M. Looby, U. Wünscher, M. Weiss
Institute of Pharmakology, Martin-Luther-University, Halle Germany
Destructive sampling (DS) is frequently used to yield data from animals in pharmacokinetic experiments. The assumption made in such experiments is that the variation between animals is small and can therefore be ignored in a naive-pooled approach. This is justified to a certain degree because variability can be largely excluded by the use of similar animals. Nevertheless not all variability can be excluded, and to which extent it might compromise parameter estimation has not yet been examined. In toxicokinetics this problem is exacerbated by the toxic effects of the administered substance which will increase interanimal kinetic variation. Determination of intra- and inter-animal variance in one-sample-per-animal designs is difficult if not impossible. Inclusion of organ concentrations in the model should stabilize the variance estimates of the parameters describing blood kinetics.
This question shall be exmained using preliminary data from a DS experiment to examine the toxicokinetics of dibromchloromethane (DBCM) in rats. Blood, fat, muscle, and heart concentrations were measured in 70 rats following single oral doses of DBCM. Since little previous information was available on this substance, this preliminary experiment was necessary to provide information for an optimal design with regard to dose and sampling strategy. The development of the optimal design shall be assisted by simulation experiments.
Analysis of the preliminary data using a semiparametric approach
show that adipose tissue represents the primary depot with a
partition coefficient to blood of about 5. A Bateman function was
sufficient to describe the blood kinetics and monoexponential
conductance functions were adequate for each tissue. The
calculated interanimal variation was very large, e.g.
, 
. It remains to be seen in
simulation experiments to which extent such variances can be
determined in a DS design, and what advantages if any a NONMEM
approach has to offer over a more classical approach.