Introduction: Enlarged perivascular spaces (EPVS) are considered early manifestations of impaired clearance mechanisms in the brain; however, it is unclear whether EPVS are associated with the development of malignant cerebral edema (MCE) after large hemispheric infarction (LHI). Therefore, we investigated the predictive value of EPVS in predicting MCE in LHI. Methods: Patients suffering from acute LHI were consecutively enrolled. EPVS were rated after the stroke with validated rating scales from magnetic resonance images. Patients were divided into two groups according to the occurrence of MCE. Logistic regression was used to analyze the relationship between EPVS and MCE in the basal ganglia (BG) and centrum semiovale (CS) regions. Receiver operating characteristic (ROC) curves assessed the ability of EPVS individually and with other factors in predicting MCE. Results: We included a total of 255 patients, of whom 98 were MCE patients (58 [59.2%] males, aged 70 [range = 61.75–78] years) and found that atrial fibrillation, National Institutes of Health Stroke Scale score, infarct volume, neutrophil-lymphocyte ratio, and moderate-to-severe CS-EPVS were positively associated with MCE. After adjusting for confounds, moderate-to-severe CS-EPVS remained independent risk factor of MCE (odds ratio = 16.212, p < 0.001). According to the ROC analysis, MCE was highly suspected when CS-EPVS >14 (sensitivity = 0.82, specificity = 0.48), and the guiding value was higher when CS-EPVS combined with other MCE predictors (area under the curve = 0.90, sensitivity = 0.74, specificity = 0.90). Conclusion: CS-EPVS were important risk factor for MEC in patients with acute LHI and can help identify patients at risk for MCE.

Large hemisphere infarction (LHI) is a severe form of ischemic stroke affecting the majority of or complete middle cerebral artery distribution area with or without anterior cerebral artery and posterior cerebral artery involvement and can develop into severe cerebral edema and be life-threatening [1]. LHI with malignant cerebral edema (MCE) demonstrates a poor prognosis with mortality rates as high as 80%. Rapid deterioration of clinical function can occur in the early stage, midline displacement of imaging, and brain hernias can form or lead to death [2, 3]. Previous studies have shown that timely decompressive craniectomy can reduce the mortality of patients with LHI with MCE [4]. Therefore, early identification of patients prone to MCE is particularly important.

Recent research has found that the cerebral glymphatic system plays an important role in the inflow and outflow of cerebrospinal fluid (CSF), and the dysfunction of glymphatic system may be an important cause of cerebral edema after ischemic stroke [5]. Cerebral glymphatic system is a fluid transport and material clearance system through the perivascular spaces (PVS) and aquaporin 4 on astrocytes found by lliff in 2012 [6]. PVS are an important part of the glymphatic system, defined as the fluid-containing space surrounding penetrating vessels [7]. PVS can be enlarged and visible with magnetic resonance imaging (MRI) due to dysfunctions in the fluid circulation of the glymphatic system and are commonly seen in the basal ganglia (BG) and centrum semiovale (CS) [8]. Studies have found that enlarged perivascular spaces (EPVS) were early manifestation of impaired fluid and waste removal from brain tissue and microvascular dysfunction associated after ischemic stroke [9, 10]; however, the effect of EPVS on edema formation after ischemic stroke remains unclear.

Therefore, in the present study, we investigated the correlation between EPVS and the occurrence of MCE by observing the distribution and severity of EPVS in patients with acute LHI. We analyzed and summarized the risk factors for the formation of MCE after acute LHI to provide a reference basis for better clinical guidance.

Study Sample

Patients with acute cerebral infarction within 24 h of onset admitted to the Department of Neurology at Xianyang Hospital of Yan’an University from January 2018 to December 2022 were consecutively enrolled. Study inclusion criteria included (1) age ≥18 years; (2) cranial computed tomography (CT) or diffusion-weighted imaging (DWI) confirmed infarction in the middle cerebral artery or internal carotid artery distribution area, and final infarct volume >1/2 of the middle cerebral artery distribution area; and (3) cranial MRI could be obtained before midline shift. Exclusion criteria included (1) CT or MRI showing intracranial hemorrhage or hemorrhagic transformation; (2) midline shift on admission; (3) previous cerebral infarction and modified Rankin Scale score ≥2; (4) bilateral cerebral hemisphere infarction; and (5) incomplete case data or poor image quality (Fig. 1). The current study was approved by the Research Ethics Committee of Yan’an University Affiliated Xianyang Hospital in accordance with the Declaration of Helsinki (revised 2013). Since this study was a retrospective cohort study, the requirement for individual informed consent was waived.

Fig. 1.

Flowchart of patients in this study.

Fig. 1.

Flowchart of patients in this study.

Close modal

Data Collection

Baseline clinical information (i.e., age, sex, systolic and diastolic blood pressure on admission, and glucose), medical history (i.e., hypertension, diabetes, coronary artery disease, atrial fibrillation [AF], history of smoking, history of alcohol consumption, and history of stroke), clinical features (i.e., baseline scores for Glasgow Coma Scale [GCS] and National Institutes of Health Stroke Scale [NIHSS]), laboratory tests (i.e., white blood cell count, neutrophil-lymphocyte ratio [NLR], platelet count, sedimentation, total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, creatinine, uric acid, homocysteine, fibrin degradation products [FDP], fibrinogen, and D-dimer), imaging features (i.e., infarct volume and EPVS count), and revascularization (i.e., thrombolysis and thrombectomy) were collected. All clinical data were obtained from the case registration database of Yan’an University Affiliated Xianyang Hospital.

Image Acquisition

Non-contrast CT and MR images were acquired using a standardized protocol as part of the routine clinical assessments. Standard non-contrast CT was conducted with a 64-slice multidetector CT machine (Maxima and ACE and Optima 680E; GE) with the following parameters: detector collimation = 64 × 0.6 mm, table feed = 10.62 mm per rotation, rotation time = 1,000 ms, tube current = 110 mA, tube voltage = 120 kV, field of view = 25 cm, matrix = 512 × 512, and slice thickness = 5 mm. MRI was performed on 1.5 T (Prodiva CX Philips) or 3.0 T Siemens (Verio, Germany) machines (24–30 slices of 5 mm thickness, 0.50 mm gaps, matrix = 256 × 256). We acquired T1- and T2-weighted, fluid-attenuated inversion recovery (FLAIR), and DWI sequences. The imaging parameters of the 1.5 T acquisition were T1 (repetition time [TR] = 565 ms; echo time [TE] = 10 ms); T2 (TR = 4,359 ms; TE = 90 ms); FLAIR (TR = 8,000 ms; TE = 100 ms); and DWI (TR = 3,454 ms; TE = 90 ms). The imaging parameters of the 3.0 T acquisition were T1 (TR = 2,430 ms; TE = 9 ms); T2 (TR = 4730 ms; TE = 99 ms); FLAIR (TR = 9,000 ms; TE = 79 ms); and DWI (TR = 5,700 ms; TE = 86 ms). Cranial NCCT/MRI was immediately repeated if the patient’s condition worsened.

Imaging Analysis

EPVS were rated on axial T2-weighted MRI using a validated visual rating scale. EPVS were defined as fluid-filled spaces surrounding the penetrating vessels with a signal intensity identical to CSF on all MRI sequences (i.e., T2 hyperintensities and T1/FLAIR hypointensities with respect to the brain along the course of penetrating arteries). They were linear, round, or ovoid in shape; generally ≤3 mm in diameter; and had no peripheral high signal rim. The distribution of EPVS is usually bilaterally symmetric. EPVS were counted in the BG and CS using the following 5-point rating scale: 0 = no EPVS, 1 = 1‒10, 2 = 11‒20, 3 = 21‒40, and 4 = ≥40, with 0–1 being mild and 2–4 being moderate-to-severe [11]. Total EPVS burden was calculated by summing the BG and CS-EPVS scores. We counted the most numerous layers on the contralateral side of the infarction. Because the structural damage caused by LHI is typically severe enough to limit the accurate score of EPVS in its immediate vicinity, only the contralateral side was rated [12]. The volume of infarction was calculated using the Pullicino formula (longest diameter of infarct × wide diameter × number of infarct layers × layer thickness × 0.5) based on the final CT/MRI infarction range [13].

All imaging data were analyzed by trained and experienced medical imaging physicians who were unaware of the patient’s clinical data. Discussion or a third party was used to reach consensus regarding disagreements.

Clinical Outcome

MCE was defined as a clinically worsening syndrome, including progressive neurological deterioration (NIHSS score increased by ≥2 points), a decrease in consciousness (score ≥1 on item 1a of the NIHSS), bilateral anisocoria with imaging evidence of brain swelling (CT or DWI showing midline shift of ≥5 mm), even resulting in cerebral herniation or death [2, 14, 15].

Statistical Analysis

SPSS (version 26.0; IBM Corp., Armonk, NY, USA) was used to conduct these analyses. Data were tested for normality; normally distributed variables are described as mean (standard deviation) and non-normally distributed variables are described by medians (interquartile ranges). Comparisons of variables were made with independent sample t tests or Mann-Whitney U tests, as appropriate. Categorical count data are described by the number of cases (percentage), with Mann-Whitney U or χ2 tests conducted to compare ordered versus unordered categorical data, respectively. A multifactorial regression was used to analyze the related influencing factors of MCE. Patients with no MCE were used as the control group. The relationship between each factor and MCE was first analyzed by univariate analysis for the MCE group. Variables with p < 0.1 from the univariate analysis were entered into the multifactorial analysis. Differences were considered statistically significant at p < 0.05. The characteristics of the two groups were analyzed, receiver operating characteristic (ROC) curves were created, and optimal cutoff values were calculated using MedCalc (version 19.8.0; MedCalc Software Ltd., Ostend, Belgium) to determine the predictive utility of EPVS among other factors. In addition, we performed sensitivity analysis on the data to verify the reliability of the results.

Clinical Characteristics

During the study period, of 351 patients screened, 96 (27.3%) patients were excluded for the following reasons: no MRI scans (n = 22), poor quality of MRI images (n = 14), midline shift on admission (n = 9), intracranial hemorrhage and hemorrhagic transformation (n = 16), previous cerebral infarction and modified Rankin Scale ≥2 points (n = 11), and incomplete or missing clinical data (n = 24). Therefore, 255 patients were included in the current study. Of these, 98 patients developed MCE (Fig. 1). Comparing age and gender, there were no significant differences between the MCE and non-MCE groups (p > 0.05; MCE: mean age = 70 [61.75, 78] years, <60 years old (23.5%), 59.2% male; non-MCE: mean age = 68 [59, 75] years, <60 years old (28.0%), 60.5% male). Compared to the MCE group, the non-MCE group had significantly fewer instances of AF history (21% vs. 36.7%, p = 0.006) and previous anticoagulation (1.3% vs. 6.1%, p = 0.031) and significantly lower NIHSS scores, admission glucose, D-dimer, FDP, white blood cell count, and NLR (p < 0.05). Thus, these may represent risk factors for the development of MCE after LHI. The number of CS-EPVS and BG-EPVS were also significantly different between the two groups in terms of imaging indices, except for infarct volume, and were significantly higher in the MCE group than in the non-MCE group (p < 0.05). Therefore, EPVS may be closely related to the development of MCE (Table 1).

Table 1.

Baseline characteristics of patients with and without MCE

CharacteristicsMED (n = 98)Without MED (n = 157)χ2/Zp value
Age, years, median (IQR) 70 (61.75–78) 68 (59–75) −1.047 0.295 
Male, n (%) 58 (59.2) 95 (60.5) 0.044 0.833 
Hypertensive disease, n (%) 49 (50) 93 (59.2) 2.086 0.149 
Diabetes, n (%) 26 (26.5) 36 (22.9) 0.425 0.514 
Coronary heart disease, n (%) 33 (33.7) 41 (26.1) 1.674 0.196 
AF, n (%) 36 (36.7) 33 (21.0) 7.55 0.006 
Alcohol consumption, n (%) 3 (3.1) 11 (7.0) 1.81 0.179 
Smoking, n (%) 33 (33.7) 43 (27.4) 1.139 0.286 
Prior stroke, n (%) 23 (23.5) 50 (31.8) 2.073 0.150 
Priors anticoagulation, n (%) 6 (6.1) 2 (1.3) 4.668 0.031 
Baseline NIHSS score, median (IQR) 18 (14–20) 10 (7–15.5) −7.181 <0.001 
Admission GCS score, median (IQR) 12 (8–13) 15 (13–15) −6.466 <0.001 
Systolic blood pressure, mm Hg, median (IQR) 150 (137.75–170) 150 (134.5–164) −0.43 0.668 
Diastolic blood pressure, mm Hg, median (IQR) 85 (75.75–94.25) 83 (73–90.5) −0.74 0.460 
Admission blood glucose, mmol/L, median (IQR) 6.38 (5.67–7.86) 5.7 (4.97–7.33) −2.491 0.013 
Thrombolysis, n (%) 25 (25.5) 36 (22.9) 0.221 0.638 
Thrombus removal, n (%) 4 (4.1) 5 (3.2) 0.143 0.706 
TC, mmol/L, median (IQR) 4.08 (3.33–4.67) 3.89 (3.31–4.63) −0.242 0.809 
TG, mmol/L, median (IQR) 1.19 (0.89–1.59) 1.21 (0.90–1.67) −0.635 0.525 
HDL-C, mmol/L, median (IQR) 1.04 (0.91–1.27) 1.03 (0.86–1.23) −1.409 0.159 
LDL-C, mmol/L, median (IQR) 2.37 (1.89–3.0) 2.46 (2.06–3.08) −0.875 0.381 
Creatinine, μmol/L, median (IQR) 71.7 (59.78–85.03) 72 (62.05–84.5) −0.286 0.775 
Uric acid, μmol/L, median (IQR) 264.5 (178.25–323.25) 270 (214–332) −1.046 0.295 
D-dimer, μg/L, median (IQR) 1.4 (0.75–2.27) 0.82 (0.47–1.61) −3.593 <0.001 
FDP, μg/L, median (IQR) 3.99 (2.21–7.83) 2.65 (1.75–5.49) −3.299 0.001 
Fib, g/L, median (IQR) 3.06 (2.56–4.37) 3.23 (2.8–4.05) −0.573 0.567 
HCY, μmol/L, median (IQR) 18.6 (15.9–21.75) 17.9 (16–22.35) −0.17 0.865 
Platelet, median (IQR) 181 (143–217) 198 (157.5–243.5) −2.366 0.018 
Hemoglobin, mm/h, median (IQR) 13 (8–28) 12 (7–27) −0.635 0.525 
Leukocyte count, median (IQR) 10.08 (7.7–11.67) 7.54 (6.25–10.3) −4.503 <0.001 
NLR, median (IQR) 7.07 (5.01–11.02) 4.73 (3.06–7.23) −4.92 <0.001 
Infarct volume, median (IQR) 170.5 (125–220.25) 95 (69–128) −8.672 <0.001 
Total EPVS, median (IQR) 37.5 (24.75–50.5) 27 (14–44.5) −3.686 <0.001 
BG-EPVS, median (IQR) 12 (6–20.5) 7 (0–14) −3.09 0.002 
Moderate-to-severe BG-EPVS, n (%) 53 (54.1) 64 (40.8) −2.072 0.038 
CS-EPVS, median (IQR) 22 (16–30) 15 (8–30) −3.549 <0.001 
Moderate-to-severe CS-EPVS, n (%) 90 (91.8) 107 (68.2) −4.38 <0.001 
CharacteristicsMED (n = 98)Without MED (n = 157)χ2/Zp value
Age, years, median (IQR) 70 (61.75–78) 68 (59–75) −1.047 0.295 
Male, n (%) 58 (59.2) 95 (60.5) 0.044 0.833 
Hypertensive disease, n (%) 49 (50) 93 (59.2) 2.086 0.149 
Diabetes, n (%) 26 (26.5) 36 (22.9) 0.425 0.514 
Coronary heart disease, n (%) 33 (33.7) 41 (26.1) 1.674 0.196 
AF, n (%) 36 (36.7) 33 (21.0) 7.55 0.006 
Alcohol consumption, n (%) 3 (3.1) 11 (7.0) 1.81 0.179 
Smoking, n (%) 33 (33.7) 43 (27.4) 1.139 0.286 
Prior stroke, n (%) 23 (23.5) 50 (31.8) 2.073 0.150 
Priors anticoagulation, n (%) 6 (6.1) 2 (1.3) 4.668 0.031 
Baseline NIHSS score, median (IQR) 18 (14–20) 10 (7–15.5) −7.181 <0.001 
Admission GCS score, median (IQR) 12 (8–13) 15 (13–15) −6.466 <0.001 
Systolic blood pressure, mm Hg, median (IQR) 150 (137.75–170) 150 (134.5–164) −0.43 0.668 
Diastolic blood pressure, mm Hg, median (IQR) 85 (75.75–94.25) 83 (73–90.5) −0.74 0.460 
Admission blood glucose, mmol/L, median (IQR) 6.38 (5.67–7.86) 5.7 (4.97–7.33) −2.491 0.013 
Thrombolysis, n (%) 25 (25.5) 36 (22.9) 0.221 0.638 
Thrombus removal, n (%) 4 (4.1) 5 (3.2) 0.143 0.706 
TC, mmol/L, median (IQR) 4.08 (3.33–4.67) 3.89 (3.31–4.63) −0.242 0.809 
TG, mmol/L, median (IQR) 1.19 (0.89–1.59) 1.21 (0.90–1.67) −0.635 0.525 
HDL-C, mmol/L, median (IQR) 1.04 (0.91–1.27) 1.03 (0.86–1.23) −1.409 0.159 
LDL-C, mmol/L, median (IQR) 2.37 (1.89–3.0) 2.46 (2.06–3.08) −0.875 0.381 
Creatinine, μmol/L, median (IQR) 71.7 (59.78–85.03) 72 (62.05–84.5) −0.286 0.775 
Uric acid, μmol/L, median (IQR) 264.5 (178.25–323.25) 270 (214–332) −1.046 0.295 
D-dimer, μg/L, median (IQR) 1.4 (0.75–2.27) 0.82 (0.47–1.61) −3.593 <0.001 
FDP, μg/L, median (IQR) 3.99 (2.21–7.83) 2.65 (1.75–5.49) −3.299 0.001 
Fib, g/L, median (IQR) 3.06 (2.56–4.37) 3.23 (2.8–4.05) −0.573 0.567 
HCY, μmol/L, median (IQR) 18.6 (15.9–21.75) 17.9 (16–22.35) −0.17 0.865 
Platelet, median (IQR) 181 (143–217) 198 (157.5–243.5) −2.366 0.018 
Hemoglobin, mm/h, median (IQR) 13 (8–28) 12 (7–27) −0.635 0.525 
Leukocyte count, median (IQR) 10.08 (7.7–11.67) 7.54 (6.25–10.3) −4.503 <0.001 
NLR, median (IQR) 7.07 (5.01–11.02) 4.73 (3.06–7.23) −4.92 <0.001 
Infarct volume, median (IQR) 170.5 (125–220.25) 95 (69–128) −8.672 <0.001 
Total EPVS, median (IQR) 37.5 (24.75–50.5) 27 (14–44.5) −3.686 <0.001 
BG-EPVS, median (IQR) 12 (6–20.5) 7 (0–14) −3.09 0.002 
Moderate-to-severe BG-EPVS, n (%) 53 (54.1) 64 (40.8) −2.072 0.038 
CS-EPVS, median (IQR) 22 (16–30) 15 (8–30) −3.549 <0.001 
Moderate-to-severe CS-EPVS, n (%) 90 (91.8) 107 (68.2) −4.38 <0.001 

MCE, malignant cerebral edema; AF, atrial fibrillation; NIHSS, National Institutes of Health Stroke Scale; GCS, Glasgow Coma Scale; TC, total serum cholesterol; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; HCY, serum homocysteine; FDP, fibrin degradation products; Fib, fibrinogen; NLR, neutrophil-to-lymphocyte ratio; BG, basal ganglia; CS, centrum semiovale; EPVS, enlarged perivascular spaces; IQR, interquartile range.

Multifactorial Regression Analysis of MCE

A multifactorial regression logistic regression analysis was performed with the presence or absence of MCE as the dependent variable and age, AF, admission NIHSS score, GCS score, admission glucose level, infarct volume, NLR, and moderate-to-severe EPVS as independent variables (we test the collinearity of NHISS and GCS, and the results show that they do not exist collinearity, VIF: 2.460, tolerance: 0.406). Our results showed that AF, NIHSS score, infarct volume, NLR, and moderate-to-severe CS-EPVS were positively associated with the occurrence of MCE (p < 0.05) and that the risk of MCE increased with the severity of CS-EPVS (odds ratio [OR] = 14.282; Table 2).

Table 2.

Multifactorial regression analysis of MCE

VariablesCoefficientSEOR95% CIp value
Age −0.014 0.017 0.986 0.954∼1.020 0.415 
AF 0.930 0.432 2.536 1.087∼5.915 0.031 
Baseline NIHSS score 0.146 0.050 1.157 1.049∼1.276 0.003 
Admission GCS score 0.115 0.097 1.122 0.929–∼1.356 0.233 
Admission blood glucose −0.003 0.064 0.997 0.879∼1.131 0.959 
NLR 0.096 0.037 1.101 1.024∼1.183 0.009 
Infarct volume 0.023 0.004 1.024 1.016∼1.031 <0.001 
Moderate-to-severe BG-EPVS 0.537 0.397 1.711 0.785∼3.730 0.177 
Moderate-to-severe CS-EPVS 2.659 0.582 14.282 4.566∼44.672 <0.001 
VariablesCoefficientSEOR95% CIp value
Age −0.014 0.017 0.986 0.954∼1.020 0.415 
AF 0.930 0.432 2.536 1.087∼5.915 0.031 
Baseline NIHSS score 0.146 0.050 1.157 1.049∼1.276 0.003 
Admission GCS score 0.115 0.097 1.122 0.929–∼1.356 0.233 
Admission blood glucose −0.003 0.064 0.997 0.879∼1.131 0.959 
NLR 0.096 0.037 1.101 1.024∼1.183 0.009 
Infarct volume 0.023 0.004 1.024 1.016∼1.031 <0.001 
Moderate-to-severe BG-EPVS 0.537 0.397 1.711 0.785∼3.730 0.177 
Moderate-to-severe CS-EPVS 2.659 0.582 14.282 4.566∼44.672 <0.001 

MCE, malignant cerebral edema; AF, atrial fibrillation; NIHSS, National Institutes of Health Stroke Scale; GCS, Glasgow Coma Scale; NLR, neutrophil-to-lymphocyte ratio; BG, basal ganglia; CS, centrum semiovale; EPVS, enlarged perivascular spaces; SE, standard error; OR, odds ratio; CI, confidence interval.

Univariate and Multifactorial Logistic Regression Analyses Affecting MCE in Patients with LHI

Compared with the non-MCE group, patients in the MCE group had a higher proportion of AF, higher NIHSS scores on admission, and larger infarct volumes (p < 0.05). Their admission GCS scores were lower (OR = 0.828, p < 0.001), which was negatively correlated with the occurrence of MCE. Moreover, the MCE group had a lower platelet count (OR = 0.995, p = 0.016) and higher D-dimer, FDP, leukocyte count, and NLR (OR = 1.195, 1.039, 1.167, 1.142, respectively; p = 0.014, 0.059, 0.001, <0.001, respectively), which may be independent risk factors for MCE (Table 3). In the multifactorial analysis, after correcting for significant factors from the univariate analysis (i.e., AF, prior anticoagulation, admission NIHSS score, admission GCS score, D-dimer, FDP, platelets, infarct volume, leukocytes, NLR, BG-EPVS, and CS-EPVS), we found that moderate-to-severe CS-EPVS were independent risk factor for the development of MCE (p < 0.05). In the MCE group, the probability of a patient’s risk of developing MCE was significantly higher for each unit increase in the patient’s CS-EPVS grade (OR = 16.212, p < 0.001; Table 4). To verify the reliability of the results, we performed a sensitivity analysis, and the results showed that CS-EPVS remained a risk factor for MCE, which was consistent with the results of our main analysis.

Table 3.

Univariate analysis of factors affecting MCE after acute LHI

VariablesOR95% CIp value
Age 1.010 0.989∼1.031 0.373 
Sex 1.057 0.632∼1.768 0.833 
Hypertensive disease 0.688 0.414∼1.144 0.149 
Diabetes 1.214 0.678∼2.174 0.515 
Coronary heart disease 1.436 0.829∼2.489 0.197 
AF 2.182 1.244∼3.828 0.007 
Alcohol consumption 0.419 0.114∼1.542 1.191 
Smoking 1.346 0.779∼2.325 0.287 
Prior stroke 0.656 0.369∼1.167 0.151 
Priors anticoagulation 5.054 0.999∼25.565 0.050 
Baseline NIHSS score 1.149 1.098∼1.202 <0.001 
Admission GCS score 0.828 0.763∼0.898 <0.001 
Systolic blood pressure 1.002 0.991∼1.013 0.726 
Diastolic blood pressure 1.004 0.985∼1.023 0.696 
Admission blood glucose 1.003 0.931∼1.080 0.947 
Thrombolysis 1.151 0.640∼2.071 0.639 
Thrombus removal 1.294 0.339∼4.939 0.706 
TC 0.972 0.773∼1.222 0.808 
TG 0.824 0.569∼1.192 0.305 
HDL-C 1.571 0.728∼3.387 0.249 
LDL-C 0.853 0.664∼1.129 0.266 
Creatinine 1.004 0.997∼1.011 0.299 
Uric acid 0.999 0.997∼1.001 0.442 
D-dimer 1.195 1.036∼1.379 0.014 
FDP 1.039 0.999∼1.081 0.059 
Fib 1.018 0.817∼1.267 0.874 
HCY 0.998 0.950∼1.047 0.920 
Platelet 0.995 0.991∼0.999 0.016 
Hemoglobin 1.004 0.985∼1.022 0.692 
Leukocyte count 1.167 1.069∼1.274 0.001 
NLR 1.142 1.073∼1.215 <0.001 
Infarct volume 1.023 1.016∼1.029 <0.001 
Mild BG-EPVS 1.000 
Moderate-to-severe BG-EPVS 1.711 1.029∼2.847 0.039 
Mild CS-EPVS 1.000 
Moderate-to-severe CS-EPVS 5.257 2.369∼11.667 <0.001 
VariablesOR95% CIp value
Age 1.010 0.989∼1.031 0.373 
Sex 1.057 0.632∼1.768 0.833 
Hypertensive disease 0.688 0.414∼1.144 0.149 
Diabetes 1.214 0.678∼2.174 0.515 
Coronary heart disease 1.436 0.829∼2.489 0.197 
AF 2.182 1.244∼3.828 0.007 
Alcohol consumption 0.419 0.114∼1.542 1.191 
Smoking 1.346 0.779∼2.325 0.287 
Prior stroke 0.656 0.369∼1.167 0.151 
Priors anticoagulation 5.054 0.999∼25.565 0.050 
Baseline NIHSS score 1.149 1.098∼1.202 <0.001 
Admission GCS score 0.828 0.763∼0.898 <0.001 
Systolic blood pressure 1.002 0.991∼1.013 0.726 
Diastolic blood pressure 1.004 0.985∼1.023 0.696 
Admission blood glucose 1.003 0.931∼1.080 0.947 
Thrombolysis 1.151 0.640∼2.071 0.639 
Thrombus removal 1.294 0.339∼4.939 0.706 
TC 0.972 0.773∼1.222 0.808 
TG 0.824 0.569∼1.192 0.305 
HDL-C 1.571 0.728∼3.387 0.249 
LDL-C 0.853 0.664∼1.129 0.266 
Creatinine 1.004 0.997∼1.011 0.299 
Uric acid 0.999 0.997∼1.001 0.442 
D-dimer 1.195 1.036∼1.379 0.014 
FDP 1.039 0.999∼1.081 0.059 
Fib 1.018 0.817∼1.267 0.874 
HCY 0.998 0.950∼1.047 0.920 
Platelet 0.995 0.991∼0.999 0.016 
Hemoglobin 1.004 0.985∼1.022 0.692 
Leukocyte count 1.167 1.069∼1.274 0.001 
NLR 1.142 1.073∼1.215 <0.001 
Infarct volume 1.023 1.016∼1.029 <0.001 
Mild BG-EPVS 1.000 
Moderate-to-severe BG-EPVS 1.711 1.029∼2.847 0.039 
Mild CS-EPVS 1.000 
Moderate-to-severe CS-EPVS 5.257 2.369∼11.667 <0.001 

MCE, malignant cerebral edema; LHI, large hemispheric infarction; AF, atrial fibrillation; NIHSS, National Institutes of Health Stroke Scale; GCS, Glasgow Coma Scale; TC, total serum cholesterol; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; HCY, serum homocysteine; FDP, fibrin degradation products; Fib, fibrinogen; NLR, neutrophil-to-lymphocyte ratio; BG, basal ganglia; CS, centrum semiovale; EPVS, enlarged perivascular spaces; OR, odds ratio; CI, confidence interval.

Table 4.

Multifactorial logistic regression affecting MCE in patients with LHI

Table 4.

Multifactorial logistic regression affecting MCE in patients with LHI

Close modal

ROC Curves of MCE-Related Risk Factors

ROC analysis was applied to assess the ability of EPVS individually and with other factors in predicting MCE after acute LHI. Based on the ROC curve, the predictive efficacy of CS-EPVS was better for MCE, with an area under the curve (AUC) of 0.632 and >14 CS-EPVS showed better predictive efficacy for MCE with a sensitivity of 0.82 and specificity of 0.48. When CS-EPVS were combined with other predictors of MCE (i.e., age, admission glucose, admission NIHSS score, NLR, infarct volume), the AUC reached 0.90, sensitivity 0.74, and specificity 0.90; while excluding CS-EPVS, the AUC of these other predictors was 0.867 (Fig. 2). Therefore, moderate-to-severe CS-EPVS were independent risk factor for the development of MCE and had higher predictive efficacy when evaluated in combination with other factors.

Fig. 2.

ROC curves of risk factors associated with MCE.

Fig. 2.

ROC curves of risk factors associated with MCE.

Close modal

The prognosis of patients with LHI caused by middle cerebral artery or internal carotid artery occlusion is poor. MCE is one of the main causes of death, with a mortality rate of approximately 80% [3]. Therefore, early prediction of MCE is necessary for the early selection of appropriate treatment options. In the present study, we investigated the value of EPVS in predicting the development of MCE in patients with acute LHI. We found that EPVS may be an independent risk factor for the development of MCE. EPVS in BG and CS were divided into mild and moderate-to-severe groups. We found that CS-EPVS were positively correlated with MCE, and the risk of MCE increased with the severity of CS-EPVS.

The glymphatic system is a novel defined brain-wide perivascular transport system between CSF and interstitial solutes that facilitates the clearance of brain metabolic wastes. Under physiological conditions, CSF from the subarachnoid space flows into the brain parenchyma through the periarterial space and is exchanged with interstitial fluid in the extracellular spaces, after which interstitial fluid and solutes move toward the perivenous space and eventually exit the brain through meningeal lymphatics. When acute LHI occurs, reduced tissue blood supply can cause spreading depolarizations of cortical neurons and astrocytes, resulting in EPVS and reduced clearance of the cerebral glymphatic system [16]. Mestre et al. [17] found in a cerebral ischemia rat model that CSF inflow along the PVS peaked at 11.4 ± 1.8 s and 5.24 ± 0.48 min after middle cerebral artery or internal carotid artery occlusion, respectively, suggesting that CSF in the cerebral glymphatic system is an important source of early edema fluid after ischemic stroke. As the disease progresses, the blood-brain barrier is damaged, metabolites accumulate in the vascular walls and PVS, and solute clearance along the glymphatic system is significantly impaired, leading to EPVS and accelerating the occurrence of edema [18]. Our subgroup study of EPVS found that CS-EPVS were positively correlated with MCE, and the risk of MCE increased with the severity of CS-EPVS. This may be related to the significantly reduced efflux rate of the cerebral glymphatic system after ischemic stroke [19], while CS region plays an important role as the primary perivenous efflux pathway of the glymphatic system [20‒22].

Consistent with previous work, we demonstrated that higher baseline NIHSS scores were a predictor of the development of MCE, suggesting that the development of MCE is associated with moderate-to-severe neurological deficits [2]. A study by Cheng [23] also showed that baseline NIHSS score was an early predictor of MCE, while the inclusion of NIHSS score in previous MCE prediction models showed better discriminatory ability and higher clinical application of the modified score. Therefore, higher NIHSS scores by patient admission can better predict the occurrence of MCE.

Previous studies have demonstrated that the inflammatory response is closely related to the progression of the course of acute cerebral infarction. When ischemic injury occurs, the expression of a large number of inflammatory-related factors in the ischemic region can lead to the development of neuronal damage and brain edema [24]. NLR is increasingly being studied as a new inflammatory marker in cerebrovascular disease. Bai et al. [25] found in a study of 257 patients with large infarcts that elevated NLR was associated with an increased risk of MCE (OR = 2.27, 95% CI: 1.11–4.62, p = 0.024) and that as the NLR ratio increased, the risk of MCE increased accordingly (p = 0.029). This shows that a higher NLR is significantly associated with a higher risk of MCE. The present study similarly concluded that NLR >5.22 was associated with the development of MCE after acute LHI. Thus, NLR, as an inflammatory factor, demonstrates that the development of MCE is associated with inflammatory responses and has a good predictive utility on the development of MCE after LHI.

Infarct volume is closely related to the malignant course, and DWI lesion volume (DHV) is considered an important predictor of MCE occurrence [26]. Its size is influenced by the time of MRI examination, and the earlier the DWI examination is performed, the smaller the DHV threshold is for predicting MCE. Our current study found that the MCE group’s infarct volume was larger than that of the non-MCE group, and the specificity and sensitivity were 0.78 and 0.74, respectively, when the infarct volume was >124 mL. The 2014 AHA/ASA [27] states that a DHV ≥80 mL on MRI-DWI within 6 h suggests the possibility of MCE.

Our present study also concluded that AF is an independent risk factor for MCE after acute LHI. This may be because AF is mostly caused by cardiac remodeling, and risk factors for cardiac remodeling, such as endothelin-1 and matrix metalloproteinases, are associated with brain edema [28].

Our current study has several limitations. First, this was a small sample, single-center study, which may lead to unavoidable bias; therefore, a larger sample size, multicenter study is needed to test this method of predicting MCE. Second, the present study was retrospective, and clinical data, such as collateral circulation and 90 d functional prognosis, was missing, and a large prospective study was warranted. Third, MRI examinations were performed after the onset of the disease, and the changes of EPVS before and after edema could not be dynamically observed. Finally, the EPVS rating assessment is inevitably subject to personal error, and more advanced quantitative assessment software is needed in the future to ensure more accurate results.

CS-EPVS may be independent risk factor of the development of MCE in patients with acute LHI. The risk of MCE increased with the severity of CS-EPVS, and they showed great predictive power when combined with other indicators to predict MCE. More prospective multicenter studies are needed in the future to confirm our findings, explore the joint predictive value of EPVS and other MCE predictors, and establish a brain edema prediction model suitable for our national population. This is of great significance for screening appropriate patients for precise intervention or treatment.

We would like to thank the Xianyang City Science and Technology Administration for funding support and Editage (www.editage.cn) for English language editing.

This study protocol was reviewed and approved by the Research Ethics Committee of Yan’an University Affiliated Xianyang Hospital (approval number: YDXY-KY-2021-026). Since this study was a retrospective cohort study, the requirement for individual informed consent was waived.

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

This work was supported by the Science and Technology Planning Project, Xianyang (L2023-ZDYF-SF-072).

Yaxin Wei and Shaojun Wang: study conception and design. Yaxin Wei, Jinhui Niu, and Rui Ma: data collection. Qingzi Zhang and Jian Miao: statistical analysis. Yaxin Wei: writing the manuscript. Shaojun Wang and Kang Huo: study supervision. The final manuscript was critically revised and approved by all authors.

1

Abbreviations: LHI = large hemisphere infarction, MCE = malignant cerebral edema, BG = basal ganglia, CS = centrum semiovale, EPVS = enlarged perivascular spaces, MRI, magnetic resonance imaging; DWI=diffusion-weighted imaging, FLAIR=fluid-attenuated inversion recovery, DHV = DWI lesion volume, ROC=receiver operating characteristic, AUC=area under the curve, CSF=cerebrospinal fluid, TR=repetition time, TE=echo time.

The data that support the findings of this study are not publicly available due to ethical reasons but are available from the corresponding author upon reasonable request.

1.
Liebeskind
DS
,
Jüttler
E
,
Shapovalov
Y
,
Yegin
A
,
Landen
J
,
Jauch
EC
.
Cerebral edema associated with large hemispheric infarction
.
Stroke
.
2019
;
50
(
9
):
2619
25
. .
2.
Wu
S
,
Yuan
R
,
Wang
Y
,
Wei
C
,
Zhang
S
,
Yang
X
, et al
.
Early prediction of malignant brain edema after ischemic stroke
.
Stroke
.
2018
;
49
(
12
):
2918
27
. .
3.
Wei
H
,
Jia
F-M
,
Yin
H-X
,
Guo
Z-L
.
Decompressive hemicraniectomy versus medical treatment of malignant middle cerebral artery infarction: a systematic review and meta-analysis
.
Biosci Rep
.
2020
;
40
(
1
):
BSR20191448
. .
4.
Beez
T
,
Munoz-Bendix
C
,
Steiger
H-J
,
Beseoglu
K
.
Decompressive craniectomy for acute ischemic stroke
.
Crit Care
.
2019
;
23
(
1
):
209
. .
5.
Chen
S
,
Shao
L
,
Ma
L
.
Cerebral edema formation after stroke: emphasis on blood-brain barrier and the lymphatic drainage system of the brain
.
Front Cell Neurosci
.
2021
;
15
:
716825
. .
6.
Iliff
JJ
,
Wang
M
,
Liao
Y
,
Plogg
BA
,
Peng
W
,
Gundersen
GA
, et al
.
A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β
.
Sci Transl Med
.
2012
;
4
(
147
):
147ra111
. .
7.
Wang
X
,
Feng
H
,
Wang
Y
,
Zhou
J
,
Zhao
X
.
Enlarged perivascular spaces and cerebral small vessel disease in spontaneous intracerebral hemorrhage patients
.
Front Neurol
.
2019
;
10
:
881
. .
8.
Doubal
FN
,
Maclullich
AMJ
,
Ferguson
KJ
,
Dennis
MS
,
Wardlaw
JM
.
Enlarged perivascular spaces on MRI are a feature of cerebral small vessel disease
.
Stroke
.
2010
;
41
(
3
):
450
4
. .
9.
Gouveia-Freitas
K
,
Bastos-Leite
AJ
.
Perivascular spaces and brain waste clearance systems: relevance for neurodegenerative and cerebrovascular pathology
.
Neuroradiology
.
2021
;
63
(
10
):
1581
97
. .
10.
Wardlaw
JM
,
Benveniste
H
,
Nedergaard
M
,
Zlokovic
BV
,
Mestre
H
,
Lee
H
, et al
.
Perivascular spaces in the brain: anatomy, physiology and pathology
.
Nat Rev Neurol
.
2020
;
16
(
3
):
137
53
. .
11.
Song
Q
,
Cheng
Y
,
Wang
Y
,
Liu
J
,
Wei
C
,
Liu
M
.
Enlarged perivascular spaces and hemorrhagic transformation after acute ischemic stroke
.
Ann Transl Med
.
2021
;
9
(
14
):
1126
. .
12.
Wu
B
,
Yao
X
,
Lei
C
,
Liu
M
,
Selim
MH
.
Enlarged perivascular spaces and small diffusion-weighted lesions in intracerebral hemorrhage
.
Neurology
.
2015
;
85
(
23
):
2045
52
. .
13.
Pullicino
P
,
Snyder
W
,
Munschauer
F
,
Pordell
R
,
Greiner
F
.
Interrater agreement of computed tomography infarct measurement
.
J Neuroimaging
.
1996
;
6
(
1
):
16
9
. .
14.
Huang
X
,
Yang
Q
,
Shi
X
,
Xu
X
,
Ge
L
,
Ding
X
, et al
.
Predictors of malignant brain edema after mechanical thrombectomy for acute ischemic stroke
.
J Neurointerv Surg
.
2019
;
11
(
10
):
994
8
. .
15.
Kim
H
,
Jin
ST
,
Kim
YW
,
Kim
SR
,
Park
IS
,
Jo
KW
.
Predictors of malignant brain edema in middle cerebral artery infarction observed on CT angiography
.
J Clin Neurosci
.
2015
;
22
(
3
):
554
60
. .
16.
Lv
T
,
Zhao
B
,
Hu
Q
,
Zhang
X
.
The glymphatic system: a novel therapeutic target for stroke treatment
.
Front Aging Neurosci
.
2021
;
13
:
689098
. .
17.
Mestre
H
,
Du
T
,
Sweeney
A
,
Liu
G
,
Samson
A
,
Peng
W
, et al
.
Cerebrospinal fluid influx drives acute ischemic tissue swelling
.
Science
.
2020
;
367
(
6483
):
eaax7171
. .
18.
Zhou
X
,
Li
Y
,
Lenahan
C
,
Ou
Y
,
Wang
M
,
He
Y
.
Glymphatic system in the central nervous system, a novel therapeutic direction against brain edema after stroke
.
Front Aging Neurosci
.
2021
;
13
:
698036
. .
19.
Gaberel
T
,
Gakuba
C
,
Goulay
R
,
Martinez De Lizarrondo
S
,
Hanouz
J-L
,
Emery
E
, et al
.
Impaired glymphatic perfusion after strokes revealed by contrast-enhanced MRI: a new target for fibrinolysis
.
Stroke
.
2014
;
45
(
10
):
3092
6
. .
20.
Jessen
NA
,
Munk
ASF
,
Lundgaard
I
,
Nedergaard
M
.
The glymphatic system: a beginner’s guide
.
Neurochem Res
.
2015
;
40
(
12
):
2583
99
. .
21.
Zhang
K
,
Zhou
Y
,
Zhang
W
,
Li
Q
,
Sun
J
,
Lou
M
.
MRI-visible perivascular spaces in basal ganglia but not centrum semiovale or hippocampus were related to deep medullary veins changes
.
J Cereb Blood Flow Metab
.
2022
;
42
(
1
):
136
44
. .
22.
Yamada
S
,
Ishikawa
M
,
Yamamoto
K
,
Yamaguchi
M
,
Oshima
M
.
Location-specific characteristics of perivascular spaces as the brain’s interstitial fluid drainage system
.
J Neurol Sci
.
2019
;
398
:
9
15
. .
23.
Cheng
Y
,
Wu
S
,
Wang
Y
,
Song
Q
,
Yuan
R
,
Wu
Q
, et al
.
External validation and modification of the EDEMA score for predicting malignant brain edema after acute ischemic stroke
.
Neurocrit Care
.
2020
;
32
(
1
):
104
12
. .
24.
Li
S
,
Bian
L
,
Fu
X
,
Ai
Q
,
Sui
Y
,
Zhang
A
, et al
.
Gastrodin pretreatment alleviates rat brain injury caused by cerebral ischemic-reperfusion
.
Brain Res
.
2019
;
1712
:
207
16
. .
25.
Bai
X
,
Wang
C
,
Wang
L
,
Jiang
S
,
Zhang
S
,
Liu
M
, et al
.
Association between neutrophil to lymphocyte ratio and malignant brain edema in patients with large hemispheric infarction
.
Curr Neurovasc Res
.
2020
;
17
(
4
):
429
36
. .
26.
Goto
Y
,
Kumura
E
,
Watabe
T
,
Nakamura
H
,
Nishino
A
,
Koyama
T
, et al
.
Prediction of malignant middle cerebral artery infarction in elderly patients
.
J Stroke Cerebrovasc Dis
.
2016
;
25
(
6
):
1389
95
. .
27.
Wijdicks
EF
,
Sheth
KN
,
Carter
BS
,
Greer
DM
,
Kasner
SE
,
Kimberly
WT
, et al
.
Recommendations for the management of cerebral and cerebellar infarction with swelling: a statement for healthcare professionals from the American Heart Association/American Stroke Association
.
Stroke
.
2014
;
45
(
4
):
1222
38
. .
28.
Nakazawa
Y
,
Ashihara
T
,
Tsutamoto
T
,
Ito
M
,
Horie
M
.
Endothelin-1 as a predictor of atrial fibrillation recurrence after pulmonary vein isolation
.
Heart Rhythm
.
2009
;
6
(
6
):
725
30
. .