A European-Multicenter Network for the Implementation of Artificial Intelligence to Manage Complexity and Comorbidities of Atrial Fibrillation Patients: The ARISTOTELES Consortium

作者：Giuseppe Boriani, Davide Antonio Mei, Gregory Y.H. Lip, on behalf of the ARISTOTELES consortium · 发表于：Thrombosis and Haemostasis · 年份：2025 · DOI：10.1055/a-2508-5708 · 被引用次数：18 · 研究领域：Atrial Fibrillation Management and Outcomes、Acute Ischemic Stroke Management、Blood Pressure and Hypertension Studies

Introduction Atrial fibrillation (AF) is the most common arrhythmia worldwide, contributing significantly to morbidity, healthcare costs, and resource utilization.[ 1 ] Patients with AF face a higher mortality and morbidity from stroke, heart failure, dementia, and hospitalizations.[ 1 ] Oral anticoagulants (OACs) are the cornerstone of AF management, as they substantially reduce the risk of stroke and mortality.[ 2 ] Nevertheless, some residual risk still remains despite anticoagulation, with most AF-related mortality linked to cardiovascular causes and comorbidities rather than stroke alone.[ 2 ] [ 3 ] AF is not a yes/no homogeneous diagnosis. AF patients are often elderly, multimorbid, and frail, with associated polypharmacy, leading to “clinically complex” phenotypes or clusters. As comorbidities often cluster in different patterns, these impact on the risk of adverse outcomes and management. In the prospective GLORIA-AF registry of AF patients, the presence of clinical complexity was associated with lower odds of being prescribed with OAC (odds ratio [OR] 0.50, 95% confidence interval [CI] 0.44–0.57), higher OAC discontinuation, and with a higher risk of adverse events (hazard ratio [HR] 1.63, 95% CI 1.43–1.86).[ 4 ] Indeed, “high clinical complexity” patients defined using latent class analysis constituted 6.6% of AF patients, and was associated with higher hazards of experiencing the primary composite outcome of all-cause death and major adverse cardiovascular events (...