# Coupled pharmacokinetic model unveils drug-drug interactions in plasma concentration

## TL;DR

A new pharmacokinetic model helps better understand how drugs interact when taken together, improving predictions of drug concentrations in the body.

## Contribution

A linearly coupled two-compartment model with interpretable drug-drug interaction terms is introduced for improved pharmacokinetic analysis.

## Key findings

- The model captures concentration-dependent changes during combined drug administration more accurately.
- The full model outperforms simplified versions in goodness-of-fit and predictive performance.
- The model bridges traditional compartmental and PBPK models for drug interaction analysis.

## Abstract

In oral drug pharmacokinetics (PK), drug-drug interactions are inevitable, yet traditional compartmental models struggle to effectively quantify such processes. This study proposes a linearly coupled two-compartment PK model, where the coupling term is defined as a linear function of another drug’s amount to strike a balance between model simplicity and physiological interpretability. The model introduces parameter heterogeneity and linear interaction terms based on the classical compartmental structure, more accurately capturing concentration-dependent dynamic changes during combined drug administration. To address the model’s nonlinear characteristics and high-dimensional parameters, a hierarchical optimization numerical solution algorithm was developed, enhancing computational efficiency while validating robustness against Gaussian noise. Through systematic analysis of key PK metrics (Cmax, Tmax, AUC, and t1/2), the study reveals the mechanisms by which absorption and clearance parameter variations influence drug distribution in vivo. Combining numerical simulations, parameter ablation experiments, and real-world data validation, the full model (retaining all linear interaction terms) outperforms the simplified model in both goodness-of-fit and information criteria, demonstrating superior interpretability and predictive performance. Overall, this model offers an intermediate solution between traditional compartmental models and PBPK models, providing a novel methodological framework for quantitative research on drug-drug interactions.