We aimed to establish the dose reduction required to reach a bioequivalent exposure compared to the standard intake of 1000 mg abiraterone acetate OD without food. Although the GMRs of the intake of 500 mg abiraterone acetate with a continental breakfast compared to 1000 mg taken fasted are within the thresholds of 0.80 and 1.25, the 90% confidence intervals do not meet the criteria for bioequivalence. However, it is known that no exposure–toxicity relation has been found for abiraterone when investigated up until doses of 2000 mg [
8]. Therefore, the strict bio-equivalence margins as defined by the FDA guidelines could be applied more flexible for abiraterone.
To the best of our knowledge, this is the first study to investigate the bioequivalence of an adjusted AA dose when taken with food. Our data demonstrated that a 50% dose reduction resulted in a GMR for the AUC
0–24h and
Ctrough below 1.0. Based on our PK data, no further dose reduction was deemed feasible. However, Szmulewitz et al. have shown that a 75% reduced AA dose taken with a low-fat breakfast resulted in similar PSA response rates compared to the standard fasted intake of AA 1000 mg OD [
9]. The descriptive PK data in their study demonstrated that the abiraterone
Ctrough levels were lower in the group treated at the reduced dose with food compared to the full dose taken fasted [
9]. Nevertheless, despite the lower
Ctrough levels, the percentage of patients reaching an adequate PSA response remained comparable between both intake regimes [
9]. The majority of the patients in the study of Szmulewitz et al. was treated prior to receiving chemotherapy and might, therefore, be more sensitive to lower abiraterone concentrations. Xu et al. showed that the EC50 of the PSA for abiraterone was 1.56 ng/mL in chemotherapy-naïve patients and 4.75 ng/mL in patients who underwent previous chemotherapy [
16]. Though, due to the large confidence intervals around the observation of Xu et al., the data of those trials should be interpreted carefully. Our population mainly consisted of patients post chemotherapy, Therefore, exposure levels similar to those reached when 1000 mg AA taken fasted were aimed for. The results of our study are based on 12 patients instead of the predefined 24 patients. Variation of the individual GMRs of the AUC
0–24h and
Cmax was larger than expected. Re-estimation of the sample, size based on the larger variability as observed in our study, learned us that bioequivalence could not be demonstrated in the predefined number of 24 patients. Therefore we considered it unethical to further conduct the study. Nevertheless, as 12 patients are the minimal number necessary for a bioequivalence study as stated by the FDA, the PK results from our study are still of value for further interpretation [
17]. Since we could not demonstrate bioequivalence, the switch of large groups of patients to an alternative intake regime of 500 mg taken with food cannot be supported. However, because the GMRs of the AUC
0–24h,
Cmax and
Ctrough were within the threshold of 0.8–1.25, the data are suggestive for bioequivalence [
18]. Therefore, in individual patients, intake of 500 mg with a CB accompanied with PK monitoring could be considered when experiencing difficulties with a fasted intake.
During our study, patients could choose between seven different types of breakfast. Though standardized for the amount of fat, differences in total caloric intake between the breakfasts were present. This could have enhanced the inter-patient variability. However, our results demonstrate that the inter-patient variability in our study was comparable to other data and was not increased when AA was ingested with food [
16,
19]. It, therefore. is not expected that the differences in breakfast contribute to the inter-patient variability.
For several other drugs with this large intra- and inter-patient PK variability, the formulation was adjusted to increase drug absorption and drug exposure to achieve a better predictable response in patients. Examples of drugs with alternative formulations due to absorption issues are regorafenib and olaparib [
20,
21]. Also for AA, a new formulation is tested, using a continuous flow precipitation technology. This new formulation shows improved bioavailability, and therefore less PK variability. In addition, food does not play a significant role in the absorption of this new abiraterone formulation [
22]. We believe that this new formulation, which is not available yet, could help to overcome the issues of highly variable PK and thereby unpredictable treatment effect for patients.
In conclusion, a bioequivalent lower dose of abiraterone taken with food could not be established in our study. Though, information on the effect of food on abiraterone pharmacokinetics as presented in our study can be used.