The diazepam (DZ)-omeprazole (OMP) interaction has been selected as a prototype for an important drug-drug interaction involving cytochrome P450 inhibition. The availability of an in vivo K(i) value (unbound K(i), 21 microM) obtained from a series of steady-state inhibitor infusion studies allowed assessment of several in vitro-derived predictions of this inhibition interaction. Studies monitoring substrate depletion with time were used to obtain in vitro K(i) values that were evaluated against the more traditional metabolite formation approach using microsomes and hepatocytes. OMP inhibited the metabolism of DZ to its primary metabolites 4'-hydroxydiazepam, 3-hydroxydiazepam, and nordiazepam to different extents over a range of concentrations (0.3-150 microM), and a competitive inhibition model best fitted the data. The K(i) values observed using the substrate depletion approach (16 +/- 3 microM and 7 +/- 2 microM in microsomes and hepatocytes, respectively) were in good agreement with the overall weighted K(i) values obtained using the standard metabolite formation approach (12 +/- 2 microM and 16 +/- 5 microM in microsomes and hepatocytes, respectively). In vitro binding and cell uptake studies as well as human serum albumin studies in hepatocytes confirmed the importance of both intracellular and extracellular unbound concentrations of inhibitor when considering inhibition predictions. Both kinetic approaches and both in vitro systems predicted the in vivo interaction well and provide a good example of the ability of in vitro inhibition studies to quantitatively predict an in vivo drug-drug interaction successfully.