Pitfalls in pharmacokinetic multicompartment analysis.

When a pharmacokinetic (PK) two-compartment body model with first-order absorption is fitted to blood levels of a drug, the estimates of the PK parameters may have considerable errors and can cause wrong predictions in other features of the system. The objectives of this report were to illustrate this problem, to provide an easy way to prevent wrong estimation, and to investigate the origin of the mistake. A simple way to prevent wrong interpretation of the calculated PK parameters is to inspect the PK profiles visually. Without observing a clear biphasic profile, one should not apply the two-compartment model if the resulting parameters are to be interpreted and used for further simulations. We investigated the origin of this ambiguity in terms of the relative order of magnitude of microconstants (ka, k12, k21, and k10) and of hybrid constants (A and B). The observed parameter errors will not be of any relevance if the calculated parameters are used only to predict future blood levels over the same time-span. However, if these parameters are used to predict any other characteristic of the system, erroneous predictions may result.