von Willebrand factor antigen, von Willebrand factor propeptide and ADAMTS13 in carotid stenosis and their relationship with cerebral micro-emboli

Summary

Background:

The relationship between von Willebrand factor antigen (VWF:Ag), von Willebrand factor propeptide (VWFpp), VWFpp/VWF:Ag ratio, ADAMTS13 activity and micro-embolic signal (MES) status in carotid stenosis is unknown.

Methods:

This prospective, multi-centre study simultaneously assessed plasma VWF:Ag, VWFpp levels and ADAMTS13 activity and their relationship with MES status in asymptomatic versus symptomatic ≥50–99% carotid stenosis patients. One-hour transcranial Doppler ultrasound of the middle cerebral arteries classified patients as MES+ve or MES-ve.

Results:

Data from 34 asymptomatic patients were compared with 43 symptomatic patients in the ‘early phase’ (≤4 weeks) and 37 patients in the ‘late phase’ (≥3 months) after TIA/ischaemic stroke. VWF:Ag levels were higher (P=0.049) and VWFpp/VWF:Ag ratios were lower (P=0.006) in early symptomatic than asymptomatic patients overall, and in early symptomatic vs. asymptomatic MES-ve subgroups (P≤0.02). There were no inter-group differences in VWFpp expression or ADAMTS13 activity (P≥0.05). VWF:Ag levels and ADAMTS13 activity (P≤0.048) decreased and VWFpp/VWF:Ag ratios increased (P=0.03) in symptomatic patients followed up from the early to late phases after TIA/stroke. Although there were no differences in the proportions of symptomatic and asymptomatic patients with blood group O, a combined analysis of early symptomatic and asymptomatic patients revealed lower median VWF:Ag levels in patients with blood group O versus those without blood group O (9.59 vs. 12.32 µg/mL, p 0.035).

Discussion:

VWF:Ag expression, a marker of endothelial +/− platelet activation, is enhanced in recently symptomatic versus asymptomatic carotid stenosis patients, including in MES-ve patients, and decreases with ADAMTS13 activity over time following atherosclerotic TIA/ischaemic stroke.

Introduction

Von Willebrand factor (VWF) is a multimeric plasma glycoprotein which is synthesised in vascular endothelial cells and megakaryocytes (Bongers et al. 2006; Nishio et al. 2004), and stored as a mixture of multimers in the α-granules of platelets, and as ultra-large multimers in Weibel-Palade bodies of endothelial cells (Dong et al. 2003). After formation, pre-pro VWF undergoes extensive post-translational modification and ultimately proteolysis in the trans-Golgi network leading to the production of mature von Willebrand factor antigen (VWF:Ag) and Von Willebrand Factor propeptide (VWFpp) (Giblin et al. 2008; Tobin et al. 2014). Upon secretion, VWFpp dissociates from VWF and circulates independently (Giblin et al. 2008). Endothelial activation results in VWF:Ag secretion which may bind to GP1b-IX-V or αIIbβ3 receptors on platelets, thus promoting platelet adhesion, aggregation and subsequent thrombosis, as well as leucocyte adhesion (De Meyer et al. 2012). If not consumed immediately, ultra-large VWF is cleaved by A Disintegrin-like And Metalloprotease with ThromboSpondin type I repeats 13 (ADAMTS13) into smaller, less adhesive multimers that circulate in plasma (McCabe et al. 2015; Nishio et al. 2004). Circulating VWF:Ag levels (Bongers et al. 2006; McCabe et al. 2015) and VWFpp levels (de Romeuf et al. 1998) may be influenced by several factors including systemic inflammation, blood type and ADAMTS13 activity. Because VWFpp has a shorter half-life of 2 to 3 hours compared with the half-life of 8 to 10 hours for VWF:Ag, VWFpp may be a potentially more sensitive marker of ‘acute endothelial activation’ than VWF:Ag in patients ischaemic cerebrovascular disease (CVD), including in the subgroup with carotid stenosis (Kinsella et al. 2014).

Elevated VWF:Ag levels have been observed in both the early (Bath et al. 1998; Bongers et al. 2006; Catto et al. 1997; Hanson et al. 2011; Kozuka et al. 2002; McCabe et al. 2004; Nadar et al. 2005; Tobin et al. 2014; van Schie et al. 2010) and late phases following a TIA or ischaemic stroke (Hanson et al. 2011; Kozuka et al. 2002; McCabe et al. 2004; Tobin et al. 2014; van Schie et al. 2010) compared with controls. Several prospective studies have also shown that plasma VWFpp levels may predict stroke risk (Folsom et al. 1999; Tzoulaki et al. 2007). However, to date, very few studies have assessed plasma VWF:Ag and VWFpp levels specifically in patients with symptomatic or asymptomatic carotid artery stenosis. A prior study by our group showed that VWFpp levels were higher in the early and late phases after TIA/ischaemic stroke in symptomatic patients with ≥50–100% stenosis than in asymptomatic patients with ≥50% carotid stenosis (Kinsella et al. 2014). VWF:Ag levels also decreased in symptomatic patients followed up from the early to late phase after symptom onset. Furthermore, VWFpp levels were higher in symptomatic patients without micro-emboli signals (MES) on Transcranial Doppler ultrasound (TCD) than in asymptomatic MES-negative patients.

A relatively recent population-based study revealed that subjects ≥55 years old with lower ADAMTS13 activity had an greater risk of incident stroke or TIA than those with higher levels (Sonneveld et al. 2015), but VWFpp levels were not analysed. One small hospital-based study revealed lower levels of ADAMTS13 activity in patients within 4 weeks of onset of ‘non-TTP related’ TIA or ischaemic stroke compared with normal controls (McCabe et al. 2015). However, to our knowledge, no study to date has simultaneously quantified VWF:Ag, VWFpp levels and ADAMTS13 activity in symptomatic versus asymptomatic carotid stenosis, including in MES+ve and MES-ve subgroups.

Therefore, the aims of this component of the HaEmostasis In carotid STenosis (HEIST) study were to determine whether:

1. There were differences in VWF:Ag levels, VWFpp levels, VWFpp/VWF:Ag ratios or ADAMTS13 activity in recently symptomatic compared with asymptomatic moderate-severe carotid stenosis patients;

2. Expression of these biomarkers changed in patients with recently symptomatic carotid stenosis during follow up from the early to the late phase after symptom onset or carotid intervention;

3. There was a relationship between VWF:Ag levels and VWFpp levels, and VWF:Ag levels and ADAMTS13 activity;

4. Observed differences in these markers varied between symptomatic and asymptomatic patients stratified according to MES status.

We hypothesised that:

1. VWF:Ag or VWFpp expression may be elevated and the VWFpp/VWF:Ag ratio or ADAMTS13 activity may be reduced in recently symptomatic versus asymptomatic carotid stenosis patients;

2. VWF:Ag or VWFpp would decrease and the VWFpp/VWF:Ag ratio or ADAMTS13 activity may increase in symptomatic patients followed up from the early to the late phase after symptom onset or carotid intervention;

3. There would be a positive relationship between VWF:Ag and VWFpp levels, and an inverse relationship between VWF:Ag levels and ADAMTS13 activity;

4. Based on pilot data from our lab (Kinsella et al. 2014), we hypothesised that one might see increased VWFpp or VWF:Ag expression in the subgroup of early symptomatic vs. asymptomatic carotid stenosis patients who were MES-ve on TCD monitoring.

Methods

Consecutive eligible patients ≥18 years old with asymptomatic or symptomatic moderate-severe (≥50–99%) carotid stenosis identified on colour Doppler ultrasound examination (CDUS) using standardised velocity criteria (Grant et al. 2003; Sidhu et al. 1997) and preferably confirmed on CT angiography (CTA) or MR angiography (MRA) (Grant et al. 2003; Silvennoinen et al. 2007) were invited to participate between October 2011 and February 2015 (Murphy et al. 2018). Patients were recruited from the Rapid Access Stroke Prevention clinics, vascular surgery and general neurology clinics, stroke service, and neurology and vascular surgery wards at AMNCH or St James’s Hospital.

Written informed consent (or proxy consent from a relative, where appropriate) was obtained in all cases. The study was approved by the Local Research Ethics Committee (Project/LREC Reference: 2011/31/02).

Inclusion/Exclusion criteria:

Asymptomatic patients:

As described previously, patients were included in the ‘asymptomatic carotid stenosis group’ if they were incidentally noted to have moderate (≥50–69%) or severe (≥70–99%) carotid stenosis on screening CDUS, and had never had TIA or stroke symptoms or had no history of TIA or stroke in the preceding 3 years e.g. following auscultation of a bruit or during a routine CDUS ‘screening examination’ (Murphy et al. 2018; Murphy et al. 2019). The degree of stenosis was confirmed on extracranial MRA or CTA as part of their routine clinical work-up in 32 of 34 (94%) asymptomatic patients.

Symptomatic patients:

Patients were included in the ‘early phase symptomatic carotid stenosis group’ if they had experienced a TIA or ischaemic stroke in the vascular territory supplied by a moderate (≥50–69%) or severe (≥70–99%) ipsilateral carotid artery stenosis within the preceding 4 weeks (Murphy et al. 2018; Murphy et al. 2019). All symptomatic patients had their stenosis confirmed on extracranial MRA or CTA to establish concordance between CDUS and another non-invasive imaging modality, and all met the TOAST classification criteria for ‘large artery atherosclerotic TIA/ischaemic stroke’(Adams et al. 1993). Symptomatic patients were followed-up and re-assessed ≥3 months after symptom onset or surgical/endovascular intervention (late phase) (Murphy et al. 2018; Murphy et al. 2019).

Exclusion criteria included ipsilateral carotid occlusion, active infection, inflammation or neoplasia; platelet count <120 or >450 × 10⁹/L; myocardial infarction, pulmonary embolism, deep vein thrombosis or major surgery within the preceding 3 months; prior primary intracerebral haemorrhage; known bleeding or clotting diathesis; ongoing unstable coronary or peripheral arterial disease; renal impairment (e.g. urea >10 mmol/l); or non-steroidal anti-inflammatory drug (NSAID) intake other than aspirin within the preceding 2 weeks. Patients were also subsequently excluded from the symptomatic group if there was evidence of a potential cardio-embolic source of embolism detected within 3 months of recruitment (Murphy et al. 2018; Murphy et al. 2019).

Clinical assessment:

As described previously, all patients underwent detailed neurovascular assessment by a neurology research registrar (SJM or STL) or supervising consultant vascular neurologist (DJHM) to confirm that asymptomatic patients met inclusion criteria, and to confirm a diagnosis of large artery atherosclerotic TIA or stroke in the symptomatic cohort (Murphy et al. 2018; Murphy et al. 2019). Detailed information regarding vascular risk factors was collected prospectively (Murphy et al. 2018; Murphy et al. 2019). Patients were asked to take their medication approximately 2 hours prior to their scheduled follow-up appointment, and the time from last dose of antiplatelet therapy to venepuncture was recorded. Details regarding antiplatelet regimens, dose and duration of therapy were recorded. Results of all available routine haematological (FBC, ESR), coagulation (PT/APTT) and biochemical tests (renal, liver, thyroid function and lipid profiles, CRP and blood glucose) were collected prospectively. CT and/or MRI brain was performed in all symptomatic patients. A chest radiograph, electrocardiograph (ECG), 24-hour Holter recording, and transthoracic or transoesophageal echocardiograph were obtained in all symptomatic patients (Murphy et al. 2018; Murphy et al. 2019).

Blood sampling and laboratory tests

All subjects were rested for at least 20 minutes, and venepuncture performed as previously described (Kinsella et al. 2013a; McCabe et al. 2004; Murphy et al. 2018). Platelet-poor plasma (PPP) was prepared by double centrifugation of three 0.105M (3.2%) buffered sodium citrate-anticoagulated blood samples at 2250G within one hour of ‘atraumatic venepuncture’. Double-spun PPP was recovered from the upper two-thirds of these samples and aliquoted into polypropylene tubes (Sarstedt®, Germany) which were immediately frozen at −70°C to −80°C. The bottom third of each sample, which was not considered to be double-spun, was also immediately frozen at −70°C to −80°C and stored for later analysis of ADAMTS13 activity. Samples were subsequently thawed for re-aliquoting, then refrozen and later thawed once more at 37°C for 20 minutes before analysis. The concentrations of VWF:Ag and VWFpp in each PPP sample were quantified as described previously (Kinsella et al. 2014; O’Donnell et al. 2005). In brief, polyclonal rabbit Anti-Human VWF antibody (DAKO, Denmark) was used as coating antibody and polyclonal rabbit anti-human VWF/HRP antibody (DAKO, Denmark) as the detection antibody for VWF:Ag. Anti-human VWF propeptide MW1939 Clone CLB-Pro 35 coating antibody and clone CLB-Pro 14.3 detection antibody (SanquinReagents) were used for VWFpp quantification. The ELISA assay result was measured by spectrophotometry at 450nm using a VERSA Max Tuneable Microplate Reader. VWF:Ag and VWFpp levels were recorded as µg/ml. ADAMTS13 activity was quantified with the FRET assay (Kokame et al. 2005) on a Varioskan Lux Plate Reader with λ_ex = 340nm and λ_em = 450 measured every 5 minutes between 0 and 60 minutes. The reaction rate was calculated by linear regression of fluorescence and expressed as a percentage ADAMTS13 activity relative to reference plasma.

Transcranial Doppler ultrasound:

Bilateral simultaneous 1-hour TCD recordings of the middle cerebral artery were performed with a Viassys Pioneer TC8080, as described previously (Kinsella et al. 2013b; Murphy et al. 2018; Murphy et al. 2019). Symptomatic patients were classified as ‘MES+ve’ if they had ≥1 MES detected ipsilateral to the stenosed carotid artery of interest. Asymptomatic patients were considered ‘MES+ve’ if ≥1 MES was detected ipsilateral to an asymptomatic ≥50% carotid stenosis. Other patients were deemed ‘MES-ve’ (Murphy et al. 2018; Murphy et al. 2019).

Statistical Methods

The primary analysis focused on ‘unmatched case-case/control comparisons’ between asymptomatic and early and late phase symptomatic carotid stenosis patients. Subgroup analyses were pre-planned for patients with severe (≥70–99%) stenosis and in patients stratified according to MES status. We also performed paired, longitudinal comparisons in a ‘nested longitudinal study’ in symptomatic patients followed up from the early to late phase after symptom onset or carotid intervention who had matched data at each timepoint. Paired or unpaired t-tests were used for comparison of parametric variables. The Wilcoxon signed rank and Wilcoxon rank sum tests were used for comparison of paired and unpaired non-parametric variables, and the Kruskal-Wallis rank sum test for comparison of multiple non-parametric variables, where appropriate. Chi-squared or Fisher exact tests were used to compare proportions between groups. Multiple linear regression analysis was performed to examine the potential influence of relevant independent variables on any observed differences between groups and associations between variables. P < 0.05 was considered statistically significant. All calculations were performed with Minitab version 16® (Minitab®, 2015).

Results

Figure 1 outlines details of patients who were initially recruited and subsequently included or excluded in our final analyses (Murphy et al. 2018). In total, data from 34 asymptomatic, 43 early phase symptomatic patients, and from 37 of these symptomatic patients who were followed-up to the late phase after symptom onset or intervention were analysed. As described previously, 33/43 symptomatic patients (77%) underwent successful carotid endarterectomy, 1 had a technically-unsuccessful carotid endarterectomy and proceeded to subsequent uncomplicated endovascular treatment and stenting (2%); the remaining 9 (21%) chose optimal medical management based on advice from their treating physician (Murphy et al. 2018; Murphy et al. 2019).

Figure 1:

Algorithm of symptomatic and asymptomatic carotid stenosis patients screened and subsequently included in or excluded from the study. AF = Atrial Fibrillation, NSAIDs = Non-steroidal anti-inflammatory drugs, DWI = Diffusion-weighted imaging, ICA = Internal carotid artery.

Early and late phase symptomatic patients were significantly younger than the asymptomatic patients (Table I). Nine (26.5%) asymptomatic patients had a remote history of a TIA or ischaemic stroke, but none had experienced any cerebrovascular symptoms in the carotid or any other cerebrovascular territory in the 3 years prior to recruitment. There was a higher prevalence of current smokers and a lower prevalence of diagnosed hyperlipidaemia amongst symptomatic than asymptomatic patients (Murphy et al. 2018). There were no other significant differences between symptomatic and asymptomatic groups (Tables I & II). Of note, because VWF:Ag levels may be influenced by blood group (Franchini et al. 2007), blood groups were analysed and the proportions of patients with blood group O were similar across groups. The mean duration of follow up in symptomatic patients was 101 +/−15.5 days (range: 87 – 155 days). Only one medically-treated symptomatic patient experienced a further dysphasic TIA during follow up. No other patients experienced recurrent cerebrovascular, cardiovascular or venous thrombotic outcome events during the peri-procedural period or during follow up (Murphy et al. 2018).

Table I:

Demographic and vascular risk factor profiles of patient groups at the time of initial study recruitment. Values are means (±SD), percentages with absolute values in parentheses, or medians [25th-75th percentile], unless otherwise specified. P values refer to comparisons between early and late symptomatic vs. asymptomatic carotid stenosis groups with chi-squared or Fisher exact tests.

Characteristic	Early Symptomatic (N = 43)	P Value	Late Symptomatic (N=37)	P Value	Asymptomatic (N=34)
Mean Age (years)	65 (± 8.51)	0.004	65.4 (±9)	0.017	71.9 (± 7.85)
Sex: % Female (number)	27.9% (12)	0.85	21.6% (8)	0.59	29.4% (10)
Stroke at Presentation *	51.2% (22)	N/A	48.6% (18)	N/A	N/A
Median Interval from Symptom Onset (days [range])	5 [1 – 28]	N/A	96 [84 – 179]	N/A	N/A
Prior TIA/Ischaemic Stroke	16.3% (7)	0.39	16.2% (6)	0.39	26.5% (9)
Ischaemic Heart Disease	27.9% (12)	0.89	32.4% (12)	0.80	29.4% (10)
Hypertension	74.4% (32)	0.27	70.3% (26)	0.16	85.3 (29)
Diabetes Mellitus	18.6% (8)	0.83	18.9% (7)	0.86	20.6% (7)
Prior CEA (any side)*	6.97% (3)	0.45	5.4% (2)	0.25	14.7% (5)
Prior DVT/PE	4.7% (2)	0.69	5.4% (2)	0.60	2.9% (1)
Peripheral Vascular Disease	4.7% (2)	0.50	5.4% (2)	0.49	0% (0)
Hyperlipidaemia*	76.7% (33)	0.056	82.4% (28)	0.048	94.1% (32)
Migraine	6.97% (3)	0.69	5.4% (2)	0.42	11.8% (4)
Smoking at Enrolment *	39.5% (17)	0.009	35.3% (20)	< 0.001	11.8% (4)
Ex-smoker	46.5% (20)	0.25	54.1% (20)	0.63	61.8% (21)
Never Smoker	13.95% (6)	0.25	10.8% (4)	0.13	26.5% (9)
Statin Therapy	90.7% (39)	0.94	86.5% (32)	0.71	91.2% (31)
Family History Stroke	39.5% (17)	0.64	37.8% (14)	0.80	32.4 (11)
Index event on Anti-Platelet Therapy *	37.2% (16)	N/A	43.2% (16)	N/A	N/A
Blood Group O *	58.1% (25)	0.48	54% (20)	0.33	67.6% (23)

^*Refers to characteristic at time of initial recruitment to the study. Significant P values highlighted in bold.

Table II:

Degree of stenosis and prescribed anti-platelet regimens in patients. P values relate to chi-squared or Fisher exact testing between asymptomatic and early and late symptomatic carotid stenosis groups. Values are percentages (absolute values) or absolute values.

Characteristic	Early Symptomatic (N = 43)	P Value	Late Symptomatic (N=37)	P Value	Asymptomatic (N=34)
Moderate stenosis (50–69%)	34.9% (15)	0.25	29.7% (11) *	0.095	50% (17)
Severe stenosis (70–99%)	65.1% (28)	0.25	70.3% (26) *	0.095	50% (17)
Aspirin Monotherapy	55.8% (24)	0.49	48.7% (18)	0.23	64.7% (22)
Aspirin-Dipyridamole Combination Therapy	30.2% (13)	0.44	35.1% (13)	0.197	20.6% (7)
Clopidogrel Monotherapy	2.3% (1)	0.58	2.7% (1)	0.60	5.9% (2)
Aspirin- Clopidogrel	11.6% (5)	0.46	13.5% (5)	0.43	5.9% (2)
Combination Therapy
No Antiplatelet Therapy	0% (0)	0.44	0% (0)	0.48	2.9% (1)
Median Daily Aspirin Dose (mg)	225	< 0.001	75	0.62	75

^*Refers to severity of baseline carotid stenosis at time of initial recruitment in late symptomatic groups, and not to distribution of stenosis severity during follow up.

VWF:Ag & VWFpp levels, VWFpp/VWF:Ag ratios and ADAMTS13 activity

Case - Case/Control Data:

VWF:Ag levels were significantly higher in early phase symptomatic than asymptomatic patients (12.53 vs. 10.57µg/mL; P = 0.049), but were similar in the late symptomatic and ‘late symptomatic post-intervention’ groups to those seen in the asymptomatic group (Table III). There was no difference in VWFpp expression between groups (P ≥ 0.15). However, the VWFpp/VWF:Ag ratio was significantly lower in early symptomatic (0.094 vs. 0.118; P = 0.006), but not in late symptomatic or late symptomatic post intervention patients compared with the asymptomatic cohort (P ≥ 0.37). There were no significant differences in ADAMTS13 activity between early symptomatic (134.71%; P = 0.79) or late symptomatic (127.84%; P = 0.33) and asymptomatic patients (132.74%; Table III).

Table III:

Comparison of VWF:Ag and VWFpp levels, VWFpp/VWF:Ag ratios and ADAMTS13 in asymptomatic versus early symptomatic, late phase symptomatic, and late symptomatic post-intervention patients. Values are means ±SD (range) or medians [25th - 75th percentiles] or absolute values. Significant P values in bold.

Marker	Early Symptomatic (N = 43)	Late Symptomatic (N = 37)	Late Symptomatic post-intervention (N = 34)	Asymptomatic (N=34)
VWF:Ag (µg/mL)	12.53 ± 5.56 (3.79 – 29.6)	10.02 ± 5.01 (3.48 – 18.38)	10.71 ± 6.09 (3.48 – 18.28)	10.47 ± 3.38 (5.8 – 18.8)
P	0.049	0.66	0.84
VWFpp (µg/mL)	1.09 [0.87 – 1.44]	1.04 [0.82 – 1.3]	1.02 [0.82 – 1.33]	1.17 [0.87 – 1.43]
P	0.59	0.15	0.14
VWFpp/VWF:Ag ratio	0.094 [0.075 – 0.117]	0.109 [0.08 – 0.145]	0.107 [0.084 – 0.147]	0.118 [0.103 – 0.138]
P	0.006	0.66	0.37
ADAMTS13 (%)	134.71 [117.3 – 161.1]	127.84 [110.19 – 148.04]	126.6 [108.44 – 148.34]	132.74 [119.66 – 163.85]
P	0.79	0.33	0.29

Nested Longitudinal Study in Symptomatic Patients:

There was a significant reduction in VWF:Ag levels in ‘matched samples’ obtained from early symptomatic patients who were followed up to the late phase after TIA/ischaemic stroke (11.36 vs. 9.05 µg/mL; P = 0.048). VWFpp levels did not change (P = 0.52), but the VWFpp/VWF:Ag ratio significantly increased during follow-up (0.095 vs. 0.109; P = 0.032). In the subgroup of symptomatic patients who underwent carotid endarterectomy or stenting, there were no significant reductions in VWF:Ag levels (P = 0.055), VWFpp levels (P= 0.4), or changes in the VWFpp/VWF:Ag ratio (P = 0.059) during longitudinal follow-up (Table V). However, ADAMTS13 activity significantly decreased during follow-up in symptomatic patients overall (143.7% vs. 130.12%; P = 0.02), and in those symptomatic patients who underwent carotid intervention (144.1% vs. 129.1%; P = 0.015) (Tables IV & V).

Table V:

Comparison of ‘matched’ VWF:Ag and VWFpp levels, VWFpp/VWF:Ag ratios and ADAMTS13 activity in early versus late symptomatic carotid stenosis patients who underwent carotid intervention. Values are either means ±SD (range), medians [25th - 75th percentile] or absolute values. Significant P values in bold.

Marker	Early Symptomatic Pre-intervention (N = 34)	Late Symptomatic Post-intervention (N = 34)	P
VWF:Ag (µg/mL)	11.41 [8.9 – 15.72]	9.27 [6.04 – 14.19]	0.055
VWFpp(µg/mL)	1.12 [0.87 – 1.42]	1.01 [0.82 – 1.31]	0.40
VWFpp/VWF:Ag ratio	0.094 [0.074 – 0.12]	0.107 [0.084 – 0.138]	0.06
ADAMTS13 (%)	144.15 ±5.08 (93 – 206.2)	129.11 ±5.60 (76 – 200.47)	0.015

Table IV:

Comparison of ‘matched’ VWF:Ag and VWFpp levels, VWFpp/VWF:Ag ratios and ADAMTS13 activity in the same symptomatic carotid stenosis patients followed up from the early to the late phases after symptom onset. Values are means ±SD (range) or medians [25th - 75th percentiles]. Significant P values in bold.

Marker	Early Symptomatic (N = 37)	Late Symptomatic (N = 37)	P
VWF:Ag (µg/mL)	11.36 [8.62 – 15.44]	9.045 [6.03 – 13.62]	0.048
VWFpp (µg/mL)	1.07 [0.87 – 1.36]	1.04 [0.82 – 1.30]	0.52
VWFpp/VWF:Ag ratio	0.095 [0.074 – 0.12]	0.109 [0.085 – 0.143]	0.03
ADAMTS13 (%)	143.7 ±4.89 (93 – 206)	130.12 ±5.14 (76 – 200.47)	0.02

Regression Analysis

There was a significant positive correlation between VWF:Ag and VWFpp levels in the early symptomatic (R² = 40.7%; P < 0.0001, Figure 1), late symptomatic (R² = 41.3%; P < 0.001) and asymptomatic cohorts (R² = 26.5%; P = 0.002). There were no significant correlations between ADAMTS13 activity and VWF:Ag levels in symptomatic or asymptomatic patients (R² ≤ 2.3%; P ≥ 0.35).

After controlling for differences in age and the proportion of active smokers at the time of enrolment between groups with multiple linear regression, VWF:Ag levels were still significantly higher (P = 0.018) and the VWFpp/VWF:Ag ratio remained lower (P = 0.015) in early symptomatic vs. asymptomatic patients; however, there was still no significant difference in VWFpp expression between groups (P = 0.93). There were no significant differences in VWF:Ag or VWFpp levels, or the VWFpp/VWF:Ag ratio between late symptomatic and asymptomatic groups after controlling for differences in age, hyperlipidemia and smoking status between groups (P ≥ 0.13).

Pre-planned Subgroup Analyses

Severe (≥ 70 – 99%) Carotid Stenosis:

There was no significant increase in VWF:Ag levels (13.26 vs. 10.51 µg/mL; P = 0.058), but there was a significantly lower VWFpp/VWF:Ag ratio in early symptomatic compared with asymptomatic severe (70–99%) carotid stenosis patients (0.085 vs. 0.134; P = 0.003) (Table VI). Furthermore, VWFpp levels were lower in late symptomatic compared with asymptomatic severe carotid stenosis patients (1.02 vs. 1.30µg/mL; P = 0.032), at which stage the vast majority (92%) of severe symptomatic patients had undergone successful carotid intervention, but the differences in the VWFpp/VWF:Ag ratio between these subgroups was not significant (P = 0.097).

Table VI:

Comparison of VWF:Ag and VWFpp levels, VWFpp/VWF:Ag ratios and ADAMTS13 in asymptomatic versus early symptomatic and late phase symptomatic subjects with ≥70% carotid stenosis. Values are means ±SD (range), medians [25th - 75th percentile] or absolute values. Significant P values in bold.

Marker	Early Symptomatic 70–99% stenosis (N = 28)	Late Symptomatic 70–99% stenosis (N = 26)	Asymptomatic 70–99% stenosis (N=17)
VWF:Ag (µg/mL)	13.26 ± 5.88 (4.87 – 24.7)	9.88 ± 4.40 (3.75 – 20.25)	10.51 ± 3.59 (5.8 – 18.37)
P	0.058	0.61
VWFpp (µg/mL)	1.04 [0.83 – 1.28]	1.02 [0.83 – 1.22]	1.30 [1.05 – 1.90]
P	0.094	0.03
VWFpp/VWF:Ag ratio	0.085 [0.064 – 0.124]	0.107 [0.083 – 0.130]	0.134 [0.104 – 0.153]
P	0.003	0.097
ADAMTS13 (%)	132.94 [112.1 – 168.3]	125.36 [106.8 – 140.7]	127.16 [119.9 – 154.81]
P	0.59	0.57

MES-Positive and MES-Negative Subgroups:

As previously reported, 28 of 34 asymptomatic patients, 35 of 44 early phase symptomatic and 30 of 37 late phase symptomatic patients had TCD data available for analysis (Murphy et al. 2018; Murphy et al. 2019). The proportion of patients who were MES+ve was higher in early symptomatic (28.5%, N = 10, P = 0.049) than asymptomatic patients (7.1%, N = 2), but had fallen to similar levels to the asymptomatic group by the late phase after symptom onset (6.7%, N = 2; P = 0.996) (Murphy et al. 2018; Murphy et al. 2019). There were no significant differences in any of these biomarkers between early symptomatic and asymptomatic MES+ve patients, but the number of patients included in this analysis was very limited (Table VIIa). However MES-ve early symptomatic patients had significantly higher VWF:Ag levels (12.72 vs. 10.12 µg/mL; P = 0.023) and a lower VWFpp/VWF:Ag ratio (0.094 vs. 0.115; P = 0.008) than asymptomatic patients (Table VIIb). There were no differences in VWFpp levels or ADAMTS13 activity in MES-ve early symptomatic vs. asymptomatic patients (P ≥ 0.05, Table VIIb).

Table VII:

VWF:Ag, VWFpp, VWFpp/VWF:Ag ratios and ADAMTS13 in early symptomatic versus asymptomatic (A) MES+ve and (B) MES −ve patients. Values are means ±SD (range) or medians [25th - 75th percentile]. Significant P values in bold.

Table VIIa Marker	Early Symptomatic MES +ve (N = 10)	Asymptomatic MES +ve (N = 2)	P value
VWF:Ag (µg/ml)	10.3 [5.8 – 16.4]	13.42	0.74
VWFpp (µg/ml)	0.94 [0.82 – 1.32]	1.75	0.33
VWFpp/VWF:Ag ratio	0.102 [0.073 – 0.153]	0.126	0.75
ADAMTS13 (%)	148.6 [132 – 175]	133.7	0.59
Table VIIb Marker	Early Symptomatic MES −ve (N = 25)	Asymptomatic MES −ve (N = 26)	P value
VWF:Ag (µg/ml)	12.72 ± 0.94 (6.2 – 24.4)	10.12 ± 0.57 (5.8 – 16.1)	0.02
VWFpp (µg/ml)	1.06 [ 0.82 – 1.27]	1.125 [0.86 – 1.32]	0.45
VWFpp/VWF:Ag ratio	0.094 [0.073 – 0.113]	0.115 [0.098 – 0.128]	0.008
ADAMTS13 (%)	122.41 [112 – 165.5]	142.43 [119.6 – 171.6]	0.42

Discussion

This innovative, prospective case-case/control and nested longitudinal study, in a very carefully clinically-phenotyped cohort of carotid stenosis patients, has revealed elevated VWF:Ag levels and a reduced VWFpp/VWF:Ag ratio in early symptomatic compared with asymptomatic carotid stenosis. The findings are robust and persisted after controlling for the potential influence of age, prevalence of hyperlipidemia or smoking status on our results. This study is in keeping with data from a prior pilot study which showed that ‘adjusted VWF:Ag levels’ were higher in early symptomatic than asymptomatic patients, thus confirming that there is evidence of increased endothelial activation in recently symptomatic compared with asymptomatic moderate-severe carotid stenosis patients (Kinsella et al. 2014).

In contrast to the prior data, we did not confirm a significant increase in VWFpp levels in early or late symptomatic vs. asymptomatic carotid stenosis patients (Kinsella et al. 2014). Therefore, we cannot confirm the prior suggestion by our group that VWFpp may be a more sensitive marker of acute endothelial activation than VWF:Ag in patients with carotid stenosis. In an attempt to explore reasons for differences in VWFpp findings between the Platelets And Carotid Stenosis (PACS) study and this study, we investigated whether there were any potential differences in study design and patient profiles between the two studies. Although the main design of both the PACS and HEIST studies was very similar, we did not include patients with carotid occlusion in HEIST, and there was a higher proportion of early symptomatic patients treated with statins in HEIST than in PACS (91% vs. 72%; chi square test: P = 0.018). Statins have been reported to decrease plasma VWF:Ag levels (Sahebkar et al. 2016), but to our knowledge, the effects of statins on VWFpp expression in patients with carotid stenosis is not known. Therefore, one cannot comment on whether or not differences in statin use may have influenced the observed difference in results between the PACS and HEIST studies.

In keeping with one of our a priori hypotheses, there was a significant reduction in VWF:Ag levels and increase in the VWFpp/VWF:Ag ratio over time in symptomatic patients who were followed up from the early to the late phases after TIA/ischaemic stroke. The non-significant reduction in VWF:Ag levels in the symptomatic subgroup who underwent intervention may well reflect a type II error because the P value of 0.055 was approaching the pre-specified level for statistical significance (Table V). VWFpp levels in the subgroup of late symptomatic severe carotid stenosis patients actually fell to below those seen in patients with asymptomatic severe stenosis. This may partly reflect resolution of the acute phase response, successful surgical/endovascular interventional treatment of the majority (92%) of symptomatic patients, but could also reflect a type I error due to the smaller number of patients involved in this subgroup analysis. Further work is required to readdress this issue.

There were no significant differences in endothelial activation markers between early symptomatic and asymptomatic MES+ve patients, but the number of patients included in this subgroup analysis was far too small to make any definitive conclusions. However, in the subgroup of MES-ve patients, early symptomatic patients had higher VWF:Ag levels and a lower VWFpp/VWF:Ag ratio than asymptomatic patients. The findings of increased endothelial activation in this subgroup of recently symptomatic MES-ve patients, in conjunction with our prior platelet biomarker data (Murphy et al. 2018; Murphy et al. 2019), may partly explain the pathogenesis of first or recurrent TIA/stroke in this patient subgroup. However, as stated above, because the majority of our symptomatic patients underwent successful carotid intervention and optimisation of secondary preventive therapy, this study does not enable us to comment on whether these findings are reflective of an acute phase response to recent cerebral or ocular ischaemia or infarction, or whether they were present before symptom onset. There is limited peer-reviewed literature on the relationship between VWF levels and MES on TCD. Markus et al. showed that the ARC1779 aptamer, which inhibits the pro-thrombotic function of von Willebrand factor by binding to its A1 domain, was associated with significantly reduced VWF activity and reduced MES compared with placebo in patients undergoing carotid endarterectomy (Markus et al. 2011). The PACS study did not find any significant differences in VWF:Ag levels, but VWFpp levels were higher in recently symptomatic versus asymptomatic MES-ve patients (Kinsella et al. 2014), a finding not replicated in this study.

In contrast to our a priori hypothesis, we did not find any significant differences in ADAMTS13 activity between early or late symptomatic and asymptomatic patients. However, ADAMTS13 activity levels fell in matched symptomatic patients between the early and late phases after TIA or stroke. The reduction in circulating VWF:Ag levels, which may also have arisen due to resolution of the acute phase response or more intense medical, surgical or endovascular intervention during follow-up, may have led to a secondary reduction in ADAMTS13 activity; however, information on ADAMTS13 activity regulation is limited (Zheng 2013). One of the limitations of this study is that we did not simultaneously assess ADAMTS13 antigen or antibody levels, but we do not think that it is likely that an acquired deficiency of ADAMTS13 antigen developed in this ‘non-TTP related’ large artery atherosclerotic TIA/ischaemic stroke patient population over time (McCabe et al. 2015).

In agreement with other studies in patients with ‘vascular disease’ (Blann et al. 1998; de Jong et al. 1997; Vischer et al. 1998), there was a significant positive correlation between VWF:Ag and VWFpp levels in all patient groups, thus adding to the literature on this topic in early and late symptomatic and asymptomatic carotid stenosis patients. Based on these data, if one is examining profiles of endothelial and/or platelet activation status in larger cohorts of patients with symptomatic and asymptomatic carotid stenosis, we advise examining all of these parameters, including the VWFpp/VWF:Ag ratio, because one or all of these biomarkers may be informative. We found no relationship between ADAMTS13 activity on the FRET assay and VWF:Ag levels which were quantified by ELISA. One prior study found evidence of reduced ADAMTS13 activity in the early but not late phase after TIA/ischaemic stroke overall versus healthy controls (McCabe et al. 2015). There was also a significant inverse relationship between ADAMTS13 activity on a collagen-binding assay and VWF:Ag levels on an automated latex agglutination assay in the early phases after TIA or ischaemic stroke; however, only 21% of patients had large artery atherosclerotic disease in that study (McCabe et al. 2015). Because collagen-binding ADAMTS13 assays utilise large VWF multimers which may be more commonly observed under conditions of higher shear stress (Gogia et al. 2015), a collagen-binding assay might be more sensitive than the FRET assay at detecting differences between asymptomatic and symptomatic carotid stenosis cohorts. Further studies with simultaneous assessment of ADAMTS13 activity using both methodologies would be needed to clarify this issue, and may be clinically relevant given the predictive role of ADAMTS13 in tPA-induced recanalisation after stroke (Bustamante et al. 2018).

This study had some other limitations. The HEIST study was not designed to assess the immediate ‘peri-procedural profile’ of these biomarkers in symptomatic patients undergoing urgent carotid intervention, because we prospectively planned to reassess symptomatic patients at least 3 months after symptom onset or intervention when the acute phase response from surgery or intervention had settled. Due to our relatively limited sample size and the fact that only one patient had a recurrent TIA during our dedicated, robust study follow-up period, we cannot comment on the value of endothelial activation markers in predicting the risk of recurrent events in this study population at present, but the predicative value of VWF:Ag has been shown by others (De Meyer et al. 2012; Sonneveld et al. 2013). Because VWF:Ag may be endothelium-or platelet-derived (Gragnano et al. 2017), we cannot comment on the relative contribution of these cell types to circulating levels of VWF:Ag in our study population, but prior data indicate that VWF:Ag is predominantly a marker of endothelial activation (Hollestelle et al. 2006).

Conclusions

These data enhance our understanding of the profiles of VWF:Ag, VWFpp and ADAMTS13 activity, and the relationship between these biomarkers in patients with carotid stenosis. There is now reproducible evidence of increased endothelial activation, as well as increased platelet activation, platelet production/secretion and/or turnover in patients with recently symptomatic compared with asymptomatic carotid stenosis despite treatment with modern secondary preventive therapy (Murphy et al. 2018; Murphy et al. 2019). Because these endothelial biomarkers decrease over time following symptom onset, it is difficult to be certain whether the longitudinal findings mainly represent resolution of an acute phase response to TIA, stroke or plaque rupture, a response to medical or surgical/endovascular intervention, or are pre-existing and ‘inherent’ in a high risk subgroup of patients with carotid stenosis who go on to develop symptoms. However, combining clinical, endothelial and platelet biomarker data with data on MES status improves our understanding of the pathogenesis of cerebrovascular events and could clearly assist in risk-stratification in this important population of CVD patients, including in the MES-ve carotid stenosis subgroup. Such comprehensive profiling has the potential to target patients who might benefit from earlier, more intensive medical or interventional treatment to prevent subsequent TIA or stroke.

Figure 2:

Relationship between VWF:Ag (µg/mL) and VWFpp (µg/mL) levels in early symptomatic patients (P < 0.0001).