Does von Willebrand factor improve the predictive ability of current risk stratification scores in patients with atrial fibrillation?

Amaya García-Fernández,Francisco Marín,Vanessa Roldán,José Miguel Rivera-Caravaca,Mariano Valdés,Diana Hernández-Romero,Vicente Vicente,Gregory Y H Lip

doi:10.1038/srep41565

Abstract

Von Willebrand factor (vWF) is a biomarker of endothelial dysfunction. We investigated its role on prognosis in anticoagulated atrial fibrillation (AF) patients and determined whether its addition to clinical risk stratification schemes improved event-risk prediction. Consecutive outpatients with non-valvular AF were recruited and rates of thrombotic/cardiovascular events, major bleeding and mortality were recorded. The effect of vWF on prognosis was calculated using a Cox regression model. Improvements in predictive accuracy over current scores were determined by calculating the integrated discrimination improvement (IDI), net reclassification improvement (NRI), comparison of receiver-operator characteristic (ROC) curves and Decision Curve Analysis (DCA). 1215 patients (49% males, age 76 (71–81) years) were included. Follow-up was almost 7 years. Significant associations were found between vWF and cardiovascular events, stroke, mortality and bleeding. Based on IDI and NRI, addition of vWF to CHA2DS2-VASc statistically improved its predictive value, but c-indexes were not significantly different. For major bleeding, the addition of vWF to HAS-BLED improved the c-index but not IDI or NRI. DCA showed minimal net benefit. vWF acts as a simple prognostic biomarker in AF and, whilst its addition to current scores statistically improves prediction for some endpoints, absolute changes and impact on clinical decision-making are marginal.

Highlights

Atrial fibrillation (AF) is increasingly more common and confers a five-fold increase in the risk of stroke[1]
We have previously shown that high plasma von Willebrand factor (vWF) predicts adverse cardiovascular events, mortality and major bleeding in anticoagulated AF patients[25]
We constructed a Receiver-operator characteristic (ROC) curve for vWF levels, that gave a median cut-off point of 190 UI/dL [areas under the two ROC curves (AUC): 0.60 (95%CI: 0.56–0.64); p < 0.001] for the composite cardiovascular end-point; 194 UI/dL [AUC: 0.60 (95%CI: 0.55–0.65); p < 0.001] for stroke; 184 UI/dL [AUC: 0.62 (95%CI: 0.59–0.65); p < 0.001] for total mortality; 184 UI/dL [AUC: 0.64 (95%CI: 0.57–0.71); p < 0.001] for cardiovascular mortality and 197 UI/dL [AUC: 0.61 (95%CI: 0.57–0.65); p < 0.001] for major bleeding

Summary

Introduction

Atrial fibrillation (AF) is increasingly more common and confers a five-fold increase in the risk of stroke[1]. Guidelines[6,7] recommend the use of the CHA2DS2-VASc score [Congestive heart failure, Hypertension, Age ≥ 75 years, Diabetes mellitus, Stroke, Vascular disease, Age 65 to 64 years, Sex category], as a simple, clinical risk factor-based approach to thromboprophylaxis[8]. This score has been validated in different cohorts and we have previously demonstrated that it is predictive for vascular events and mortality in AF9. The predictive value of the CHA2DS2-VASc and other clinical factor-based risk stratification schemes for identifying ‘high risk’ patients that develop events remains modest[10]. Continuous efforts have been made to improve stroke and bleeding risk stratification in AF and various studies point out to a promising role of cardiac biomarkers to further refine these risks[17,18]

Methods

Results

Conclusion