P4805A new score for assessing bleeding risk in patients with atrial fibrillation treated with NOACs

Abstract

Background: Atrial fibrillation (AF) significantly increases the risk of embolic stroke and death. Treatment with oral anticoagulants (OAC) effectively reduces risk of stroke, but this effect comes at a cost of significantly increased risk of bleeding. Non-vitamin K oral anticoagulants (NOACs) are gradually replacing vitamin K antagonists as the drugs of choice. Studies on risk factors for bleeding have mainly been performed on warfarin- treated patients. In the current era, information is needed on bleeding risk factors specifically for patients on NOACs. Purpose: The aim of this study was to identify risk factors for bleeding in patients with atrial fibrillation being treated with NOACs, and to create a simple bedside tool to assess bleeding risk in these patients. Methods: Using nationwide registries (Norwegian Patient Registry and Norwegian Prescription Database), we identified AF patients with a first prescription of a NOAC between January 2013 and June 2015. Patients were followed until discontinuation or switching of oral anticoagulants, death, or end of follow-up (June 30, 2015). The primary endpoint was major or clinically relevant non-major (CRNM) bleeding. Cox proportional hazards analyses were used to identify risk factors for bleeding, and a bleeding score was developed based on the ten strongest risk factors. Results: A total of 21 248 patients were included in the cohort; 7925 were treated with dabigatran, 6817 with rivaroxaban, and 6506 with apixaban. The median age was 73 years and 57.4% of patients were male. After a median follow-up time of 183 days, 1257 (5.9%) patients experienced a major or CRNM bleeding. The strongest prediction model included the variables age, male sex, history of stroke/TIA, history of bleeding, history of anaemia, hypertension, heart failure, non-bleeding related hospitalisation within the last 12 months, chronic kidney disease, and chronic obstructive pulmonary disease, and showed good discriminative ability, with a Harrell's c - index of 0.68. A bleeding risk score was then created with weights proportional to the model coefficients. From our cohort we also calculated a modified HAS-BLED score, which achieved a c - index of 0.59. A simplified score was finally derived from the full score, including only age, history of bleeding, and non-bleeding related hospitalisation within the last 12 months, that reached a Harrell's c - index of 0.66. Conclusions: In this nationwide cohort study of real-life AF patients being prescribed NOACs, a bleeding risk score was created that showed a high c- index, requiring no blood sampling or imaging for its calculation. The simplified version of the bleeding risk score would help clinicians quickly assessing bleeding risk bedside, aiding in planning anticoagulation in patients with AF.