Low cellular fibronectin levels are associated with bleeding in patients with atrial fibrillation: a long-term observational study

Abstract Introduction An estimated 5% (ca. 3 mln per year worldwide) of anticoagulated patients with atrial fibrillation (AF) experience clinically relevant bleeding every year while receiving the recommended pharmacotherapy for stroke prevention, and identifying those at high bleeding risk remains challenging. Elevated levels of cellular fibronectin (cFn), a marker of endothelial dysfunction and vascular injury, were found to promote arterial thrombosis and thrombo-inflammation. However, to the best of our knowledge, there have been no studies exploring cFn as a marker of bleeding risk in AF patients. Purpose The purpose of this study was to evaluate a possible association between cFn and bleeding risk in anticoagulated AF patients in long-term follow-up. Methods In this prospective cohort study, we enrolled 185 consecutive AF patients treated with rivaroxaban aged 70.0 (62.0-76.0) years with a median DOAC Score of 4.0 (3.0-6.0). We determined plasma cFn levels, along with demographic, clinical, and laboratory parameters, including fibrin clot permeability (Ks) and fibrinolysis time. Clinically relevant and major bleeding in accordance with the International Society on Thrombosis and Haemostasis (ISTH) were recorded during a median follow-up of 49.0 (46.0-51.0) months. Results The median cFn levels were 3.5 (2.8-4.5 µg/ml). cFn showed a positive correlation with fibrinogen and CRP, and an inverse correlation with Ks. cFn in the lowest quartile (3.0 µg/ml) was associated with a higher bleeding risk, as compared to the top quartile (4.7 µg/ml). On multivariate analysis, low cFn level was an independent predictor of both clinically relevant bleeding and major bleeding according to ISTH. Based on ROC analysis, the optimal cut-off value for baseline cFn to predict bleeding in long-term observation was 3.7 µg/m. A 1 µg/ml decrease in cFn was associated with a 5.1-fold increase in the risk of major ISTH bleeding (OR = 5.1, 95% CI: 1.3–33.8, p = 0.037). A mathematical simulation showed that incorporating cFn into the DOAC Score enhanced its predictive performance for bleeding risk assessment. Conclusions Our results show, for the first time, that low cFn levels are associated with bleeding risk during long-term follow-up in rivaroxaban-treated AF patients, which is likely related to unfavorably modified fibrin clot properties. Our observations suggest that incorporating plasma cFn levels into the DOAC Score could enhance its predictive value, improving bleeding risk stratification in anticoagulated AF patients. However, larger validation studies are needed to confirm cFn utility in routine clinical practice.