Modelling tissue pharmacokinetics: A comparison of empiric and mechanistic PBPK modelling approaches of different complexity exemplified with moxifloxacin
Khalid Iqbal (1), Hartmuth Nowak (2), Caroline Weismann (2), Stefan Martini (2), Matthias Unterberg (2), Michael Adamzik (2), Christoph Dorn (3), Frieder Kees (3), Zoe A. Österreicher (4), Markus Zeitlinger (4), Sebastian G. Wicha (1)
(1) Dept. of Clinical Pharmacy, Institute of Pharmacy, University of Hamburg, Hamburg, Germany (2) Universitätsklinikum Knappschaftskrankenhaus Bochum, Klinik für Anästhesiologie, Intensivmedizin und Schmerztherapie, Bochum, Germany (3) Universität Regensburg, Institut für Pharmazie, Regensburg, Germany (4) Medizinische Universität Wien, Allgemeines Krankenhaus (AKH), Universitätsklinik für Klinische Pharmakologie, Vienna, Austria
Objectives: Suboptimal tissue site exposure is a critical factor for therapeutic failure, potentially stimulating antibiotic resistance and is hence crucial to investigate using a pharmacometric approach. Thereby, a pharmacometric model shall provide good predictive performance and provide mechanistic insight into the tissue pharmacokinetics (PK). The aim of this study was to (i) develop an empirical, a whole-body physiologically-based (WBPBPK) and minimal-PBPK model for moxifloxacin in septic patients integrating microdialysis data from the target sites subcutaneous and muscle tissue and (ii) to compare their predictive performance and mechanistic inferences.
Methods: Plasma (total and unbound), subcutaneous and muscle tissue samples from 10 septic patients were collected at day 1, 3 and 5 after intravenous once daily dosing of 400 mg moxifloxacin. Microdialysis was used to collect samples from subcutaneous and muscle tissue.
In the first step, an integrated empiric microdialysis pharmacometric model [1] was developed on the data where all compartment parameters were empirically estimated.
In the second step, a WBPBPK model consisting of fifteen physiologically-motivated compartments including plasma, skin and muscle compartments was built. Blood flows and tissue volumes were calculated from the ICRP report [2], scaled by sex and body weight. Tissue partition coefficients (KP) were calculated based upon their physicochemical properties [3] and used as informative priors (uncertainty: 25%) during estimation from the available clinical data.
In the third step, a minimal-PBPK model [4] was built by lumping various tissue compartments into central and peripheral compartment while retaining the clinically important muscle and subcutaneous tissues unlumped. KP values were calculated as for the WPBPK model.
All models included estimation of recovery measurements from the microdialysis data [1]. Visual predictive checks (VPC’s) were used to compare the predictive performance of all the models. NONMEM® 7.4 was used for all estimation and simulation tasks.
Results: For the empiric approach, a four compartment model with total body weight as a covariate on clearance (CL) and central volume of distribution (V1) best described the plasma and tissue PK of moxifloxacin. For a 75 kg patient, CL was 12.6 L/h (interindividual variability, IIV: 21.5 %CV), V1 was 15.5 L (IIV: 44.8 %) and peripheral volume of distribution (V2) was 98. 7 L (IIV: 93.8 %). The distribution volumes of the tissues were only imprecisely estimated and differed substantially from the volume of the interstitial space fluid (ca. 12 L).
The WBPBPK model predicted the typical concentration time profiles of plasma and target tissues adequately and the compartments were physiologically meaningful. Yet, variability was overpredicted if inter-individual variability on KP values was estimated and the model was unstable and exhibited very long run times.
The minimal PBPK model combined the good predictive performance, high model stability and short run times of the empirical model with the physiologically plausible description of the tissue PK. Additionally, the PBPK models indicated blood-flow mediated distribution to muscle and permeation-restricted distribution to subcutaneous tissue –a mechanistic insight that was not provided by the empiric approach.
Conclusions: Moxifloxacin displayed good tissue penetration to muscle and skin tissue. The minimal PBPK modelling approach provided an attractive framework for modelling of the tissue PK of moxifloxacin in a mechanistically plausible and efficient way and combined the favourable properties of the empiric and WBPBPK approach. Application of the presented approach to further tissue PK studies is warranted.
References:
[1] Tunblad, K. Udenaes, H. M. and Jonsson, N. E. An integrated model for the analysis of pharmacokinetics data from microdialysis experiments. Pharm. Res. (2004); 21 (9): 1698-1707
[2] Basic anatomical and physiological data for use in radiological protection: reference values. A report of age- and gender-related differences in the anatomical and physiological characteristics of reference individuals. ICRP Publication 89. Ann ICRP, (2002); 32 (3–4): 5–265
[3] Rodgers, T. and Rowland, M. Physiologically based pharmacokinetic modelling 2: Predicting the tissue distribution of acids, very weak bases, neutrals and zwitterions. J. Pharm. Sci. (2006); 95 (6): 1238-57
[4] Cao Y. and Jusko W. J. Applications of minimal physiologically-based pharmacokinetic models. J Pharmacokinet Pharmacodyn. (2012); 39 (6): 711-23