What is PAGE?

We represent a community with a shared interest in data analysis using the population approach.

   Uppsala, Sweden

Evaluation of a transit compartment model versus a lag time model for describing drug absorption delay

Radojka Savic(1), DaniŽl M. Jonker(1), Thomas Kerbusch(2), Mats O. Karlsson(1)

(1) Div. of Pharmacokinetics and Drug Therapy, Dept of Pharmaceutical Biosciences, Faculty of Pharmacy, Uppsala University, Sweden. (2) Dept. Clinical PK/PD M&S, Pfizer Global R&D, Sandwich, Kent, UK

PDF of presentation

Background: Traditionally, a delay before the commencement of drug absorption has been modeled as a discrete lag time. At this lag time, the concentration-time profile is discontinuous, which is unattractive from a computational point of view and which does not have physiological basis. An analytical solution of a delay model which approximates a series of transit compartment may avoid these drawbacks of the discrete lag model.

Aim: To compare the performance of the standard lag-time model with an analytical solution of the transit compartment model, which allows the (non-integer) number of transit compartments to be estimated, on data from 4 different compounds.

Methods: Concentration-time data on glibencamide, furosemide, amiloride, and moxonidine was analyzed. In the transit compartment model, absorption delay was described by the passage of drug through a series of transit compartments with a single transfer rate. Drug transfer to the central compartment was described by a second rate constant. The optimal number of transit compartments was estimated from the data. The population pharmacokinetic analysis was performed in NONMEM using the FOCE method with interaction. Goodness-of-fit was assessed by the decrease in OFV value and by inspection of diagnostic graphs.

Results: For all investigated drugs, a significant absorption delay was estimated with both the transit compartment model and the lag time model. With the transit compartment model, the goodness-of-fit was visibly better in the absorption phase and around the concentration peak compared to the lag time model. A significant drop in the OFV up to 530 points was observed with all compounds. The estimated number of transit compartments for glibenclamide, furosemide, amiloride and moxonidine were 22.9, 9.7, 12.4 and 7.2 respectively. The estimates of the remaining PK parameters were similar between the two models.

Conclusion: Based on the results from this comparison with four drugs, the transit compartment model is an attractive alternative for modeling drug absorption delay, especially when the drug absorption phase is poorly described by a lag time model.