Predicting human pharmacokinetics from preclinical data.

Approaches used for the prediction of pharmacokinetics in relevant populations of human patients mostly rely on in vivo data from animals, using allometric scaling or time-invariant methods. The growth of in vitro and, more recently, in silico screens for evaluating pharmaceutical, pharmacokinetic and toxicity properties can also be used to predict complex in vivo behavior in humans. In most cases, careful and educated application of available approaches provides predictions of pharmacokinetic parameters within 2- or 3-fold of that observed. Attention should now be directed toward integrating information from different sources to increase the precision and accuracy of these pharmacokinetic predictions and to enable a better understanding of the processes underlying ADME behavior in humans.