Neutrophil-to-Lymphocyte Ratio Is an Independent Predictor of 30-Day Mortality of Intracerebral Hemorrhage Patients: a Validation Cohort Study

In a previous study in patients with intracranial hemorrhage (ICH), we found an association between high neutrophil-to-lymphocyte ratio (NLR) with poor short-term mortality. In the current study, this preliminary finding was validated using an independent patient cohort. A total of 181 ICH patients (from January 2016 to December 2017) were included. Diagnosis was confirmed using computed tomography (CT) in all cases. Patient survival (up to 30 days) was compared between subjects with high NLR (above the 7.35 cutoff; n = 74) versus low NLR (≤ 7.35; n = 107) using Kaplan-Meier analysis. A multivariate logistic regression was performed to identify factors that influenced the 30-day mortality. Correlation between NLR with other relevant factors (e.g., C-reactive protein (CRP) and fibrinogen) was examined using Spearman correlation analysis. The 30-day mortality was 19.3% (35/181) in the entire sample, 37.8% (28/74) in the high-NLR group, and 6.5% (7/107) in the low-NLR group (P < 0.001). In comparison to the low-NLR group, the high-NLR group had higher rate of intraventricular hemorrhage (29.7 vs. 16.8%), ICH volume (median 23.9 vs. 6.0 cm3) and ICH score (median 1.5 vs. 0), and lower GCS score (9.4 ± 4.5 vs. 12.9 ± 3.2). An analysis that divided the samples into three equal parts based on NLR also showed increasing 30-day mortality with incremental NLR (1.6, 15.0, and 41.7% from lowest to highest NLR tertile, P for trend < 0.001). Kaplan-Meier curve showed higher 30-day mortality in subjects with high NLR than those with low NLR (P < 0.001 vs. low-NLR group, log-rank test). High NLR (> 7.35) is associated with poor short-term survival in acute ICH patients.


Introduction
Acute intracerebral hemorrhage (ICH) is associated with high disability and mortality. Increasing evidence suggests that inflammation contributes significantly to tissue damage caused by ICH. Specifically, activated inflammatory cells could release a variety of proinflammatory cytokines and proteases (Zhao et al. 2007), which in turn cause secondary brain injury. Edema, typically the result of inflammatory responses and mechanical compression by hematoma, is a major clinical feature of secondary brain injury and contributes to neurological deterioration (Babu et al. 2012).
Neutrophil-to-lymphocyte ratio (NLR) increases with increasing severity of inflammatory response and has been associated with poor patient outcomes in cancers (Grenader et al. 2016;Ojerholm et al. 2017), cardiovascular diseases (Kurtul et al. 2016;Sari et al. 2015), ischemic diseases Fei Wang and Li Wang contributed equally to this work. (Aktimur et al. 2016;Qun et al. 2017), and a variety of other conditions (Ozcicek et al. 2017;Pan et al. 2017;Senturk et al. 2016). A recent study from this research group  showed an association of high NLR with 30-day mortality in ICH patients. High NLR has also been suggested to be predictive for 90-day prognosis (Lattanzi et al. 2016b and early neurological deterioration in patients with acute ICH (Lattanzi et al. 2017b). In a study by Lattanzi et al. (2018), NLR improved the accuracy of outcome prediction when added to the Modified ICH score. In patients with ischemic stroke, high NLR has also been associated with bleeding after thrombolysis (Guo et al. 2016). In the current study, we used an independent cohort of ICH patients to validate our previous finding that high NLR (> 7.35) is associated with 30-day mortality in ICH patients.

Study Sample
Consecutive adult ICH patients receiving treatment for acute ICH at the Emergency Department of Jiading District Central Hospital Affiliated Shanghai University of Medicine & Health Sciences between January 2016 and December 2017 were retrospectively reviewed. The diagnosis of ICH was established with CT scan in all subjects. The inclusion and exclusion criteria are listed in Table 1.
The study protocol was approved by the Ethics Review Board of Jiading District Central Hospital (No.2017-KY-09). All subjects were de-anonymized. Written informed consent was waived by the Ethics Review Board.

Data Collection
Demographic information, past medical history, clinical data, and laboratory measures were collected from medical records. Hypertension was defined using the 2013 ESH/ESC Guidelines (Mancia et al. 2014): resting systolic pressure (SBP) at ≥ 140 mmHg and/or diastolic pressure (DBP) at ≥ 90 mmHg on three separate occasions or regular use of antihypertension medications. Diabetes was defined using the 2016 American Diabetes Association Guidelines (Chamberlain et al. 2016). All laboratory tests were carried out using venous blood collected after over-night fasting. Patient management was, in principle, based on the 2015 American Heart Association/American Stroke Association Guidelines (Hemphill 3rd et al. 2015).

Imaging Analysis
The ICH diagnosis was based on clinical features and confirmed by a post hoc assessment of CT images by an experienced neurologist. The following features were extracted using the CT slice with the largest ICH area: (A) the largest diameter of the hematoma; (B) the dimension of the hemorrhage perpendicular to the largest diameter as the second diameter; (C) the height of the hematoma, as calculated by multiplying the number of slices involved by the slice thickness. ICH volume was calculated as follows: ABC/2 (Kothari et al. 1996). Intraventricular hemorrhage (IVH) was defined as hyperdense intraventricular signal not attributable to calcification or choroid plexus.

Statistical Analysis
Based on our previous study ), the study sample was divided using NLR at a cutoff of 7.35. Continuous variables were analyzed using Student's t test if normally distributed and with Mann-Whitney U test if otherwise. Categorical variables were analyzed using χ 2 test. Potential association between NLR and 30-day mortality was also assessed by dividing the sample into three parts of equal size followed by P for trend analysis using the Jonckheere-Terpstra test. Spearman correlation analysis was used to determine the correlation of NLR with other factors. Multiple logistic regression was conducted to identify the factors that influenced the 30-day mortality. P < 0.05 was considered statistically significant. All statistical analyses were performed using IBM SPSS 19.0 (IBM, Armonk, New York, USA).

Results
A total of 213 patients with acute ICH sought emergency care at our department during the study period; 32 patients were excluded due to treatment discontinuation within 24 h (n = 19), hospital admission at > 24 h after the first symptom (n = 2), infection within 2 weeks before ICH (n = 5), anticoagulant use within 3 months (n = 5), and leukemia (n = 1). The final analysis included 181 patients (112 men; age 65.8 ± Patients with hematologic disorders, immunosuppressant drug users (steroids), those with a history of infection within 2 weeks before ICH, a stroke history within 6 months, patients with a history of malignancy, and those using anticoagulants. Patients who refused treatment.
The Spearman correlation analysis showed an association between NLR with the presence of IVH, ICH volume, GCS score, ICH score, and 30-day mortality as well as CRP (Table 4).
The Kaplan-Meier analysis showed that patients with high NLR had significantly higher 30-day mortality than those with low NLR (log-rank test, P < 0.001, Fig. 2).

Discussion
Previous studies indicated that NLR is closely related to the prognosis of stroke patients (Aktimur et al. 2016;Qun et al. 2017). High NLR is associated with 30-day mortality ) and in-hospital mortality (Giede-Jeppe et al. 2017), as well as 90-day mortality (Lattanzi et al. 2016a, b;Tao et al. 2017) in ICH patients. In patients with ischemic stroke, high NLR has also been associated with hemorrhagic transformation upon thrombolysis (Guo et al. 2016). In the current study, we found a close association of high NLR (> 7.35) with IVH, ICH volume, and ICH score. We also identified a negative correlation between NLR and GCS score. Multivariate logistic regression showed that high NLR is an independent risk for 30-day mortality.
In summary, we found higher 30-day mortality in ICH patients with high NLR (> 7.35). Multivariate regression showed that high NLR is an independent risk for 30-day mortality.
Author Contributions Fei Wang and Li Wang: carried out the studies, participated in collecting data, and drafted the manuscript. Ting-ting Jiang, Jian-jun Xia, and Wen-hui Kang: participated in collecting data and helped to draft the manuscript. Li-juan Shen: participated in collecting data and tested the blood samples. Feng Xu: performed the statistical analysis. Yong Ding, Li-xia Mei, and Xue-feng Ju: participated in collecting data and followed up patients. Shan-you Hu and Xiao Wu: design, review, and edit the manuscript. All authors read and approved the final manuscript.

Compliance with Ethical Standards
Conflicts of Interest The authors declare that they have no conflict of interest.