O. Sayar(1), L. Bueno(1), M.J. Garrido(1), F. Rombout(2), I.F. Trocóniz(1)
(1) Department of Pharmacy, School of Pharmacy, University of Navarra, Pamplona, Spain; (2) Department for Modelling and Simulation, Grünenthal GmbH, Aachen, Germany
Introduction: The two most popular models used to fit drug and metabolite kinetics are: Model I, where the drug is absorbed and once it has reached systemic circulation, metabolite is appearing in plasma due to the continuos pass through the liver; and model II, where both, parent drug and metabolite are simultaneously absorbed with equal or different input rates. Model II would reflect a lost of parent drug due to a hepatic first past effect. Models I and II have the advantage that if the analyst chooses to first fit the parent data, the model parameters can be used in the analysis of the metabolite data. However, there are cases where those two models, despite of increasing the number of disposition compartments, and extra routes of elimination do not provide an adequate fit. In those situations, a model incorporating a liver compartment (model III) has been shown to improve model performance. Such model has been used rarely and still is not very well documented.
Purpose: To explore by means of computer simulations situations where the liver compartment model can provide better fits, which trends appearing in the goodness of fit plots suggest the inclusion of liver compartment, and interpretation of the model parameter estimates.
Methods: Simulations were performed assuming first order input kinetics, one compartment distribution kinetics from both drug and metabolites, and only one elimination route for the parent and metabolite. Combinations of restricted and non-restricted parent compound and metabolite characteristics together with different input rates values were explored. Models II and III were used to simulate concentration vs time profiles, which were analysed by models I to III. For the case of the simulations with model III, different distribution rates from and to the liver compartment were also used. Simulations and analyses were performed with NONMEM, and model discrimination was based on the minimum value of the objective function and visual inspection of the goodness of fit plots.
Results and conclusions: In general the kinetics of the parent drug was adequately described for the three models. With regard to the metabolite, Model I performed usually very poorly, performance of model II was affected by the ratio between the clearance of the drug and metabolite. Model III performed the best in most of the scenarios. After single administration, delayed predicted Cmax for the metabolite was the major indication for choosing model III.
Reference: PAGE 13 (2004) Abstr 530 [www.page-meeting.org/?abstract=530]
Poster: poster