Clinical predictors of warfarin-associated bleeding: A case-control study.

Abstract

Despite the advent of directly acting oral anticoagulants, warfarin is still widely used in resource-poor settings. Bleeding as a result of warfarin overdosage is common and risk factors seem to vary between patient populations. Predictors of bleeding in patients using warfarin were studied using a case-control design. We calculated the Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly, Drugs/Alcohol (HAS-BLED) score (a bleeding risk prediction tool). Cases chosen were adults prescribed warfarin presenting with bleeding and controls were likewise taking warfarin but had no bleeding in the previous six months. The most common indication for warfarin use was venous thromboembolism. Recent decrease in food intake was strongly associated with bleeding. Older age, higher warfarin doses, recent illness in the past fortnight and prior bleeding were also independent predictors. A higher HAS-BLED score was not predictive independently. Closer monitoring of international normalised ratio is therefore advisable during acute illnesses and in those with decreased food intake.