OBJECTIVES: Recent anecdotal, unvalidated case reports have suggested potentiation of warfarin-induced anticoagulation by cranberry juice, possibly through inhibition of human cytochrome P450 (CYP) 2C9, the enzyme responsible for the clearance of the active S-enantiomer of warfarin. To address this question, the effect of cranberry juice and other beverages on CYP2C9 activity was evaluated in vitro and in vivo.

METHODS: The effects of 4 beverages on CYP2C9 activity were studied in human liver microsomes, by use of flurbiprofen hydroxylation as the index reaction. In a clinical study 14 healthy volunteers received 100 mg flurbiprofen on 5 occasions in a crossover fashion, with at least 1 week separating the 5 trials. Flurbiprofen was preceded in random sequence by the following: (1) cranberry juice placebo (8 oz), (2) cranberry juice (8 oz), (3) brewed tea (8 oz), (4) grape juice (8 oz), and (5) fluconazole, a CYP2C9 inhibitor serving as a positive control, with 8 oz of water.

RESULTS: Flubiprofen hydroxylation in vitro was reduced to 11% +/- 8% of control by 2.5% (vol/vol) brewed tea, to 10% +/- 7% of control by grape juice, to 56% +/- 16% of control by cranberry juice, to 85% +/- 5% of control by cranberry juice placebo, and to 21% +/- 6% of control by the index inhibitor sulfaphenazole (2.5 micromol/L) (P .01 for all comparisons versus control). Flurbiprofen clearance (29-33 mL/min) and elimination half-life (3.3-3.4 hours) did not differ significantly among trials 1, 2, 3, and 4. However, clearance in the fluconazole treatment condition (trial 5) was significantly reduced compared with the placebo control (17 +/- 5 mL/min versus 31 +/- 8 mL/min, P .05), and the half-life was prolonged (5.3 +/- 1.6 hours versus 3.3 +/- 0.8 hours, P .05). Formation of 4-hydroxyflurbiprofen was correspondingly reduced by fluconazole (P .05).

CONCLUSIONS: Although grape juice and tea impaired CYP2C9 activity in vitro, none of the 3 beverages altered CYP2C9-mediated clearance of flurbiprofen in humans, making a pharmacokinetic interaction with warfarin highly unlikely.