Predictors of unstable anticoagulation in African Americans

Objective We sought to identify contributors to unstable anticoagulation in African Americans. Patients and methods Sixty African Americans on warfarin were enrolled. Cytochrome P450 2C9 and vitamin K epoxide reductase genotypes and vitamin K intake were assessed, and clinical and dietary data during the 12 months prior to enrollment were collected. Data were compared between stable and unstable patients, classified based on the proportion of international normalized ratio (INR) values outside the therapeutic range. Results The median proportion of out-of-range INRs among study participants was 44%; 28 patients had a higher proportion of INRs out-of-range and were included in the unstable group, with the remaining constituting the stable group. The median (IQR) number of clinic visits/year was higher among unstable versus stable patients [18 (15-22) vs. 16 (13-19); P = 0.03]. Higher warfarin doses, lower adherence, vomiting or diarrhea, and use of antiinfective agents were more common among unstable patients. Genotype was not associated with anticoagulation stability. After regression analysis, only poor adherence and gastrointestinal illness remained predictive of unstable anticoagulation. In a control group of Caucasians of similar age and sex distribution, poor adherence, but not gastrointestinal illness, was associated with unstable anticoagulation. Conclusion We conclude that poor warfarin adherence and gastrointestinal illness are major contributors to unstable anticoagulation in African Americans. Our data suggest that, similar to Caucasians, improving warfarin adherence rates may be an important mean to improve anticoagulation control in African Americans. In addition, close monitoring during acute illness may be particularly important in this population.