Pharmacokinetic/pharmacodynamic drug–drug interactions of avatrombopag when coadministered with dual or selective CYP2C9 and CYP3A interacting drugs

Abstract

Aims

Avatrombopag, a thrombopoietin receptor agonist, is a substrate of cytochrome P450 (CYP) 2C9 and CYP3A. We assessed three drug–drug interactions of avatrombopag as a victim with dual or selective CYP2C9/3A inhibitors and inducers.

Methods

This was a three‐part, open‐label study. Forty‐eight healthy subjects received single 20 mg doses of avatrombopag alone or with one of 3 CYP2C9/3A inhibitors or inducers: fluconazole 400 mg once daily for 16 days, itraconazole 200 mg twice daily on Day 1 and 200 mg once daily on Days 2–16, or rifampicin 600 mg once daily for 16 days. Pharmacokinetics, pharmacodynamics (platelet count) and safety of avatrombopag were evaluated.

Results

Coadministration of a single 20‐mg dose of avatrombopag with fluconazole at steady‐state resulted in 2.16‐fold increase of AUC of avatrombopag, prolonged terminal elimination phase half‐life (from 19.7 h to 39.9 h) and led to a clinically significant increase in maximum platelet count (1.66‐fold). Itraconazole had a mild increase on both avatrombopag pharmacokinetics and pharmacodynamics compared to fluconazole. Coadministration of rifampicin caused a 0.5‐fold decrease in AUC and shortened terminal elimination phase half‐life (from 20.3 h to 9.84 h), but has no impact on maximum platelet count. Coadministration with interacting drugs was found to be generally safe and well‐tolerated.

Conclusions

The results from coadministration of fluconazole or itraconazole suggest that CYP2C9 plays a more predominant role in metabolic clearance of avatrombopag than CYP3A. To achieve comparable platelet count increases when avatrombopag is coadministered with CYP3A and CYP2C9 inhibitors, an adjustment in the dose or duration of treatment is recommended, while coadministration with strong inducers is not currently recommended.

What is Already Known about this Subject

• Avatrombopag is a thrombopoietin receptor agonist that increases platelet count in healthy subjects and patients with chronic liver disease undergoing elective invasive procedures.

• For the assessment of the worst‐case scenario for a dual substrate, regulatory guidance recommends investigation using a dual inhibitor/inducer approach.

What this Study Adds

• The results from this clinical study suggest that CYP2C9's contribution to the metabolism of avatrombopag predominates.

• When avatrombopag is coadministered with CYP3A and CYP2C9 inhibitors, dose/duration adjustment is recommended, while coadministration with strong inducers is not currently recommended.

Introduction

Avatrombopag (previously known as AKR‐501, YM‐477, YM‐301477 or E5501 monomaleate) is an orally administered, small molecule, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1722 agonist that is believed to act on the http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5063, activate the intracellular signalling system and promote production of platelets and megakaryocytes from haemopoietic precursor cells 1, 2. Avatrombopag is being investigated for its potential as an alternative to platelet transfusions in chronic liver disease patients with thrombocytopenia and scheduled to undergo a procedure. Avatrombopag has also been evaluated in patients with immune thrombocytopenic purpura.

A major route of excretion of avatrombopag and its metabolites is by the faecal route, accounting for approximately 88% of the administered dose. The primary metabolite of avatrombopag, 4‐hydroxy derivative, was detected only in faeces and accounted for 44% of administered dose. No metabolites of avatrombopag were detectable in plasma. Avatrombopag metabolism is mediated by cytochrome P450 (CYP) 3A4 and CYP2C9. The relative percentage contribution of the two major CYP enzymes toward the CYP dependent metabolism of avatrombopag was assessed using the relative activity factor in human liver microsomes and recombinant CYP enzyme 3. The data suggested that avatrombopag metabolism is mediated by CYP2C9 and CYP3A equally to form 4‐hydroxy metabolite (Appendix S1). These CYP enzymes seemed likely to contribute ≥25% of total clearance each based on in vitro data, and drug–drug interaction potential was predicted by physiologically based pharmacokinetic (PBPK) modelling and simulation (Appendix S2).

Avatrombopag showed linear pharmacokinetics (PK) and increases platelet count in a dose‐dependent manner across a range of 20 mg to 60 mg dose in healthy subjects 4. Following single dosing under fasted and fed conditions, mean peak concentrations occurred at 6–8 h and subsequently declined with a half‐life of 16–19 h in healthy subjects. Food intake did not alter the rate or extent of avatrombopag absorption, but substantially reduced PK variability relative to the fasted condition. Therefore, avatrombopag is recommended to be taken with a meal. Increases in platelet count are evident as early as 3–5 days after avatrombopag administration, and the highest changes in platelet count were observed by approximately 6–10 days 4. The safety and tolerability of avatrombopag were confirmed up to single doses of 80 mg in healthy subjects.

CYP2C9 and CYP3A‐mediated interactions were considered as the highest‐risk metabolism‐based interactions that might affect the exposure of avatrombopag according to in vitro metabolism data and PBPK simulations. To evaluate the maximum effect of metabolism‐based inhibition, fluconazole was selected as a moderate dual inhibitor of CYP2C9 and CYP3A. Fluconazole is the only reference CYP2C9 inhibitor recommended in the European Medicines Agency guideline 5, and among those noted in the Food and Drug Administration 6, 7, and Ministry of Health, Labour and Welfare draft guidance 8. Itraconazole was selected as a strong CYP3A inhibitor to evaluate the impact of potent CYP3A inhibition, and has been shown not to impact the CYP2C9 activity 9, 10. Although the inhibition potential of itraconazole is considered to be less than that of ketoconazole 11, itraconazole has been used in clinical drug–drug interaction studies as a typical inhibitor of CYP3A since ketoconazole was reported to cause serious safety concerns in liver toxicities 12. To evaluate the maximum effect of metabolism‐based induction, rifampicin was selected as a strong CYP3A inducer with moderate CYP2C9 inducing potency.

We have conducted a clinical drug interaction study with the above‐mentioned three drugs that are selective and dual CYP2C9‐ and CYP3A‐interacting in their effects to evaluate the impact of those drugs on PK and pharmacodynamics (PD) and safety of avatrombopag and provide appropriate guidance to the prescribing physicians on concomitant administrations of avatrombopag with drugs that are CYP2C9‐ and/or CYP3A‐interacting drugs.

Methods

This study was approved by the Institutional Review Board of Worldwide Clinical Trials Early Phase Services (IntegReview IRB, TX, USA), and conducted in accordance with the principles of the Declaration of Helsinki and Good Clinical Practice guidelines. All participants were given the explanation of the study, and written informed consent was obtained prior to screening.

Subjects

A total of 48 healthy, nonsmoking subjects, male and female aged 18–55 years with body mass index 18–28 kg/m² were enrolled. Subjects were of nonchildbearing potential or practicing highly effective contraception during the study. Medical history, vital signs, physical examination, 12‐lead electrocardiogram (ECG), serology (viral hepatitis B surface antigen, hepatitis C virus and anti‐human immunodeficiency antibody), and laboratory tests were evaluated prior to study entry. Subjects were excluded if: their platelet count was not between the lower limit of normal and 300 × 10⁹ l^–1 at the time of screening and at baseline of each period; they used any prescription drugs or St. John's Wort within 4 weeks before first dosing of avatrombopag; they took over‐the‐counter medications within 2 weeks before; or they took any food or beverages (e.g. alcohol, grapefruit juice) that might have affected the various drug metabolizing enzymes and transporters within 1 week before first dosing of avatrombopag.

Study design

This study was an open‐label drug–drug interaction study at Worldwide Clinical Trials Early Phase Services, LLC (San Antonio, TDX, USA; Clinical Trials.gov identifier NCT02809768). The study was divided into three parts. In Part A, the effects of steady‐state dosing of a moderate inhibitor of CYP2C9 and CYP3A (i.e. fluconazole) on single dose PK of avatrombopag were assessed. In Part B, the effects of steady‐state dosing of a strong CYP3A inhibitor (itraconazole) on the single dose PK of avatrombopag were assessed. In Part C, the effects of steady‐state dosing of a strong CYP3A and moderate CYP2C9 inducer (rifampicin) on the single dose PK of avatrombopag were assessed (Figure 1). There were 16 subjects enrolled in each study part, and each part was conducted in parallel. Avatrombopag was administrated under fed conditions (approximately 800–1000 calories, containing fat content of approximately 50% of total caloric content of the meal). Each dose of avatrombopag was given 30 min after the start of the meal.

Figure 1.

Study design. Ava = avatrombopag 20 mg dose, D = day. Part A: fluconazole 400 mg once daily on Days 1–16. Part B: itraconazole 200 mg twice daily on Day 1 and 200 mg once daily on Days 2–16. Part C: rifampicin 600 mg once daily on Days 1–16

Each part of the study consisted of two treatment periods: Period 1 (administration of a single oral dose of avatrombopag 20‐mg alone) and Period 2 (administration of oral doses of each inhibitor or inducer alone and concomitant administration of a single oral dose of avatrombopag 20‐mg with each inhibitor or inducer [fluconazole in Part A, itraconazole in Part B, or rifampicin in Part C]). Each period was separated by a washout interval of at least 28 days. Subjects whose platelet count exceeded 600 × 10⁹ l^–1 during the study were to be discontinued. Subjects whose platelet count exceeded 400 × 10⁹ l^–1 were to be administered low‐dose aspirin at the discretion of the investigator.

Part A:

Subjects were administered a single oral dose of avatrombopag 20 mg on Day 1 of Period 1. Subjects were administered fluconazole (Diflucan 200‐mg tablet; Pfizer, NY, USA) 400 mg once daily on Days 1 to 16 and a single dose of avatrombopag 20 mg on Day 7 in Period 2. Each dose of avatrombopag was administered 30 min after the start of the meal.

Part B:

Subjects were administered a single oral dose of avatrombopag 20 mg on Day 1 of Period 1. Subjects were administered itraconazole (Sporanox 100‐mg capsule; Janssen Pharma, NJ, USA) 200 mg twice daily on Day 1 and 200 mg once daily on Days 2 to 16 of Period 2. A single dose of avatrombopag 20 mg was administered on Day 7 of Period 2. Each dose of avatrombopag was administered 30 min after the start of the meal.

Part C:

Subjects were administered a single oral dose of avatrombopag 20 mg on Day 1 of Period 1. In Treatment Period 2, rifampicin (Rifadin 200‐mg capsule; Sanofi‐aventis U.S. LLC, Bridgewater NJ, USA) 600 mg was administered once daily on Days 1 to 16. To avoid a food effect on rifampicin absorption 13, each dose was administered 1 h before subjects consumed a meal. On Day 7 of Period 2, rifampicin 600 mg and avatrombopag 20 mg were administered 1 h before starting meal consumption and 30 min after starting meal consumption, respectively.

Serial blood samples for analysis of plasma concentration of avatrombopag were collected before dosing and at 1, 2, 3, 4, 5, 6, 7, 8, 12, 24, 48, 72, 96, 144 and 216 h after avatrombopag administration. Blood samples for measurement of platelet count were collected before dosing and on Days 3, 4, 5, 7, 10, 12, 14, 21 and 28 after avatrombopag administration. A blood sample in Period 2 of each part was also collected to confirm an adequate and expected concentration of inhibitor or inducer. Plasma was separated by centrifugation at 1500 × g for 10 min at 4°C. The plasma samples were stored in polypropylene tubes at approximately –20°C until analysis. The plasma concentrations of avatrombopag were quantified by using a validated liquid chromatography followed by tandem mass spectrometry (LC–MS/MS) at XenoBiotic Laboratories, Inc. (Plainsboro, NJ, USA) in human plasma with sodium heparin as anticoagulant. The method utilized a protein precipitation procedure to extract avatrombopag and deuterated‐internal standard (IS, YM‐385029; avatrombopag‐d10) from 100 μl of human plasma, a reversed‐phase high‐performance liquid chromatography C8 column, 5 μm, 50 × 2.0 mm column, with a mobile phase gradient using 10 mmol l^–1 ammonium formate in water, pH4 and methanol:acetonitrile (1:1) to separate the analyte from the matrix and an LC–MS/MS instrument with positive electrospray ionization multiple reaction monitoring mode to quantify the analyte. The mass spectrometer was operated in positive electrospray ionization mode and the resolution setting used was unit for both Q1 and Q3. The multiple reaction monitoring transition was m/z 649.2 → 267.1 for avatrombopag and m/z 659.4 → 267.1 for the IS. The interrun accuracy was <4.50%; interrun precision was <10.2% and intrarun accuracy was <6.50%; intrarun precision was <12.3% with all met the acceptance criteria within ±15%. At the lower limit of quantitation of 1.00 μg l^–1, the interrun accuracy was <–7.20%; interrun precision was <15.1%; intrarun accuracy was <–13.1%; intrarun precision was <14.8% with all met the acceptance criteria within ±20%. The calibration curve ranged from 1.00 μg l^–1 to 500 μg l^–1 for avatrombopag. Lack of bioanalytical interference between avatrombopag and the inhibitor or inducer (fluconazole, itraconazole and rifampicin) were documented prior to plasma sample analysis. The measurement of platelet count was performed at Worldwide Clinical Trials Early Phase Services (San Antonio, TX, USA).

PK and PD analysis

Noncompartmental PK and PD analysis were performed to derive PK parameters of avatrombopag in plasma and PD parameters from platelet count using Phoenix WinNonlin software (Version 6.4; Pharsight Corporation, Mountain View, CA, USA). The following PK parameters were derived: the maximum observed concentration (C_max); time at which the highest drug concentration occurs (t_max); area under the concentration‐time curve from 0 time extrapolated to infinite time (AUC_(0‐inf)); and terminal elimination phase half‐life (t_½).

Statistical analyses were performed using Phoenix WinNonlin and SAS (Version 9.3, SAS institute Inc., Cary, NC, USA) software. PK data were summarized using descriptive statistics. Log transformed C_max and AUC_(0‐inf) and untransformed t_½ were analysed using mixed effect models to yield the ratios (test/reference) of geometric least square means and 90% confidence intervals (CIs), and the difference between test and reference and 90% CIs, respectively. The mixed model included period as a fixed effects and subjects as a random effect. A nonparametric (Hodges–Lehmann) method was used to estimate the median difference between test and reference for t_max.

The following PD parameters were estimated: the maximum platelet count (E_max); observed time of maximum increase in platelet count (TE_max); and area under the effect curve for platelet count from 0 to 28 days after avatrombopag administration [AUEC_(0–28d)]. AUEC_(0–28d) was calculated by the area under or over the baseline value of platelet count. The statistical analysis for PD parameters was performed on untransformed data due to the presence of negative values using a mixed effect model with period as a fixed effect and subject as a random effect. Treatment means, difference of treatment means and 90% CIs were presented on the untransformed data.

Genotyping of CYP2C9 (*2, *3) and CYP3A5 (*3)

Genomic DNA from blood samples was isolated and purified using a solution‐based methodology followed by genotyping. The genotyping was conducted by Cancer Genetics Inc. (formerly Gentris; Morrisville, NC, USA). The CYP2C9 (*2, *3) was identified using a TaqMan gene expression system, and the CYP3A5*3 alleles was identified using sequencing method. Genotypes reported as *1/*1 refer only to the alleles tested. A subject reported as *1/*1 may actually carry a variant for which testing was not performed.

Safety assessments

All adverse events (AEs) observed during the study were collected. Monitoring of platelet count, clinical laboratory results (including haematology, serum chemistry, urinalysis, serology, pregnancy test and drug screen), and measurement of vital signs (diastolic and systolic blood pressure, pulse rate, respiration rate and temperature) were monitored physical examinations and 12‐lead ECGs were assessed throughout the study.

Sample size determination

Sample size estimates were based on an estimation approach rather than an approach to assess equivalence because PBPK simulation predicted 2‐fold and above of drug interaction (Appendix S2). Assuming a within‐subject standard deviation (SD) of 0.35 for the logarithmically transformed AUC_(0‐inf) of avatrombopag, as an example if the observed ratio of geometric means for test compared to reference is estimated to be 2.0, with 12 completed subjects, the 90% CI for this ratio would extend from 1.58 to 2.53. Allowing for a 25% drop‐out rate, a sample size of 16 subjects was required in each of Parts A, B and C to ensure 12 subjects complete each part the study.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 14, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 15.

Results

Among 48 subjects assigned to study, 42 received both treatments and completed the study. All subjects were aged between 19 and 51 years with body mass index between 18 and 32 kg/m². Thirty‐three (69%) subjects were male and 25 (52%) were white.

CYP2C9 and 3A5 alleles genotyping was conducted in all subjects. The CYP2C9 genotypes were distributed into one subject with *3/*3, nine subjects with *1/*2 or *1/*3, and 38 subjects with *1/*1. The CYP3A5*3 genotypes were distributed into 17 subjects with *3/*3, 23 subjects with *1/*3 and eight subjects with *1/*1.

Fluconazole; a dual moderate inhibitor of CYP2C9 and CYP3A (part A)

A total of 16 subjects were randomized and completed both treatment periods. There were 14 (87.5%) male subjects and two (12.5%) female subjects, with a mean age of 38.4 years. There were eight (50.0%) Black or African American subjects, six (37.5%) White, one (6.3%) Asian and one (6.3%) multiple race subject.

Avatrombopag plasma concentrations were elevated when avatrombopag was coadministered with fluconazole compared to avatrombopag alone (Figure 2). Coadministration of avatrombopag and fluconazole increased C_max and AUC of avatrombopag by 1.17‐fold and 2.16‐fold, respectively. Coadministration of fluconazole extended t_1/2 of avatrombopag by approximately 20 h (19.7 h in avatrombopag alone vs. 39.8 h in avatrombopag+fluconazole) but t_max was unaffected by fluconazole coadministration (Table 1).

Figure 2.

Mean (+ standard deviation) plasma concentration–time profiles of avatrombopag following administration of avatrombopag alone and concomitantly with fluconazole. n = 16

Table 1.

Summary of pharmacokinetic parameters of avatrombopag and statistical comparisons following administration of avatrombopag 20 mg with or without fluconazole, itraconazole and rifampicin

Treatment (Test:Reference)	Parameter	n		Arithmetic mean (SD)		LS Mean Ratio	90% CI
		Test	Reference	Test	Reference
Avatrombopag + fluconazole: avatrombopag	Cmax (μg l–1)	16	16	115 (34.5)	101 (37.1)	1.17	0.964, 1.42
	AUC(0‐inf) (μg h l–1)	15	16	6800 (2340)	3170 (1160)	2.16	1.71, 2.72
	t½ (h)	15	16	39.8 (5.12)	19.7 (2.44)	20.2a	18.4, 22.1a
	tmax (h)	16	16	7 (5, 24)	7 (4, 12)	ND	ND
Avatrombopag + itraconazole: avatrombopag	Cmax (μg l–1)	13	16	123 (55.0)	106 (31.2)	1.07	0.855, 1.35
	AUC(0‐inf) (μg h l–1)	13	16	4990 (3070)	3380 (1530)	1.37	1.10, 1.72
	t½ (h)	13	16	28.0 (9.77)	19.6 (3.25)	8.47a	5.25, 11.7a
	tmax (h)	13	16	6 (3, 12)	6 (4, 12)	ND	ND
Avatrombopag + rifampicin: avatrombopag	Cmax (μg l–1)	14	16	108 (33.5)	103 (37.9)	1.04	0.882, 1.23
	AUC(0‐inf) (μg h l–1)	13	16	1790 (776)	3340 (1760)	0.568	0.465, 0.623
	t½ (h)	13	16	9.77 (1.49)	20.3 (5.34)	–10.5a	–12.7, –8.21a
	tmax (h)	14	16	5 (2, 12)	6 (3, 8)	ND	ND

C_max, AUC_(0‐inf) and t_½ presented as arithmetic means (SD), t_max presented as median (range).

AUC_(0‐inf), area under the concentration‐time curve from 0 time extrapolated to infinite time; C_max, maximum observed concentration; CI, confidence interval; LS, least square; ND, not determined; SD, standard deviation; t_½, terminal elimination phase half‐life; t_max, time at which the highest drug concentration occurs.

^aLS mean difference (90% CI)

Consistent with the PK effect of fluconazole, the mean profile of platelet count over time was greater during coadministration with fluconazole compared with avatrombopag alone. Coadministration of avatrombopag and fluconazole increased E_max by 1.66‐fold (calculated as a ratio of mean value for coadministration of fluconazole and avatrombopag to mean for avatrombopag alone) and the mean difference in E_max between two treatments was 21.19 × 10⁹ l^–1. There was approximately a 1.47‐fold increase in AUEC_(0‐28d) with coadministration of avatrombopag and fluconazole (Table 2). Relative to the magnitude of PK interaction, coadministration with fluconazole resulted in a PD interaction to a lesser degree; however, the >20 × 10⁹ l^–1 mean difference in E_max with fluconazole coadministration is considered a clinically significant effect.

Table 2.

Summary of pharmacodynamic parameters and statistical comparisons following administration of avatrombopag 20 mg with or without fluconazole, itraconazole and rifampicin

Treatment (Test:Reference)	Parameter	n		Arithmetic mean (SD)		LS mean difference	90% CI
		Test	Reference	Test	Reference
Avatrombopag + fluconazole: avatrombopag	Emax (109 l–1)	16	16	307 (64.5)	285 (59.9)	21.19	2.89, 39.49
	TEmax (day)	16	16	10 (7, 12)	10 (7, 12)	ND	ND
	AUEC(0‐28d) (h·109 l–1)	16	16	23 300 (14100)	15 800 (7790)	7480	2400, 12 560
Avatrombopag + itraconazole: avatrombopag	Emax (109 l–1)	13	16	307 (84.1)	314 (73.2)	–5.92	–30.98, 19.14
	TEmax (day)	13	16	10 (3, 21)	10 (4, 28)	ND	ND
	AUEC(0‐28d) (h·109 l–1)	12	14	25 100 (18100)	18 100 (11800)	6658	1200, 12 116
Avatrombopag + rifampicin: avatrombopag	Emax (109 l–1)	14	16	307 (40.4)	317 (42.1)	–9.73	–19.50, 0.05
	TEmax (day)	14	16	7 (5, 12)	8.5 (5, 28)	ND	ND
	AUEC(0‐28d) (h·109 l–1)	14	15	3460 (10300)	15 300 (8600)	–12 001	–14 905, –9096

E_max and AUEC_(0‐28d) presented as arithmetic means (SD), TE_max presented as median (range). ND = not determined.

AUEC_(0‐28d), area under the effect curve for platelet count; CI, confidence interval; E_max, maximum platelet counts; LS, least square; SD, standard deviation; TE_max, observed time of maximum increase in platelet counts

Itraconazole, a strong CYP3A inhibitor (part B)

A total of 16 subjects were randomized and 14 (87.5%)completed Period 1 and 12 (75%) completed Period 2. One subject discontinued due to elevated platelet count (>300 × 10⁹ l^–1) at baseline of Period 2. There were 11 (68.8%) male subjects and five (31.3%) female subjects with a mean age of 32.5 years. There were 10 (62.5%) White subjects, 5 (31.3%) Black or African American subjects and one (6.3%) multiple race subject.

Mean plasma concentrations of avatrombopag during coadministration of itraconazole were slightly higher than after avatrombopag alone (Figure 3). Following coadministration of itraconazole, C_max and t_max of avatrombopag did not show apparent difference whereas AUC_(0–inf) increased by 1.37‐fold and t_½ extended by 8.47 h compared to avatrombopag alone (Table 1).

Figure 3.

Mean (+ standard deviation) plasma concentration–time profiles of avatrombopag following administration of avatrombopag alone and concomitantly with itraconazole. Avatrombopag alone: n = 16, avatrombopag + Itraconazole: n = 13

The mean profile of platelet count over time was similar between coadministration with itraconazole and avatrombopag alone. The mean difference in E_max for avatrombopag alone and for coadministration with itraconazole was not statistically significant (Table 2), indicating that itraconazole has no impact on the PD effect of avatrombopag.

Rifampicin; a strong CYP3A and moderate CYP2C9 dual inducer (part C)

A total of 16 subjects were randomized and 15 (93.8%)completed Period 1 and 14 (87.5%)completed Period 2. One subject discontinued due to elevated platelet count (>300 × 10⁹ l^–1) at Baseline of Period 2. There were eight (50.0%) male subjects and eight (50.0%) female subjects with a mean age of 29.8 years. There were nine (56.3%) White subjects, and seven (31.3%) Black or African American subjects.

After C_max was reached, mean plasma concentrations of avatrombopag during coadministration with rifampicin declined more rapidly than those of avatrombopag administered alone (Figure 4). Coadministration of rifampicin resulted in no apparent difference in C_max or t_max but did lead to an approximately 0.5‐fold decrease in AUC and a shortening of t_½ (20.3 h for avatrombopag alone vs. 9.77 h for avatrombopag+rifampicin) were observed (Table 1).

Figure 4.

Mean (+ standard deviation) plasma concentration–time profiles of avatrombopag following administration of avatrombopag alone and concomitantly with rifampicin. Avatrombopag alone: n = 16, avatrombopag + rifampicin: n = 14

The platelet count on coadministration with rifampicin returned to baseline level by approximately Day 14, i.e. 7 days after administration of avatrombopag with rifampicin, while the platelet count on administration avatrombopag alone returned to baseline by 27 days after administration of avatrombopag. Further to this faster return to baseline platelet levels on coadministration with rifampicin, there was approximately a 5‐fold reduction in AUEC_(0‐28d) with coadministration of avatrombopag and rifampicin compared to that of avatrombopag alone. However, the mean difference in E_max for avatrombopag alone and coadministration with rifampicin was not statistically significant (Table 2).

Effects of CYP2C9 (*2, *3) and CYP3A5*3

AUC_(0‐inf) following single dose of avatrombopag 20 mg alone on Day 1 of Period 1across three parts were plotted for each genotype (Figure 5). There appeared to be of no apparent impact of CYP2C9 or CYP3A5 genotype on the PK of avatrombopag after administration of avatrombopag alone.

Figure 5.

Plot of area under the curve extrapolated to infinity [AUC_(0‐inf)] of avatrombopag alone by CYP2C9 genotype (left) or CYP3A5 genotype (right)

Safety results

The incidence of AEs was similar between avatrombopag alone and coadministration. All treatment‐emergent AEs were mild or moderate in severity. There were no changes of clinical importance in mean vital signs and no clinically significant, abnormal ECG findings during the study.

Discussion

The overall purpose of this study was to investigate the effect of concomitant administration of avatrombopag with CYP2C9 and CYP3A interacting drugs on PK, PD and safety of avatrombopag. The study used fluconazole, a dual moderate inhibitor of CYP2C9 and CYP3A; itraconazole, a strong CYP3A only inhibitor; and rifampicin, a dual inducer having strong inducing potency of CYP3A and moderate potency of CYP2C9. In vitro data using recombinant CYP enzyme indicated that CYP2C9 and CYP3A contributed equally to the metabolism of avatrombopag; no other major enzymes were identified. The evaluations of the interaction with the three different interacting drugs in this study were expected not only to provide insights into the relative role of either CYP2C9 or CYP3A in metabolic clearance of avatrombopag, but also to be helpful in providing dosing recommendations on concomitant use of such interacting drugs with avatrombopag.

In accordance with guidance from European Medicines Agency, Food and Drug Administration, and Ministry of Health, Labour and Welfare for drug–drug interaction studies 5, 6, 7, 8, the study designs proposed for each of these interaction were standard procedures maximizing the extent of inhibition/induction of the interacting drugs in steady‐state conditions 16, 17, 18 and assessing those effects on a single dose profile of the avatrombopag. Since there is no standard CYP2C9‐specific strong inhibitor available, fluconazole was selected as a dual inhibitor of CYP2C9 and CYP3A to assess the effects on both enzymes simultaneously. Separately, the effect of a strong CYP3A inhibitor, itraconazole, was also assessed. It was considered that comparing the results obtained using fluconazole and itraconazole would allow us to estimate contribution of each CYP enzyme. Rifampicin was selected as the most suitable inducer to assess dual induction. Fluconazole is known to reach steady‐state by the sixth daily administration 16. Itraconazole 200 mg twice daily on the first day as loading dose, then 200 mg once daily from the next day was employed to bring itraconazole to steady‐state rapidly 17. Rifampicin 600 mg once daily has previously been shown to give maximum CYP3A induction on the 7^th day of dosing 18. Since absorption of rifampicin is reduced under fed condition, rifampicin was administered under fasted condition 13.

Results demonstrated a significant increase in systemic exposure to avatrombopag (AUC: 2.16‐fold) when coadministered with a dual inhibitor of CYP2C9 and CYP3A, fluconazole, compared to the C_max (1.17‐fold). Among PD parameters such as E_max and AUEC_(0‐28d), for clinical practice, the physicians refer to the maximum increase in platelet count, E_max as the determining factor for making decisions on whether the patients would require platelet transfusion 19, 20, 21. However, AUEC_(0‐28d) is also useful when considering the clinical relevance. More than 20 × 10⁹ l^–1 increase in E_max caused by fluconazole coadministration might become a trigger of the event of portal vein thrombosis and thrombophlebitis septic, and AUEC_(0‐28d) was significantly increased. Accordingly, dose adjustment of avatrombopag might be needed when coadministration with dual inhibitors of CYP3A and CY2C9. In contrast, a strong CYP3A inhibitor, itraconazole had milder effects on the avatrombopag PK and PD than those of fluconazole. Even though inhibition by itraconazole is known less than the extent of inhibition produced by ketoconazole 11, considering together with PBPK simulation results (Appendix S2), ketoconazole would also indicate milder effects on the avatrombopag than those of fluconazole. As itraconazole is known to have no impact on CYP2C9 activity 9, 10, the substantial impact of fluconazole suggests that CYP2C9 plays a more dominant role in metabolic clearance of avatrombopag than CYP3A. Coadministration with a dual inducer of CYP2C9 and CYP3A, rifampicin, demonstrated an approximately 0.5‐fold decrease in AUC without any effect on C_max. Avatrombopag t_½ was shortened by approximately 10 h when coadministered with rifampicin. Taken together with drug interactions seen with inhibitors, the results of coadministration with rifampicin is likely to be driven primarily by CYP2C9 induction. Regarding the impact of coadministration of rifampicin on the platelet count profile, there was approximately 5‐fold reduction in AUEC_(0‐28d) without any impact on E_max.. Even though E_max is more relevant in influencing the decisions on need of platelet transfusions for a patient requiring a surgical procedure, 5‐fold reduction in AUEC_(0‐28d) has significant clinical impact.

Given the primary role of CYP2C9 in the metabolic clearance of avatrombopag, further exploration of the relationship between CYP2C9 genotypes and avatrombopag PK was performed. Because of the very low prevalence of *3/*3 allele in western population for CYP2C9 22, only one subject in the current study was a homozygote of mutant alleles (*2 or *3), and hence a conclusive assessment of the potential relationship between CYP2C9 genetic polymorphism and avatrombopag PK could not be made. However, this subject showed highest exposure in this study, thus the possibility of an effect of CYP2C9 polymorphism on PK profile of avatrombopag cannot be excluded and further assessment is needed. For CYP3A5, there were also no apparent relationship between CYP3A5 genotype and avatrombopag PK.

Conclusion

The results of fluconazole and itraconazole coadministration suggest that CYP2C9 plays a more predominant role in metabolic clearance of avatrombopag than CYP3A, and the drug interaction observed with fluconazole and rifampicin is likely to be driven primarily by CYP2C9. Although administration of avatrombopag with concomitant use of CYP2C9 and CYP3A inhibitors and inducers was found to be generally safe and well‐tolerated, coadministration has an impact on the platelet response, and therefore a dose adjustment is recommended when avatrombopag is coadministered with CYP3A and CYP2C9 inhibitors. Coadministration with strong inducers is not currently recommended for chronic liver disease patients scheduled to undergo a procedure; however, a dose adjustment may be considered for other indications that dictate chronic administration (e.g. immune thrombocytopenic purpura).

Competing Interests

M.N., E.S., P.B., M.K.C., J.A., Y.S.A., W.G.L., S.Y. and J.F. are current employees of Eisai Co., Ltd and Eisai Inc., and may have stock and/or stock options. C.A.Z. does not have stock or stock options.

We thank all subjects, investigators, and all staff of Worldwide Clinical Trials for their contribution to the study. We thank David Critchley, a former employee of Eisai Ltd, who reviewed the draft manuscript. The work of this study was funded by PBM Pharmaceuticals Inc., Eisai Co., Ltd., Eisai Inc. and Eisai Ltd.

Contributors

M.N., E.S., B.R., P.B., J.A., S.Y. and J.F. participated in the study design, data analyses, interpretation of study results, and development of the manuscript. M.K.C. participated as a clinical study operation. C.A.Z. was involved in conduct of clinical study and the assessment of safety. PBPK analyses were performed by E.S. Y.A.S. and W.G.L. conducted the in vitro study. All authors revised the manuscript and approved the final version.

Supporting information

Appendix S1 In vitro study results

Appendix S2 Detail physiologically based pharmacokinetic modelling and simulation results

Nomoto, M. , Zamora, C. A. , Schuck, E. , Boyd, P. , Chang, M.‐K. , Aluri, J. , Siu, Y. A. , Lai, W. G. , Yasuda, S. , Ferry, J. , and Rege, B. (2018) Pharmacokinetic/pharmacodynamic drug–drug interactions of avatrombopag when coadministered with dual or selective CYP2C9 and CYP3A interacting drugs. Br J Clin Pharmacol, 84: 952–960. doi: 10.1111/bcp.13517.