Eleni Karatza, Vangelis Karalis
Department of Pharmacy, School of Health Sciences, National and Kapodistrian University of Athens, Greece
Objectives: Develop a joint pharmacokinetic model, describing the kinetics of losartan (LOS) and its active metabolite (EXP-3174) that seem to be significantly affected by gastric emptying.
Methods: LOS and EXP-3174 plasma concentration (C) – time (t) data were obtained from a single dose, 2×2 bioequivalence study comparing two immediate release oral products in 25 men and women. Concentrations were expressed in pmoles per mL for the joint modeling of LOS and EXP-3174. Non-linear mixed-effect modeling was applied and a variety of pharmacokinetic models were examined. In order to mathematically describe the delayed and variable appearance of LOS and EXP-3174 in plasma, resulting in oscillations of the concentration, several approaches were explored including: the use of gastrointestinal pseudo-compartments, delay differential equations, time-dependent gastric emptying, and periodic gastric release using sinusoidal functions. Models tested were evaluated in terms of their physiological relevance and goodness-of-fit criteria. Several error models were evaluated, whereas the period and treatment effects were tested as potential covariates. The entire computational work was implemented in Monolix 2016R1.
Results: The disposition of LOS was best described by a two-compartment model preceded by another compartment representing a pre-absorption gastro-intestinal compartment. For the metabolite EXP-3174, a one-compartment model led to the optimum performance. Delay differential equations were found to be the most appropriate approach to mathematically describe the oscillations [1] observed in the plasma concentration of LOS and EXP-3174. Indeed, both the absorption rate of LOS and the metabolite formation rate were expressed by first order constants integrated in a system of delay differential equations. The pharmacokinetic parameters derived for losartan were the first order gastric emptying constant (Kg = 3.98 h-1), the absorption rate constant in the central compartment (Ka = 1.38 h-1), the absorption lag time (tau1 = 0.131 h), the apparent volume of distribution of the central (Vp1/F = 50.3 L) and peripheral (Vp2/F = 211 L) compartment, the apparent clearance from the central compartment (CL/F = 125 L/h), and the inter-compartmental clearance (Q/F = 165 L/h). For the active metabolite, EXP-3174, the formation rate constant (Km = 0.455 h-1), the formation lag time (tau2 = 0.281 h), the apparent volume of distribution (Vm/F = 16.4L), and clearance (CLm/F = 4.14 L/h) were derived. Based on the findings of this study, it can be generally deduced that similar models with delay differential equations can describe the pharmacokinetics of orally administered drugs with high solubility and low permeability, which are affected by gastric emptying as LOS. Application of a combined error model led to the optimum performance for both LOS and EXP-3174, whereas no statistically significant difference was observed in the performances of the two drug products.
Conclusions: The most appropriate joint pharmacokinetic model for LOS and EXP-3174 plasma C-t data consisted of a two-compartment model for LOS and a one-compartment model for EXP-3174, using delay differential equations for both the absorption rate and the metabolite formation rate.
References:
[1] Raphaël Kuate, Marc Lavielle, Eric Blaudez, Kaelig Chatel, Jerome Marquet, et al.. A delay differential equation solver for MONOLIX & MLXPLORE. [Research Report] RR-8489, INRIA. 2014, pp.19.
Reference: PAGE 27 (2018) Abstr 8560 [www.page-meeting.org/?abstract=8560]
Poster: Drug/Disease Modelling - Absorption & PBPK