Malcolm Rowland, Robert L. Dedrick, in Principles of Clinical Pharmacology (Third Edition), 2012
Microdosing
Microdosing is a recent, novel, and essentially empirical approach to human pharmacokinetic prediction, which has been made possible by the development of the ultrasensitive analytical techniques discussed in Chapter 12 [10]. It involves administering to humans a minute, safe, subpharmacologic dose (a microdose) of the test compound (not greater than 0.1 mg) and proportionally scaling the observed pharmacokinetic profile, and hence dose, to the desired concentration–time profile intended to be evaluated in the Phase I study. Because microdosing is strictly neither a preclinical activity nor part of the normal Phase I program, it has been termed a Phase 0 study.
The microdose approach is based on two simple ideas: first, that there is no better pharmacokinetic predictor of humans than humans; and second, the assumption that the pharmacokinetics of the compound is dose proportional – that is, parameters, such as clearance, volume of distribution, and bioavailability, do not change with dose. Although essentially all processes within the body are eventually saturable if the dose is large enough, which clearly would violate the assumption of dose proportionality, it turns out that this assumption holds reasonably well for many compounds over the dose range of interest, as shown in Figure 32.5 for midazolam [11]. Exceptions to this assumption occasionally are encountered, and attempts are being made to improve predictions by coupling microdosing results with in vitro characterization of the saturable process. Another limitation of microdosing predictions is due to the fact that, at the minute doses administered, test compounds are invariably sufficiently soluble to dissolve in the dosing solution. As such, microdosing cannot be used to assess the performance of solid formulations, including tablets and capsules, a particular issue when dissolution of the drug critically determines its absorption. Nevertheless, early warning of a poor pharmacokinetic profile when a compound is given in solution is useful as it generally portends serious problems that are unlikely to be readily overcome by formulation efforts.

Neither microdosing nor allometry can predict pharmacokinetic events that may be encountered in studies beyond Phase I, such as the impact of genetics, disease, and age. This limitation does not apply in principle to the third approach, namely physiologically based prediction of human pharmacokinetics.