A patient was treated with warfarin for atrial fibrillation. During treatme
nt, an increase in international normalized ratio (INR) and her admission t
hat she was taking a variety of natural products, to include boldo and fenu
greek, led us to suspect that some of these natural products could alter th
e effect of warfarin. When she stopped the culpable products, the INR retur
ned to normal after I week. The herb-drug interaction was observed a second
time after both products were reintroduced a few days later. The imputabil
ity of this interaction to both natural products, as determined by the Nara
njo algorithm, suggests a probable association between boldo-fenugreek and
increased bleeding time in patients treated with warfarin. No undesirable r
eaction was reported during telephone discussions with the patient. Neverth
eless, we recommend that clinicians treating patients with anticoagulant th
erapy be vigilant when patients also take herbal agents.