Abstract
BACKGROUND: Atrial fibrillation (AF) is a condition with heterogeneous underlying causes, often involving multiple cardiovascular comorbidities. Large-scale studies examining the heterogeneity of patients with AF in the Asian population are limited. OBJECTIVES: The purpose of this study was to identify distinct phenotypic clusters of patients with AF and evaluate their associated risks of ischemic stroke, heart failure hospitalization, cardiovascular mortality, and all-cause mortality. METHODS: We analyzed 5002 adult patients with AF from the National Taiwan University Hospital between 2014 and 2019 using an unsupervised hierarchical cluster analysis based on the CHA(2)DS(2)-VASc score. RESULTS: We identified 4 distinct groups of patients with AF: cluster I included diabetic patients with heart failure preserved ejection fraction as well as chronic kidney disease (CKD); cluster II comprised older patients with low body mass index and pulmonary hypertension; cluster III consisted of patients with metabolic syndrome and atherosclerotic disease; and cluster IV comprised patients with left heart dysfunction, including reduced ejection fraction. Differences in the risk of ischemic stroke across clusters (clusters I, II, and III vs cluster IV) were statistically significant (hazard ratio [HR] 1.87, 95% confidence interval [CI] 1.00-3.48; HR 2.06, 95% CI 1.06-4.01; and HR 1.70, 95% CI 1.02-2.01). Cluster II was independently associated with the highest risk of hospitalization for heart failure (HR 1.19, 95% CI 0.79-1.80), cardiovascular mortality (HR 2.51, 95% CI 1.21-5.22), and overall mortality (HR 2.98, 95% CI 1.21-4.2). CONCLUSION: A data-driven algorithm can identify distinct clusters with unique phenotypes and varying risks of cardiovascular outcomes in patients with AF, enhancing risk stratification beyond the CHA(2)DS(2)-VASc score.