The feasibility of model-based exposure calculations in preclinical toxicology
Núria Buil-Bruna (1) *, Tarjinder Sahota (1), Meindert Danhof (1), Oscar Della-Pasqua (1,2)
(1) Division of Pharmacology, LACDR, Leiden University, Leiden, The Netherlands; (2) Clinical Pharmacology Modelling & Simulation, GlaxoSmithKline, Stockley Park, UK; *Current Institution: Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Navarra, Pamplona, Spain
Objectives: An important part of drug development involves the establishment of safe exposure levels in humans. Preclinical experiments often use an empirical non-compartmental approach to the calculation of drug exposure. This has a variety of limitations including the difficulty to extrapolate to clinical exposure levels. Pharmacokinetic modelling can address these shortcomings, however the sparse pharmacokinetic sampling that if often used in these experiments calls into question the feasibility of a model-based approach. The aim of this study was to evaluate the expected model parameter precision from preclinical designs and explore the possibility of reductions to the numbers of animals used.
Methods: Expected parameter precision was determined by calculation of the expected Fisher Information Matrix (FIM) in the software PopED, for 9 different hypothetical drugs. These drugs were described by a variety of structural PK models and parameter values. Models tested included non-linear kinetics and distribution to a peripheral compartment. Experimental designs matched standard practices and included sparse and serial sampling designs. Reduced designs with fewer animals were also tested. The secondary parameters, AUC and CMAX were calculated from the primary PK parameters using NONMEM VI. Acceptable precision was defined as CV < 30% for fixed effects and CV < 50% for random effects.
Results: The results show that secondary parameter precision remained below 35% in all cases. A reduction of animals used in composite designs by 2/3 yielded no significant loss in expected precision.
Conclusions: The use of pharmacokinetic modelling to characterise toxicokinetic exposure is feasible without changes to toxicity study protocols. Significant reductions to the numbers of animals for sparse designs may be possible using an integrated approach. The FIM-based approach used here, can be used as alternative to lengthy simulation/re-estimation for the calculation of expected parameter precision.
