Study objective: To conduct a population pharmacokinetic analysis of carbamazepine (CBZ).
Design: Retrospective chart review.
Setting: Ambulatory neurology clinics at three medical centers.
Patients: Patients diagnosed with epilepsy from 1991-1995. The index set included 829 adults receiving CBZ. A separate validation set consisted of 50 patients.
Interventions: None.
Measurements and main results: Final regression equations were apparent oral clearance (Cl/F) (L/hr) = (0.0134 x TBW + 3.58), x 1.42 if receiving phenytoin only; x 1.17 if receiving phenobarbital or felbamate; x 1.62 if receiving phenytoin and phenobarbital or felbamate; x 0.749 if age > or = 70 years; apparent volume of distribution (Vd/F) (L) = 1.97 x total body weight; absorption rate constant [hr(-1)] = 0.441. Interindividual variability in Cl/F and Vd/F was 26% and 82%, respectively. Residual variability was 1.8 mg/L. Predictive performance analysis of the validation set provided a mean prediction error of 0.6 mg/L and median absolute error of 2.4 mg/L.
Conclusions: These routinely collected data provided quantitative estimates of changes in CBZ Cl/F due to comedication and an age-related decrease in Cl/F The derived regression equations reasonably predicted concentrations in a separate validation set.