Monika Twarogowska (1), Geraldine Cellière (1), Pauline Traynard (1), Claude Magnard (1)
(1) Simulations Plus, Lixoft division, Antony, France
Objectives: For most drugs, a total or part of a dose undergoes a biotransformation before elimination from the human body. It can lead to the presence of pharmacologically active metabolites, which can drive partial or total drug activity and contribute to the therapeutic effect. But they can also be toxic and so bring safety concerns. Joint parent-metabolite modeling helps in understanding and predicting key metabolites, which can be critical during a drug development process [1].
We have developed a modular parent-metabolite model library implemented in MonolixSuite to simplify the selection and testing of different hypotheses. The library helps to fit a large variety of parent-metabolite data. The models are defined using the Mlxtran macros and the implementation includes analytical solutions, whenever they exist, to make the model estimation algorithms faster and less prone to numerical instabilities.
Methods: The library is organized into a clear model selection workflow to aid modelers when selecting the most appropriate model for their needs. It allows to choose:
- different structural models for parent and metabolite (number of compartments, elimination process)
- uni- or bi-directional transformation between parent and metabolite drug
- various drug administration routes (iv/oral/iv+oral, no delay/delay/transit compartments, zero/first order absorption)
- a first-pass-effect with or without dose apportionment
The modular selection via the graphical interface allows to combine easily all the above options into a wide variety of models and is incorporated into a step-by-step modeling workflow with the MonolixSuite applications. The workflow includes visualizing the data set to characterize the parent and metabolite dynamics, setting up and estimating several models in Monolix, assessing the uncertainty of the parameter estimates, comparing the different runs in Sycomore to select the best solution, and simulating the final model with Simulx to investigate the impact of different treatments.
Results: The use of the library is exemplified on a dataset for oxycodone and its metabolite noroxycodone. The library was able to easily fit step-by-step the parent-metabolite pharmacokinetics, starting with a simple PK model for the parent, before moving to a joint parent-metabolite model. Different model hypotheses were tested, in particular first-pass-effect and back-transformation. In addition, the possibility to estimate inter-individual variability was precisely assessed for each parameter. The final model properly captures the data for both the parent and the metabolite. Over-parameterization of the model was verified using the convergence assessment and diagnostic tools.
Secondly, the advantage of using the analytical solution compared to solving the ODE system has been investigated. Models with 1, 2 and 3 compartments for the parent and the metabolite, and with a single dose and multiple doses, have been benchmarked. The comparison shows that the analytical solution is 2.5 to 5 times faster than the ODE solver.
Conclusions: The MonolixSuite and the new parent-metabolite library allow an efficient modeling and diagnosis of parent-metabolite data assuring optimal numerical performance. The modularity of the library gives the user the possibility to test many options without spending time writing equations.
References:
[1] J. Bertrand et al. (2011) Development of a complex parent-metabolite joint population pharmacokinetic model., The AAPS journal, 13(3), pp. 390–404.
Reference: PAGE 30 (2022) Abstr 10169 [www.page-meeting.org/?abstract=10169]
Poster: Methodology - Other topics