A new population pharmacokinetic model for dosing optimization of zonisamide in patients with refractory epilepsy

Zonisamide exhibits significant pharmacokinetic variability, demanding for the development of population pharmacokinetic (PopPK) models to identify key factors influencing drug disposition. This study aimed to develop and validate a PopPK model to optimize zonisamide posology in patients with refractory epilepsy. A total of 114 plasma concentrations of zonisamide, obtained from 64 patients, were used for PopPK model development, employing the nonlinear mixed-effects modelling approach. The final model was evaluated by visually inspecting the goodness-of-fit plots and the visual predictive check plot and by the bootstrap resampling method. A one-compartment model with first-order elimination was the one that best described the pharmacokinetic profile of zonisamide. Between-patient variability (BPV) was included on clearance (CL/F), volume of distribution (Vd/F) and absorption rate constant (ka). The residual error (RE) was modeled as proportional. The final model estimates for CL/F, Vd/F and ka were 0.761 L/h, 48.10 L and 0.671 h⁻¹, respectively. The BPV associated with CL/F, Vd/F, and ka was 43.93%, 52.06%, and 91.27%, respectively, while the proportional RE was 7.18%. The concomitant administration of enzyme-inducing antiseizure drugs (EIASDs), included in the model as inducer drug load (INDDL), significantly accounted for BPV associated with CL/F and led to increased CL/F in patients receiving EIASDs compared to the others. Consequently, patients receiving EIASDs require higher daily doses of zonisamide to achieve therapeutic plasma concentrations compared to those not treated with EIASDs. Model validation, using bootstrap and visual predictive checks, confirmed its stability and robustness, making it a valuable tool for individualized zonisamide dosing in adults with refractory epilepsy.