OBJECTIVES: (S)-Mephenytoin is selectively metabolised to (S)-4'-hydroxymephenytoin by CYP2C19. The urinary excretion of 4'-hydroxymephenytoin reflects the activity of individual enzymes. We evaluated fractioned urinary collection and beta-glucuronidase pre-treatment in order to determine the optimal CYP2C19 metrics. We also assessed whether urinary excretion of N-desmethylmephenytoin (nirvanol) might be a useful CYP2B6 metric in in vivo studies. METHODS: A 50-mg dose of mephenytoin was administered to 52 volunteers as a component of phenotyping cocktails in four separate studies. Urine was collected up to 166 h post-dose. Urinary excretion of 4'-hydroxymephenytoin and nirvanol was quantified by liquid chromatography-tandem mass spectrometry, and common CYP2C19 and CYP2B6 genotypes were determined. RESULTS: Cumulative excretion of 4'-hydroxymephenytoin in urine with beta-glucuronidase treatment collected from before mephenytoin administration up to 12-16 h thereafter showed the greatest difference between CYP2C19 genotypes and the lowest intra-individual variability (7%). Renal elimination of nirvanol was highest for a *4/*4 individual and lowest for individuals carrying the *5/*5 and *1/*7 genotype, but lasted for several weeks, thus making its use in cross-over studies difficult. CONCLUSION: Cumulative urinary excretion of 4'-hydroxymephenytoin 0-12 h post-administration is a sensitive and reproducible metric of CYP2C19 activity, enabling the effect of a drug on CYP2C19 to be assessed in a small sample size of n=6 volunteers. While nirvanol excretion may reflect CYP2B6 activity in vivo, it is not useful for CYP2B6 phenotyping.