Bleeding in Patients with Atrial Fibrillation Treated with Non Vitamin K Antagonist Oral Anticoagulants: A Population-Based Study

INTRODUCTION Recent large randomized controlled trials (RCTs) have shown non-vitamin K antagonist oral anticoagulants (NOACs) are at least as effective as vitamin K antagonists (VKAs) for prevention of stroke or systemic embolism in patients with non-valvular atrial fibrillation (AF) and are associated with similar or lower rates of bleeding. The results for bleeding seen in Phase 3 trials might not be applicable to real world practice. We performed a large population-based study to determine the incidence of bleeding in patients with AF beginning treatment with dabigatran, rivaroxaban or a VKA. METHODS From the computerized database of the 4.5 million member Israeli Clalit Health Services health care provider, consecutive patients initiating a VKA or NOAC for AF between January 1, 2011 and December 31, 2013 were studied. For prevention of embolism in AF, dabigatran had regulatory approval for 36 months and rivaroxaban for 24 months. Bleeding patients who required hospitalization were identified and key clinical and laboratory data were recorded. Because patients received different anticoagulants for different durations, time-to-event analyses were performed and bleeding incidences were calculated and reported as rates per 100 patient-years of treatment. Bleeding sites and all-cause mortality within 30 days were recorded and case fatality rates were calculated as the proportions of bleeding patients who died within 30 days. RESULTS (Table1) 18249 patients initiated anticoagulants for AF: 9564 received VKA, 4170 received dabigatran 110 mg bid , 1806 received dabigatran 150 mg bid and 2709 received rivaroxaban. The bleeding rates per 100 patient-years were 3.9 in VKA-treated patients, 2.8 in dabigatran 150 mg patients, 4.6 in dabigatran 110 mg patients and 4.3 in rivaroxaban patients. The intracranial hemorrhage (ICH) rates per 100 patient-years were 0.70 in VKA-treated patients, 0.37 in dabigatran 150 mg patients, 0.49 in dabigatran 110 mg patients and 0.27 in rivaroxaban patients. The gastrointestinal (GI) hemorrhage rates per 100 patient-years were 1.88 in VKA-treated patients, 1.85 in dabigatran 150 mg patients, 3.36 in dabigatran 110 mg patients and 2.39 in rivaroxaban patients. The case fatality rate for any bleed was 21%; for ICH 28.8%, and for GI bleeds it was 11.1%. Multivariate analysis revealed that increased age and increased serum creatinine were risk factors for bleeding in NOAC-treated patients. CONCLUSIONS The results of our population-based non-randomized study of AF patients are consistent with the RCTs in showing similar rates of overall bleeding, an increase in GI bleeding associated with dabigatran and a reduction in ICH seen with both dabigatran and rivaroxaban. Table 1: Clinical profile of patient cohort VKA Dabigatran 150 mg Dabigatran 110 mg Rivaroxaban Overall Number of patients 9564 1806 4170 2709 18249 Age in years) Median (Range) 79 (27-99) 78 (52-89) 82 (55-95) 82 (58-91) 80 (27-99) Women % 43.8 45.1 47 38.6 43.7 Serum creatinine mg/dL Median (Range) 1.2 (0.3-11.6) 1.0 (0.5-4.4) 1.2 (0.4-4.1) 1.3 (0.5-3.5) 1.2 (0.3-11.6) CHADS2 score Median (Range) 3 (0-6) 3 (1-6) 4 (1-6) 4 (2-6) 3 (0-6) Anti aggregant use (%) 52 50 35 55 48 Bleeds per 100 patient years (N) 3.9 (372) 2.8 (50) 4.6 (191) 4.3 (116) (729) Fatalities within 1 month of hemorrhage N 44 8 15 3 70 Intracranial hemorrhage N 67 4 16 3 90 Gastrointestinal hemorrhage N 178 20 108 26 332 Disclosures Ellis: Boehringer Ingelheim: Honoraria. Eikelboom:Bayer: Honoraria, Research Funding; Bristol Meyers Squibb: Honoraria, Research Funding; Boehringer Ingelheim: Honoraria, Research Funding; Janssen: Honoraria, Research Funding; Daiichi Sankyo: Honoraria, Research Funding.