Background
Management of metastatic castration-resistant prostate cancer (mCRPC) has dramatically changed over the past 5 years [
1]. Until 2011, the standard of care in first line was the addition of docetaxel, a tubulin poison chemotherapy (CT), to LHRH analogue considering that hormonal treatments alone are no longer efficient in this setting [
2]. As second line treatment, the only published phase III trial compared the use of cabazitaxel, another tubulin poison, to mitoxantrone after progression and led to the approval of cabazitaxel [
3].
New paradigms have emerged in the last decade with initial studies showing that Abiraterone Acetate (AA) may reverse hormonal resistance by specifically inhibiting 17 α-hydroxylase/C17,20 lyase (CYP17A1) involved in the androgen synthesis pathway [
4]. Indeed, CYP17A1 is expressed in testicular, adrenal and prostatic tumor tissues, which explain why mCRPC tumor growth still relies on androgen. AA can overcome both “standard” and “backdoor” pathway of androgen synthesis and may result in a drastic decrease of testosterone circulating levels [
5]. In that context, the phase III trial COU-AA-301 demonstrated a significant overall survival benefit of AA/prednisone compared to placebo/prednisone (14.8 vs 10.9 months) [
6].
In the meantime, France was one of the first countries to make AA available to mCRPC patients after docetaxel CT through a Temporary Authorization for Use (TAU). This type of program allowed patients to have access to the drug from the time of European Medicines Agency approval and the reimbursement approval by the French National Health Services.
In December 2010, we undertook an observational study evaluating safety and long-term efficacy of AA in the daily clinical practice.
Methods
Data collection
The French Agency for National Medical Security allowed patients with mCRPC that progressed during or after docetaxel, to access AA before its commercial availability from December 2010 to September 2011.
This ambispective observational cohort study was conducted in 20 centres that accepted to record AA safety and efficacy data for all their patients enrolled in the TAU.
Data collection was done on site from medical records of all the TAU patients. Data were updated in April 2014.
Ethics statements
This study was approved by the French data protection authorities (CNIL) and the Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la Santé (CCTIRS # 11.545 approved on September 29th 2011). Written informed consent was waived because this is a retrospective study. The study was undertaken in accordance with the ethical standards of the World Medical Association Declaration of Helsinki.
Patients and treatment
Inclusion criteria were as follows: men with mCRPC and documented disease progression during or after a docetaxel-containing regimen. Progression was defined by clinical progression, PSA progression and/or radiographic progression on bone scan or CT scan, as defined by the Prostate cancer Working Group 2 (PCWG2) criteria [
7]. Patients should be under androgen deprivation and had castration level of testosterone (<50 ng/ml). Before AA delivering, patients should had potassium level >3.5 mmol/l, ASAT/ALAT <5 UNL in case of liver metastasis, or <2.5 UNL in the absence of metastases, and total bilirubin <1.5 UNL. Regarding the toxicity profile, exclusion criteria included uncontrolled hypertension, severe or unstable angina, and myocardial infarction within 6 months, heart failure, arterial or venous thromboembolic events, or clinically significant ventricular arrhythmias.
The recommended dose of AA was 1 g per day, as 4 tablets of 250 mg in one administration one hour before or two hours after a meal, in combination with oral prednisone 5 mg twice a day. Patients were treated until clinical, biological or radiological progression according to PCWG2 criteria, death, unacceptable toxicity, or physician’s or patient’s decision to stop the treatment.
Outcomes measures
In the context of this TAU, clinical and biological follow-up were scheduled every 15 days within the first three months of treatment and monthly afterwards until treatment discontinuation. Radiological evaluation during follow-up was not mandatory. All selected variables were collected in the medical report i.e. patients’ characteristics, disease description at diagnosis (Gleason score, tumour classification, metastasis sites before chemotherapy and before AA, PSA kinetics, number of prior docetaxel cycles, duration of treatments and reasons for treatment discontinuation), and follow-up.
For the efficacy analysis, survival time were calculated in two different manners: from the beginning of CT, defined as the time interval between the start of first line chemotherapy and the date of death; and from the initiation of AA and the date of death. Patients alive were censored at the last known follow-up date. AA treatment duration was classified in three categories (≤3 month, 3–6 months, and > 6 months), according to the biological and radiological assessment planned in the TAU program, and in two categories (≤3 month, >3 months) for multivariate analysis. PSA was measured at the time of inclusion, at 3 and 6 months as suggested in the TAU. Adverse events were followed on a monthly basis and graded according to the NCI-CTCAE v3.0.
Statistical methods
Qualitative variables were described by frequency of modalities and percentage. Continuous variables were described by mean, median, and range. Data are presented with 95% confidence intervals (95% CI), calculated with the use of exact methods based on the binomial distribution for discrete variables.
Median follow-up was calculated with the use of the Kaplan-Meier reverse method.
Predictive factors of AA treatment duration (in two categories: ≤3 month, >3 months) were identified with the use of univariate and multivariate logistic regression using a backward selection method, including the following variables: age, Gleason score, duration of CT before AA treatment, baseline PSA, PSA before AA treatment, site of metastasis before AA treatment initiation, duration of hormone therapies, and number of CT lines.
Overall survival rates were estimated using the Kaplan–Meier method.
A Cox proportional-hazards model was used to estimate the hazard ratios indicating the effects of prognostic factors on the risk of death. Three and 6 months landmark analyses were performed to explore the association between duration of AA treatment and overall survival.
All tests were two-sided, with a P value of less than 0.05 considered as statistically significant. Analyses were performed using the Stata software, v13.
Discussion
This ambispective observational cohort study enrolled all the eligible mCRPC patients of the 20 centres which agreed to participate. This was rapidly followed by the prescription of AA by other centres leading to a national TAU cohort of a total of 1629 patients over nine months. When the TAU was initiated, no other treatment was available besides docetaxel or experimental treatments accessible in clinical trials. Therefore, a high number of patients were allowed to receive AA treatment. Consequently, the population of this study is a “real-life”, non-selective population that includes a large number of patients with advanced disease (48.7% with multiple sites of metastases) who received up to 5 lines of chemotherapy.
In terms of safety, the pivotal COU-AA-301 study demonstrated that AA was associated with elevated mineral corticoids levels, aminotransferase level affecting liver function, urinary tract infections, fluid retentions, and oedema [
6]. In our study, a high proportion of included patients presented an advanced disease, but no new adverse event was recorded, confirming the safety of AA usage.
Median treatment duration was three months shorter than the one observed in the COU-AA-301 trial (5 versus 8 months). Though patients were more heavily pre-treated and the duration of treatment by AA was much shorter, we did not observe a significant change in OS (14.6 months in the present study versus 14.8 months in the COU-AA-301). In concordance with the OS that is observed from the introduction of first line CT, it may reflect the evolution of care in the management of mCRPC patients.
We found that the duration of AA treatment was significantly associated with prolonged survival. Two third of the patients received more than 3 months of AA, whereas the other third received less than 3 months of AA, indicating that these patients rapidly developed a resistance to the drug. This resistance is mainly due to an alteration of the androgen receptor (AR) axis by several mechanisms including changes in AR expression levels, occurrence of AR mutations, interactions of AR with co-activators or co-repressors, or increase in the expression of the CYP17A1 target itself [
8]. In these patients, a fatal issue is rapidly observed despite the use of cabazitaxel or of the AR antagonist enzalutamide. Indeed, several retrospective studies showed that enzalutamide had modest clinical activity in patients with mCRPC who previously received docetaxel and AA [
9,
10]. For patients where resistance is due to an overexpression of CYP17A1, it is however possible to envisage an increase in AA dosage in order to prolong survival [
11].
Prior to our study, the only relevant predictive factor of response to AA was the baseline level of testosterone as determined in the post hoc exploratory analysis of COU-AA-301 data, the OS being significantly longer in patients with high androgen levels [
12,
13]. Interestingly, we found that the main predictive factor of AA benefit was the difference in PSA values between baseline and 3 months of treatment. PSA flare up described previously concerns a minority of patients (less than 10%) [
14], so determination of PSA levels could help the early monitoring of AA benefit and avoid maintaining an ineffective costly treatment. When localized to the bone only, presence of metastases was a good prognostic factor significantly associated with prolonged OS. A former retrospective analysis with 116 patients treated with AA at the Princess Margaret Hospital in Toronto showed that bone localisation could impact PSA response [
15] further strengthening our current observation. We also report for the first time that duration of hormonal sensitivity was associated with prolonged survival following AA treatment, asking the question whether AA should be prescribed for patients with particularly high levels of resistance to hormonal therapy.
Recent studies have shown new options for the treatment of mCRPC: the use of AA as first line treatment in chemo-naïve patients [
13] or the use of enzalutamide as first [
16] or second line [
17] treatment. However, there is no study evaluating the different possible sequences with the three drugs that are currently approved or going to be approved as first line treatment. Preclinical data showed impaired efficacy of docetaxel and cabazitaxel in abiraterone-resistant prostate cancer cell lines [
18]. These data were reinforced by clinical studies evidencing a lower activity of docetaxel in patients pre-treated with AA [
19,
20]. Thus, the question of AA positioning in terms of clinical benefit in a chronic disease where patients could live up to 3 years remains open. The results of our study tend to suggest that using AA post docetaxel is an excellent option with a median OS of 37 months.
Acknowledgements
The authors are grateful to Dr Mariella Lomma for her editorial assistance and Dr A. Kramar for insightful comments. This work was supported by Janssen. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests
David Azria, Philippe Beuzeboc, Aude Flechon and Nadine Houédé had a consultant or advisory role for Janssen. All remaining authors have declared no conflicts of interest.
Authors’ contributions
NH conceived the study, provided with patients' data, analyzed the results and wrote the manuscript. PB conceived the study, provided with patients' data and helped to draft the manuscript. SG conceived the study, performed statistical analyses and helped to draft the manuscript. DT, LM, GG, RD, AF, IL, JMF, SO, ST, BL, DT, GR, and XR provided with patients data and helped to draft the manuscript. HA contributed to the establishment of the cohort database. DA conceived and coordinated the study, obtained funding, analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.