Population Pharmacokinetics of Brexpiprazole in Japanese Healthy Subjects and Patients with Schizophrenia

doi:10.21203/rs.3.rs-5352712/v1

Download PDF

Research Article

Population Pharmacokinetics of Brexpiprazole in Japanese Healthy Subjects and Patients with Schizophrenia

https://doi.org/10.21203/rs.3.rs-5352712/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Brexpiprazole, widely approved for the treatment of schizophrenia, is an atypical antipsychotic that modulates serotonin–dopamine activity. To better understand the pharmacokinetics (PK) of brexpiprazole in Japanese patients, a population PK model was constructed and used to estimate steady state PK profiles and parameters as well as dopamine D₂/D₃ receptor occupancy profiles after repeated oral administrations of brexpiprazole at 1 and 2 mg/day. Nonlinear mixed effects modelling was used to analyse data from a total of 398 healthy subjects and patients with schizophrenia who received brexpiprazole in three Japanese clinical trials. The PK of brexpiprazole were well described by a two-compartment disposition model with transit absorption compartments. Estimated glomerular filtration rate, age and cytochrome P450 2D6 phenotype were identified as significant covariates on CL/F only. The model predicted that, at a dose of 2 mg/day, trough plasma concentration (90% prediction interval) of brexpiprazole is 77.1 (22.4–173) ng/mL and that dopamine D₂/D₃ receptor occupancy is > 80% over one day for most patients at steady state. This suggests the recommended maintenance dose of 2 mg/day of brexpiprazole leads to clinically useful dopamine D₂/D₃ receptor occupancy at steady state in Japanese patients.

Pharmacokinetics

schizophrenia

Japanese patients

dopamine D2/D3 receptor occupancy

dopamine D2 receptor partial agonist

antipsychotic

nonlinear mixed effects

Schizophrenia is a serious and often disabling psychiatric condition that imposes a substantial burden on patients, including worsening quality of life, financial disadvantage, and reduced life expectancy [1–4]. The disorder affects over 20 million people worldwide [2, 4], and the lifetime prevalence in Japanese patients is estimated to be 0.59% [1].

Antipsychotics are the cornerstone of treatment for patients diagnosed with schizophrenia [5]. Brexpiprazole (OPC-34712) is an atypical antipsychotic approved for the treatment of schizophrenia in over 60 countries, including the United States since 2015 and Japan since 2018 [6–8]. Brexpiprazole is a serotonin-dopamine activity modulator that acts as a partial agonist at serotonin 5-HT_1A and dopamine D₂ receptors, and as an antagonist at serotonin 5-HT_2A and α_1B/2C adrenergic receptors, all with similar subnanomolar potency [9].

The target dose range of brexpiprazole is 2–4 mg/day [10], which is typically achieved by uptitrating from 1 mg/day [8]. The effect of different doses on striatal dopamine D₂/D₃ receptor occupancy in the brain has previously been studied. In a positron emission tomography (PET) study of 15 healthy subjects, the mean dopamine D₂/D₃ receptor occupancy in the putamen and caudate nucleus increased with brexpiprazole dose [11]. At 4 hours post-dose, receptor occupancy levelled out at 77–88% with brexpiprazole 5 mg and 6 mg and was maintained at a similar level until 23.5 hours post-dose. The maximum obtainable receptor occupancy was 89.2% for the putamen and 95.4% for the caudate nucleus [11].

Two Phase 3 randomised, double-blind, placebo-controlled trials (ClinicalTrials.gov identifier: NCT01396421, NCT01393613) demonstrated that oral brexpiprazole significantly improved Positive and Negative Syndrome Scale (PANSS) total score over 6 weeks in patients with acute schizophrenia [12, 13]. Similarly, a 6-week Phase 2/3 trial specifically in Japanese patients with schizophrenia (NCT01451164) supported the efficacy of brexpiprazole at a dose of 2 mg/day with an improvement in PANSS score versus placebo (least squares mean difference − 7.32, p = 0.0124) [14]. A long-term, open-label study in Japanese patients (NCT01456897) also reported favourable results for brexpiprazole, including stable PANSS scores up to 52 weeks and a tolerable safety profile [15]. Additionally, a global 52-week maintenance study (NCT01668797) demonstrated an improvement in time to impending relapse with brexpiprazole versus placebo, suggesting that brexpiprazole is appropriate for maintenance treatment of schizophrenia [16].

Brexpiprazole and its metabolite, DM-3411, are metabolised by cytochrome P450 (CYP) enzymes, namely CYP2D6 and CYP3A4. Therefore, dose adjustments are warranted in patients who are known CYP2D6 poor metabolisers and those concomitantly taking CYP3A4 inhibitors, CYP2D6 inhibitors, or strong CYP3A4 inducers [8, 17].

The pharmacokinetics (PK) of brexpiprazole in Japanese patients was previously explored in a small (n = 21), 14-day, multiple-dose (1, 4, and 6 mg/day) study [17]. The study found that the maximum plasma concentration (C_max) and area under plasma concentration–time curve from 0–24 hours (AUC_24h) of brexpiprazole and DM-3411 were dose-proportional. Across all doses on Day 14, the median time to peak plasma concentration (t_max) of brexpiprazole was 4–5 hours, and the mean elimination half-life was 52–92 hours. The concentration of brexpiprazole at 24 hours post-dose reached steady state after Day 10 in all dose groups.

To better understand the PK of brexpiprazole in Japanese patients, the present study analysed data from healthy subjects and patients with schizophrenia collected in two Phase 1 trials and one Phase 2/3 trial. The objectives of the study were to: develop a population PK model to characterise brexpiprazole PK after single and multiple once-daily administration of oral tablets; evaluate and quantify possible covariate effects on brexpiprazole population PK parameters; and evaluate the adequacy of the final population PK model. We also reported the steady state PK profiles and parameters as well as dopamine D₂/D₃ receptor occupancy profiles based on the exposure-response relationship from a previous PET study [11], after repeated oral administration of brexpiprazole 1 and 2 mg/day and simulated using the final population PK model.

Clinical studies and population

The analysis used data from three prior clinical studies of brexpiprazole carried out in Japan, all of which were conducted in accordance with Good Clinical Practice guidelines and local regulatory requirements. The study protocols were approved by relevant institutional review boards, and all patients provided written informed consent prior to participation. The included trials were: (i) a randomised, placebo-controlled, single-dose, Phase 1 study of 46 healthy adult males who received brexpiprazole doses ranging from 0.2 to 6 mg (trial 331-07-002); (ii) a Phase 1 repeated-dose study of 21 adult patients with schizophrenia receiving multiple once-daily doses of 1, 4, and 6 mg brexpiprazole (trial 331-10-001); and (iii) a Phase 2/3 trial of 331 adult patients with schizophrenia who received multiple once-daily doses of 1, 2, and 4 mg brexpiprazole (trial 331-10-002; Japan Registry for Clinical Trials identifier: jRCT2080221591).

Dosing regimens and PK sampling schedules for the clinical studies included in the model are shown in Table S1. Blood samples for PK analysis (~ 5 mL per sample) were drawn by venous puncture, and the date and time were recorded in the case report form. Each blood tube was inverted and centrifuged, and the obtained plasma was frozen at approximately − 20°C (or − 80°C in trial 331-07-002). Frozen samples were sent to a bioanalytical laboratory where drug concentrations were measured using liquid chromatography–tandem mass spectrometry.

In all trials, subjects were included for population PK analyses if they provided at least one measurable post-dose brexpiprazole PK measurement; and all venous plasma concentrations of brexpiprazole with a valid date and time were included in the final PK dataset. Concentration measurements below the lower limit of quantification (LLOQ; 0.5 ng/mL) were excluded from the analyses.

Population PK model development and analysis

In order to identify and characterise the key model features needed to accurately describe the typical (i.e. population average) brexpiprazole exposure, PK data were analysed through nonlinear mixed effects modelling (NLME). The first-order conditional estimation method with interaction was used for all stages of the model development process. The precision of model parameter estimates, the magnitude of residual variability, and the appearance of trends in goodness-of-fit plots were all considered when assessing the suitability of structural NLME models. Unless otherwise noted, additional structural parameters were incorporated into the model only if the resulting improvement in the objective function was significant at the α = 0.01 level and did not result in any instability in model convergence.

In summary, a stepwise model development approach was applied, involving exploratory data analysis and data visualisation, development of a base PK model, stepwise covariate method (SCM) search, and evaluation of the final PK model.

Base structural model

Initial model development was conducted using Phase 1 data (trials 331-07-002 and 331-10-001) following a single dose. One-, two- and three-compartment structural models were evaluated. To further improve the description of the absorption phase, a series of transit compartments were evaluated to see whether they could improve the prediction. As a next step in the model building process, the best model describing the single-dose data was rerun on the multiple-dose Phase 1 data combined with the single-dose Phase 1 data. The best model identified for the Phase 1 data was then rerun on the combined data from the Phase 1 and Phase 2/3 studies.

The base model was parameterised in terms of apparent clearance (CL/F), apparent central volume of distribution (Vc/F), apparent inter-compartmental clearance (Q/F), apparent peripheral volume of distribution (Vp/F), and absorption rate constant (KA). Transit compartments were added in a stepwise manner until no further improvement in the model diagnostics was observed.

Stochastic model

Between subject variability (BSV) was incorporated into the NLME model through inclusion of random effects on model parameters in an exponential form. Then, the random effects model was optimised by evaluating the effect of including BSV on several structural parameters (i.e. CL/F, Vc/F, and KA). The residual error structure for plasma concentration data was assumed to follow a proportional error model.

Covariate selection

Once the structural/stochastic model was defined, potential covariate effects were tested to identify observable patient characteristics with a meaningful impact on exposure. The set of covariates evaluated for inclusion were selected based upon clinical relevance. Potential covariates were tested in an SCM search process, which included both forward inclusion and backward deletion steps, with a required significance level of p < 0.01 and p < 0.001, respectively. The covariates selected for use in the model are shown in Table 1.

Table 1

Covariates selected for evaluation in population PK analysis
Covariate	PK parameter
• Body weight • Age • Sex • Dose • Population (healthy subjects or patients with schizophrenia) • Kidney parameter (eGFR^a) • Liver parameter (ALT) • CYP2D6 inferred metabolic status (extensive, intermediate, or poor metabolisers) • CYP3A4/CYP2D6 inhibitor use (none, 1 strong/moderate CYP3A4/CYP2D6 comedication, or > 1 strong/moderate CYP3A4/CYP2D6 comedication) • Moderate CYP3A4/CYP2D6 inhibitor use (none, 1 moderate CYP3A4/CYP2D6 comedication, or > 1 moderate CYP3A4/CYP2D6 comedication) • Strong CYP3A4/CYP2D6 inhibitor use (none, 1 strong CYP3A4/CYP2D6 comedication, or > 1 strong CYP3A4/CYP2D6 comedication)	CL/F
• Body weight • Sex • Age	Vc/F
^aThe following equation was used to estimate glomerular filtration rate:
eGFR = 194 × serum creatinine^− 1.094 × age^− 0.287 (× 0.739 if female).
ALT, alanine aminotransferase; CL/F, apparent clearance; CYP, cytochrome P450; eGFR, estimated glomerular filtration rate; PK, pharmacokinetic; Vc/F, apparent central volume of distribution.

The magnitudes of the categorical covariates retained in the final model were reported as the percentage change in the relevant parameter relative to a typical subject. The impact of continuous covariates on the relevant parameter in the measured range of the covariate was explored.

Sensitivity analyses

Sensitivity analyses were performed to further confirm the validity of the final covariate model: (i) the final covariate model was rerun with the exclusion of outliers; (ii) an alternative model with allometrically scaled body weight on the population PK parameters CL/F, Q (exponent set to 0.75), Vc/F, and Vp/F (exponent set to 1.0) was run [18]; (iii) SCM searches with other measures of body size (lean body weight and body mass index [BMI]) were performed; and (iv) the final covariate model with three separate categories defined for CYP2D6 metabolisers (extensive, intermediate, and unknown) was rerun.

Model evaluation

The following four approaches were applied to perform an evaluation of the final PK model: (i) standard goodness-of-fit plots; (ii) bootstrap resampling including a total of 1000 bootstrapped replications of the original data; (iii) predictioncorrected visual predictive checks; and (iv) assessment of ETA (η) and EPSILON (ε) shrinkage estimates for all variability terms.

Simulations using the final PK model

The final model was used to simulate steady state PK profiles for repeated oral administrations of 1 and 2 mg/day of brexpiprazole. One thousand virtual patients were generated by resampling from the PK modelling dataset. Additionally, exposure-response relationship data from a PET study [11] were used to simulate dopamine receptor occupancy using the following formula:

Receptor occupancy = E_max × C / (EC₅₀ + C)

where E_max = estimated maximum receptor occupancy in the caudate nucleus, 95.4%; EC₅₀ = estimated concentration of brexpiprazole to half maximum receptor occupancy in the caudate nucleus, 7.75 ng/mL; and C = simulated plasma concentration of brexpiprazole.

Statistical software and analysis

NONMEM software (version 7.2.0, ICON plc) was used for NLME modelling. Statistical and graphical interpretations of the NONMEM output were performed by PSN 4.2.0 R, version 3.1 and higher, and/or S-PLUS 8.2.

Data availability and patient characteristics

The final dataset, formed from all available data from the three included studies, contained 2,374 evaluable PK samples from 398 subjects. A total of 160 (6.7%) PK samples were below the LLOQ and were, therefore, excluded from the analysis. Summary statistics for covariates included in the population PK analysis using the final dataset, including demographic characteristics, are summarised in Table 2. The total analysis sample was 55.3% male, with a median age of 42 years and a median body weight of 61.8 kg. Half of the subjects (50.8%) were extensive CYP2D6 metabolisers, and the remainder were intermediate CYP2D6 metabolisers (17.1%) or had missing/inconclusive data (32.2%).

Table 2

Summary statistics of covariates included in population PK analysis (final dataset, n = 398)
Continuous variables
Covariate	Mean (SD)	Min	Median		Max
Age (years)	42 (13.1)	18.0	42.0		64.0
Body weight (kg)	62.1 (12.9)	31.9	61.8		127.0
Lean body weight (kg)	45.4 (10.2)	22.1	46.2		77.3
BMI (kg/m²)	22.9 (4.1)	12.6	22.2		39.7
eGFR (mL/min)	90.0 (17.8)	51.7	88.4		207.0
ALT (U/L)	22.1 (13.3)	7 .0	18 .0		93.0
Categorical variables
Covariate	Category			n		%
Sex	Male			220		55.3
Sex	Female			178		44.7
Inferred CYP2D6 metabolic status	Poor			0		0
	Intermediate			68		17.1
	Extensive			202		50.8
	Ultra rapid			0		0
	Missing/inconclusive			128		32.2
Population	Healthy subjects			46		11.6
Population	Patients with schizophrenia			352		88.4
Concomitant CYP2D6/CYP3A4 moderate/strong inhibitor use	None			380		95.5
	One			17		4.3
	More than one			1		0.3
Concomitant CYP2D6/CYP3A4 moderate inhibitor use	None			382		96.0
	One			16		4.0
	More than one			0		0
Concomitant CYP2D6/CYP3A4 strong inhibitor use	None			395		99.2
	One			3		0.8
	More than one			0		0
ALT, alanine aminotransferase; BMI, body mass index; CYP, cytochrome P450; eGFR, estimated glomerular filtration rate; PK, pharmacokinetic; SD, standard deviation.

Initial exploration of data

Output from graphical exploration of the available data is shown in Figure S1. Approximately 4-fold accumulation in exposure following multiple daily dosing was observed, and PK profiles between healthy subjects and patients across Day 1 appeared comparable (Figure S1A). Higher exposure of brexpiprazole was observed in the group of intermediate CYP2D6 metabolisers compared with extensive and unknown metabolisers, which showed a similar level of exposure (Figure S1B). Analysis of trough samples of brexpiprazole after multiple daily dosing showed that the steady state was achieved by the final day of dosing for most of the subjects in Phase 1 studies (Figure S1C). When compared with plasma concentrations measured in Phase 2/3 patients, a trend towards lower exposure was observed for the Phase 1 patients under the assumption of steady state by Day 14 (Figure S1D).

Population PK model

Model development

Initially, a one-compartment model with first-order absorption was fitted to the Phase 1 data, but since the graphical analysis showed bi-exponential disposition, a second compartment was added, resulting in a significant decrease in the objective function (OBJF). The addition of a third compartment resulted in model instability, leading to the rejection of a three-compartment model. Thus, the final base model was a two-compartment population PK model that described the absorption with first-order transfer processes (Fig. 1) and a dose dependency on absorption transit rate constant (KTR) described using a power function (Figure S2).

The model was rerun on combined Phase 1 and 2/3 study data, and bias was observed between the Phase 1 and Phase 2/3 data for BSV on CL/F and Vc/F (Figure S3). A bioavailability scaling factor (F1) on relative oral bioavailability described the observed difference in exposure between Phase 1 (F1 = ~ 0.6) and Phase 2/3 (F1 = 1). BSV was identified on CL/F, Vc/F, and KTR with an OMEGA block between BSV terms on CL/F and Vc/F. A proportional residual error was used to describe the residual variability.

The covariate effects of estimated glomerular filtration rate (eGFR), age, and CYP2D6 inferred metabolic status (intermediate versus extensive/unknown metabolisers) on CL/F were retained in the final population PK model; a univariate–covariate analysis step and a forward inclusion–backward elimination procedure determined that these covariates had a statistically significant effect on CL/F. Subjects with missing inferred CYP2D6 metaboliser status were grouped with the CYP2D6 extensive metabolisers for covariate analysis as their CL/F did not appear to differ. Of note, a trend towards lower CL/F and Vc/F was observed for subjects with strong CYP3A4/CYP2D6 inhibitor use (n = 3); however, this effect was not retained in the model during the forward inclusion step of the covariate analysis.

The final population PK parameter estimates are shown in Table 3. In the final population PK model, the typical values for CL/F, Q/F, Vc/F, and Vp/F were estimated to be 0.74 L/h (intermediate metabolisers), 1.0 L/h, 63.0 L, and 24.6 L, respectively. The typical value for CL/F for the extensive and unknown metabolisers was calculated to be 1.15 L/h. The scaling factor, F1, was considered relevant in order to get an unbiased model towards the Phase 2/3 patients and was, therefore, retained in the final model. Including this factor led to an estimated 38%, on average, smaller oral bioavailability (F) in Phase 1 compared with Phase 2/3, which was included to describe the structurally lower observed concentrations in Phase 1.

Table 3

Parameter estimates and point estimates from the bootstrap for the final brexpiprazole population PK model
Parameter description	Population estimate (%SE)	Inter-individual variability (%SE)	Bootstrap final population PK model median (5th, 95th percentiles)
Apparent clearance intermediate metaboliser, CL/F (L/h)	0.74 (6.4%)	52.8% (7.9%)	0.73 (0.638, 0.838)
Apparent clearance extensive metaboliser, CL/F (L/h)	1.15^a	-	-
Apparent volume of distribution in central compartment, Vc/F (L)	63.0 (5.1%)	40.1% (11.7%)	62.6 (56.8, 71.7)
Absorption rate constant, KTR (h-1)	1.9 (5.1%)	48.8% (9.9%)	1.89 (1.71, 2.12)
Apparent inter-compartmental clearance, Q/F (L/h)	1.0 (17.8%)	-	0.989 (0.701, 1.47)
Apparent volume of distribution in peripheral compartment, Vp/F (L)	24.6 (10.7%)	-	24.7 (19.8, 30.4)
Power function on KTR	−0.228 (20.0%)	-	−0.228 (-0.337, − 0.133)
Phase 1 scaling factor (versus Phase 2/3)	0.62 (6.1%)	-	0.61 (0.548, 0.709)
Effect of extensive CYP2D6 inferred metabolic status on CL/F	0.56 (17.9%)	-	0.565 (0.361, 0.802)
Effect of age on CL/F	−0.30 (26.1%)	-	−0.293 (-0.46, − 0.133)
Effect of eGFR on CL/F	0.4 (42.8%)	-	0.374 (0.0527, 0.781)
Correlation coefficient CL/F x Vc/F	0.776 (10.7%)	-	-
Proportional residual error (%)	18.5 (9.6%)	-	0.0329 (0.0237, 0.0491)
^aValue calculated based on the model parameter estimates [(1 + 0.56) × 0.74 L/h]
ALT, alanine aminotransferase; CL/F, apparent clearance; CYP, cytochrome P450; eGFR, estimated glomerular filtration rate; KTR, absorption transit rate constant; PK, pharmacokinetic; Q/F, apparent inter-compartmental clearance; SE, standard error; Vc/F, apparent central volume of distribution; Vp/F, apparent peripheral volume of distribution.

Magnitude of covariate effect

Age was found to be one of the statistically significant covariates for CL/F, with CL/F tending to be lower in older patients (Table S2). At the 5th (22 years) and 95th (63 years) percentiles of age, CL/F was 21% higher and 11.3% lower, respectively, as compared with a 42-year-old reference patient.

eGFR was also a statistically significant covariate for CL/F, with a higher expected CL/F at higher eGFR values on average (Table S2). Compared with a reference subject with an eGFR of 88.4 mL/minute, subjects at the 5th percentile (65.5 mL/minute) and 95th percentile (119.7 mL/minute) had a 11.4% lower and 13.0% higher CL/F, respectively.

In terms of CYP2D6 inferred status (Table S3), intermediate metabolisers had a 35.9% lower CL/F and a 56% higher expected AUC when compared with extensive/unknown metabolisers.

Sensitivity analysis

No relevant changes in the estimates of the relevant PK parameters were observed when the model was rerun with the exclusion of outliers (note that two PK samples were marked as outliers). None of the evaluated allometric models showed a statistically significant drop in the OBJF (> 6.63) when compared with the final PK model. SCM searches with alternative measures of body size (lean body weight and BMI) were not found to have a statistically significant effect on CL/F and/or Vc/F. Rerunning the model with each of the categories defined for CYP2D6 metabolisers showed that CL/F values estimated with the final covariate model (intermediate and extensive/unknown metabolisers: 0.74 and 1.15 L/h, respectively) were very similar to those estimated with the model with three separate categories of CYP2D6 metabolisers (intermediate, extensive, and unknown metabolisers: 0.74, 1.16, and 1.14 L/h, respectively).

Model evaluation

Model evaluation techniques indicated that the identified final population PK model had no significant biases and represented an adequate fit to the data. Goodness-of-fit plots are shown in Figures S4 and S5. The plots revealed that individual predictions fit well along the identity line, and the conditionally weighted residuals were relatively low and well distributed along the zero line relative to population predictions and time. The median values of all parameters from bootstrapping were consistent with the original population PK estimates (Table 3). Prediction-corrected visual predictive checks showed good agreement between the simulated and observed data (Fig. 2). The shrinkage of BSV for CL/F, Vc/F, and KTR were estimated at 2.1%, 18.9%, and 50.0%, respectively (Figure S6).

Simulations using the final PK model

Simulated steady state PK profiles for repeated oral administrations of 1 and 2 mg/day of brexpiprazole in Japanese patients with schizophrenia are shown in Fig. 3. PK parameters derived from simulated PK profiles are shown in Table 4. Trough plasma concentrations (90% prediction interval) of brexpiprazole at steady state were 38.5 (10.5–87.0) and 77.1 (22.4–173) ng/mL at doses of 1 and 2 mg/day, respectively.

Table 4

Simulated PK parameters at steady state after repeated administrations of 1 and 2 mg/day of brexpiprazole based on final population PK model
Parameters	AUC_0-24,ss	C_max,ss	C_trough,ss	C_trough,ss^a
	(ng·hr/mL)	(ng/mL)	(ng/mL)	(ng/mL)
1 mg/day	1084 (377–2308)	66.9 (23.9–141)	38.5 (10.5–87.0)	45.6 (33.9)
2 mg/day	2170 (726–4591)	134 (46.3–272)	77.1 (22.4–173)	78.7 (47.6)
Shown as mean with 90% prediction intervals.
^aObserved trough concentrations from Phase 2/3 study, shown as mean with standard deviation. Trough concentration is defined as concentrations with time after last dose within 20 to 28 hours. N = 71 for 1 mg/day, N = 79 for 2 mg/day.
AUC_{0 − 24,ss}, area under the curve from 0 to 24 hours after last dose at steady state; C_max,ss, maximum plasma concentration of brexpiprazole at steady state; C_trough,ss, trough plasma concentration of brexpiprazole at steady state; PK, pharmacokinetic.

Simulated dopamine D₂/D₃ receptor occupancy profiles for repeated oral administrations of 1 and 2 mg/day of brexpiprazole in Japanese patients with schizophrenia are shown in Fig. 4. The simulation suggests that the maintenance dose (i.e. 2 mg/day of brexpiprazole) would achieve > 80% of receptor occupancy over a day for most patients at steady state.

A robust two-compartment population PK model was developed based on the results from two Japanese Phase 1 studies and one Japanese Phase 2/3 study. This model was successfully used to simulate clinically relevant steady state PK profiles for repeated oral administrations of brexpiprazole in Japanese patients with schizophrenia. It predicted that, at a target dose of 2 mg/day, trough plasma concentration (90% prediction interval) of brexpiprazole is 77.1 (22.4–173) ng/mL, and dopamine D₂/D₃ receptor occupancy is > 80% over a day for most patients at steady state.

The covariate effects of age, eGFR, and CYP2D6 inferred metabolic status on population PK model parameters, in particular CL/F, were identified and quantified. Although they were retained as covariates, age and eGFR were not considered clinically relevant due to their small magnitude of effects on AUC. Inclusion of age and CYP2D6 inferred metabolic status only slightly reduced the unexplained random variability on CL/F by less than 3%. Including a covariate effect of eGFR on CL/F increased the unexplained variability on this parameter, a phenomenon that has been reported previously [19]. The eGFR covariate effect was kept in the final model as the magnitude and direction of its effect on CL/F were plausible from a clinical perspective.

A scaling factor on F was implemented and retained in the final model to ensure that it was unbiased towards the Phase 2/3 patients. Including this factor led to an estimated 38% smaller F, on average, in Phase 1 compared with Phase 2/3. Although assuming a difference between healthy subjects and patients also led to a reduction in OBJF, this did not take away the observed bias between patients in Phase 1 and patients in Phase 2/3. However, assuming a ‘phase’ effect removed the observed bias, this approach was, therefore, preferred. The reason for this ‘phase’ difference is unknown but may be a result of low subject numbers in the Phase 1 study and thus considered a coincidence. The model with a scaling factor was deemed adequate to be used in the covariate analysis.

Population PK analysis was separately reported in a non-Japanese population [6]. This analysis supported the dose setting for adolescents with schizophrenia. The reported model includes body weight as a covariate on clearance and volume of distribution to account for the impact of the growth of body size on PK of brexpiprazole. While this analysis was conducted using data from different age distribution, the reported typical value of CL/F and Vc/F for body weight = 70 kg are similar with our estimates in Japanese population.

Existing literature suggests that the therapeutic effects of most antipsychotics are elicited when striatal dopamine D₂/D₃ receptor occupancy rate is at 65–80%; a rate below this range would result in a reduced likelihood of response, and exceeding this range would lead to an increased risk of unwanted extrapyramidal side effects [20, 21]. However, dopamine D₂ receptor partial agonists such as brexpiprazole, aripiprazole, and cariprazine have been reported to show therapeutic effects when the dopamine D₂/D₃ receptor occupancy rate is over 80–90%, without the same risk of extrapyramidal effects [21, 22]. For most patients receiving 2 mg/day of brexpiprazole, the final model produced by our study predicted > 80% of dopamine D₂/D₃ receptor occupancy over a day at steady state. This supports the use of 2 mg/day, the recommended maintenance dose of brexpiprazole in Japanese patients with schizophrenia [10], to achieve a therapeutic effect.

The expected AUC for CYP2D6 intermediate metabolisers was 56% higher than that for extensive/unknown metabolisers. This suggests that dose adjustment is not necessary for CYP2D6 intermediate metabolisers as the increase in AUC is less than two-fold. It should be noted that brexpiprazole dosing instructions only recommend dose reduction for patients who are poor CYP2D6 metabolisers, taking either strong CYP2D6 or CYP3A4 inhibitors, or taking a combination of strong/moderate CYP2D6 inhibitors with strong/moderate CYP3A4 inhibitors [8]. The findings of the present study are consistent with the existing knowledge about the PK of brexpiprazole in the Japanese population in terms of CYP2D6 metabolism. Notably, in a small Phase 1 PK study by Ishigooka et al. [17], brexpiprazole showed dose proportionality, and dose-normalised C_max and AUC_24h of brexpiprazole after 14 days of treatment were higher in intermediate CYP2D6 metabolisers than extensive CYP2D6 metabolisers, demonstrating the role of CYP2D6 function in the elimination of brexpiprazole.

A population PK model was successfully developed and used to describe the effects of brexpiprazole on Japanese patients with schizophrenia. The simulations carried out using the model suggest that a 2 mg/day dose of brexpiprazole leads to appropriate dopamine D₂/D₃ receptor occupancy at steady state in this patient population.

Acknowledgments

The authors would like to thank all the participants who provided data for this analysis. Support for the initial population pharmacokinetic analysis, excluding the simulation, was provided by LAP&P Consultants BV and funded by Otsuka. Medical writing support was provided by Highfield Communication and funded by Otsuka in accordance with Good Publication Practice (GPP) guidelines (Available at: https://www.acpjournals.org/doi/epdf/10.7326/M22-1460).

Declarations of interest

All authors are full-time employees of Otsuka Pharmaceutical Co., Ltd.

Funding

This analysis and the included clinical trials (331-07-002, 331-10-001, and 331-10-002 [clinical trial identifier: jRCT2080221591]) were supported by Otsuka Pharmaceutical Co., Ltd.

Author contributions

Higashi K: Conceptualisation, Methodology, Formal analysis and interpretation of data; Writing – review and editing; Sasaki T: Conceptualisation, Methodology, Formal analysis and interpretation of data; Writing – review and editing; Aoki K: Formal analysis and interpretation of data; Writing – review and editing; Sekine D: Formal analysis and interpretation of data; Writing – review and editing; Maeda K: Formal analysis and interpretation of data; Writing – review and editing; Shiomi Y: Formal analysis and interpretation of data; Writing – review and editing; Kawai Y: Formal analysis and interpretation of data; Writing – review and editing.

Baba K et al (2022) Burden of schizophrenia among Japanese patients: a cross-sectional National Health and Wellness Survey. BMC Psychiatry 22(1):410. 10.1186/s12888-022-04044-5
Charlson FJ et al (2018) Global Epidemiology and Burden of Schizophrenia: Findings From the Global Burden of Disease Study 2016. Schizophr Bull 44(6):1195–1203. 10.1093/schbul/sby058
He H et al (2020) Trends in the incidence and DALYs of schizophrenia at the global, regional and national levels: results from the Global Burden of Disease Study 2017. Epidemiol Psychiatr Sci 29:e91. 10.1017/s2045796019000891
Solmi M et al (2023) Incidence, prevalence, and global burden of schizophrenia - data, with critical appraisal, from the Global Burden of Disease (GBD) 2019. Mol Psychiatry 28(12):5319–5327. 10.1038/s41380-023-02138-4
Owen MJ, Sawa A, Mortensen PB (2016) Schizophrenia Lancet 388(10039):86–97. 10.1016/s0140-6736(15)01121-6
Wang Y et al (2023) Population Pharmacokinetic Analysis of Brexpiprazole to Support its Indication and Dose Selection in Adolescents With Schizophrenia. J Clin Pharmacol 63(11):1290–1299. https://doi.org/10.1002/jcph.2307
Otsuka Pharmaceutical Co., L. Otsuka Receives Approval in Japan for the Manufacture and Sale of New Antipsychotic Drug Rexulti® Tablets for Schizophrenia (2018) ; https://www.otsuka.co.jp/en/company/newsreleases/2018/20180119_1.html. [accessed October 2024]
Otsuka Pharmaceutical Co., L. REXULTI® (brexpiprazole) Prescribing Information (2024) ; https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/205422s011lblcorrection.pdf. [accessed October 2024]
Maeda K et al (2014) Brexpiprazole I: in vitro and in vivo characterization of a novel serotonin-dopamine activity modulator. J Pharmacol Exp Ther 350(3):589–604. 10.1124/jpet.114.213793
Leucht S et al (2020) Dose-Response Meta-Analysis of Antipsychotic Drugs for Acute Schizophrenia. Am J Psychiatry 177(4):342–353. 10.1176/appi.ajp.2019.19010034
Wong DF et al (2021) An open-label, positron emission tomography study of the striatal D(2)/D(3) receptor occupancy and pharmacokinetics of single-dose oral brexpiprazole in healthy participants. Eur J Clin Pharmacol 77(5):717–725. 10.1007/s00228-020-03021-9
Correll CU et al (2015) Efficacy and Safety of Brexpiprazole for the Treatment of Acute Schizophrenia: A 6-Week Randomized, Double-Blind, Placebo-Controlled Trial. Am J Psychiatry 172(9):870–880. 10.1176/appi.ajp.2015.14101275
Kane JM et al (2015) A multicenter, randomized, double-blind, controlled phase 3 trial of fixed-dose brexpiprazole for the treatment of adults with acute schizophrenia. Schizophr Res 164(1–3):127–135. 10.1016/j.schres.2015.01.038
Ishigooka J, Iwashita S, Tadori Y (2018) Efficacy and safety of brexpiprazole for the treatment of acute schizophrenia in Japan: A 6-week, randomized, double-blind, placebo-controlled study. Psychiatry Clin Neurosci 72(9):692–700. 10.1111/pcn.12682
Ishigooka J, Iwashita S, Tadori Y (2018) Long-term safety and effectiveness of brexpiprazole in Japanese patients with schizophrenia: A 52-week, open-label study. Psychiatry Clin Neurosci 72(6):445–453. https://doi.org/10.1111/pcn.12654
Fleischhacker WW et al (2017) Efficacy and Safety of Brexpiprazole (OPC-34712) as Maintenance Treatment in Adults with Schizophrenia: a Randomized, Double-Blind, Placebo-Controlled Study. Int J Neuropsychopharmacol 20(1):11–21. 10.1093/ijnp/pyw076
Ishigooka J et al (2018) Pharmacokinetics and Safety of Brexpiprazole Following Multiple-Dose Administration to Japanese Patients With Schizophrenia. J Clin Pharmacol 58(1):74–80. https://doi.org/10.1002/jcph.979
West GB, Brown JH, Enquist BJ (1997) A general model for the origin of allometric scaling laws in biology. Science 276(5309):122–126. 10.1126/science.276.5309.122
Lagishetty CV, Vajjah P, Duffull SB (2012) A reduction in between subject variability is not mandatory for selecting a new covariate. J Pharmacokinet Pharmacodyn 39(4):383–392. 10.1007/s10928-012-9256-2
Uchida H et al (2011) Dopamine D2 receptor occupancy and clinical effects: a systematic review and pooled analysis. J Clin Psychopharmacol 31(4):497–502. 10.1097/JCP.0b013e3182214aad
Hart XM, Schmitz CN, Gründer G (2022) Molecular Imaging of Dopamine Partial Agonists in Humans: Implications for Clinical Practice. Front Psychiatry 13:832209. 10.3389/fpsyt.2022.832209
Mamo D et al (2007) Differential effects of aripiprazole on D(2), 5-HT(2), and 5-HT(1A) receptor occupancy in patients with schizophrenia: a triple tracer PET study. Am J Psychiatry 164(9):1411–1417. 10.1176/appi.ajp.2007.06091479

The authors declare potential competing interests as follows: All authors are full-time employees of Otsuka Pharmaceutical Co., Ltd.

BrexpiprazolepopPKmanuscriptsupplementaryinformation.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Population Pharmacokinetics of Brexpiprazole in Japanese Healthy Subjects and Patients with Schizophrenia

Status:

Version 1

Abstract

Figures

Introduction

Methods

Clinical studies and population

Population PK model development and analysis

Base structural model

Stochastic model

Covariate selection

Sensitivity analyses

Model evaluation

Simulations using the final PK model

Receptor occupancy = E_max × C / (EC₅₀ + C)

Statistical software and analysis

Results

Data availability and patient characteristics

Initial exploration of data

Population PK model

Model development

Magnitude of covariate effect

Sensitivity analysis

Model evaluation

Simulations using the final PK model

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

Population Pharmacokinetics of Brexpiprazole in Japanese Healthy Subjects and Patients with Schizophrenia

Status:

Version 1

Abstract

Figures

Introduction

Methods

Clinical studies and population

Population PK model development and analysis

Base structural model

Stochastic model

Covariate selection

Sensitivity analyses

Model evaluation

Simulations using the final PK model

Receptor occupancy = Emax × C / (EC50 + C)

Statistical software and analysis

Results

Data availability and patient characteristics

Initial exploration of data

Population PK model

Model development

Magnitude of covariate effect

Sensitivity analysis

Model evaluation

Simulations using the final PK model

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

Receptor occupancy = E_max × C / (EC₅₀ + C)