A PBPK approach for simulating the effect of liver cirrhosis on drug PK
Annika R. P. Schneider (1,2), Rebekka Fendt (1,2), Jan-Frederik Schlender (2), Lars Kuepfer (1,2)
(1) Institute of Applied Microbiology, RWTH Aachen University, Germany; (2) Systems Pharmacology & Medicine, Bayer AG, Leverkusen, Germany
Objectives:
Liver cirrhosis is a progressive disease which is associated with severe morphological and physiological changes. Those alterations are known to substantially affect drug pharmacokinetics (PK). Physiologically based pharmacokinetic (PBPK) modeling is a valuable tool to link knowledge on physiological changes in diseases and to predict their influence on drug PK. Currently, the suitability of PBPK models for the prediction of alterations in PK profiles of cirrhotic patients is still under discussion. Especially the common Child-Pugh grading system complicates model construction due to limited translatability of the classification rules into actual physiological parameters. The objective of the presented approach was to assess and refine a current PBPK model for liver cirrhosis. The improvement achieved with these physiological model extensions is exemplary shown for ascites and liver enzyme activity.
Methods:
Model implementation and simulations were performed using PK-Sim® as part of the Open Systems Pharmacology Suite [1] and the included MATLAB toolbox. As a starting point, an existing liver cirrhosis model [2] was used. This model includes changes in blood flows, plasma protein concentrations, hematocrit, liver enzyme activities and glomerular filtration rate. A comprehensive literature research was conducted to identify further physiological changes and their potential influence on drug PK.
Based on this literature research, ascites was additionally implemented into the PBPK framework. This was done by increasing the interstitial fluid of the large intestine. Tobramycin was chosen as a test compound due to its hydrophilicity and ability to enter the ascitic fluid [3]. Using predicted PK parameters like the volume of distribution, the correct model implementation as well as the sensitivity for the ascites volume were systematically evaluated. The generated information was implemented into the existing liver cirrhosis model and was used to generate in silico populations with different grades of liver cirrhosis.
Not only new physiological changes were added to the existing model, but also a continuous approach for altered liver enzyme activity was developed. To inform the influence of liver cirrhosis on four different CYP enzymes, data was compiled in a comprehensive literature search. The data was then used in a Markov chain Monte Carlo (MCMC) approach [4] to inform a function (and a corresponding prediction interval) that describes the loss of enzyme activity in relation to disease progression. As a marker for disease progression the Child-Pugh Score was used.
Results:
The literature analysis revealed several potential physiological changes that could have an impact on drug PK and are not yet included in the existing liver cirrhosis model. These changes relate to ascites, porto-systemic shunting, intestinal liver enzyme activity, transporter activities and sinusoidal capillarization. As one example, ascites was successfully integrated into the model. Simulations with the hydrophilic test compound tobramycin revealed a strong correlation between the implemented ascites volume and the predicted volume of distribution. This is in line with data from literature [3].
The literature research on CYP1A2, CYP3A4, CYP2C19 and CYP2E1 activity in liver cirrhosis patients resulted in a dataset for each enzyme, showing continuous loss in activity on increasing Child Pugh Score. The functions resulting from the MCMC analysis show a decrease of activity throughout disease progression for all enzymes but with different slopes at different disease stages. This is in line with literature describing different CYP enzymes being differently affected in liver cirrhosis [5].
Conclusions:
In this study the impact of various pathophysiological alterations associated with liver cirrhosis was analyzed and implemented into a PBPK modeling framework. The changes like ascites and the liver enzyme activity, exemplary demonstrate the effect of an increased physiological level of detail compared to the already existing model. Hereby, continuous disease progression was taken into account while keeping the relation to the Child-Pugh grading system. In the future, this model will allow improved in silico trial simulations in cirrhotic patient cohorts and, thus, will be a helpful tool for the evaluation of drug efficacy and safety.
References:
[1] http://www.open-systems-pharmacology.org/
[2] Edginton, A.N. and S. Willmann, Clin Pharmacokinet., 2008; 47(11):743-52.
[3] Sampliner, R. et al., J Clin Pharmacol., 1984; 24(1):43-6.
[4] PAGE 27 (2018) Abstr 8583 [www.page-meeting.org/?abstract=8583]
[5] Murray, M., Clin Pharmacokinet., 1992; 23(2):132-46.