Plasma von Willebrand Factor and Soluble P-Selectin as Indices of Endothelial Damage and Platelet Activation in 1321 Patients With Nonvalvular Atrial Fibrillation
Relationship to Stroke Risk Factors
Background— Epidemiological studies have identified clinical and echocardiographic factors associated with increased stroke risk in atrial fibrillation (AF), but mechanisms linking these factors to stroke in AF are incompletely understood. We hypothesized that stroke risk factors may be associated with increased endothelial damage/dysfunction and platelet activation among patients with AF.
Methods and Results— We measured plasma levels of von Willebrand factor (vWF, a marker of endothelial damage/dysfunction) and soluble P-selectin (sP-sel, a marker of platelet activation) by ELISA in 1321 participants in the Stroke Prevention in Atrial Fibrillation (SPAF) III study and related these indices to the presence of stroke risk factors and cardiovascular disease. Age (P<0.001), prior cerebral ischemia (P<0.01), recent heart failure (P<0.001), diabetes (P<0.001), and body mass index (P<0.001) were independently associated with increased vWF (r2 adjusted=9%). Independent associates of increased sP-sel were diabetes (P=0.01), peripheral vascular disease (P<0.001), and current smoking (P=0.01), whereas prior cerebral ischemia (P=0.002) and female sex (P<0.001) were associated with reduced sP-sel (r2 adjusted=4%). Using prospectively validated stroke risk stratification criteria, we observed a significant stepwise increase in vWF from low- to moderate- to high-risk groups (r2 adjusted=3%, P<0.001), whereas sP-sel remained constant (P= 0.24).
Conclusions— Four recognized risk factors for stroke in AF (advancing age, prior cerebral ischemia, recent heart failure, and diabetes) were independently associated with raised plasma vWF (or endothelial damage/dysfunction), whereas only 1 (diabetes) was associated with increased sP-sel (platelet activation). Further longitudinal studies are now needed to confirm relationships between endothelial damage/dysfunction, platelet activation, and stroke in AF.
Received May 22, 2002; revision received July 29, 2002; accepted July 29, 2002.
Nonvalvular atrial fibrillation (AF), the most common sustained cardiac arrhythmia, is associated with an approximately 5-fold increase in stroke risk1 because of the embolization of thrombus formed within the left atrial appendage (LAA). The absolute risk of stroke varies widely according to the presence or absence of certain clinical and echocardiographically identifiable factors,2 but the mechanisms whereby these factors influence the risk of stroke are incompletely understood.3
Abnormal plasma markers of coagulation, endothelial function, and platelet activation have been described in AF,4–8⇓⇓⇓⇓ but the relationship between these markers and stroke risk is unproven9–11⇓⇓ because many analyses have been based on small, underpowered case-controlled studies. Furthermore, AF is frequently a manifestation of additional (possibly undiagnosed) cardiovascular disease,1 but it is unclear to what extent additional disease may influence the abnormal prothrombotic plasma markers in AF.12
We hypothesized that (1) plasma levels of von Willebrand factor (vWF, a marker of endothelial damage/dysfunction13) and soluble P-selectin (sP-sel, a marker of platelet activation13) could be related to recognized stroke risk factors in AF; (2) these indices could also be related to additional stroke risk factors and other cardiovascular disease; and (3) these indices could be related to stroke risk stratification. To test this hypothesis, we measured plasma levels of vWF and sP-sel in 1321 participants in the Stroke Prevention in Atrial Fibrillation (SPAF) III study and related levels to the presence of stroke risk factors and cardiovascular disease among this large AF cohort. Furthermore, using the SPAF III stroke risk stratification criteria, which have been prospectively validated in 3 subsequent AF cohorts, 14–16⇓⇓ we examined the relationship between plasma vWF and sP-sel levels and estimated stroke risk.
All patients were participants in the SPAF III study, which was performed at 20 clinical sites in the United States and Canada between 1993 and 1997; the design and main results have been reported previously.17 In brief, patients with nonvalvular AF were stratified as having either low, moderate, or high risk for stroke based on clinical and echocardiographic features found to be predictive of thromboembolic risk in the earlier SPAF I and II studies. Specifically, those with any of 4 high-risk criteria (women >75 years of age, systolic hypertension >160 mm Hg, impaired left ventricular function [clinical heart failure within 100 days of entry or M-mode fractional shortening ≤25%], or previous thromboembolism) were randomized to receive either adjusted-dose warfarin (target international normalized ratio [INR] 2 to 3) or fixed, low-dose warfarin (target INR 1.2 to 1.5) plus aspirin 325 mg/d (termed “combination therapy”). Participants without any of the 4 specific risk factors were classified as being at low or moderate risk (depending on the absence or presence of hypertension, respectively) and received aspirin 325 mg/d alone.
An earlier study of different hemostatic markers among the same cohort has been reported previously.9 Blood samples were collected within 30 days of enrollment or after 3 months in the study. One sample per patient (sample collected at study entry, or if not available, then sample collected 3 months after entry) was included in these analyses. Sixty-nine percent (1339/1936) of SPAF III participants had a sample collected appropriately for these analyses, but because of natural sample wastage over time, specimens were available for only 1321 patients in the present cross-sectional analysis.
Blood Collection and Laboratory Analysis
Blood collection materials were prepared at the Laboratory for Clinical Biochemistry Research, Department of Pathology, University of Vermont. Blood for vWF and sP-sel assays was drawn into 3.8% sodium citrate tubes (Becton Dickinson), immediately mixed by gentle inversion, stored on melting ice, and centrifuged at 4°C for 30 000g-minutes within 1 hour of phlebotomy, and plasma was separated for vWF and sP-sel assays. Measurements of sP-sel and vWF were performed with ELISA with reagents from R&D Systems and Dako-Patts, respectively. The unit for vWF is international units per deciliter, which was standardized by reference vWF from the National Institute for Biological Standards and Controls, Blanche Lane, South Mimms, Potters Bar, Hertfordshire, UK. Intra-assay coefficients of variation for all ELISA assays were <5%, and interassay variances were <10%.
Patient characteristics were compared between groups with a χ2 test for categorical variables and Student’s t test or ANOVA for continuous data. Forward and backward stepwise linear regression analyses (with residual analysis) were used to identify features independently associated with marker levels. Statistical analyses were undertaken with SPSS software (SPSS Inc). Statistical significance was accepted at the 0.05 level (2-sided).
Levels of vWF for the 1321 participants were approximately normally distributed, with a mean value of 147 IU/dL (SD 31) and a range of 47 to 240 IU/dL. Levels of sP-sel were slightly skewed right, with a mean value of 34 ng/mL (SD 13) and a median of 32 ng/mL (range 3 to 140 ng/mL).
Relationship of vWF and sP-sel to Risk Factors for Stroke and Cardiovascular Disease
Univariate analyses revealed associations between increased vWF levels and the following stroke risk factors: advancing age (r=0.18, P<0.001), female sex (P=0.01), prior cerebral ischemia (P<0.001), history of hypertension (P<0.01), recent heart failure (P<0.001), diabetes (P<0.001), and left ventricular fractional shortening ≤25% (P<0.01; Table 1). Furthermore, moderate-to-severe versus normal or mild left ventricular dysfunction as assessed by 2D echocardiography (P< 0.01), peripheral vascular disease (P=0.005), ischemic heart disease (P<0.01), and increased body mass index (r=0.07, P=0.01) were also associated with increased vWF levels.
After adjustment for age, statistical significance (P<0.05) remained for all these variables except female sex (P>0.05; Table 1). There was no apparent association between vWF and systolic blood pressure >160 mm Hg (149 versus 146 IU/dL, P=0.17). Age, prior cerebral ischemia, recent heart failure, diabetes, and body mass index remained independently associated with increased vWF levels (r2 adjusted=9%; Table 2) in a multivariate model.
Increased sP-sel levels were univariately associated with diabetes (P=0.01), current smoking (P=0.004), and peripheral vascular disease (P<0.001), whereas lower sP-sel levels were found in women (P<0.001) and those with a history of prior cerebral ischemia (P<0.01; Table 1). Statistical significance (P<0.05) remained for these variables after adjustment for age. Adjustment for age also revealed a borderline significant relationship between sP-sel and ischemic heart disease (P=0.046; Table 1). However, only the original 5 factors emerged as independent associates of sP-sel levels (increased in association with diabetes, peripheral vascular disease, and current smoking and decreased in association with female sex and prior cerebral ischemia) in a multivariate model (r2 adjusted=4%; Table 2). Restriction of our analyses to groups stratified according to sex, age, presence of additional vascular diseases (ischemic heart disease, prior cerebral ischemia, or peripheral vascular disease), or use of hormone replacement therapy did not result in any appreciable change in the coefficients for vWF and sP-sel in the multivariate models (P=NS, data not shown).
Levels of vWF increased significantly with increasing stroke risk (r2 adjusted=3%, P<0.001) by the SPAF risk stratification criteria (Figure). Expected differences for the moderate- versus low-risk and high- versus low-risk groups, respectively, were 5.4 IU/dL (P=0.02) and 12.0 IU/dL (P<0.001). However, there was no difference in sP-sel levels between different strata of stroke risk (P=0.24).
Relationship of vWF and sP-sel to Antithrombotic Therapy
To assess for differences in vWF and sP-sel levels due to antithrombotic therapy, we compared values by assigned antithrombotic group in the high-risk participants with samples obtained at 3 months into the study (n=276). No differences in vWF (151 versus 149 IU/dL, P=0.6), sP-sel (35 versus 32 ng/mL, P=0.07), or any of the patient characteristics were observed between those randomized to adjusted-dose warfarin versus low-dose warfarin plus aspirin. The multivariate models were also refit to the data, adjusting for warfarin or aspirin at the time of sample, and no appreciable change in model coefficients occurred for either the vWF or the sP-sel model.
The aim of the present cross-sectional study was not to examine these markers as predictors of stroke but to examine the relationship between known stroke risk factors and separate, well-described potential mechanisms of thrombosis (endothelial damage/dysfunction or platelet activation) in AF. We believe the observed relationships reveal important differences between atherosclerotic and thromboembolic stroke risk factors.
Among a large cohort of patients with AF, 4 major clinical risk factors for stroke in AF (advancing age, prior cerebral ischemia, recent heart failure, and diabetes) were independently associated with an elevated plasma level of vWF (a marker of endothelial damage/dysfunction). Furthermore, a stepwise increase in vWF was found in association with increasing strata of thromboembolic risk when SPAF III risk stratification criteria were used. Conversely, although several atherosclerosis risk factors (male sex, current smoking, peripheral vascular disease, and diabetes) were independently associated with elevated levels of sP-sel (a marker of platelet activation), sP-sel was not associated with degree of thromboembolic risk in AF on the basis of the same risk stratification criteria. Indeed, a prior history of cerebral ischemia, the strongest single clinical predictor of future thromboembolism in AF,2 was independently associated with lower sP-sel levels.
Despite the recognition of several clinical and echocardiographic factors associated with increased stroke risk in AF,2 the mechanisms whereby these factors predispose to intra-atrial thrombogenesis remain poorly understood.3 Intra-atrial blood pool stasis is undoubtedly an important factor; in previous studies arising from SPAF III, age and hypertension were independent predictors of dense spontaneous echo contrast on transesophageal echocardiography,18 whereas age, systolic blood pressure, and a history of ischemic heart disease were independent predictors of reduced LAA flow velocities on transesophageal echocardiography.19 Both of these transesophageal echocardiography features reflect intra-atrial stasis and have been shown to be associated with increasing level of stroke risk (by the SPAF III risk stratification criteria)20 and to independently predict thromboembolism.19,21⇓
However, Virchow’s triad of thrombogenesis states that abnormalities of blood flow, vessel wall (endothelium), and intravascular coagulation factors must be present for thrombus formation to occur, and thus intra-atrial blood pool stasis alone may be insufficient to explain intra-atrial thrombogenesis. Furthermore, the effect on thromboembolic risk of factors such as diabetes and the anatomic location of complex aortic plaque (another transesophageal echocardiography feature independently predictive of thromboembolic events and localized LAA thrombus21 that is associated with age, hypertension, and diabetes22) remote from the LAA in the descending thoracic aorta support the possibility of additional mechanisms, such as a widespread abnormal endothelial or intravascular prothrombotic process, involved in LAA thrombogenesis.
P-selectin is a component of platelet α-granules that is expressed on the platelet surface membrane and shed into the plasma (as sP-sel) on platelet activation.23 P-selectin is also found in Weibel-Palade bodies of endothelial cells, but the bulk of circulating sP-sel appears to be platelet derived.13,24,25⇓⇓ Conversely, although vWF is similarly found in platelet α-granules and endothelial cell Weibel-Palade bodies, the majority of circulating vWF appears to be derived from the endothelium.13,25,26⇓⇓ An advantage of vWF and sP-sel as vascular markers for the present analysis is that they are relatively unaffected by use of antithrombotic therapy.7,27,28⇓⇓ Because of the design of SPAF III, antithrombotic therapy was often closely related to patient characteristics (including prior thromboembolism), and because few (5%) of the patients were taking no antithrombotic therapy at the time of the sample, confounding could be a problem for other vascular markers sensitive to antithrombotic therapy (such as β-thromboglobulin29 or fibrin D-dimer27,29⇓). However, although we again found no statistically significant effect of antithrombotic therapy on vWF and sP-sel in the present study, our ability to fully exclude small effects is limited by the cross-sectional design, and we cannot exclude the possibility that the nonsignificant trend toward lower sP-sel levels among high-risk patients treated with combination therapy compared with adjusted-dose warfarin might have reached statistical significance if sample sizes had been larger.
Our finding of an association between vWF levels and stroke risk stratification provides evidence of endothelial damage/dysfunction in AF patients at increased risk of stroke, although it should be recognized that the risk stratification scheme accounted for only 3% of the observed variability of vWF levels in our cohort (r2 adjusted=3%). Indeed, because the adjusted r2 values for the vWF and sP-sel multivariate models were each below 10%, it is important to understand that vWF and sP-sel levels may be affected by many unmeasured factors beyond the scope of the present study,30,31⇓ and the cross-sectional design of this study does not allow us to positively establish or refute either endothelial damage/dysfunction or platelet activation as mechanisms of thromboembolism in AF. However, previous studies have found increased endocardial expression of vWF in association with thrombus formation in the overloaded human atrial appendage,32 a correlation between raised plasma vWF levels and ultrastructural damage to the surface of the LAA endocardium in patients with mitral stenosis (many of whom were in AF),33 and plasma vWF as an independent predictor of the presence of LAA thrombus on transesophageal echocardiography (in patients with AF),34 which supports endothelial damage/dysfunction (or vWF itself) as a plausible mechanism/marker of increased thrombotic risk.
The absence of a relationship between sP-sel and stroke risk factors (except diabetes) in AF, the inverse relationship between sP-sel and a history of prior cerebral ischemia, and the strong association between sP-sel and several risk factors for atherosclerosis/atherothrombosis suggest that stroke risk factors in AF may promote left atrial thrombus formation by mechanisms quite different from the platelet-related mechanisms of atherothrombotic disease. However, an alternative marker of platelet activation (β -thromboglobulin) has also been shown to independently predict LAA thrombus on transesophageal echocardiography,34 yet it did not predict thromboembolic events among the SPAF III population.15 Prospective studies are therefore needed to fully evaluate vWF (endothelial damage/dysfunction) and sP-sel (platelet activation) as potential mechanisms/markers of stroke risk and cardiovascular outcome in AF.
An unfortunate limitation of the present study is that because of the design of the SPAF III trial, we lack a sinus rhythm control group from the same population and thus are unable to address the effect of AF itself on vWF and sP-sel levels. However, this large study was designed to assess the relationships between known thromboembolic risk factors and possible mechanisms of thrombosis in AF rather than the effect of AF per se on these mechanisms, although we note that vWF levels even in low-risk AF patients in the present study are higher than the usual healthy control levels obtained from our laboratory in previous studies.6,13,27⇓⇓ Previous smaller studies have found abnormal plasma markers of thrombogenesis, endothelial function (including vWF), and platelet activation (including sP-sel) among patients with AF compared with healthy controls in sinus rhythm,4–8⇓⇓⇓⇓ but our finding that several cardiovascular conditions exert important (and diverse) effects on levels of vWF and sP-sel in AF raises the possibility that some studies may have overstated the effect of AF itself by failing to adequately account for conditions that are commonly found among patients with AF.1,12⇓ Indeed, the community-based Framingham Offspring Study found that after adjustment for the presence of such additional factors, any effect of AF itself on prothrombotic markers disappeared12; yet even in that study, the number of AF cases was small (n=47), which makes interpretation difficult. Therefore, evidence is needed from large, well-matched case-control studies to establish the true relationship between AF itself and these prothrombotic plasma markers, in addition to the need for prospective studies to establish the true prognostic importance of vWF (endothelial damage/dysfunction) and sP-sel (platelet activation) as potential mechanisms of thromboembolic stroke in AF.
We acknowledge the support of the Dowager Countess Eleanor Peel Trust and the City Hospital Research and Development program. The SPAF-III investigators are listed in Reference 21.
- ↵Hart RG, Halperin JL. Atrial fibrillation and stroke: concepts and controversies. Stroke. 2001; 32: 803–808.
- ↵Gustafsson C, Blomback M, Britton M, et al. Coagulation factors and the increased risk of stroke in nonvalvular atrial fibrillation. Stroke. 1990; 21: 47–51.
- ↵Lip GYH, Lowe GD, Rumley A, et al. Increased markers of thrombogenesis in chronic atrial fibrillation: effects of warfarin treatment. Br Heart J. 1995; 73: 527–533.
- ↵Feinberg WM, Pearce LA, Hart RG, et al. Markers of thrombin and platelet activity in patients with atrial fibrillation: correlation with stroke among 1531 participants in the stroke prevention in atrial fibrillation III study. Stroke. 1999; 30: 2547–2553.
- ↵Stroke Prevention in Atrial Fibrillation Investigators. 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.
- ↵Asinger RW, Koehler J, Pearce LA, et al. Pathophysiologic correlates of thromboembolism in nonvalvular atrial fibrillation, II: dense spontaneous echocardiographic contrast (The Stroke Prevention in Atrial Fibrillation [SPAF-III] study). J Am Soc Echocardiogr. 1999; 12: 1088–1096.
- ↵Goldman ME, Pearce LA, Hart RG, et al. Pathophysiologic correlates of thromboembolism in nonvalvular atrial fibrillation, I: reduced flow velocity in the left atrial appendage (The Stroke Prevention in Atrial Fibrillation [SPAF-III] study). J Am Soc Echocardiogr. 1999; 12: 1080–1087.
- ↵Blackshear JL, Pearce LA, Hart RG, et al. Aortic plaque in atrial fibrillation: prevalence, predictors, and thromboembolic implications. Stroke. 1999; 30: 834–840.
- ↵Michelson AD, Barnard MR, Hechtman HB, et al. In vivo tracking of platelets: circulating degranulated platelets rapidly lose surface P-selectin but continue to circulate and function. Proc Natl Acad Sci U S A. 1996; 93: 11877–11882.
- ↵Li-Saw-Hee FL, Blann AD, Lip GYH. Effects of fixed low-dose warfarin, aspirin-warfarin combination therapy, and dose-adjusted warfarin on thrombogenesis in chronic atrial fibrillation. Stroke. 2000; 31: 828–833.
- ↵Lip GYH, Lip PL, Zarifis J, et al. Fibrin D-dimer and beta-thromboglobulin as markers of thrombogenesis and platelet activation in atrial fibrillation: effects of introducing ultra-low-dose warfarin and aspirin. Circulation. 1996; 94: 425–431.
- ↵Barbaux SC, Blankenberg S, Rupprecht HJ, et al. Association between P-selectin gene polymorphisms and soluble P-selectin levels and their relation to coronary artery disease. Arterioscler Thromb Vasc Biol. 2001; 21: 1668–1673.
- ↵Heppell RM, Berkin KE, McLenachan JM, et al. Haemostatic and haemodynamic abnormalities associated with left atrial thrombosis in non-rheumatic atrial fibrillation. Heart. 1997; 77: 407–411.