P2700Fatal risk of warfarin in asian very elderly patients with atrial fibrillation in direct oral anti-coagulant era

Abstract

Background: The prevarence rate of atrial fibrillation (AF) is 10% in people over 80 years-old in Japan. For the aging countries such as Japan, AF management including anticagulant therapy is a major concern because of its serious complication and poor prognosis. In particular, Asian elderly patients have a higher bleeding risk. Though direct oral anti-coagulant (DOAC) became available for use in such patients, there is no reliable evidence of not only DOAC but also warfarin (WF). Purpose: To clarify the efficacy and safety of anticoagulant drug in Asian very elderly patients, we evaluated the clinical events, including systemic thromboembolism (SSE), bleeding, and death in very elderly AF patients requiring anticoagulant therapy. Method: We retrospectively investigated consecutive 540 Japanese nonvalvular AF (NVAF) patients who were over 80 years-old and have been prescribed WF or DOAC in the period from March 2011 to October 2016. We compared the incidence of SSE, major bleeding, and either SSE or bleeding related death between patients taking WF (WF group; n=248) and patients taking DOAC (DOAC group; n=292). Result: CHADS-VASc score and HASBLED score were significantly higher in the DOAC group than in the WF group (4.6±1.2 vs. 3.9±1.6, 2.9±1.1 vs. 2.5±0.9, p<0.001, for all). Creatinine-clearance was higher in the DOAC group, but was not statistical signifcant difference between the two group (49±17 vs. 43±20 ml/min, p=0.31). The incidence of SSE, major bleeding, and death were significantly higher in the WF group than in the DOAC group [25 (10.1%) vs. 5 (1.7%), P<0.001, 29 (11.7%) vs. 11 (3.8%), P=0.005, 13 (5.2%) vs. 1 (0.4%), p<0.001, respectively]. Multivariate analysis for all patients revealed that predictors of SSE were WF intake (p<0.001), and prior SSE (p=0.01), and that predictor of either major bleeding or death was only WF intake (p<0.001, p=0.02, respectively). Multivariate analysis for each group patients revealed that only prior SSE was determined as predictor of SSE in both group (p=0.03, p=0.01, respectively), and that anti-platelet drug intake was determined as predictor of major bleeding only in the DOAC group (p=0.02). In contrast, no predictor of either major bleeding or death was determined in the WF group. Conclusion: WF was continued to use frequently in DOAC era, and had a higher fatal risk despite it being used in lower risk patients. Moreover, WF had not predictive factors of the events, except for prior SSE. These findings suggest that WF should be avoided in AF patients over 80 years-old.