2.2.1 Population PK/PD model and simulation
A previously published population PK/PD model of edoxaban was used in our study 23. . To the best of our knowledge, this is the only published combined PK and PD model. This model was developed based on phase I and phase II studies. A total of 1,624 healthy subjects and NVAF patients were enrolled in the PK analysis, and a total of 585 NVAF patients were enrolled in the PD analysis. PK was described by a two-compartment model with first-order absorption with lag time and first-order elimination. The intrinsic FXa activity (iFXa) was selected as a PD marker and was modelled using a dynamic binding and equilibrium model. The PD data were modeled using dynamic binding model and equilibrium model. A summary of the demographic information of the modelling population is presented in Supplementary Table 1. The PK/PD model and all parameter estimates employed in the Monte Carlo simulation are listed in Table 2.
Monte Carlo simulations were performed with RXODE (version 0.9.1-8) in R (version 3.6.1) 24. Under each non-adherence scenario, one thousand virtual patients were simulated to obtain both the PK and PD profiles. The patients were assumed to reach a steady state before the delayed dose. Additionally, it was assumed that the current dose was sufficient to produce the desired anticoagulation effect without undesirable adverse effects.