Enhancing the predictive accuracy of PBPK models for drug concentrations in tissues: accounting for the impact of relative distribution of blood flow
Manuel Ibarra, Marianela Lorier, Pietro Fagiolino
Department of Pharmaceutical Sciences, Faculty of Chemistry, Universidad de la República. Uruguay.
Introduction: Physiologically based pharmacokinetic (PBPK) models are a crucial tool for model-informed drug development and are increasingly being explored for dose optimization in clinical settings. One of the key advantages of PBPK models is the ability to predict drug concentrations in different tissues, as these are explicitly described using a systems framework. However, the validation of these predictions is challenging due to the difficulty of measuring drug concentrations in tissues. The predicted outcome can vary significantly depending on the choice of the distribution model (i.e., the method to predict the tissue-to-plasma partition coefficient) [1], and there is a need to improve such predictions [2,3]. This is particularly true when there is variability in the clinical condition affecting physiological factors relevant for drug distribution and elimination. In our opinion, one important limitation of current PBPK models is not considering that molecules return to blood from the tissues by a constant intrinsic clearance that depends on the organ size instead of the supplied blood flow [4].
Objectives: The aim of this work is to propose an approach that enhances the predictive accuracy of PBPK models for drug distribution in tissues by accounting for the impact of the relative distribution of blood flow to each organ.
Methods: The following equation proposes a method to model drug distribution to tissues under perfusion rate-limited kinetics, assuming a constant clearance for tissue drug output:
VT*(dCT/dt) = CO*σT*CA – CO*ωT*(fuT/fuV)*CT
In this equation, CO represents the cardiac output, CA total drug concentration in arterial blood, CT total drug concentration in tissue, σT fraction of cardiac output to the tissue, ωT the tissue size expressed as a fraction of the whole extravascular water, and fuX the unbound fraction in either the blood (vein) or the tissue. The equation is derived by adjusting the tissue-to-blood partition coefficient (KT:B) by considering the ratio σT/ωT, which reflects the balance between organ size and blood flow to the organ. Under this approach, drug exposure in tissue becomes sensitive to σT, a phenomenon that has been observed, for instance, under physical activity [5]. We evaluated the approach using data from Joukhadar et al. [6], who observed changes in the ciprofloxacin tissue-to-plasma AUC ratio (T/P) in the adipose tissue of thigh before and after angioplasty in patients with peripheral arterial occlusive disease. Using a previously reported ciprofloxacin whole-PBPK model [7] implemented in NONMEM 7.5 (ICON plc), we simulated the reported scenarios using the traditional perfusion rate-limited model and the proposed adjustment, to evaluate the impact of changing the blood flow delivered to each extremity (healthy and ischemic) on T/P.
Results: Under the current framework, the T/P is not altered by the blood flow to the tissue (CO*) in non-eliminating organs, and therefore the widely used perfusion rate-limited organ model fails to describe how the changes in the blood flow distribution can alter the tissue-to-blood partition. Using the proposed approach, we were able to describe the reported T/P before and after angioplasty in both healthy and ischemic tissue considering a 20% reduction in the blood flow to the compromised extremity.
Conclusions: The proposed approach was successfully implemented to describe the reported changes for ciprofloxacin T/P in adipose tissue before and after angioplasty. The equation allows to further evaluate the impact of changes in blood flow distribution on T/P in the context of PBPK models.
References:
[1] Utsey K et al. Quantification of the Impact of Partition Coefficient Prediction Methods on Physiologically Based Pharmacokinetic Model Output Using a Standardized Tissue Composition Drug Metabolism and Disposition (2020) 48(10): 903-916.
[2] De Sutter PJ et al. Predictive Performance of Physiologically Based Pharmacokinetic Modelling of Beta-Lactam Antibiotic Concentrations in Adipose, Bone and Muscle Tissues. Drug Metab Dispos (2023).
[3] Muliaditan M et al. Prediction of lung exposure to anti-tubercular drugs using plasma pharmacokinetic data: Implications for dose selection. European Journal of Pharmaceutical Sciences (2022) 173, 106163
[4] Ibarra M et al. Current PBPK Models: Are They Predicting Tissue Drug Concentration Correctly? Drugs (2020) 20:295-299.
[5] Lenz T. The effects of high physical activity on pharmacokinetic drug interactions. Expert Opinion on Drug Metabolism and Toxicology (2011) 7(3):257-266
[6] Joukhadar C et al. Angioplasty increases target site concentrations of ciprofloxacin in patients with peripheral arterial occlusive disease. Clinical Pharmacology and Therapeutics (2001) 70(6): 532-539.
[7] Sadiq MW et al. A whole-body physiologically based pharmacokinetic (WB-PBPK) model of ciprofloxacin: a step towards predicting bacterial killing at sites of infection. Journal of Pharmacokinetics and Pharmacodynamics (2017) 44:69–79.
