Integrated pharmacokinetic and metabolic modeling of ipriflavone and metabolites after oral administration.

OBJECTIVE

Disposition of ipriflavone, an agent indicated for the treatment of osteoporosis, is highly variable after oral administration. Ipriflavone has five major metabolites (M1, M2, M3, M4, and M5). The metabolites M2 and M5 have activity and are major metabolic constituents in humans. Hence, it is important to characterize both ipriflavone and metabolites simultaneously. Our purpose was to develop an integrated pharmacokinetic/metabolic model that simultaneously predicts plasma concentrations of ipriflavone and metabolites after a single oral administration.

METHODS

The model was based on the reported metabolic conversion of ipriflavone to M1, M3, and M4; subsequent conversion of M4 to M5; and conversion of both M1 and M3 to M2 in rats. The further conversion of M5 to M6 and M7 was ignored, as this conversion represents an insignificant portion of the total metabolite pool. The input function was described by a first-order constant. Each analyte required two-compartment disposition. The elimination/nonmetabolic constants for each analyte accounted for urinary elimination. Plasma concentration data from a pilot pharmacokinetic study in which 16 healthy male volunteers were administered 200 mg of an ipriflavone corn suspension were used to examine the predictability of this model.

RESULTS

The coefficient of determination was 0.99, and model selection criterion was 3.7 for mean data fits supporting the goodness-of-fit and predictability of the model. The model also predicted negligible urinary recoveries for ipriflavone, M1, M3, and M4; M2 and M5 had high urinary recoveries. The metabolic conversion constant from M3 to M2 was negligible. Divergence from the proposed pathway may be attributed to the species differences in metabolism between humans and rats.

CONCLUSIONS

Model predictions supported the improvement in bioavailability with corn-oil suspension compared to the conventional oral tablet. Future model applications may also help identify significant covariates (i.e., age, gender, and disease state) in proposed clinical trials.