found that rifampicin significantly enhanced the total body clearance of
midazolam, from 5.5 mL/min/kg on day 1 to 10.1 mL/min/kg on day 15.
Notably, the twofold induction of CYP3A4-dependent activity measured by
the systemic clearance of midazolam was consistent with the twofold increase
in erythromycinN-demethylation, measured by the erythromycin breath test
performed in the same subjects. A similar study by Backman et al., utilizing
midazolam clearance as a measure of CYP3A4 induction, demonstrated that
rifampicin treatment (600 mg daily for 5 days) reduced Cmax plasma
concentrations of midazolam (15 mg administered PO) from 55 to 3.5 ng/
mL, reduced the AUCtotal from 10.2 to 0.4mg.min/mL, and reduced the
midazolam half-life from 3.1 to 1.3 h (Backman et al., 1996).
Theoretically, the least hazardous method of directly assessing CYP3A4
inductionin vivoin humans would be to measure the CYP3A4-dependent
generation of a metabolite of an endogenous substrate, which is predominantly
excreted into urine. This would provide a noninvasive method that would not
involve administration of a xenobiotic or a radiolabeled compound. Recently,
the clearance of 4b-hydroxycholesterol, an endogenous CYP3A4-dependent
metabolite with an apparent half-life of 52 h, was found to be a good indicator
of CYP3A4 induction (Bodin et al., 2001, 2002). For example, the plasma
levels of 4b-hydroxycholesterol in humans treated with antiepileptic drugs
phenobarbital, carbamazepine, or phenytoin were significantly increased by
7- to 20-fold over normal levels (Bodin et al., 2001, 2002). In addition,
6 b-hydroxycortisol is a metabolite of cortisol that is believed to be formed
exclusively by CYP3A4, and excreted into the urine. Rifampicin treatment
(600 mg daily for a week) increased the renal excretion of 6b-hydroxycortisol
by more than threefold (Dilger et al., 2005). In addition, 25- or 6b-
hydroxylated bile acids, also believed to be endogenous metabolites produced
by CYP3A4, may be useful markers for CYP3A4 induction, because they are
excreted in urine; and therefore, easily measured (Araya and Wikvall, 1999;
Furster and Wikvall, 1999; Handschin et al., 2002; Ourlin et al., 2002).
Thus far, a qualitative correlation betweenin vitroandin vivoestimates of
CYP3A4 induction has been established, but only in the case of a few, already
well-characterized inducers, and quantitative correlations are less clear-cut.
Several potent inducers of CYP3A4, including St. John’s wort, efavirenz,
progestins, dexamethasone, and rifampicin produced a positive induction
response in bothin vitroassay systems, and with the erythromycin breath test
in vivo, as described above in this section. Also, a number of compounds that
induce CYP3A4in vitroalso reducein vivothe AUC of coadministered drugs
that are CYP3A4 substrates. These CYP3A4 inducers include carbamazepine,
phenobarbital, phenytoin, rifampin, St. John’s wort, topiramate, and
troglitazone (Luo et al., 2004; Ripp et al., 2006). Several factors should be
considered when comparing and extrapolatingin vitroandin vivoresults. These
factors include, but are not limited to, the compound’s dosage, route and
duration of administration, pharmacokinetic properties, tissue distribution,
and potential to produce mechanism-based inactivation of CYP3A4, all of
564 TESTING DRUG CANDIDATES FOR CYP3A4 INDUCTION