Characterization of non-linear relationship between total and unbound serum concentrations of valproic acid in epileptic children

Objective: To establish a regression equation to properly estimate the unbound serum concentration of valproic acid (VPA) from its total serum concentration; the relationship between total and unbound serum VPA concentrations was retrospectively characterized. Methods: Data were obtained from the clinical examination records that were routinely archived during therapeutic drug monitoring. The screening encompassed 342 records of 108 paediatric patients whose total and unbound VPA concentrations had been determined. The relationship between total and unbound VPA concentrations was characterized according to the Langmuir equation by taking account of inter-individual variability with the NONMEM program. Results: The total VPA concentration (C-t) in the screened patients ranged from 5.5 to 179.8 mu g/mL, and the unbound VPA concentration (C-f) increased in a non-linear manner as the total VPA concentration increased. Taking account of the effects of antiepileptics concurrently administered, the VPA dissociation constant (K-d) and maximum binding site concentration (B-m) were 7.8 +/- 0.7 and 130 +/- 4.5 mu g/mL respectively, for the regression equation, C-t = C-f + B-m.C-f/(K-d + C-f). An alteration in the unbound concentration was seen in patients who were treated with the combination of VPA and ethosuximide and in those who received two additional antiepileptics. Conclusions: A regression equation for estimation of the unbound VPA concentration, based on total VPA concentration collected during routine therapeutic drug monitoring was established. Use of two additional antiepileptics and ethosuximide treatment was considered as potential factors affecting unbound VPA concentration.