Abstract 15523: Biomarkers Predict Cause of Death More Accurately Than Clinical Variables in Patients With Atrial Fibrillation on Oral Anticoagulation - Results From the ARISTOTLE Trial
Introduction: Mortality is the most frequent adverse outcome in anticoagulated patients with atrial fibrillation (AF). Several cardiac biomarkers are associated with increased mortality, but the associations between biomarkers and specific cause of death in AF patients have not been investigated.
Methods: In the ARISTOTLE trial, baseline plasma samples were available in 14,798 patients, with median follow up 1.9 years, for measurement of troponin T (cTnT-hs), NT-proBNP, and GDF-15 (cardiovascular); and interleukin-6 (inflammatory). Deaths were centrally adjudicated for specific causes. Multivariable adjusted Fine-Gray subdistribution hazards models, with death from any other cause as competing risk were used.
Results: Of the 1,075 deaths in this cohort, 547 (51%) were cardiovascular (CV). The majority of CV mortality was sudden death, 215 (39%), followed by heart failure 140 (26%), and stroke/systemic embolism/TIA 92 (17%). Fatal bleeding accounted for 31 (2.9%) deaths. Overall, the biomarkers cTnT-hs and NT-proBNP showed the strongest associations with CV and sudden death. NT-proBNP had the strongest association with heart failure death, followed by GDF-15, prior heart failure, and cTnT-hs. The strongest indicators of stroke-related mortality were history of stroke, cTnT-hs, age, and NT-proBNP. Although based on few events, the strongest associations with fatal bleeding were found for GDF-15 and age. For the majority of causes of death, the biomarkers showed stronger associations than clinical variables (Table).
Conclusions: In anticoagulated patients with AF, the majority of deaths were CV, most commonly sudden death and heart failure. Biomarkers indicating CV dysfunction had stronger association with CV, sudden and heart failure death than any clinical variable. GDF-15 levels had the strongest association with fatal bleeding. These results support the potential role of using biomarkers to more accurately assess the risk of adverse outcomes.
Author Disclosures: The Disclosure Block has exceeded its maximum limit. Please call Tech support at (217) 398-1792 for more information.
- © 2016 by American Heart Association, Inc.