Consecutive AIS patients aged 18 years or more, eligible for thrombolysis, admitted to a single Stroke Center (Department of Neurology, Faculty of Medicine, University of Debrecen, Hungary) were enrolled in the study. All patients were within 4.5 h of their symptom onset at the time of admission. Patient enrollment lasted for 22 months starting in March 2011. Intravenous thrombolytic therapy was applied according to the European Stroke Organization (ESO) guidelines using rt-PA (Alteplase, Boehringer Ingelheim, Germany) (18). Inclusion and exclusion criteria of patients were identical to that of thrombolysis eligibility as described in the ESO 2008 guideline (18). The diagnosis of IS was based on clinical symptoms and brain imaging using computer tomography (CT) scan and CT angiography (CTA). Admission CTA was used to identify the level of vessel occlusion in every patient. A control CT was performed for every patient 24 h after thrombolysis. All CT images were analyzed by four different investigators blinded to the clinical state of the patients and the Alberta Stroke Program Early CT Scores (ASPECTS) were calculated (19). Neurological deficit of patients was determined by the NIHSS (20) at various time points: on admission and at 2 h, 24 h, and 7 days post-lysis. Stroke etiology was determined according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) criteria (21). Hemorrhagic events were classified as symptomatic or asymptomatic intracranial hemorrhage (SICH or aSICH, respectively) according to the European Cooperative Acute Stroke Study (ECASS) II criteria (22). For each patient a detailed list of clinical parameters was recorded including demographic characteristics, neurological status, time of symptom onset, cardiovascular risk factors (arterial hypertension, atrial fibrillation, hyperlipidemia, diabetes mellitus, smoking status), history of previous cardiovascular events, and medications. Patients were followed and long-term functional outcomes were determined at 90 days post-event using the modified Rankin Scale (mRS) (23).

The following outcomes were investigated: (1) short-term functional outcome at 7 days post-event: favorable outcome was defined as a decrease in NIHSS score by at least 4 points or to 0 by day 7, unfavorable outcome was defined as an increase in NIHSS score by at least 4 points by day 7. (2) The presence of therapy-associated SICH or aSICH was defined according to ECASS II criteria. (3) Long-term functional outcome at 90 days post-event: poor long-term outcome was specified as an mRS greater than 2.

The study was approved by the Ethics Committee of the University of Debrecen, Hungary. All patients or their relatives provided written informed consent.

Peripheral blood samples were drawn from patients into vacutainer tubes on three different occasions: upon hospital admission (before thrombolysis), immediately after the administration of rt-PA infusion (~1 h after the initiation of thrombolysis) and approximately 24 h after the administration of thrombolytic therapy. Routine laboratory tests were performed from blood samples taken before thrombolysis and included the measurements of complete blood count, serum ions, glucose levels, basic kidney function tests, liver function test, high-sensitivity C-reactive protein (hsCRP), and screening tests of coagulation (prothrombin time, activated partial thromboplastin time and thrombin time). Blood samples anticoagulated with sodium citrate, theophylline, adenosine and dipyridamole (Vacuette CTAD Tubes, Greiner Bio-One, Austria) were centrifuged at 1,220 g, room temperature for 15 min. Plasma aliquots were labeled with a code and stored at −70°C until further analysis of FVIII activity and VWF antigen levels. FVIII activity, determined by chromogenic method and VWF antigen level, determined by immunoturbidimetric assay were measured from coded plasma samples on a BCS coagulometer by standard methods (Siemens Healthcare Diagnostic Products, Marburg, Germany), by an investigator blinded to patient identification and clinical information. Batches of plasma samples (approximately 20 samples) were thawed for 10 min at 37°C followed by immediate analysis. According to the manufacturer, precision of the assays are the following: coefficient of variation within run (intra-assay precision) for FVIII activity assay: 4–9%, for VWF antigen level test: 1.4–4.2%; between run (inter-assay precision) for FVIII activity assay: 4–10%, for VWF antigen level test: 0.9–2.6%.

Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS, Release 22.0, Chicago, IL, USA) and GraphPad Prism 5.0 (GraphPad Prism Inc., La Jolla, CA, USA) softwares. Normality of the data was evaluated by the Shapiro–Wilk test. As FVIII activity and VWF antigen levels were not normally distributed at any time points measured, the Mann–Whitney U test was applied for all two-group analyses and the Kruskal–Wallis analysis with Dunn–Bonferroni post hoc test was used for multiple comparisons. Differences between categorical variables were assessed by the Fisher’s exact or χ² test. Friedman’s two-way ANOVA with Dunn–Bonferroni post hoc test was applied to investigate the effect of thrombolysis on FVIII activity and VWF levels. Strength of association between FVIII activity and VWF antigen levels was tested using Spearman’s correlation test. In order to test for differences between adjusted means, univariate analysis incorporating covariate testing (one-way ANCOVA) was performed after logarithmic transformation of data. Positive predictive values (PPVs) and negative predictive values (NPVs) of the studied parameters were assessed using contingency tables and the Fisher’s exact test. A binary backward logistic regression model was used to determine whether elevated FVIII and VWF levels of different time points are independent predictors of poor functional outcomes at 90 days post-event. Adjustment of the models were based on the results of previous statistical analyses (Mann–Whitney U test, Fisher’s exact, or χ² test), previous literature and methodological principles (dichotomized variables wherever possible). Results of the logistic regression analysis were expressed as odds ratio (OR) and 95% confidence interval (CI). A p-value of <0.05 was considered statistically significant.

During the study period, 131 consecutive AIS patients receiving intravenous rt-PA treatment were enrolled. Baseline demographic variables, main stroke characteristics and details of thrombolysis treatment are listed in Tables 1 and 2. In case of six patients, intravenous thrombolytic therapy was supplemented with intra-arterial thrombolysis using rt-PA according to standard protocol; the duration of thrombolysis and the final dose of rt-PA applied did not significantly differ for these patients. Median age of the patient cohort was 69 (IQR: 59–79) years, 60.3% were men. The most common cerebrovascular risk factor was arterial hypertension in this patient cohort (n = 100, 76.3%). Median time from symptom onset to treatment was 155 (IQR: 125–180) min. Median NIHSS before stroke treatment was 8 (IQR: 5–14). According to the TOAST criteria, etiology of stroke was most commonly large vessel disease (n = 49, 37.4%), followed by 27 (20.6%) patients with cardioembolic stroke. As detected by CTA on admission, 70 patients (53.4%) had a vessel occlusion, and 27 patients (20.6%) stenosis. Poor functional outcome at 7 days post-event was observed in 20 cases (15.3%), while poor outcome at 90 days (mRS ≥ 3) was observed in case of 51 (38.9%) patients. Therapy-associated intracranial hemorrhage was detected in 13 cases, of which 6 cases (4.6%) were symptomatic according to ECASS II. Mortality rates by day 7, 14, and day 90 post-event were 3.8, 6.1, and 22.1%, respectively.

The median and IQR values of FVIII activity and VWF antigen levels are shown at all investigated time points in Figures 1A,B, respectively. In the samples taken on admission, the median values of both hemostasis parameters were above the upper limit of the respective reference interval in the whole patient cohort (FVIII activity median: 188.0%, IQR: 153.0–242.0%, VWF antigen level median: 201.3%, IQR: 169.1–259.6%). FVIII activity dropped significantly in the samples obtained immediately after thrombolysis as compared to the initial values (median: 102.0%, IQR: 62.0–155.5%, p < 0.001) and showed an increase 24 h after the event (median: 166%, IQR: 130.0–209.0%, p = 0.014) (Figure 1A). On the contrary, VWF levels increased steadily post-lysis (median VWF levels immediately after lysis: 229.1%, IQR: 157.6–293.3%, at 24 h post-lysis: 231.6%, IQR: 176.8–284.8%, Friedman’s two-way ANOVA p = 0.002). Notably, VWF median and IQR values were above the upper limit of the reference interval at all investigated time points in the study cohort (Figure 1B).

Factor VIII activity and VWF antigen levels showed good correlation on admission (r = 0.748, p < 0.001) (Figure 2A), but no significant correlation was found immediately after thrombolysis (r = 0.093, p = 0.299, Figure 2B), most probably due to plasmin-mediated FVIII degradation. Fair correlation was observed between the two parameters in the samples obtained 24 h after thrombolysis (r = 0.420, p < 0.001, Figure 2C).

von Willebrand factor antigen levels were gradually and significantly elevated in case of more severe AIS (NIHSS 6–16 and NIHSS >16) at all investigated time points, but no such significant association was observed for FVIII activity levels (Table 3). The association between VWF antigen levels and stroke severity remained significant after adjustments for confounders (current smoking, hsCRP, age) in the statistical model.

The association of elevated VWF levels and more severe AIS was also proved as significantly higher VWF levels were found at all investigated time points in patients presenting with worse 24 h post-lysis CT scans (ASPECTS score 7–0) (Table 4). Similar association was observed for FVIII levels, except for the samples investigated immediately after lysis. Associations for VWF levels and FVIII activity remained significant after adjustments for confounders (current smoking, hsCRP, age) in the statistical model (Table 4). As CT scans on admission are not indicative of stroke severity and have less predictive values as the ASPECTS at 24 h post-lysis, it was not surprising that no significant association was found between the investigated hemostasis parameters and the ASPECTS on admission (Table 4). No association was found between FVIII activity and VWF antigen levels at any time points and stroke subtypes according to TOAST criteria (data not shown). FVIII activity at 24 h post-lysis was significantly elevated in patients with vessel occlusion (median: 175.0%, IQR: 151.5–227.0%) as compared to those with stenosis only (median: 137.0%, IQR: 98.5–175.0%) or without occlusion/stenosis (median: 142.0%, IQR: 115.0–177.0%) (p = 0.001) (Table S1 in Supplementary Material), while such association was not observed for VWF levels.

Factor VIII activity and VWF antigen levels were not associated at any investigated time points with short-term therapy outcomes as assessed by the changes in NIHSS score by day 7 post-lysis (data not shown).

No association was found between FVIII activity or VWF antigen levels at any investigated time points and therapy-associated ICH, except for higher VWF at 24 h post-lysis in patients presenting with SICH (median: 226.8%, IQR: 176.5–279.4% vs. median: 347.5%, IQR: 263.3–372.1% for no bleeding or aSICH vs. SICH, p = 0.017) (Figure S1 in Supplementary Material).

Poor functional outcome (mRS ≥3) at 90 days post-event was associated with traditional risk factors including advanced age, increased NIHSS on admission, elevated hsCRP, and the presence of diabetes/diabetes treatment (Table 5). Moreover, as expected, ASPECTS at 24 h post-lysis and the level of vessel occlusion as detected by CTA was also indicative of the long-term outcome. Among the hemostasis parameters investigated at various time points, elevated FVIII activity 24 h post-lysis and elevated VWF antigen level measured immediately after lysis and 24 h after therapy showed significant association with poor outcomes (Table 5). Both parameters, as measured immediately post-lysis and 24 h post-lysis conferred a significant PPV and NPV for poor functional outcomes (highest PPV: VWF 24 after thrombolysis: 0.83; 95% CI: 0.59–0.96, p = 0.009 and highest NPV: FVIII immediately after thrombolysis: 0.73; 95% CI: 0.50–0.89, p = 0.009).

A binary backward logistic regression model including age, gender, elevated hsCRP, active smoking, diabetes mellitus, and NIHSS >5 on admission revealed that a FVIII activity and VWF antigen level above the upper limit of the reference interval (168 and 160%, respectively) as measured immediately after lysis and 24 h after thrombolysis significantly and independently increase the risk of unfavorable functional outcomes at 90 days (Table 6). In this model, FVIII activity and VWF antigen levels on admission did not prove to have an independent prognostic value regarding poor functional outcomes at 90 days, while elevated FVIII and VWF levels immediately after thrombolysis conferred an independent OR: 7.09 (IQR: 1.77–28.38, p = 0.006) and OR: 6.31 (IQR: 1.83–21.7, p = 0.003), respectively. Elevated levels of both factors 24 h after lysis were also found to have a significant predictive value (OR: 4.67, IQR: 1.42–15.38, p = 0.011 for FVIII activity and OR: 19.02, IQR: 1.39–187.0, p = 0.012 for VWF antigen level). Besides these hemostasis parameters, only hsCRP >5.2 mg/L and NIHSS >5 on admission remained in the stepwise backwards regression analysis model as independent risk factors for poor outcomes at 90 days (OR: 4.85, 95% CI: 1.64–14.33, p = 0.004 and OR: 3.51, 95% CI: 1.17–10.57, p = 0.026, respectively).

Our study has limitations. The sample size is limited, but in the light of other studies on AIS patients treated by thrombolysis it is considered representative. Due to the limited number of patients with SICH and with poor outcomes in this cohort, despite the significant associations found, results presented here need to be verified by larger studies.

We did not investigate AIS patients who were not suitable for rt-PA therapy. In theory, measuring FVIII/VWF levels of patients receiving and not receiving rt-PA and comparing the results with outcomes might seem useful. However, due to the important baseline differences between the two groups (e.g., the group not receiving rt-PA might be highly heterogeneous regarding time window from symptom onset, baseline coagulation screening tests, effective anticoagulation, age, etc.), which may significantly affect the results, such comparison might fail to support relevant conclusions.

The lack of advanced neuro-imaging (e.g., perfusion and collateral circulation imaging) limits the application of this study.

Finally, as the study was designed to find potential biomarkers with predictive values for long-term outcomes following thrombolysis, we did not perform any functional characterization studies on the plasmin-mediated effect of FVIII and VWF proteins. In case of FVIII activity, we assumed, based on previous studies, that the reduction in activity levels as detected immediately after thrombolysis is due to plasmin-mediated degradation and did not perform any biochemical tests to prove this hypothesis. In case of VWF, functional activity tests, including ristocetin-induced activity tests, collagen-binding assay, multimer testing, etc. were not measured in the patient population. Future studies are required to elucidate whether the qualitative changes of both proteins following stroke and thrombolysis would have any pathophysiological relevance and prognostic value.