Drug Metabolism in Drug Design and Development Basic Concepts and Practice

with respect to time and enzyme concentrations. The 2006 draft guidance for
drug interaction studies adds CYP2B6 and CYP2C8 as required enzymes for
in vitroandin vivostudies.

7.4.1.2 Time-Dependent Inhibition Time-dependent inhibition of CYP
enzymes should be conducted as a routine part ofin vitroscreening studies,
because the prediction of the extent of drug–drug interactions by the new drug
may be underestimated significantly by using competitive models if the new
drug is a mechanism based inhibitor. For instance, theKivalue for inhibiting
midazolam metabolism by erythromycin is quite large, and the prediction using
[I]/Kiwould far underestimate the extent of inhibitory drug interaction between
these two drugs (Table 7.2). In general, a various time of preincubation with a
new drug before the addition of substrate is recommended. Any time-dependent
loss of initial product formation rate indicates mechanism-based inhibition.
Detection of time-dependent inhibition kineticsin vitromay suggest the need for
follow-up within vivostudies in humans after multiple dose administration.

7.4.1.3 [ I]/Ki In the past few years, NDA and IND submissions increas-
ingly contain prediction of in vivo drug interactions using in vitro drug
metabolism data, although a complete understanding of the relationship
between the in vitrofindings and in vivo results of metabolism/drug–drug
interaction studies is still emerging. The most popular and simplest approach is
to directly apply [I]/Kito the prediction of fold increase in exposure when an
inhibitor is present. With this approach, [I] represents the mean steady-state
Cmaxvalue for total drug (bound plus unbound) following administration
of the highest therapeutic clinical dose and theKiis the inhibition constant
(Table 7.3). The consensus cut-off value for the [I]/Kiratio is 0.1. When [I]/Kiis
greater than 0.1, a clinically relevant drug interaction may occur, and a follow-
upin vivodrug interaction study is recommended. When [I]/Kiis<0.1, an

TABLE 7.2 Thein vitroprediction and actualin vivodrug interaction between
erythromycin and midazolam.

Kia Cmaxb [I]/Ki Actual AUCI/AUCCONc

8.8mM 2.9mM 0.33 >fourfold

aAtkinson et al. (2005).
bYu et al. (2001).
cOlkkola et al. (1993).

TABLE 7.3 Prediction of CYP inhibition.

[I]/Ki Prediction

[I]/Ki> 1 Likely

1 >[I]/Ki>0.1 Possible

0.1>[I]/Ki Remote

220 REGULATORY CONSIDERATIONS OF DRUG METABOLISM AND DRUG