Possible interaction between warfarin and cranberry juice

Despite the lame seasonal reference, the report of Suvarna et al (BMJ
2003;327:1454)implicating cranberry juice in the causation of fatal
bleeding in a patient taking Warfarin is unconvincing. Two other factors
well known to interact dangerously with coumarin anticoagulation - vitamin
K deficiency due to starvation and broad-spectrum oral antibiotics - were
operating in the case in question. Cranberry juice was therefore the least
likely contributor to the bleeding episode.

The patient ate 'next to nothing' in the two weeks prior to the
bleeding. Starvation abolishes vitamin K intake. Warfarin's anticoagulant
effect is highly dependent on vitamin K status and strongly potentiated by
deficiency. The patient was also treated with the antibiotic cephalexin
which reduces non-dietary vitamin K absorption by its effect on the
intestinal flora. These factors acting in concert were entirely sufficient
to cause extreme sensitivity to Warfarin and resulting haemorrhage.

The authors enlist reports of International Normalised Ratio
variation outside the 'therapeutic' range in other patients on Warfarin
who consumed cranberry juice. This suggests unfamiliarity with the
clinical reality of Warfarin monitoring. Anticoagulation with this agent
is inherently variable: typically, only 50-60% of patients at a clinic
show INRs within the target range (1). This largely unexplained variation
is the reason we hold these huge and costly clinics. In the UK, it is a
racing certainty that 80% of patients found to be outside target INR will
have consumed tea - another flavonoid containing beverage - in the two
weeks before their visit. Would the authors like me to send in 100 or so
yellow cards each month to alert them to this striking and plausible
association? Advising patients to avoid cranberry juice seems about as
logical as warning them off tea.

The yellow card reporting system is of great value despite its
vulnerability to reporting bias. Publishing this case may attract more
such bias: I find it odd that affiliates of the Committee for Safety of
Medicines would want to do this.

1. Oral Anticoagulation Monitoring Study Group (2000). Prothrombin
time measurement using a patient self-testing system. American Journal of
Clinical Pathology; 115, 280-287.

Competing interests:
A company that markets cranberry juice has asked for the author's advice on the subject. No fee has been charged or paid.