Aymara Sancho Araiz1
1Division of Systems Pharmacology and Pharmacy, Leiden Academic Centre for Drug Research, Leiden University, 2Roche Pharma Research and Early Development, Roche Innovation Center Basel, 3Department of Clinical Pharmacy, St. Antonius Hospital
Introduction: Physiologically-based pharmacokinetic (PBPK) models are powerful tools for predicting drug absorption, distribution, metabolism, and excretion [1]. A key aspect of PBPK models for orally administered drugs is the estimation of first-pass metabolism, which includes both intestinal (FG) and hepatic (FH) availability [2]. The extended well-stirred liver model [3], which includes the hepatic blood flow (QH) and accounts for permeability, clearance and fraction of drug unbound in blood and tissue, adequately characterizes FH. The QGUT model is typically applied to estimate FG [4], however this approach has limited generalizability to drugs with low permeability [2]. Despite great advantages of PBPK models, the high complexity and long run times limit their utility for parameter estimation [1]. Our aim is therefore to characterize FG and FH (through the estimation of intestinal intrinsic clearance (CLint,G) and hepatic intrinsic clearance (CLint,H)) using different mathematical simplifications of a base PBPK model. The model includes a generalized QGUT (gQGUT) model, allowing the applicability of the structure to drugs with varied physicochemical properties. Methods: The base PBPK model was built on a previous framework [5, 6] and included five physiological compartments, gut wall (enterocytes and enterocyte blood), portal vein, liver (liver tissue and liver blood), and three empirical compartments (a central blood compartment, and two peripheral compartments for distribution). The physiological tissues were connected by blood flow, with no active drug transport mechanisms, and unbound concentrations drive distribution and metabolism. Physiological blood flows, tissue volumes, and tissue/blood concentration ratios were taken from literature [2, 3, 5, 6, 7]. Systemic clearance was assumed to occur exclusively in the liver and pre-systemic metabolism in the gut wall and the liver. Four model simplifications were explored by progressively reducing the number of ordinary differential equations (ODEs) describing the physiological compartments: (A) ODEs for gut wall, portal vein, and liver, (B) quasi-steady state approximation for gut wall, portal vein, and liver, (C) ODE-based model with steady-state solutions for FG and FH and, (D) analytical form of (C) with steady-state solutions for FG and FH. For all models, drug distribution was modelled empirically with a two-compartmental model. The generalized Qgut model was derived for the estimation of FG. Midazolam plasma concentrations from 12 healthy individuals following both oral and intravenous administration [8] were analysed using NONMEM 7.4. Estimated parameters include: CLint,H, CLint,G, absorption rate constant, volume of distribution of central and peripheral compartments, and inter-compartmental clearance. Final model estimates and model performance were compared. Analytical and numerical solutions for the estimates FG and FH were assessed for possible discrepancies among the models. Results: Parameter estimation of the four different model reductions were compared to the base PBPK model, the objective function values (OFVs) were: OFVbasePBPK=664.87, OFVA=663.81, OFVB=669.16, OFVC=669.17, and OFVD=669.20. Parameter estimates were remarkably similar. For CLint,G and CLint,H, the base PBPK model estimates (relative standard error %) were 603. L/h (8) and 11.2 L/h (7), respectively; and for Model D were 603 L/h (8) and 10.9 L/h (7). Individual CLint estimates were consistent across the different models (R2 = 0.95). Model run times were improved from 8071.99 seconds of the estimation step and 12386.65 seconds of the covariance step for the base model to 17.65 and 11.52 seconds, respectively, in Model D. FH and FG were adequately estimated in all scenarios. Conclusions: By comparing ODE-based versus analytical PBPK models using midazolam as case example, we evaluated the feasibility of a reduced approach without compromising physiological relevance. Moreover, the proposed gQgut model enhances applicability to drugs with varied physicochemical properties. The ODE-based approaches provide a mechanistic foundation, while the analytical form (applicable to linear kinetics) significantly reduce computational demand without compromising parameter estimation accuracy, allowing broader application to address complex clinical pharmacology questions (e.g., paediatrics and drug-drug interactions).
[1] Madabushi, R., Seo, P., Zhao, L. et al. Review: Role of Model-Informed Drug Development Approaches in the Lifecycle of Drug Development and Regulatory Decision-Making. Pharm Res 39, 1669–1680 (2022). https://doi.org/10.1007/s11095-022-03288-w [2] Gertz M, Houston JB, Galetin A. Physiologically based pharmacokinetic modeling of intestinal first-pass metabolism of CYP3A substrates with high intestinal extraction. Drug Metab Dispos. 2011 Sep;39(9):1633-42. doi: 10.1124/dmd.111.039248. Epub 2011 Jun 1. PMID: 21632965. [3] Pang KS, Han YR, Noh K, Lee PI, Rowland M. Hepatic clearance concepts and misconceptions: Why the well-stirred model is still used even though it is not physiologic reality? Biochem Pharmacol. 2019 Nov;169:113596. doi: 10.1016/j.bcp.2019.07.025. Epub 2019 Aug 6. PMID: 31398312. [4] Yang J, Jamei M, Yeo KR, Tucker GT, Rostami-Hodjegan A. Prediction of intestinal first-pass drug metabolism. Curr Drug Metab. 2007 Oct;8(7):676-84. doi: 10.2174/138920007782109733. PMID: 17979655. [5] Brill MJ, Välitalo PA, Darwich AS, van Ramshorst B, van Dongen HP, Rostami-Hodjegan A, Danhof M, Knibbe CA. Semiphysiologically based pharmacokinetic model for midazolam and CYP3A mediated metabolite 1-OH-midazolam in morbidly obese and weight loss surgery patients. CPT Pharmacometrics Syst Pharmacol. 2016 Jan;5(1):20-30. doi: 10.1002/psp4.12048. Epub 2015 Dec 18. PMID: 26844012; PMCID: PMC4728292. [6] Brussee JM, Yu H, Krekels EHJ, Palic S, Brill MJE, Barrett JS, Rostami-Hodjegan A, de Wildt SN, Knibbe CAJ. Characterization of Intestinal and Hepatic CYP3A-Mediated Metabolism of Midazolam in Children Using a Physiological Population Pharmacokinetic Modelling Approach. Pharm Res. 2018 Jul 30;35(9):182. doi: 10.1007/s11095-018-2458-6. PMID: 30062590; PMCID: PMC6096899. [7] Gertz M, Houston JB, Galetin A. Physiologically based pharmacokinetic modeling of intestinal first-pass metabolism of CYP3A substrates with high intestinal extraction. Drug Metab Dispos. 2011 Sep;39(9):1633-42. doi: 10.1124/dmd.111.039248. Epub 2011 Jun 1. PMID: 21632965. [8] Cleary Y, Gertz M, Jaminion F, Lennon-Chrimes S, Bentley D, Derks M. Physiologically-based pharmacokinetic modelling of basmisanil to evaluate CYP3A mediated victim and perpetrator drug-drug interaction propensity. PAGE meeting 2022 Ljubljana, Slovenia (https://wwwpage-meetingorg/defaultasp?abstract=9979). Accessed March 2025.
Reference: PAGE 33 (2025) Abstr 11661 [www.page-meeting.org/?abstract=11661]
Poster: Methodology - New Modelling Approaches