in vivodrug interaction study may not be needed. The ratio of [I]/Kionly
predicts the likelihood of anin vivointeraction, not the exact magnitude of the
interaction. As the ratio increases, the likelihood of an interaction increases.
However, both [I] andKican be quite variable depending on the methods used
for each individual drug. A short survey shows that theKivalues can be up to
48-fold apart in experiments using the same substrate and the same inhibitor
toward the same CYP enzyme (Fig. 7.1). The variation inKivalues using the
same inhibitor but different substrates is far greater across various sponsors
and laboratories (Table 7.4). Differential sensitivity of CYP3A4 substrates
toward inhibition by the same inhibitor indicates that enzyme–substrate
interactions at the CYP3A4 active site(s) are substrate dependent (Wang et al.,
2000). The importance of estimating inhibitor potency using the drug of
interest as the substrate has been emphasized. Use of alternate substrate drugs
for the same enzyme can potentially yield inaccurate predictions ofin vivodrug
interactions.
The regulatory agencies recommend the use of total drug concentrations of
the mean steady-stateCmax(bound plus unbound) of the highest proposed
1 2 3 4 5 6 7 8 9 1011050100150 Max/min >48K(μM)iFIGURE 7.1 Kiin human liver microsomes for ketoconazole to inhibit the formation
of 1-OH-midazolam fromin vitro studies (data from 11 published articles from
University of Washington Drug Metabolism database).
TABLE 7.4 KetoconazoleKivalues determined with different CYP3A4 substrates.
Substrate Ki(mM) Reference
Midazolam 0.011 Li et al. (2004)
Triazolam 0.026 von Moltke et al. (1996)
Alprazolam 0.046 von Moltke et al. (1994)
Nifedipine 0.05 Wandel et al. (2000)
Trazodone 0.12 Zalma et al. (2000)
Lovastatin 0.4 Jacobsen et al. (1999)
Cisapride 1.5 Desta et al. (2000)
Terfenadine 3 Jurima-Romet et al. (1994)
Tacrolimus 11 Christians et al. (1996)
DRUG–DRUG INTERACTION STUDIES 221