Early neutrophil count relates to infarct size and fatal outcome after large hemispheric infarction

Abstract Aims To investigate the relationship between peripheral leukocyte dynamics and the outcome of large hemispheric infarction (LHI) patients. Methods Patients with acute LHI admitted to the neuro‐intensive care unit of Xuanwu Hospital from 2013 to 2017 were prospectively enrolled and followed up for 6 months after LHI. Results A total of 84 LHI patients were included, 38 patients suffered brain herniation and 20 patients died from stroke. Compared to patients with benign course, LHI patients with fatal outcome showed larger infarcts and more severe brain edema (P < .01), as well as increased WBC and neutrophil counts throughout the first week after stroke (P < .05). Correlation analysis revealed that neutrophil counts on D2 after LHI positively correlated with infarct and edema volumes measured from CT/MRI (R2 = 0.22 and R 2 = 0.15, P < .01) and negatively correlated with Glasgow Coma Scale (ρ = −0.234, P < .05). Patients with D2 neutrophils > 7.14 × 109/L had higher risk of brain herniation [odds ratio (OR) = 7.5, 95% CI: 2.0‐28.1, P = .001], and patients with D2 neutrophils > 7.79 × 109/L had a higher risk of death (OR = 5.8, 95% CI: 1.2‐27.0, P = .015). Conclusion Early peripheral neutrophil count after stroke relates to infarct size and the fatal outcome of LHI patients, which might help guiding acute LHI management such as reduction of intracranial pressure and potential antiinflammatory therapy in the future.


| INTRODUC TI ON
Large hemispheric infarction (LHI) accounts for approximately 10% to 15% of ischemic stroke events and is associated with high mortality and poor outcome. 1 Malignant brain edema is frequently seen in LHI, which could cause brain herniation and potentially death. 1,2 Brain herniation can hardly be reversed by medication alone, and timely decompressive craniectomy is often needed to reduce mortality. 3,4 It is thus important to identify indicators of malignant brain edema and potential brain herniation to prevent severe neurological dysfunction and death by early intervention.
Leukocytes, particularly neutrophils, have been reported to play an important role in experimental stroke. [5][6][7] However, the evidence in stroke patients is still limited. 8 Moreover, few studies have investigated the dynamics of leukocyte after LHI and its relationship with infarct and edema volumes, or fatal outcomes including brain herniation and death. Therefore, we aimed to investigate the potential relationship between leukocyte changes of LHI patient within one week after symptom onset and the outcome including brain herniation and death.

| ME THODS
This observational, prospective study was approved by the Ethics Committee of Xuanwu Hospital, Capital Medical University (No. 2008-1) and complied with the Declaration of Helsinki.

| Patient inclusion and exclusion criteria
Patients with acute LHI admitted to the neuro-intensive care unit (NICU) of Xuanwu Hospital, Capital Medical University from January 2013 to December 2017 were enrolled in this study. Informed consents were obtained from patients or their relatives. LHI diagnosis and management were performed according to the established guidelines. 2,9 The inclusion criteria were (1) 18-85 years old; (2) unilateral LHI involving at least 2/3 of the middle cerebral artery (MCA) territory, confirmed by computed tomography (CT) or magnetic resonance imaging (MRI); (3) within 48 hours after LHI onset; and (4) modified Rankin Scale (mRS)≤2 before symptom onset.
Exclusion criteria were (1) comorbidities of serious organ dysfunction or cancer; (2) infection within 48h after symptom onset; (3) death due to reasons other than brain herniation; and (4) cerebral lobe hemorrhage > 30 mL or ventricular hemorrhage. Brain herniation (in particular uncal herniation) was defined by deterioration of consciousness, a failure of upper brain stem function, and CT or MRI findings including severe cerebral edema, cerebral ventricular, and brain stem compression with midline shift. 9

| Clinical assessment and data collection
Blood samples were collected on the 2nd (ie, 24h-48h), 4th, and 7th day (D2, D4, and D7) after LHI onset for whole blood cell test, including white blood cell (WBC), neutrophil, and lymphocyte counts as well as neutrophil-to-lymphocyte ratio (NLR). Changes in comparison to D2 were shown as Δ. We also recorded (1) age; (2) sex; (3) history of F I G U R E 1 Patient enrollment flow chart smoking and drinking alcohol; (4) occurrence of headache, vomiting, disturbance of consciousness, and gaze palsy; (5) body temperature from D2 to D7; (6) blood pressure (BP), National Institutes of Health Stroke Scale (NIHSS), Glasgow Coma Scale (GCS), levels of procalcitonin (PCT), C-reactive protein (CRP), neuron-specific enolase (NSE), and blood glucose on D2 after symptom onset; (7) volume of cerebral infarction, which was measured semi-automatically from head CT or diffusion-weighted imaging (DWI) sequence of MRI on D2 after LHI using Fiji software and corrected with indirectly measured brain edema 10 ; (8) stroke etiology according to the TOAST classification (large artery atherosclerosis, cardiogenic cerebral embolism, smallartery occlusion, other determined and undetermined etiology) 11 ; (8) history of atrial fibrillation, cardiac dysfunction, hypertension, and diabetes mellitus; and (9) thrombolysis, endovascular treatment, and/or decompressive craniectomy. 2,9 Patients were followed up for 6 months after symptom onset, and mRS at 6 months was recorded.

| Statistical analysis
Statistical analyses were performed with SPSS 22.0 (IBM Corporation, Armonk, NY). Kolmogorov-Smirnov test was used to check the normality of continuous data. Normally distributed data were expressed as mean ± standard deviation (SD), while the nonnormally distributed data were expressed as median (interquartile range, IQR). Student's t-test, Mann-Whitney U-test, or chi-square test was used for intergroup comparisons, when appropriate.
Repeated measures ANOVA was used to explore the group-time effects regarding the changes of peripheral leukocyte counts within the first week poststroke. Pearson's correlation analysis was performed to explore the correlation between the peripheral leukocyte count and infarct/edema volumes. Spearman's analysis was used to investigate the correlation between peripheral leukocyte count and NIHSS, mRS, or GCS. Logistic regression analysis was used to investigate the risk factors of brain herniation and death. Receiver operating characteristic (ROC) curve was performed to find the cutoff value to predict brain herniation and death. All analyses were 2-tailed, and P < .05 was considered statistically significant. Subgroup analysis revealed that patients with brain herniation were younger (57.6 ± 13.6 vs 65.3 ± 13.7 years old), presented with lower initial GCS score (median: 7 vs 9, P < .01), larger cerebral infarction, and more severe brain edema compared to those without brain herniation (P < .01, Table 2). Patients that died had higher NIHSS (21.4 ± 6.5 vs 17.7 ± 5.9) and lower GCS scores (median: 6 vs 9), higher incidence of coronary heart disease upon admission (25.0% vs 6.3%), more often received intravenous thrombolysis (30.0% vs 10.9%) or decompressive craniectomy (55.0% vs 28.1%, P < .05), and exhibited larger infarcts and more severe brain edema compared to the survivors (P < .01, Table 3).

| Leukocyte dynamics in LHI patients
Compared to the patients with benign courses, patients developing herniation or dying from stroke showed higher WBC and Note:: The continuous data were described as mean ± SD for normally distributed data and as median/IQR for non-normally distributed data. Categorical data were shown as N (%).

TA B L E 1 Baseline characteristics of included LHI patients
neutrophil counts throughout the first week poststroke, as well as higher body temperature on D4 and D7 (P < .05, Figure 2).
No significant difference in percentages of neutrophils/lymphocytes or NLR was found between groups ( Figure S1). ΔWBC and Δneutrophil counts are presented in Figure S2. Repeated measures ANOVA revealed significant group-time effects between WBC/neutrophil and death (P < .01 and P < .05, respectively), but not between WBC/neutrophil and brain herniation (P > .05).

| D ISCUSS I ON
Our study found that LHI patients developing brain herniation or succumbing to stroke showed higher WBC and neutrophils counts within the first week after stroke. Neutrophil count on D2 after stroke significantly related to infarct and edema volumes as well as GCS, and also related to fatal outcome (brain herniation and death).
Animal studies revealed that neutrophils rapidly increase within a few hours after stroke and peak at 24-48h before gradually returning to baseline. [12][13][14] They contribute to the disruption of bloodbrain barrier and aggravate ischemic injury and edema by releasing reactive oxygen species, proteases (metalloproteinases, elastase, cathepsin G, etc), cytokines (eg, IL-6, IL-8, TNF-α), and chemokines (CCL2, CCL3, CCL5, etc). 13,15 They also play a multifaceted role in various stroke-related aspects such as thrombosis and poststroke infection. 12,13 Our results showed peripheral neutrophil count peaked on D2 after stroke and then gradually decreased on D4 in LHI patients without brain herniation. Similar changes were observed in stroke survivors. In contrast, neutrophil count further increased until D4 in patients with brain herniation. Hence, it may be proposed that neutrophil infiltration is more severe and persists for a longer time in malignant cerebral edema or prior to death ( Figure S3).
In our study, D7 neutrophil counts peaked across groups, which might be related to the frequent poststroke pneumonia.
Poststroke pneumonia might be a respiratory syndrome due to complex bacterial, chemical, and immunological factors, and could not be prevented by prophylactic antibiosis alone, even for patients at high aspiration risk. 16 All patients with known infection upon admission were excluded in our study. Moreover, body temperature on D2, as well as the levels of PCT and CRP, was comparable between groups. Thus, potential effects of early infection on D2 neutrophil count are unlikely. However, patients that developed pneumonia thereafter were not excluded due to high incidence in each group (>90%), and exclusion of such patients would therefore dramatically decrease sample size and might further cause a selection bias.
Our results provide preliminary evidence for a linear relationship between neutrophil count and infarct/edema volumes, which has not been established previously. Furthermore, neutrophil count was found to significantly relate to brain herniation and death after LHI. Various imaging modalities such as perfusion CT, PET, and MRI within 6-24h also have been reported to reliably predict a malignant LHI course mainly through evaluating infarct and edema volumes. [17][18][19] However, availability and feasibility of these imaging modalities are limited due to practical reasons early after symptom onset in many stroke units. Other biomarkers such as NLR, PCT, F I G U R E 2 Dynamics of peripheral leukocytes and body temperature in different groups of LHI patients. * /# P < .05, ** /## P < .01, ***P < .001

F I G U R E 3 (A) Linear fitting of WBC and neutrophil counts on D2 in relation to volumes of cerebral infarction and edema in LHI patients;
(B) ROC curves of WBC and neutrophil counts on D2 to predict brain herniation (i and iii, respectively) and death (ii and iv, respectively) and NSE have been suggested to predict the severity of ischemic stroke, [20][21][22]  primarily to the growth of infarct or edema, or the frequently observed pneumonia.
In conclusion, our study provides preliminary evidence that increased peripheral neutrophil count early after LHI relates to infarct and edema volumes, and can indicate fatal outcome of LHI patients, thus may help guiding acute management of LHI such as reduction of intracranial pressure. As a promising target in stroke therapy, neutrophils may also be used to identify patients that could benefit from adjunctive antiinflammatory therapies. 15 Future studies with larger numbers of patients are required to verify this hypothesis and shall also focus specifically on earlier time points after stroke (≤24h).

ACK N OWLED G M ENTS
The authors thank Dr Xiao-wen Wu and Dr He Miao for their assistance with imaging data analysis. This project was supported by Chinese Natural Science Fund (81671037) to Dr Ying-ying Su, Xuanwu Hospital Fund (XWJL-2019008) and Capital Medical University Science Cultivation Fund (PYZ19129) to Dr Li-li Cui.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.