Carbamazepine population pharmacokinetics in children: Mixed-effect models

The aim of the authors' study was to investigate the factors affecting carbamazepine (CBZ) clearance (CL) in children with epilepsy. The factors evaluated were total body weight (TBW), age, dose, sex, and phenobarbital (PB) and valproic acid (VA) comedication. A total of 387 steady-state serum concentration samples was analyzed. These were collected during CBZ therapy from 201 children, aged 1-14 years and weighting 9-78 kg. Population CL was calculated by using NONMEM, with a one-compartment model with first-order absorption and elimination. The absorption rate, bioavailability, and volume of distribution were set at values found in the literature. The model found best to describe the data was CL = (0.0122 TEW + 0.0467 Dose) Age(0.331) (1.289 PB). The interindividual variability in CL had a variation coefficient (CV) of 11.8%, and the residual error, described by using an additive model, was 1.5 mg/l. The results show that CL increases linearly with TEW and nonlinearly with age; thus older children have a lower CL with respect to TEW than do younger ones. Likewise CL was seen to increase with the increase in the CBZ dose, suggesting a dose-dependent autoinduction of CBZ metabolism. Concomitant PB administration affected CL; however, sex and VA comedication did not affect it significantly. The final regression model for CL was validated in a different group of 74 children. The standarized prediction error (SPE) was not significantly different from zero (SPE = 0.028), indicating that the model proposed for CL can be used to make accurate dosage recommendations. With these population estimates, CBZ doses that would be suitable for pediatric patients of different ages are proposed.