Pharmacokinetic theory dictates that the extent of ensuing metabolism of a formed metabolite during drug transit through the liver is influenced by the number of consecutive reactions required for its genesis and the total intrinsic clearances of the precursors. This hypothesis was tested in the perfused murine liver by examining the successive conversion of the precursor diazepam (DZ) to its primary metabolite nordiazepam (NZ), and then the secondary metabolite oxazepam (OZ) and, finally, the tertiary metabolite, the oxazepam glucuronides. The concomitant C3-hydroxylation of DZ to temazepam, which can also be N-demethylated to form OZ, was minimal. The hepatic extraction ratios of NZ (E[NZ,DZ]) and OZ (E[OX,DZ]) after administration of [14C]DZ were compared to those obtained previously from [14C]NZ (E[NZ] and E[OZ,NZ]) and [3H]OZ (E[OZ]). The ability of three hepatic clearance models, the well-stirred, parallel-tube and dispersion models, to predict the experimental E[NZ,DZ] and E[OZ,DZ] was evaluated. DZ was highly extracted by the murine liver (E[DZ] = 0.95). The metabolism of NZ, generated in situ from DZ, was greater than that of preformed NZ (E[NZ,DZ] = 0.51; E[NZ] = 0.4), whereas E[OZ,DZ] (0.066) was similar to E[OZ,DZ] (0.056) and less than E[OZ] (0.0125). The unexpected observation of E[NZ,DZ] > E[NZ] may be explained by the coupling of N-demethylation and C3-hydroxylation/glucuronidation reactions or by a sequestration of hydrophobic substrates within the enzymic space, favoring sequential metabolism of products formed in situ. The atypical kinetic behavior of generated NZ may have also influenced the ensuing metabolic fate of its product, OZ, such that E[OZ,NZ] approximately E[OZ,DZ].(ABSTRACT TRUNCATED AT 250 WORDS)