Intra- and extra-hospital improvement in ischemic stroke patients: influence of reperfusion therapy and molecular mechanisms

Neuroprotective treatments in ischemic stroke are focused to reduce the pernicious effect of excitotoxicity, oxidative stress and inflammation. However, those cellular and molecular mechanisms may also have beneficial effects, especially during the late stages of the ischemic stroke. The objective of this study was to investigate the relationship between the clinical improvement of ischemic stroke patients and the time-dependent excitotoxicity and inflammation. We included 4295 ischemic stroke patients in a retrospective study. The main outcomes were intra and extra-hospital improvement. High glutamate and IL-6 levels at 24 hours were associated with a worse intra-hospital improvement (OR:0.993, 95%CI: 0.990–0.996 and OR:0.990, 95%CI: 0.985–0.995). High glutamate and IL-6 levels at 24 hours were associated with better extra-hospital improvement (OR:1.13 95%CI, 1.07–1.12 and OR:1.14, 95%CI, 1.09–1.18). Effective reperfusion after recanalization showed the best clinical outcome. However, the long term recovery is less marked in patients with an effective reperfusion. The variations of glutamate and IL6 levels in the first 24 hours clearly showed a relationship between the molecular components of the ischemic cascade and the clinical outcome of patients. Our findings suggest that the rapid reperfusion after recanalization treatment blocks the molecular response to ischemia that is associated with restorative processes.

The ischemic cascade is a complex molecular and cellular process that is triggered immediately after an ischemic stroke (IS). During this process, several mechanisms are developed sequentially as a consequence of the deprivation of blood and oxygen supply to the ischemic core [1][2][3][4][5][6][7] . Neuroprotective treatments in acute IS are mainly focused to reduce the pernicious effect of excitotoxicity, oxidative-nitrosative stress, and inflammation 7,8 , however, these cellular and molecular mechanisms could also have reparative and beneficial effects for the patients. The reduction of glutamate concentration in the synaptic cleft has been the target of several clinical trials using antagonists of glutamate N-methyl-D-aspartate receptors to alleviate excitotoxicity, however those trials have failed to show clinical efficacy 8,9 . Targeting glutamate receptors can also block the synaptic transmission, and even though it could protect from glutamate excitotoxicity, it could also hinder the promotion of neuronal survival and cerebral plasticity in the late phase of IS 9 . Inflammation can also amplify the ischemic lesion and a considerable number of clinical trials were designed to target the inflammatory response 10 . Nonetheless, it is now clear that inflammation also promotes critical events necessary for tissue repair during the late stages of the IS 11 .
Recanalization treatments aimed to achieve reperfusion, either with the tissue Plasminogen Activator (tPA) or by mechanical thrombectomy, are the only treatments that have shown clinical benefit in IS, with a better clinical outcome observed in patients treated with recanalization therapies compared with untreated patients,

Results
The IS patients included in the analysis were classified into 3 groups: patients who did not receive recanalization treatment (PNR); patients with recanalization treatment and effective reperfusion (PER) and patients with recanalization treatment and uneffective reperfusion (PWER). The results are summarized in Table 1, including clinical and analytical variables.
The selection criteria of patients for recanalization partially determined the statistical differences between the groups. In this regard, PER showed the highest clinical improvement, and the benefit persists during the first three months after the treatment (mRS at 3 months: 0[0,1] for PER vs 2 [1,3] for PNR and 1[0,3] for PWER. p < 0.0001).
The type of recanalization treatment did not influence the intra-hospital improvement (Table S1, p = 0.064), but intra-hospital improvement was more frequent in PER compared with PWER (10.9% vs 9.2%, p < 0.0001).
The data was also analyzed in a logistic regression model (Table S1), using the intra-hospital improvement as the dependent variable and the relationship of effective reperfusion therapy with intra-hospital improvement was confirmed (OR: 38.13, 95%CI: 5.26-276.64; p < 0.0001). Likewise, an ineffective reperfusion therapy was associated with a negative intra-hospital evolution (OR: 0.26, 95%CI: 0.19-0.35; p < 0.0001).
Extra-hospital improvement. Extra-hospital improvement was defined as a positive value of the NIHSS scale at discharge minus the value of the NIHSS scale at three months. Extra-hospital improvement was analyzed in 3853 patients (Table 3). Among those patients, 1326 (34.4%) did not show extra-hospital improvement while 2527 (65.6%) showed an improvement of at least 1 point in the NIHSS 3 ± 1 months after the hospital discharge. A bivariate analysis was carried out classifying the patients into two groups depending on the presence or absence of extra-hospital improvement (Table 3), and it was observed that the extra-hospital improvement was not related with a good outcome at 3 months, defined as mRS values ≤2 (64.9% vs 60.5%, p = 0.051). Extra-hospital improvement was associated with higher values of inflammatory markers (leukocytes 8.7 ± 3.1×10 3 mL vs 9.4 ± 3.1 × 10 3 mL, p = 0.001; fibrinogen 437.6 ± 94.9 mg/dL vs 445.3 ± 98.6 mg/dL, p = 0.002; and sedimentation rate 23.2 ± 23.6 mm vs 26.5 ± 22.8 mm, p < 0.0001). Also, extra-hospital improvement was higher in patients with more serious infarcts (infarct volume 40.8 ± 69.1 mL vs 11.2 ± 31.3 mL, p < 0.0001). The type of reperfusion treatment did not influence the extra-hospital improvement, however, it was less frequent in PER (2.7%) compared with PNR (79.7%) and with PWER (17.6%).
A logistic regression analysis was carried out using the extra-hospital improvement as a dependent variable ( www.nature.com/scientificreports www.nature.com/scientificreports/ Molecular mechanisms (excitotoxicity and inflammation) associated with intra and extra-hospital improvement. In order to evaluate the influence of excitotoxicity and inflammation on the intra and extra-hospital improvement, an analysis of glutamate and IL6 levels was carried out in patients on admission and after 24 h (Table S3). Serum glutamate levels decreased significantly in the first 24 hours for the three groups of patients, although this decrease was more pronounced in PER (Fig. 1). The difference between the glutamate levels on admission in the three groups of patients is attributed to the selection criteria of patients for reperfusion treatments, the onset-inclusion time and the severity of the neurological deficit (both listed in Table S3), demonstrated by the correlations between the NIHSS and the glutamate levels on admission in the three groups of patients.
Serum IL6 levels on admission were similar for the three groups of patients, however, IL6 levels were significantly higher in PNR and PWER at 24 hours compared with PER. Moreover, only PER showed a decrease in IL6 levels after 24 h.
Surprisingly, and opposed to the trend observed for intra-hospital improvement, the extra-hospital improvement (Table S5)  The association of glutamate and IL6 levels with intra and extra-hospital improvement was further confirmed using a multiple linear regression analysis (Table S6). Glutamate levels on admission were correlated with positive intra and extrahospital improvement (β:0.025, p = 0.017 and β:0.019, p = 0.002 respectively), however, there was a negative correlation of high glutamate levels at 24 hours and intrahospital improvement (β −0.042, p = 0.046). Regarding IL6, high levels of this marker on admission had a negative correlation with extrahospital improvement (β −0.305, p < 0.0001), while high levels of IL6 at 24 hours correlates with a negative intrahospital improvement and a positive extrahospital improvement (β: −0.121, p < 0.0001 vs β: 0.131, p < 0.0001). These analyses showed the association between glutamate and IL6 and are further discussed below.

Discussion
To the best of our knowledge, this is the first study aimed to analyse the early (intra-hospital) and late (extra-hospital) evolutionary outcome of IS patients in relation to the recanalization treatment.
Recanalization treatments have demonstrated high survival rates and good outcome in clinical trials [13][14][15] . In our study, we also observed the best clinical outcome in patients with effective reperfusion as a response to the most appropriate recanalizing treatment, both at the time of hospital discharge and after 3 months. Indeed, the multivariate analysis demonstrates that early improvement is 38 times more frequent in patients with effective reperfusion compared with patients without recanalization treatment. Patients who received an uneffective recanalization treatment were the ones with the worst clinical situation. These results are in agreement with other clinical studies, as it is now recognized that effective reperfusion is an accurate indicator for clinical outcome in IS patients 24,25 .
The main factors independently associated with a negative intra-hospital evolution are hyperthermia, infarct volume and especially the uneffective reperfusion. However, effective reperfusion is related to worst extra-hospital improvement. Logically, the extra-hospital evolution is greater in patients who present a higher NIHSS at discharge, and therefore, the margin for improvement is lower for PER (Fig. 2). However, we hypothesize that the effective early reperfusion could also hinder neuroprotective molecular mechanisms. Although this hypothesis cannot be demonstrated in this study, the variations in glutamate and IL6 levels in the first 24 hours clearly show a relationship between the molecular components of the ischemic cascade and the clinical outcome of patients after 3 months, as discussed next.
Serum glutamate levels on admission are related to the stroke severity, and high levels of glutamate at and 24 h after admission were independently associated with a worse intra-hospital evolution. This is in agreement with previous studies 26,27 , although the clinical significance of these findings is rather low taking into account the nature of these variables and its units. Surprisingly, we observed an association between high glutamate levels at 24 hours and better extra-hospital evolution, and we determined that high glutamate levels at 24 hours are independently associated with an 11% better extra-hospital improvement, with both statistical and clinical significance. These observations suggested that the intense reduction of glutamate in the first 24 hours may be a factor that leads to a lower late recovery after an effective reperfusion (Fig. 3).
The inflammatory response has also been associated with early clinical deterioration 28,29 , and a worse prognosis after IS 30 . In our study, the inflammatory response was similar in the three groups of patients at admission; however we found significant differences on IL6 levels in serum at 24 h. IL6 levels significantly increased after 24 hours in PNR and PWER, but decreased in PER. Similar to the trend observed with glutamate, high levels of IL6 at 24 hours are independently associated with a worse intra-hospital evolution, but also with an 11% increase in extra-hospital improvement. A reduction in IL6 levels after 24 hours is another factor correlated with a worse extra-hospital improvement in IS patients (Fig. 4). This is a retrospective and single-center study. However, the study comprises a large and unbiased sample (5306 patients) for the purposes of this analysis. The use of a clinical criterion to define effective reperfusion could lead to a bias in the classification in therapeutic response groups. The Doppler and MR angiography study after recanalization treatment was limited to a small group of patients. The narrow therapeutic margin available (2020) 10  www.nature.com/scientificreports www.nature.com/scientificreports/ for patients with good therapeutic response could reduce the benefits in this group of patients; however, to avoid this effect we used the percentage deviation, not absolute, between the score of the NIHSS at discharge and at 3 months. The 3-month follow-up time is probably too short to evaluate the long-term effects. We have made an analysis in a subgroup of 287 patients that we have followed for a year, and the results were similar (data not shown).

Conclusions
Effective reperfusion after recanalization treatments provided the best clinical outcome for IS patients. However, this clinical improvement is usually achieved at an early stage, and long term recovery is less marked in PER. It is likely that the rapid reperfusion blocks the molecular response to ischemia, which is associated with restorative processes and cerebral plasticity, and some degree of excitatory and inflammatory mechanisms may be associated with the late recovery of patients.
The results of this work may involve the following practical suggestions: treatments targeting excitotoxicity may be optimal if administered in very short therapeutic windows and the inhibition of excitotoxicity after the first hours could block molecular responses for brain healing. The use of peripheral glutamate grabbers, which do not affect the mechanisms of the excitatory response in the brain, could be therefore an interesting therapeutic alternative 22 . The administration of anti-inflammatory drugs in large therapeutic windows could have negative effects. In this sense, the use of inflammatory markers like IL6 could be used not only for diagnosis but also as neuroprotectors in the management of IS 31,32 .

Methods and patients
Study design. This is a retrospective analysis conducted on a prospective registry of patients with acute cerebrovascular disease consecutively admitted to the Stroke Unit of the University Clinical Hospital of Santiago de Compostela (Spain) from September 2007 to March 2018. We prospectively included all IS patients admitted to the Stroke Unit in the BICHUS registry. The study was carried out in accordance with the Declaration of Helsinki of the World Medical Association and approved by the Research Ethics Committee of Galicia. Written informed consent was obtained on admission from each patient or from their relatives after full explanation of the procedures.
The clinical questions addressed in this study expressed in PICO (Patient, Intervention, Comparison, Outcome) format were: Among IS patients, do high levels of glutamate, IL6 and effective reperfusion compared with low levels of the mentioned makers and with uneffective reperfusion and non recanalization treatments improve the functional outcome at short and long term after hospitalization? Clinical variables. All patients were handled by the same stroke team according to the Stroke Unit protocol.
The variables of our registry included the date, onset-inclusion and wake-up strokes. Moreover, demographic data, previous modified Rankin scale score (mRS) and vascular risk factors were included. Data on previous carotid disease and processes of carotid revascularization were also collected, and in patients with a previous transient ischemic attack (TIA), the time between the TIA and stroke, as well as the coincidence or not with the topography. Medical history recording potential vascular risk factors, blood and coagulation test, 12-lead ECG and chest x-ray were collected. Carotid and transcranial Doppler were performed on admission; the transcranial Doppler study was repeated as a control of the recanalization treatment in 18% of the treated patients. Axillary temperature was obtained on admission and every 4 hours during the first 48 hours. The outcome clinical   www.nature.com/scientificreports www.nature.com/scientificreports/ modified Ranking Scale (mRS), which is commonly used for measuring the degree of disability or dependence in the activities of stroke patients, ranging from 0 (no symptoms) to 6 (dead), measured at discharge and at 3 ± 1 months. Good outcome was considered for patients with a mRS value of 0,1 and 2. Both scales were performed by accredited neurologist.
On the other hand, IS stroke subtypes were classified according to the TOAST criteria and hemorrhagic strokes were classified as hypertensive, amyloid or related to antiplatelet/anticoagulant. The type of reperfusion treatment, its characteristics and complications were also recorded.
At discharge, 28.7% of the patients were institutionalized and the rest were reintegrated into their family environment. During hospitalization, candidate patients for rehabilitation were evaluated by the Rehabilitation Service according to a common protocol. The outpatient rehabilitation was coordinated by the primary care physician, but it was not protocolized in an even way.
Analytical determinations. Biochemistry, haematology and coagulation test were assessed in the central laboratory of the hospital. For the molecular determinations, venous blood samples were collected in Vacutainer tubes (Becton Dickinson, San Jose, CA, USA) on admission, and 24 ± 12 hours and/or 48 ± 12 hours from stroke onset. After clotting for 60 minutes, blood samples were centrifuged at 3000 × g for 10 minutes, and the serum was immediately aliquoted, frozen and stored at −80 °C until analysis.
Glucose levels, glycosylated hemoglobin, leukocytes, red blood cells, platelets, fibrinogen, C-reactive protein, total and fractionated cholesterol, triglycerides, proBNP, vitamin D and cholecalciferol were measured on admission. Serum levels of glutamate were determined on admission and after 24 hours by high performance liquid chromatography (HPLC) analysis (1260 Infinity II; Agilent Technologies) using the AccQ-Tag derivatization method for amino acid analysis following the manufacturer´s technical specifications (Waters) 33 . Serum levels of interleukin 6 (IL6) were also measured on admission and after 24 hours using an immunodiagnostic IMMULITE 1000 System (Diagnostic Products Corporation, California, USA).
Neuroimaging studies. All patients underwent cerebral computed tomography (CT) on admission and between days 4 and 7 after IS. Multimodal magnetic resonance imaging (MRI) was also performed on admission in IS patients candidates to recanalization treatments. Lesion volumes were measured using ABC/2 method 34 in DWI-MRI on admission and betwen the 4 th and 7 th day in IS patients. In 2012, recanalization by endovascular treatment was protocolized in our hospital. Since 2012, a MRI angiography is performed in all patients candidate for reperfusion treatments. Neuroradiologists blinded to clinical and analytical data performed all imaging studies. Patients. We included 6565 stroke patients. Excluded patients were: ICH (1011), transient ischemic attack patients (512), patients transferred to other hospital during the acute phase (44), dead by other causes (40), no final stroke diagnosis (85), admitted outside the acute phase (97) and lacked follow-up at 3 months (480). Finally, 4295 patients were valid for the analysis.
Early neurological improvement (the reduction ≥8 points in the NIHSS in the first 24 hours 35 ) was used as a clinical marker of effective reperfusion.
Outcome endpoints. We have defined the intra-hospital improvement as the normalized percentage difference between the basal and discharge NIHSS through the formula: [(NIHSS on admission-NIHSS at discharge)/42 × 100], and the extra-hospital improvement between the NIHSS at discharge and at 3 ± 1 months, using the formula [(NIHSS at discharge-NIHSS at 3 months)/42 × 100] (42 was used to normalize the values as the maximum variation of the NIHSS).
We categorized the intra-hospital and extra-hospital improvement as a decrease ≥1 in the NIHSS between admission and discharge and between discharge and at 3 months, respectively. The deceased patients were given the highest NIHSS value. Statistical analysis. Firstly, an analysis was performed to describe the sample. Frequencies and percentages were calculated for categorical variables while mean and standard deviation [SD] or median and interquartile range were used for continuous variables depending on their adjustment to normality. Kolmogorov-Smirnov test with Lilliefors was applied to assess it. ANOVA and squared test were performed to determine differences in patients according with the reperfusion group (PNR, PER, PWER). The main objective of the study was to evaluate the intra and extra hospital improvement. In order to achieve it, multivariable models were built using a lineal regression analysis. The dependent variables were adjusted by clinical significant variables calculated in the bivariate analysis. Also, logistic regression analyses were carried out for patients with the same dependent variables but categorized. These results were shown as odds ratios (ORs) with 95% confidence intervals (CI 95%). P-value <0.05 was considered to be statistically significant. The statistical analyses were conducted in SPSS 21.0 (IBM, USA).

Data availability
All data are available within the text of the manuscript. Further anonymized data could be made available to qualified investigators upon reasonable request.