Risk Factors Analysis and Prediction Model for Malignant Cerebral Edema Following Mechanical Thrombectomy in Patients With Large Vessel Occlusive Stroke

Abstract

Background

To investigate the risk factors for malignant cerebral edema (MCE) following mechanical thrombectomy (MT) in patients with large vessel occlusion (LVO) stroke, construct a risk prediction model, and provide evidence for early intervention.

Methods

A retrospective study was conducted on 248 patients with LVO who underwent MT between January 2021 to December 2024 (60 cases in the MCE group and 188 cases in the non-MCE group). Independent predictive factors were identified through univariate analysis, LASSO regression, and multivariate logistic regression. A nomogram model was constructed and validated internally using the Bootstrap method (1,000 resampling).

Results

Independent predictive factors for MCE after MT in LVO patients were identified, including high baseline NIHSS score (OR = 1.805), infarct core volume (OR = 1.235), good collateral circulation was a protective factor (OR = 0.287, indicating that each one-point increase in the collateral score reduced the risk of MCE by 71.3%), high (neutrophil-to-lymphocyte ratio) NLR (OR = 5.312), and hyperglycemia (OR = 15.445) were significant risk factors, while successful reperfusion (mTICI ≥ 2b grade, OR = 0.068) was a key protective factor. The nomogram prediction model constructed based on the above six factors demonstrated excellent discriminative ability (AUC = 0.902, 95%CI: 0.844 - 0.959) and good calibration (Hosmer-Lemeshow test, P = 0.906).

Conclusion

NIHSS score, infarct core volume, collateral circulation, NLR, and blood glucose are independent risk factors for MCE after MT, while successful reperfusion is a key protective factor. The constructed nomogram model can effectively identify high-risk patients and guide individualized interventions.

Keywords

large vessel occlusive stroke mechanical thrombectomy malignant cerebral edema nomogram

Introduction

Large vessel occlusion stroke (LVO) is the most severe subtype of acute ischemic stroke, accounting for approximately 24% to 46% of all acute ischemic strokes.¹ Malhotra et al.² reported that although LVO patients account for only 38.7% of all acute ischemic strokes, they represent 61.6% of stroke-related deaths or dependency, indicating that LVO has become a major global public health challenge. In recent years, mechanical thrombectomy (MT) has emerged as a key intervention for acute ischemic stroke caused by occlusion of large vessels in the anterior circulation, thanks to its ability to rapidly reopen occluded vessels, significantly improving functional outcomes.³ Although MT can rapidly restore cerebral blood flow, 10%-20% of patients develop malignant cerebral edema (MCE) postoperatively.⁴ This severe complication manifests as progressive intracranial pressure elevation, midline shift, and brain herniation, and is the primary cause of early mortality (24-72 hours) following MT. Even if vascular recanalization is successful, MCE may still result in irreversible neurological dysfunction.

Currently, the mechanisms underlying MCE remain unclear, and effective predictive tools are lacking. This makes it challenging for clinicians to identify high-risk patients in advance when treating LVO patients, thereby preventing timely implementation of targeted preventive and interventional measures.⁵ Therefore, analyzing the risk factors for malignant cerebral edema following mechanical thrombectomy and constructing a risk prediction model are of utmost importance for early identification of high-risk patients, optimization of treatment strategies, and improvement of patient outcomes.

This study aims to systematically analyze clinical data, identify relevant risk factors, and construct a predictive model to provide scientific evidence for clinical decision-making. We will collect a large amount of clinical data from LVO patients, including age, gender, underlying conditions (such as hypertension, diabetes, hyperlipidemia, etc.), onset time, location of vascular occlusion, pre-thrombectomy neurological function scores (such as NIHSS scores), and various parameters during thrombectomy (such as the number of thrombectomies, thrombectomy time, etc.). Through in-depth analysis of these data, we will identify risk factors associated with MCE occurrence and use statistical methods to construct a risk prediction model. This model will assist clinicians in assessing patient risk preoperatively or in the early postoperative period, enabling them to develop personalized treatment plans in advance, such as enhancing intracranial pressure monitoring, optimizing drug therapy, or promptly implementing surgical interventions. Ultimately, we hope that this study will provide more precise guidance for the treatment of LVO patients, reduce the incidence of MCE, improve patient outcomes, and alleviate the burden on society and families.

Methods

Patient Selection

Retrospectively collect data on patients with acute ischemic stroke admitted to the hospital from January 2021 to December 2024, and divide them into an MCE group and a non-MCE group based on whether MCE occurred after MT surgery. Inclusion criteria: (1) Meet the criteria of the “Chinese Guidelines for the Diagnosis and Treatment of Acute Ischemic Stroke 2018” and have neurological deficits; (2) Age ≥ 18 years; (3) Preoperative imaging studies (MRA/CTA/DSA) confirmed occlusion of the internal carotid artery or middle cerebral artery; (4) Both the patient and their family members signed informed consent forms. Exclusion criteria: (1) Multiple vessel occlusions, intracranial hemorrhage, or subarachnoid hemorrhage; (2) Baseline Alberta Stroke Project Early CT (ASPECT) score < 6 points, or baseline National Institutes of Health Stroke Scale (NIHSS) score < 6 points; (3) Severe cardiac, pulmonary, hepatic, or renal dysfunction; (4) Severe allergy to the contrast agent; (5) Severe allergy to the contrast agent. (6) Brain herniation present at admission or early signs of cerebral edema (such as midline shift >5mm with effacement of sulci/cisterns). This study was approved by the Ethics Committee of The First Hospital of Tsinghua University (approval number: (R) 2025-050-02).

MCE Diagnostic Criteria

(1) Neurological deterioration(an increase of ≥ 1 point in the NIHSS consciousness assessment section or an increase of ≥ 2 points in the NIHSS score). (2) Imaging findings: > 50% low-density lesions in the middle cerebral artery territory accompanied by signs of cerebral edema (compression of the lateral ventricles/disappearance of sulci); midline shift > 5 mm (at the level of the septum pellucidum/pineal gland) accompanied by disappearance of the interpeduncular cistern.

Data Collection

The study retrospectively collected patients’ clinical data, including: ① demographic characteristics (age, gender); ② risk factors (smoking history, drinking history) and underlying diseases (diabetes, hypertension, coronary heart disease, atrial fibrillation, history of stroke); ③ clinical parameters: systolic blood pressure, diastolic blood pressure; ④ site of vascular occlusion: determined by MRA/CTA/DSA (internal carotid artery, middle cerebral artery M1 segment, M2 segment and beyond); ⑤ laboratory indicators (blood glucose, CRP, fibrinogen, NLR); ⑥ Clinical scores [NIHSS, baseline Glasgow Coma Scale (GCS) score, ASPECT score]; ⑦Imaging parameters (MRA/CTA/DSA to assess vascular status, DWI/CTP to assess tissue injury, including infarct core volume and collateral circulation status); ⑧ Treatment-related indicators [TOAST classification, time from onset to puncture/reperfusion, thrombectomy method and number of procedures, modified thrombolysis in cerebral infarction (mTICI) grading and reperfusion status, perioperative complications]. Collateral circulation was scored using the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology (ASITN/SIR) collateral flow grading system.

Related Definitions

Successful recanalization: mTICI ≥ 2b grade on angiography after thrombectomy; time from onset to recanalization: the time from onset to the first achievement of mTICI ≥ 2b grade; perioperative complications: including vascular perforation, embolism, arterial dissection, and perforating artery injury.

Statistical Analysis

Statistical analysis was performed using SPSS 27.0 and R 4.4.3. Count data are expressed as n (%) and analyzed using the chi-square test; continuous data with a normal distribution are expressed as mean ± standard deviation (t-test), while those with a non-normal distribution are expressed as median (interquartile range) (Mann-Whitney U test). Variables were screened using univariate Logistic regression and LASSO regression, and multivariate logistic regression was used to identify independent predictors of MCE and construct a nomogram model. Internal validation was performed using the Bootstrap method (1,000 resamples), and model performance was assessed using ROC curves, calibration curves, and the Hosmer-Lemeshow test. The significance level was set at α = 0.05. No prior sample size calculation was performed, but it is noted that the number of MCE events (n=60) relative to the number of candidate predictor variables carries a risk of overfitting, which is acknowledged as a limitation in the Discussion.

Results

Patient Characteristics

According to the inclusion and exclusion criteria, a total of 248 patients were finally included in the study, including 60 patients with MCE (MCE group) and 188 patients without MCE (non-MCE group). The clinical conditions of the two groups are shown in Table 1.

Table 1.

Comparison of Clinical Data Between the Two Groups of Patients

	Non-MCE group(n=188)	MCE group(n=60)	t/z/x ²	P
Age (years)	63.03±9.04	63.62±9.42	-0.432	0.666
Gender[n(%)]			3.111	0.078
Male	106(56.38)	26(43.33)
Female	82(43.62)	34(56.67)
Underlying medical conditions or past medical history[n(%)]
Diabetes	12(6.38)	21(35.00)	-5.682	<0.001
Hypertension	41(21.81)	10(16.67)	0.736	0.391
Coronary heart disease	14(7.45)	11(18.33)	5.947	0.015
History of stroke	36(19.15)	13(21.67)	0.182	0.670
Atrial fibrillation	26(13.83)	5(8.33)	1.256	0.262
Smoking history[n(%)]	38(20.21)	16(26.67)	1.112	0.292
Drinking historyn(%)]	36(19.15)	10(16.67)	0.186	0.667
Systolic blood pressure(mmHg)	153.27±10.83	156.00±13.02	-1.618	0.714
Diastolic blood pressure(mmHg)	77.19±7.01	77.38±6.01	-0.212	0.107
Baseline NIHSS score[M(Q1,Q3)]	13.0(12.0,15.0)	15.0(14.0,17.0)	-4.976	<0.001
Baseline GCS score[M(Q1,Q3)]	10.0(10.0,11.0)	10.0(9.0,11.0)	-0.033	0.973
Preoperative ASPECTS score[M(Q1,Q3)]	8.0(8.0,8.0)	7.0(6.0,8.0)	-5.021	<0.001
Infarction core volume(ml)	24.07±6.40	31.20±3.49	-10.988	<0.001
TOAST Etiological Classification[n(%)]			3.644	0.162
Atherosclerosis of the large arteries	104(55.32)	35(58.33)
Cardioembolic	76(40.43)	19(31.67)
Other etiologies	8(4.26)	6(10.00)
Vascular occlusion site[n(%)]			7.312	0.026
Internal carotid artery	90(47.87)	18(30.00)
Middle cerebral artery M1 segment	90(47.87)	36(60.00)
Middle cerebral artery segment M2 and beyond	8(4.26)	6(10.00)
Collateral circulation score[M(Q1,Q3)]	2.0(1.0,2.0)	1.0(1.0,2.0)	-4.840	<0.001
Thrombolysis[n(%)]	50(26.60)	17(28.33)	0.070	0.792
Mechanical thrombectomy[n(%)]			4.499	0.105
Stent thrombectomy	128(68.09)	32(53.33)
Thrombectomy	24(12.77)	10(16.67)
Stent-assisted thrombectomy	36(19.15)	18(30.00)
Number of thrombectomy procedures ≥ 3 [n(%)]	21(11.17)	24(40.00)	25.43	<0.001
Time from onset to femoral artery puncture(min)	261.90±72.25	272.63±64.52	-1.027	0.306
Time from onset to recanalization(min)	327.46±82.26	345.25±91.78	-1.418	0.158
mTIC I≥ 2b[n(%)]	175(93.09)	37(61.67)	36.182	<0.001
Vessel recanalization successful[n(%)]	177(94.15)	48(80.00)	10.822	0.001
Perioperative complications[n(%)]	12(6.38)	5(8.33)	0.271	0.603
Glucose(mmol/L)	5.54±0.45	8.30±1.86	-11.389	<0.001
CPR(mg/L)	2.77±0.59	3.31±0.38	-8.220	<0.001
Fibrinogen(g/L)	3.10±0.68	3.23±0.36	-1.957	0.052
D-dimer(mg/L)	1.17±0.30	1.25±0.33	-1.784	0.076
NLR	2.64±0.78	6.44±0.75	-33.252	<0.001

Note. mTICI ≥ 2b indicates the reperfusion grade assessed by postoperative angiography; successful vessel recanalization was defined as the first achievement of mTICI ≥ 2b sustained until the end of the procedure, treated as a binary variable. The definit ions differ slightly, hence the discrepancy in data.

Variable Selection

Using the occurrence of MCE as the dependent variable, variables with P < 0.1 in the univariate logistic regression analysis were selected as independent variables for Lasso regression variable selection (Figure 1). Variable values are shown in Table 2. The non-zero variables selected by Lasso regression analysis were included in the multivariate logistic regression analysis, ultimately identifying six independent predictors of MCE: baseline NIHSS score, infarct core volume, collateral circulation score, mTICI grade ≥ 2b, NLR, and blood glucose. Among these, mTICI grade ≥ 2b and good collateral circulation are protective factors, while the remaining factors are risk factors. The results of the multivariate logistic regression analysis are shown in Table 3.

Figure 1.

Coefficient profiles of lasso regression screening for predicting IAPI influences

Table 2.

Assignment of Independent Variables

Independent variable	Variable description	Assignment
mTICI≥2b	Categorical variable	YES=1, NO=0
Baseline NIHSS score	Continuous variables	Original value input
Infarction core volume	Continuous variables	Original value input
Collateral circulation score	Continuous variables	Original value input
Glucose	Continuous variables	Original value input
NLR	Continuous variables	Original value input

Table 3.

Multivariate Logistic Regression Analysis

	β	SE	z	Wald χ ²	P	OR(95% CI)
Baseline NIHSS score	0.590	0.196	3.006	9.038	0.003	1.805(1.228 ∼ 2.652)
Collateral circulation score	-1.248	0.627	-1.992	3.969	0.046	0.287(0.084 ∼ 0.980)
Glucose	2.737	0.681	4.022	16.180	0.000	15.445(4.069 ∼ 8.618)
Infarction core volume	0.211	0.076	2.773	7.688	0.006	1.235(1.064 ∼ 1.433)
mTICI ≥ 2b	-2.683	0.872	-3.077	9.469	0.002	0.068(0.012 ∼ 0.377)
NLR	1.670	0.731	2.284	5.216	0.022	5.312(1.267 ∼ 22.265)
Constant	-34.268	7.253	-4.725	22.324	0.000

Construction and Validation of the Regression Plot Model for MCE Prediction

Based on the results of multi-factor logistic regression analysis, a nomogram prediction model for MCE risk was constructed (Figure 2). The ROC curve results showed an AUC of 0.902 (95% CI: 0.844 - 0.959) (Figure 3). Calibration curve analysis indicated that the prediction curve of the regression model closely aligned with the ideal curve, and the predicted values were consistent with the observed values, indicating that the model has high predictive accuracy (Figure 4). Through the Hosmer-Lemeshow goodness-of-fit test, there was no significant difference between the predicted results of the nomogram model and the actual occurrence (χ² = 3.405, P = 0.906), confirming the model’s good calibration performance.

Figure 2.

Linear model of MCE occurrence after MT surgery in LVO patients

Figure 3.

ROC curve for MCE occurrence after MT surgery in LVO patients

Figure 4.

Calibration curve for MCE occurrence after MT surgery in LVO patients

Discussion

This study identified six predictive factors influencing the occurrence of NCE in LVO patients (all P < 0.05): baseline NIHSS score, infarct core volume, collateral circulation score, mTICI grade ≥ 2b, NLR, and blood glucose. Among these, mTICI grade ≥ 2b is a protective factor, while the remaining factors are risk factors.

A high NIHSS score (especially ≥15 points) is an important indicator for predicting MCE, reflecting severe neurological damage caused by extensive infarction or involvement of functional areas.⁶ An increased score suggests neuronal necrosis in the ischemic core, disruption of the blood-brain barrier, and worsening edema. Multiple studies have confirmed its significant association with the risk of malignant cerebral edema. When the NIHSS score exceeds 18 points, the predictive value becomes more pronounced.^7,8 Chinese studies have shown that incorporating NIHSS scores >15 into a predictive model for anterior circulation infarction can improve efficacy.⁹ This study demonstrates that the NIHSS score at admission has predictive value for malignant cerebral edema and is a key indicator for assessing disease progression.

Infarction core volume is quantified through imaging (e.g., CTP or MRI-DWI) and directly reflects the extent of irreversible ischemic brain tissue.¹⁰ Studies show that patients with core volumes >70 ml have a malignant cerebral edema risk as high as 68% after endovascular therapy. Large core infarcts cause an imbalance in transcapillary hydrostatic pressure gradients, leading to a sharp increase in blood-brain barrier permeability and plasma component extravasation, resulting in significant edema. Additionally, neuronal death in the core releases inflammatory factors (such as IL-1β and TNF-α), further activating microglia and exacerbating edema.^11,12 Therefore, in clinical assessment, the size of the infarct core volume is crucial for determining whether a patient is suitable for endovascular therapy and predicting potential postoperative complications. Physicians should comprehensively consider the infarct core volume alongside other factors to develop individualized treatment plans.

The status of collateral circulation is an independent predictor of MCE following reperfusion in acute large-vessel occlusive stroke. Good collateral circulation maintains blood flow in the penumbra, reduces neuronal apoptosis, and prevents blood-brain barrier disruption,¹³ while poor collateral circulation accelerates infarct progression and increases the risk of MCE.¹⁴ Therefore, even after successful recanalization, patients with poor collateral circulation should remain vigilant for MCE and prepare for early decompression.

Timely reperfusion (≤6 hours) can restore energy metabolism, replenish ATP supply, reduce the accumulation of anaerobic glycolysis products (lactic acid), and thereby alleviate cytotoxic edema.^15,16 Additionally, mTICI 2b/3-grade reperfusion protects the blood-brain barrier (BBB), downregulates matrix metalloproteinase (MMP-9) expression, and reduces the degradation of tight junction proteins (occludin/ZO-1).¹⁶ Animal models have demonstrated that BBB permeability (Evans Blue leakage) can be reduced by 60% following reperfusion.¹⁶ Furthermore, successful reperfusion can regulate the inflammatory response, with a 40% decrease in the NLR, inhibit elastase and reactive oxygen species-induced damage to vascular endothelium, and block cytokine cascade reactions, resulting in a 50% reduction in serum TNF-α and IL-1β levels, thereby alleviating neuroinflammatory-driven vascular edema.¹⁷ An mTICI grade of ≥ 2b indicates successful reperfusion. The HERMES meta-analysis showed that patients with mTICI 2b/3 grades had a 55% lower risk of malignant cerebral edema compared to those with 0 - 2a grades.³ This study shows that successful reperfusion (≥ mTICI 2b) is the most significant modifiable protective factor for malignant cerebral edema. Clinically, the reperfusion time window can be managed with precision, shortened to < 6 hours for patients with poor collateral circulation and cautiously considered for appropriate extension in those with good collateral circulation,¹⁶ although the data from this study are insufficient to support a specific recommendation of extension to 24 hours, which requires further research for confirmation.The limitations of this study include its retrospective single-center design, unvalidated cutoff values for some parameters, lack of in-depth exploration into biomarker mechanisms, and the potential risk of overfitting due to the limited sample size.

Following acute ischemic stroke, inflammatory responses and immune imbalance jointly contribute to blood-brain barrier disruption, promoting brain edema formation. Neutrophil infiltration releases MMP-9, exacerbating blood-brain barrier damage, while lymphocytes exert a protective effect.^18,19 The neutrophil-to-lymphocyte ratio (NLR) serves as a comprehensive inflammatory marker that better reflects this pathological process. Studies have shown that an NLR > 7 (AUC = 0.7) has predictive value for prognosis.^20,21 This study found that the preoperative NLR was significantly elevated in the MCE group, suggesting that systemic inflammatory response is associated with malignant cerebral edema, and that NLR can serve as a predictive indicator for cerebral edema following endovascular therapy.

Hyperglycemia increases intracellular osmotic pressure through the polyol pathway while activating protein kinase C to disrupt vascular endothelial adhesion molecules and impair the blood-brain barrier.²² Clinical studies have also shown that patients with a history of diabetes are more prone to inflammatory cascades following ischemia-reperfusion injury.²² Therefore, regardless of whether patients have a history of diabetes, acute hyperglycemia (typically defined as blood glucose levels > 7.8 -10.0 mmol/L) is an independent risk factor for MCE following MT surgery. Strict perioperative blood glucose control is an important intervenable measure to reduce the risk of MCE.

Conclusion

The findings of this study indicate that baseline NIHSS scores, large infarct core volume, poor collateral circulation, high NLR, and hyperglycemia are independent risk factors for MCE after MT in LVO patients, while successful reperfusion (mTICI ≥ 2b grade) is a key protective factor. The results of this study emphasize the need for comprehensive assessment of the aforementioned indicators at admission and implementation of individualized interventions for high-risk patients: optimizing the reperfusion time window (collateral-dependent management), strictly controlling blood glucose levels, dynamically monitoring inflammatory markers, and preparing for decompressive craniectomy. The limitations of this study include its retrospective single-center design, unvalidated cutoff values for some parameters, and the absence of exploration into biomarker mechanisms. Future research should establish a weighted predictive model through a multicenter prospective cohort study and integrate radiomics with molecular biomarkers to deepen understanding of the pathophysiological mechanisms.

Footnotes

ORCID iD

Yujuan Jiao

Ethical Considerations

The experiment was approved by the ethics committee of The First Hospital of Tsinghua University(approval number: (R) 2025-050-02). All patients signed informed consent. All methods were carried out in accordance with Declaration of Helsinki.

Author Contributions

GJ Z, D W: conceptualization, methodology and wrote the original draft.

F C, YN L: performed the experiments, the data validation and analyse.

YJ J, L W: edited and reviewed the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This Study is supported by First Hospital of Tsinghua University Navigation Fund (NO. 2024-LH-12) and China International Medical Foundation (NO. Z-2016-20-2101).

Declaration of Conflicting Interests

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

Data Availability Statement

The data analyzed are available from the corresponding authors upon reasonable request.*

References

Rennert

Wali

Steinberg

, et al. Epidemiology, Natural History, and Clinical Presentation of Large Vessel Ischemic Stroke. Neurosurgery. 2019;85(suppl_1):S4-S8. doi:10.1093/neuros/nyz042.

Malhotra

Gornbein

Saver

. Ischemic Strokes Due to Large-Vessel Occlusions Contribute Disproportionately to Stroke-Related Dependence and Death: A Review. Front Neurol. 2017;8:651. doi:10.3389/fneur.2017.00651.

Goyal

Menon

van Zwam

, et al. HERMES collaborators. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet. 2016;387(10029):1723-1731. doi:10.1016/S0140-6736(16)00163-X.

Thorén

Escudero-Martínez

Andersson

, et al. Reperfusion by endovascular thrombectomy and early cerebral edema in anterior circulation stroke: Results from the SITS-International Stroke Thrombectomy Registry. Int J Stroke. 2023;18(10):1193-1201. doi:10.1177/17474930231180451.

Liu

Yang

, et al. Prediction models for post-thrombectomy brain edema in patients with acute ischemic stroke: a systematic review and meta-analysis. Front Neurol. 2023;14:1254090. doi:10.3389/fneur.2023.1254090.

Bajgur

Kim

, et al. A simple prediction score system for malignant brain edema progression in large hemispheric infarction. PLoS One. 2017;12(2):e0171425. doi:10.1371/journal.pone.0171425.

Wei

Zhang

Niu

, et al. Enlarged perivascular spaces predict malignant cerebral edema after acute large hemispheric infarction. Cerebrovasc Dis Extra. 2024;14(1):30-38. doi:10.1159/000536592.

Kim

Jin

Kim

, et al. Predictors of malignant brain edema in middle cerebral artery infarction observed on CT angiography. J ClinNeurosci. 2015;22(3):554-560. doi:10.1016/j.jocn.2014.08.021.

Cheng

Wang

, et al. External validation and modification of the EDEMA score for predicting malignant brain edema after acute ischemic stroke. Neurocrit Care. 2020;32(1):104-112. doi:10.1007/s12028-019-00844-y.

10.

Cheripelli

Huang

Maclsaac

, et al. Interaction of recanali. zation, intracerebral hemorrhage, and cerebral edema after intravenous thrombolysis. Stroke. 2016;47(7):1761-1767. doi:10.1161/STROKEAHA.116.013142.

11.

Chen

Fang

Chen

, et al. Microglia-derived TNF-α mediates endothelial necroptosis aggravating blood brain-barrier disruption after ischemic stroke. Cell Death Dis. 2019;10(7):487. doi:10.1038/s41419-019-1716-9.

12.

Wang

Zhang

Sheng

, et al. Immunoreactive Cells After Cerebral Ischemia. Front Immunol. 2019;10:2781. doi:10.3389/fimmu.2019.02781

13.

Xiong

Song

Huang

, et al. Exosomes Derived from Mesenchymal Stem Cells: Novel Effects in the Treatment of Ischemic Stroke. Front Neurosci. 2022;16:899887. doi:10.3389/fnins.2022.899887.

14.

Ginsberg

. Expanding the concept of neuroprotection for acute ischemic stroke: The pivotal roles of reperfusion and the collateral circulation. Prog Neurobiol. 2016;145-146:46-77. doi:10.1016/j.pneurobio.2016.09.002.

15.

Sacks

Baxter

Campbell

BCV

, et al. Multisociety Consensus Quality Improvement Revised Consensus Statement for Endovascular Therapy of Acute Ischemic Stroke. Int J Stroke. 2018;13(6):612-632. doi:10.1177/1747493018778713.

16.

Campbell

Mitchell

Kleinig

, et al. Endovascular therapy for ischemic stroke with perfusion-imaging selection. N Engl J Med. 2015;372(11):1009-1018. doi:10.1056/NEJMoa1414792.

17.

Yuan

Wang

, et al. Early Prediction of Malignant Brain Edema After Ischemic Stroke. Stroke. 2018;49(12):2918-2927. doi:10.1161/STROKEAHA.118.022001.

18.

Gautier

Ouk

Petrault

, et al. Neutrophils contribute to intracerebral haemorrhages after treatment with recombinant tissue plasminogen activator following cerebral ischaemia. Br J Pharmacol. 2009;156(4):673-679. doi:10.1111/j.1476-5381.2009.00068.x.

19.

Yang

Hawkins

Dore

, et al. Neuroinflammatory mechanisms of blood-brain barrier damage in ischemic stroke. Am J Physiol Cell Physiol. 2019;316(2):C135-C153. doi:10.1152/ajpcell.00136.2018.

20.

Petrone

Eisenman

Steele

, et al. Temporaldynamics of peripheral neutrophil and lymphocytes following acute ischemic stroke. Neurol Sci. 2019;40(9):1877-1885. doi:10.1007/s10072-019-03919-y.

21.

Ferro

Matias

Neto

, et al. Neutrophil‐to‐lymphocyteratio predicts cerebral edema and clinical worsening early after reperfusion therapy in stroke. Stroke. 2021;52(3):859-867. doi:10.1161/STROKEAHA.120.032130.

22.

Shao

Bayraktutan

. Hyperglycaemia promotes human brain microvascular endothelial cell apoptosis via induction of protein kinase C-ßI and prooxidant enzyme NADPH oxidase. Redox Biol. 2014;2:694-701. doi:10.1016/j.redox.2014.05.005.