Abiraterone for the Treatment
`of mCRPC
`
`9
`
`Zafeiris Zafeiriou, Niven Mehra and Johann S. de Bono
`
`Introduction
`
`(PCa) was a
`identified that prostate cancer
`Huggins and Hodges first
`hormone-driven disease and established as standard treatment androgen deprivation
`by surgical castration or medical castration (administration of estrogens) [1]. Vir-
`tually all metastatic PCa respond to castration but disease invariably progresses
`after a median of 18–24 months. This state was initially characterized as “hormone
`insensitive” or “hormone resistant” PCa but these terms have now been abandoned
`for the term “castration-resistant PCa (CRPC)”.
`It was hypothesized early that the effect of castration could be augmented by
`suppression of adrenal androgen biosynthesis as after castration androgens of
`extragonadal origin were still existent at about 10 % of the pre-castration level. For
`that purpose adrenalectomy [2] and hypophysectomy [3] were attempted with poor
`
`Z. Zafeiriou N. Mehra
`Prostate Cancer Targeted Therapy Group, Institute of Cancer Research/the Royal
`Marsden NHS Foundation Trust, Sycamore House, Downs Road,
`Sutton SM2 5PT, Surrey, UK
`e-mail: Zafeiris.Zafeiriou@icr.ac.uk
`
`N. Mehra
`e-mail: niven.mehra@icr.ac.uk
`
`J.S. de Bono (&)
`Department of Drug Development Unit, Institute of Cancer Research/the Royal
`Marsden NHS Foundation Trust, Sutton, Surrey, UK
`e-mail: johann.de-bono@icr.ac.uk
`
`© Springer International Publishing Switzerland 2016
`K.C. Balaji (ed.), Managing Metastatic Prostate Cancer In Your
`Urological Oncology Practice, DOI 10.1007/978-3-319-31341-2_9
`
`125
`Amerigen Exhibit 1134
`Amerigen v. Janssen
`IPR2016-00286
`
`

`
`126
`
`Z. Zafeiriou et al.
`
`outcome but provided evidence of activity and support for the rationale of sup-
`pression of extra-testicular androgens. Later this was pursued with administration of
`corticosteroids to suppress the hypothalamic ACTH axis [4]. Ketoconazole, an
`imidazole antifungal, was subsequently used [5] which inhibits enzymes involved
`in adrenal steroid synthesis, including CYP17, albeit non-selectively and incom-
`pletely [6], with PSA responses in approximately 60 % of patients [7], but also
`grade 3–4 toxicities in approximately 20 % of them [8].
`
`Early Development of Abiraterone
`
`Ketoconazole’s efficacy in CRPC sparkled the development of more potent and
`selective inhibitors of CYP17A1 [9, 10]. Pioneering work at the ICR in Sutton, UK,
`identified in the early 1990s that both steroidal [11] and nonsteroidal pyridyl esters
`[12, 13] exhibited selectivity for CYP17A1 inhibition.
`The steroidal inhibitors are derived from the natural substrates of CYP17A1,
`pregnenolone or progesterone [14], bind irreversibly to CYP17A1 and have
`increased inhibitory properties over many nonsteroidal agents. The 3-pyridyl ana-
`logues (Figs. 9.1 and 9.2) result in more potent inhibition of CYP17 relative to the
`2-pyridyl and 4-pyridyl analogues [12] and the 16,17-double bond adjacent to the
`pyridyl residue is necessary for the irreversible binding [15].
`One such steroidal compound, CB7630, later known as abiraterone acetate
`(AA) or JNJ-212082 is a 3ß-O-acetated version and a prodrug of CB7598, with
`better bioavailability and easier formulation than the 3ß-hydroxy compound
`(Figs. 9.1 and 9.2) and its administration in mice was able to decrease circulating
`levels of testosterone [10].
`
`Fig. 9.1 Abiraterone structure. Abiraterone (CB7598) for R=H and abiraterone acetate (CB7630)
`for R=Ac. Adapted from O’Donnell [16]. With permission from Nature Publishing Group
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`127
`
`Fig. 9.2 The parent ABCD
`steroid ring system
`(hydrocarbon framework) is
`shown with IUPAC-approved
`ring lettering and atom
`numbering
`
`Phase I and II Development
`
`In 2004 the first in human trial of AA in PCa was reported and consisted of three
`small phase I studies of AA without the use of concomitant corticosteroids in
`castrate and non-castrate patients [16]. Administration was continuous in one of the
`arms and once-off in the other two and AA was dose-escalated from 10 to 800 mg
`with pharmacokinetic (PK) studies indicating good bioavailability and supporting
`once-daily dosing. AA was well tolerated without documented grade 3–4 toxicities.
`Abiraterone plasma concentrations were detected at doses of AA ≥ 200 mg with
`consistent effects on testosterone levels when dose-escalated to the 500 and 800 mg
`dose level. In castrate males suppression of testosterone was sustained, while in
`non-castrate males a compensatory increase in LH levels overcame inhibition of
`gonadal testosterone synthesis. Boehringer Ingelheim, to which abiraterone was
`licensed at that time, suspended further clinical development despite these initial
`data. Later on, the interest in abiraterone was re-sparkled, probably because of
`accumulating data supporting an important
`role of CYP171A in castration
`resistance.
`In 2008 and 2010 two additional phase I trials were reported respectively
`[17, 18]. In the first study, twenty-one chemotherapy-naive CRPC patients were
`dose-escalated from 250 to 2000 mg once daily, with 3 patients per cohort and 6
`additional patients in the 1000 mg group to complete PK and pharmacodynamic
`(PD) studies. The recommended dose for the phase II development was determined
`at 1000 mg OD as above that level there was a plateau of the endocrine effect of
`AA. In the 2010 study 33 chemotherapy-naïve CRPC patients were enrolled, of
`which 19 had received prior ketoconazole therapy. Patients were dose-escalated
`from 250 to 1000 mg once daily; all patients received a single dose at day 1, and
`continuous dosing from day 7 with fed and fasted cohorts at each dose level. PSA
`
`

`
`128
`
`Z. Zafeiriou et al.
`
`and objective responses were commonly witnessed and warranted further
`evaluation.
`The first phase II study on AA was reported in 2009 on 42 chemotherapy-naïve
`patients [19], and was a phase II expansion of the UK phase I study reported in
`2008 [17]. Encouraging PSA declines of ≥50 % were seen in 28 of 42 CRPC
`patients, with objective responses by RECIST in 9 of 24 patients with measurable
`disease. In this study the median time to PSA progression was 225 days (95 % CI,
`162 to 287 days). Subsequently further phase II trials were performed which are
`summarized in Table 9.1, both in the pre- as well as in the post-chemotherapy
`setting and strengthened proof of activity across the whole spectrum of mCRPC
`with a frequency of PSA declines of ≥50 % between 36 % and 51 % which
`appeared higher in ketoconazole-naïve rather than in ketoconazole-pretreated
`patients. In the first reported study AA was given without the use of concomitant
`steroids [20] whereas later prednisone 5 mg was prescribed twice daily from the
`beginning of the treatment [21].
`
`Pharmacokinetics and Pharmacodynamics of Abiraterone
`and Interaction with Other Drugs
`
`Pharmacokinetics
`
`PK data of Abiraterone are available from the first phase I studies [17, 18] as well as
`from two post-licensing studies: a PK study in healthy men [22] and a population
`PK analysis with 359 subjects including 62 healthy volunteers [23].
`Pharmacokinetics of Abiraterone are strongly influenced by food and therefore it
`is administered routinely in the fasted state, 2 h after a previous meal and 1 h before
`the next meal. Following oral administration AA is converted to its active
`metabolite Abiraterone by rapid hydrolysis taking place in the intraluminal envi-
`ronment of the intestine [24] but also in the liver involving hydroxylation mediated
`by esterases [25]. Therefore, AA is below detectable levels in the plasma [22] and
`Abiraterone is detectable instead at doses of AA ≥200 mg. Abiraterone reaches its
`maximum concentration (Cmax) of 1.2–5 μΜ in approximately 1–2 h in fasting
`patients and 4 h following a high-fat meal (Tmax) [17] and subsequently follows a
`biphasic elimination [22] with a terminal half-life of 5–16 h [17, 18, 22]. Further
`metabolic reactions generate its two main but inactive circulating metabolites,
`abiraterone sulphate and N-oxide abiraterone sulphate. The main metabolite
`excreted in urine is N-oxide abiraterone sulfate. Faeces is the primary route of
`excretion (87.9 %), with major components unchanged AA (55.3 %) and abi-
`raterone (22.3 %) [23, 26]. Interestingly, the apparent clearance of Abiraterone is
`lower in CRPC patients compared to healthy subjects [23].
`There is a significant inter-subject and within-subject variability regarding Cmax
`and drug exposure, as measured with the area under the concentration-time curve
`(AUC) of approximately 40–70 % [22, 23]. There is also variability between cycles,
`with the Cmax observed in cycle 1 being higher by 10–15 % than the Cmax at later
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`129
`
`Danila (2010)
`[21]
`
`Satoh (2014) [28]
`
`Kwak (2014)
`
`Prior
`ketoconazole
`(%)
`
`2 (4.8 %)
`
`0
`
`NR
`
`PSA ≥50 %
`(%)
`
`28 (67 %)
`
`26 (79 %)
`
`30 (63 %)
`
`ORR (%,
`evaluable
`patients)
`
`Time to PSA progression,
`median days (95 % CI)
`
`Population
`
`First author (year
`of publication)
`
`9/24 (37.5 %)
`
`225 (162–287)
`
`caucasian
`
`Attard (2009) [19]
`
`9/13 (69 %)
`
`4/18 (22 %)
`
`496 (280–NE)
`NEa
`
`caucasian
`
`Ryan (2011) [41]
`
`asian
`
`Matsubara (2014)
`[27]
`
`8 (17 %)
`
`29 (51 %)
`
`8/30 (27 %)
`
`169 (113–281)
`
`caucasian
`
`Reid (2010) [20]
`
`27 (46.6 %)
`
`22 (36 %)
`
`4/22 (18 %)
`
`169 (82–200)
`
`caucasian
`
`Table 9.1 Phase II studies of AA in CRPC
`
`Phase II studies of abiraterone acetate in CRPC
`
`Patients
`
`Chemotherapy
`setting
`
`Naïve
`
`Naïve
`
`Naïve
`
`Pretreated
`
`Pretreated
`
`42
`
`33
`
`48
`
`47
`
`58
`
`46
`
`Pretreated
`
`0
`
`Pretreated
`35
`aMedian time not reached
`Abbreviations: NE not estimable; NR not reported
`
`0
`
`16 (35 %)
`
`35 (43 %)
`
`1/22 (4 %)
`
`2/50 (4 %)
`
`142 (85–NE)
`
`143 (113–252)
`
`asian
`
`asian
`
`

`
`130
`
`Z. Zafeiriou et al.
`
`cycles [17, 18]. AUC and Cmax increase with incremental dose proportionally
`between 750 and 1000 mg but more than proportionally between 500 and 750 mg
`[17, 22]. When administered with high-fat content food, drug exposure is increased
`by 4.4-fold compared to fasting achieved by extended absorption without any
`observed increase of Cmax [17, 18]. Beyond food intake and cancer status, other
`covariates such as age, prior chemotherapy status, testosterone level, body mass
`index and total plasma proteins were not found to impact the PK of Abiraterone [23].
`
`Pharmacodynamics
`
`The changes of hormone levels of the steroidogenesis pathway upstream and
`downstream of CYP17 (Fig. 9.3) served as pharmacodynamic markers for the
`phase I trials. Both testosterone and androstenedione were suppressed to unde-
`tectable levels (<1 ng/dL) at all Abiraterone dose levels already by day eight for the
`study by Attard et al. and by day 28 in Ryan et al. (Fig. 9.4) [17]. By day 28,
`DHEA was decreased by a median of 3.4-fold but was still at detectable levels in
`contrast to the other androgens; this was caused by interference of DHEA with the
`used assay. ACTH was increased up to six-fold, as well as the levels of steroid
`precursors upstream of CYP17A1, namely deoxycorticosterone (median 10-fold)
`and corticosterone (median 40-fold) with a plateau of endocrine effects observed at
`the dose level of 750 mg. Administration of dexamethasone, given at PSA pro-
`gression, resulted in suppression of ACTH and decrease of upstream steroids. The
`recommended phase II dose was selected at 1000 mg once daily based on these PD
`endocrine studies [17]. The phase I study by Ryan et al. confirmed the aforemen-
`tioned endocrine changes and concluded to the same recommended dose for the
`phase II [29].
`
`Drug Interactions
`
`AA has been shown to be a strong inhibitor of CYP1A2 and CYP2D6 and a
`moderate inhibitor of CYP2C9, CYP2C19, CYP3A4 and CYP3A5 in vitro [30].
`A phase I drug-drug interaction (DDI) study in CRPC indicated only a relevant
`in vivo interaction between AA-P and CYP2D6, but not with CYP1A2, after single
`dose co-administration with dextromethorphan, a CYP2D6 substrate, and theo-
`phylline, a CYP1A2 substrate [31]. An in vitro study indicated Abiraterone could
`its substrates 1α,25-dihydrox-
`decrease CYP3A4-dependent hydroxylation of
`yvitamin D3 (calcitriol, active vitamin D) and midazolam [32]. In view of the
`reported in vivo interactions caution especially is indicated with concomitant
`administration of AA with known CYP2D6 substrates.
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`131
`
`Fig. 9.3 CYP17 has 17a-hydroxylase and C17,20-lyase activity necessary for the conversion of
`21-carbon pregnanes into 19-carbon sex steroid precursors. a Androgen biosynthesis pathway.
`b Abiraterone inhibits 17a-hydroxylase resulting in feedback increase in ACTH with subsequent
`increase in deoxycorticosterone and corticosterone and symptoms of mineralocorticoid excess.
`Inhibition of C17,20-lysase results in suppression of DHEA, androstenedione and testosterone.
`Adapted from Attard et al. [17]. With permission from American Society of Clinical Oncology
`
`Mechanisms of Action of Abiraterone
`
`Abiraterone blocks the CYP17A1 17α-hydroxylase/17,20-lyase enzyme via for-
`mation of a covalent bond between the nitrogen of its pyridine group and the haem
`iron of CYP17 (Fig. 9.5a,b) while a hydrogen bond network contributes to the
`interaction [33]. Inhibition of the hydroxylase activity suppresses hydroxylation of
`pregnenolone and progesterone at C17 whereas inhibition of the lyase activity limits
`
`

`
`132
`
`Z. Zafeiriou et al.
`
`Fig. 9.4 Pharmacodynamic end points of the treatment with single agent Abiraterone Acetate.
`Treatment with Abiraterone Acetate results in significant suppression of testosterone, dehy-
`droepiandrostenedione (DHEA), and androstenedione (a–c) and an increase of corticosterone and
`deoxycorticosterone (d–e). Plateau of endocrine effect after the dose of 750 mg (f). Adapted from
`Attard et al. [17]. With permission from American Society of Clinical Oncology
`
`Fig. 9.5 Representation of CYP17A1-Abiraterone bound structure. a CYP17A1 from the N
`terminus (blue) to the C terminus (red). b Abiraterone binds at an angle of approximately 60° from
`haem against helix I (yellow). Adapted from DeVore and Scott [33]. With permission from Nature
`Publishing Group
`
`the subsequent conversion of the hydroxylated metabolites to dehydroepiandros-
`terone and androstenedione, respectively, resulting in decreased testosterone and
`In addition to its 17α-hydroxylase/C17,20-lyase
`DHT levels (Fig. 9.3a, b).
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`133
`
`Fig. 9.6 D4A, an Abiraterone metabolite, has the ability to interact with multiple enzymes of the
`androgen biosynthesis pathway. a Conversion from Abiraterone to D4A by 3ßHSD. b D4A IC50
`inhibitory activity in comparison to Abiraterone and Enzalutamide. Abbreviations: AD
`androstenedione; A5diol ∆5-androstenediol; D4A ∆4-Abiraterone. Adapted from Li et al. [36].
`With permission from Nature Publishing Group
`
`inhibiting activity, Abiraterone also was shown to block 3-hydroxysteroid dehy-
`drogenase (3ßHSD) [34] which transforms DHEA to androstenedione. Abiraterone
`can also interact with the AR directly acting in vitro as an antagonist of both the wild
`type and mutant AR but at the dose of 5 μM which is not achieved in plasma under
`normal administration [35]. It was also found to undergo a 3ßHSD enzymatic
`conversion into ∆4-Abiraterone, in which the double bond is moved from C5 to C4
`(D4A; Figs. 9.2, 9.6a) [36]. D4A shares identical steroid A and B rings with
`testosterone, enabling a strong antagonistic interaction with wildtype and mutant
`AR. D4A also inhibits both 3ßHSD and 5α-reductases (SRD5A) (Fig. 9.6b) while it
`retains CYP171A inhibitory activity comparable to Abiraterone.
`Mineralocorticoids are able to inhibit AR transcriptional activity in vitro in the
`presence of
`androgens,
`at
`concentrations
`similar
`to those measured in
`Abiraterone-treated patients [37]. In the phase III studies a significant proportion of
`patients experienced symptoms associated with mineralocorticoid overload, and it
`is not inconceivable, that in this group of patients the mineralocorticoid antagonistic
`effects on AR could contribute to Abiraterone’s clinical activity.
`
`Efficacy of Abiraterone
`
`After the promising results of the phase I and II trials AA was further evaluated in
`the phase III trials COU-AA-301 and COU-AA-302 in patients with mCRPC who
`had progressed on docetaxel in the former and were chemotherapy naive in the
`latter. In both trials patients were blindly randomized between the combination of
`AA 1000 mg once a day with prednisone 5 mg twice a day in the experimental arm
`—from now on in this work designated as AAP—and placebo plus prednisone
`
`

`
`134
`
`Z. Zafeiriou et al.
`
`5 mg twice a day in the control arm respectively—from now on designated as
`PP. Inclusion criteria, randomization scheme, numbers of patients and stratification
`factors are summarized in Table 9.2.
`The first interim analysis of COU-AA-301 was announced after 552 events and a
`median follow-up of 12.8 months (m) and showed an absolute improvement in the
`median Overall Survival (mOS) of 3.9 m in AAP compared to PP which remained
`significant after adjusting for stratification factors [38]. Subsequently the trial was
`un-blinded, crossover of patients from placebo to active drug was allowed and the
`drug gained regulatory approval for mCRPC patients who progressed after treatment
`with docetaxel. In the final survival analysis performed with the data available before
`crossover [39], after a median follow up of 20.2 months and 775 death events both
`the primary and secondary endpoints remained in favour of the AAP arm with an
`absolute benefit of 4.6 m in mOS (Table 9.3; Fig. 9.7). In multivariate analysis all
`the pre-specified stratification factors proved to be prognostic for survival and AAP
`showed superior survival compared to PP in all subgroups, while in some of them it
`did not reach statistical significance probably due to their small size (Fig. 9.8).
`The COU-AA-302 trial enrolled chemo-naive patients with good prognostic
`parameters (Table 9.2) i.e. absence of significant pain, liver disease, low hae-
`moglobin or albumin. The primary endpoints of the trial were radiographic PFS and
`OS and clinically meaningful secondary endpoints were captured: time to cytotoxic
`chemotherapy initiation, time to opiate use for cancer related pain, time to prostate
`specific antigen progression and time to performance status deterioration.
`Two interim analyses (IA) were published on this trial and neither could
`demonstrate a survival benefit consistent with the pre-specified criterion for efficacy
`[40, 41]. This was subsequently met in the final overall analysis [42]. The first IA
`was performed after 43 % of the death events and showed a beneficial effect of
`AAP in the risk for radiographic progression or death with a hazard ratio of 0.49 as
`well as in the risk for death with a HR of 0.75[41] (Table 9.5). All clinically
`meaningful secondary endpoints-median time to cytotoxic chemotherapy, median
`TTPP, median time to performance status deterioration by 1 grade [41]
`(Table 9.4)-favoured AAP despite the fact that the median OS of 27.2 m reported in
`the PP group was the largest ever reported in that patient population. As a result the
`study was un-blinded and crossover allowed and AA’s regulatory approval in the
`US and Europe was expanded to the pre-chemotherapy setting. The second pub-
`lished IA was performed shortly after the un-blinding of the study when only 3
`patients had crossed over to active treatment from placebo. Again AAP showed an
`improved trend in OS compared to PP and an impressively improved time to rPFS
`of 16.5 m compared to 8.2 m of PP and this favourable effect was consistent in all
`stratification subgroups irrespective of age, baseline PSA, serum LDH or ALP,
`ECOG PS, presence or not at entry of pain or bone metastases [40]. In the final
`overall analysis after 49.2 months of follow up, 65 % of patients had died in AAP
`compared to 71 % in PP [42]. Only 67 % of the patients in the Abiraterone group
`compared to 80 % of patients in the placebo group had received further treatments.
`The final analysis showed increased mOS in the Abiraterone group of 34.7 months
`versus 30.3 months in the placebo group which met the pre-specified statistical
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`135
`
`Table 9.2 Inclusion criteria and stratification factors in COU-AA-301 and COU-AA-302
`
`COU-AA-301
`
`Number of patients
`
`1195
`
`Randomization
`ratio
`
`Inclusion criteria
`
`2:1
`
`Prior treatment
`
`Post docetaxel
`
`ECOG PS
`
`≤2
`
`ALT and AST
`
`<2.5 times ULN or < 5 times the ULN if liver
`lesions present
`
`Albumin
`
`Haemoglobin
`
`>3 g/dL
`
`≥9.0 g/dL
`
`Platelet Count
`
`≥100,000/µL
`
`Visceral disease
`
`Allowed
`
`Score of BPI-SF
`question 3
`
`Any
`
`Type of progression
`to prior treatment
`
`Primary endpoints
`
`Radiologic or PSA progression
`
`Overall survival
`
`Secondary endpoints
`
`Time to PSA progression
`
`Radiologic progression free survival
`
`PSA response rate
`
`Stratification factors
`
`ECOG performance status 0–1 versus 2
`
`Presence of significant pain in the past 24 h before
`randomization as captured by the Brief Pain
`Inventory-Short Form (BPI-SF) question 3 (Yes vs.
`No)
`
`Number of previous chemotherapy regimens (1 vs.
`2)
`
`PSA progression only versus radiographic
`progression regardless of PSA progression
`
`COU-AA-302
`
`1088
`
`1:1
`
`Pre docetaxel
`
`≤1
`
`<2.5 times ULN
`
`>3.5 g/dL
`
`≥10.0 g/dL
`
`≥100,000/µL
`
`Not allowed
`
`0–3
`
`Radiologic or PSA
`progression
`
`Overall survival
`
`Radiologic progression
`free survival
`
`Time to PSA progression
`(TTPP)
`
`Time to opiate use for
`cancer-related pain
`
`Time to initiation of
`cytotoxic chemotherapy
`
`Time to ECOG
`performance status
`deterioration
`
`ECOG performance
`status 0 versus 1
`
`

`
`136
`
`Z. Zafeiriou et al.
`
`Table 9.3 Outcomes of efficacy endpoints in COU-AA-301
`
`COU-AA-301
`
`Duration of drug exposure
`(months)
`
`Median overall survival
`(months)
`
`Time to PSA progression
`(months)
`
`rPFS (months)
`
`PSA response rate
`
`Response rate by RECIST in
`patients with measurable
`disease
`
`COU-AA-301 1st
`analysis [38]
`
`interim
`
`COU-AA-301 final analysis [39]
`
`AAP PP HR [95 % CI]
`
`AAP
`
`PP
`
`HR [95 % CI]
`
`8
`
`4
`
`7.4
`
`3.6
`
`14.8
`
`10.9 0.65 [0.54–0.77]
`
`15.8
`
`11.2
`
`0.74[0.64–0.86]
`
`10.2
`
`6.6
`
`0.58; [0.46–0.73]
`
`8.5
`
`6.6
`
`0.63[0.52–0.78]
`
`5.6
`
`3.6
`
`0.67[0.58–0.78]
`
`5.6
`
`3.6
`
`0.66[0.58–0.76]
`
`29 % 6 %
`
`14 % 3 %
`
`29.5 % 5.5 %
`
`14.8 % 3.3 %
`
`Fig. 9.7 Overall survival in the COU-AA-301 trial as evaluated in the last analysis at the time
`point of crossover. Adapted from Fizazi et al. [39]. With permission from Elsevier
`
`boundary for efficacy. The radiographic progression free survival endpoint was also
`met by the trial [42] (Table 9.5).
`It is noteworthy that the Kaplan-Meier curves in COU-AA-302 started sepa-
`rating at 12 months and most clearly so after 18 months [42] which also explains
`why there was no significant survival benefit evident between AAP and P in the
`interim analyses (Fig. 9.9). This can be attributed to the inclusion of patients with
`good prognosis and the resulting small number of death events in the beginning of
`the trial. Still, during the first 12 months of the trial approximately 10 % of the
`patients died and 25–30 % of these deaths could be attributed to non cancer-related
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`137
`
`Fig. 9.8 Overall survival in COU-AA-301 based on duration of previous docetaxel chemother-
`apy (a ≤3 months and b >3 months). The group of ≤3 months seems to have a non-statistically
`significant HR of 0.76 (95 % CI 0.53–1.08) and could be considered as not deriving any benefit
`from treatment with Abiraterone. Nevertheless, the 209 patients represented in this group are
`substantially less compared to the 981 of the >3 months group. Additionally, the survival curves
`appear to have a divergent course. Adapted from Fizazi et al. [39]. With permission from Elsevier
`
`Table 9.4 Outcomes of secondary endpoints in COU-AA-301 and COU-AA-302
`
`AAP [95 %
`CI] (months)
`
`PP [95 % CI]
`
`COU-AA-301
`
`Improvement of fatigue intensity (%) [43]
`
`Improvement of fatigue interference (%) [43]
`
`Median time to fatigue improvement (days) [43]
`
`Time to fatigue intensity progression [25th percentile]
`(days) [43]
`
`Pain palliation (%) [44]
`
`Median time to palliation of pain intensity (months) [44]
`
`Median duration of palliation of pain intensity (months)
`[44]
`
`Median time to occurrence of first skeletal related event
`(months) [44]2
`
`Improvement in the FACT-P total score (%) [46]
`
`Median time to deterioration of FACT-P(weeks) [46]
`
`COU-AA-302
`
`Median time to cytotoxic chemotherapy [40]
`
`Median time to decline in ECOG PS
`
`Pain
`
`Median time to opiate use (months) [40]
`
`Median time to pain progression (months) [41]
`
`58.1
`
`55
`
`59
`
`232
`
`45
`
`5.6
`
`4.2
`
`25
`
`48
`
`59.9
`
`26.5
`
`12.3
`
`33.4
`
`26.7 [19.3–
`not estimable]
`
`Median time to progression of the pain interference with
`daily activities (months) [45]
`
`(10.3 [9.3–
`13.0]
`
`40.3
`
`38
`
`194
`
`139
`
`28.8
`
`13.7
`
`2.1
`
`20.3
`
`32
`
`36.1
`
`16.8
`
`10.9
`
`23.4
`
`18.4 [14.9–not
`estimable]
`
`7.4 [6.4–8.6]
`
`(continued)
`
`

`
`138
`
`Table 9.4 (continued)
`
`Z. Zafeiriou et al.
`
`AAP [95 %
`CI] (months)
`
`PP [95 % CI]
`
`Median time to progression of the worst pain (months)
`[45]
`
`26.7 [19.4–not
`estimable]
`
`(19.4 [16.6–not
`estimable])
`
`QOL
`
`Median time to deterioration of QOL (as measured by
`FACT-P) (months) [45]
`
`Median time to deterioration of QOL (as measured by
`PCa specific FACT-P subscale) (months) [45]
`
`12.7 [11.1–
`14.0]
`
`11.1 [8.6–
`13.8]
`
`8.3 [7.4–10.6]
`
`5.8 [5.5–8.3]
`
`Table 9.5 Outcomes of efficacy endpoints in COU-AA-302
`
`COU-AA-302 1st
`analysis [41]
`
`interim
`
`COU-AA-302 2nd
`interim analysis [40]
`
`COU-AA-302 final
`analysis [42]
`
`AAP
`
`PP
`
`HR [95 %
`CI]
`
`AAP PP
`
`HR [95 %
`CI]
`
`AAP PP
`
`HR [95 %
`CI]
`
`Median overall
`survival (months)
`
`Not
`reached
`
`27.2
`
`0.75
`[0.61–0.93]
`
`35.3
`
`30.1
`
`0.79
`[0.66–0.95]
`
`34.7
`
`30.3 0.81 [95 %
`CI 0 ·70–
`0 · 93]
`
`Time to PSA
`progression (months)
`
`11.1
`
`5.6
`
`rPFS (months)
`
`16.5
`
`8.3
`
`0.49
`[0.42–0.57]
`
`0.53
`[0.45–0.62]
`
`NR
`
`NR
`
`NR
`
`NR
`
`NR NR
`
`16.5
`
`8.2
`
`0.52
`[0.45–0.61]
`
`NR
`
`NR NR
`
`62 %
`
`36 %
`
`24 %
`
`16 %
`
`68 % 29 %
`
`NR
`
`NR
`
`NR
`
`NR
`
`NR NR
`
`NR
`
`PSA response rate
`
`Response rate by
`RECIST in patients
`with measurable
`disease
`
`NR not reported
`
`morbidity [42]. The rest of the patients who died in this time period, approximately
`7 % of the total patients, potentially reflect patients with disease of bad prognosis
`that was not captured by the prognostic factors applied in the inclusion criteria and
`who did not respond either to Abiraterone or to prednisone.
`
`Effect of Abiraterone on Secondary Endpoints
`
`In both COU-AA-301 and COU-AA-302 trials secondary endpoints were captured
`to substantiate the clinical effect of Abiraterone and all of them favoured AAP over
`PP (Table 9.4).
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`139
`
`Fig. 9.9 Overall survival curves in COU-AA-302. Noteworthy is the fact that the curves started
`separating only after 12–18 months of treatment. Adapted from Ryan et al. [42]. With permission
`from Elsevier
`
`Fatigue
`
`In COU-AA-301 fatigue was evaluated at each cycle of treatment by assessment of
`the Brief Fatigue Inventory questionnaire which evaluates fatigue intensity and
`fatigue interference with aspects of everyday life. Improvement of fatigue intensity
`and fatigue interference was significantly higher in AAP compared to PP (58.1 %
`vs. 40.3 % and 55 % vs. 38 % respectively). Additionally, in the AAP arm the time
`to improvement of fatigue was substantially shorter and the time to fatigue pro-
`gression delayed compared to the PP arm [43]. In COU-AA-302 fatigue was not
`prospectively captured.
`
`Pain Palliation and Progression
`
`Pain was evaluated throughout COU-AA-301 and COU-AA-302 by means of the
`BPI-SF. Pain intensity was defined as the score of the BPI-SF question 3 (worst
`pain in the previous 24 h) and pain interference was defined as the mean of all
`seven questions assessing pain interference with everyday activities on the BPI-SF.
`Additionally, in COU-AA-301 the mean analgesic use was scored as per WHO
`criteria. In a post hoc analysis of COU-AA-301 with the data available at the
`time-point before the crossover to active drug the median time to pain progression
`was not reached at that time point, yet significantly more patients in the AAP group
`had experienced pain palliation and improvement of pain interference than in the PP
`group and pain palliation occurred faster in AAP (Table 9.4) [44].
`
`

`
`140
`
`Z. Zafeiriou et al.
`
`In COU-AA-302 patients did not have significant pain at treatment initiation but
`AA delayed pain occurrence compared to placebo as indicated by the delayed time
`to all pain related parameters: median time to pain progression, median time to
`opiate use [41, 42], median time to progression of the pain interference with daily
`activities and the median time to progression of the worst pain [45] (Table 9.4).
`
`Skeletal Related Events
`
`In COU-AA-301 the proportion of patients with skeletal related events was similar
`in both groups but there was a significantly longer median time to skeletal related
`events in Abiraterone (25 m) compared to placebo (20.3 m)(Table 9.4) [44].
`Skeletal related events were not captured in COU-AA-302.
`
`Quality of Life
`
`In COU-AA-301 and 302 health related quality of life (HRQoL) was assessed by
`means of the Functional Assessment of Cancer Therapy-Prostate (FACT-P) ques-
`tionnaire which has subscales to assess different aspects of QoL. In COU-AA-301,
`AA improved quality of life in more patients (48 % vs. 32 %) and in a shorter time
`compared to prednisone and delayed functional status deterioration by 6 months
`despite the fact that prednisone was able to produce substantial improvements in
`QOL as well [46]. In the minimally symptomatic population of COU-AA-302, AA
`delayed median time to deterioration of quality of life both as measured by the
`FACT-P total score and by the prostate-cancer-specific subscale [45] (Table 9.4).
`
`Toxicity of Abiraterone
`
`AA is a well tolerated drug as indicated by the fact that in COU-AA-301 the
`frequency of adverse events (AE) leading to treatment discontinuation was lower in
`AAP compared to PP (19 % vs. 23 %) [38] and in COU-AA-302 [42] the same
`frequency was again at the level of 19 % for AAP but only 12 % for PP [41], the
`latter reflecting a population with less advanced and symptomatic disease.
`The most frequent adverse events in both trials were fatigue, back pain, con-
`stipation, diarrhoea, nausea, bone pain and arthralgia. These occurred in similar
`frequency in the experimental and control groups [38, 39, 41, 42] and constitute
`disease related symptoms rather than Abiraterone related AE (Table 9.6). A con-
`tribution of Abiraterone to arthralgia, fatigue and diarrhoea cannot be excluded as
`these AEs showed consistently a mildly increased frequency in both trials in the
`AAP over the PP arm (Table 9.6).
`
`

`
`9 Abiraterone for the Treatment of mCRPC
`
`141
`
`Table 9.6 Adverse events grade 1–4 in AA phase III trials
`
`Adverse event
`
`COU-AA-301
`
`COU-AA-302
`
`AAP (%)
`
`PP (%)
`
`AAP (%)
`
`PP (%)
`
`Anaemia
`
`Fatigue
`
`Back pain
`
`Arthralgia
`
`Nausea
`
`Vomiting
`
`Constipation
`
`Hot flush
`
`Diarrhoea
`
`Bone pain
`
`Pain in extremity
`
`Cough
`
`Urinary tract infection
`
`25
`
`47
`
`33
`
`30
`
`33
`
`24
`
`28
`
`NR
`
`20
`
`27
`
`20
`
`NR
`
`13
`
`Abbreviations: NR not reported
`Based on data from Refs. [38, 41, 42]
`
`28
`
`44
`
`36
`
`24
`
`33
`
`26
`
`32
`
`NR
`
`15
`
`30
`
`21
`
`NR
`
`7
`
`NR
`
`39
`
`32
`
`28
`
`22
`
`NR
`
`23
`
`22
`
`22
`
`20
`
`17
`
`17
`
`2
`
`NR
`
`34
`
`32
`
`24
`
`22
`
`NR
`
`19
`
`18
`
`18
`
`19
`
`16
`
`14
`
`<1
`
`On the other hand, AEs related to mineralocorticoid excess (ME) (hypokalemia,
`fluid retention, hypertension) were more frequent in AAP compared to PP in both
`trials [38, 42] (Table 9.7). ME observed in CRPC patients treated with single agent
`AA can be managed either by the addition of eplerenone, a mineralocorticoid
`receptor antagonist or by exogenous corticosteroids [17, 47] that inhibit secretion of
`corticotrophin-releasing hormone and ACTH, therefore exogenous corticosteroids
`are routinely co-administered with Abiraterone. Use of spironolactone for the
`management of ME should be avoided as it can act as an agonist for wildtype AR
`[35, 48]. Hypertension that arises despite steroid use can be managed symp-
`tomatically by addition of an antihypertensive agent but diuretics should be avoided
`as they may exacerbate Abiraterone induced hypokalemia. Hypokalemia can be
`treated by supplementation with potassium.
`Cardiac events were consistently more frequent in AAP rather than in PP in both
`trials although this difference did not reach statistical significance. Most frequent
`were grade 1 or 2 tachycardia and grade 3 or less atrial fibrillation [39] and fatal
`cardiac events were rare and approximately 1 % for each arm in both trials [38, 42]
`(Table 9.7). No effect of AA could be detected on the QTcF and QRS interval by a
`further small study evaluating ECG changes with AA [49].
`Early in the course of the COU-AA-301 trial an incident of grade 4 liver function
`test (LFT) elevation instigated patients’ vis