High Plasma Glutamate Levels are Associated with Poor Functional Outcome in Acute Ischemic Stroke

Abstract

The aim of the present study was to investigate the relationship between acute ischemic stroke and glutamate levels and to determine the prognosis value of plasma glutamate levels to predict the functional outcome. Two hundred and forty-two patients with acute ischemic stroke and 100 sex- and age-matched controls were included in the study. Plasma glutamate levels were determined by HPLC at admission in both groups. Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS). The modified Rankin Scale (mRS) scores at 3 months was determined to outcomes, and unfavorable outcomes were defined as mRS at 3–6. The prognostic value analyzed by logistic regression analysis, after adjusting for the possible confounders. In the 94 patients with an unfavorable functional outcome, plasma glutamate levels were higher compared with those in patients with a favorable outcome [221(IQR, 152–321) μM; 176(IQR, 112–226) μM, respectively; P < 0.0001). In multivariate logistic regression analysis, glutamate was an independent predictor of functional outcome, with an adjusted OR of 6.99 (95 % confidence interval [CI] 2.21–21.23). Receiver operating characteristics to predict functional outcome demonstrated areas under the curve of glutamate of 0.821 (95 % CI 0.733–0.878; P < 0.0001) and combined model (glutamate and NIHSS) improved the NIHSS score alone. Plasma glutamate levels can be seen as an independent short-term prognostic marker of functional outcome in Chinese patients with acute ischemic stroke even after correcting for possible confounding factors.

Introduction

Stroke is the second commonest cause of death and leading cause of adult disability in China (Tu et al. 2014). Early detection and control of risk factors are thought to be crucial in reducing the risk of stroke and providing effective care (Zhang and Zhang 2014). It is well established that during ischemia, glutamate acts as an important mediator of neuronal degeneration, being released in large amounts from neurons and astrocytes, causing cellular overload of calcium, mainly through its action on calcium-permeable NMDA (N-methyl-d-aspartate) receptors (Campos et al. 2011a).

Previous studies have suggested that altered glutamate metabolism is associated with schizophrenia (Bustillo et al. 2014), retinal diseases (Ishikawa 2013), lung cancer (Mohamed et al. 2014), and depression (Cheng et al. 2014). Animal models and human clinical studies reveal the association of pathologically elevated glutamate levels and several acute and chronic neurodegenerative disorders, including stroke (Leibowitz et al. 2012). Clinical studies in healthy volunteers and stroke patients showed a close correlation between plasma glutamate concentrations and cerebrospinal fluid (CSF) concentrations (Leibowitz et al. 2012). Previous studies have demonstrated that neurological deterioration of patients with acute ischemic stroke (AIS) was associated with higher glutamate levels in blood and CSF (Castillo et al. 1996; 1997). However, no significant associations between glutamate concentrations in CSF and stroke characteristics were found in another study (Brouns et al. 2010).

In addition, the decrease of blood glutamate levels leads to a larger glutamate gradient between the brain and blood, facilitating the lowering of extracellular levels of brain glutamate (Gottlieb et al. 2003; Teichberg et al. 2009). Studies in animal models of traumatic brain injury have reported that treatment with recombinant glutamate–oxaloacetate transaminase (GOT) or glutamate–pyruvate transaminase (GPT) (Zlotnik et al. 2007; Boyko et al. 2011) caused a decrease in blood glutamate levels inducing a neuroprotective effect. Human plasma glutamate levels during the first 24 h of stroke were shown to correlate with the volume of ischemic lesion on CT scans or MRI and neurological outcomes (Campos et al. 2011b; Aliprandi et al. 2005). To our knowledge, no studies have evaluated glutamate as a prognostic marker after stroke in Chinese sample. Thus, the aim of this study was to assess the short-term prognostic value of blood glutamate in AIS patients during first attack periods.

Materials and Methods

From May 2012 to December 2013, consecutive first-ever AIS patients admitted to the Department of Emergency of our hospital were identified. All patients were admitted within 24 h of experiencing a new focal or global neurological event. Brain imaging (either CT or MRI) was performed routinely within 24 h after admission. Patients with malignant tumor, intracerebral hemorrhage, head trauma, severe edema, significant acute medical illness (which would be associated with significant inflammatory burden of its own; e.g., infection, autoimmune disease, and cancer), significant acute neurological illness other than stroke, and autoimmune diseases were excluded. The patients taken any medication, which could modify the plasma levels of glutamate (e.g., glutamine and recombinant GOT or GPT), were also excluded.

The control cases (N = 100) were of similar age and gender distribution to the AIS patients. They had no known diseases and were not using any medication. The mean age of controls included in this study was 66 (SD 12) years and 39 % were women. A detailed medical history was taken, and clinical and laboratory examinations were performed on all participants in both groups. The present study has been approved by the ethics committee of the Navy General Hospital. All participants were informed of the study protocol and their written informed consents were obtained, according to the Declaration of Helsinki.

Clinical information was collected. Demographic data (age and sex) and history of risk factors (hypertension, diabetes mellitus, atrial fibrillation, hyperlipidemia, smoking habit, and alcohol abuse) were obtained at admission. Stroke subtype was classified according to TOAST (Trial of Org 10172 in Acute Stroke Treatment) criteria (Adams et al. 1993). The clinical stroke syndrome was determined by applying the criteria of the Oxfordshire Community Stroke Project (Bamford et al. 1991). The National Institutes of Health Stroke Scale (NIHSS) score was assessed on admission (Brott et al. 1989). Functional outcome was obtained according to the modified Rankin Scale (mRS) score (Bonita 1988). The primary end point of this study was favorable functional outcome of stroke patients after 3 months from baseline (mRS 0–2). Secondary end point in stroke patients was death from any cause within a 3-month follow-up. Outcome assessment was performed by one trained medical student with a structured follow-up telephone interview with the patient or, if not possible, with the closest relative.

Brain imaging (either CT or MRI) was performed routinely within 24 h after admission. MRI with diffusion-weighted imaging (DWI) was available in some patients. In those patients, DWI lesion volumes were determined by one experienced neurologist unaware of the clinical and laboratory results. The infarct volume was calculated using the formula 0.5 × a × b × c (where a is the maximal longitudinal diameter, b the maximal transverse diameter perpendicular to a, and c is the number of 10-mm slices containing infarct) (Sims et al. 2009).

Fasting blood was collected via venipuncture in EDTA (ethylenediaminetetraacetic acid) BD Vacutainer ^® (New Jersey, USA) tubes at 7:30 am ± 30 min on the morning after the clinical assessments were conducted (within 0–6 [n = 55], 6–12 [n = 75], 12–24 [n = 60], and 24–72 [n = 52] hours from symptom onset). Blood samples were centrifuged at 1,000×g for 10 min at 4 °C, and plasma was separated and stored at −80 °C until the time of assay. Glutamate levels were determined by HPLC, using the Waters Pico Tag^® Chemistry Package for HPLC amino acids analysis. The intra-and inter-assay coefficients of variation were 1.8–2.7 % and 2.3–3.9 %, respectively. The median value of morning plasma glutamate level in our 100 normal cases is 72 μM, and the 97.5 percentile was 90 μM. The detection limit was 1 μM, and the detection range was 1–2000 μM. Other biochemical measurements were done using standard laboratory methods. All determinations were carried out in a laboratory blind to the clinical outcome and neuroimaging findings.

Results are expressed as percentages for categorical variables and as mean (SD) or median (quartiles) for continuous variables, depending on the normal or non-normal distribution of data. Proportions were compared using the χ ² test, and Student’s t test or the Mann–Whitney test to compare continuous variables between groups. Spearman’s analysis was used for bivariate correlations. The influence of glutamate levels on functional outcome was assessed by logistic regression analysis, after adjusting for the main baseline variables related to outcome in the univariate analyses. Common logarithmic transformation (i.e., log) was performed to obtain normal distribution for skewed variables (i.e., glutamate concentrations). We used crude models and multivariate models adjusted for all significant results were expressed as adjusted OR (odds ratios) with the corresponding 95 % CIs (confidence intervals). Receiver operating characteristic curves (ROC) were used to test the overall prognostic accuracy of glutamate, the NIHSS and other serum biomarkers and results were reported as area under the curve (AUC). To test whether the glutamate levels improve score performance, we considered the nested models with NIHSS and MBL as compared with NIHSS only. All statistical analysis was performed with SPSS for Windows, version 19.0 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined as P < 0.05.

Results

In our study, from 456 screened patients, the study cohort consisted of 305 patients with ischemic stroke at baseline. By the time of hospital follow-up, at 3-month post-stroke, 12.8 % (n = 39) declined the invitation to participate and 7.9 % (n = 24) lost follow-up, leaving 242 individuals. See the Fig. 1. However, these 242 patients were similar in terms of baseline characteristics [age (P = 0.77), gender (P = 0.84), NIHSS (P = 0.67), hospital stays (P = 0.43), and weight (P = 0.79)] compared to the overall cohort. In the study population, 38.8 % were female and the average age was 65.6 ± 12.2 years. The median (quartiles) NIHSS score on admission was 7 (4, 11), and the median time from symptom recognition on admission to hospital was 4.8 h (IQR 2.6–7.9). The number of tissue plasminogen activator-treated patients was 97 (40.1 %). The baseline characteristics of the 242 patients presenting with AIS are described in Table 1.

Fig. 1.

Study profile/flow sheet of the study

Table 1.

Basal characteristics of patients with acute ischemic stroke

Characteristics	All	Good outcome (mRS0-2)	Poor outcome (mRS3-6)	P a
N	242	148	94
Female sex (n)	94	40	25	NS
Median age, yr (SD)	65.6 (12.2)	63.5 (10.1)	70.9 (12.5)	0.006
NIHSS score at admission, (IQR)	7 (4–11)	4 (2–8)	9 (5–15)	0.008
Time to inclusion, hours(IQR)	4.8 (2.6–7.9)	4.7 (2.5–7.6)	4.8 (2.7–8.3)	NS
Hospital stays, (IQR)	26 (16–45)	27 (16–48)	25 (16–44)	NS
Lesion volumes(mL), (n = 199; median, IQR)	29 (18–56)	21 (12–41)	39 (27–79)	0.006
Received thrombolytic therapy(n)	97	84	13	<0.0001
Hypertension, (%)	64.0	67.6	58.5	NS
Diabetes at baseline, (%)	40.0	32.4	51.1	0.004
Hypercholesterolemia, (%)	40.5	40.5	40.4	NS
Atrial fibrillation, (%)	36.0	37.2	34.0	NS
Coronary heart disease, (%)	26.8	27.0	26.6	NS
Family history of stroke, (%)	22.7	22.3	23.4	NS
Cigarette smoking, (%)	21.9	20.3	24.5	NS
Alcohol drinking, (%)	21.5	20.3	23.4	NS
TOAST classification (%)				NS
a. Large artery	25.6	24.3	27.7
b. Small artery	23.1	22.3	24.5
c. Cardioembolism	38.4	40.5	35.1
d. Other cause	6.3	6.8	5.3
e. Unknown	6.6	6.1	7.4
Stroke syndrome (%)				NS
TACS	14.5	13.5	16.0
PACS	18.6	18.9	18.1
LACS	38.0	37.2	39.4
POCS	28.9	30.4	26.5
Laboratory findings (median, IQR)
Glucose, mmol/L	6.4 (5.5–7.5)	6.2 (5.3–7.0)	6.6 (5.6–7.9)	0.043
White cell count, ×109/L	7.4 (5.5–8.4)	7.1 (4.9–8.1)	7.7(5.7–9.0)	0.035
Hs-CRP (mg/dL)	1.44 (0.49–3.29)	0.88 (0.19–2.04)	2.45 (1.02–3.99)	0.001
HCY (μmol/L)	16.5 (10.8–19.2)	13.6(9.6–16.8)	20.2 (12.6–22.5)	0.003
GOT levels (U/L)	18 (13–29)	20 (16–32)	14 (11–22)	<0.0001
GPT levels(U/L)	20 (16–26)	21 (17–29)	18 (15–24)	0.013
Glutamate levels (μM)	187 (128–267)	176 (112–226)	221 (152–321)	<0.0001

IQR interquartile range, NIHSS National Institutes of Health Stroke Scale, TACS total anterior circulation syndrome, LACS lacunar syndrome, PACS partial anterior circulation syndrome, POCS posterior circulation syndrome, Hs-CRP high-sensitivity-C respond protein, HCY homocysteine, GOT glutamate-oxaloacetate transaminase, GPT glutamate-pyruvate transaminase, NS not significant

^a P value was assessed using Mann–Whitney U test or Chi-Square test

The results indicated that the plasma glutamate levels were significantly (P < 0.0001) higher in acutely ischemic stroke patients as compared to healthy controls [187 (IQR, 128–267) μM; 72 (IQR, 64–78) μM, respectively). The plasma glutamate levels were not significantly different between patients receiving thrombolytic therapy and those without treatment [183 (IQR, 119–262) μM; 192 (IQR, 132–273) μM, respectively; P = 0.124). Plasma glutamate levels increased with the increasing severity of stroke as defined by the NIHSS score. There was a positive correlation between levels of plasma glutamate and NIHSS score (r [spearman] = 0.205, P = 0.001). See the Fig. 2a. There was a modest correlation between levels of plasma glutamate and Hs-CRP (r = 0.152, P = 0.031). There was no correlation between levels of plasma glutamate and sex, age, white blood cells count, homocysteine, and glucose (P > 0.05, respectively). Interestingly, GOT and GPT levels were negatively correlated with plasma levels of glutamate at admission (r = −0.264, P < 0.0001, see the Fig. 2b; r = −0.193, P = 0.003, see the Fig. 2c). This correlation was higher for GOT levels than for GPT levels. In addition, in patients for whom MRI data were available (n = 199), there was a positive correlation between levels of glutamate and the infarct volume (r = 0.579, P < 0.0001; Fig. 2d).

Fig. 2.

The correlation between plasma glutamate levels and other predictors. a Correlation between the National Institutes of Health Stroke Scale (NIHSS) and plasma glutamate levels; b Correlation between the Glutamate-oxaloacetate transaminase (GOT) and plasma glutamate levels. c Correlation between the glutamate-pyruvate transaminase(GPT) and plasma glutamate levels. d Correlation between the infract volumes and plasma glutamate levels

In the 94 patients with an unfavorable functional outcome, plasma glutamate levels were higher compared with those in patients with a favorable outcome [221 (IQR, 152–321) μM; 176 (IQR, 112–226) μM, respectively; P < 0.0001). See the Fig. 2. In univariate logistic regression analysis, we calculated the odds ratio (OR) of log-transformed glutamate levels as compared with the NIHSS score and other risk factors as presented in Table 2. With an unadjusted OR of 13.62 (95 % CI 3.53–65.49), glutamate had a strong association with unfavorable functional outcome. After adjusting for all other significant outcome predictors, glutamate remained an independent outcome predictor with an adjusted OR of 6.99 (95 % CI 2.21–21.23). In the subgroup of patients (n = 199) in whom MRI evaluations were performed, glutamate was an independent unfavorable outcome predictor with an OR of 8.12 (95 % CI 1.88–30.52; P < 0.0001), after adjustment for both lesion size and the NIHSS score. In addition, age, thrombolytic therapy, the NIHSS score, and laboratory findings, such as homocysteine level and Hs-CRP remained significant outcome predictors (Table 2).

Table 2.

Unadjusted and adjusted analysis for outcomes

Parameter	Unadjusted analysis			Adjusted analysis
	ORa	95 % CIa	P	ORa	95 % CIa	P
Glutamateb	13.62	3.53–65.49	<0.0001	6.69	2.11–21.23	<0.0001
Age	1.20	1.05–1.87	0.004	1.07	1.01–1.24	0.001
NIHSS	1.48	1.30–1.86	<0.001	1.31	1.21–1.40	<0.001
Infarct volume	1.22	1.11–1.35	0.003	1.15	1.06–1.29	0.009
Glucose	1.09	1.02–1.34	0.037	1.06	1.01–1.42	0.044
Homocysteine	1.44	1.28–1.67	<0.001	1.30	1.21–1.40	<0.001
Hs-CRP	1.22	1.11–1.38	0.003	1.11	1.01–1.13	0.012
Thrombolytic therapy	0.13	0.07–0.25	<0.0001	0.14	0.08–0.25	<0.0001

OR odds ratio, CI confidence interval, Hs-CRP high-sensitivity-C-reactive protein, NIHSS National Institutes of Health Stroke Scale

^aNote that the odds ratio corresponds to a unit increase in the explanatory variable

^bLog-transformed to achieve normal distribution

With an AUC of 0.821 (95 % CI 0.733–0.878), glutamate showed a significantly greater discriminatory ability as compared with Hs-CRP (AUC 0.649; 95 % CI 0.573–0.726; P = 0.001) and age (AUC 0.562; 95 % CI 0.511–0.644; P = 0.007), while was in the range of NIHSS score (AUC, 0.838; 95 % CI 0.738–0.899; P < 0.0001). Interestingly, combined model (glutamate and NIHSS) improved the NIHSS score alone (AUC of the combined model, 0.889; 95 % CI 0.804–0.943; P < 0.0001). This improvement was stable in an internal 5-fold cross-validation that resulted in an average AUC (standard error) of 0.84 (0.028) for the NIHSS and 0.89 (0.017) for the combined model, corresponding to a difference of 0.05 (0.011).

Discussion

Largely in accord with previous findings (Castillo et al. 1996, 1997), our data showed that plasma glutamate levels were higher in stroke patients compared with normal cases. Although majority of results are confirmatory data (Campos et al. 2011a; Castillo et al. 1996, 1997; Campos et al. 2011b), our study was well designed and showed an important role of glutamate in the prognosis of patients with AIS in Chinese samples. In this study, we firstly assessed plasma glutamate levels with regard to their accuracy to predict the short-term functional outcomes in patients with AIS in Chinese sample. Our main finding was that glutamate can be seen as an independent short-term prognostic marker of functional outcomes in Chinese sample with AIS even after correcting for possible confounding factors, and plasma glutamate levels increased per log unit were associated with a 6.99-fold increase in unfavorable outcome. Similarly, Campos et al. (2011b) found that there was an association between high blood glutamate levels and poor outcomes in ischemic stroke patients. Besides, this study confirmed the possibility of GOT or GPT enzymes as possible therapeutic targets in AIS.

We also found that plasma glutamate levels increased with infarct volume and neurological deficit (assessed by the NIHSS). Consistent with our results, previous studies have shown that high glutamate levels in blood and in CSF are associated with large infarct volume and greater stroke severity (Castillo et al. 1996), higher frequency of early neurological deterioration (Campos et al. 2011a) and infarct growth (Castellanos et al. 2008), which leads to poor functional outcome. Since glutamate plays a central role in the ischemic cascade, this neurotransmitter represents a good target in the search for neuroprotective agents in ischemic stroke. Blood glutamate scavenging may serve as a new neuroprotective strategy for the treatment of ischemic stroke (Boyko et al. 2011). Thus, it may open the way to the proposal of new therapeutic options in patients with ischemic stroke.

Interestingly, GOT and GPT are two enzymes that metabolize glutamate levels in the peripheral blood, and the decrease of blood glutamate levels facilitates the lowering of extracellular levels of brain glutamate (Teichberg et al. 2009). In our study, we also found that GOT and GPT levels were negatively correlated with plasma levels of glutamate at admission. In addition, the role of neuroprotection by GOT in ischemic stroke had been suggested (Campos et al. 2011b, c). Therefore, the properties of these enzymes to metabolize glutamate in blood show that they could be used as new neuroprotective tools against excitotoxic neuronal injury after ischemic stroke. Pérez-Mato et al. (2014) found that a reduction in serum and brain glutamate levels, resulting in a reduction in infarct volume and sensorimotor deficit, suggesting that the combination of human rGOT1 with low doses of oxaloacetate seems to be a successful approach for stroke treatment. More work should be done before it can be conducted.

Whether the high plasma glutamate is just an epiphenomenon to stroke severity or independently contributes to prognosis remains uncertain. A severe stroke per se implicates a poor outcome. However, there are several other reasons explaining unfavorable outcome in patients with higher plasma glutamate levels. Glutamate excitotoxicity is one cause of post-ischemic neuronal death (Sims et al. 2009). Glutamine synthetase (GS) is an enzyme that is expressed in glial cells and may affect glutamate excitotoxicity. Second, in the brain, glutamine synthetase (GS) rapidly converts blood-borne ammonia into glutamine which in high concentrations may cause mitochondrial dysfunction and osmolytic brain edema (Fries et al. 2014). Third, Gómez-Galán et al. (2012) suggested that the dysfunctional astrocytic glutamate reuptake triggers a succession of events, including the reduction of d-serine production as a safety mechanism to avoid NMDA receptor overactivation, which in turn causes the synaptic plasticity and memory impairments observed. Forth, many types of glutamate receptors or transporters appear to be involved in the etiology of stroke. Activation of kainate (KA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptors (GluRs) increase inflammatory cytokines release (Bonnet et al. 2013). Microglia activated by excess inflammation, astroglial loss, and inappropriate glutamate receptor activation ultimately disrupts the delicate balance of neuroprotective versus neurotoxic effects in the brain. Similarly, in our study, we found a modest correlation between the levels of plasma glutamate and Hs-CRP. Lastly, glutamate is secreted from platelets upon activation together with other bioactive substances that can all together contribute to poor outcomes.

Some limitations of this observational study should be considered. Firstly, without serial measurement of the circulating glutamate levels, this study yielded no data regarding when and how long glutamate is elevated in these patients. A three-fold increase of plasma glutamate compared with baseline values has been shown to begin 4–6 h after ischemic induction, reaching peak values at 8–24 h and returning to pre-ischemic values by 48–72 h (Leibowitz et al. 2012). Secondly, measurements were performed after the stroke, so they may not accurately reflect pre-stroke exposure. Thirdly, the relatively small sample size may limit the generalization of the results of this study. Before broad implementation, additional studies are needed for external validation. In addition, no power calculation for the sample size was made, and this fact raises serious concerns that interpretations were being skewed by artifact. Lastly, the effects of circulating glutamate on long-term clinical outcome were not included in the study protocol. Besides, infarct volume based on the formula for hematoma volumetry (0.5 × A × B × C) in our study protocol was suboptimal.

Conclusions

In conclusion, the results of the present study showed a good association between high plasma levels of glutamate with poor outcome in AIS patients, although further studies were necessary to demonstrate if glutamate really can be used as therapeutic tools for ischemic stroke.