M. Simeoni(1), F. Fiorentini (2), M. Germani(2), I. Poggesi(2), M. Rocchetti (2)
(1) Dipartimento di Informatica e Sistemistica, Pavia University, Pavia, Italy (2) Global Drug Metabolism, Pharmacia, Nerviano, Italy
Objectives:In a typical toxicokinetic program, the pharmacokinetic information is obtained from different studies characterized by a limited number of observations (in terms of number of experimental units and plasma samples/experimental unit). The possibility of performing meta-analyses as soon as new data are becoming available during the development of a new drug is one of the most interesting application of the population pharmacokinetic analysis. In this communication, this approach was applied to a real toxicokinetic program, in which resampling was applied to one of the studies to simulate a sparse sampling strategy.
Methods: Two toxicity studies of a new intravenous anticancer drug were retrospectively considered: a single and a cyclic dose study. A 2-compartment model with elimination from the central compartment and a model of errors including random effects and fixed effects for gender, dose and cycle was used. Two population analyses were performed using NONMEM: in the first case both the complete data sets from the two studies were considered; in a second analysis only the complete data set of the single dose study and a subset (approximately 10%) of the data in the cyclic study were used. The results of both population analyses were compared with those obtained from a conventional non-compartmental analysis performed on the two complete data sets.
Results: The results obtained from the two different population approaches were in excellent agreement. Compared with the non-compartmental approach, both population analyses gave similar results in terms of average pharmacokinetic parameters. Similar conclusions were also obtained in terms of dose-, time- and gender-dependencies.
Conclusion: Overall, population approaches can be efficiently applied to preclinical programs, allowing the incorporation into the models of the growing body of preclinical information. With some prior knowledge, such as that obtained in the first studies, sparse sample analysis can be applied in a toxicokinetic program, with a consequent substantial reduction of costs.
Reference: PAGE 12 (2003) Abstr 410 [www.page-meeting.org/?abstract=410]
Poster: poster