The aims of this study were to define the relationship between total and unbound valproic acid (VPA) concentrations, to compare pharmacokinetic parameters of total and unbound VPA, and to determine the difference in pharmacokinetic parameters between the seizure-controlled and -uncontrolled groups. Total and unbound steady-state serum concentration of VPA were determined at trough and at 5 hours after the morning dose from 29 pediatric patients with epilepsy who were receiving valproic acid as monotherapy every 8 or 12 hours. Pharmacokinetic parameters were calculated and compared by t test (alpha=0.05). A strong relationship between total and unbound concentrations was found. The equation for predicting unbound concentration from total concentration was unbound concentration=-6.01+0.18 * total concentration (r=0.78, P<0.001). A higher percentage of free fraction was found at higher concentrations. There were significant differences between total and unbound VPA pharmacokinetic parameters. Both the elimination rate constant (Ke) and volume of distribution (Vd) of the unbound VPA were much higher than those of the total VPA; clearance (CI) was approximately 10-fold higher. The seizure-uncontrolled group could eliminate VPA much faster than the seizure-controlled group, as indicated by a much higher Ke and a much shorter t1/2 (P<0.05). Vd of the unbound VPA was also much smaller in the seizure-uncontrolled group (P<0.05). There were significant differences in pharmacokinetic parameters between seizure- controlled and -uncontrolled pediatric epileptic patients. Further study with a larger sample size is strongly recommended.