2025 年 32 巻 11 号 p. 1416-1424
Aims: Branch atheromatous disease (BAD)-related stroke, caused by atherosclerotic occlusion at the origin of a deep penetrating artery, are prone to early neurological deterioration (END). This study aimed to assess the association between hemostatic activation markers and occurrence of END in patients with BAD-related stroke.
Methods: This prospective observational study included 88 patients with BAD-related stroke within 7 days of onset. On admission, plasma beta-thromboglobulin (beta-TG), platelet factor 4 (PF4), and D-dimer levels were measured. END was defined as an increase of ≥ 2 points in the total National Institutes of Health Stroke Scale (NIHSS) score or ≥ 1 point in the motor items of the NIHSS within 7 days of admission.
Results: Of the 88 patients, 34 (38.6%) experienced END. Mean beta-TG (158 ng/mL vs. 102 ng/mL; P = 0.021), PF4 (61 ng/mL vs. 35 ng/mL; P = 0.024), and D-dimer (2.0 µg/mL vs. 1.2 µg/mL; P = 0.037) levels were significantly higher in patients with END than in those without END. Multivariate analysis revealed that beta-TG and PF4 levels were independently associated with the occurrence of END, with an adjusted odds ratio per 10 ng/mL increase (95% confidence interval) of 1.09 (1.01-1.20) and 1.21 (1.02-1.49), respectively. In contrast, D-dimer levels were not independent predictors. The optimal cutoff values for predicting END were 130 and 55 ng/mL for beta-TG and PF4, respectively.
Conclusions: Elevated beta-TG and PF4 levels were independent predictors of END in patients with BAD-related stroke. Hence, the measurement of these platelet activation markers helps improve the risk assessment of BAD-related stroke and may provide management implications.
Intracranial branch atheromatous disease (BAD), first proposed in 1989 by Caplan, is a pathological concept characterized by stenosis or occlusion at the origin of a deep penetrating artery1). Unlike classical lacunar infarcts caused by lipohyalinotic degeneration, BAD is caused by atherosclerotic plaque in the vicinity of the orifice of the parental artery, resulting in infarction in the entire area of the perforating artery. Typical BAD-related stroke presents an infarct size of ＞15 mm with no severe stenosis in the parental artery, which corresponds to “stroke of undetermined cause” in the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification system2). Because routine imaging techniques are usually unable to depict microatheroma in the perforating artery, BAD is almost indirectly diagnosed based on some morphological features of the ischemic lesion presumed to be produced by it3). Such difficulties in diagnosing BAD may have hindered clear definition, data collection, or large clinical trials of BAD-related stroke, and consequently, there is still a critical knowledge gap.
From a clinical perspective, the high incidence of early neurological deterioration (END) is a major concern in BAD3, 4). For example, observational studies from Japan reported that 40%-46% of patients with BAD-related stroke experience early neurological worsening within several days after symptom onset, which is significantly higher than 11%-19% incidence in patients with lacunar stroke5, 6). Because END is strongly associated with more severe disbilities in the chronic phase7), reliable predictors of END would be informative for risk assessment and management implications. Although the mechanisms underlying END are yet incompletely understood, it is plausible that the thrombus growth can play a central role, leading to the occlusion of the perforator orifice, distal embolization, or hypoperfusion4). Beta-thromboglobulin (beta-TG) and platelet factor 4 (PF4) are known to serve as markers for platelet activation8), and D-dimer for a coagulation and fibrinolysis activation8, 9). To date, limited data are available on the association between these hemostatic markers and END10, 11).
This study aimed to assess the predictive value of beta-TG, PF4, and D-dimer for the occurrence of END in patients with acute BAD-related stroke.
The Tokyo Women’s Medical University (TWMU) Stroke Registry is an ongoing, prospective, single-center, observational study that enrolled patients ≥ 20 years of age with acute ischemic stroke or high-risk transient ischemic attack within 1 week of onset, hospitalized at our center12, 13). This study adhered to the ethical principles of the 1975 Declaration of Helsinki, the Ethical Guidelines for Epidemiological Research by the Japanese government, and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. The study protocol was approved by the Ethics Committee of TWMU Hospital (approval no. 2955-R2). Written informed consent was obtained from all the patients. The TWMU Stroke Registry is registered at https://upload.umin.ac.jp (UMIN000031913).
The present study was a retrospective analysis of prospectively collected data. Patient data were collected on demographics, clinical symptoms during the qualifying event, medical history, investigations (including standard blood chemistry, brain and cerebral artery imaging, and cardiac workup), management (medical treatment, revascularization procedure, and surgery), and occurrence of clinical events after the qualifying event using a structured case report form. Upon admission, plasma levels of beta-TG and PF4 were measured in 79 and 79 patients, respectively, using an enzyme-linked immunosorbent assay (Asserachrom, Diagnostica Stago, Asniere sur Seine, France). D-dimer levels were measured in 88 patients using the latex agglutination method (Lias Auto D-dimer Neo, Sysmex, Kobe, Japan). Neurological symptoms were assessed using National Institutes of Health Stroke Scale (NIHSS) scores. NIHSS scores were measured twice daily. Treatment decisions during the acute phase were always made by board-certified neurologists in accordance with the guideline recommendations14).
Definitions of BAD-Related Stroke and ENDBAD was diagnosed on the basis of findings from magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA), according to the previously reported criteria15). Briefly, BAD-related stroke was defined as having either (i) a single acute ischemic lesion more than 15 mm in diameter (i.e., more than 3 axial pictures at a slice thickness of 5 mm) in the lenticulostriate artery (LSA) region or (ii) a single acute ischemic lesion extended from the ventral pontine surface to the deep pons in the paramedian pontine artery (PPA) region, with no evidences of significant (＞50%) stenosis in the responsible parental artery, high-risk cardioembolic sources (e.g., atrial fibrillation), or other specific causes for stroke. END was defined as an increase of ≥ 2 points in total NIHSS score or ≥ 1 point in the motor items of the NIHSS within 7 days after admission16). END did not include transient worsening, but was defined as persistent deterioration.
Statistical AnalysisQuantitative variables were expressed as mean (standard deviation) in cases of normal distribution or median (interquartile range). Qualitative variables were expressed as frequencies (percentages). Bivariate comparisons were conducted using the t-test or Mann–Whitney U test for quantitative variables and χ2 test for categorical variables. We established cut-off values for each of the three hemostatic markers based on receiver operating characteristic (ROC) curves. The area under the curve, also referred to as the C-statistic, represent the ability of the scoring system to correctly predict the occurrence of END. Multivariate logistic regression analyses were performed to assess the predictive value of each of the three hemostatic markers for END. Age, sex, and all variables with P＜0.10 in the univariate analysis (body mass index, systolic blood pressure, diastolic blood pressure, and low-density lipoprotein cholesterol levels) were entered into the models. In each model, analyses were performed using the three hemostatic markers as continuous variables and binary variables dichotomized by the optimal cutoff values in the ROC analyses. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. For all analyses, statistical significance was set at P＜0.05.
Of the 1180 patients consecutively assessed for eligibility between December 2013 and December 2021, 25 patients who met the exclusion criteria (final diagnosis of stroke mimic, enrollment more than 1 week after stroke onset, missing data due to transfer to another hospital, or duplicate enrollment) were excluded. Subsequently, 96 patients with TIA and 971 patients who did not meet the diagnostic criteria for BAD-related stroke were excluded, leaving 88 patients with BAD-related stroke included in the analysis (Supplemental Fig.1).
BAD indicates branch atheromatous disease; TIA, transient ischemic attack; TWMU, Tokyo Women’s Medical University.
Of the 88 patients (mean age, 69.7 years; male, 58.0%), 34 (38.6%) had END. Table 1 shows the baseline characteristics of patients with and without END. Patients with END were older and had lower blood pressure and lower low-density lipoprotein cholesterol levels than those without END. The admission NIHSS scores did not differ between the two groups.
| Overall (n = 88) | No END (n = 54) | END (n = 34) | P value | |
|---|---|---|---|---|
| Age, years, mean (SD) | 69.7 (13.9) | 66.7 (14.7) | 74.4 (11.1) | 0.011 |
| Male, n (%) | 51 (58.0) | 34 (63.0) | 17 (50.0) | 0.23 |
| BMI, kg/m2, mean (SD) | 23.1 (3.9) | 23.7 (4.1) | 22.2 (3.4) | 0.077 |
| Medical history, n (%) | ||||
| Hypertension | 72 (81.8) | 45 (83.3) | 27 (79.4) | 0.64 |
| Dyslipidemia | 46 (52.2) | 30 (55.6) | 16 (47.1) | 0.44 |
| Diabetes mellitus | 35 (39.8) | 20 (37.0) | 15 (44.1) | 0.51 |
| Chronic kidney disease | 21 (23.9) | 14 (25.9) | 7 (20.6) | 0.56 |
| Chronic heart failure | 4 (4.5) | 2 (3.7) | 2 (5.9) | 0.64 |
| Stroke | 11 (12.5) | 6 (11.1) | 5 (14.7) | 0.62 |
| Coronary artery disease | 4 (4.5) | 3 (5.6) | 1 (2.9) | 0.56 |
| Peripheral artery disease | 4 (4.5) | 3 (5.6) | 1 (2.9) | 0.56 |
| Current smoking | 23 (26.1) | 16 (29.6) | 7 (20.6) | 0.34 |
| Excessive alcohol | 3 (3.4) | 1 (1.9) | 2 (5.9) | 0.32 |
| Systolic BP, mm Hg, mean (SD) | 165 (28) | 169 (30) | 159 (25) | 0.088 |
| Diastolic BP, mm Hg, mean (SD) | 91 (17) | 94 (19) | 86 (11) | 0.028 |
| Laboratory data | ||||
| LDL-C, mg/dL, mean (SD) | 125 (29) | 131 (31) | 115 (22) | 0.011 |
| HDL-C, mg/dL, mean (SD) | 59 (21) | 60 (21) | 58 (23) | 0.72 |
| TG, mg/dL, median (IQR) | 110 (83-152) | 112 (87-157) | 106 (74-142) | 0.25 |
| Glucose, mg/dL, mean (SD) | 126 (40) | 121 (42) | 133 (37) | 0.19 |
| HbA1c, %, median (IQR) | 6.0 (5.7-6.9) | 5.9 (5.6-6.5) | 6.1 (5.8-7.5) | 0.18 |
| Initial NIHSS, median (IQR) | 4 (3-6) | 4 (2-5) | 4 (3-7) | 0.14 |
| Ischemic lesion, n (%) | ||||
| Lenticulostriate artery | 57 (64.8) | 37 (68.5) | 20 (58.8) | 0.36 |
| Paramedian pontine artery | 31 (35.2) | 17 (31.5) | 14 (41.2) | 0.36 |
BMI indicates body mass index; BP, blood pressure; END, early neurological deterioration; HbA1c, hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range; LDL-C, low-density lipoprotein cholesterol; NIHSS, National Institute of Health Stroke Scale; SD, standard deviation; and TG, triglycerides.
Table 2 shows comparisons in baseline characteristics by stroke location. Compared to patients with BAD-related stroke in LSA region, those in PPA region tended to be older and were significantly more often had diabetes mellitus and lower high-density lipoprotein cholesterol levels. The frequency of END did not differ between BAD-related strokes with LSA and PPA.
| Ischemic lesion | P value | ||
|---|---|---|---|
| LSA (n = 57) | PPA (n = 31) | ||
| Age, years, mean (SD) | 67.6 (14.8) | 73.5 (11.3) | 0.055 |
| Male, n (%) | 31 (54.4) | 20 (64.5) | 0.36 |
| BMI, kg/m2, mean (SD) | 23.1 (4.1) | 23.1 (3.5) | 0.92 |
| Medical history, n (%) | |||
| Hypertension | 49 (86.0) | 23 (74.2) | 0.18 |
| Dyslipidemia | 34 (59.7) | 12 (38.7) | 0.060 |
| Diabetes mellitus | 15 (26.3) | 20 (64.5) | ＜0.001 |
| Chronic kidney disease | 16 (28.1) | 5 (16.1) | 0.20 |
| Chronic heart failure | 3 (5.3) | 1 (3.2) | 0.65 |
| Stroke | 4 (7.0) | 7 (22.6) | 0.040 |
| Coronary artery disease | 3 (5.3) | 1 (3.2) | 0.65 |
| Peripheral artery disease | 2 (3.5) | 2 (6.5) | 0.54 |
| Current smoking | 14 (24.6) | 9 (29.0) | 0.65 |
| Excessive alcohol | 2 (3.5) | 1 (3.2) | 0.94 |
| Systolic BP, mm Hg, mean (SD) | 167 (31) | 162 (23) | 0.43 |
| Diastolic BP, mm Hg, mean (SD) | 93 (19) | 87 (12) | 0.10 |
| Laboratory data | |||
| LDL-C, mg/dL, mean (SD) | 129 (30) | 118 (26) | 0.11 |
| HDL-C, mg/dL, mean (SD) | 64 (20) | 50 (22) | 0.003 |
| TG, mg/dL, median (IQR) | 104 (882-153) | 123 (86-152) | 0.57 |
| Glucose, mean (SD) | 120 (39) | 135 (40) | 0.10 |
| HbA1c, %, median (IQR) | 5.9 (5.6-6.4) | 6.2 (5.8-8.4) | 0.017 |
| Initial NIHSS, median (IQR) | 4 (2-6) | 4 (3-6) | 0.94 |
| END, n (%) | 20 (35.1) | 14 (45.2) | 0.36 |
BMI indicates body mass index; BP, blood pressure; END, early neurological deterioration; HbA1c, hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range; LDL-C, low-density lipoprotein cholesterol; LSA, lenticulostriate artery; NIHSS, National Institute of Health Stroke Scale; PPA, paramedian pontine artery; SD, standard deviation; and TG, triglycerides.
The medications administered before admission and acute treatment are summarized in Table 3. The usage rates of antithrombotic agents, intravenous alteplase, and statins did not differ between the patients with and without END. More than 85% of patients received dual antiplatelet therapy (DAPT) in both groups.
| Pre-admission medication | P value | Acute management | P value | |||
|---|---|---|---|---|---|---|
| No END (n = 54) | END (n = 34) | No END (n = 54) | END (n = 34) | |||
| Antiplatelet agents | ||||||
| Aspirin | 7 (13.0) | 3 (8.8) | 0.55 | 46 (85.2) | 29 (85.3) | 0.99 |
| Clopidogrel | 4 (7.4) | 5 (14.7) | 0.28 | 50 (92.6) | 31 (91.2) | 0.64 |
| Cilostazol | 2 (3.7) | 0 | 0.16 | 3 (5.6) | 3 (8.9) | 0.56 |
| Dual antiplatelet therapy | 0 | 0 | - | 47 (87.0) | 29 (85.3) | 0.82 |
| Aspirin + clopidogrel | - | - | - | 45 (83.3) | 26 (76.5) | 0.43 |
| Aspirin + cilostazol | - | - | - | 0 | 1 (2.9) | 0.17 |
| Clopidogrel + cilostazol | - | - | - | 2 (3.8) | 2 (5.9) | 0.65 |
| Anticoagulant agents | ||||||
| Warfarin or DOAC | 5 (9.3) | 2 (5.9) | 0.56 | 5 (9.3) | 2 (5.9) | 0.56 |
| Argatroban | - | - | - | 44 (81.5) | 30 (88.2) | 0.39 |
| Alteplase | - | - | - | 2 (3.7) | 1 (2.9) | 0.85 |
| Edaravone | - | - | - | 32 (59.3) | 25 (73.5) | 0.17 |
| Statin | 12 (22.2) | 12 (35.3) | 0.18 | 40 (74.0) | 24 (70.6) | 0.72 |
Values are expressed as n (%).
DOAC indicates direct oral anticoagulant; END, early neurological deterioration.
At enrollment, the beta-TG, PF4, and D-dimer levels were measured in 79, 79, and 88 patients, respectively. Plasma levels of beta-TG (mean, 158 ng/mL versus 102 ng/mL; P = 0.021), PF4 (mean, 61 ng/mL versus 35 ng/mL; P = 0.024), and D-dimer (mean, 2.0 µg/mL versus 1.2 µg/mL; P = 0.037) were significantly higher in patients with END than those without END (Fig.1). According to the ROC analysis, optimal cut-off values for predicting END were 130 ng/mL, 55 ng/mL, and 2.0 µg/mL for beta-TG (C-statistic, 0.648), PF4 (C-statistic, 0.678), and D-dimer (C-statistic, 0.571), respectively. After adjustments for potential confounders, including age, sex, body mass index, systolic blood pressure, diastolic blood pressure, and low-density lipoprotein cholesterol levels, beta-TG (adjusted OR per 10 ng/mL increase [95% CI], 1.09 [1.01-1.20]; adjusted OR for beta-TG ≥ 130 ng/mL [95% CI], 4.63 [1.45-14.75]) and PF4 (adjusted OR per 10 ng/mL increase [95% CI], 1.21 [1.02-1.49]; adjusted OR for PF4 ≥ 55 ng/mL [95% CI], 3.63 [1.12-11.77]) were independent predictors of END, whereas D-dimer was not predictive of END (Table 4). Supplemental Table 1 shows the predictive value of each of the three hemostatic markers according to the location of the infarct.
END indicates early neurological deterioration.
|
Crude OR (95% CI) |
P value |
Age- and sex- adjusted OR (95% CI) |
P value |
Fully-adjusted OR (95% CI)＊ |
P value＊ | |
|---|---|---|---|---|---|---|
| Beta-thromboglobulin | ||||||
| Per 10 ng/mL increase | 1.08 (1.01-1.17) | 0.043 | 1.10 (1.02-1.21) | 0.023 | 1.09 (1.01-1.20) | 0.047 |
| ≥ 130 ng/mL | 3.64 (1.36-10.11) | 0.011 | 5.15 (1.67-15.86) | 0.004 | 4.63 (1.45-14.75) | 0.010 |
| Platelet factor 4 | ||||||
| Per 10 ng/mL increase | 1.21 (1.03-1.47) | 0.044 | 1.21 (1.03-1.48) | 0.049 | 1.21 (1.02-1.49) | 0.050 |
| ≥ 55 ng/mL | 2.91 (1.03-8.25) | 0.043 | 3.45 (1.16-10.94) | 0.026 | 3.63 (1.12-11.77) | 0.032 |
| D-dimer | ||||||
| Per 1.0 μg/mL increase | 1.27 (1.01-1.68) | 0.050 | 1.22 (0.96-1.61) | 0.12 | 1.14 (0.90-1.51) | 0.31 |
| ≥ 2.0 μg/mL | 3.02 (0.91-10.87) | 0.070 | 2.75 (0.78-10.49) | 0.11 | 2.13 (0.57-7.98) | 0.26 |
CI indicates confidence interval; OR, odds ratio.
＊Adjusted for age, sex, body mass index, systolic blood pressure, diastolic blood pressure, and low-density lipoprotein cholesterol levels.
|
Crude OR (95% CI) |
P value |
Age- and sex- adjusted OR (95% CI) |
P value | |
|---|---|---|---|---|
| BAD in LSA region(n = 57) | ||||
| Beta-thromboglobulin | ||||
| Per 10 ng/mL increase | 1.09 (1.00-1.21) | 0.090 | 1.11 (1.00-1.26) | 0.063 |
| ≥ 130 ng/mL | 2.70 (0.80-9.14) | 0.11 | 3.54 (0.91-13.80) | 0.069 |
| Platelet factor 4 | ||||
| Per 10 ng/mL increase | 1.28 (1.03-1.69) | 0.039 | 1.26 (0.95-1.66) | 0.10 |
| ≥ 55 ng/mL | 3.18 (0.88-12.03) | 0.079 | 3.47 (0.87-15.17) | 0.078 |
| D-dimer | ||||
| Per 1.0 μg/mL increase | 1.29 (1.00-1.78) | 0.068 | 1.25 (0.96-1.77) | 0.12 |
| ≥ 2.0 μg/mL | 4.86 (1.12-25.67) | 0.035 | 5.58 (1.13-34.37) | 0.035 |
| BAD in PPA region (n = 31) | ||||
| Beta-thromboglobulin | ||||
| Per 10 ng/mL increase | 1.06 (0.95-1.21) | 0.29 | 1.09 (0.96-1.28) | 0.19 |
| ≥ 130 ng/mL | 6.50 (1.18-44.72) | 0.031 | 11.64 (1.56-167.13) | 0.015 |
| Platelet factor 4 | ||||
| Per 10 ng/mL increase | 1.13 (0.88-1.51) | 0.35 | 1.18 (0.89-1.58) | 0.25 |
| ≥ 55 ng/mL | 2.47 (0.43-15.85) | 0.31 | 2.96 (0.44-19.86) | 0.26 |
| D-dimer | ||||
| Per 1.0 μg/mL increase | 1.36 (0.80-2.96) | 0.31 | 1.26 (0.72-2.77) | 0.44 |
| ≥ 2.0 μg/mL | 1.25 (0.15-10.22) | 0.84 | 1.02 (0.10-10.48) | 0.98 |
BAD indicates branch atheromatous disease; CI, confidence interval; LSA, lenticulostriate artery; OR, odds ratio; PPA, paramedian pontine artery.
In this observational study, including Japanese patients with acute BAD-related stroke, END occurred in 38.6% of patients during the first week of hospitalization. Patients with END had higher levels of beta-TG, PF4, and D-dimer than those without END. Furthermore, elevated beta-TG and PF4 levels were independently associated with the risk of END; the optimal cutoff values for predicting END were 130 ng/mL and 55 ng/mL, respectively. The relationship between hemostatic activation markers and END may provide important insights into stroke pathophysiology and potential interventions.
The frequency of END in BAD-related stroke varies widely among previous studies, with the reported incidence ranging from 15% to 75%3, 4, 7). This variation could be partly explained by differences in patient demographics, risk factor prevalence, or acute treatment between the studies. In addition, the definition of END has been inconsistent. Although most studies have used the change in the NIHSS score to define END, the cutoff varied from an increase of at least 1 point in the motor items of the NIHSS to at least 4 points in the overall score. The elapsed time to define “early” also differed by study, ranging from hours to weeks. Thus, interpretation and comparison of the results of the available studies remain difficult.
The positive association between platelet activation markers and END suggests the benefit of aggressive antiplatelet therapy in preventing END. Nearly 90% of our patients received DAPT with aspirin and clopidogrel, which represents current guideline-based strategies14). However, the antiplatelet effect of clopidogrel, a prodrug metabolized by the hepatic cytochrome P450 (CYP) into its active form, is attenuated in carriers of CYP2C19 loss-of-function alleles17). In particular, East Asian populations much more often have this genetic variant than white populations, with approximately 60% and 25% prevalence in East Asian and White patients with stroke, respectively17). Furthermore, previous studies of East Asian patients have indicated a potential benefit of other P2Y12 receptor antagonists such as ticagrelor17) and prasugrel18), which are not dependent on CYP2C19 for activation, over clopidogrel in secondary stroke prevention. Cilostazol, a phosphodiesterase III inhibitor, may also be an alternative or additional antiplatelet medication for BAD. Yamamoto et al. reported that combined antiplatelet therapy using cilostazol significantly improved functional outcomes in BAD-related stroke18). Given the pleiotropic effects of cilostazol, including not only antiplatelet action but also vasodilatation, anti-inflammation, and endothelial protection19), its administration is reasonable in preventing END. It remains to be elucidated which combination of drugs is best, but DAPT would be the first choice20). Approaches that modify or intensify therapy according to the change of platelet activation markers after treatment initiation may also be worth testing in the future.
Both beta-TG and PF4 are structurally similar, platelet-specific proteins stored in alpha-granules and released during platelet activation, but they may have different pathophysiological roles in thrombosis. Namely, beta-TG modulates coagulation by shortening thrombin generation time and directly interacting with factor X, promoting its activation21). On the other hand, PF4 inhibits heparin-antithrombin interactions, which can neutralize heparin’s anticoagulant effect and thereby promote thrombosis22). This mechanism is particularly relevant in heparin-induced thrombocytopenia, where PF4-heparin complexes trigger an immune response that enhances platelet activation. These differences highlight their unique roles in thrombosis, and the use of beta-TG and PF4 in clinical practice may guide a personalized antithrombotic strategy.
Although we failed to find an independent association, patients with END had higher D-dimer levels than those without END. This is in line with a previous study that found increased levels of D-dimer and thrombin-antithrombin complex in BAD-related stroke than in lacunar stroke11). It is considered that, in cases with stroke of atherothrombotic origin, a thrombus mainly composed of platelets is formed first23). When the thrombus causes blood flow stagnation, the coagulation system is activated and a fibrin-rich thrombus is secondarily formed, resulting in the expansion of the infarct. In fact, studies have demonstrated that anticoagulant therapy with argatroban in addition to DAPT is more effective in preventing END than DAPT alone in patients with BAD-related stroke18, 24). These suggest the importance of coagulation activation in the pathophysiology of END.
LimitationsThis study had several limitations that were mainly related to the inherent nature of observational studies. Because we included a single-center cohort of Japanese patients, our findings cannot be generalizable to other populations. Due to the small sample size, the cutoff values based on the ROC analysis results should only be used as a reference. In a number of cases, it was difficult to pinpoint the exact time of stroke onset because of wake-up stroke, slow or progressing symptoms, absence of witnesses, cognitive impairment that makes it difficult to recall or communicate the time of onset, or else. Thus, the current analysis could not take into account the time lapse between onset and hospitalization, and any worsening of symptoms prior to hospital visit may have been missed. Another important limitation is that we did not establish a unified treatment protocol for this study. The combination of antithrombotic agents and the duration of administration were determined at the discretion of the attending physician based on the patient’s condition and could have been more intensified or prolonged than usual, especially in the presence of worsening symptoms. In addition, acute blood pressure and glycemic control also substantially affect stroke outcomes25, 26), but information on these acute managements was not available. Future research can better elucidate the complex interplay between hemostatic disturbances, acute stroke treatments, and both early and long-term neurological outcomes.
When measuring beta-TG and PF4 levels, it is crucial to minimize platelet activation during blood collection by using a large-bore needle and avoiding excessive venous stasis27). In addition, blood samples must be processed immediately with gentle, low-speed centrifugation. Although we made efforts to standardize blood collection and handling procedures, there were challenges in strictly adhering to optimal protocols. Specifically, the use of a large-bore needle and the avoidance of excessive venous stasis were not consistently achieved. Furthermore, while we aimed to process blood samples as quickly as possible, delays were occasionally encountered. These deviations from the recommended protocol could introduce variability in the measurement of beta-TG and PF4 levels and may affect the reproducibility and accuracy of our results.
With regard to the diagnosis of BAD-related stroke, we only used routine MRI and MRA in accordance with previously established criteria. Recently, high-resolution MRI has been applied to the visualization of inner atherosclerotic lesions in middle cerebral arteries or basilar arteries of patients with BAD-related stroke, whose routine MRA had showed normal vessel’s profile28, 29). The use of such advanced imaging techniques would enable more accurate diagnosis of BAD along its original, pathological concept.
In patients with acute BAD-related stroke, elevated levels of beta-TG and PF4 were independently predictive of END, suggesting that the measurement of these platelet activation markers may help refine the risk assessment of these patients. Furthermore, more intensive antiplatelet regimens may be reasonable for BAD-related strokes with elevated hemostatic activation markers, although further studies are warranted.
None.
None.
The authors declare no conflicts of interest associated with this manuscript.
