Abstract
Poor functional outcome is associated with perihematomal edema (PHE) expansion after intracerebral hemorrhage (ICH). The inflammatory response is crucial for the onset and progression of PHE. This study aimed to determine the connection between admission neutrophil-lymphocyte ratio (NLR) and early PHE development. We retrospectively analyzed patients with ICH admitted to the Chaohu Affiliated Hospital of Anhui Medical University from January 2021 to December 2022. The primary outcome measure was absolute PHE, defined as the volume of the follow-up PHE minus admission PHE. A semiautomated measurement tool (3D Slicer) was used to calculate the volumes of cerebral hematoma and cerebral edema. Spearman’s correlation analysis determined the relationship between NLR and absolute PHE. The multiple linear regression model was constructed to analyze the predictive relation of admission NLR on early PHE expansion. A total of 117 patients were included. The median hematoma and PHE volumes on admission were 9.38 mL (interquartile range [IQR], 4.53–19.54) and 3.54 mL (IQR, 1.33–7.1), respectively. The median absolute PHE was 2.26 mL (IQR, 1.25–4.23), and the median NLR was 3.10 (IQR, 2.26–3.86). Spearman’s correlation test showed a positive correlation between admission NLR and absolute PHE (r = .548, P < .001). Multiple linear regression analyses suggested that for every 1-unit increase in admission NLR (B = .176, SE = .043, Beta = .275, P < .001), there was a 0.176 mL increase in absolute PHE. Admission neutrophil-to-lymphocyte ratio (NLR) significantly and positively predicted early perihematomal edema (PHE) expansion.
Keywords: brain edema, inflammation, intracerebral hemorrhage, leukocyte, neutrophil, neutrophil–lymphocyte ratio, perihematomal edema
1. Introduction
Intracerebral hemorrhage (ICH) is a devastating stroke characterized by rupturing of a cerebral vessel and bleeding into the brain parenchyma.[1,2] Despite making up just 15% to 20% of all stroke cases, ICH has a high disability and mortality rate, and there were 42 instances of ICH per 100,000 person-years, globally.[2,3] After cerebral hemorrhage, the hematoma and its degrading components induce inflammatory and cytotoxic responses that cause further damage to the surrounding parenchyma in the following days to weeks.[4] Perihematomal edema (PHE) is a quantitative marker of secondary brain damage after cerebral hemorrhage and is associated with a poor prognosis.[4,5] Recently, a systemic inflammatory response with altered white blood cell counts has been shown to correlate with the severity and prognosis of cerebral edema.[6,7] The neutrophil-lymphocyte ratio (NLR), as an easily accessible inflammatory marker, may reflect the degree of inflammation and predict stroke outcome.[8]
In this study, we aimed to explore the relationship between the NLR on admission and PHE growth, and whether NLR can predict PHE growth.
2. Methods
This retrospective study was approved by the Ethics Committee of Chaohu Affiliated Hospital of Anhui Medical University (Ethics Code: KYXM-202211-013), which was conducted in accordance with the principles of the Declaration of Helsinki. Because this was an anonymized retrospective analysis, the ethics committee did not require patients’ informed consent.
2.1. Study design and population
Patients with ICH were retrospectively selected from January 2021 to December 2022 in the Chaohu Affiliated Hospital of Anhui Medical University.
Inclusion criteria: first attack of intracerebral hemorrhage; age > 18 years; a CT scan and peripheral blood examination were performed within 24 hours of the onset of stroke symptoms.
Exclusion criteria: history of infection within 2 weeks before the onset of stroke; surgical treatment such as hematoma removal and cranial decompression; secondary causes of ICH (traumatic injury, arteriovenous malformation, tumor, conversion of hemorrhage after cerebral infarction, drug-induced cerebral hemorrhage, coagulation abnormality, and other reasons found during hospitalization); history of taking immunosuppressive agents; patients suffering from cancer, auto-immune diseases, severe liver and kidney diseases; prior to the commencement of the condition, there was a history of atrial fibrillation, myocardial infarction, dementia, and the presence of significant sequela of previous stroke.
2.2. Clinical data
We saved Digital Imaging and Communications in Medicine (DICOM) format data of both admission and follow-up CT scans which were collected for further review. Hematoma and PHE volumes on admission and follow-up were measured by a semiautomated measurement tool (3D Slicer, version 5.3.0) (Fig. 1). Baseline data were collected from all enrolled subjects, including demographics (gender composition, age), lifestyle risk factors (smoking, alcohol consumption), past medical history (hypertension, diabetes mellitus), admission blood pressure, GCS scores within 3 hours of admission, baseline biochemistry on admission (blood glucose, electrolytes), and routine blood test. Our primary predictor variable was NLR. The primary outcome measure was absolute PHE, defined as the volume of the difference between the follow-up PHE and the admission PHE.
Figure 1.
Panels 1 and 2 show the CT images on admission. Panel 3 shows the 3D images generated by the 3D Slicer, admission ICH volume was 21.31 mL, and admission PHE volume was 6.98 mL. Panels 4, 5, and 6 show the CT and 3D images at the follow-up examination after 16.75 h, the follow-up ICH volume was 24.20 mL, and the follow-up PHE volume was 9.38 mL.
2.3. Statistical analysis
Data were analyzed using IBM SPSS Statistics software (version 26.0; IBM Corp). Continuous data were expressed as mean ± standard deviation or the median and interquartile range (IQR). Discrete variables were expressed as percentages (%). Correlation analysis was performed by Spearman’s method. Multiple linear regression analyses were performed on multiple variables with high correlation coefficients. P < .05 was taken as the difference was statistically significant.
3. Results
3.1. Baseline characteristics
A total of 117 patients were included in the final analysis of this study (the enrollment flow chart of the study population is shown in Fig. 2). There were more men (n = 77, 65.8%) than women (n = 40, 34.2%) enrolled, and the average age was 62.01 (±11.85) years. The median admission ICH volume was 9.38 mL (IQR, 4.53–19.54), the median admission PHE volume was 3.54 mL (IQR, 1.33–7.16), the median time from admission to follow-up CT scan was 7.03 hours (IQR, 4.58–13.17), the median absolute PHE volume was 2.26 mL (IQR, 1.25–4.23), and the median NLR was 3.10 (IQR, 2.26–3.86). The remaining baseline demographic and clinical characteristics are shown in Table 1.
Figure 2.
. Enrollment process.
Table 1.
Demographic and clinical characteristics.
| Demographic characteristics | ICH cases, N = 117 |
|---|---|
| Age (yr), mean (SD) | 62.01 (11.85) |
| Gender | |
| Male, n (%) | 77 (65.80%) |
| Female, n (%) | 40 (34.20%) |
| Smoking, n (%) | 45 (45.30%) |
| Alcohol consumption, n (%) | 47 (40.20%) |
| Medical history | |
| Diabetes mellitus, n (%) | 10 (8.50%) |
| Hypertension, n (%) | 86 (73.50%) |
| Admission clinical characteristics | |
| Blood pressure | |
| SBP (mm Hg), mean (SD) | 169.94 (25.06) |
| DBP (mm Hg), mean (SD) | 96.5 (25.03) |
| GCS score, median (IQR) | 15 (13–15) |
| Location of ICH | |
| Deep, n (%) | 94 (80.30%) |
| Brain stem, n (%) | 1 (0.90%) |
| Lobar, n (%) | 19 (16.20%) |
| Cerebellum, n (%) | 3 (2.60%) |
| Admission ICH volume (mL)a | 9.38 (4.53–19.54) |
| Admission PHE volume (mL)a | 3.54 (1.33–7.16) |
| Follow-up ICH volume (mL)a | 9.70 (4.9–21.27) |
| Follow-up PHE volume (mL)a | 5.60 (2.90–11.53) |
| Absolute PHE increase (mL)a | 2.26 (1.25–4.23) |
| Time interval from admission CT to review (h)a | 7.03 (4.58–13.17) |
| Admission blood examination | |
| INRa | 0.91 (0.87–0.95) |
| D-dimer (µg/mL)a | 0.34 (0.22–0.59) |
| Fibrinogen (g/L)a | 2.85 (2.53–3.39) |
| Prothrombin time (s)a | 12.30 (11.60–12.8) |
| Blood glucose (mmol/L)a | 6.70 (5.80–7.95) |
| Serum kalium (mmol/L)a | 3.65 (3.35–3.91) |
| Serum natrium (mmol/L)a | 139.00 (137.50–141.45) |
| Serum calcium (mmol/L )a | 2.17 (2.11–2.23) |
| Total leukocyte count (109/L)a | 7.35 (6.16–8.71) |
| Neutrophil count (109/L)a | 5.21 (4.13–6.42) |
| Lymphocyte count (109/L)a | 1.24 (0.99–1.85) |
| Monocyte count (109/L)a | 0.46 (0.35–0.59) |
| Neutrophil-to-lymphocyte ratioa | 3.10 (2.26–3.86) |
CT = computerized tomography, DBP = diastolic blood pressure, GCS = Glasgow coma scale, ICH = intracerebral hemorrhage, INR = international normalized ratio, IQR = interquartile range, mL = milliliters, PHE = perihematomal edema, SBP = systolic blood pressure, SD = standard deviation.
Values represented as the median (interquartile range).
3.2. Spearman’s correlation analysis
Spearman’s correlation analysis showed that admission systolic blood pressure (r = .191, P = .039), diastolic blood pressure (r = .197, P = .034), admission ICH volume (r = .667, P < .001), admission PHE volume (r = .667, P < .001), time from admission to follow-up CT scan (r = .203, P = .028), admission blood glucose (r = .277, P = .002), white blood cell count (r = .186, P = .044), neutrophil count (r = .386, P < .001), and NLR (r = .548, P < .001) were positively correlated with absolute PHE, and admission GCS score (r = −.469, P < .001), and lymphocyte count (r = −.478, P < .001) were negatively correlated with absolute PHE (Table 2).
Table 2.
Spearman correlation analysis.
| Variable | Absolute PHE | NLR | ||
|---|---|---|---|---|
| r | P | r | P | |
| Age | −.106 | .254 | .012 | .901 |
| Gender (male) | −.015 | .871 | .005 | .959 |
| Smoking | .113 | .225 | .011 | .908 |
| Alcohol consumption | .042 | .652 | .015 | .871 |
| Medical history | ||||
| Diabetes mellitus | −.095 | .308 | −.065 | .485 |
| Hypertension | .083 | .373 | .058 | .535 |
| Admission clinical characteristics | ||||
| Systolic blood pressure | .191* | .039 | .128 | .168 |
| Diastolic blood pressure | .197* | .034 | .074 | .425 |
| Glasgow coma scale score | −.469** | <.001 | −.275** | .003 |
| Admission ICH volume | .667** | <.001 | .236* | .010 |
| Admission PHE volume | .667** | <.001 | .138 | .137 |
| Time interval from admission CT to review | .203* | .028 | −.046 | .620 |
| Admission blood examination | ||||
| INR | −.023 | .809 | .054 | .573 |
| D-dimer | .175 | .064 | .249** | .008 |
| Fibrinogen | .172 | .069 | .231* | .014 |
| Prothrombin time | .039 | .678 | −.008 | .933 |
| Blood glucose | .277** | .002 | .156 | .093 |
| Serum kalium | −.090 | .336 | −.113 | .226 |
| Serum natrium | .055 | .554 | −.152 | .101 |
| Serum calcium | −025 | .786 | −.153 | .100 |
| Total leukocyte count | .186* | .044 | .372** | <.001 |
| Neutrophil count | .386** | <.001 | .700** | <.001 |
| Lymphocyte count | −478** | <.001 | −.831** | <.001 |
| Monocyte count | −.126 | .177 | −.119 | .201 |
| Neutrophil-to-lymphocyte ratio | .548** | <.001 | 1.000 | |
Bold values represents the results are statistically significant.
CT = computerized tomography, ICH = intracerebral hemorrhage, INR = international normalized ratio, PHE = perihematomal edema.
.01 < P < .05.
P < .01.
3.3. Multiple linear regression analysis
Variables with significant correlation with absolute PHE were studied using multiple linear regression analysis, which strongly correlated with NLR were removed to avoid the presence of collinearity. The results of the variance inflation factor indicate that there is no significant problem of multicollinearity between the variables. The outcome is shown in Table 3. In model 1, admission systolic blood pressure (Beta = .100, P = .198), diastolic blood pressure (Beta = −.051, P = .509), blood glucose (Beta = −.015, P = .816) did not predict early PHE growth, removing these four variables multiple linear regression was performed on the remaining variables. Table 4 displays the results. NLR (B = .176, SE = .043, Beta = .275, P < .001) significantly and positively predicts early PHE growth, with 0.176 mL of absolute PHE growth for every 1 unit increase in NLR. A combination of the NLR, admission ICH volume, admission GCS score, and the time from admission to follow-up CT scan explained 56.7% of the variation in absolute PHE.
Table 3.
Multiple linear regression analysis model 1.
| Variable | B | SE | Beta | t | P | VIF | |
|---|---|---|---|---|---|---|---|
| NLR | .168 | .045 | .245 | 3.761 | <.001 | 1.117 | |
| Admission GCS score | −.194 | .081 | −.161 | −2.388 | .019 | 1.192 | |
| Admission SBP | .010 | .008 | .100 | 1.294 | .198 | 1.569 | |
| Admission DBP | −.009 | .013 | −.051 | −.663 | .509 | 1.557 | |
| Admission ICH volume | .139 | .029 | .553 | 4.841 | <.001 | 3.428 | |
| Admission PHE volume | −.011 | .062 | −.020 | −.181 | .856 | 3.182 | |
| Time from admission to follow-up CT scan | .112 | .029 | .244 | 3.894 | <.001 | 1.032 | |
| Admission blood glucose | −.018 | .077 | −.015 | −.233 | .816 | 1.082 | |
| F | 19.317*** | Durbin–Watson | 1.619 | ||||
| R-squared | .589 | Adj. R-squared | .558 | ||||
Bold values represents the results are statistically significant.
DBP = diastolic blood pressure, GCS = Glasgow coma scale, ICH = intracerebral hemorrhage, NLR = neutrophil-to-lymphocyte ratio, PHE = perihematomal edema, SBP = systolic blood pressure, SE = standard error, VIF = variance inflation factor.
P < .001.
Table 4.
Multiple linear regression analysis model 2.
| Variable | B | SE | Beta | t | P | VIF | |
|---|---|---|---|---|---|---|---|
| NLR | .176 | .043 | .257 | 4.056 | <.001 | 1.075 | |
| Admission ICH volume | .134 | .016 | .534 | 8.162 | <.001 | 1.149 | |
| Admission GCS score | −.202 | .078 | −.168 | −2.585 | .011 | 1.134 | |
| Time from admission to follow-up CT scan | .113 | .028 | .246 | 3.989 | <.001 | 1.017 | |
| F | 39.009*** | Durbin–Watson | 1.624 | ||||
| R-squared | .582 | Adj. R-squared | .567 | ||||
Bold values represents the results are statistically significant.
GCS = Glasgow coma scale, ICH = intracerebral hemorrhage, NLR = neutrophil-to-lymphocyte ratio, SE = standard error, VIF = variance inflation factor.
P < .001.
4. Discussion
Vladimir et al showed a close correlation between NLR and neuroimaging markers (hematoma volume, especially edema volume) in their study.[9] Similar results were obtained in our research. After further analysis, we found that admission NLR could be a predictor of early PHE growth. Admission NLR positively predicts the growth of early PHE. Higher admission NLR was related to larger brain edema volume. NLR is an accessible biomarker of systemic inflammation after ICH.[7,10,11] Therefore, these results suggest a potential role of inflammatory pathways in the development of early PHE.
After intracerebral hemorrhage, both the local central nervous system and systemic inflammatory responses will be activated.[12] Neutrophils are the first leukocytes to actively migrate from the peripheral blood to the brain within an hour after ICH.[10] Infiltrating leukocytes release inflammatory and cytotoxic mediators which exacerbate the peri-lesional edema by favoring capillary permeability, cell swelling, and blood–brain barrier damage.[13] Lymphocytes are the main regulators of immunity, and it has been shown that there will be lymphocyte reduction and functional inactivation after brain injury.[11] Therefore, inflammatory pathways may be potential therapeutic targets for brain edema. Further research is needed on how to control the inflammatory response after cerebral hemorrhage.
Consistent with previous findings,[14,15] we also observed that admission ICH volume significantly and positively predicted PHE expansion. In addition, we performed a multiple linear regression analysis, but the results suggested that the variables NLR, admission ICH volume, admission GCS score, and the time from admission to follow-up CT scan explained only 56.7% of the variation in absolute PHE. Considering that PHE expansion is the consequence of multifactorial effects, additional research on its predictors is necessary so that the risk of PHE expansion can be more accurately predicted and stratified.
There are several noteworthy limitations of this study. First, because it was a retrospective study, our results may have been limited by selection bias, a single center, and a small sample size; we only analyzed imaging at the time of patient admission and to the nearest 24 hours, but there were temporal differences in CT images between patients. Second, by excluding patients who underwent surgical treatment, our findings were limited to patients who were treated conservatively. Again, our study was limited to changes in PHE in the early stages of ICH, and the NLR was only collected on admission. Whereas NLR and PHE are constantly changing, changes in late PHE and NLR still require studies with larger sample sizes and data collection studies at different time points to further elucidate the relationship between PHE and NLR at different time points.
5. Conclusions
In conclusion, our study suggests that elevated admission NLR predicts early PHE expansion in patients with ICH. It provides a reference for early identification of patients at risk of PHE growth and early implementation of treatment programs. Due to the small sample size of this study and the limitations of NLR and CT data collection, subsequent studies need to conduct larger prospective studies on additional time points of PHE to confirm the role of NLR in PHE growth.
Author contributions
Conceptualization: Yirong Mao, Lumao Huang.
Data curation: Yirong Mao, Gengsheng Ji, Liang Wang, Xiang Wang, Xinyi Zheng.
Formal analysis: Yirong Mao.
Investigation: Yirong Mao, Lumao Huang.
Methodology: Yirong Mao.
Project administration: Yirong Mao, Lumao Huang.
Validation: Yirong Mao.
Writing—original draft: Yirong Mao.
Writing—review & editing: Yirong Mao, Lumao Huang.
Supervision: Lumao Huang.
Abbreviations:
- CT
- computerized tomography
- DBP
- diastolic blood pressure
- GCS
- Glasgow coma scale
- ICH
- intracerebral hemorrhage
- INR
- international normalized ratio
- IQR
- interquartile range
- mL
- milliliters
- NLR
- neutrophil-lymphocyte ratio
- PHE
- perihematomal edema
- SBP
- systolic blood pressure
- SD
- standard deviation
- SE
- standard error
- VIF
- variance inflation factor
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
How to cite this article: Mao Y, Huang L, Ji G, Wang L, Wang X, Zheng X. Neutrophil-to-lymphocyte ratio on admission predicts early perihematomal edema growth after intracerebral hemorrhage. Medicine 2024;103:12(e37585).
Contributor Information
Yirong Mao, Email: mao_yirong@163.com.
Gengsheng Ji, Email: 1070437254@qq.com.
Liang Wang, Email: 26937890400@qq.com.
Xiang Wang, Email: 26937890400@qq.com.
Xinyi Zheng, Email: 1973945625@qq.com.
References
- [1].Magid-Bernstein J, Girard R, Polster S, et al. Cerebral hemorrhage: pathophysiology, treatment, and future directions. Circ Res. 2022;130:1204–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Puy L, Parry-Jones AR, Sandset EC, et al. Intracerebral haemorrhage. Nat Rev Dis Primers. 2023;9:14. [DOI] [PubMed] [Google Scholar]
- [3].Campbell BCV, Khatri P. Stroke. Lancet. 2020;396:129–42. [DOI] [PubMed] [Google Scholar]
- [4].Ironside N, Chen CJ, Ding D, et al. Perihematomal edema after spontaneous intracerebral hemorrhage. Stroke. 2019;50:1626–33. [DOI] [PubMed] [Google Scholar]
- [5].Chen Y, Chen S, Chang J, et al. Perihematomal edema after intracerebral hemorrhage: an update on pathogenesis, risk factors, and therapeutic advances. Front Immunol. 2021;12:740632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Ferro D, Matias M, Neto J, et al. Neutrophil-to-lymphocyte ratio predicts cerebral edema and clinical worsening early after reperfusion therapy in stroke. Stroke. 2021;52:859–67. [DOI] [PubMed] [Google Scholar]
- [7].Menon G, Johnson SE, Hegde A, et al. Neutrophil to lymphocyte ratio: a novel prognostic marker following spontaneous intracerebral haemorrhage. Clin Neurol Neurosurg. 2021;200:106339. [DOI] [PubMed] [Google Scholar]
- [8].Zou F, Wang J, Han B, et al. Early neutrophil-to-lymphocyte ratio is a prognostic marker in acute ischemic stroke after successful revascularization. World Neurosurg. 2022;157:e401–9. [DOI] [PubMed] [Google Scholar]
- [9].Rendevski V, Aleksovski B, Mihajlovska Rendevska A, et al. Inflammatory and oxidative stress markers in intracerebral hemorrhage: relevance as prognostic markers for quantification of the edema volume. Brain Pathol. 2023;33:e13106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Mracsko E, Javidi E, Na SY, et al. Leukocyte invasion of the brain after experimental intracerebral hemorrhage in mice. Stroke. 2014;45:2107–14. [DOI] [PubMed] [Google Scholar]
- [11].Lattanzi S, Cagnetti C, Provinciali L, et al. Neutrophil-to-lymphocyte ratio and neurological deterioration following acute cerebral hemorrhage. Oncotarget. 2017;8:57489–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Askenase MH, Sansing LH. Stages of the inflammatory response in pathology and tissue repair after intracerebral hemorrhage. Semin Neurol. 2016;36:288–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Lee KR, Kawai N, Kim S, et al. Mechanisms of edema formation after intracerebral hemorrhage: effects of thrombin on cerebral blood flow, blood-brain barrier permeability, and cell survival in a rat model. J Neurosurg. 1997;86:272–8. [DOI] [PubMed] [Google Scholar]
- [14].Ye G, Huang S, Chen R, et al. Early predictors of the increase in perihematomal edema volume after intracerebral hemorrhage: a retrospective analysis from the Risa-MIS-ICH Study. Front Neurol. 2021;12:700166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Arima H, Wang JG, Huang Y, et al. INTERACT Investigators. Significance of perihematomal edema in acute intracerebral hemorrhage: the INTERACT trial. Neurology. 2009;73:1963–8. [DOI] [PMC free article] [PubMed] [Google Scholar]