Bayesian estimation of six different sets of carbamazepine pharmacokinetic parameters in Egyptian adult epileptic patients

Abstract

The individualization of carbamazepine (CBZ) dosage regimen based on estimation of pharmacokinetic (PK) parameters and measurement of serum drug concentration in epileptic patients can help to control epilepsy. In a retrospective study, the predictive performance of six different sets of CBZ PK parameters selected according to the literature was evaluated in 60 adult epileptic patients. Patients were administered controlled release CBZ (dose range: 200–1200 mg day−1) as monotherapy and one steady state serum concentration of the drug was available for each patient. The Bayesian Program of Abbott (PKS system; Abbott Laboratories, Wiesbaden, Germany) was used in the prediction process. Predictive measures included estimation of mean prediction error (mpe) for bias, mean squared prediction error (mspe) and root mean squared prediction error (rmspe) for precision. The analysis showed that three of the investigated six sets achieved the best predictive performance in Egyptian patients and consequently, the PK parameters of any of these three sets can be used by the Bayesian approach as prior information for CBZ dose optimization among the Egyptian adult population.