The pharmacokinetics of many drugs often vary considerably among individuals, largely because of variations in the expression of different cytochrome P-450 (CYP) enzymes in the liver and other tissues. Relatively selective substrate probes in vivo have been discovered for several major CYP isoforms involved in oxidative drug metabolism. Regarding isoforms that show genetic polymorphism (CYP2C19 and CYP2D6), genotyping as well as phenotyping with appropriate probe drugs can be used to distinguish between "poor" and "extensive" metabolizers. Measurement of CYP2D6 activity, which is being performed increasingly by means of genotyping, has an established role in the individualization of the dosage of selected CYP2D6 substrates. However, the therapeutic implications of extremely high CYP2D6 activity in some patients (ultrarapid metabolizers) need more attention. The therapeutic consequences of CYP2C19 polymorphism are not as well characterized as those of CYP2D6 polymorphism, but are likely to be of little significance with most CYP2C19 substrates. Probe-based assays are also available for measurement of in vivo activity of CYP1A2, CYP2E1 and CYP3A4; those will be discussed in detail in this review. These tests can be used, for example, to compare the activity of a specific isoform among patients and to characterize effects of such environmental factors as drugs and compounds in the diet on enzyme activity. However, it should be recognized that attempts to develop valid probe-based assays of in vivo activity of specific, nonpolymorphic CYP isoforms have proved relatively difficult; for example, none of the several putative probes of CYP3A4, the most important drug-metabolizing CYP isoform, is completely satisfactory. It is now clear that many diverse factors must be considered in the validation of these tests.