OBJECTIVES : To evaluate the effectiveness of edoxaban compared to warfarin in prevention of stroke and systemic embolic events (SEE) using observational healthcare data on patients with atrial fibrillation in a general practice setting.

METHODS : Using IBM EXPLORYS^® Electronic Healthcare Record (EHR) data (~55 million unique patient records), patient cohorts with conditions of non-valvular atrial fibrillation (NVAF) with no pre-existing deep vein thrombosis (DVT), pulmonary embolism (PE), or mitral valve stenosis were constructed. The cohorts were categorized into a new-user cohort of edoxaban (target) and a warfarin users cohort (comparator). Propensity score matching using patient baseline covariates including patient demographics, medical history including medication and surgical history was performed. The matched cohorts were checked for residual bias. Incidence rates of stroke or SEE were measured as primary outcomes for matched target and comparator cohort. Cox proportional hazard models were applied to estimate the risk of stroke and SEE associated with edoxaban and warfarin users.

RESULTS : Within NVAF adult patients (N = 1,012,245), we identified 744 qualified edoxaban users and 549 propensity score matched warfarin users between Jan. 2015 and Dec. 2017. Compared to warfarin, edoxaban is associated with similar risks in stroke (hazard ratio [HR] 0.89, 95% CI 0.58- 1.39, p = 0.605), and with lower risks in systemic embolic events (hazard ratio [HR] 0.58, 95% CI 0.37- 0.90, p = 0.014)

CONCLUSIONS : The edoxaban is non-inferior in prevention of stroke but reduces the risk of embolism in comparison to warfarin in our studies using EHR data.