An influencing factor to predict plasma valproate concentrations: Circadian stage-dependent kinetics.

Abstract

This study was designed to examine how the dosing time of the day influenced the accuracy in predicting plasma valproate (VPA) concentrations at steady state. Eight healthy male volunteers took 400-mg doses of VPA for 9 days on a twice-daily basis (08: 30 and 20: 30). The circadian changes in VPA kinetics occurred at the absorption phase. The prediction of plasma VPA concentrations at 2hr (around Cmax) and at 12hr after both the morning and the evening dose on the ninth day was performed using the individual subject's pharmacokinetic parameters and population pharmacokinetic parameters (Bayesian method) obtained from the morning trial on the eighth day. The predictive accuracy for VPA concentration around Cmax after the morning dose was better, but that for VPA concentration after the evening dose was significantly biased toward overestimation. The individual pharmacokinetic parameters obtained from the morning trial became better sources for the prediction of VPA concentrations around Cmax after the morning dose, but worse sources for that after the evening dose. The predictive performance based on the Bayesian approach also showed the similar finding to that based on individual pharmacokinetic parameters. The overestimation of the target points around Cmax after the evening dose is closely related to that of Ka value produced by using population parameters obtained from morning trial. Therefore, these results suggest that the time in the circadian stage at which VPA is administered is important for evaluating exactly the predictive accuracy.