Article Text

Download PDFPDF

Original article
Increased thrombin generation and platelet activation are associated with deficiency in high molecular weight multimers of von Willebrand factor in patients with moderate-to-severe aortic stenosis
  1. Joanna Natorska1,2,
  2. Ksenia Bykowska3,
  3. Marta Hlawaty2,
  4. Grzegorz Marek1,
  5. Jerzy Sadowski1,2,
  6. Anetta Undas1,2
  1. 1John Paul II Hospital, Krakow, Poland
  2. 2Institute of Cardiology, Jagiellonian University School of Medicine, Krakow, Poland
  3. 3Institute of Hematology and Transfusion Medicine, Warsaw, Poland
  1. Correspondence to Anetta Undas, Institute of Cardiology, Jagiellonian University School of Medicine, 80 Pradnicka St, 31-202 Krakow, Poland; mmundas{at}


Background High molecular weight von Willebrand factor (vWF) multimers (HMWM) are often deficient in patients with severe aortic stenosis (AS) owing to shear stress-enhanced proteolysis of vWF. It has also been reported that AS is associated with increased activation of blood coagulation.

Objective To investigate whether patients with AS with a deficiency in vWF HMWM have enhanced thrombin generation and platelet activation in vivo.

Design Based on the analysis of vWF HMWM performed using immunolocalisation, 11 subjects with vWF HMWM deficiency (low %HMWM group) were identified and compared with 42 patients with AS with a normal distribution of vWF HMWM (normal %HMWM group). Plasma thrombin markers thrombin-antithrombin complexes (TAT) and prothrombin factor 1+2 (F1.2) plus platelet activation markers soluble CD40 ligand (sCD40L), β-thromboglobulin and P-selectin were also measured.

Patients 48 consecutive patients with severe AS and five with moderate AS, free of angiographically-proven coronary artery disease and clinically overt bleeding, were studied.

Results Patients in the low %HMWM group had 34.8% higher maximal transvalvular gradient (p=0.0003) and 44.8% higher mean gradient (p=0.0002) than those in the normal %HMWM group. Thrombin formation was enhanced in the low %HMWM group (F1.2, 284.5±63.7 vs 216.9±62.5 pmol/l, p=0.004; thrombin-antithrombin, 4.89±1.3 vs 4.06±0.9 μg/l, p=0.02) and both markers showed inverse correlations with the percentage of vWF HMWM (r=−0.59, p=0.002; r=−0.42, p=0.03, respectively). In the low %HMWM group sCD40L (279.4±60.7 vs 221.4±41.7 pmol/l, p=0.003) and β-thromboglobulin (73.1±9.2 vs 64.5±8.5 IU/ml, p=0.04), but not P-selectin, were also higher than in the remaining patients with AS.

Conclusion Patients with advanced AS deficient in vWF HMWM are characterised by enhanced thrombin formation and platelet activation. This observation indicates the ambivalent impact of high shear stress in AS on haemostasis and might help explain two aspects of AS—Heyde syndrome and increased risk of thromboembolism.

  • Aortic stenosis
  • von willebrand factor high-molecular-weight multimers deficiency
  • thrombin formation
  • platelet activation markers
  • platelet activation
  • platelet adhesion
  • aortic valve disease
  • inflammation
View Full Text

Statistics from


Aortic stenosis (AS) is the most commonly acquired valvular lesion in adults. The prevalence of critical AS is 1–2% at age 75 years.1 AS is a multifactorial atherosclerosis-like process, but the precise mechanisms of AS are largely unknown. The pathological changes observed at the early phase of AS include valve thickening, accumulation of irregular fibrocalcific masses and inflammatory infiltrates involving predominantly macrophages and lymphocytes.1 Risk factors for AS are similar to those for atherosclerosis.2

von Willebrand factor (vWF), synthesised in endothelial cells and megakaryocytes, is a multimeric plasma glycoprotein that is essential for platelet-subendothelium adhesion and platelet-to-platelet interactions, as well as platelet aggregation under high shear stress conditions.3 vWF binds to the platelet receptor glycoprotein Ib/IX and to injured vessel walls by interacting with collagen.4 The largest high molecular weight multimers (HMWM) of vWF are most effective in platelet-mediated haemostasis.5 It is estimated that up to 20% of patients with severe AS experience mucocutaneous bleeding,6 whereas life-threatening haemorrhages—mainly from the gastrointestinal tract, termed Heyde syndrome, usually associated with intestinal angiodysplasia5 6—are rare.

The precise mechanism underlying the link between AS and bleeding tendency was shown by Warkentin et al who detected a deficiency in HMWM of vWF acquired type 2A von Willebrand syndrome due to proteolysis of vWF multimers.7 High shear stress can alter the structure of the vWF molecule leading to exposure of the bond between Tyr842 and Met843, which is sensitive to the action of a specific metalloprotease (ADAMTS 13; A Disintegrin And Metalloproteinase with ThromboSpondin).8 Loss of the largest vWF multimers leads to impairment of primary haemostasis and the resultant bleeding tendency.9 In a study of 50 consecutive patients with severe AS, 67–92% of the patients had primary haemostasis abnormalities which correlated with the severity of AS. Surgical correction of AS has been shown to normalise the distribution of vWF HMWM.9 10

Dimitrow et al have recently reported that moderate-to-severe AS with a mean transvalvular gradient of 46±12 mm Hg is associated with increased thrombin generation and platelet activation in circulating blood.11 It has been shown that shear stress caused by the constricted aortic orifice of the aortic valve contributes to the generation of microparticles and thereby systemic inflammation and coagulation activation.12 On the other hand, tissue factor—the main initiator of blood coagulation in vivo—is highly expressed in stenotic aortic valves and its presence is associated with in vivo thrombin formation and the resultant enhanced calcification.13 14 The factors that determine the balance between prothrombotic and haemorrhagic effects in patients with severe AS remain unknown. It has been hypothesised that coagulation activation might compensate, to some extent, the deficiency in vWF HMWM and thus reduces the incidence of bleeding in AS.15 The aim of the current study was to investigate whether patients with AS with vWF HMWM deficiency have more pronounced activation of blood coagulation and platelets.



A total of 53 consecutive patients with moderate-to-severe AS (35 men and 18 women) of mean±SD age 65.6±10.7 years were recruited prior to valve replacement. The exclusion criteria were diabetes, autoimmune disorders, acute infection, Valsalva sinus aneurysm, angiographically documented epicardial artery stenosis (>20% diameter), anaemia (haemoglobin <11 g/dl), known cancer, endocarditis, previous cardiac surgery, common or internal carotid artery stenosis, peripheral artery disease, a history of myocardial infarction, stroke or venous thromboembolism. Patients who had other heart defects were ineligible for the study.

Arterial hypertension was diagnosed based on a history of hypertension or antihypertensive treatment. Diabetes mellitus was defined as a previous diagnosis of diabetes or at least two random fasting glucose levels of >7 mmol/l. Hyperlipidaemia was diagnosed based on medical records, statin therapy or total cholesterol of ≥5.2 mmol/l. Current smoking was defined as smoking at least one cigarette daily.

Nine plasma samples from healthy subjects, analysed at the same time as AS patients, served as controls to define the normal range for the percentage of vWF HMWM.6


Transthoracic echocardiography was performed in each patient using a MargotMac 5000 ultrasound machine within 1–3 days prior to valve replacement using conventional techniques in accordance with the European Society of Cardiology guidelines.16 The aortic valve area (AVA) was calculated using the standard continuity equation. The AVA was divided by the body surface area to obtain the aortic valve area index (AVAI).17 The transvalvular gradient was measured by Doppler echocardiography using the modified Bernoulli equation.17

Laboratory investigations

Fasting blood samples were collected between 07.00 to 10.00 h. Routine blood tests, including lipid profile, glucose and creatinine in serum, were assayed by routine laboratory techniques. High-sensitive C reactive protein (hsCRP) was measured in serum by immunoturbidimetry (Siemens Healthcare Diagnostics, Deerfield, Illinois, USA). Commercially available ELISAs were used to determine two plasma thrombin generation markers—prothrombin fragments 1+2 (F1.2) and thrombin-antithrombin complex (TAT) (Siemens Healthcare Diagnostics)—and three platelet activation markers—β-thromboglobulin (β-TG) (Diagnostica Stago, Asnieres, France), soluble CD40 ligand (sCD40L) and P-selectin (both R&D Systems, Minneapolis, Minnesota, USA), together with serum interleukin 6 (IL-6) (R&D Systems.). All measurements were performed by technicians blinded to the origin of the samples. Intra-assay and inter-assay coefficients of variation for all assays were <8%.

Plasma vWF antigen (vWF:Ag) was measured by enzyme-linked fluorescent assay (Vidas, bioMerieux, Marcy I'Etoile, France). Functional analysis of vWF was performed by measuring its collagen-binding activity (vWF:CB) with the use of type III collagen (Technoclone, Vienna, Austria) and ristocetin cofactor activity (vWF:RCo) (Siemens Healthcare Diagnostics) in plasma, as described previously.18 The ratio of vWF:CB to vWF:Ag and that of vWF:RCo to vWF:Ag were calculated (vWF:CB/Ag and vWF:RCo/Ag, normal values >0.7).9 The multimeric structure of plasma vWF was analysed by electrophoresis with 0.1% sodium dodecyl sulphate and 1.5% agarose gel.18 19 The percentage of vWF HMWM (%HMWM >15 mers) was determined based on densitometry as described previously.6 19 Normal plasma samples (n=10) were used as a reference for each gel electrophoresis. The mean value of %HMWM >15 mers for normal plasma samples was 22.37±6.03%, as described elsewhere.6 The lower limit of the normal range for the %HMWM was 10.3, calculated as 2×SD below the mean value for normal plasma samples.6

Statistical analysis

Statistical analysis was performed with Statsoft 7.1 PL package (StatSoft Inc, 2005). Values are expressed as mean±SD, median (IQR) or percentage. The normality of distribution was tested with the Kolmogorov–Smirnov test for each variable in each group. Comparisons between groups were performed with the χ2 test, Student test or Mann–Whitney U test, as appropriate. The Spearman correlation coefficient was calculated to evaluate associations between values. A value of p<0.05 was considered statistically significant.


Patient groups

The demographic, clinical and echocardiographic data for the entire population and for the patients with AS stratified according to the percentage of vWF HMWM in the plasma are summarised in table 1. The maximal transvalvular gradient in the patient group was 88±21.9 mm Hg and the mean gradient was 63.7±16.3 mm Hg. Patients with a percentage of vWF HMWM <10.3% (low %HMWM group) had 34.8% higher maximal transvalvular gradient and 44.8% higher mean gradient (p=0.0003 and p=0.0002, respectively) than those with a percentage of vWF HMWM >10.3% (normal %HMWM group). Patients with low %HMWM more frequently had a history of arterial hypertension and current smoking. The groups did not differ with regard to other cardiovascular risk factors (table 1). None of the patients reported previous clinically overt bleeding.

Table 1

Characteristics of patients with aortic stenosis and their two subgroups stratified according to the percentage of high molecular weight multimers (%HMWM) of von Willebrand factor

Laboratory parameters

Patients in the low and normal %HMWM groups did not differ with regard to routine laboratory tests, including hsCRP and IL-6 (table 1).

vWF function

As shown in table 2, the percentage of vWF HMWM was below the normal range in 11 (20%) patients (median 8.88%, range 5.2–10.3%) and was lower by 54.8% (p=0.001) compared with the remaining patients in the normal %HMWM group. vWF function was significantly impaired in the low %HMWM group, as reflected by a decrease in both vWF:RCo and vWF:RCo/Ag ratios compared with the remaining patients in the group. The vWF:CB/Ag ratio was below the normal value of 0.7 in 13 (24.5%) patients, including seven subjects (63.6%) from the low %HMWM group and six (14.3%) from the normal %HMWM group (p=0.03). The binding of vWF to platelets in the presence of collagen was similar in both groups (table 2).

Table 2

Haemostatic variables

There were no significant correlations between the percentage of vWF HMWM and transvalvular gradients in the normal %HMWM group (figure 1). Interestingly, the percentage of vWF HMWM in the low %HMWM group inversely correlated with both maximal (r=−0.54, p=0.029) and mean (r=−0.58, p=0.034) transvalvular gradients (figure 1). The percentage of vWF HMWM in the low %HMWM group was associated with the occurrence of arterial hypertension (r=0.69, p=0.038). In the low %HMWM group the percentage of HMWM tended to be associated with left ventricular ejection fraction (r=0.51, p=0.058), AVA (r=−0.25, p=0.064) or AVAI (r=−0.26, p=0.07). There were also inverse correlations between the maximal transvalvular gradient and vWF:RCo (r=−0.6, p=0.047) and vWF:CB (r=−0.49, p=0.045) in the low %HMWM group. No such associations were observed for a mean gradient in this group.

Figure 1

Relation between the percentage of von Willebrand factor high molecular weight multimers (vWF HMWM) and maximal (A, C) or mean (B, D) transvalvular gradients in patients with aortic stenosis with a normal distribution of vWF HMWM (n=42) (left panels) and with a low percentage of vWF HMWM (n=11) (right panels).

Thrombin generation

In the whole group, vWF:Ag correlated inversely with both thrombin formation markers TAT and F1.2 (r=−0.36, p=0.024 and r=−0.55, p=0.028, respectively). Moreover, TAT and F1.2 were elevated in patients with low %HMWM compared with those with normal %HMWM (table 2). The percentage of vWF HMWM was not associated with thrombin generation markers in the whole patient group (figure 2A,B). However, in subjects with low %HMWM there were inverse correlations between the percentage of vWF HMWM and both TAT and F1.2 (r=−0.42, p=0.03 and r=−0.59, p=0.002, respectively; figure 2D,E). In the whole group there were positive correlations between a mean (but not maximal) transvalvular gradient and thrombin generation markers TAT and F1.2 (r=0.39, p=0.031 and r=0.55, p=0.042, respectively). However, in the low %HMWM group we also observed positive correlations between maximal transvalvular gradient and both F1.2 (r=0.49, p=0.036) and TAT (r=0.51, p=0.018).

Figure 2

Relation between the percentage of von Willebrand factor high molecular weight multimers (vWF HMWM) and thrombin-antithrombin complex (TAT) (A, D), prothrombin fragment 1+2 (F1.2) (B, E) and soluble CD40 ligand (sCD40L) (C, F) in the entire aortic stenosis group (n=53) (left panels) and in the patients with a low percentage of vWF HMWM (n=11) (right panels).

Platelet activation

Of the three platelet markers measured, only sCD40L and β-TG were elevated in the low %HMWM group compared with those from the normal %HMWM group (table 2). In the low %HMWM group there was a positive correlation between the maximal transvalvular gradient and β-TG (r=0.4, p=0.01), but not sCD40L or P-selectin.

In the whole group the percentage of HMWM was not associated with platelet activation markers (figure 2C). However, patients with low %HMWM showed inverse associations between the percentage of HMWM and sCD40L (r=−0.43, p=0.004; figure 2F). No such associations were observed for β-TG or P-selectin.

In the low %HMWM group F1.2 was positively correlated with both β-TG and sCD40L (r=0.66, p=0.006 and r=0.69, p=0.003, respectively), but not P-selectin. The same correlations were observed between TAT and both β-TG and sCD40L (r=0.57, p=0.003 and r=0.64, p=0.005, respectively).


The major finding of this study was that the deficiency of large vWF multimers is accompanied by increased thrombin generation and platelet activation in moderate-to-severe AS, as reflected by increased thrombin generation markers TAT and F1.2 and platelet activation markers β-TG and sCD40L. We confirmed that the percentage of vWF HMWM is lower when maximal and mean transvalvular gradients are higher in patients with AS.9 The prothrombotic changes, which are in line with our previous findings,11 were also more pronounced in patients with AS with a higher pressure gradient. This phenomenon shifts the haemostatic balance towards thrombosis in severe AS and might affect the risk of bleeding in this disease.

Acquired type 2 von Willebrand syndrome is a recognised risk factor for cutaneous or mucosal bleeding.6 7 Vincentelli and colleagues9 reported that 31 patients (74%) with AS had the percentage of vWF HMWM below the normal range and only 9 patients (21%) had a clinically overt bleeding tendency. In the present study of a similar size performed on patients with AS with a higher mean transvalular gradient (88.2±21.9 vs 57.3±12.7 mm Hg9), we have shown that 20% of patients with moderate-to-severe AS had a marked reduction in the %HMWM and acquired type 2A von Willebrand syndrome can be diagnosed. Of note, our patients did not report overt bleeding within the 6 months prior to enrolment, which might contribute to a lower prevalence of subjects with vWF HMWM deficiency in the present study.

The current findings might have practical implications. Available data suggest that, in patients with severe AS, the risk of bleeding is lower than would be expected based on previous reports,6 20 21 which might be explained, at least in part, by enhanced thrombin generation and release of active platelet granule proteins observed simultaneously. Moreover, the current study points out that the effect of high shear stress in patients with AS is ambivalent and comprises two aspects—Heyde syndrome, largely related to the presence of bleeding prone lesions, and increased risk of thromboembolism, demonstrated in 1999.22 Prothrombotic aspects of severe AS, documented in the present study performed in patients without significant coronary or carotid stenosis, highlight the role of AS-associated high shear stress in the risk of thrombotic events.

From the pathophysiological point of view, it is unclear to what extent increased thrombin production and platelet activation might affect vWF-mediated platelet effects. vWF HMWM-mediated platelet adhesion cannot be corrected by more vigorous thrombin production and platelet granule release. However, enhanced thrombin formation leading to fibrin clot formation and platelet activation might promote efficient haemostasis despite a reduced proportion of the largest vWF multimers in patients with AS. These prothrombotic changes result in faster fibrin-platelet thrombus formation at the site of vascular injury.23 Higher concentrations of thrombin, a potent platelet agonist, potentiate platelet activation and aggregation.23 Increased thrombin levels, as well as platelet activation, lead to the formation of tight fibrin clot networks resistant to lysis.24 25 Given the current results, it might be hypothesised that increased coagulation activation, together with facilitated interaction of platelets with endothelial cells in bleeding-prone lesions, may counteract a haemorrhagic tendency which does not always show an association with the deficiency in vWF HMWM.26 We may conclude that a high transvalvular gradient leads to high shear stress and vWF proteolysis as well as a hypercoagulable state, indicating the ambivalence of the impact of AS on the haemostatic balance.

It has recently been reported that low levels of HMWM vWF in severe AS are associated with a significant impairment of platelet aggregation and ADP-inducible P-selectin expression.27 In contrast to the current study, circulating plasma P-selectin levels were not determined. Our findings indicate that the amounts of P-selectin released from activated platelets are not related to the deficiency in HMWM vWF in severe AS. As suggested by Panzer et al,27 reduced P-selectin platelet expression in vitro might be just the result of increased release of P-selectin in patients with AS.

Our study has several limitations. First, the number of patients enrolled in this study was small, mostly due to numerous exclusion criteria, which may have introduced type II errors, especially in calculations of the correlation coefficients. Given the fact that there were 48 patients with severe AS in the current study (n=53), our findings might not be extrapolated to subjects with a lower mean transvalvular gradient. Second, all variables were analysed at a single time point before the corrective surgery and it remains to be established if there is a significant increase in vWF HMWM percentage and reduction in coagulation activation after reduction in the pressure gradient following valve replacement. Third, exclusion of comorbidities—especially documented coronary artery disease and stroke—could lead to the omission of patients with AS with a more pronounced bleeding tendency.

In conclusion, our results indicate that patients with advanced AS and deficiency of vWF HMWM have enhanced in vivo thrombin formation and platelet-derived protein release. This phenomenon might alter the clinical presentation of patients with severe AS. Further studies with a larger population of patients with AS followed for a few years are required to explore the actual impact of thrombin and platelet markers on bleeding tendency in AS.


View Abstract


  • See Editorial, p 1997

  • Funding This study was supported by a grant from the Polish Ministry of Science to AU(no N N402 383338).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the Jagiellonian University Krakow Poland.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles