Phenotypic clustering of patients with atrial fibrillation at high risk of bleeding and stroke: insights from three global registries

Patients with atrial fibrillation (AF) who are at increased risk of both thrombotic and haemorrhagic events represent a heterogeneous and broad clinical entity, characterised by the presence of several Janus-faced risk factors. However, how these risk factors tend to cluster into distinct clinical phenotypes remains largely unknown.

In this observational study, using three prospective registries from Europe (EORP-AF), East Asia (APHRS_AF) and Middle East (IRAF-AF), we selected only AF patients who had HAS-BLED score ≥ 3 and CHA2DS2-VASc score ≥2. Hierarchical clustering method using Ward’s linkage criterion was applied to identify phenotypic subgroups. Univariable and multivariable logistic regression were employed to assess the one-year risk of all-cause death, thromboembolic events, major adverse cardiovascular events (MACE), acute coronary syndrome, and major bleeding across identified clusters.

A total of 2535 patients (mean age = 75.4 ± 7.8, 58.3% male) were enrolled. Three different clusters were identified: Cluster I: characterised by the highest proportion of prior thromboembolic events (43.1%) and prior haemorrhagic events (25.0%) along with the lowest prevalence of diabetes (18.8%), coronary artery disease (16.8%), heart failure (6.0%), COPD (6.2%), and peripheral artery disease (5.0%); Cluster II: the oldest patients (76.7 ± 7.6 years), with the highest prevalence of dyslipidaemia (55.5%), heart failure (67.8%), and peripheral artery disease (15.7%); Cluster III: The youngest group (72.1 ± 7.2 years) with the highest BMI (28.2 ± 4.8) and the highest alcohol use (77.9%).

Cluster II (aOR 1.99, 95%CI 1.42-2.85), and Cluster III (aOR 1.62, 95%CI 1.08-2.45) were associated with a higher risk of all-cause death compared to cluster I. Cluster II (aOR 1.48, 95%CI 1.03-2.15) and Cluster III (aOR 1.67, 95%CI 1.10-2.55) demonstrated a higher risk of MACE compared to cluster I.

Three distinct phenotypic clusters were identified among AF patients at high risk of both bleeding and thromboembolic events. These clusters differed markedly in clinical characteristics and outcomes, underscoring the need for more tailored management strategies in this complex patient population.Table 1Table 2