Neutrophil-to-lymphocyte ratio is associated with increased cerebral blood flow velocity in acute bacterial meningitis

In community-acquired bacterial meningitis (CABM) intracranial vascular alterations are devastating complications which are triggered by neuroinflammation and result in worse clinical outcome. The Neutrophil-to-Lymphocyte ratio (NLR) represents a reliable parameter of the inflammatory response. In this study we analyzed the association between NLR and elevated cerebral blood flow velocity (CBFv) in CABM-patients. This study included all (CABM)-patients admitted to a German tertiary center between 2006 and 2016. Patients’ demographics, in-hospital measures, neuroradiological data and clinical outcome were retrieved from institutional databases. CBFv was assessed by transcranial doppler (TCD). Patients’, radiological and laboratory characteristics were compared between patients with/without elevated CBFv. Multivariate-analysis investigated parameters independently associated with elevated CBFv. Receiver operating characteristic(ROC-)curve analysis was undertaken to identify the best cut-off for NLR to discriminate between increased CBFv. 108 patients with CABM were identified. 27.8% (30/108) showed elevated CBFv. Patients with elevated CBFv and normal CBFv, respectively had a worse clinical status on admission (Glasgow Coma Scale: 12 [9–14] vs. 14 [11–15]; p = 0.005) and required more often intensive care (30/30 [100.0%] vs. 63/78 [80.8%]; p = 0.01).The causative pathogen was S. pneumoniae in 70%. Patients with elevated CBFv developed more often cerebrovascular complications with delayed cerebral ischemia (DCI) within hospital stay (p = 0.031). A significantly higher admission-NLR was observed in patients with elevated CBFv (median [IQR]: elevated CBFv:24.0 [20.4–30.2] vs. normal CBFv:13.5 [8.4–19.5]; p < 0.001). Multivariate analysis, revealed NLR to be significantly associated with increased CBFv (Odds ratio [95%CI] 1.042 [1.003–1.084]; p = 0.036). ROC-analysis identified a NLR of 20.9 as best cut-off value to discriminate between elevated CBFv (AUC = 0.713, p < 0.0001, Youden's Index = 0.441;elevated CBFv: NLR ≥ 20.9 19/30[63.5%] vs. normal CBFv: NLR > 20.9 15/78[19.2%]; p < 0.001). Intracranial vascular complications are common among CABM-patients and are a risk factor for unfavorable outcome at discharge. Elevated NLR is independently associated with high CBFv and may be useful in predicting patients’ prognosis.

www.nature.com/scientificreports/ in damage to cortical and subcortical structures 7,8 which further result in edema, hydrocephalus and increased intracranial pressure 9 . Therefore they represent a devastating complication leading to unfavorable outcome and increased morbidity 10 . The inflammatory process in bacterial meningitis can be monitored by numerous biochemical markers 11,12 . The Neutrophil-to-Lymphocyte ratio (NLR) represents information on both the innate and adaptive immune system and is a reliable parameter for the general immune response to various stimuli. Neutrophil-to-Lymphocyte ratio (NLR), calculated by absolute neutrophil count divided by absolute lymphocyte count, is easy to perform in routine practice and cost-effective. Neutrophil-to-Lymphocyte ratio (NLR) has proven its prognostic value in cerebrovascular 13,14 and inflammatory diseases 15 , several types of tumors [16][17][18] , infections 19,20 and also in the differential diagnosis between viral and bacterial meningitis 21,22 . This study aimed to analyze the association between NLR and elevated cerebral blood flow velocity (CBFv) which represents a devastating complication in bacterial meningitis.

Methods
Patients und Inclusion criteria. All consecutive patients with diagnosis of CABM admitted to the Department of Neurology, University Hospital Erlangen, Germany were included in a prospective longitudinal institutional database, which was approved by the institutional ethics committee. Out of this database, all patients (N = 141) admitted between 2006 and 2016 for treatment of CABM have been recruited to this study. We excluded patients receiving permanent immunomodulatory treatment (e.g. corticosteroids, methotrexate, cytostatic drugs and biologicals) on admission (n = 14). Patients without follow-up data or refused consent were also excluded (n = 19). 108 patients remained for final analysis (Fig. 1).

Data collection.
Data of all patients were retrieved from our institutional prospective database: patients' history (hypertension, diabetes mellitus, alcohol abuse or other comorbidities), GCS (Glasgow Coma Scale) on admission, clinical symptoms on admission, laboratory findings on admission (cerebrospinal fluid and blood work, causative pathogen), clinical course, outcome and neurologic findings at discharge, and treatment. Arterial vascular alterations and incidence of cerebral ischemia were assessed using computed tomography imaging (CT-scan, (SIEMENS Somatom Volume Zoom, Somatom Sensation 64, Somatom Definition AS+; Siemens Healthcare, Forchheim, Germany) or magnetic resonance imaging (MRI) (SIEMENS Magnetom Sonata 1.5T, Magnetom Aera 1.5T, Siemens Healthcare, Erlangen, Germany).

Figure 1.
Flowchart of patients. 141 patients with community-acquired bacterial meningitis were identified during the study period. After exclusion of 33 patients 108 patients remained for further analysis. Patients were dichotomized according to increased cerebral blood flow velocity (N = 30) and normal cerebral blood flow velocity (N = 78). Abbreviations: CABM community acquired bacterial meningitis, CBFv cerebral blood flow velocity, GOS Glasgow outcome scale (range, 5 no or mild deficit, to 1, death), NLR Neutrophil-to-Lymphocyte ratio. www.nature.com/scientificreports/ Diagnosis of acute bacterial meningitis (ABM). The diagnosis of acute bacterial meningitis was confirmed by identification of the causative pathogen in cerebrospinal fluid and/or blood via microscopic Gram stain, CSF/blood cultured isolates and PCR on CSF 23 . In case the causative pathogen could not be identified bacterial meningitis was diagnosed by typical CSF findings as mainly granulocytic CSF white blood cell count greater than 1000 cells/μl, an increase of CSF protein of more than 100 mg/dl, and a CSF/serum glucose ratio less than 0.3 and/or presence of clinical symptoms as fever, neck stiffness, headache, impaired consciousness 24 . CSF lactate level was measured using standard enzymatic test.

Detection of increased cerebral blood flow velocity. Presence of vascular alterations was assessed by
TCD as a routine clinical practice termed as cerebral blood flow velocity (CBFv) of the anterior cerebral artery (ACA), middle cerebral artery (MCA), internal cerebral artery (ICA), posterior cerebral artery (PCA), and the basilar artery (BA). Systolic CBFv greater than 150 cm/s were considered increased 25,26 . Outcome. Outcome at discharge was evaluated according to the Glasgow Outcome Scale (1-5) by two physicians, trained and certified for data collection: A score of 1 on this scale indicates death; a score of 2, a vegetative state; a score of 3, severe disability (the patient is not able to live independently but can follow commands); a score of 4, moderate disability (the patient is capable of living independently but unable to return to work); and a score of 5, mild or no disability (the patient is able to return to work). A favorable outcome was defined as a score of 5, and an unfavorable outcome as a score of 1-4. The Glasgow Outcome Scale is a well-validated instrument with high interobserver agreement 27,28 . Long-term outcome was evaluated using the modified Rankin Scale (mRS) 3 and 12 months after onset. Favorable outcome representing an independent clinical status was defined as mRS score of 0-2, unfavorable outcome as mRS score of 3-6 29 .
Statistical analysis. Statistical analysis was performed using SPSS 22.0 (IBM Analytics, Armonk, New York, United States) and GraphPad Prism 8 (GraphPad Software, San Diego, California, United States). Categorial variables were presented as frequency and percentage, Pearson chi square and Fisher's exact test were used to compare between these groups. For continuous variables, the Kolmogorov-Smirnov test was used to test the distribution of data. If data showed normal distribution, data was presented with mean ± SD and the Student t test was used for analysis. Data lacking normal distribution, median and interquartile range were shown and the Mann-Whitney U test was used for comparison. Significance level was set at α = 0.05. All parameters showing a statistical trend (p < 0.1) were included in a multivariate model to identify parameters independently associated with elevated cerebral blood flow. A Receiver Operating Characteristic (ROC) curve and Youden's J statistic was used to determine the cut-off value for NLR 30 . Then, patients were dichotomized according to the identified cut-off value.
Ethics approval and consent to participate. All procedures involving human participants were in accordance with the ethical standards of the institutional research committee of the University of Erlangen-Nuremberg and with the 1964 Helsinki declaration and its later amendments. The Institutional Ethics Committee of the University of Erlangen-Nuremberg had approved the study protocol (304_16 B).
Informed consent. Informed consent was obtained from all subjects.

Patient characteristics.
Overall 108 patients with CABM remained for final analysis (Fig. 1 [8.4-19.5]; p < 0.001). Parameters with tendency for significance (p < 0.1) in univariate testing were included into a multivariable model. In this model NLR levels on admission were associated with increased cerebral blood flow velocity (1.042 (1.003-1.084); p = 0.036; Table 2) as well as need of osmotherapy (p = 0.026). Need of catecholamine therapy (p = 0.561), clinical status on admission (p = 0.116) and need of ventilation (p = 0.178) were not significantly associated with increased CBFv.
Association of Neutrophil-to-Lymphocyte ratio with increased cerebral blood flow velocity. ROC Fig. 3). Table 2. Multivariate analysis of parameters associated with increased cerebral blood flow velocity. Multivariable regression analysis was calculated for the association with increased cerebral blood flow velocity. Only parameters showing a statistical trend (p < 0.1) in prior univariate testing were included in the multivariable model. For each parameter risk ratio and 95% confidence interval are provided. Significant findings are expressed in bold. NLR Neutrophil-to-Lymphocyte ratio, 95% CI Confidence Interval.

Discussion
There are three major findings from this cohort study on CABM-patients: (1) We identified elevated NLR on admission to be associated with increased CBFv within hospital stay. (2) ROC-analysis revealed a NLR of 20.9 as the best cut-off threshold on admission to discriminate between increased CBFv within hospital stay. (3) These patients showed an unfavorable outcome (GOS 1-4) at discharge. According to published data 21,22 NLR represents a promising parameter in the differential diagnosis of acute bacterial meningitis predominantly in patients aged 15 and older. As this dataset does not provide CSF counts on lymphocytes and neutrophils we are unable to reproduce these findings.
Cerebral vasospasms represent a devastating complication in acute bacterial meningitis contributing to an unfavorable functional outcome 10,31 . Our results imply that patients with an increased NLR on admission may yield a potential risk to develop an elevated CBFv which contributes to a potentially severe course of disease. Of 108 patients 30 (27.8%) developed increased cerebral blood flow velocity within hospital stay due to infection with S. pneumoniae in 70% in agreement of published data 31,32 . Among these patients ischemic complications were observed in 21.1%. TCD was performed routinely whereas cerebral imaging was performed according to clinical signs and symptoms. As not every patient with increased cerebral blood flow velocity underwent CT/ MRI-imaging subclinical strokes may remained undetected. Further, increased flow velocities do not necessarily cause impending ischemic complications 26 .
In our cohort of patients with community acquired bacterial meningitis we identified a predictive NLR-cut off level to be ≥ 20.9. This cut-off value is increased compared to published data for e.g. prediction of ICH after endovascular thrombectomy in acute ischemic stroke 33 .
Unexpectedly and contrary to published data 34,35 we did not identify dexamethasone administration to be independently associated with increased cerebral blood flow velocity in multivariate analysis. Prior to the wide spread use of adjunctive steroids in bacterial meningitis 36,37 cerebral vascular alterations were devastating but common complications. Reasons may be cerebral vasculitis, septic emboli, intraarterial thrombosis or disseminated intravascular coagulation 35,[38][39][40] . Thus specific anti-inflammatory regimes are desirable 41,42 beyond adjuvant corticosteroid treatment which has been proven beneficial on case fatality rates in adult patients with pneumococcal meningitis but only in high-income countries 37 .
Taken together, increased cerebral blood flow velocity is associated with neuroinflammation and represents a devastating complication in bacterial meningitis. NLR is an easy accessible and cost-effective biomarker that has proven its prognostic value in several conditions among cerebrovascular 14,47 as well as infectious diseases 19,48 and also in the differential diagnosis between viral and bacterial meningitis 21,22 . As a reliable parameter for the general immune response to various stimuli, NLR does not represent an inflammatory marker within a specific entity. Therefore, it remains unclear whether increased NLR-levels on admission reflect the state of bacterial infections or if high NLR values represent an independent surrogate to predict a severe course of the disease www.nature.com/scientificreports/ with intracranial vascular alterations 34 as a devastating complication in bacterial meningitis. Future studies should investigate whether elevated NLR-levels in CABM-patients help to identify patients suitable for immune modulating treatments. Our study has several limitations. First of all the small cohort size comprising only 108 patients weakens our data. It lacks a prospective and multi-center design. Except for admission laboratory data were not available every day within hospital stay. Bloodwork and CSF stains were done as required by clinical practice and not by scheduled timing. Further assessment of CBFv was conducted by different examiners within the inclusion period. As critical care procedures improved within the last years this may impose bias to the reported data. Finally, above findings do not provide any mechanistic evidence confirmed by specific blood sampling analyses or detailed immunoprofiling, and therefore depict a phenomenological finding.

Conclusions
Among CABM-patients intracranial vascular complications are a risk factor for unfavorable outcome at discharge. Elevated NLR is independently associated with high CBFv and may be useful in predicting patients' prognosis. Future studies should investigate whether the impact of NLR represents an independent clinical implication or a preexisting comorbidity.

Data availability
All data generated or analysed during this study are included in this published article.