimensions (EQ-5D) and Functional Assessment of Cancer Therapy – Prostate (FACT-P), while pain and analgesic use diaries were used to facilitate changes in participants’ pain perception and management. An example of both the QoL booklet and pain diary are part of the protocol in Appendix 1. A QoL booklet and pain diary were completed at baseline and subsequently prior to each treatment and follow-up visit. Completion of these documents remained voluntary and continued throughout patient follow-up (pre and post clinical progression), irrespective of any further therapy a patient may have received.

Monitoring

The study was conducted under the auspices of the Cancer Research Clinical Trials Unit (CRCTU) according to current guidelines for good clinical practice. Participating centres were monitored by CRCTU staff to confirm compliance with the protocol and the protection of patients’ rights as detailed in the Declaration of Helsinki.

Participating centres were monitored by checking incoming forms for compliance against the protocol, consistent data, missing data and timing. CRCTU onsite monitoring was carried out as detailed by the trial’s risk assessment, primarily of all sites that had enrolled four or more patients into the trial. Patients’ records to be audited at such visits were selected randomly from the different treatment arms.

Stratified randomisation

Stratification was used to ensure the balance of participant characteristics as well as numbers within each treatment group. Patients were randomised to treatment arms in a 1 : 1 : 1 : 1 allocation ratio using a computerised minimisation algorithm. If the minimisation is balanced, then allocation is random with equal chance of allocation to all arms. Randomisation was stratified by centre and ECOG performance score (0, 1 or 2) to avoid imbalance.

Implementation

Prior to randomisation, patients gave their informed consent to take part in the trial and the clinician or research nurse completed a pre-randomisation checklist to ascertain that the patient met all the entry criteria.

The process of entering a patient into the trial was conducted by telephone with the CRCTU randomisation office. Using either a computerised randomisation program or a paper equivalent should the computer system be out of commission, the CRCTU randomisation officer re-ascertained the patient’s eligibility, after which the computer program allocated the next available trial number and randomised treatment arm for the participant. When the randomisation was conducted while the computer was out of commission, systems were in place to allocate the next available trial number and random treatment.

The allocated trial number and treatment arm were communicated to the site by telephone and confirmed by fax.

Trial Management Group

The Trial Management Group comprised the chief investigator, a few co-investigators and members of the CRCTU, as detailed in the front sleeve of the protocol (see Appendix 1). The Trial Management Group was responsible for the day-to-day running and management of the trial and met by teleconference or in person, as required.

Data Monitoring Committee

An independent Data Monitoring Committee (DMC; see Appendix 2), comprising an independent statistician, oncologist and urologist, met approximately annually to review the accumulating confidential trial data. Their main objective was to advise the Trial Steering Committee (TSC) whether or not there was any evidence or reason to amend or terminate the trial based on the recruitment rate or safety. Reports to the DMC were produced by the CRCTU.

Trial Steering Committee

An independent TSC (see Appendix 2) provided overall supervision for the trial and advised the trial management group. Members included an independent statistician, oncologist and two urologists. The ultimate decision regarding continuation of the trial lay with the TSC, based on the advice received from the DMC. The TSC met approximately annually, shortly after the DMC met.

Primary end points

Secondary end points

Ancillary end points

Bone mineral density changes and biomarker substudies are detailed in the protocol (see Appendix 1); tertiary end points will not be presented at this time in this report.

Summary of changes to the trial protocol

Phase II treatment consisted of six cycles of docetaxel chemotherapy plus an additional four cycles off-study at the discretion of the treating physician. NICE, however, recommended that up to 10 cycles of docetaxel chemotherapy should be administered in one treatment block. This was not stated clearly in the Phase II protocol and the previous trial design had the inadvertent effect of preventing some patients from receiving cycles 7 to 10 at a later stage because of local policy. Adopting the NICE recommendation formally into the clinical trial design ensured that all patients had access to the NICE-recommended schedule of chemotherapy and that the control treatment arm was considered the true ‘standard of care’ (Tables 3 and 4).

Ineligible

In total, 27 patients were found to be ineligible following randomisation. Five were randomised to docetaxel alone, 10 to docetaxel + ZA, seven to docetaxel + Sr-89 and five to docetaxel, ZA and Sr-89. All ineligible patients are included in intention-to-treat analysis.

There were three main categories of ineligibilities. These were (1) pre-randomisation blood pressure and blood tests were missed or performed outside of the allowed time frame, (2) progression on trial entry was not appropriately documented and (3) hormone therapies were not stopped at the appropriate time point, for example if bicalutamide had been stopped within 4 weeks of starting trial treatment rather than within 4 weeks of randomisation as stipulated in the eligibility criteria.

Protocol deviations

In total, 71 protocol deviations were reported: 15 in the docetaxel arm, 17 in the docetaxel and ZA arm, 18 in the docetaxel and Sr-89 arm and 21in the docetaxel, ZA and Sr-89 arm.

Table 5 provides a complete summary of all protocol deviations reported during the course of the trial.

Patient withdrawal of consent

Full consent for any further participation in the trial, including follow-up, has been withdrawn by 21 patients. In addition, 28 patients have withdrawn from one or more of the trial substudies. Table 6 contains a breakdown of all non-treatment withdrawals by randomisation arm and Table 7 by comparison group; a complete list of all patients who have withdrawn full consent can be found in Appendix 5.

Withdrawal of trial treatment

Lost to follow-up

Six patients in total have been reported as being lost to follow-up by site: three randomised to docetaxel alone, one randomised to docetaxel and Sr-89 and two randomised to docetaxel, ZA and Sr-89. Two of these reached the primary end point prior to being lost, one subsequently died and, although some follow-up information remains missing, the death information was obtained.