Study rationale

Carbamazepine (CBZ) and valproic acid (VPA) are long-established treatments for paediatric epilepsy. However, their dosing presents challenges due to their complex pharmacokinetics, especially in children. Both drugs also exhibit high inter-individual variability. VPA is extensively bound to albumin in a concentration-dependent manner, while CBZ undergoes auto-induction by inducing its own major metabolizing enzyme, both of which complicate dose optimization. This study aimed to evaluate the appropriateness of the current Dutch Paediatric Formulary (DPF) dosing guidelines to achieve therapeutic levels for CBZ and VPA. If room for optimization was observed, alternative dosing regimens were explored. In addition, the impact of altered albumin levels impact on VPA exposure was explored.

The role of PBPK modelling in supporting clinical decision-making

PBPK modelling takes the body’s physiological processes into account, providing a more comprehensive understanding of how drugs behave across different age groups. In this study, PBPK models for CBZ and VPA (immediate and extended release) were used to simulate paediatric drug levels under current dosing guidelines. These models helped evaluate whether existing guidelines were adequate for achieving therapeutic drug levels in children and where dosing can be optimized. The PBPK models also provided deeper mechanistic insights into how factors like hypoalbuminemia may affect VPA exposure.

Evaluation of paediatric dosing

The simulations showed that therapeutic levels of both drugs are reached within 1–2 weeks of treatment, and therefore supported the current dosing strategy, where treatment is started with a low dose and increased weekly, depending on the child’s age.

For CBZ, PBPK modelling provided valuable mechanistic insights into metabolite exposure. The results suggested a lower starting dose for neonates. However, it is difficult to compare this with other recommendations because clear guidelines for this age group are limited. For children aged 12–18 years, a higher starting dose could be considered based on pharmacokinetics viewpoint. Still, the standard of care is to initiate treatment with a low dose to reduce side effects.

For VPA, the current dosing strategies seem to result in appropriate drug levels across all age groups. However, total drug levels do not always reflect the amount of active drug in the body. The model showed that factors such as low albumin levels can affect VPA levels. This means that in some patient groups, dosing or monitoring may need further optimisation.

PBPK modelling proved to be a valuable tool to evaluate current dosing guidelines, and how it can support clinical decision-making.

The full paper can be accessed here.