Asian Hematology Research Journal

Background: Prevention of stroke and thrombo-embolism in patients with atrial fibrillation(AF) using oral vitamin K antagonists is closely related to risk of bleeding. The HAS-BLED score has been evaluated as a predictor of major bleeding risk in various populations. The present study was conducted to evaluate the usefulness of the HAS-BLED score for predicting 1-year major bleeding risk in a cohort of Sri Lankan patients with non valvular atrial fibrillation on warfarin.

Materials and Methods: This was a retrospective cohort study conducted at anticoagulation outpatient clinics of a tertiary care teaching hospital, which included ninety one patients with non valvular atrial fibrillation on warfarin for at least 18 months. The relevant data were collected using an interviewer administered questionnaire and medical records of the patient. HAS-BLED score was calculated 6 months after the initial diagnosis and start of warfarin, using the retrospective information. The occurrence of major bleeding events during the period of following 01year was documented. The results were analyzed by obtaining a p value through Fisher exact test and using logistic regression analysis to assess the independent risk factors.

Results: Out of 91 patients, 47.25% (43/91) were males and 52.74% (48/91) were females. The median HAS-BLED score was 2. In this cohort, 06(6.6%) were in a low risk, 59(64.8%) were in an intermediate risk and 26(28.6%) were in a high risk of bleeding according to HAS-BLED score. There were 2 major bleeding events in the one year period with one each occurring in intermediate and high risk groups. But, the association was not statistically significant(Fisher exact test is 1.000, P=1.000). Logistic regression analysis did not show any significance of individual components of the score as independent risk factors for bleeding.

Conclusion: The data of our study showed a limited value of HAS-BLED score, as it had no statistically significant relationship with the major bleeding events in patients in our cohort. However, further studies using greater samples are recommended to draw further conclusions in our population.