Some Comments on the Evaluation of Bioavailability Data

Abstract

Intrasubject variability in the systemic clearance of a drug from one treatment to the next in a randomized crossover bioavailability study reduces the ability to differentiate dosage forms with respect to the extent or relative extent of drug absorption. The most commonly used method to reduce this intrasubject variability is a half-life correction first proposed in 1967. The mathematical basis of the half-life correction is reviewed, and it is shown that this correction requires the implicit assumptions that the change in half-life from one treatment to the next quantitatively reflects the change in systemic clearance and that there is no change in apparent volume of distribution. The limitations of this assumption are considered. It is concluded that the half-life correction should be evaluated in all cases where the variability in half-life is random, and applied if it results in a significant decrease in the standard deviation associated with the average measure of bioavailability. On the other hand, if there is a significant difference or strong indication of a difference in half-life between treatments, the half-life correction should not be applied, for it can lead to misinterpretation. Examples illustrating this problem are presented.