Abstract
Aim: Various embolic sources and pathogenetic mechanisms underlie cryptogenic stroke (CS). We investigated the association of etiological diversity with short-term outcomes in patients with CS using a modified atherosclerosis (A), small-vessel disease (S), cardiac pathology (C), other causes (O), and dissection (D) (ASCOD) system.
Methods: Patients with CS who underwent transesophageal echocardiography were registered in this multicenter, observational study. In the modified classification system, O and D were inapplicable and thus excluded. Instead, atherosclerosis, small-vessel disease, cardiac pathology-CS classification was specifically constructed for the etiological diagnosis of CS. We utilized this system to explore the mechanism of CS by grading each pathology and evaluated its association with poorer modified Rankin Scale scores of 3–6 at hospital discharge.
Results: A total of 672 patients (68.7±12.8 years, 220 females) were analyzed. In the multiple logistic regression model, female sex (odds ratio [OR], 1.87 [1.15–3.04]; P =0.012), body mass index (OR, 0.93 [0.88–0.99]; P =0.025), National Institute of Health Stroke Scale score (OR, 1.16 [1.12–1.21]; P＜0.001), CHADS2 score (OR, 1.56 [1.30–1.86]; P＜0.001), D-dimer (OR, 1.04 [1.01–1.08]; P =0.015), diffusion-weighted image (DWI) lesion size (OR, 1.44 [1.10–1.89]; P =0.009), and S＋C score (OR, 1.26 [1.03–1.56]; P =0.029) were associated with poor functional outcome at discharge whereas the S＋C score was marginally associated with poor functional outcome after excluding 137 patients with a premorbid modified Rankin Scale score of ≥ 3.
Conclusions: The coexistence of small-vessel disease and cardiac pathology might be associated with poor in-hospital functional outcome in CS.
Introduction
About 25% of ischemic stroke cases are classified as cryptogenic stroke (CS). In 2014, Hart et al 1). reported that patients with ischemic stroke with unknown embolic origins were classified as having embolic stroke of undetermined source (ESUS). Antiplatelet therapy has been the preferred treatment for patients with CS2), but the subgroup analysis of the Warfarin-Aspirin Recurrent Stroke Study indicated the efficacy of anticoagulant therapy3). Two clinical trials compared the therapeutic efficacy of aspirin and direct oral anticoagulants (DOACs) in the secondary prevention for patients with ESUS; however, it did not show any benefits after treatment with DOACs4, 5). Meanwhile, various potential embolic pathologies are involved in CS and ESUS, such as cardiogenic (paroxysmal atrial fibrillation [PAF], patent foramen ovale [PFO], and atrial septal aneurysm [ASA]) and arteriogenic (non-stenotic atherosclerotic carotid artery plaque and aortic atherosclerotic plaques) embolic sources. Additionally, various pathologic conditions, such as cerebral white matter diseases and microbleeds, can also be involved6-8).These underlying pathologies could contribute to the recurrence of ischemic stroke in patients with ESUS and CS and could be related to DOAC responsiveness9-11).
The atherosclerosis (A), small-vessel disease (S), cardiac pathology (C), other causes (O), and dissection (D) (ASCOD) classification, a method based on the quantitative evaluation of each stroke mechanism, is a useful system for classifying ischemic stroke subtypes because it can capture the overlap between underlying diseases. It varies from the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification that stratifies embologenic cardiac diseases as high or medium risk but does not include other etiologies12, 13). The grading of stroke phenotypes A and C of the ASCOD classification is considered homologous to potential embolic sources in CS1). Furthermore, in CS, severe white matter diseases and cerebral microbleeds (CMBs) can be classified as “causal link is unlikely, but the disease is present” under the S phenotype, whereas O phenotypes, such as antiphospholipid syndrome and thrombocythemia, and D phenotypes were essentially less applicable. In a previous single-center study that used this phenotyping, C and S factors and a number of ASC categories were highly related to stroke recurrence in patients with ESUS14).
Functional outcomes after stroke vary among stroke subtypes. Patients with cardioembolic stroke showed the worst functional outcomes, whereas lacunar infarctions had the mildest prognosis at discharge15, 16). Long-term functional outcomes were found to be milder in CS and ESUS than in cardioembolic stroke, similar to that observed in large-artery atherosclerosis17). However, to date the association of etiologic diversity with functional outcome in CS has not been fully studied.
Aim
We sought to investigate the diagnostic efficacy of a new diagnostic classification, atherosclerosis, small-vessel disease, cardiac pathology-cryptogenic stroke (ASC-CS) phenotyping, to clarify the mechanisms of CS with grading each pathology and the association of this new classification with short-term functional outcome at hospital discharge. The investigation evaluated CS records from the CHALLENGE ESUS/CS registry, a multicenter registry that cataloged CS cases assessed using transesophageal echocardiography (TEE).
Methods
Study Population
The CHALLENGE ESUS/CS registry is a multicenter, retrospective registry which enrolled consecutive patients with CS who underwent TEE among eight hospitals in Japan from April 2014 to December 2016. The diagnosis of CS was based on the elicitation of patient medical history; diagnostic modalities including computed tomography (CT)/magnetic resonance imaging (MRI), carotid duplex ultrasonography, 12-lead electrocardiography, and electrocardiographic monitoring for the first 24 h; blood examinations, and chest X-ray, which were performed on admission. The institutional review boards of all eight participating centers approved the study protocol. We used clinical information obtained from medical records; thus, the need to obtain written informed consent from each patient was waived in this retrospective study. This study was conducted in accordance with the Declaration of Helsinki and was registered with http://www.umin.ac.jp/ctr/ (UMIN000032957).
Inclusion and Exclusion Criteria for the CHALLENGE ESUS/CS Registry
The inclusion criteria for the current study were as follows: 1) ischemic stroke within seven days of onset, 2) non-lacunar stroke on neuroradiological imaging, 3) absence of arterial stenosis of ≥ 50% or occlusion in a corresponding large artery, 4) absence of major embologenic cardiac disease, and 5) absence of other determined stroke etiologies. At admission, we excluded patients with major embologenic cardiac diseases that were in the exclusion criteria of ESUS based on medical history and the current study evaluations1). Additionally, cardiac monitoring for ＞24 h is recommended in the diagnostic criteria of ESUS1) and PAF detected ＜24 h after admission is not suitable for the diagnosis of ESUS; therefore, patients fulfilling this criterion were not registered in the CHALLENGE ESUS/CS registry. In contrast, TEE, transthoracic echocardiography (TTE), and continuous cardiac monitoring with 24-h Holter electrocardiography were performed after admission and covert embolic sources detected during hospitalization, such as PAF detected by cardiac monitoring at 48 h after admission or Holter electrocardiography, low left ventricle ejection fraction (LVEF) on TTE, and intracardiac thrombus on TEE, were regarded as potential embolic sources in originally classified CS and were analyzed in the current study.
Atherosclerotic Risk Factors
Hypertension, diabetes mellitus, dyslipidemia, current smoking, and history of ischemic heart disease and ischemic stroke were defined according to our previous work8).
TEE Study
TEE was performed as described in our previous work8). ASA was defined as a ≥ 10-mm excursion into either the left or the right atrium or a total excursion of ≥ 15 mm into the left or right atrium. The right-to-left shunt was assessed by injecting agitated saline and having patients perform the Valsalva maneuver. Subsequently, the number of microbubbles with and without contrast agents was compared, and the number transiting from the right to the left atrium was counted. PFO was diagnosed when microbubbles were visualized in the left atrium ≤ 3 cardiac cycles after the Valsalva maneuver, whereas pulmonary arteriovenous fistula was diagnosed when microbubbles were visualized in the left atrium ＞3 cardiac cycles after the Valsalva maneuver or when microbubbles were visualized without the Valsalva maneuver. Aortic arch plaque thickness ≥ 4 mm, presence of ulcerations, and mobile components were evaluated8).
ASC Phenotyping
We constructed ASC-CS as a modification of the ASCOD system12). First, O (other causes) and D (dissection) were not suitable for CS and were thus excluded. Grade 9 (incomplete workup) was excluded because our registry comprehensively collected data, including echocardiographic and carotid ultrasonographic investigations for cardiac and arteriogenic embolic sources and cardiac electrophysiological analyses to detect PAF ≥ 24 h after admission. Grading for A and C phenotyping was assigned 1 (potentially causal: A1 and C1), 2 (causal link is uncertain: A2 and C2), 3 (causal link is unlikely, but disease is present: A3 and C3), and 0 (disease is absent: A0 and C0). In S phenotyping, only S3 and S0 were used in the ASC-CS, since S1 and S2 were not suitable for CS and were excluded.
Table 1 lists the embologenic diseases and pathologic conditions under the ASC-CS criteria. Each subset of the grades in A, S, and C was determined according to modifications of the ASCOD system: A was graded as A1 (mobile and ulcerative lesions in the aortic arch), A2 (aortic arch plaque ≥ 4 mm without mobile and ulcerative lesions), and A3 (stenosis ≥ 50% in an intra- or extracranial artery contralateral to the infarct area or history of ischemic heart disease), and A0. S was graded as S3 (corresponding to a Fazekas grade ≥ 3 or CMBs with deep and diffuse localization)7) and S0. C phenotype was graded as C1 (PAF, intracardiac thrombus, mitral stenosis, or LVEF ＜35%), C2 (PFO with ASA, spontaneous echo contrast, or 35% ≤ LVEF ＜50%)18), C3 (lone PFO, lone ASA, or mitral annulus calcification), and C0. Points were then assigned to each of the ASC-CS grades (Grade 1, 3 points; Grade 2, 2 points; Grade 3, 1 point). History of myocardial infarction and multiple brain infarcts on both sides or both the anterior and posterior circulation, which were defined in C2 in the original ASCOD criteria12), were not included in the ASC-CS system, because multiple infarctions in different arterial territories are principal radiological findings of cancer-related stroke and aortogenic brain embolism19, 20) and myocardial infarction resulting in cardiac ventricular dysfunction can be detected by TTE and categorized as LVEF ＜35% in C1 or 35% ≤ LVEF ＜50% in C2.
Table 1.
Grades of modified ASC phenotypes
| A: Atherosclerosis |
| Grade 1 mobile or ulcerative lesions in the aortic arch |
| Grade 2 aortic plaque ≥ 4 mm without mobile or ulcerative lesions |
| Grade 3 (1) stenosis ≥ 50% in an intra- or extracranial artery contralateral to the infarct area |
| (2) history of myocardial infarction |
| S: Small-vessel disease |
| Grade 3 (1) severe (confluent, Fazekas grade ≥ 3) leukoaraiosis |
| (2) cerebral microbleeds visible on T2＊-weighted MRI§
|
| C: Cardiac pathology |
| Grade 1 (1) paroxysmal atrial fibrillation |
| (2) intracardiac thrombus |
| (3) mitral stenosis |
| (4) left ventricle ejection fraction ＜35% |
| Grade 2 (1) PFO＋ASA |
| (2) spontaneous echo contrast |
| (3) 35% ≤ left ventricle ejection fraction ＜50% |
| Grade 3 (1) lone PFO |
| (2) lone ASA |
| (3) mitral annulus calcification |
ASA, atrial septal aneurysm; ASC, atherosclerosis, small-vessel disease, cardiac pathology; MRI, magnetic resonance imaging; PFO, patent foramen ovale
Grade 0, absence of disease
§Deep and diffuse type but not lobar cerebral microbleed is defined as Grade 3 in S.
Baseline clinical information, laboratory and radiological data on admission, echocardiographic findings, and cardiac electrophysiological data were collected from hospital charts or through database reviews. In the laboratory data, cutoff values of ＞100 and ＞300 pg/mL were used for brain natriuretic peptide and N-terminal pro-brain natriuretic peptide, respectively, based on studies showing the utility of these values in excluding the diagnosis of acute heart failure21, 22). Infarct size (＜1.5, 1.5–3.0, ＞3.0 cm) and multiple lesions (single lesion, multiple lesions in single, two, and three vascular territories among bilateral anterior, middle, and posterior cerebral arteries) on diffusion-weighed images were evaluated, and therapy for secondary prevention, including antiplatelet agents or anticoagulant agents, were analyzed. Short-term functional outcomes at hospital discharge were determined using the modified Rankin Scale (mRS)23), with mRS scores 3–6 considered as poor functional outcomes.
Statistical Analysis
Data were analyzed using the chi-squared test for categorical variables and the Mann–Whitney test for nonparametric analyses. All variables with P value of ＜0.05 on univariate analyses were entered into stepwise logistic regression analysis with a backward elimination (Wald test) to identify independent variables for patients with poor functional outcomes (mRS score 3–6 at discharge). Multiple logistic regression models were adjusted for variables, including ASC-CS scores considered as related to the functional outcome. All data were analyzed using the SPSS for Windows version 26.0 software (SPSS, Chicago, IL, USA). A P value of ＜0.05 was considered statistically significant.
Data Availability
The datasets used and analyzed during the current study are available through the corresponding author on reasonable request.
Results
A total of 677 patients with CS were enrolled in the CHALLENGE ESUS/CS registry. Five patients were excluded because of missing MRI data, and a final total of 672 patients were analyzed. The median length of hospital stay was 17 days, and TTE was performed in 576 patients (85.7%). The mean age was 68.7±12.8 years, 220 patients were females, and the median National Institute of Health Stroke Scale (NIHSS) score was 2. Fig. 1 shows a Venn diagram depicting the distributions of patients in the simplex pathology subsets of A, S, and C and complex subsets of A＋S, S＋C, A＋C, and A＋S＋C.
Clinical Characteristics of Patients with CS according to Functional Outcome at Discharge
Table 2 lists the baseline characteristics of the two groups with 3–6 (poor; 147 patients) and 0–2 (good; 525 patients) mRS scores at discharge. Patients with poor functional outcomes were older (73.4±13.3 vs. 67.4±12.6 years, P＜0.001) and had lower body mass index (BMI; 22.46±4.03 vs. 23.48±3.81 kg/m2; P =0.006) and NIHSS scores (median [interquartile range], 6 [3–15] vs. 2 [1–3]; P＜0.001). Furthermore, the frequency of previous history of ischemic stroke was higher in patients with poor functional outcome (27.2% vs. 15.8%; P =0.002). No significant differences were observed among the atherosclerotic vascular risk factors. In laboratory data, high-density lipoprotein cholesterol (HDL-C) levels were lower in patients with poor functional outcome (48.44±13.88 mg/dL; vs. 51.92±15.30 mg/dL; P=0.015), whereas D-dimer levels (6.92±33.97 vs. 1.89±4.04 µg/mL; P =0.002) were higher. In radiological findings, patients with poor functional outcomes displayed larger infarcts (＜1.5 cm, 31.3% vs. 47.8%; 1.5–3.0 cm, 19.0% vs. 27.8%; ＞3.0 cm, 49.7% vs. 24.4%; P＜0.001). Patients with poor functional outcomes tended to present multiple lesions. As for therapy for secondary prevention of stroke, the frequency of antiplatelet agent medication was higher in patients with good functional outcomes (55.1% vs. 71.6%; P＜0.001), whereas the frequency of anticoagulant therapy was higher in patients with poor functional outcome (46.3% vs. 30.1%; P＜0.001).
Table 2.
Baseline characteristics, laboratory data, and radiological observations in patients with good and poor functional outcomes
|
Total study subjects N = 672 |
Good functional outcome (mRS score, 0–2)
N = 525
|
Poor functional outcome (mRS score, 3–6)
N = 147
|
P value
|
| Age, years mean±SD |
68.7±12.8 |
67.4±12.6 |
73.4±13.3 |
＜0.001 |
| Sex, female, no (%) |
220 (32.7) |
162 (30.9) |
58 (39.5) |
0.0495 |
| Body mass index, kg/m2
|
23.03±3.88 |
23.48±3.81 |
22.46±4.03 |
0.010 |
| NIHSS score on admission, median (IQR) |
2 (1–5) |
2 (1–3) |
6 (3–15) |
＜0.001 |
| Hypertension, no (%) |
480 (71.4) |
372 (70.9) |
108 (73.5) |
0.536 |
| Diabetes mellitus, no (%) |
170 (25.3) |
126 (24.0) |
44 (29.9) |
0.144 |
| Dyslipidemia, no (%) |
343 (51.0) |
266 (50.7) |
77 (52.4) |
0.713 |
| Cigarette smoking, no (%) |
180 (26.8) |
141 (26.9) |
39 (26.5) |
0.937 |
| Coronary artery disease, no (%) |
68 (10.1) |
58 (11.1) |
10 (6.8) |
0.131 |
| Previous history of ischemic stroke, no (%) |
123 (18.3) |
83 (15.8) |
40 (27.2) |
0.002 |
| Heart failure, no (%) |
17 (2.5) |
11 (2.1) |
6 (4.1) |
0.175 |
| CHADS2 score, median (IQR)
|
2 (1–2) |
1 (1–2) |
2 (1–3) |
＜0.001 |
| HbA1c (%) a
|
6.13±1.15 |
6.08±1.02 |
6.29±1.54 |
0.097 |
| HDL cholesterol (mg/dL) b
|
51.17±48.20 |
51.92±15.30 |
48.44±13.88 |
0.026 |
| LDL cholesterol (mg/dL) b
|
112.39±33.94 |
112.36±33.46 |
112.52±35.77 |
0.896 |
| Triglyceride (mg/dL) c
|
133.08±110.0 |
134.34±113.58 |
128.57±84.43 |
0.915 |
| eGFR |
67.38±25.96 |
67.59±24.80 |
66.64±29.79 |
0.063 |
|
BNP ＞100 pg/mL or NT-proBNP ＞300
pg/mL, no (%) d
|
171 (26.4) |
134 (26.5) |
37 (26.1) |
0.952 |
| D-dimer (μg/mL) |
2.99±16.38 |
1.89±4.04 |
6.92±33.97 |
＜0.001 |
| DWI lesion size |
|
|
|
＜0.001 |
| ＜1.5 cm |
297 (44.2) |
251 (47.8) |
46 (31.3) |
|
| 1.5–3.0 cm |
174 (25.9) |
146 (27.8) |
28 (19.0) |
|
| ＞3.0 cm |
201 (29.9) |
128 (24.4) |
73 (49.7) |
|
| Multiple DWI lesions |
|
|
|
0.011 |
| Single lesion |
241 (35.9) |
204 (38.9) |
37 (25.2) |
|
| Multiple lesions in one vascular territory |
250 (37.2) |
186 (35.4) |
64 (43.5) |
|
| Multiple lesions in two vascular territories |
124 (18.4) |
96 (18.3) |
28 (19.0) |
|
|
Multiple lesions in three vascular
territories
|
57 (8.5) |
39 (7.4) |
18 (12.3) |
|
| Therapy for secondary prevention |
|
|
|
|
| Antiplatelet agents |
457 (68.0) |
376 (71.6) |
81 (55.1) |
＜0.001 |
| Anticoagulants |
226 (33.6) |
158 (30.1) |
68 (46.3) |
＜0.001 |
Chi-square test and the Man-Whitney test were used for comparisons.
mRS, modified Rankin Scale; NIHSS, National Institutes of Health stroke scale; IQR, interquartile range; HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular fibrillation rate; BNP, brain natriuretic peptide; NT-proBNP, N-terminal pro-brain natriuretic peptide; CRP, C-reactive protein; DWI, diffusion-weighted image Missing values: a, n = 9; b, n = 8; c, n = 5; d, n = 25
Table 3 shows the distribution of the phenotype grades. In S phenotype, S3 grades were more frequent in patients with poor functional outcomes (40.8% vs. 27.2%, P=0.014). In C phenotype, the frequency of the frequencies of grades indicating etiological causality tended to be higher in patients with poor functional outcomes (Grade 1, 14.3% vs. 9.1%; Grade 2, 18.4% vs. 12.0%; Grade 3, 41.5 vs. 40.8%; P=0.010).
Table 3.
Grades and complexity of ASC phenotyping in patients with good and poor functional outcomes
|
Total study population N = 672 |
Good functional outcome (mRS score, 0–2)
N = 525, 78.1%
|
Poor functional outcome (mRS score, 3–6)
N = 147, 21.9%
|
P value
|
| Grading in classification |
|
|
|
|
| Atherothrombosis (A) |
|
|
|
0.421 |
| A1 |
83 (12.4) |
65 (12.4) |
18 (12.3) |
|
| A2 |
168 (25.0) |
133 (25.3) |
35 (23.8) |
|
| A3 |
63 (9.4) |
44 (8.4) |
19 (12.9) |
|
| A0 |
358 (53.3) |
283 (53.9) |
75 (51.0) |
|
| Small-vessel disease (S) |
|
|
|
0.014 |
| S3 |
203 (30.2) |
143 (27.2) |
60 (40.8) |
|
| S0 |
469 (69.8) |
382 (72.8) |
87 (59.2) |
|
| Cardiac pathology |
|
|
|
0.010 |
| C1 |
69 (10.3) |
48 (9.1) |
21 (14.3) |
|
| C2 |
90 (13.4) |
63 (12.0) |
27 (18.4) |
|
| C3 |
275 (40.9) |
214 (40.8) |
61 (41.5) |
|
| C0 |
238 (35.4) |
200 (38.1) |
38 (25.9) |
|
| Complexity of pathogenesis |
|
|
|
|
| Atherothrombosis (A) |
|
|
|
0.032 |
| Sole A |
89 (13.2) |
77 (14.7) |
12 (8.2) |
|
| A with S or C |
225 (33.5) |
165 (31.4) |
60 (40.8) |
|
| A0 |
358 (53.3) |
283 (53.9) |
75 (51.0) |
|
| Small-vessel disease (S) |
|
|
|
0.007 |
| Sole S |
23 (3.4) |
16 (3.0) |
7 (4.8) |
|
| S with A or C |
180 (26.8) |
127 (24.2) |
53 (36.1) |
|
| S0 |
469 (69.8) |
382 (72.8) |
87 (59.2) |
|
| Cardiac pathology |
|
|
|
＜0.001 |
| Sole C |
206 (30.7) |
167 (31.8) |
39 (26.5) |
|
| C with A or S |
228 (33.9) |
158 (30.1) |
70 (47.6) |
|
| C0 |
238 (35.4) |
200 (38.1) |
38 (25.9) |
|
The chi-square test was used for comparisons.
mRS, modified Rankin Scale
With regard to ASC phenotyping complexity, Fig. 1 shows that duplicate or triplicate pathologies were not uncommon in both good and poor functional outcomes. However, the frequency of coexistence with other pathologies was more common in patients with poor functional outcomes in A (40.8% vs. 31.4%; P=0.032), S (36.1% vs. 24.2%; P =0.007), and C (47.6% vs. 30.1%; P＜0.001, Table 3). Conversely, single pathologies were more frequent in patients with good functional outcomes.
Fig. 2 shows that patients with poor functional outcomes had higher scores for S (0.41±0.49 vs. 0.27±0.45; P =0.002), C (1.21±0.99 vs. 0.92±0.93; P =0.001), S＋C (1.62±1.11 vs. 1.19±1.01; P＜0.001), A＋C (2.18±1.42 vs. 1.88±1.33; P =0.021), and A＋S＋C (2.59±1.56 vs. 2.16±1.48; P =0.003).
Underlying Clinical Characteristics and ASC Phenotyping Associated with Poor Functional Outcomes at Discharge
Age, female sex, BMI, NIHSS score, previous history of ischemic stroke, CHADS2 score, HDL-C and D-dimer levels, diffusion-weighted image (DWI) lesion size, multiple DWI lesions, antiplatelet therapy, and anticoagulant therapy were evaluated in the stepwise logistic regression analysis as covariates. The S, C, S＋C, A＋C, and A＋S＋C categories were entered in separate models, as these covariates confounded each other. In the model including the S＋C category, female sex (odds ratio [OR], 1.87; 95% confidence interval [CI], 1.15–3.04; P =0.012), BMI (OR, 0.93; 95% CI, 0.88–0.99; P =0.025), NIHSS score (OR, 1.16; 95% CI, 1.12–1.21; P＜0.001), CHADS2 score (OR, 1.56; 95% CI, 1.30–1.86; P＜0.001), D-dimer (OR, 1.04; 95% CI, 1.01–1.08; P =0.015), DWI lesion size (OR, 1.44; 95% CI, 1.10–1.89; P =0.009), and S＋C score (OR, 1.26; 95% CI, 1.03–1.56; P =0.029) were associated with poor functional outcomes (Table 4). The models including the S, C, A＋C, and A＋S＋C categories did not show significant associations with poor functional outcomes.
Table 4.
Multiple logistic regression analyses for factors associated with poor functional outcome
| Variables, S＋C score model |
OR |
95% CI |
P value
|
| Female |
1.87 |
1.15–3.04 |
0.012 |
| BMI |
0.93 |
0.88–0.99 |
0.025 |
| NIHSS score |
1.16 |
1.12–1.21 |
＜0.001 |
| CHADS2 score
|
1.56 |
1.30–1.86 |
＜0.001 |
| HDL cholesterol |
0.98 |
0.97–1.00 |
0.055 |
| D-dimer |
1.04 |
1.01–1.08 |
0.015 |
| DWI lesion size |
1.44 |
1.10–1.89 |
0.009 |
| S＋C score (0–4) |
1.26 |
1.03–1.56 |
0.029 |
OR, odds ratio; CI, confidence interval; BMI, body mass index; NIHSS, National Institutes of Health stroke scale;
HDL, high-density lipoprotein; DWI, diffusion-weighted image
To exclude the influence of a high premorbid mRS score on poor functional outcomes at hospital discharge, we further analyzed 635 patients with a premorbid mRS score of 0–2 by categorizing into two groups based on mRS score at discharge: poor (mRS score, 3–6; n=121) and good (mRS score, 0–2; n=514). Age, NIHSS score, previous history of ischemic stroke, CHADS2 score, D-dimer level, DWI lesion size, presence of multiple DWI lesions, antiplatelet therapy, and anticoagulant therapy, which were significantly different between the two groups (P＜0.05), were selected as variables for multiple logistic regression model (Table 5). The C, S＋C, and A＋S＋C categories were entered in separate models, as these covariates confounded each other. NIHSS score (OR, 1.16; 95% CI, 1.12–1.21; P＜0.001), CHADS2 score (OR, 1.43; 95% CI, 1.19–1.73; P＜0.001), D-dimer (OR, 1.06; 95% CI, 1.02–1.10; P=0.004), and DWI lesion size (OR, 1.65; 95% CI, 1.23–2.31; P=0.001) were significantly associated with poor functional outcomes, whereas the S＋C score was relatively associated with poor functional outcomes in patients with a CS and a premorbid mRS score of 0–2 (OR, 1.24; 95% CI, 0.99–1.54; P =0.059) (Table 6).
Table 5.
Baseline characteristics, laboratory data, and radiological observations in patients with good and poor functional outcomes among those with a premorbid mRS score of 0–2
|
Total study subjects N= 635 |
Good functional outcome (mRS score, 0–2)
N= 524
|
Poor functional outcome (mRS score, 3–6)
N= 111
|
P value
|
| Age, years, mean±SD |
68.2±12.8 |
67.4±12.7 |
71.7±13.1 |
＜0.001 |
| Sex, female, no (%) |
204 (32.1) |
161 (30.7) |
43 (38.7) |
0.101 |
| Body mass index, kg/m2
|
23.32±3.85 |
23.48±3.81 |
22.6±3.97 |
0.059 |
| NIHSS score on admission, median (IQR) |
2 (1–5) |
2 (1–3) |
6 (3–15) |
＜0.001 |
| Hypertension, no (%) |
452 (71.2) |
372 (71.0) |
80 (72.1) |
0.820 |
| Diabetes mellitus, no (%) |
159 (25.0) |
126 (24.0) |
33 (29.7) |
0.209 |
| Dyslipidemia, no (%) |
320 (50.4) |
265 (50.6) |
55 (49.5) |
0.845 |
| Cigarette smoking, no (%) |
177 (27.9) |
141 (26.9) |
36 (32.4) |
0.238 |
| Coronary artery disease, no (%) |
66 (10.4) |
58 (11.1) |
8 (7.2) |
0.226 |
| Previous history of ischemic stroke, no (%) |
109 (17.2) |
82 (15.2) |
27 (24.3) |
0.028 |
| Heart failure, no (%) |
16 (2.5) |
11 (2.1) |
5 (4.5) |
0.142 |
| CHADS2 score, median (IQR)
|
2 (1–2) |
1 (1–2) |
2 (1–3) |
0.008 |
| HbA1c (%) a
|
6.13±1.17 |
6.08±1.02 |
6.38±1.72 |
0.084 |
| HDL cholesterol (mg/dL) b
|
51.74±15.17 |
51.94±15.30 |
50.74±14.53 |
0.726 |
| LDL cholesterol (mg/dL) b
|
112.88±33.47 |
112.40±33.47 |
115.21±35.3 |
0.434 |
| Triglyceride (mg/dL) c
|
134.26±109.51 |
134.51±113.62 |
133.04±87.68 |
0.747 |
| eGFR |
67.58±24.95 |
67.56±24.82 |
67.68±25.65 |
0.364 |
|
BNP ＞100 pg/mL or NT-proBNP ＞300 pg/
mL, no (%) d
|
159 (26.0) |
132 (26.2) |
27 (25.29 |
0.773 |
| D-dimer (μg/mL) |
3.03±16.84 |
1.89±4.05 |
8.40±39.00 |
＜0.001 |
| DWI lesion size |
|
|
|
＜0.001 |
| ＜1.5 cm |
279 (43.9) |
251 (47.9) |
28 (25.2) |
|
| 1.5–3.0 cm |
168 (26.5) |
145 (27.7) |
23 (20.7) |
|
| ＞3.0 cm |
188 (29.6) |
128 (24.4) |
60 (54.1) |
|
| Multiple DWI lesions |
|
|
|
0.026 |
| Single lesion |
234 (36.9) |
203 (38.7) |
31 (27.3) |
|
| Multiple lesions in one vascular territory |
231 (36.4) |
186 (35.5) |
45 (40.0) |
|
| Multiple lesions in two vascular territories |
116 (18.3) |
96 (18.3) |
20 (19.8) |
|
| Multiple lesions in three vascular territories |
54 (8.5) |
39 (7.4) |
15 (13.2) |
|
| Therapy for secondary prevention |
|
|
|
|
| Antiplatelet agents |
430 (67.7) |
375 (71.6) |
55 (49.5) |
＜0.001 |
| Anticoagulants |
211 (33.2) |
153 (29.2) |
58 (52.3) |
＜0.001 |
| Subset of ASC-CS, points |
|
|
|
|
| A |
0.95±1.12 |
0.96±1.14 |
0.89±1.07 |
0.670 |
| S |
0.29±0.45 |
0.27±0.45 |
0.34±0.48 |
0.141 |
| C |
0.98±0.96 |
0.92±0.93 |
1.27±1.04 |
0.001 |
| A＋S |
1.24±1.30 |
1.24±1.32 |
1.23±1.23 |
0.765 |
| S＋C |
1.27±1.06 |
1.19±1.01 |
1.61±1.18 |
0.001 |
| A＋C |
1.93±1.35 |
1.89±1.33 |
2.16±1.42 |
0.056 |
| A＋S＋C |
2.22±1.50 |
2.16±1.48 |
2.50±1.57 |
0.037 |
The chi-square and the Mann–Whitney tests were used for comparisons.
mRS, modified Rankin Scale; NIHSS, National Institutes of Health stroke scale; IQR, interquartile range; HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular fibrillation rate; BNP, brain natriuretic peptide; NT-proBNP, N-terminal pro-brain natriuretic peptide; CRP, C-reactive protein; DWI, diffusion-weighted image
Missing values: a, n= 7; b, n= 5; c, n= 4; d, n= 24
Table 6.
Multiple logistic regression analyses for factors associated with poor functional outcome in patients with a premorbid mRS score of 0–2
| Variables, S＋C score model |
OR |
95% CI |
P value
|
| NIHSS |
1.16 |
1.12–1.21 |
＜0.001 |
| CHADS2 score
|
1.43 |
1.19–1.73 |
＜0.001 |
| D-dimer |
1.06 |
1.02–1.10 |
0.004 |
| DWI lesion size |
1.65 |
1.23–2.21 |
0.001 |
| S＋C score (0–4) |
1.24 |
0.99–1.54 |
0.059 |
mRS, modified Rankin Scale; OR, odds ratio; CI, confidence interval; NIHSS, National Institute Health stroke scale; DWI, diffusion-weighted image
Discussion
The present study clarified the contribution of various pathogeneses of CS to functional outcome at discharge. Our analyses revealed that atherothrombosis, small-vessel disease, and cardiac pathology were implicated as the pathogenesis in 46.7%, 30.2%, and 64.6% patients initially diagnosed as CS, respectively. According to the ACS-CS criteria, the S＋C complex phenotype might be an independent predictor for poor functional outcome at discharge in patients with CS.
The ASCOD classification systemically evaluates strokes according to the causality and stratifies each pathogenesis using Grades 1–3, thus enabling the classification of mixed pathologies with quantifications12). Sirimarco et al 24). found that 89.8%, 65.8%, and 51.8% of 405 patients with ischemic stroke had A, S, and C pathologies, respectively, 80% of whom had multiple pathogeneses (A＋S, 59.5%; S＋C, 34.1%; A＋C, 49.1%) in overall ischemic stroke. In the current study, the ASC-CS system, modified from the ASCOD system, classified 672 patients with CS into the following categories: A, 46.7%; S, 30.2%; C, 64.6%; A＋S, 19.0%; S＋C, 19.5%; A＋C, 26.2%; and A＋S＋C, 11.8%. The findings showed that A, S, and complex pathologies were less frequent, but C phenotype was more frequent. Therefore, the data on the CHALLENGE ESUS/CS registry, which enrolled patients who were initially diagnosed as CS based on extensive cardiac investigations, likely reflect a higher frequency of cardiac pathologies. Additionally, not a few patients with potential cardiac embolic sources such as mitral stenosis, low LVEF detected by TTE, and intracardiac thrombus detected by TEE, might not have been identified in initial stroke examinations and might have been enrolled as CS.
We also found that female sex, BMI, NIHSS and CHADS2 scores, D-dimer levels, and large infarcts on DWI were related to poor functional outcome. Cardioembolic stroke had higher NIHSS scores and larger infarction sizes than other stroke subtypes and showed worse functional prognoses at discharge25, 26). Female patients with cardioembolic stroke were previously reported to exhibit greater neurological severity27). CHADS2 score is an index to objectively evaluate stroke risk28), whereas plasma D-dimer level is higher in patients with cardioembolic stroke compared to those with other stroke subtypes29, 30). Thus, the association of cardiac pathology reflecting these independent variables with poor functional outcome at discharge is unsurprising. Additionally, BMI was recently shown to be related to disability based on mRS scores, exhibiting a U-shaped trend31). More importantly, our data indicated that mixed pathologies, especially the S＋C phenotype, were related to poor functional outcomes at hospital discharge and that this association was attenuated after excluding patients with a premorbid mRS score of ≥ 3. Patients with cardioembolic stroke exhibit severe stroke symptoms and worse functional outcomes, whereas small-vessel disease is associated with cognitive dysfunction and frailty, implying that both S and C factors impact mRS scores17, 32). Moreover, white matter lesions was reported to be associated with the inhibition of short-term recovery from unilateral spatial neglect after stroke33). Several lines of emerging evidence reveal the association between small-vessel disease and cardiac pathology: (1) the degree of cerebral white matter disease is significantly high in patients with AF34); (2) the degree of cerebral white matter disease is related to the prevalence of new-onset AF in patients with ESUS6); and (3) CHA2DS2-VASc score is related to leukoaraiosis on the Fazekas scale in patients with stroke with AF35). Collectively, our data indicate that the coexistence of cardiac pathology and small-vessel disease in patients with CS might contribute to poor functional outcomes based on the mRS score at hospital discharge.
As for the secondary prevention of ESUS and CS, the NAVIGATE ESUS (New Approach Rivaroxaban Inhibition of Factor Xa in a Global Trial vs. ASA to Prevent ESUS)4) and RE‐SPECT ESUS (Randomized, double‐blind, Evaluation in secondary Stroke Prevention comparing the EfficaCy and safety of the oral Thrombin inhibitor dabigatran etexilate vs. acetylsalicylic acid in patients with Embolic Stroke of Undetermined Source)5) trials did not indicate a reduction in recurrent stroke risk in ESUS after treatments with DOACs over aspirin. Meanwhile, meta-analyses including subanalysis of the NAVIGATE ESUS and CS trials showed that ESUS with PFO and dilation of the left atrium responded well to anticoagulant therapy, whereas those with aortic arch plaques did not9-11). In the current study, anticoagulant therapy tended to be selected as secondary preventive therapy for patients with CS exhibiting poor functional outcomes. Our data also indicated that patients with CS exhibiting poor functional outcomes at discharge presented with not only cardiac pathology but also small-vessel disease. Meanwhile, small-vessel disease, including extensive white matter disease and CMBs (based on hypertensive arteriopathy such as lipohyalinosis and microaneurysm), might pose a risk for intracerebral hemorrhage after long-term treatment with oral anticoagulants36). Therefore, anticoagulant therapy was suitable for the prevention for cardiac pathology whereas the development of hemorrhagic stroke should be noted in patients with CS exhibiting poor functional outcomes, especially those with CMBs37). Future studies are warranted to determine optimal antithrombotic therapy for patients with CS exhibiting the complex S＋C phenotype.
Our study has several limitations that must be considered when interpreting the results. First, this registry was evaluated by retrospective analysis. Medications, especially acute therapy for ischemic stroke were nonrandomized, which might have affected functional outcomes at discharge. The hospitalization period varied among patients. Second, the selection of patients with CS for TEE in participating hospitals could be biased. In particular, the median NIHSS score on admission was 2, raising the possibility of selection bias in classifying good and poor functional outcomes. Third, mRS is commonly used to assess functional outcome after discharge such as three months after stroke onset. Therefore, in the present study the mRS score might have been higher at hospital discharge than at later time points. Fourth, small proportions of radiological, echocardiographic, and laboratory data were missing. Finally, although major embologenic cardiac diseases were excluded from the current study which was based on the exclusion criteria of ESUS, not a few potential cardiac diseases could not be detected in initial examinations. On the other hand, TTE was performed in 85.7% of the enrolled patients; therefore, potential cardiac embolic sources such as low LVEF might not have been evaluated in the remaining patients.
Conclusion
The ASC-CS classification is useful in stratifying the pathogenesis of CS and clarifying multiple pathologies. Mixed pathology is not uncommon in CS, especially in cases with poor functional outcomes at hospital discharge. The coexistence of cardiac pathology and small-vessel disease in patients with CS might be associated with poor in-hospital functional outcomes.
Sources of Funding
The authors did not receive any specific grants from any funding agency in the public, commercial, or not-for-profit sectors for conducting this research.
Conflict of Interest Statement
YU received personal fees from OHARA Pharmaceutical Co., Ltd. and grants from Bristol-Myers Squibb outside the submitted work. HT received grants from Pfizer Japan Inc. and Daiichi Sankyo Co., Ltd. outside the submitted work. YK received personal fees from Daiichi Sankyo Co., Ltd. and grants from Bristol-Myers Squibb Co., Ltd. and Nippon Boehringer Ingelheim Co., Ltd. outside the submitted work. MK received honoraria from Bayer Pharmaceutical Co. and Daiichi Sankyo Co., Ltd.; consultant fees from Ono Pharmaceutical Co., Ltd.; and research funds from Takeda, Daiichi Sankyo, Nippon Boeringer Ingelheim, Astellas, and Shionogi. MI received grants from Shimadzu Corporation, Otsuka Pharmaceutical, and Panasonic Corporation and personal fees from Daiichi Sankyo Co., Ltd., Eisai Co., Ltd., and Bayer Pharmaceutical Co. outside the submitted work. KH received personal fees from MSD Co., Ltd., Eisai Co., Ltd., Otsuka Pharmaceutical Co., Ltd., Takeda Pharmaceutical Co., Ltd., Pfizer Japan Inc., Novartis Pharma K.K, AbbVie GK, Kyowa Hakko-Kirin Co., Eli Lilly Japan K.K, Amgen K.K., and Lundbeck Japan K.K. and grants from Eisai Co., Ltd., Pfizer Japan Inc., Novartis Pharma K.K., Takeda Pharmaceutical Co., Ltd., TAIYO Co., Ltd., Kyowa Minami Hospital, Shirasawa Hospital, Shiobara Onsen Hospital, Utsunomiya Chuo Hospital, Nishikata Hospital, and Moka Hospital outside the submitted work. YH received personal fees from Bayer Pharmaceutical Co. and Nippon Boehringer Ingelheim, Co., Ltd. during the conduct of the study. NH was an advisory member of Dai-Nippon Sumitomo Pharma Co., Ltd., Hisamitsu Pharmaceutical Co., Inc, and Biogen Idec Japan Ltd.; received lecture fees from Dai-Nippon Sumitomo Pharma Co., Ltd., Otsuka Pharmaceutical, Co., Ltd., Takeda Pharmaceutical Co., Ltd., Kyowa Hakko-Kirin Co., Ltd., FP Pharmaceutical Corporation, Eisai Co., Ltd., Novartis Pharma K.K., and AbbVie; and received departmental endowments by commercial entities from Kyowa Hakko-Kirin Co., Ltd., Nippon Boehringer Ingelheim, Co., Ltd., AbbVie GK, FP Pharmaceutical Corporation, Otsuka Pharmaceutical, Co., Ltd., Dai-Nippon Sumitomo Pharma Co., Ltd., Eisai Co., Ltd., Nihon Medi-physics Co., Ltd., Asahi Kasei Medical Co., Ltd., Ono Pharmaceutical Co., Ltd., MiZ Co., Ltd., AbbVie GK, OHARA Pharmaceutical Co., Ltd., Nihon Pharmaceutical Co., Ltd., Mitsubishi Tanabe Pharma Corporation, Boston Scientific Corporation, and Medtronic Inc. TU received lecture fees from Daiichi Sankyo Co., Ltd., Boehringer Ingelheim, Otsuka Pharmaceutical Co., Ltd., Bayer Pharmaceutical Co., and AstraZeneca K.K. and research funds from Otsuka Pharmaceutical Co., Ltd. and AbbVie GK.
References
- 1) Hart RG, Diener HC, Coutts SB, Easton JD, Granger CB, O’Donnell MJ, Sacco RL, Connolly SJ and Cryptogenic Stroke EIWG: Embolic strokes of undetermined source: the case for a new clinical construct. Lancet Neurol, 2014; 13: 429-438
- 2) Hart RG, Catanese L, Perera KS, Ntaios G and Connolly SJ: Embolic Stroke of Undetermined Source: A Systematic Review and Clinical Update. Stroke, 2017; 48: 867-872
- 3) Sacco RL, Prabhakaran S, Thompson JL, Murphy A, Sciacca RR, Levin B, Mohr JP and Investigators W: Comparison of warfarin versus aspirin for the prevention of recurrent stroke or death: subgroup analyses from the Warfarin-Aspirin Recurrent Stroke Study. Cerebrovasc Dis, 2006; 22: 4-12
- 4) Hart RG, Sharma M, Mundl H, Kasner SE, Bangdiwala SI, Berkowitz SD, Swaminathan B, Lavados P, Wang Y, Wang Y, Davalos A, Shamalov N, Mikulik R, Cunha L, Lindgren A, Arauz A, Lang W, Czlonkowska A, Eckstein J, Gagliardi RJ, Amarenco P, Ameriso SF, Tatlisumak T, Veltkamp R, Hankey GJ, Toni D, Bereczki D, Uchiyama S, Ntaios G, Yoon BW, Brouns R, Endres M, Muir KW, Bornstein N, Ozturk S, O’Donnell MJ, De Vries Basson MM, Pare G, Pater C, Kirsch B, Sheridan P, Peters G, Weitz JI, Peacock WF, Shoamanesh A, Benavente OR, Joyner C, Themeles E, Connolly SJ and Investigators NE: Rivaroxaban for Stroke Prevention after Embolic Stroke of Undetermined Source. N Engl J Med, 2018; 378: 2191-2201
- 5) Diener HC, Sacco RL, Easton JD, Granger CB, Bernstein RA, Uchiyama S, Kreuzer J, Cronin L, Cotton D, Grauer C, Brueckmann M, Chernyatina M, Donnan G, Ferro JM, Grond M, Kallmunzer B, Krupinski J, Lee BC, Lemmens R, Masjuan J, Odinak M, Saver JL, Schellinger PD, Toni D, Toyoda K, Committee R-SES and Investigators: Dabigatran for Prevention of Stroke after Embolic Stroke of Undetermined Source. N Engl J Med, 2019; 380: 1906-1917
- 6) Kashima S, Shimizu T, Akiyama H and Hasegawa Y: Magnetic Resonance Imaging White Matter Hyperintensity as a Predictor of Stroke Recurrence in Patients with Embolic Stroke of Undetermined Source. J Stroke Cerebrovasc Dis, 2018; 27: 3613-3620
- 7) Kikuno M, Ueno Y, Shimizu T, Kuriki A, Tateishi Y, Doijiri R, Shimada Y, Takekawa H, Yamaguchi E, Koga M, Kamiya Y, Ihara M, Tsujino A, Hirata K, Toyoda K, Hasegawa Y, Aizawa H, Hattori N, Urabe T and collaborators CECS: Underlying embolic and pathologic differentiation by cerebral microbleeds in cryptogenic stroke. J Neurol, 2020; 267: 1482-1490
- 8) Ueno Y, Tateishi Y, Doijiri R, Kuriki A, Shimizu T, Kikuno M, Shimada Y, Takekawa H, Yamaguchi E, Koga M, Kamiya Y, Ihara M, Tsujino A, Hirata K, Toyoda K, Hasegawa Y, Hattori N and Urabe T: Large aortic arch plaques correlate with CHADS2 and CHA2DS2-VASc scores in cryptogenic stroke. Atherosclerosis, 2019; 284: 181-186
- 9) Ntaios G, Pearce LA, Meseguer E, Endres M, Amarenco P, Ozturk S, Lang W, Bornstein NM, Molina CA, Pagola J, Mundl H, Berkowitz SD, Liu YY, Sen S, Connolly SJ and Hart RG: Aortic Arch Atherosclerosis in Patients With Embolic Stroke of Undetermined Source: An Exploratory Analysis of the NAVIGATE ESUS Trial. Stroke, 2019; 50: 3184-3190
- 10) Healey JS, Gladstone DJ, Swaminathan B, Eckstein J, Mundl H, Epstein AE, Haeusler KG, Mikulik R, Kasner SE, Toni D, Arauz A, Ntaios G, Hankey GJ, Perera K, Pagola J, Shuaib A, Lutsep H, Yang X, Uchiyama S, Endres M, Coutts SB, Karlinski M, Czlonkowska A, Molina CA, Santo G, Berkowitz SD, Hart RG and Connolly SJ: Recurrent Stroke With Rivaroxaban Compared With Aspirin According to Predictors of Atrial Fibrillation: Secondary Analysis of the NAVIGATE ESUS Randomized Clinical Trial. JAMA Neurol, 2019; 76: 764-773
- 11) Kasner SE, Swaminathan B, Lavados P, Sharma M, Muir K, Veltkamp R, Ameriso SF, Endres M, Lutsep H, Messe SR, Spence JD, Nedeltechev K, Perera K, Santo G, Olavarria V, Lindgren A, Bangdiwala S, Shoamanesh A, Berkowitz SD, Mundl H, Connolly SJ, Hart RG and Investigators NE: Rivaroxaban or aspirin for patent foramen ovale and embolic stroke of undetermined source: a prespecified subgroup analysis from the NAVIGATE ESUS trial. Lancet Neurol, 2018; 17: 1053-1060
- 12) Amarenco P, Bogousslavsky J, Caplan LR, Donnan GA, Wolf ME and Hennerici MG: The ASCOD phenotyping of ischemic stroke (Updated ASCO Phenotyping). Cerebrovasc Dis, 2013; 36: 1-5
- 13) Adams HP, Jr., Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL and Marsh EE, 3rd: Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke, 1993; 24: 35-41
- 14) Shimizu T, Kashima S, Akiyama H, Isahaya K and Hasegawa Y: The ASCOD Phenotyping of Embolic Strokes of Undetermined Source. J Stroke Cerebrovasc Dis, 2020; 29: 104491
- 15) Sandercock P, Bamford J, Dennis M, Burn J, Slattery J, Jones L, Boonyakarnkul S and Warlow C: Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long term prognosis (Oxfordshire community stroke project). BMJ, 1992; 305: 1460-1465
- 16) Suto Y, Nakayasu H, Maeda M, Kusumi M, Kowa H, Awaki E, Saito J and Nakashima K: Long-term prognosis of patients with large subcortical infarctions. Eur Neurol, 2009; 62: 304-310
- 17) Ntaios G, Papavasileiou V, Milionis H, Makaritsis K, Vemmou A, Koroboki E, Manios E, Spengos K, Michel P and Vemmos K: Embolic Strokes of Undetermined Source in the Athens Stroke Registry: An Outcome Analysis. Stroke, 2015; 46: 2087-2093
- 18) Goda A, Yamashita T, Suzuki S, Ohtsuka T, Uejima T, Oikawa Y, Yajima J, Koike A, Nagashima K, Kirigaya H, Sagara K, Ogasawara K, Isobe M, Sawada H and Aizawa T: Heart failure with preserved versus reduced left ventricular systolic function: a prospective cohort of Shinken Database 2004-2005. J Cardiol, 2010; 55: 108-116
- 19) Nouh AM, Staff I and Finelli PF: Three Territory Sign: An MRI marker of malignancy-related ischemic stroke (Trousseau syndrome). Neurol Clin Pract, 2019; 9: 124-128
- 20) Ueno Y, Kimura K, Iguchi Y, Shibazaki K, Inoue T, Hattori N and Urabe T: Mobile aortic plaques are a cause of multiple brain infarcts seen on diffusion-weighted imaging. Stroke, 2007; 38: 2470-2476
- 21) Maisel AS, Krishnaswamy P, Nowak RM, McCord J, Hollander JE, Duc P, Omland T, Storrow AB, Abraham WT, Wu AH, Clopton P, Steg PG, Westheim A, Knudsen CW, Perez A, Kazanegra R, Herrmann HC, McCullough PA and Breathing Not Properly Multinational Study I: Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med, 2002; 347: 161-167
- 22) Januzzi JL, Jr., Chen-Tournoux AA and Moe G: Amino-terminal pro-B-type natriuretic peptide testing for the diagnosis or exclusion of heart failure in patients with acute symptoms. Am J Cardiol, 2008; 101: 29-38
- 23) Suda S, Muraga K, Kanamaru T, Okubo S, Abe A, Aoki J, Suzuki K, Sakamoto Y, Shimoyama T, Nito C and Kimura K: Low free triiodothyronine predicts poor functional outcome after acute ischemic stroke. J Neurol Sci, 2016; 368: 89-93
- 24) Sirimarco G, Lavallee PC, Labreuche J, Meseguer E, Cabrejo L, Guidoux C, Klein IF, Olivot JM, Abboud H, Adrai V, Kusmierek J, Ratani S, Touboul PJ, Mazighi M, Steg PG and Amarenco P: Overlap of diseases underlying ischemic stroke: the ASCOD phenotyping. Stroke, 2013; 44: 2427-2433
- 25) Fujimoto S, Toyoda K, Jinnouchi J, Yasaka M, Kitazono T and Okada Y: Differences in diffusion-weighted image and transesophageal echocardiographical findings in cardiogenic, paradoxical and aortogenic brain embolism. Cerebrovasc Dis, 2011; 32: 148-154
- 26) Kimura K, Kazui S, Minematsu K, Yamaguchi T and Japan Multicenter Stroke Investigator’s C: Analysis of 16,922 patients with acute ischemic stroke and transient ischemic attack in Japan. A hospital-based prospective registration study. Cerebrovasc Dis, 2004; 18: 47-56
- 27) Nezu T, Hosomi N, Kondo K, Aoki S, Matsumoto M, Kobayashi S and Japan Standard Stroke Registry Study G: Greater Severity of Neurological Defects in Women Admitted With Atrial Fibrillation-Related Stroke. Circ J, 2016; 80: 250-255
- 28) Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW and Radford MJ: Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA, 2001; 285: 2864-2870
- 29) Takano K, Yamaguchi T, Kato H and Omae T: Activation of coagulation in acute cardioembolic stroke. Stroke, 1991; 22: 12-16
- 30) Dougu N, Takashima S, Sasahara E, Taguchi Y, Toyoda S, Hirai T, Nozawa T, Tanaka K and Inoue H: Differential diagnosis of cerebral infarction using an algorithm combining atrial fibrillation and D-dimer level. Eur J Neurol, 2008; 15: 295-300
- 31) Liu Z, Sanossian N, Starkman S, Avila-Rinek G, Eckstein M, Sharma LK, Liebeskind D, Conwit R, Hamilton S, Saver JL, Investigators F-M and Coordinators: Adiposity and Outcome After Ischemic Stroke: Obesity Paradox for Mortality and Obesity Parabola for Favorable Functional Outcomes. Stroke, 2021; 52: 144-151
- 32) Appleton JP, Woodhouse LJ, Adami A, Becker JL, Berge E, Cala LA, Casado AM, Caso V, Christensen HK, Dineen RA, Gommans J, Koumellis P, Szatmari S, Sprigg N, Bath PM, Wardlaw JM and Investigators E: Imaging markers of small vessel disease and brain frailty, and outcomes in acute stroke. Neurology, 2020; 94: e439-e452
- 33) Kamakura CK, Ueno Y, Sakai Y, Yoshida H, Aiba S, Hayashi A, Shimura H, Takeda K, Kamakura K, Hattori N and Urabe T: White matter lesions and cognitive impairment may be related to recovery from unilateral spatial neglect after stroke. J Neurol Sci, 2017; 379: 241-246
- 34) Kobayashi A, Iguchi M, Shimizu S and Uchiyama S: Silent cerebral infarcts and cerebral white matter lesions in patients with nonvalvular atrial fibrillation. J Stroke Cerebrovasc Dis, 2012; 21: 310-317
- 35) Hwang J, Chung JW, Lee MJ, Kim GM, Chung CS, Lee KH and Bang OY: Implications of CHA2DS2-VASc Score in Stroke Patients with Atrial Fibrillation: An Analysis of 938 Korean Patients. Eur Neurol, 2017; 77: 307-315
- 36) Marti-Fabregas J, Medrano-Martorell S, Merino E, Prats-Sanchez L, Marin R, Delgado-Mederos R, Martinez-Domeno A, Camps-Renom P, Jimenez-Xarrie E, Zedde M, Gomez-Choco M, Lara L, Boix A, Calleja A, De Arce-Borda AM, Bravo Y, Fuentes B, Hernandez-Perez M, Canovas D, Llull L, Zandio B, Freijo M, Casado-Naranjo I, Sanahuja J, Cocho D, Krupinski J, Rodriguez-Campello A, Palomeras E, De Felipe A, Serrano M, Zapata-Arriaza E, Zaragoza-Brunet J, Diaz-Maroto I, Fernandez-Dominguez J, Lago A, Maestre J, Rodriguez-Yanez M, Gich I and investigators Hs: MRI predicts intracranial hemorrhage in patients who receive long-term oral anticoagulation. Neurology, 2019; 92: e2432-e2443
- 37) Shoamanesh A, Hart RG, Connolly SJ, Kasner SE, Smith EE, Marti-Fabregas J, Liu YY, Uchiyama S, Mikulik R, Veltkamp R, O’Donnell MJ, Ntaios G, Muir KW, Field TS, Santo GC, Olavarria V, Mundl H, Lutsep H, Berkowitz SD and Sharma M: Microbleeds and the Effect of Anticoagulation in Patients With Embolic Stroke of Undetermined Source: An Exploratory Analysis of the NAVIGATE ESUS Randomized Clinical Trial. JAMA Neurol, 2021; 78: 11-20
