Discrimination of kinetic models in heteroscedastic parametric estimation of bentazepam following multiple dosing

The aim of the present study was to attempt to discriminate between single- and two-compartment kinetic models used for calculating the pharmacokinetic parameters of bentazepam when the plasma concentrations of different administrations are used as initial data during multiple dosage regimes. Determination of the best estimated pharmacokinetic parameters was performed using non-linear regression analysis, weighting the data as a function of the error of the analytical technique. Bentazepam was administered at a dose of 25 mg orally at intervals of 8, 12 or 24 h to a total of 9 patients. The mean values of the parameters established for the single-compartment model were Ka = 2.024 h-1; Vd = 2.198 l/kg and Ke = 0.130 h-1. For the two-compartment model these values were: Ka = 2.134 h-1, Vc = 2.049 l/kg, K10 = 0.154 h-1, K12 = 0.042 h-1 and K21 = 0.103 h-1. By application of the MAICE test mean AIC values of 41.62 and 42.52 were obtained for the single- and two-compartment models, respectively. The most suitable kinetic model was determined for each patient according to the predictive nature of the individual parameters of the two kinetic models and by analysis of the residuals of the non-linear regressions of the parametric estimation.