Pitfalls and valid approaches to pharmacokinetic analysis of mean concentration data following intravenous administration.

Pharmacokinetic analysis fo arithmetic mean concentration data can lead us to selection of an inappropriate deterministic compartmental model and biased pharmacokinetic parameter estimates. The terminal phase disposition rate constant estimated by fitting a deterministic model to mean data is in all cases an underestimate of the expected value of this rate constant. The area under the mean data curve calculated via the linear trapezoidal rule from time zero to the last detectable concentration sampling point is equal to the mean of the individual subject areas under the curve for the same time span. This equality supports the use of mean data for determination of model-independent pharmacokinetic parameters.