Multiplex protein screening of biomarkers associated with major bleeding in patients with atrial fibrillation treated with oral anticoagulation

• Published in: Journal of thrombosis and haemostasis 2021-11, Vol.19 (11), p.2726-2737
• Main Authors: Siegbahn, Agneta, Lindbäck, Johan, Hijazi, Ziad, Åberg, Mikael, Alexander, John H., Eikelboom, John W., Lopes, Renato D., Pol, Tymon, Oldgren, Jonas, Granger, Christopher B., Yusuf, Salim, Wallentin, Lars
• Format: Article
• Language: English
• Subjects: Administration, Oral; anticoagulants; Anticoagulants - adverse effects; Apoptosis; atrial fibrillation; Atrial Fibrillation - complications; Atrial Fibrillation - diagnosis; Atrial Fibrillation - drug therapy; Biomarkers; Bleeding; Calcification (ectopic); Cardiac arrhythmia; Fibrillation; Fibrinolysis; Hemorrhage - chemically induced; Hemorrhage - diagnosis; Hemorrhage - drug therapy; hemorrhages; Humans; Inflammation; Osteoprotegerin; Oxidative stress; proteomics; Replication; Risk Factors; Stroke - drug therapy; Thromboembolism; Thromboembolism - diagnosis; Thromboembolism - drug therapy; Tumor necrosis factor
• ISSN: 1538-7933; 1538-7836; 1538-7836
• DOI: 10.1111/jth.15498

Background Oral anticoagulants (OAC) in patients with atrial fibrillation (AF) prevent thromboembolic events, but are associated with significant risk of bleeding. Objectives To explore associations between a wide range of biomarkers and bleeding risk in patients with AF on OAC. Method Biomarkers were analyzed in a random sample of 4200 patients, 204 cases with major bleedings, from ARISTOTLE. The replication cohort included 344 cases with major bleeding and 1024 random controls from RE‐LY. Plasma samples obtained at randomization were analyzed by the Olink Proximity Extension Assay cardiovascular and inflammation panels and conventional immunoassays. The associations between biomarker levels and major bleeding over 1 to 3 years of follow‐up were evaluated by random survival forest/Boruta analyses and Cox regression analyses to assess linear associations and hazard ratios for identified biomarkers. Results Out of 268 proteins, nine biomarkers were independently associated with bleeding in both cohorts. In the replication cohort the linear hazard ratios (95% confidence intervals) per interquartile range were for these biomarkers: TNF‐R1 1.748 (1.456, 2.098), GDF‐15 1.653 (1.377, 1.985), EphB4 1.575 (1.320, 1.880), suPAR 1.548 (1.294, 1.851), OPN 1.476 (1.240, 1.757), OPG 1.397 (1.156, 1.688), TNF‐R2 1.360 (1.144,1.616), cTnT‐hs 1.232 (1.067, 1.423), and TRAIL‐R2 1.202 (1.069, 1.351). Conclusions In patients with AF on OAC, GDF‐15, cTnT‐hs, and seven novel biomarkers were independently associated with major bleedings and reflect pathophysiologic processes of inflammation, apoptosis, oxidative stress, vascular calcification, coagulation, and fibrinolysis. Investigations of the utility of these markers to refine risk stratification and guide the management of patients at high risk of bleeding are warranted.