Mgambi Gideon Gamba 1,2, Ayatallah Saleh 1,2, Amrei Konrad 1, David Busse 5, Gerd Mikus 3, Wilhelm Huisinga 2,4, Charlotte Kloft 1,2, Robin Michelet 1
1 Graduate Research Training program PharMetrX (Berlin, Germany), 2 Freie Universität Berlin, Germany (Berlin, Germany), 3 University Hospital Heidelberg, Germany (Heidelberg, Germany), 4 University of Potsdam (Potsdam, Germany), 5 Boehringer Ingelheim Pharma GmbH & Co. KG (, Germany)
Voriconazole (VRC) is a second-generation triazole antifungal with broad-spectrum activity, and remains a cornerstone of antifungal therapy, particularly for invasive fungal infections [1]. It is characterized by nonlinear pharmacokinetics (PK), pronounced inter- and intraindividual variability, and a narrow therapeutic index [2]. The nonlinear PK of VRC arises largely from its metabolism by CYP2C19, CYP3A4, and to a lesser extent by CYP2C9, enzymes for which VRC serves as both a substrate and an inhibitor. Its main circulating metabolites, VRC N-oxide (NO) and hydroxy-VRC (OHVRC), further inhibit these same isoenzymes, adding additional complexity to nonlinear PK [3]. Because CYP2C19 and CYP3A4 are central to the metabolism of many concomitant medications, VRC therapy carries substantial potential for drug-drug and drug-drug-gene interactions (DDGI). Importantly, clinically relevant differences exist between paediatric and adult populations, with markedly reduced oral bioavailability reported in paediatric patients (44.6%) compared with healthy adults (96.0%) [4].
This work aimed to understand the oral bioavailability of VRC by expanding a developed intravenous parent-metabolite PBPK model with newly generated in vitro data and mechanistically extending it to capture oral absorption, thereby establishing a robust foundation towards a comprehensive CYP2C19- and CYP3A4-mediated DDGI network.
Methods
A previously validated intravenous parent–metabolite PBPK model of VRC, NO, and OH-VRC served as the structural foundation [5]. The model was refined by incorporating newly generated in vitro plasma protein binding (PPB) data for VRC and its primary metabolites, replacing literature-derived and previously assumed PPB. It was subsequently mechanistically extended to describe oral absorption. As tablet formulations were simulated, dissolution kinetics were implemented using a Weibull function. Dissolution parameters were adopted from published PBPK models, with the dissolution time for 50% of the dose set to 30 minutes and the Weibull shape parameter fixed at 1.29 [6]. Gastrointestinal solubility, effective intestinal permeability, and segment-specific CYP expression were integrated to capture absorption and pre-systemic metabolism. Nonlinear elimination of NO was preserved via a hypothetical hepatic enzyme and additionally expressed in the intestine to account for potential intestinal metabolism. Model verification was performed using PK data from 72 healthy volunteers across different CYP2C19 genotype-predicted phenotypes (Rapid, Normal, intermediate, and poor metabolizers) following single-dose oral administration of 50 mg and 400 mg in studies conducted at Heidelberg University Hospital [7]. Finally, sensitivity analysis was conducted in the oral model to inform subsequent modelling steps.
Results
PPB in vitro values were 56% for VRC, 14.4% for NO, and 87.7% for OHVRC. Following mechanistic extension to the oral setting, the model successfully reproduced systemic exposures across VRC 400mg single dose and CYP2C19 genotype-predicted phenotype normal metabolisers, demonstrating acceptable predictive performance for both parent drug and metabolites. Model performance varied across analytes but remained overall satisfactory. For VRC, 76.9% of measured plasma concentrations were within two-fold of model predictions, with 53.8% within 1.25-fold (AFE = 0.744; AAFE = 1.58; RMSE = 0.769). Predictive performance was strongest for NO, with 84.6% concentrations within two-fold and 53.8% within 1.25-fold (AFE = 0.763; AAFE = 1.48; RMSE = 1.10). Performance was comparatively lower for OH-VRC, where 66.7% of observations were within two-fold and 33.3% within 1.25-fold (AFE = 0.883; AAFE = 1.76; RMSE = 0.172). Subsequent sensitivity analysis of the oral model identified VRC solubility, relative enzyme expression levels in the intestine, and intestinal permeability as key determinants of oral exposure predictions, highlighting the mechanistic relevance of absorption and metabolic processes in shaping systemic PK [8].
Conclusions
By integrating newly generated in vitro data with a mechanistic extension of an established intravenous parent–metabolite PBPK model, the oral bioavailability of voriconazole across CYP2C19 genotype-predicted phenotypes was successfully characterized. The model demonstrated good predictive performance for VRC and NO and to a lesser extent on OHVRC following oral dosing, supporting its structural adequacy and indicating areas for further model refinement. This work provides a quantitatively framework that captures the nonlinear disposition and complex CYP2C19–CYP3A4 interplay of VRC. As such, it establishes a robust foundation for the systematic investigation of CYP-mediated drug-drug interactions and represents a key step towards building a comprehensive DDGI network to inform precision dosing of oral VRC. Future work will involve parameter estimation in MoBi, followed by systematic DDGI simulations to support precision dosing, therapeutic drug monitoring, and dose optimisation in paediatric population.
References:
[1] M. Resztak, J. Sobiak, A. Czyrski. Recent Advances in Therapeutic Drug Monitoring of Voriconazole, Mycophenolic Acid, and Vancomycin: A Literature Review of Pediatric Studies. Pharmaceutics 13: 1991 (2021).
[2] J. Schulz, F. Kluwe, G. Mikus, R. Michelet, C. Kloft. Novel insights into the complex pharmacokinetics of voriconazole: a review of its metabolism. Drug Metabolism Reviews 51: 247–265 (2019).
[3] J. Schulz, A. Thomas, A. Saleh, G. Mikus, C. Kloft, R. Michelet. Towards the Elucidation of the Pharmacokinetics of Voriconazole: A Quantitative Characterization of Its Metabolism. Pharmaceutics 14: 477 (2022).
[4] M.O. Karlsson, I. Lutsar, P.A. Milligan. Population Pharmacokinetic Analysis of Voriconazole Plasma Concentration Data from Pediatric Studies. Antimicrob Agents Chemother 53: 935–944 (2009).
[5] A. Saleh, J. Schulz, J.-F. Schlender, L.B.S. Aulin, A.-P. Konrad, F. Kluwe, G. Mikus, W. Huisinga, C. Kloft, R. Michelet. Understanding Voriconazole Metabolism: A Middle-Out Physiologically-Based Pharmacokinetic Modelling Framework Integrating In Vitro and Clinical Insights. Clin Pharmacokinet 63: 1609–1630 (2024).
[6] M. Li, P. Zhao, Y. Pan, C. Wagner. Predictive Performance of Physiologically Based Pharmacokinetic Models for the Effect of Food on Oral Drug Absorption: Current Status. CPT Pharmacom & Syst Pharma 7: 82–89 (2018).
[7] I. Scholz, H. Oberwittler, K. Riedel, J. Burhenne, J. Weiss, W.E. Haefeli, G. Mikus. Pharmacokinetics, metabolism and bioavailability of the triazole antifungal agent voriconazole in relation to CYP2C19 genotype. Brit J Clinical Pharma 68: 906–915 (2009).
[8] H.M. Lazarus, J.L. Blumer, S. Yanovich, H. Schlamm, A. Romero. Safety and Pharmacokinetics of Oral Voriconazole in Patients at Risk of Fungal Infection: A Dose Escalation Study. The Journal of Clinical Pharma 42: 395–402 (2002).
Reference: PAGE 34 (2026) Abstr 12271 [www.page-meeting.org/?abstract=12271]
Poster: Drug/Disease Modelling - Absorption & PBPK