Blood Glutamate Levels Are Closely Related to Acute Lung Injury and Prognosis after Stroke

Background Acute lung injury (ALI) is a serious complication of stroke that occurs with a high incidence. Our preclinical results indicated that ALI might be related to blood glutamate levels after brain injury. The purpose of this study was to assess dynamic changes in blood glutamate levels in patients with stroke and to determine the correlation between blood glutamate levels, ALI, and long-term prognosis after stroke. Methods Venous blood samples were collected from controls and patients with stroke at admission and on the third and seventh day after the onset of stroke. Patients were followed for 3 months. The correlations among blood glutamate levels, severities of stroke and ALI, and long-term outcomes were analyzed, and the predictive values of blood glutamate levels and severity scores for ALI were assessed. Results In this study, a total of 384 patients with stroke were enrolled, with a median age of 59 years. Patients showed significantly increased blood glutamate levels within 7 days of stroke onset (p < 0.05), and patients with more severe injuries showed higher blood glutamate levels. Moreover, blood glutamate levels were closely related to the occurrence (adjusted odds ratio, 3.022, p = 0.003) and severity (p < 0.001) of ALI and the long-term prognosis after stroke (p < 0.05), and they were a more accurate predictor of ALI than the more commonly used severity scores (p < 0.01). Conclusion These results indicated that an increased blood glutamate level was closely related to the development of ALI and a poor prognosis after stroke. Clinical Trial Registration http://www.chictr.org.cn, identifier ChiCTR-RPC-15006770.

after brain injury have been identified, including intracranial (e.g., midline shift and increased intracranial pressure) and extracranial (e.g., administration of vasoactive drugs and a history of drug abuse) factors; however, these factors have either been nonspecific or difficult to measure, and an accurate, simple, and stable marker for ALI following brain injury is urgently needed (2,3,5). Stroke, including primarily acute ischemic stroke (AIS), intracerebral hemorrhage (ICH), and spontaneous subarachnoid hemorrhage (SAH), is a serious threat to human health and life (6). After the occurrence of a stroke, elevated glutamate-induced excitotoxicity in the brain is known to severely affect patient prognosis (7,8). Nevertheless, the role of glutamate levels in the blood has not been well characterized. Glutamate is a non-essential amino acid that participates in a series of basic metabolic reactions, such as deamination and gluconeogenesis (9). In addition, under physiological conditions, the blood levels of glutamate remain relatively stable, and a normal diet prevents significant fluctuations in blood glutamate levels (10); however, certain diets [e.g., rich in monosodium glutamate (MSG)] will cause a significant increase in blood glutamate levels (11). Moreover, researchers have recently found that blood glutamate levels are greatly increased in many patients who experience stroke (12,13). Our previous animal experiments have indicated that elevated blood glutamate levels might be related to the development of ALI after brain injury (14), although there was no evidence of a close relationship in a large scale clinical study in patients with stroke. Furthermore, since these subtypes of stroke were very different in terms of their pathophysiology, evaluation of whether this is a shared phenomenon between ischemic and hemorrhagic stroke is necessary.
The aim of this study was to evaluate the relationship between blood glutamate levels and the incidence of ALI in patients with stroke, the ability of blood glutamate levels to predict the occurrence of ALI and the relationship between glutamate levels and the long-term prognosis following stroke. The results of this study increase our understanding of this common complication following brain insult.

Patients
Hospitalized patients were consecutively enrolled from Daping Hospital (Chongqing, China), which has over 2,500 beds and is ranked among the top 100 hospitals in China. It is also a teaching hospital affiliated with the Third Military Medical University and serves as a base for stroke screening and prevention in China. We enrolled patients admitted to the hospital within 6 h of presenting with clinical stroke symptoms and who received a diagnosis of ischemic or hemorrhagic stroke. A control group included 120 subjects who received a routine physical examination, were not diagnosed with stroke, and had no history of stroke/ transient ischemic attack (TIA), hypertension, hyperlipidemia, hypercholesterolemia, diabetes mellitus, myocardial infarction, or atrial fibrillation. All participants were Han Chinese living in Chongqing and the surrounding provinces with an age ≥18 years. The diagnosis of stroke was confirmed in all patients by computed tomography (CT) or magnetic resonance imaging. The key exclusion criteria were as follows: transferred from other hospitals or departments; the presence of severe hepatic disease [e.g., aspirate transaminase (AST) levels >40 U/L or alanine aminotransferase (ALT) levels >40 U/L], current or previous lung disease (e.g., tuberculosis, pneumonia, or chronic obstructive pulmonary disease), or abnormal clinical manifestations (e.g., dyspnea, shortness of breath, or cyanosis), and failure of one or more organs. Stroke severity was quantified by a neurologist according to the National Institutes of Health Stroke Scale (NIHSS) score, ICH score, or World Federation of Neurosurgical Surgeons (WFNS) scale at admission and after treatment for patients with AIS, ICH, or SAH, respectively; based on the scores at admission, the patients were divided into mild (NIHSS score: ≤8; ICH score: ≤2; and WFNS scale: I), moderate (NIHSS score: 9-15; ICH score: 3-4; and WFNS scale: II-III), and severe (NIHSS score: ≥16; ICH score: 5-6; and WFNS scale: IV-V) groups (15)(16)(17)(18). The detailed sociodemographic and baseline characteristics of patients and controls are listed in Table 1.
This study was performed in accordance with the recommendations of international ethical guidelines for biomedical research involving human subjects, and the ethics committee of the Research Institute of Surgery and Daping Hospital approved this study (No. 2015-12). All participants (or legal guardians) in this study provided written informed consent in accordance with the Declaration of Helsinki and its later amendments. This clinical trial was registered at the Chinese Clinical Trial Registry (www.chictr.org.cn; unique identifier: ChiCTR-RPC-15006770).

Diagnosis and classification of ali
According to the Berlin Definition for ALI/ARDS (19), after admission, stroke patients with ALI would quickly present (usually within 72 h) with new or worsening respiratory symptoms (e.g., shortness of breath and dyspnea), an acute onset of hypoxemia (PaO2/FiO2 ≤ 300 mm Hg), and a radiographic examination (representative images, e.g., chest X-ray or CT, are shown in Figure S1 in Supplementary Material) indicating high-density patchy shadows or edema not fully explained by effusion, lobar/ lung collapse, or nodules; the respiratory failure symptoms could not be fully explained by cardiac failure, which was excluded by echocardiographic measurements (20,21). In addition, based on the degree of hypoxemia, the severity of ALI could be classified as mild (200 mm Hg < PaO2/FiO2 ≤ 300 mm Hg), moderate (100 mm Hg < PaO2/FiO2 ≤ 200 mm Hg), or severe (PaO2/ FiO2 ≤ 100 mm Hg) (19,20).
clinical Management of Patients with stroke Treatment protocols were implemented by a single attending neurologist. For patients with AIS, intravenous thrombolysis (IVT) or endovascular treatment (ET) was performed after rigorous assessment, while for patients with ICH or SAH, surgical (decompressive craniectomy) or conservative treatment (e.g., reduction of intracranial pressure via mannitol) was used accordingly. In addition, at the moment the diagnosis of ALI was made, the patients were transferred to the ICU, and treatment was based primarily on ventilation with positive end-expiratory pressure (PEEP), following the ARDSnet protocol (22,23).

Blood collection, Processing, and glutamate assay
Venous blood samples were collected from the patients at the time of admission (without any treatment and within 6 h from stroke onset) and on the mornings of the third day and seventh day. Blood was obtained from controls during routine physical examination. All samples were anticoagulated with heparin; then, using standard clinical laboratory methods, the samples were evaluated for inflammatory markers, such as IL-6, C-reactive protein, and procalcitonin, and routine markers, including hepatic function markers, such as ALT and AST, and nerve injury markers, such as neuron-specific enolase (NSE) and S-100B. Blood glutamate levels were assayed using high-performance liquid chromatography with spectrofluorometric detection (Beckman Coulter, CA, USA) at a fixed excitation wavelength of 330 nm and a fixed emission wavelength of 450 nm, according to our previously published methodology (24).

Outcome Measurement
Mortality data were collected over a follow-up period of 3 months. The long-term outcomes were evaluated by modified Rankin scale (mRS) at 90 days by phone interview by an experienced neurologist. Favorable outcomes and functional independence were defined as an mRS score of 0-2 (25).

statistical analysis
The justification of sample sizes in each group was performed as previously described (26) (27). Areas under receiver-operating characteristic curves were utilized to evaluate and compare the accuracy of blood glutamate levels and severity scores (NIHSS score, ICH score, and WFNS scale) for predicting the occurrence of ALI (28). In addition, the possibility of defining a "cutoff value" of blood glutamate level associated with survival and poor functional outcome was explored using Kaplan-Meier survival curve analysis, followed by Log Rank testing and multivariate logistic regression, respectively. Two-tailed p values <0.05 represent statistically significant differences. All statistical analyses were performed using Sigma Plot (version 12.5; Systat Software Inc., San Jose, CA, USA).

resUlTs
Between August 2015 and July 2016, a total of 450 patients were consecutively screened in our institution's Department of Neurology and Department of ICU. Of these, 49 patients were excluded, and 17 patients discharged or died within 7 days. Ultimately, 384 patients were included and analyzed in this study, including 132 patients with AIS, 124 patients with ICH, and 128 patients with SAH (Figure 1).   stroke, there were relatively few changes in glutamate levels within 7 days of admission (Figure 2A). We further found that patients with severe or moderate stroke exhibited significantly increased blood glutamate levels; these levels were higher with increased stroke severity (Figures 2B-D).
Blood glutamate levels but not severity scores Were closely related to the incidence and severity of ali For all of the stroke patients, the median severity scores of patients who developed ALI were much higher than those without ALI (Figures 3A-C). However, logistic regression analysis showed that these were not closely related with the occurrence of ALI when adjusted for confounding factors (

Blood glutamate level analysis as a sensitive and specific Method for Predicting ali
To obtain an early and accurate diagnosis of ALI, we next generated receiver-operating characteristic curves to assess the predictive accuracy of both blood glutamate levels and severity scores at admission. For all the patients with stroke, the optimal cutoff value for blood glutamate as an indicator of ALI was projected to be 87.6 µmol/L, which yielded an area under the curve (AUC) of 0.931 (95% CI: 0.905-0.957), with a sensitivity of 0.878 and a specificity of 0.892 (Figure 4A). The likelihood ratio test showed a significant increase in the predictive value of blood glutamate levels over severity scores (p = 1.665E−04, p = 0.002, and p = 0.009 for patients with AIS, ICH, and SAH, respectively). Moreover, the combination of glutamate levels and severity scores resulted in AUCs that were superior to severity scores but not superior to blood glutamate levels alone for predicting ALI (Figures 4B-D).

association between Blood glutamate levels and long-term clinical Outcomes
Thirty-one patients (8.7%) were lost to follow-up during the 90 days following discharge from the hospital. The analysis of blood glutamate level cutoff suggested that the survival rate after a stroke (within 7 days), increased disease severity was associated with higher levels of blood glutamate. Patients with AIS exhibited higher levels of blood glutamate than patients with ICH or SAH; the proportion of patients in the AIS group who were diagnosed with ALI was also much higher than that in the ICH group and the SAH group. We speculate that this was partly because the severity of patients with AIS that we enrolled in this study was higher than that previously reported (12). Patients with ALI showed higher levels of brain injury markers (NSE and S-100B) ( Table S1 in Supplementary Material) and severity scores than patients without ALI. Nevertheless, we found that blood glutamate levels but not the severity scores of stroke patients at admission could serve as a risk factor for the development of ALI, which was demonstrated by regression analysis after controlling for many potential confounding factors (including age, sex, BMI, and OTT, current smoking status, MSG dietary preference, medical history, treatment, and inflammatory markers). It was noteworthy that patients with a history of myocardial infarction showed a marginal association (p = 0.050) with ALI, partly because these patients always showed other risk factors for stroke (29). The severity scores were always influenced by these confounding factors, was significantly lower for patients with high blood glutamate levels (AIS, p = 0.039; ICH, p = 1.325E−05; and SAH, p = 0.008) (Figures 5B-D). The distributions of mRS scores at 90 days in patients with the three stroke subtypes are shown in Figure 5A. For all three subtypes, after adjustment for potential confounders, patients with high blood glutamate levels were less likely to report functional independence [AIS (43.9 versus 60.0%;

DiscUssiOn
In this preliminary clinical study, we found that patients who experienced stroke (both hemorrhagic and ischemic) exhibited significantly increased blood glutamate levels; these levels could rise for approximately 7 days or more, according to previous work (12). In addition, we found that during hospitalization especially the OTT (30,31), which may partly account for these discrepancies.
As the clinical manifestations of ALI lack specificity and are often overshadowed by primary disease, and because ALI typically develops rapidly (32), there is an urgent need to identify an objective and easily measurable parameter that can predict the occurrence of ALI. Several scores at admission have been commonly used to evaluate the severity and predict prognosis after brain injury (33)(34)(35); however, we found that blood glutamate levels were more accurate and reliable for predicting the occurrence of ALI and were closely related to the long-term prognosis. In addition, we found blood glutamate levels were always much higher in patients with ALI than in patients without ALI; in patients with ALI, blood glutamate levels gradually decreased over the length of the hospital stay, whereas few changes in blood glutamate levels were observed in patients without ALI ( Figure S2B in Supplementary Material). These results collectively demonstrate the potential of using blood glutamate levels in predicting ALI after stroke and indicate that although the pathogenesis is different, the predictive value of blood glutamate levels in different types of stroke is worthy of further investigation.
The inflammatory reaction is an important facet of the occurrence and development of ALI (21). In this study, we also measured the levels of inflammatory markers in blood; we found that patients with ALI had much higher levels of inflammatory markers than patients without ALI (Table S1 in Supplementary Material), and additional analysis showed a positive correlation between blood glutamate levels and IL-6 levels during hospitalization ( Figure S3 in Supplementary Material). After brain injury, the activation of the sympathetic nervous system and the hypothalamic-pituitary-adrenal axis has been shown to be associated with the development of systemic inflammatory complications (36,37); however, it is noteworthy that in patients without ALI but who were diagnosed with a complicated chest infection, even though inflammatory mediator levels in these patients were similar to those in patients with ALI, the mean glutamate levels in these patients were much lower ( Figure S2A in Supplementary Material). In addition, our previous animal experiments demonstrated that high concentrations of blood glutamate mediate interactions between the adenosine A2A receptor and the metabotropic glutamate receptor 5 on neutrophils, thereby inducing the release of inflammatory mediators (14). These results indicate not only that the elevated blood glutamate and inflammatory marker levels were not simply a concomitant phenomenon but also that blood glutamate might play a precipitating role in the occurrence of ALI, and this role may be related to the pro-inflammatory effects of glutamate. However, inflammatory cells, including neutrophils (38) and platelets (12,39), also release glutamate into the blood after brain injury, thereby forming a mutually amplifying system that strengthens the effects of blood glutamate. These hypotheses and the specific mechanism underlying this phenomenon require further validation in animal studies. Nevertheless, as several studies have reported that hemodialysis and peritoneal dialysis are effective ways of lowering blood glutamate levels (40,41), these might serve as a potential and effective measure for reducing blood glutamate levels to prevent or treat ALI after stroke.
This study has several limitations. First, this was a singlecenter observational study with a relatively small sample size. Although the interference of several known factors was controlled for statistically based on logistic regression, such as age, sex, BMI, OTT, current smoking status, MSG dietary preference, medical history, treatment, and inflammatory markers, which have been previously shown to be a risk factor for stroke (29), a potential modulating factor for blood glutamate levels (10), or disease progression (42), it is possible that hidden confounding factors exist, and our results should be interpreted carefully. Therefore, additional larger prospective studies would strengthen our conclusions. Moreover, patients were recruited over a relatively long period of time, which may have introduced a diagnostic bias for ALI based on the individual experiences of different clinicians. Finally, on the first day, only one blood sample but multiple severity scores were often obtained; therefore, we analyzed the relationship between these variables only at admission. Besides, in this study, we found acute treatment could obviously alleviate the severity of brain injury (especially in patients with AIS), which indicated the possible effects of acute treatment in the levels of blood glutamate and the occurrence of ALI. However, the analysis of additional blood samples taken at different time points through one-to-one correlation analyses and a strict randomized control trial in a future study was needed to test and confirm these effects.

cOnclUsiOn
This study further confirms the results in our previous animal studies and demonstrates that blood glutamate levels are closely related to the incidence of ALI in three subtypes of stroke patients. Moreover, our findings suggest that high blood glutamate levels are a potential effective predictor for ALI and a risk factor for the development of ALI and poor prognosis after stroke.

eThics sTaTeMenT
This study was performed in accordance with the recommendations of international ethical guidelines for biomedical research involving human subjects, and the ethics committee of the Research Institute of Surgery and Daping Hospital approved this study (No. 2015-12). All participants (or legal guardians) in this study provided written informed consent in accordance with the Declaration of Helsinki and its later amendments. aUThOr cOnTriBUTiOns WB, WL, Y-LN, and Y-GZ designed the research and drafted the manuscript. PL and YZ analyzed the data. WB, NY, and Y-LJ performed the experiments. Z-PL, D-PJ, YW, and MZ helped perform the experiments. Y-GZ revised the manuscript and contributed reagents, materials, and analysis tools. All the authors read and approved the final manuscript.