First-order Conditional Estimation Method Research Articles

An interactive dose optimizer based on population pharmacokinetic study to guide dosing of methotrexate in Chinese patients with osteosarcoma.

Osteosarcoma is a rare tumor with an incidence of 4.4 cases per million per year in adolescent. High-dose methotrexate (HD-MTX) is the standard first-line chemotherapeutic agent for osteosarcoma. However, its efficacy can vary significantly among individuals due to wide pharmacokinetic variability. Despite this, only a few population pharmacokinetics (popPK) models based on Chinese patients with osteosarcoma have been reported. Thus, this study aimed to develop a HD-MTX popPK model and an individual model-based dose optimizer for osteosarcoma therapy. A total of 680 MTX serum concentrations from 57 patients with osteosarcoma were measured at the end of MTX infusion and 10h, 24h, 48h, and 72h after the start of infusion. Using the first-order conditional estimation method with NONMEM, a popPK model was estimated. Goodness-of-fit plots, visual predictive checks, and bootstrap analysis were generated to evaluate the final model. A dose optimizer tool was developed based on the validated models using R Shiny. Additionally, clinical data from 12 patients with newly diagnosed osteosarcoma were collected and used as the validation set to preliminarily verify the predictive ability of the popPK model and the dose optimizer tool. Body surface area (BSA) was the most significant covariate for compartment distribution. Creatinine clearance (CrCL) and co-administration of NSAIDs were introduced as predictors for central compartmental and peripheral compartmental clearance, respectively. Co-administration of NSAIDs was associated with significantly higher MTX concentrations at 72h (p = 0.019). The dose optimizer tool exhibited a high consistency in predicting MTX AUC compared to the actual AUC (r = 0.821, p < 0.001) in the validation set. The dose optimizer tool could be used to estimate individual PK parameters, and optimize personalized MTX therapy in particular patients.

Population Pharmacokinetics of Sertraline in Psychiatric and Substance Use Disorders.

This study aimed to characterize the population pharmacokinetics of sertraline in Mexican patients with psychiatric and substance use disorders. Fifty-nine patients (13 to 76 years old) treated with doses of sertraline between 12.5 and 100 mg/day were included. Plasma sertraline concentrations were determined in blood samples and five of the main substances of abuse were determined by rapid tests in urine samples. Demographic, clinical, and pharmacogenetic factors were also evaluated. Population pharmacokinetic analysis was performed using NONMEM software with first-order conditional estimation method. A one-compartment model with proportional residual error adequately described the sertraline concentrations versus time. CYP2D6*2 polymorphism and CYP2C19 phenotypes significantly influenced sertraline clearance, which had a population mean value of 66 L/h in the final model. The absorption constant and volume of distribution were fixed at 0.855 1/h and 20.2 L/kg, respectively. The model explained 11.3% of the interindividual variability in sertraline clearance. The presence of the CYP2D6*2 polymorphism caused a 23.1% decrease in sertraline clearance, whereas patients with intermediate and poor phenotype of CYP2C19 showed 19.06% and 48.26% decreases in sertraline clearance, respectively. The model was internally validated by bootstrap and visual predictive check. Finally, stochastic simulations were performed to propose dosing regimens to achieve therapeutic levels that contribute to improving treatment response.

Population Pharmacokinetic Modeling of Posaconazole in Japanese Patients Receiving Fungal Prophylaxis.

Posaconazole is a vital drug to treat and prevent invasive fungal infections. Several factors, such as sex, body weight, total serum proteins, dietary intake, and severe mucositis, affect posaconazole pharmacokinetics (PKs). However, the relevance of other factors that affect the PKs of posaconazole in hematopoietic stem cell transplantation (HSCT) is unknown. This study explored factors influencing the PKs of posaconazole in HSCT recipients and nontransplant patients with hematological diseases. The authors conducted a single-institution, retrospective study. Forty-two Japanese inpatients receiving oral posaconazole tablets as prophylaxis for fungal infections were enrolled in this study. A one-compartment model with first-order absorption was used as the structural pharmacokinetic model. A population PK (PopPK) analysis was performed using a nonlinear mixed-effects modeling program, using a first-order conditional estimation method with interactions. Perl-speaks-NONMEM and R were used to evaluate the goodness of fit and visualize the output. In 29% of the enrolled patients, the serum concentration of posaconazole was <0.5 mcg/mL, considered the effective range. PopPK analysis revealed that the patient had undergone HSCT within 1 year, diarrhea occurred more than 5 times a day, and aspartate aminotransferase were covariates that influenced apparent clearance (CL/F). The CL/F of posaconazole was 1.43-fold higher after HSCT and 1.26-fold higher during diarrhea. PopPK analysis revealed that HSCT, diarrhea, and aspartate aminotransferase were factors associated with the CL/F of posaconazole. The trough concentration of posaconazole may be below the therapeutic range in a few patients with diarrhea and/or after HSCT. As invasive fungal infections in patients with hematologic diseases can be life-threatening, therapeutic drug monitoring of posaconazole is strongly recommended, and patients should be carefully monitored.

Development of extended pharmacokinetic models for propofol based on measured blood and brain concentrations

Propofol’s pharmacokinetics have been extensively studied using human blood samples and applied to target-controlled infusion systems; however, information on its concentration in the brain remains scarce. Therefore, this study aimed to simultaneously measure propofol plasma and brain concentrations in patients who underwent awake craniotomy and establish new pharmacokinetic model. Fifty-seven patients with brain tumors or brain lesions who underwent awake craniotomy were sequentially assigned to model-building and validating groups. Plasma and brain (lobectomy or uncapping margins) samples were collected at five time-points. The concentration of propofol was measured using high-performance liquid chromatography. Population pharmacokinetic analysis was conducted through a nonlinear mixed-effects modeling program using a first-order conditional estimation method with interactions. Propofol’s brain concentrations were higher than its plasma concentrations. The measured brain concentrations were higher than the effect site concentrations using the previous models. Extended models were constructed based on measured concentrations by incorporating the brain/plasma partition coefficient (Kp value). Extended models showed good predictive accuracy for brain concentrations in the validating group. The Kp value functioned as a factor explaining retention in the brain. Our new pharmacokinetic models and Kp value can predict propofol’s brain and plasma concentrations, contributing to safer and more stable anesthesia.

Understanding hemoglobin contribution to high-dose methotrexate disposition-population pharmacokinetics in pediatric patients with hematological malignancies.

The aim of the present study was to develop a population pharmacokinetic model for methotrexate (MTX) during high-dose treatment (HDMTX) in pediatric patients with acute lymphoblastic leukemia (ALL) and non-Hodgkin's lymphoma (NHL) and to describe the influence of variability factors. The study included 50 patients of both sexes (aged 1-18years) who received 3 or 5g/m2 of HDMTX. A nonlinear mixed effect modeling approach was applied for data analysis. Parameter estimation was performed by first-order conditional estimation method with interaction (FOCEI), whereas stepwise covariate modeling was used to assess variability factors. The final model is a two-compartment model that incorporates the effect of body surface area and the influence of hemoglobin and serum creatinine on MTX clearance (CL). Population pharmacokinetic values for a typical subject were estimated at 5.75 L/h/m2 for clearance (CL), 21.3 L/m2 for volume of the central compartment (V1), 8.2 L/m2 for volume of the peripheral compartment (V2), and 0.087 L/h/m2 for intercompartmental clearance (Q). According to the final model, MTX CL decreases with increasing serum creatinine, whereas a positive effect was captured for hemoglobin. A difference of almost 32% in MTX CL was observed among patients' hemoglobin values reported in the study. The developed population pharmacokinetic model can contribute to the therapy optimization during HDMTX in pediatric patients with ALL and NHL. In addition to renal function and body weight, it describes the influence of hemoglobin on CL, allowing better understanding of its contribution to the disposition of HDMTX.

Effects of Letermovir and/or Methylprednisolone Coadministration on Voriconazole Pharmacokinetics in Hematopoietic Stem Cell Transplantation: A Population Pharmacokinetic Study.

ObjectiveThe aim of this study was to identify factors affecting blood concentrations of voriconazole following letermovir coadministration using population pharmacokinetic (PPK) analysis in allogeneic hematopoietic stem cell transplant (allo-HSCT) recipients.MethodsThe following data were retrospectively collected: voriconazole trough levels, patient characteristics, concomitant drugs, and laboratory information. PPK analysis was performed with NONMEM® version 7.4.3, using the first-order conditional estimation method with interaction. We collected data on plasma voriconazole steady-state trough concentrations at 216 timepoints for 47 patients. A nonlinear pharmacokinetic model with the Michaelis–Menten equation was applied to describe the relationship between steady-state trough concentration and daily maintenance dose of voriconazole. After stepwise covariate modeling, the final model was evaluated using a goodness-of-fit plot, case deletion diagnostics, and bootstrap methods.ResultsThe maximum elimination rate (Vmax) of voriconazole in patients coadministered letermovir and methylprednisolone was 1.72 and 1.30 times larger than that in patients not coadministered these drugs, respectively, resulting in decreased voriconazole trough concentrations. The developed PPK model adequately described the voriconazole trough concentration profiles in allo-HSCT recipients. Simulations clearly showed that increased daily doses of voriconazole were required to achieve an optimal trough voriconazole concentration (1–5 mg/L) when patients received voriconazole with letermovir and/or methylprednisolone.ConclusionsThe development of individualized dose adjustment is critical to achieve optimal voriconazole concentration, especially among allo-HSCT recipients receiving concomitant letermovir and/or methylprednisolone.Supplementary InformationThe online version contains supplementary material available at 10.1007/s40268-021-00365-0.

Nonlinear Mixed Effects Modeling of Deterministic and Stochastic Dynamical Systems in Wolfram Mathematica

Nonlinear mixed effects (NLME) modeling is a powerful tool to analyze time-series data from several individual entities in an experiment. In this paper, we give a brief overview of a package for NLME modeling in Wolfram Mathematica entitled NLMEModeling, implementing the first-order conditional estimation method with sensitivity equation-based gradients for parameter estimation. NLMEModeling supports mixed effects modeling of dynamical systems where the underlying dynamics are described by either ordinary or stochastic differential equations combined with observation equations with flexible observation error models. Moreover, NLMEModeling is a user-friendly package with functionality for parameter estimation, model diagnostics (such as goodness-of-fit analysis and visual predictive checks), and model simulation. The package is freely available and provides an extensible add-on to Wolfram Mathematica.

Population pharmacokinetics of azathioprine active metabolite in patients with inflammatory bowel disease and dosage regimens optimisation.

Azathioprine is a first-line drug used to maintain the remission of inflammatory bowel disease (IBD). As a prodrug, azathioprine is metabolised to produce active 6-thioguanine nucleotides (6-TGN). There are large individual variations in the pharmacokinetics/pharmacodynamics of 6-TGN in patients with IBD. Here, we aimed to develop a model to quantitatively investigate factors that affect 6-TGN pharmacokinetics to formulate a dosage guideline for azathioprine. Data were collected prospectively from 100 adult patients with IBD who were receiving azathioprine. Patients were genotyped for two single-nucleotide polymorphisms (TPMT*3C c.719A>G and NUDT15 c.415C>T). Using high-performance liquid chromatography, we measured 156 steady-state trough concentrations of 6-TGN within the range 0.09 to 1.16mg/L (ie 133-1733pmol per 8×108 RBC). The covariates analysed included sex, age, body-weight, laboratory tests and concomitant medications. A population pharmacokinetic model was established using "non-linear mixed-effects modelling" software and the "first-order conditional estimation method with interaction." Body-weight, TPMT*3C polymorphisms and co-therapy with mesalazine were found to be important factors influencing the clearance of 6-TGN. A dosage guideline for azathioprine was developed based on the PPK model that enables individualised azathioprine dosing in adult patients with different body-weights, TPMT*3C genotypes and co-administration with mesalazine.

Population pharmacokinetics of high‐dose methotrexate in Chinese pediatric patients with medulloblastoma

Methotrexate (MTX) pharmacokinetics has substantial inter-individual variability and toxicity. In children with medulloblastoma treated with high-dose methotrexate (HD-MTX), the pharmacokinetic properties of methotrexate have not been established. A total of 660 serum samples from 105 pediatric patients with medulloblastoma were included in a population pharmacokinetic (PPK) analysis of methotrexate by using the nonlinear mixed-effects modeling method. The basic one-compartment population pharmacokinetic model was established by NONMEM software and the first-order conditional estimation (FOCE) method, and the final covariate model was obtained by the stepwise regression method. Weight (WT), creatinine clearance (CrCL), and whether the treatment was combined with dexamethasone (DEX) were covariates that had significant effects on the clearance rate (CL) of the model. The pharmacokinetic equation of CL in the final covariate model was as follows: CLi = 9.23× (1 + 0.0005× (θCrCL -105.78)) × (1 + 0.0017× (θWT -16)) × eηcl,i (L/h), IF (θDEX ) CLi = 1.19× CLi (L/h). The estimation accuracy of all pharmacokinetic parameters were acceptable (relative standard error < 14.74%). The goodness-of-fit diagram and bootstrap tests indicated that the final PPK model was stable with acceptable predictive ability. The PPK model may be useful for determining personalized medication levels in pediatric medulloblastoma patients undergoing HD-MTX therapy.

Population pharmacokinetics of cyclosporine A in Chinese patients with nephrotic syndrome in individualized drug administration

To establish a population pharmacokinetic (PopPK) model of cyclosporine A (CsA) in Chinese patients with nephrotic syndrome (NS) and to use the model to guide the adjustment of individualized dosage regimens. 216 CsA therapeutic drug monitoring (TDM) concentration observations were collected from 127 Chinese patients with NS. The basic model was developed as a one-compartment PK model with first-order absorption and linear elimination. The first-order conditional estimation (FOCE) method was applied to establish the final model with covariates using NONMEM software. The final model was evaluated through internal validation including goodness-of-fit analysis and bootstrap method as well as external validation using 39 additional PK observations from 35 patients with NS. A PopPK model of CsA was established in Chinese NS patients with influence of body weight on clearance. The internal and external validation results showed that the final model was stable. The established population model adequately characterized the PK of CsA in Chinese patients and could support individualized medication during treatment of NS with CsA.

Population Pharmacokinetic Analysis of the MDM2 Inhibitor KRT-232 (formerly AMG 232) in Subjects with Advanced Solid Tumors, Multiple Myeloma or Acute Myeloid Leukemia

Introduction KRT-232 is a potent and selective, targeted small molecule inhibitor of human mouse double minute 2 (MDM2) homolog interactions with tumor protein 53 (p53). MDM2 prevents p53 activation and reduces p53-mediated transcription and cell cycle control. KRT-232 is under development by Kartos Therapeutics for treatment of myelofibrosis, polycythemia vera, acute myeloid leukemia (AML) and Merkel cell carcinoma (see NCT03662126, NCT03669965, NCT03787602). Study 20120106 was a 2-part Phase 1 dose-exploration/dose-expansion study of KRT-232 in patients with advanced solid tumors or multiple myeloma (Gluck et al. Invest New Drugs in press; NCT01723020). Study 20120234 was a Phase 1b study evaluating KRT-232 alone and combined with trametinib, in patients with relapsed/refractory AML (Erba et al. Blood Adv 2019; NCT02016729). This population PK analysis quantified the KRT-232 PK time-course and variability, and the contribution of subject covariates to PK variability, using data from Amgen studies 20120106 and 20120234. Methods Study 20120106: Subjects received KRT-232 doses of 15 mg (n=3), 30 mg (n=3), 60 mg (n=4), 120 mg (n=7), 240 mg (n=76), 300 mg (n=4), 360 mg (n=4) and 480 mg (n=6). Doses were administered once daily (QD) for 7 days in 21-day cycles. PK sampling occurred on Cycle 1 days 1 and 7, at nominal sampling times pre-dose and 1, 3, 5, 7, and 24 h post-dose, and 72 h after the day 7 dose. Study 20120234: Subjects received KRT-232 doses of 60 mg (n=14; n=10 co-administered with 2 mg trametinib once daily, n=4 as single agent), 180 mg (n=5), 240 mg (n=3), and 360 mg (n=10). KRT-232 was administered QD for 7 days in 14-day cycles. PK sampling occurred on days 1 and 7, at nominal sampling times pre-dose and 1, 2, 4, 6, and 24 h post-dose, and 72 h after the day 7 dose. Population PK modeling was conducted using the first-order conditional estimation (FOCE-I) method in NONMEM® 7.3. Model covariates were selected using a forward addition and backward elimination method, based on significance levels of p&lt;0.05 and p&lt;0.01, respectively. Tested covariates included age, sex, weight, race, ethnicity, creatinine clearance, total bilirubin, AST, ALT, albumin, tumor type, disease stage, ECOG performance status, and the presence/absence of trametinib. Model quality was checked by inspection of model parameters and their confidence intervals, and standard residual-based and simulation-based diagnostics. Results KRT-232 plasma concentrations from 141 subjects with 1783 samples could be described by a two-compartment model with first-order absorption. Apparent oral clearance (CL/F) of KRT-232 was estimated to be 24.9 L/h (CV 61.5%) in subjects with solid tumors. The apparent oral central volume (Vc/F) and peripheral volume (Vp/F) were 62.9 L and 333 L, respectively. The terminal half-life was 17.1 hr. The prediction-corrected visual predictive check in Figure A suggested that the median, 5th and 95th percentiles of the measured plasma KRT-232 concentrations aligned with modeled concentrations for 24 h post-dose; the modeled variability was larger than the measured variability at 72 h post-dose. Residual diagnostics confirmed an adequate model fit. Apparent oral clearance did not change with dose over the studied dose range of 15-480 mg QD, indicating AUC increases were dose-proportional (Figure B). Relative to solid tumor subjects, AML subjects had 61.6% greater steady-state AUC (Figure C). A subject with decreased albumin at the 5th percentile (30 g/L) was modeled to have a 47.7% increase in AUC and an 87.2% increase in Cmin at steady state, relative to a typical subject with a median albumin level of 39 g/L. Trametinib and multiple myeloma model covariates had confidence intervals overlapping the line of unity. Covariates weight, sex, age, race and ethnicity did not affect CL/F or Vc/F. No effects of renal parameters, hepatic parameters, or ECOG performance status on CL/F were detected. Conclusion A two-compartment linear PK model with first-order absorption adequately described KRT-232 PK. Apparent oral clearance did not change with dose, indicating AUC increases were dose-proportional. AML subjects and subjects with decreased albumin had greater steady-state AUC values relative to subjects with solid tumors and normal albumin levels, although the magnitude of these effects were less than 2-fold. Disclosures Ma: Certara Strategic Consulting: Consultancy, Employment. Wada:Certara Strategic Consulting: Consultancy, Employment. Allard:Certara Strategic Consulting: Consultancy, Employment. Slatter:Kartos Therapeutics: Employment, Equity Ownership. OffLabel Disclosure: KRT-232 (formerly AMG 232) is an investigational small molecule MDM2 inhibitor.

Population Pharmacokinetic Analysis of Phenytoin After Intravenous Administration of Fosphenytoin in Adult and Elderly Epileptic Patients.

Fosphenytoin, the diphosphate ester salt of phenytoin, is widely used to treat status epilepticus. The aim of this study was to develop a population pharmacokinetic (PPK) model to describe serum phenytoin concentrations after the intravenous administration of fosphenytoin in adult and elderly epileptic patients. Patient backgrounds, laboratory tests, and prescribed drugs were retrospectively collected from electronic medical records. Patients who received fosphenytoin were enrolled. The PPK analysis was performed using NONMEM 7.3.0 with the first-order conditional estimation method with interaction. Age, sex, laboratory tests, and coadministered drugs were selected as candidates for covariates. Significance levels for forward inclusion and backward elimination were set at 0.05 and 0.01, respectively. The study protocol was approved by the Fukuoka Tokushukai Ethics Committee. A total of 340 serum phenytoin concentrations from 200 patients treated with fosphenytoin were available. The median age and body weight of the population were 71 years and 53.4 kg, respectively. A linear 1-compartment model with the conversion rate of fosphenytoin to phenytoin clearly described the pharmacokinetics of phenytoin after the intravenous administration of fosphenytoin. Age was detected as a covariate of clearance (CL): CL (L/h) = 1.99 × (body weight/53.4) × (age/71). Goodness-of-fit plots revealed the high-predictive performance of the final PPK model, and systematic deviations were not observed. The final model was validated by a prediction-corrected visual predictive check and bootstrap analysis. We herein developed a PPK model to describe phenytoin concentrations after the intravenous administration of fosphenytoin. Age was identified as a significant covariate for CL.

Model-Based Conditional Weighted Residuals Analysis for Structural Model Assessment

Nonlinear mixed effects models are widely used to describe longitudinal data to improve the efficiency of drug development process or increase the understanding of the studied disease. In such settings, the appropriateness of the modeling assumptions is critical in order to draw correct conclusions and must be carefully assessed for any substantial violations. Here, we propose a new method for structure model assessment, based on assessment of bias in conditional weighted residuals (CWRES). We illustrate this method by assessing prediction bias in two integrated models for glucose homeostasis, the integrated glucose-insulin (IGI) model, and the integrated minimal model (IMM). One dataset was simulated from each model then analyzed with the two models. CWRES outputted from each model fitting were modeled to capture systematic trends in CWRES as well as the magnitude of structural model misspecifications in terms of difference in objective function values (ΔOFVBias). The estimates of CWRES bias were used to calculate the corresponding bias in conditional predictions by the inversion of first-order conditional estimation method’s covariance equation. Time, glucose, and insulin concentration predictions were the investigated independent variables. The new method identified correctly the bias in glucose sub-model of the integrated minimal model (IMM), when this bias occurred, and calculated the absolute and proportional magnitude of the resulting bias. CWRES bias versus the independent variables agreed well with the true trends of misspecification. This method is fast easily automated diagnostic tool for model development/evaluation process, and it is already implemented as part of the Perl-speaks-NONMEM software.

Exact Gradients Improve Parameter Estimation in Nonlinear Mixed Effects Models with Stochastic Dynamics

Nonlinear mixed effects (NLME) modeling based on stochastic differential equations (SDEs) have evolved into a promising approach for analysis of PK/PD data. SDE-NLME models go beyond the realm of standard population modeling as they consider stochastic dynamics, thereby introducing a probabilistic perspective on the state variables. This article presents a summary of the main contributions to SDE-NLME models found in the literature. The aims of this work were to develop an exact gradient version of the first-order conditional estimation (FOCE) method for SDE-NLME models and to investigate whether it enabled faster estimation and better gradient precision/accuracy compared to the use of gradients approximated by finite differences. A simulation-estimation study was set up whereby finite difference approximations of the gradients of each level were interchanged with the exact gradients. Following previous work, the uncertainty of the state variables was accounted for using the extended Kalman filter (EKF). The exact gradient FOCE method was implemented in Mathematica 11 and evaluated on SDE versions of three common PK/PD models. When finite difference gradients were replaced by exact gradients at both FOCE levels, relative runtimes improved between 6- and 32-fold, depending on model complexity. Additionally, gradient precision/accuracy was significantly better in the exact gradient case. We conclude that parameter estimation using FOCE with exact gradients can successfully be applied to SDE-NLME models.

Prediction of Methionine and Homocysteine levels in Zucker diabetic fatty (ZDF) rats as a T2DM animal model after consumption of a Methionine-rich diet

BackgroundAlthough alterations in the methionine metabolism cycle (MMC) have been associated with vascular complications of diabetes, there have not been consistent results about the levels of methionine and homocysteine in type 2 diabetes mellitus (T2DM). The aim of the current study was to predict changes in plasma methionine and homocysteine concentrations after simulated consumption of methionine-rich foods, following the development of a mathematical model for MMC in Zucker Diabetic Fatty (ZDF) rats, as a representative T2DM animal model.MethodThe model building and simulation were performed using NONMEM® (ver. 7.3.0) assisted by Perl-Speaks-NONMEM (PsN, ver. 4.3.0). Model parameters were derived using first-order conditional estimation method with interactions permitted among the parameters (FOCE-INTER). NCA was conducted using Phoenix (ver. 6.4.0). For all tests, we considered a P-value < 0.05 to reflect statistical significance.ResultsOur model featured seven compartments that considered all parts of the cycle by applying non-linear mixed effects model. Conversion of S-adenosyl-L-homocysteine (SAH) to homocysteine increased and the metabolism of homocysteine was reduced under diabetic conditions, and consequently homocysteine accumulated in the elimination phase.Using our model, we performed simulations to compare the changes in plasma methionine and homocysteine concentrations between ZDF and normal rats, by multiple administrations of the methionine-rich diet of 1 mmol/kg, daily for 60 days. The levels of methionine and homocysteine were elevated approximately two- and three-fold, respectively, in ZDF rats, while there were no changes observed in the normal control rats.ConclusionThese results can be interpreted to mean that both methionine and homocysteine will accumulate in patients with T2DM, who regularly consume high-methionine foods.

Influence of genetic co-factors on the population pharmacokinetic model for clopidogrel and its active thiol metabolite

PurposeA high interindividual variability is observed in the pharmacokinetics of clopidogrel, a widely used antiplatelet drug. In the present study, a joint parent-metabolite population pharmacokinetic model was developed to adequately describe observed concentrations of clopidogrel and its active thiol metabolite (H4).MethodsThe study included 63 patients undergoing elective coronarography or percutaneous coronary intervention. The population pharmacokinetic model was developed in the NONMEM 7.3 software, and first-order conditional estimation method with interaction was applied. Also, the influence of covariates was evaluated (age, weight, body mass index (BMI), obesity defined as BMI ≥ 30 kg/m2, sex, diabetes mellitus, co-administration of PPI or statins, presence of CYP2C19*2, CYP2C19*17, CYP3A4*1G alleles, and ABCB1 3435 TT genotype).ResultsIt was found that the only significant covariate was the presence of CYP2C19*2 allele, which had an impact on lower conversion of clopidogrel to H4. As a result, predicted area under the time-concentration curve values was lower in carriers of this allele, with median 5.94 ng h/ml (interquartile range 3.92–12.51 [ng∙h/ml]) vs. 12.70 ng h/ml in non-carriers (interquartile range, 7.00–19.39 [ng∙h/ml]), respectively (p = 0.004).ConclusionsDeveloped model predicts that the only significant covariate influencing the observed concentrations and therefore the exposure to the active H4 metabolite is the presence of CYP2C19*2 allele.

Population Pharmacokinetics Of Propranolol in Patients with Liver Cirrhosis

Since propranolol, a nonselective β-blocker, is known to a high extraction drug and mainly metabolized in the liver after oral administration, the pharmacokinetics may be markedly affected by the hepatic function. The aim of this study is to develop a population pharmacokinetic model and evaluate the effects of covariates including hepatic impairment. Single oral dose of 40mg propranolol was administered and plasma concentrations of propranolol and metabolite concentrations were measured over 8h after dosing. The population pharmacokinetics was analyzed using NONMEM for 468 plasma concentrations of propranolol and metabolite from 42 noncirrhotic and 36 cirrhotic subjects. Volume of distribution of parent drug (Vp/F) was set equal to volume of distribution of metabolite. The parameters were estimated using the first-order conditional estimation method with interaction. Covariates were tested with exponential relationship by stepwise covariate model building. The final model was a one-compartment for both parent and metabolite model with two depot compartments: the dose enters not only into the parent compartment at a rate Ka with a fraction F1(0.25 fix) but also into the metabolite compartment at a rate Kam with a fraction F2(0.315 fix). Final parameter estimates (%inter-individual variability, IIV) were as follows; Vp/F = 14.8 L (187.8%), clearance of propranolol (CL) = 44.8 L/hr (130.4%), clearance of 4-OH propranolol (CLm) = 1060 L/hr (75.6%), and the metabolic rate constant of propranolol to 4-OH propranolol (Km) = 0.215 hr-1 (47.6%). The platelet count showed an influence on propranolol PK parameters; Vp/F, CL, CLm and Km. The visual predictive check and the fits of a bootstrap with 1000 replicates of the data set demonstrated the adequateness and robustness of the developed model. A population PK model reflecting the first-pass effect was successfully developed and we evaluated the effects of covariates on the pharmacokinetic parameters.

Population Pharmacokinetic Modeling of Armodafinil and Its Major Metabolites.

Population pharmacokinetic models for armodafinil and its major metabolites, R-modafinil acid and modafinil sulfone, were developed, and selected covariates were investigated. Data from 583healthy subjects and patients with bipolar I disorder in 11 phase 1-3 studies (8027 concentrations) of armodafinil, given as single or multiple once-daily doses (50- to 400-mg tablet or capsule), were pooled. A previously developed 1-compartment model with first-order absorption without covariate effects was initially applied to pooled phase 1 and 2 data. After covariate analysis, the phase 3 data were pooled with the phase 1 and 2 data set and the model was refined again using a second backward elimination step. Population modeling was performed with NONMEM version 7 with the first-order conditional estimation method. Estimated armodafinil apparent oral clearance (CL/F), volume of distribution (Vc/F), and absorption t½ were 2.01L/h, 45L, and 0.226hours, respectively. Armodafinil CL/F and Vc/F increased with weight; predicted steady-state area under the curve was 16.4% higher and 29.1% lower in a patient weighing 50 or 150 kg, respectively, relative to a 70-kg patient. Female participants had 10.2% lower armodafinil Vc/F compared with male participants. Age, race (white vs nonwhite), health status (healthy vs bipolar I disorder), liver function, and renal function were not statistically significant predictors of armodafinil pharmacokinetics. CL/F and Vc/F for R-modafinil acid and modafinil sulfone were 16.7 L/h and 8.95 L and 6.82 L/h and 12.4 L, respectively. Weight did not affect exposure of either metabolite. These population pharmacokinetic models were from the largest population of adults reported to date and provide a robust characterization of the pharmacokinetics of armodafinil, R-modafinil acid, and modafinil sulfone in adults.

Abstract B19: Population pharmacokinetic-pharmacodynamic (PKPD) modeling of ibrutinib in patients with B-cell malignancies.

Abstract Ibrutinib (IBRU) is an oral Bruton's tyrosine kinase (BTK) inhibitor, approved by US FDA for the treatment of chronic lymphocytic leukemia (CLL/SLL) and mantle cell lymphoma (MCL) patients having received at least one prior therapy. A nonlinear mixed-effects population model was developed to describe the PK of IBRU in patients with B-Cell malignancies and to establish the effect of pathophysiological covariates on its PK behavior. The relationship between PK and BTK engagement in peripheral blood mononuclear cells (PBMC) was also explored. IBRU PK data (3477 observations in 245 patients) were available in patients with MCL, CLL/SLL and recurrent B-cell malignancies at dose levels from 1.25 to 12.5 mg/kg and at fixed doses from 420 to 840 mg once daily. An additional phase 2 study in 119 patients with MCL (772 observations) treated at 560 mg once daily was used to validate the PK model. BTK occupancy was assessed (694 observations in 127 patients) in PBMCs using a fluorescent affinity probe. Various models were tested on the data using the first-order conditional estimation method as implemented in NONMEM version 7.1. A 2-compartment linear model with sequential zero-first order absorption and first order elimination was able to accommodate available PK data, including those of the validation dataset (prediction errors &amp;lt;15%). PK was dose- and time- independent. IBRU was rapidly absorbed, extensively distributed (volume of distribution at steady-state ~ 10,000 L) and cleared (apparent oral clearance ~1000 L/h). Relative bioavailability in the fasting state was about one third lower compared to the fed condition used in the clinical trials. No significant effect of other pathophysiological covariates on the PK was found (including sex, age or indication) except for body weight and coadministration of antacids, which had a marginal effect on the volume of distribution and duration of absorption, respectively. Analysis of PK-BTK engagement suggested that IBRU is a potent inhibitor of the BTK activity and that its interaction with BTK is rapid and durable. Citation Format: Italo Poggesi, Maria Luisa Sardu, Eleonora Marostica, Juthamas Sukbuntherng, Betty Y. Chang, Jan de Jong, Xavier Woot de Trixhe, An Vermeulen, Giuseppe De Nicolao, Susan Mary O'Brien, John C Byrd, Ranjana H Advani, Danelle Frances James, William Deraedt, Darrin Beaupre, Michael Wang. Population pharmacokinetic-pharmacodynamic (PKPD) modeling of ibrutinib in patients with B-cell malignancies. [abstract]. In: Proceedings of the AACR Special Conference on Hematologic Malignancies: Translating Discoveries to Novel Therapies; Sep 20-23, 2014; Philadelphia, PA. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(17 Suppl):Abstract nr B19.

Pooled post hoc analysis of population pharmacokinetics of oxycodone and acetaminophen following a single oral dose of biphasic immediate-release/extended-release oxycodone/acetaminophen tablets.

This analysis evaluated the single-dose population pharmacokinetics (PK) of biphasic immediate-release (IR)/extended-release (ER) oxycodone (OC)/acetaminophen (APAP) 7.5/325 mg tablets administered under fasted conditions and the effects of a meal on their single-dose population PK. Data were pooled from four randomized, single-dose crossover trials enrolling healthy adult (18–55 years old) participants (three trials) and nondependent recreational users of prescription opioids (one trial) with a body weight of ≥59 kg. Participants received IR/ER OC/APAP 7.5/325 mg tablets in single doses of 7.5/325 mg (one tablet), 15/650 mg (two tablets), or 30/1,300 mg (four tablets) under fasted or fed conditions. Six variables were examined: sex, race, age, weight, height, and body mass index. Single-dose population PK was analyzed using first-order conditional estimation methods. A total of 151 participants were included in the analysis under fasted conditions, and 31 participants were included in the fed analysis. Under fasted conditions, a 10% change in body weight was accompanied by ~7.5% change in total body clearance (CL/F) and volume of distribution (V/F) of OC and APAP. Black participants had 17.3% lower CL/F and a 16.9% lower V/F of OC compared with white participants. Under fed conditions, the absorption rate constant of OC and APAP decreased significantly, although there was no effect on CL/F and V/F. Considering that the recommended dose for IR/ER OC/APAP 7.5/325 mg tablets is two tablets every 12 hours, adjustments of <50% are not clinically relevant. Dose adjustment may be necessary for large deviations from average body weight, but the small PK effects associated with race and consumption of a meal are not clinically relevant.