Sparse sampling for assessment of drug exposure in toxicological studies.

In support of animal toxicity testing of new drugs, toxicokinetics is designed to assess the systemic exposure of the animals to the drug across dose levels, genders, and periods of the study. In small rodents, repeated sampling may alter the health of the animals and jeopardize the toxicity evaluation. One conventional way to circumvent this limitation is to collect serial samples from satellite animals maintained as the main study animals but not monitored for toxicity. We evaluated, on a real example, whether the exposure could be assessed in the main animals from sparse samples. The only acceptable designs consisted of one single sample per animal repeated on two or three study days. In the rat 13-week oral toxicity study of a new chemical entity, both serial sampling in the satellite animals and sparse sampling in the main animals were applied. Similar measures of exposure and qualitative conclusions were derived from the two groups of animals. The very sparse design applied to the main group yielded adequate estimation of the animal exposure, even with a very simple non-compartmental approach. The population pharmacokinetics analysis of the sparse data with NONMEM provided additional information about drug disposition and the influence of the covariates.