Objective: To develop a population pharmacokinetic model of valproic acid (VPA) and to evaluate factors influencing VPA pharmacokinetic parameters in epileptic patients.

Material and Method: Population pharmacokinetics analysis was performed using 215 VPA concentrations of 136 epileptic outpatients receiving oral VPA from routine clinical data at Prasat Neurological Institute. A non-linear mixed effects one compartment model with first order absorption and elimination was fit to the data using the nonlinear mixed effects model (NONMEM®) software package. A first-order conditional estimation method with interaction (FOCEI) was used for parameter estimation. Interindividual variability and residual variability were estimated using proportional model. Model validation was performed using 1,000 simulations based on the bootstrap method.

Results: Base on the final model, the apparent oral clearance (CL/F) of VPA was 0.585 L/hr, apparent oral volume of distributon (Vd/F) was 23 L and absorption rate constant (Ka) was fixed to 0.35 hr-1. Significant covariates on CL/F were the daily dose of VPA, co-administered phenytoin and body weight. The final model for CL/F of VPA was CL/F = 0.585 x [1+0.000362(DOSE-1,000)] x [1+0.379PHT] x [1+0.00798(WT-61.9)]. The interindividual variability in CL/F was 16.46% and residual variability was 28.05%. The 95% CI of all parameters were similar to the range obtained from the bootstrap, indicating that the final model estimates were reliable.

Conclusion: VPA pharmacokinetics from routine clinical data was best described by one compartment model with first order absorption and elimination. The daily dose of VPA, co-administered phenytoin and body weight were significant covariates influencing CL/F of VPA. These factors should be considered when defining VPA dosage regimens in epileptic patients.