Abiraterone acetate has significant activity in patients with CRPC, as evidenced by a PSA response rate of 58% in this phase I trial and declines in PSA on all dose levels. Although the utility of using PSA reductions as a marker of clinical activity is debated, in general, most investigators agree that it is a reasonable tool to screen for activity. The activity of abiraterone acetate is attributed to the reduction of the total androgen pool, with a reduction in levels of both adrenal androgens and testosterone, thereby inhibiting persistent signaling through the AR. It is of considerable interest that abiraterone acetate has demonstrated activity in patients previously treated with ketoconazole.
Dose escalation was not discontinued as a result of the presence of DLTs. Abiraterone acetate was well tolerated up through the highest dose level evaluated (1,000 mg/d) with no MTD observed and no apparent toxicity differences among patients who had or had not received prior ketoconazole use. On the basis of safety, endocrinologic and pharmacokinetic parameters, and indications of activity, an abiraterone acetate dose of 1,000 mg/d is recommended for further study.
The adverse event and steroid endocrine profiles were consistent with anticipated outcomes of selective CYP17 inhibition (decrease in androgens and concomitant increase in upstream mineralocorticoid production). At the 1,000-mg dose level, hypertension and fatigue were the most commonly observed toxicities. The use of beta-blockers, diuretics, and eplerenone (often at doses > 25 mg daily) was modestly effective in managing abiraterone-induced hypertension. Administration of a corticosteroid was associated with normalization of mineralocorticoid levels and improvements in blood pressure. Of interest, two patients who received dexamethasone, which has little mineralocorticoid activity, experienced orthostatic hypotension, possibly related to a rapid decline in mineralocorticoid levels. This was not seen in patients who received hydrocortisone or prednisone. Therefore, we suggest that a low dose of a corticosteroid, such as prednisone, be administered with abiraterone acetate in future studies to optimize the safety profile.
The endocrine changes observed with abiraterone acetate were also consistent with its known inhibitory effects on 17α-hydroxylase and C17,20-lyase activities. Serum DHEA-S and testosterone declined substantially at all dose levels. At day 28, increased levels of adrenal steroids and upstream precursor molecules, including mineralocorticoids (corticosterone and deoxycorticosterone) were seen, consistent with a feedback mechanism resulting from increasing ACTH levels. Toxicities of mineralocorticoid excess (hypertension and potassium wasting) were mitigated with corticosteroid use as a result of suppression of the hypothalamic-pituitary adrenal axis.
No differences were observed in toxicity or PSA declines between fasted and fed patients. Interestingly, a high-fat meal substantially increased total drug exposure (AUC) in the fed patients, whereas an increased clearance and volume of distribution were observed in the fasted state. Although not directly observed, the potential for greater drug exposure to increase toxicities cannot be ruled out. To minimize diet-related variability in drug exposure, we suggest that future studies administer abiraterone acetate in a fasted state.
Confirmed PSA declines of ≥ 50% were seen in 58% of patients overall, including a 47% PSA response rate in patients with disease progression on prior ketoconazole therapy; in phase III trials of ketoconazole, the PSA response rate was approximately 30%,11
and phase II studies showed a 45% to 65% rate of response.17
Of note, at the time of disease progression on ketoconazole,11
adrenal androgen levels were increased relative to their nadir, suggesting that the suppressive effects of ketoconazole on adrenal androgen synthesis are not durable. Furthermore, prolonged disease suppression by ketoconazole may be complicated by induction of metabolism or drug-drug interactions. By contrast, the suppressive effects of abiraterone on adrenal androgen synthesis are durable because progression of disease on abiraterone is not accompanied by an increase in adrenal androgens (), again potentially reflecting the more potent CYP17 inhibitory activity of abiraterone compared with ketoconazole. These observations may explain why a high proportion of patients were able to benefit from abiraterone after ketoconazole. Also of note, the patients with prior ketoconazole treatment had experienced longer than average responses to ketoconazole (15 months v
reported median responses of approximately 5 months) and thus may represent a group of patients more likely to harbor disease that is dependent on CYP17-mediated androgen production. Although this may be reflective of an unintended selection bias, the high rate of response to abiraterone in ketoconazole-exposed patients is further evidence of the superior potency of abiraterone compared with ketoconazole. More analysis of this relationship in future studies is warranted.
From our data, it cannot be determined whether pretreatment adrenal androgen levels can identify patients more likely to respond to abiraterone, as has been suggested with ketoconazole,18,19
or whether changes in adrenal androgen levels correlate with clinical outcomes. However, such correlations are of interest and should be pursued. The exclusion of patients with prior chemotherapy for CRPC in this study was intended to allow for preliminary evaluation of abiraterone acetate in a patient population typically treated with secondary hormonal manipulation, chiefly as a means of delaying use of traditional cytotoxics. Ongoing studies are evaluating abiraterone in patients with both chemotherapy-refractory and chemotherapy-naïve CRPC.
Despite these findings, several unaddressed issues remain. First, although abiraterone did not cause adrenal insufficiency in any patients, the impact of concurrent prednisone on response proportion and durability or long-term toxicity is not known. Second, despite the high PSA decline rate, the relationship between PSA decline and survival in the prechemotherapy setting is undefined. Finally, the observation of abiraterone efficacy in a ketoconazole-refractory CRPC population requires prospective validation. Such studies are ongoing.
In summary, these data suggest that abiraterone is an active agent for CRPC and is associated with acceptable toxicity. Mineralocorticoid-induced hypertension is a unique toxicity but is controlled by corticosteroids. Preliminary observations of responses to abiraterone in patients with prior ketoconazole therapy suggest that cross-resistance between these therapies may not exist. Further definitive studies of this agent are warranted.