Access to the PDF text

Free Article !

Archives of cardiovascular diseases
Volume 111, n° 5
pages 357-369 (mai 2018)
Doi : 10.1016/j.acvd.2017.08.004
Received : 6 April 2017 ;  accepted : 31 August 2017
Clinical research

Is von Willebrand factor associated with stroke and death at mid-term in patients with non-valvular atrial fibrillation?
Le facteur von Willebrand est-il associé à l’incidence d’infarctus cérébral et de décès à moyen terme chez les patients avec une fibrillation atriale non valvulaire ?

Yann Ancedy a, b, Emmanuelle Berthelot a, b, Sylvie Lang a, b, Stéphane Ederhy a, b, Louise Boyer-Chatenet a, b, Emanuele Di Angelantonio c, Laurie Soulat-Dufour a, b, Arnaud Etienney a, b, Saroumadi Adavane-Scheublé a, b, Franck Boccara b, c, d, Ariel Cohen a, b, d, e,
a Service de cardiologie, hôpital Saint-Antoine, hôpitaux de l’Est parisien, AP–HP, 75012 Paris cedex 12, France 
b Université Pierre-et-Marie-Curie (UPMC), Sorbonne universités, 75005 Paris, France 
c Department of Public Health & Primary Care, Strangeways Research Laboratory, CB1 8RN Cambridge, UK 
d Centre de Recherche Saint-Antoine, Inserm, UMR S 938, 75012 Paris, France 
e Inserm, U856, « Thrombose, Athérothrombose et Pharmacologie Appliquée », 75012 Paris, France 

Corresponding author. Service de cardiologie, hôpital Saint-Antoine, AP–HP, 184, rue du Faubourg-Saint-Antoine, 75012 Paris cedex 12, France.Service de cardiologie, hôpital Saint-Antoine, AP–HP, 184, rue du Faubourg-Saint-Antoine, 75012 Paris cedex 12, France.

Heart failure and atrial fibrillation share common mechanisms that may contribute to hypercoagulability and thrombotic risk. Elevated von Willebrand factor (vWF) concentration has been associated with increased risk of thromboembolism and cardiovascular events.


To investigate whether increased vWF plasma concentration predicts occurrence of a composite endpoint (all-cause death and stroke) in patients with non-valvular atrial fibrillation (NVAF).


We prospectively studied 122 patients (mean age 70±14years; 46% men) hospitalized with NVAF, and followed over a median (interquartile range) of 5.4 (2.3–9.0)years. Cox proportional models were used to estimate the association of vWF concentration with time to stroke and death.


Forty-three patients (35%) had at least a stroke or died during the 5-year follow-up. Kaplan-Meier curves using vWF plasma concentration tertiles (≤191IU/dL;>191 to295IU/dL;>295IU/dL) showed that vWF plasma concentrations discriminated groups of patients with higher cardiovascular event rates (log-rank P =0.01). In the multivariable analysis, higher vWF concentrations (middle tertile hazard ratio [HR] 4.59, 95% confidence interval [CI] 1.55–13.50 [P =0.006]; upper tertile HR 4.10, 95% CI 1.43–11.75 [P =0.009]), age75years (HR 5.02, 95% CI 1.53–16.49; P =0.008), heart failure (HR 2.05, 1.01–4.19; P =0.048), C-reactive protein, log2 per unit increase (HR 1.29, 95% CI 1.04–1.61; P =0.021), no warfarin at discharge (HR 4.96, 95% CI 2.02–12.20; P <0.0001) and no aspirin at discharge (HR 4.41, 95% CI 1.71–11.97; P =0.002) were independently associated with an increased risk of stroke and all-cause death, whereas female sex was a protective factor (HR 0.35, 0.16–0.78; P =0.01).


High vWF plasma concentrations may discriminate patients with NVAF at greater risk of stroke or all-cause death.

The full text of this article is available in PDF format.

L’insuffisance cardiaque et la fibrillation atriale (FA) partagent des mécanismes communs qui contribueraient à l’hypercoagulabilité et au risque thrombotique. L’élévation du facteur von Willebrand (vWF) a été associée à une augmentation du risque thromboembolique et de l’incidence d’évènement cardiovasculaire.


L’objectif de l’étude était de déterminer si l’augmentation du taux plasmatique de vWF pouvait prédire un critère primaire composite associant : mortalité toute cause et infarctus cérébrale (IC) chez des patients en FA non valvulaire.


Nous avons évalué prospectivement 122 patients (46 % sexe masculin ; age moyen 70±14ans) hospitalisé pour FA et suivi pendant une durée médiane (interquartile) de 5,4 (2,3–9,0) années. Un modèle proportionnel de Cox a été utilisé pour évaluer l’association entre le taux plasmatique de vWF et le délai de survenu d’un IC ou décès.


Quarante-trois patients (35 %) ont présenté un IC ou sont décédé au cours du suivi de 5ans. L’analyse des courbes de Kaplan-Meier selon les tertiles du taux plasmatique de vWF (≤191UI/dL ; >191 et295UI/dL ;>295UI/dL) a montré que le taux plasmatique de vWF permettait de discerner les groupes de patients avec une incidence augmentée d’évènements cardiovasculaires (log-rank p <0,01). En analyse multivariée, les taux plasmatiques élevées de vWF (tertile moyen hazard ratio [HR] 4,59, intervalle de confiance à 95 % [IC95 %] 1,55–13,60 [p =0,006] ; tertile supérieur HR 4,10, IC95 % 1,43–11,75 [p =0,009]), age75ans (HR 5,02, IC95 % 1,53–16,49 ; p =0,008), l’insuffisance cardiaque (HR 2,05, IC95 % 1,01–4,19 ; p =0,048) le taux de CRP, log2  par unités d’augmentation (HR 1,29, IC95 % 1,04–1,61 ; p =0,021), l’absence de warfarin à la sortie d’hospitalisation (HR 4,96, IC95 % 2,02–12,20 ; p <0,0001) et l’absence d’aspirine à la sortie d’hospitalisation (HR 4,41, IC9 5 % 1,71–11,97 ; p =0,002) étaient associés de manière indépendante à une augmentation du risque d’IC et de mortalité toute cause alors que le sexe féminin était un facteur protecteur (HR 0,35, IC95 % 0,16–0,78 ; p =0,01).


Le taux plasmatique élevé de vWF pourrait permettre de discerner les patients en FA avec un risque élevé d’IC et de mortalité toute cause.

The full text of this article is available in PDF format.

Keywords : Non-valvular atrial fibrillation, Prognosis, Von Willebrand factor, Stroke, Cardiovascular events

Mots clés : Fibrillation atriale non valvulaire, Pronostic, Facteur von Willebrand, Infarctus cérébral, Évènements cardiovasculaires

Abbreviations : AF, CHADS2 , CHA2 DS2 -VASc, LVEF, NVAF, OAC, vWF


Atrial fibrillation (AF) is the most common sustained arrhythmia. The estimated global number of people with AF in 2010 was 33.5 million [1]. Patients with AF who have a stroke are at higher risk of death and greater disability than those without AF [2, 3]. The presence of AF confers a hypercoagulable state, with abnormalities in haemostasis, thrombosis and platelet function [4, 5]. The beneficial role of oral anticoagulant therapy (OAC) in reducing the risk of stroke and thromboembolic events in non-valvular AF (NVAF) has been confirmed in large-scale studies [6, 7].

Risk scores such as CHADS2 (congestive heart failure, hypertension, age, diabetes, stroke/transient ischaemic attack/thromboembolism [doubled]) [8] and CHA2 DS2 -VASc (congestive heart failure, hypertension, age75years [doubled], diabetes, stroke/transient ischaemic attack/thromboembolism [doubled]–vascular disease, age 65–74years and sex category [female]) [9], among many others, were developed with the aim of stratifying patients according to their level of thromboembolic risk, and are widely used. Compared with the CHADS2 score, the CHA2 DS2 -VASc score provides better risk stratification of “truly” low-risk patients [10].

Numerous studies have identified plasma markers of coagulation, endothelial function and platelet activation, which may be of interest in understanding the pathophysiology of thrombogenesis [5]. Von Willebrand factor (vWF), a plasma glycoprotein, has an important role in haemostasis, promoting platelet adhesion and aggregation at sites of vascular injury. vWF release is increased when endothelial cells are activated or damaged. Elevated vWF plasma concentrations are considered a marker of endothelial dysfunction, a condition that predisposes to atherosclerosis and thrombosis [5]. Increased concentrations of vWF are found in patients with cardiovascular disease, and particularly in AF [11]. Higher vWF concentration is a marker of increased risk of stroke in NVAF [12, 13], and of myocardial infarction in vascular disease [14, 15]. It is also elevated in patients with congestive heart failure and AF [11, 16, 17, 18, 19] or who are in sinus rhythm [20], suggesting that it may be linked to poor prognosis in heart failure. Several studies have investigated vWF plasma concentrations for predicting cardiovascular events in anticoagulated patients with AF [21], and their results are discordant [21, 22, 23]. It is not clear whether vWF is an independent physiological factor in cardiovascular events or a measure of the severity of known cardiovascular predictors.

The aim of this analysis from the ‘Atrial fibrillation and flutter and thromboembolism” (AFFLUATE) study [24] was to determine whether increased plasma concentrations of vWF could independently predict a composite endpoint, comprising all-cause death and stroke, in a cohort of patients with NVAF. We also sought to clarify whether this biomarker is influenced by, or is independent of, other known predictors of cardiovascular events.


The AFFLUATE study is a prospective cohort study designed to evaluate predictors of stroke in patients with atrial arrhythmia [24]. For the purpose of this analysis, patients with infection, chronic or systemic illness, including renal failure, hepatic impairment, cancer, inflammatory connective tissue disease and inflammatory bowel disease, and AF caused primarily by valvular or pulmonary disease were excluded. From 2002 to 2004, vWF concentrations were determined at baseline among consecutive patients hospitalized with NVAF and recruited into the AFFLUATE study [24].

Clinical evaluation and echocardiography

Data on medical history, baseline risk factors, physical examination, electrocardiography and echocardiography were recorded routinely for all patients at admission to hospital. Hypertension was considered present if the systolic blood pressure was140mmHg or the diastolic blood pressure was90mmHg or if the patient was receiving antihypertensive treatment. Diabetes was defined as a fasting blood glucose >7.0mmol/L or if the patient was being treated with insulin or an oral hypoglycaemic agent. Dyslipidaemia was defined as low-density lipoprotein cholesterol >5.2mmol/L and/or triglycerides >1.7mmol/L and/or high-density lipoprotein cholesterol <1.0mmol/L or if the patient was being treated for dyslipidaemia. Heart failure was defined as congestive heart failure or left ventricular dysfunction with a left ventricular ejection fraction (LVEF)<40% without congestive symptoms. As patients were enrolled between 2002 and 2004, before publication of the CHA2 DS2 -VASc score [9], their scores were calculated retrospectively to provide an estimate of clinical thromboembolic risk. The decision about whether to initiate OAC therapy was taken by the physician in charge of the patient. AF management was based on the European Society of Cardiology guidelines at the time of inclusion [25].

Transthoracic echocardiography was performed in all subjects using the Vivid™ 7 (General Electric, Horten, Norway) and the ACUSON Sequoia™ C512 (Siemens Healthcare, Mountain View, CA, USA) ultrasound machines. Studies were performed and measurements were taken according to the guidelines of the American Society of Echocardiography [26]. In particular, left atrial area, right atrial area and LVEF (according to Simpson's biplane method) were measured using apical views. AF was considered paroxysmal if it lasted<48hours; AF was considered persistent if medical or electrical cardioversion was successful; and AF was considered permanent if the cardioversion failed or was not tried, and the patient remained in AF at discharge [25]. Data were collected on treatments given at discharge. The ethics committee of Saint-Antoine University Hospital approved the study. All patients provided informed consent to participate.

Laboratory tests

Venous blood samples were obtained after a 12-hour overnight fast for measurement of clinical chemistry profiles and haemostatic variables. Blood samples were drawn atraumatically and without stasis into syringes preloaded with sodium citrate. A successful blood-sample collection was one in which no more than one puncture was necessary. The platelet-poor plasma fraction was obtained from venous blood by centrifugation at 3000rpm (1000g) for 20minutes at 4°C. Aliquots of plasma were stored at 70°C to allow for batch analysis. vWF plasma concentrations were measured by an established enzyme-linked immunosorbent assay (Dako, Glostrup, Denmark). The unit for reference vWF is international units per decilitre (IU/dL), which was standardized. The coefficients of variation for all clinical chemistry variables were 5%. All laboratory work was performed in a blinded fashion with regard to the origin of the samples.

Patient follow-up and study endpoints

All patients were followed up, and the occurrence of stroke, a new episode of AF, heart failure, acute coronary syndrome and all-cause death was documented. Clinical follow-up after admission for NVAF was performed via patients’ medical records obtained from hospitals or referring doctors, and through structured telephone contacts. At the time of each follow-up, data pertaining to the patient's clinical status, interim occurrence of any adverse events and anticoagulant, antiplatelet and antiarrhythmic drug therapy were collected.

The study endpoint was a composite of all-cause death and stroke. A secondary endpoint was a composite of death, stroke and heart failure. Only the first event occurring within the first 5years after inclusion into the study was taken into account.

Statistical analysis

Data are presented as mean±standard deviation, median [interquartile range] or count and percentage, as appropriate. Event frequencies were compared using the χ2 test. Other comparisons between two groups of data were made using the unpaired Student's t test or the Mann-Whitney U test, as appropriate.

The aim was to identify whether plasma vWF concentrations were an independent predictor of stroke and all-cause death. Plasma vWF concentrations were defined as a categorical variable according to tertile concentration at the time of enrolment (lower tertile:191IU/dL; middle tertile:>191 to295IU/dL; upper tertile:>295IU/dL). Survival curves were computed using the Kaplan-Meier method, and were compared between groups using the log-rank test.

Cox proportional hazards models were used to determine the association of vWF concentrations with time to the composite endpoint. We used univariate analyses to study patient characteristics previously identified as predictive of the primary composite endpoint of stroke or all-cause death, and the secondary composite endpoint of stroke or all-cause death or heart failure (factors defining the CHA2 DS2 -VASc score [age, sex, hypertension, diabetes, cardiac failure or dysfunction, stroke, vascular disease], dyslipidaemia, smoking status, chronic kidney disease, AF type, C-reactive protein after log2 transformation to obtain a distribution close to a normal distribution, echocardiographic variables and treatments at discharge). Variables with a P value<0.2 were included in the multivariable analysis, except for those defining the CHA2 DS2 -VASc score, which were kept in the model regardless of the P value, as these variables are considered to potentially increase the risk of stroke.

The vWF concentration with the best sensitivity and specificity for the endpoint (stroke and all-cause death) was chosen as cut-off. A new risk-stratification scheme was created by adding one point to the CHA2 DS2 -VASc risk score when the vWF concentration was204IU/dL. To compare the predictive ability of risk-stratification schemes before and after adding vWF concentrations to the CHA2 DS2 -VASc risk score, the statistical significance of the difference between the areas under the two receiver operating characteristic curves were calculated. Also, to test the predictive accuracy of the models, C-indexes were calculated and compared.

A P value of<0.05 was considered statistically significant. All analyses were performed using Stata® 12 statistical software (StataCorp, College Station, TX, USA).


Among 128 patients hospitalized with NVAF, six were lost to follow-up in the 6 months after discharge. Thus, we prospectively followed 122 patients (mean age 70±14years; 46% of whom were men) for a median [interquartile range] of 5.4 [2.3–9.0]years. Of these 122 patients, three were lost to follow-up after 2years, four after 3years and two after 4years of follow-up. Patient baseline clinical characteristics and selected treatments at discharge from hospital, overall and according to vWF plasma concentration tertiles, are depicted in Table 1.

Patients in the middle or upper tertile for vWF plasma concentrations, versus those in the lower tertile, presented a higher prevalence of hypertension (P <0.017), congestive heart failure (P <0.0001), chronic kidney disease (P =0.003), non-paroxysmal AF (P =0.008), higher median CHA2 DS2 -VASc scores (P =0.0002), LVEF<40% (P =0.03) and a higher median C-reactive protein concentration (P =0.0001). At discharge from hospital, the rate of anticoagulation using warfarin was not significantly different between the three groups according to vWF plasma concentration tertile (P =0.06). However, patients in the upper vWF plasma concentration tertiles were more likely than those in the bottom tertile to be given aspirin (P =0.027), but were less likely to be on an antiarrhythmic drug (P =0.005) (Table 1).

Factors predictive of primary composite endpoint (stroke or all-cause death)

During the 5-year follow-up period, 43 patients (35%) had at least a stroke or died (Table 2). Patients in the upper and middle vWF plasma concentration tertiles had a higher rate of the primary composite endpoint (44% and 50%, respectively) than those in the lower tertile (12%).

When considering the composite endpoint of all-cause death or stroke, the factors associated with an increased risk of events were the upper and middle vWF plasma concentration tertiles (respectively, P =0.006 and P =0.009), age75years (P =0.008), cardiac failure or dysfunction (P =0.048), C-reactive protein, log2 per unit increase (P =0.021), no warfarin at discharge (P <0.0001) and no aspirin at discharge (P =0.002), whereas female sex was a protective factor (P =0.01). Chronic kidney disease was not a predictive of the primary composite endpoint (P =0.10) (Table 3).

Kaplan-Meier survival curves (Figure 1) showed that vWF plasma concentrations discriminated groups of patients with higher rates of the composite endpoint of stroke or all-cause death (log-rank P =0.01).

Figure 1

Figure 1. 

Kaplan-Meier curves showing the relationship between von Willebrand factor (vWF) concentration and outcome (all-cause death and stroke) during 5years of follow-up.


Factors predictive of secondary composite endpoint (stroke, all-cause death or heart failure)

Patients in the upper and middle vWF plasma concentration tertiles had a higher rate of the secondary composite endpoint (61 and 61%, respectively) than those in the lower tertile (12%) (Table 2).

When considering the secondary composite endpoint of all-cause death, stroke or acute heart failure, our results demonstrated that factors associated with an increased risk of events were the upper and middle vWF plasma concentration tertiles (respectively, P =0.001 and P =0.002), age75years (P =0.026), C-reactive protein, log2 per unit increase (P =0.009), no warfarin at discharge (P <0.0001) and no aspirin at discharge (P =0.001) (Table 4) (Appendix A)

Prognostic value of CHA2 DS2 -VASc associated with vWF

When comparing the predictive ability of risk-stratification schemes before and after adding vWF concentrations to the CHA2 DS2 -VASc risk score, results demonstrated a slight but a significative statistical improvement in the predictive value of CHA2 DS2 -VASc associated with vWF, and area under the receiver operating characteristic curve values of 0.6559 and 0.6902, respectively (P =0.0033), considering a cut-off value of 204IU/dL (Appendix A). When using the model C-indexes to test predictive accuracy, comparison of the results showed: hazard ratio 1.27, 95% confidence interval 1.08–1.49 for CHA2 DS2 -VASc; and hazard ratio 1.29 95% confidence interval 1.11–1.50 for CHA2 DS2 -VASc+vWF; P =0.0009.


In this prospective longitudinal study of 122 patients hospitalized with NVAF, high plasma concentrations of vWF were a strong independent predictor of cardiovascular events. Moreover, we demonstrated that increased vWF plasma concentrations were associated with a higher risk of adverse vascular events, especially when considering the cut-off value of 191IU/dL (lower tertile).

The presence of NVAF is associated with a wide range of adverse events, including stroke, death, hospitalization and reduced quality of life and functional capacity. Patients who have a stroke are twice as likely to die within 12months as patients with AF who do not have a stroke [27]. Yet, only limited prospective data are available on predictors of mortality in this population [2, 3]. OAC treatment is effective in reducing the risk of AF-related death [6, 7, 28], and is recommended for stroke prevention in moderate- to high-risk patients with AF [25, 29]. However, physicians must balance the risk of a thrombotic event with the increased risk of bleeding associated with OAC therapy. Use of the CHA2 DS2 -VASc score is therefore recommended in contemporary clinical practice guidelines for risk stratification in AF [10, 25, 29, 30].

vWF is involved in the prothrombotic state in AF, which is associated with endothelial dysfunction or damage [4]. The vWF plasma concentration is easy to measure, thus raising the possibility of including it in the assessment of thromboembolic risk in patients with AF. Indeed, vWF plasma concentrations have been proposed as an additional factor for risk stratification, independent of the CHADS2 score, in patients with NVAF treated with OAC [21]. Several studies have addressed this issue and the results are discordant (Table 5). Our results concur with the findings of Conway et al. [12], who demonstrated that vWF plasma concentrations were an independent predictor of vascular events (i.e. composite of stroke, myocardial infarction or vascular death) in AF, with a relative risk of 1.2 (95% confidence interval 1.0–1.4) per 20IU/dL increase in vWF plasma concentration (P <0.02). In the Stroke Prevention in Atrial Fibrillation (SPAF) study, vWF plasma concentrations were higher in patients with acute or recent decompensated heart failure, who were deemed a particularly high-risk subgroup for AF-related stroke [31]. The underlying causes of a poor prognosis are not well understood. vWF plasma concentration as a marker of endothelial damage or dysfunction is raised in heart failure and AF, and increases further when the conditions co-exist [11, 19]. When we analysed the incidence of the secondary composite endpoint (stroke, all-cause death or acute heart failure), taking into account vWF plasma concentrations, patients who had a marker value in the middle or upper tertile were at higher risk of heart failure, stroke or all-cause death, suggesting that vWF plasma concentrations could be a predictor of heart failure in NVAF (Appendix A). Of note, in our study, 37 (46%) patients who had vWF plasma concentrations in the middle and upper tertiles also had a history of heart failure, suggesting a central role for this biomarker of thrombogenesis in the prognosis of AF associated with heart failure [18], independent of the presence of other cardiovascular risk factors. Roldán et al. [21] demonstrated that elevated vWF plasma concentrations were associated with adverse prognosis, including cardiovascular events and death, in patients with permanent AF on OAC therapy. The results from our study therefore lend further support to the proposal to include elevated vWF plasma concentration as a predictor of cardiovascular events for risk stratification in AF. The inclusion of vWF plasma concentration may provide additional stratification in terms of stroke risk, particularly among patients with a low CHA2 DS2 -VASc score, for whom the risk-benefit ratio of OAC therapy is less clear (Appendix A). In fact, there are poor data concerning the rationale for adding vWF to the usual score to predict thromboembolic risk. García-Fernández et al. [32] showed that vWF addition to CHA2 DS2 -VASc and HAS-BLED scores statistically improved prediction of stroke/transient ischaemic attack, systemic and peripheral embolism, acute coronary syndrome, acute heart failure, cardiac death and major bleeding. However, the impact on clinical decision-making remains non-significant. It remains to be elucidated whether AF per se increases the concentration of vWF in patients with stroke or heart failure, or whether the presence of other major cardiovascular risk factors also has an effect. We also identified other predictors of adverse cardiovascular outcome, including increased C-reactive protein concentration, non-paroxysmal AF and lack of OAC therapy or aspirin at discharge.

Conversely, some studies have not reported the same conclusion. An analysis from the prospective population-based Rotterdam study showed that in subjects aged>55years included in a group with AF and a group in sinus rhythm, there was no association between vWF plasma concentrations and subsequent stroke [33]. Heeringa et al. concluded that vWF plasma concentrations (age- and sex-adjusted) were associated with cardiac mortality in the studied population (relative risk 1.16, 95% confidence interval 1.06–1.27, per 10IU/dL increase); nevertheless, statistical significance was lost after additional adjustment [33]. Furthermore, in a meta-analysis that aimed to evaluate the association of inflammatory and haemostatic markers with stroke and thromboembolic events in patients with AF, Wu et al. [23] showed that circulating vWF plasma concentrations were not associated with stroke, with a pooled hazard ratio of 1.14 (95% confidence interval 0.76–1.72; P =0.53).

Study limitations

This analysis is limited by the lack of a control group in sinus rhythm. The small number of patients in the study results in a lack of precision in the results and wide confidence intervals. However, when comparing our study with others in the same setting, we have a follow-up of 5years, which is one of the most important follow-ups (Table 5). Further studies are needed, therefore, to establish the mechanism of cardiovascular events associated with raised vWF plasma concentrations in NVAF, to evaluate the endothelium/endocardium as a target for novel antithrombotic therapies in NVAF, to investigate other indices of endothelial dysfunction in NVAF and to examine the potential for vWF plasma concentrations as an aid to clinical risk stratification in NVAF, particularly in patients with a low CHA2 DS2 -VASc risk score, in whom OAC treatment is not recommended (Appendix A).


These prospective data suggest that the presence of elevated vWF plasma concentrations, when combined with known cardiovascular risk factors, can better predict clinical outcome (including stroke or all-cause death) in patients with NVAF. Further studies should be performed to assess whether vWF plasma concentrations can help to identify patients with a low CHA2 DS2 -VASc score who are at moderate risk of stroke or other cardiovascular events, and would therefore benefit from OAC therapy (Appendix A).

Sources of funding


Disclosure of interest

A.C. Research grants from the companies ARS, RESICARD (research nurses) and Boehringer-Ingelheim; consultant and lecture fees from the companies AstraZeneca, Bayer Pharma, BMS-Pfizer alliance, Boehringer-Ingelheim, Daiichi Sankyo and Novartis, unrelated to the present work.

Yann Ancedy, Emmanuelle Berthelot, Sylvie Lang, Stéphane Ederhy, Louise Boyer-Chatenet, Emanuele Di Angelantonio, Laurie Soulat-Dufour, Arnaud Etienney, Saroumadi Adavane-Scheublé, Franck Boccara, Ariel Cohen declare that they have no competing interest.


Sophie Rushton-Smith, PhD, provided editorial assistance, including editing, checking content and language, formatting and referencing, and was funded by the authors.

Appendix A

Figure A.1 Receiver operating characteristic (ROC) curve testing the predictive value of CHA2 DS2 -VASc score alone and associated with von Willebrand factor plasma concentration to identify the composite endpoint: stroke and all-cause death during follow-up. Area under the ROC curve: CHA2 DS2 -VASc=0 6559; CHA2 DS2 -VASc+vWF=0.6902; P =0.0033 for comparison. C-index: CHA2 DS2 -VASc=0.620 [interquartile range 0.534–0.707], hazard ratio 1.27, 95% confidence interval 1.08–1.49 (P <0.00010); CHA2 DS2 -VASc+vWF=0.646 [interquartile range 0.564–0.729], hazard ratio 1.29, 95% confidence interval 1.11–1.50 (P <0.0001); P =0.0009 for comparison. CHA2 DS2 -VASc: congestive heart failure, hypertension, age75years (doubled), diabetes, stroke/transient ischaemic attack/thromboembolism (doubled)–vascular disease, age 65–74years and sex category (female).

Variable HF at follow-up No HF at follow-up P a 
 (n =22) (n =100)  
Clinical characteristics    
Men 9 (40.9) 47 (47.0) 0.89 
Age (years) 74±12 70±15 0.12 
Age>75years 11 (50.0) 37 (37.0) 0.12 
Hypertension 16 (72.7) 61 (61.0) 0.31 
Diabetes 4 (18.2) 18 (18.0) 0.97 
Dyslipidaemia 8 (36.4) 31 (31.0) 0.65 
Current smoker 3 (13.6) 16 (16.0) 0.74 
Medical history    
AF 10 (45.5) 33 (33.0) 0.25 
ACS 6 (27.3) 17 (17.0) 0.21 
Congestive HF 13 (59.1) 28 (28.0) 0.001 
Stroke or TIA 2 (9.1) 7 (7.0) 0.74 
Non-paroxysmal AF 15 (68.2) 53 (53.0) 0.16 
CHA2 DS2 -VASc risk score 4 [3–6] 3 [1–5] 0.014 
1 (4.6) 11 (11.0)  
1 (4.6) 15 (15.0) 0.19 
20 (90.9) 74 (74.0)  
CRP concentration (mg/L) 20.0 [9.2–55.0] 5.0 [1.7–25.0] 0.002 
vWF concentration (IU/dL) 264 [230–347] 227 [164–328] 0.039 
Echocardiographic characteristics    
Left atrial area (cm220 [19–29] 20 [16–24] 0.13 
Right atrial area (cm216 [15–22] 16 [13–20] 0.35 
LVEF<40% 7 (31.8) 9 (9.0) <0.0001 
Treatment at discharge    
Warfarin 9 (40.9) 76 (76.0) <0.0001 
Aspirin 10 (45.5) 27 (27.0) 0.12 
Antiarrhythmic drugb 14 (63.6) 80 (80.0) 0.022 

Data are expressed as number (%), mean±standard deviation or median [interquartile range]. ACS: acute coronary syndrome; AF: atrial fibrillation; CHA2 DS2 -VASc: congestive heart failure, hypertension, age75years (doubled), diabetes, Stroke/transient ischaemic attack/thromboembolism (doubled)–vascular disease, age 65–74years and sex category (female); CRP: C-reactive protein; HF: heart failure; LVEF: left ventricular ejection fraction; TIA: transient ischaemic attack; vWF=von Willebrand factor.

[a]  P values were determined with univariate Cox analysis.
[b]  Including amiodarone.

Variable CHA2 DS2 -VASc=CHA2 DS2 -VASc=CHA2 DS2 -VAScP  
 (n =12) (n =16) (n =94)  
vWF concentration (IU/dL) 212 [137–312] 186 [148–250] 258 [190–340] 0.0447 
CRP concentration (mg/L) 4.1 [0.7–40.2] 3.3 [1.5–13.0] 9.1 [3.3–45.6] 0.06 
Non-paroxysmal AF 4 (33) 7 (44) 57 (61) 0.12 
Echocardiographic characteristics     
Left atrial area (cm218 [15–20] 21 [17–24] 20 [16–27] 0.23 
Right atrial area (cm218 [15–20] 15 [12–18] 16 [14–20] 0.30 
LVEF<40% 1 (6) 15 (16) 0.29 
Cardiovascular events     
All-cause death 1 (8) 2 (13) 38 (40) 0.0255a 
Stroke 4 (4) 0.50a 
Heart failure 1 (8) 1 (6) 20 (17) 0.27a 
Composite endpoints     
Stroke or death 1 (8) 2 (13) 40 (43) 0.0186a 
Stroke, death or heart failure 1 (8) 3 (19) 50 (53) 0.0037a 

Data are expressed as number (%) or median (interquartile range). AF: atrial fibrillation; CHA2 DS2 -VASc: congestive heart failure, hypertension, age75years (doubled), diabetes, stroke/transient ischaemic attack/thromboembolism (doubled)–vascular disease, age 65–74years and sex category (female); CRP: C-reactive protein; LVEF: left ventricular ejection fraction.

[a]  Log-rank P at 5-year follow-up.

 Univariate analysis Multivariable analysis 
191<vWF295IU/dL 4.41 (1.64–11.82); 0.003 2.96 (1.07–8.21); 0.036 
vWF>295IU/dL 5.33 (1.99–14.28); 0.001 3.17 (1.13–8.93); 0.029 
CHA2 DS2 -VASc risk score   
1.64 (0.15–18.06); 0.69 2.13 (0.19–24.12); 0.61 
6.25 (0.86–45.45); 0.07 3.22 (0.40–25.71); 0.27 
Chronic kidney disease, no 
Chronic kidney disease, yes 2.20 (1.19–4.09); 0.012 1.25 (0.63–2.48); 0.52 
CRP, log2 per unit increase 1.41 (1.18–1.67);<0.0001 1.23 (1.01–1.51); 0.043 
Warfarin at discharge 
No warfarin at discharge 3.41 (1.87–6.23);<0.0001 3.43 (1.46–8.07); 0.005 
Aspirin at discharge 
No aspirin at discharge 0.55 (0.30–1.02); 0.06 2.21 (0.93–5.25); 0.07 
Antiarrhythmics at discharge 
No antiarrhythmics at discharge 2.42 (1.29–4.55); 0.006 1.31 (0.65–2.63); 0.45 

Data are expressed as hazard ratio (95% confidence interval); P value. CHA2 DS2 -VASc: congestive heart failure, hypertension, age75years (doubled), diabetes, stroke/transient ischaemic attack/thromboembolism (doubled)–vascular disease, age 65–74 years and sex category (female); CRP: C-reactive protein; vWF: von Willebrand factor.

Variable Univariate analysis Multivariable analysis 
Female sex 16.33±20.4; (−24.12–56.78); 0.43 – 
Age, per unit increase 1.21±0.71 (−0.21–2.62); 0.09 −0.89±0.82 (−2.50–0.73); 0.28 
Hypertension 59.32±20.45 (18.83–99.81); 0.004 33.16±20.30 (−7.05–73.36); 0.11 
Diabetes −17.86±26.50 (−70.33–34.61); 0.50 – 
Dyslipidaemia 1.21±21.89 (−42.13–44.54); 0.96 – 
Current smoker 29.79±28.02 (−25.69–85.26); 0.29 – 
History of AF 7.06±21.36 (−35.23–49.34); 0.74 – 
History of ACS 7.14±26.09 (−44.51–58.80); 0.79 – 
History of congestive HF 58.26±20.94 (16.79–99.72); 0.006 12.19±20.84 (−29.10–53.47); 0.56 
History of stroke or TIA 16.70±39.02 (−60.55–93.96); 0.67 – 
Chronic kidney disease 44.86±20.03 (5.21–84.51); 0.027 34.98±21.75 (−8.09–78.06); 0.11 
CRP, after log transformation 28.97±4.83 (19.40–38.54);<0.0001 25.76±5.29 (15.29–36.24);<0.0001 
Non-paroxysmal AF 12.41±20.52 (−28.22–53.04); 0.55 – 
CHA2 DS2 -VASc=Ref – 
CHA2 DS2 -VASc=−14.56±42.32 (−98.36–69.24); 0.73 – 
CHA2 DS2 -VASc45.49±33.97 (−21.78–112.76); 0.18 – 
LVEF<40% 21.85±30.17 (−37.89–81.59); 0.47 – 

Data are expressed as β±standard error (95% confidence interval); P value. ACS: acute coronary syndrome; AF: atrial fibrillation; CHA2 DS2 -VASc: congestive heart failure, hypertension, age75years (doubled), diabetes, stroke/transient ischaemic attack/thromboembolism (doubled)–vascular disease, age 65–74years and sex category (female); CRP: C-reactive protein; HF: heart failure; LVEF: left ventricular ejection fraction; TIA: transient ischaemic attack.


Chugh S.S., Havmoeller R., Narayanan K., and al. Worldwide epidemiology of atrial fibrillation: a global burden of disease 2010 study Circulation 2014 ;  129 : 837-847 [cross-ref]
Benjamin E.J., Wolf P.A., D’Agostino R.B., Silbershatz H., Kannel W.B., Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study Circulation 1998 ;  98 : 946-952 [cross-ref]
Kannel W.B., Wolf P.A., Benjamin E.J., Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates Am J Cardiol 1998 ;  82 : 2N-9N
Kahn S.R., Solymoss S., Flegel K.M. Nonvalvular atrial fibrillation: evidence for a prothrombotic state CMAJ 1997 ;  157 : 673-681
Watson T., Shantsila E., Lip G.Y. Mechanisms of thrombogenesis in atrial fibrillation: Virchow's triad revisited Lancet 2009 ;  373 : 155-166 [cross-ref]
Hart R.G., Pearce L.A., Aguilar M.I. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation Ann Intern Med 2007 ;  146 : 857-867 [cross-ref]
Hylek E.M., Go A.S., Chang Y., and al. Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation N Engl J Med 2003 ;  349 : 1019-1026 [cross-ref]
Gage B.F., Waterman A.D., Shannon W., Boechler M., Rich M.W., Radford M.J. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation JAMA 2001 ;  285 : 2864-2870 [cross-ref]
Lip G.Y., Frison L., Halperin J.L., Lane D.A. Identifying patients at high risk for stroke despite anticoagulation: a comparison of contemporary stroke risk stratification schemes in an anticoagulated atrial fibrillation cohort Stroke 2010 ;  41 : 2731-2738 [cross-ref]
Camm A.J., Lip G.Y., De Caterina R., and al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation--developed with the special contribution of the European Heart Rhythm Association Europace 2012 ;  14 : 1385-1413
Freestone B., Gustafsson F., Chong A.Y., and al. Influence of atrial fibrillation on plasma von willebrand factor, soluble E-selectin, and N-terminal pro B-type natriuretic peptide levels in systolic heart failure Chest 2008 ;  133 : 1203-1208 [cross-ref]
Conway D.S., Pearce L.A., Chin B.S., Hart R.G., Lip G.Y. Prognostic value of plasma von Willebrand factor and soluble P-selectin as indices of endothelial damage and platelet activation in 994 patients with nonvalvular atrial fibrillation Circulation 2003 ;  107 : 3141-3145 [cross-ref]
Roldan V., Marin F., Garcia-Herola A., Lip G.Y. Correlation of plasma von Willebrand factor levels, an index of endothelial damage/dysfunction, with two point-based stroke risk stratification scores in atrial fibrillation Thromb Res 2005 ;  116 : 321-325 [cross-ref]
Morange P.E., Simon C., Alessi M.C., and al. Endothelial cell markers and the risk of coronary heart disease: the Prospective Epidemiological Study of Myocardial Infarction (PRIME) study Circulation 2004 ;  109 : 1343-1348 [cross-ref]
Spiel A.O., Gilbert J.C., Jilma B. von Willebrand factor in cardiovascular disease: focus on acute coronary syndromes Circulation 2008 ;  117 : 1449-1459 [cross-ref]
Chong A.Y., Blann A.D., Patel J., Freestone B., Hughes E., Lip G.Y. Endothelial dysfunction and damage in congestive heart failure: relation of flow-mediated dilation to circulating endothelial cells, plasma indexes of endothelial damage, and brain natriuretic peptide Circulation 2004 ;  110 : 1794-1798 [cross-ref]
Gombos T., Mako V., Cervenak L., and al. Levels of von Willebrand factor antigen and von Willebrand factor cleaving protease (ADAMTS13) activity predict clinical events in chronic heart failure Thromb Haemost 2009 ;  102 : 573-580 [cross-ref]
Kistorp C., Chong A.Y., Gustafsson F., and al. Biomarkers of endothelial dysfunction are elevated and related to prognosis in chronic heart failure patients with diabetes but not in those without diabetes Eur J Heart Fail 2008 ;  10 : 380-387 [cross-ref]
Lip G.Y., Pearce L.A., Chin B.S., Conway D.S., Hart R.G. Effects of congestive heart failure on plasma von Willebrand factor and soluble P-selectin concentrations in patients with non-valvar atrial fibrillation Heart 2005 ;  91 : 759-763 [cross-ref]
Gibbs C.R., Blann A.D., Watson R.D., Lip G.Y. Abnormalities of hemorheological, endothelial, and platelet function in patients with chronic heart failure in sinus rhythm: effects of angiotensin-converting enzyme inhibitor and beta-blocker therapy Circulation 2001 ;  103 : 1746-1751 [cross-ref]
Roldan V., Marin F., Muina B., and al. Plasma von Willebrand factor levels are an independent risk factor for adverse events including mortality and major bleeding in anticoagulated atrial fibrillation patients J Am Coll Cardiol 2011 ;  57 : 2496-2504 [cross-ref]
Krishnamoorthy S., Khoo C.W., Lim H.S., and al. Prognostic role of plasma von Willebrand factor and soluble E-selectin levels for future cardiovascular events in a ‘real-world’ community cohort of patients with atrial fibrillation Eur J Clin Invest 2013 ;  43 : 1032-1038 [cross-ref]
Wu N., Chen X., Cai T., and al. Association of inflammatory and hemostatic markers with stroke and thromboembolic events in atrial fibrillation: a systematic review and meta-analysis Can J Cardiol 2015 ;  31 : 278-286 [cross-ref]
Di Angelantonio E., Ederhy S., Benyounes N., Janower S., Boccara F., Cohen A. Comparison of transesophageal echocardiographic identification of embolic risk markers in patients with lone versus non-lone atrial fibrillation Am J Cardiol 2005 ;  95 : 592-596 [inter-ref]
Fuster V., Ryden L.E., Cannom D.S., and al. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: full text: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 guidelines for the management of patients with atrial fibrillation) developed in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society Europace 2006 ;  8 : 651-745
Lang R.M., Badano L.P., Mor-Avi V., and al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging J Am Soc Echocardiogr 2015 ;  28 : 1–39 e14.
Fitzmaurice D.A., Hobbs F.D., Jowett S., and al. Screening versus routine practice in detection of atrial fibrillation in patients aged 65 or over: cluster randomised controlled trial BMJ 2007 ;  335 : 383 [cross-ref]
ACTIVE writing group of the ACTIVE investigatorsConnolly S., Pogue J., and al. Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the atrial fibrillation clopidogrel trial with irbesartan for prevention of vascular events (ACTIVE W): a randomised controlled trial Lancet 2006 ;  367 : 1903-1912
Skanes A.C., Healey J.S., Cairns J.A., and al. Focused 2012 update of the Canadian Cardiovascular Society atrial fibrillation guidelines: recommendations for stroke prevention and rate/rhythm control Can J Cardiol 2012 ;  28 : 125-136 [cross-ref]
Kirchhof P., Benussi S., Kotecha D., and al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS Eur Heart J 2016 ;  37 : 2893-2962 [cross-ref]
Adjusted-dose warfarin versus low-intensity, fixed-dose warfarin plus aspirin for high-risk patients with atrial fibrillation: Stroke Prevention in Atrial Fibrillation III randomised clinical trial Lancet 1996 ;  348 : 633-638
Garcia-Fernandez A., Roldan V., Rivera-Caravaca J.M., and al. Does von Willebrand factor improve the predictive ability of current risk stratification scores in patients with atrial fibrillation? Sci Rep 2017 ;  7 : 41565 [cross-ref]
Heeringa J., Conway D.S., van der Kuip D.A., and al. A longitudinal population-based study of prothrombotic factors in elderly subjects with atrial fibrillation: the Rotterdam Study 1990-1999 J Thromb Haemost 2006 ;  4 : 1944-1949 [cross-ref]
Alonso A., Tang W., Agarwal S.K., Soliman E.Z., Chamberlain A.M., Folsom A.R. Hemostatic markers are associated with the risk and prognosis of atrial fibrillation: the ARIC study Int J Cardiol 2012 ;  155 : 217-222 [cross-ref]
Pinto A., Tuttolomondo A., Casuccio A., and al. Immuno-inflammatory predictors of stroke at follow-up in patients with chronic non-valvular atrial fibrillation (NVAF) Clin Sci (Lond) 2009 ;  116 : 781-789 [cross-ref]

© 2017  Published by Elsevier Masson SAS.
EM-CONSULTE.COM is registrered at the CNIL, déclaration n° 1286925.
As per the Law relating to information storage and personal integrity, you have the right to oppose (art 26 of that law), access (art 34 of that law) and rectify (art 36 of that law) your personal data. You may thus request that your data, should it be inaccurate, incomplete, unclear, outdated, not be used or stored, be corrected, clarified, updated or deleted.
Personal information regarding our website's visitors, including their identity, is confidential.
The owners of this website hereby guarantee to respect the legal confidentiality conditions, applicable in France, and not to disclose this data to third parties.
Article Outline