LAUSR

Direct oral anticoagulants (DOACs) are indicated for the prevention of stroke in non-valvular atrial fibrillation (NVAF). Although FDA labeling for DOACs utilizes estimated creatinine clearance (CrCL) according to the Cockcroft-Gault (C-G) equation, estimated glomerular filtration rate (eGFR) according to the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation is often reported. The objectives of this study were to evaluate DOAC dosing discordance and to determine whether discordance based on various estimates of kidney function is associated with bleeding or thromboembolism. The study was an IRB-approved retrospective analysis of patients at UPMC Presbyterian Hospital from 01/01/2010-12/31/2016. Data were obtained through electronic medical records. Adults who received a medication charge for rivaroxaban or dabigatran, had a diagnosis code for atrial fibrillation, and a serum creatinine (SCr) within 3 days of DOAC initiation were included. Doses were considered discordant if the calculated dose based on CKD-EPI did not match the patient's dose during index admission, if dosed correctly utilizing C-G. Association of discordance with dabigatran, rivaroxaban, and clinical outcomes was determined using odds ratios and 95% confidence intervals. Rivaroxaban discordance was present among 49 of the 644 (8%) patients that were dosed correctly with C-G. Dabigatran discordance was present among 17 of the 590 (3%) patients that were dosed correctly. Discordance with rivaroxaban was found to increase the risk of thromboembolism when utilizing CKD-EPI (OR = 2.83; 95% CI, 1.02 to 7.79, p = 0.045) versus C-G. Our findings emphasize the need to dose DOACs, specifically rivaroxaban, appropriately in patients with NVAF. This article is protected by copyright. All rights reserved.