Biomarker panels for improved risk prediction and enhanced biological insights in patients with atrial fibrillation

Abstract

Abstract Atrial fibrillation (AF) increases the risk of adverse cardiovascular events, yet the underlying biological mechanisms remain unclear. We evaluate a panel of 12 circulating biomarkers representing diverse pathophysiological pathways in 3817 AF patients to assess their association with adverse cardiovascular outcomes. We identify 5 biomarkers including D-dimer, growth differentiation factor 15 (GDF-15), interleukin-6 (IL-6), N-terminal pro-B-type natriuretic peptide (NT-proBNP), and high-sensitivity troponin T (hsTropT) that independently predict cardiovascular death, stroke, myocardial infarction, and systemic embolism, significantly enhancing predictive accuracy. Additionally, GDF-15, insulin-like growth factor-binding protein-7 (IGFBP-7), NT-proBNP, and hsTropT predict heart failure hospitalization, while GDF-15 and IL-6 are associated with major bleeding events. A biomarker model improves predictive accuracy for stroke and major bleeding compared to established clinical risk scores. Machine learning models incorporating these biomarkers demonstrate consistent improvements in risk stratification across most outcomes. In this work, we show that integrating biomarkers related to myocardial injury, inflammation, oxidative stress, and coagulation into both conventional and machine learning-based models refine prognosis and guide clinical decision-making in AF patients.