Baseline demographic and clinical characteristics as predictors of adverse outcomes to improve management of patients with AF receiving edoxaban: a subanalysis of the ETNA-AF registry

Abstract Background/Introduction Atrial fibrillation (AF) is associated with substantial morbidity and mortality for patients, and significant burden for societal health and health economy. Even optimal oral anticoagulation and treatment of concomitant conditions leave an unacceptable remaining morbidity and mortality. Aim The purpose of this analysis was to identify clinical parameters associated with cardiovascular events in patients with AF receiving adequate anticoagulation. Methods The Global Edoxaban Treatment in routiNe clinical prActice (ETNA)-AF observational, non-interventional programme prospectively collects data on characteristics and clinical outcomes in patients treated with the direct oral factor Xa inhibitor edoxaban for stroke prevention in AF from Europe, Japan, and other Asian countries. A risk prediction model was built and evaluated for two dependent variables: composite endpoint 1, consisting of cardiovascular (CV) death, all stroke or systemic embolism or transient ischaemic attack (TIA), myocardial infarction (MI), and venous thromboembolism (VTE); and composite endpoint 2, consisting of CV death, all stroke, and MI. Independent variables used to identify predictors of the composite endpoints included the following parameters at baseline: age, history of any stroke/TIA, peripheral arterial disease (PAD), heart failure (HF), coronary artery disease (CAD), major or clinically relevant non-major gastrointestinal bleeding, chronic obstructive pulmonary disease, body mass index <18.5 kg/m2, weight, creatinine clearance (CrCl), history of valvular disease, history of diabetes mellitus, smoking history, age ≥85 years, and hepatic impairment. Multivariable Cox analysis with stepwise selection was used to identify baseline predictors of the composite endpoints. Results Overall, 26,580 edoxaban patients were included in the analysis. At baseline, 3034 (11.4%) patients were ≥85 years. Additional baseline demographics and clinical characteristics are reported in Table 1. All clinical variables were significant predictors (P <0.05) of both composite endpoints. Stronger predictors of composite endpoint 1 included history of any stroke/TIA, PAD, HF, higher weight, and lower CrCl (all P <0.0001; Figure 1). Most significant predictors of composite endpoint 2 included a history of ischaemic stroke, PAD, HF, and weight (P <0.0001; Figure 1). Conclusion A history of ischemic stroke, peripheral artery disease, heart failure, and weight identify patients with atrial fibrillation at risk of cardiovascular events on optimal current therapy. These patients may benefit from targeted interventions including early rhythm control and intensified, integrated management. Additional external validation will be performed to further the additive value of the analysis presented here.Table 1Figure 1