Correlation between serum S100β protein level with neurological deficit in patients with acute intracerebral hemorrhage

Introduction: S100β protein has shown its potential as biomarker of brain injury. However its efficacy is less known in intracerebral hemorrhage (ICH) cases. This study aimed to assess the relationship between S100β protein with neurological deficit, hematoma volume and mortality within 1 week among patients with acute


INTRODUCTION
Stroke is a disease caused by the disruption of cerebral blood flow. Currently, stroke has been one of the major causes of mortality and disability throughout the world. In the United States, overall total stroke incidence was 3.73 per 1,000 persons per year. Incidence of ischemic stroke was 3.29 per 1,000 persons per year, whereas for hemorrhagic stroke was 0.49 per 1,000 persons per year. 1 In Indonesia National Health Survey (RISKESDAS) conducted by Ministry of Health in 2007, the prevalence of stroke is 8.3 per 1000 persons; it increases to 12.1 per 1000 persons in RISKESDAS 2013. 2 Intracerebral hemorrhage (ICH) is the second most common type of stroke. 3 According to the American Heart Association (AHA), the most common type of stroke is an ischemic stroke of 87%, followed by intracerebral hemorrhage by 10-15% and subarachnoid hemorrhage by 3%. 4 However, the mortality rate in ICH (68%) was higher compared with ischemic stroke (57%). 1 Patients with ICH, who have same hematoma volume at the same location and received similar appropriate treatments, could have different neurological deficits and different prognoses. Brain hematoma can damage cellular structure due to laceration, compression, and distortion. 5 Nowadays, it is believed that many damages occurred in the subacute period were due to a complex pathophysiology pathway that still needs to be proven. In recent years, several studies had been conducted on biomarkers potential to help asses patient prognosis. The levels of some biomarkers may involve in the process of inflammation, the blood-brain barrier impairment, endothelial dysfunction, and neuron toxicity that cause secondary brain injury. Although some plasma biomarkers have been found to rise in response to acute ICH, but no blood biomarkers been routinely used, up until this day for clinical prognostication.
The S100β protein is a member of S100 protein, an acidic protein with a molecular weight of 21 kilodaltons and consist of two subunits α and β chain. The combination of these subunits was varied and differed among locations within the human body. The location of S100β protein is in the glial cell and Schwann cell. 6 S100β is primarily ORIGINAL ARTICLE expressed in astrocytes and has been implicated in the regulation of cell cycle progression and differentiation. 7 In adults, S100β protein level was found to be elevated in the presence of nervous system damages. In the presence of metabolic changes such as lack of oxygen and glucose in the brain, the initial response occurring in glia is the removal of the S100β protein. 8,9 In head injury, cerebral hemorrhage, and other conditions that affecting the central nervous system that results in damage to the astroglial cellular structure can cause leakage of S100β protein into the extracellular matrix, cerebrospinal fluid and finally enter the bloodstream. According to Lai et al. study, S100β was a potential prognostic marker for the outcome of subarachnoid hemorrhage. 7 The S100β protein may be used as a prognostic biomarker as well as a target of pharmacological therapy. 10 S100β protein has been used as a marker of cell damage and severity of brain injury, but the studies mostly focused on ischemic stroke, head injury, and subarachnoid hemorrhage. Although the S100β protein has been shown to predict prognosis after those conditions, it is still unusual for S100β protein or other biomarker proteins to be used in daily practice and has not been incorporated in standard prognostication.

Study subjects
A prospective study was performed in 46 consecutive patients who were hospitalized at Dr. Sardjito General Hospital, Yogyakarta, Atma Jaya Hospital, Jakarta and Mitra Keluarga Kelapa Gading Hospital, Jakarta within 48 hours after onset of ICH between August 2016 and December 2017. The patients met the following inclusion criteria: (1) ICH noted on brain computed tomography (CT) scan; (2) age ≥ 30 years old; (3) admission time ≤ 48 hours; (4) no other previous systemic diseases including malignancy, chronic heart or lung disease, liver cirrhosis; (5) give a consent to join this study. Informed consent was obtained from study participants or their families. The exclusion criteria include (1) history of head trauma; (2) history of previous stroke or TIA or subarachnoid hemorrhage. Histories of the illness, vital signs, level of consciousness assessed by Glasgow Coma Scale (GCS) and neurological deficits assessed by the National Institute of Health Stroke Scale (NIHSS) were obtained from all patients. To determine the reference value for S100β protein, measurement of S100β protein level in 42 healthy individuals were also performed. This study was approved by the Medical and Health Research Ethics Committee (MHREC) Faculty of Medicine Universitas Gadjah Mada -Dr. Sardjito General Hospital.

Laboratory evaluations
Blood samples (10 cc) were drawn from all ICH patients via antecubital vein into heparin tubes on the day of admission and on the 7 th day of onset. One-time blood samples were also drawn from 42 healthy individual controls. Samples were centrifuged (1000 rpm) for 10 minutes to obtain serum and stored at -20°C until analyzed. Serum S100β protein levels were determined using a Human S100β ELISA kit (BioVendor, Cat.no RD192090100R, Lot.no E17-092) according to the manufacturer's protocol.

Statistical Analysis
Statistical analysis was conducted using SPSS version 21. Continuous data were presented as mean±standard deviations or medians (minimum-maximum), as appropriate. A parametric test was used on data with a normal distribution, and the nonparametric Mann-Whitney U test was used for non-normally distributed data. Correlation between two variables was tested by Spearman's rank-order correlation coefficient analysis. A Receiver Operating Characteristics (ROC) curve was generated to identify serum S100β protein cutoff levels that predict a neurological deficit. A p-value of less than 0.05 was considered statistically significant.
Blood glucose level was significantly lower in the survival group than in the non-survival group (p=0.027). The Mann-Whitney U test revealed that serum S100β protein level was significantly higher in the non-survival group than in the survival group, 25 (Table 3). Based on the results of the CT scans between survival and non-survival groups, there were statistically significant differences in the hematoma volume (p=0.006), the presence of edema (p<0.001) and presence of intraventricular hemorrhage (p<0.001) ( Table 4).

The relationship between serum S100β protein levels and neurological deficit (NIHSS score)
Spearman's rank-order correlation coefficient analysis revealed that there was a significant positive correlation between serum S100β protein levels and NIHSS (r=0,418; p=0,004). The area under the curve (AUC) of serum protein S100β was 0.839±0.103 (95% CI 0.638-1.000), with cut off level was 28.505 pg/mL to distinguish the degree of stroke severity with 80% sensitivity and specificity of 87.8% (p=0.014). When the relationship between serum S100β protein levels and mortality was evaluated, it was shown that serum S100β protein levels were significantly associated with mortality within 1 week after the onset (p=0.032).
The relationship between GCS score, hematoma volume, and serum S100β protein levels Spearman's rank-order correlation coefficient analysis revealed that there was a significant negative correlation between serum S100β protein levels and GCS on admission (r=-0.376; p=0.010) and a significant positive correlation between serum S100β protein levels with hematoma volume (r=0.678; p=0.000).

Changes in serum S100β protein levels
In this study, there were 28 patients with ICH whose blood samples were taken twice. The median level of S100β protein on admission to hospital was 22.70 (19.31-96,74) pg/mL and on day 7 of onset was 21.24 (19.27-141.83) pg/mL. Serial median protein S100β serum changes in 28 patients were

ORIGINAL ARTICLE
taken at <48 hours and at the 7th day of onset was as shown in Table 5. Based on the Mann-Whitney U test, the median level of S100β protein on day 7 of onset was significantly different with control (<0.001). With Spearman's rank-order correlation coefficient analysis, serum S100β protein level on day 7 of onset was significantly correlated with NIHSS score (r=0.613; p=0.001).

DISCUSSION
S100β protein can be detected at a very low amount in the blood of normal individuals. Neither gender nor age affects S100β protein, but the race influences it. 11,12 S100β protein levels in healthy individuals ranged from 0.02 to 0.15 μg / L (i.e. 20-150 pg / mL) depending upon the race. 13 18 Acute ICH is a severe condition with a poor prognosis. After hematoma, the neurological function deteriorates rapidly over a short period of time. The hematoma activates astrocytes to secrete the S100β protein which then leads to vasogenic edema that negatively affects the neurological deficits or patient mortality. S100β protein is a biomarker for severity of brain damage and has been shown in acute ischemic stroke and traumatic brain injury. In many clinical trials of the traumatic brain injury, the S100β protein was increased and associated with the extension of primary brain injury, the presence of secondary brain injury and a poor prognosis. 19 In this study, serum S100β protein were significantly correlated with NIHSS (r=0.418; p=0.004), with AUC = 0.839 ± 0.103 (95% CI 0.638-1.000). Serum S100β protein was also correlated with patient mortality within 1 week (p=0.032). The results of this study are supported by several studies. In the study of Brea et al. involving 44 patients with acute ICH also found a correlation between serum S100β protein levels and NIHSS (r=0.607; p<0.0001). 20 Similar results were also obtained in a study conducted by Alatas et al. serum S100β protein levels were positively correlated with   Figure 1 Receiver operating characteristics curve of the serum S100 protein measurement on admission for the neurological deficit. A cut off level of admission S100 protein of 28.505 pg/mL predicts neurological deficits (NIHSS) with 80% sensitivity and 87.8% specificity. AUC=0.839 (95% CI, 0.638-1.000; p=0.014) NIHSS score (r=0.397; p<0.05). 15 Cut off for S100β protein in this study to distinguish stroke severity degree with 80% sensitivity and specificity 87,8% was 28.505 pg/mL (p=0.014). Cut off levels in this study were lower than in other studies. The previous study by James et al. also had a similar result that S100β protein may become an independent and strong predictor of neurological outcome. 21 According to Hu et al. S100β protein with cut off level of 192.5 pg/mL, predict the mortality within 1 week with a sensitivity 93.8% and a specificity 70.4% (p<0.001). 16 According to Delgado et al. S100β protein levels at hospital admission of >118 pg/mL, may predict early neurological deterioration with a sensitivity of 92% and a specificity of 70%. 22 In the study of Huang et al. cut off levels of S100β protein in plasma of 173.8 pg/mL had a sensitivity 80.0% and specificity 69.6% to predict mortality in 1 week. 23 A study conducted by Dilek at al predicted the mortality of patients with ICH by measuring serum S100β protein. The cutoff value of serum the S100β on the day of admittance for ICU to predict mortality was 50.9 pg/mL (AUC = 0.776, 95% CI=0.6700.882) with 73% sensitivity and 73% specificity. 24 In a study conducted by Alatas et al. serum S100β protein levels were negatively correlated with GCS score (r=0.467; p<0.01). 15 According to research by Huang et al. there was a significant correlation between GCS scores with plasma S100β levels in patients with ICH (r=-0.451; p=0.004). 23 While research conducted by James et al. it was found that the S100β protein was not correlated with the initial GCS score (r 2 =0.32). 21 According to research by Hu et al. GCS score at hospital admission, day 1 and 2 was negatively correlated with plasma S100β protein levels (r=0.588; p=0.000, r=0.457; p=0,000, r=0.399; p=0.001 respectively) but unrelated on day 3, day 5 and day 7 (r=-0.185; p=0.155, r=-0.168; p=0.208, r=-0.147; p=0.289 respectively). 16 In this study, S100β protein content significantly shown a negative correlation with baseline GCS score (r=-0.376; p=0.010). This finding was similar to that of Alatas In an intracerebral hemorrhage, the hematoma will activate astrocytes to secrete the S100β protein. There is a study linking S100β protein levels with hematoma volume, for example, a study conducted by Alatas et al. the results of serum S100β protein levels were positively correlated with hematoma volume (r=0,564; p<0.001). 15 According to Huang et al. there was a significant correlation between hematoma volume with the S100β level in plasma (r=0.380; p=0.019) and cerebrospinal fluid (r=0.468; p=0.003) among patients with intracerebral hemorrhage. 23 A study conducted by James et al. obtained results that the S100β protein content correlated with hematoma volume (r2=0.48; p=0.01). 21 According to Hu et al. volume of bleeding at the time of admission, day 1 and 2 was positively correlated with plasma S100β protein levels (r=0.568, p<0,0001, r=0.455; p<0,0001, r=0.363; p=0.002, respectively) but no correlation was found on day 3, day 5 and day 7 (r=0.157; p=0.228, r=0.034; p=0.803, r=0.099; p=0.474, respectively). 16 The results of the study by Yoon et al. there was a positive correlation between volume of hematoma and serum S100β protein level at hospital admission (r=0.663; p=0.001) and on day 3 (r=0.609; p=0.021), but no correlation was shown on day 7. 18 According to research by Delgado et al. there was a positive correlation between the volume of hematoma at admission and plasma S100β protein levels (r=0.45; p<0.0001) and between perihematomal edema and plasma S100β protein levels (r=0.27; p=0.033). However, the median level of baseline S100β protein was not associated with the possibility of further bleeding or perihematomal edema expansion. 22 A similar result was also shown by Brea et al. with serum S100β protein levels correlated with intracerebral hematoma within the first 24 h (r=0.607; p<0.0001). 20 A somewhat different study was conducted by Li et al. in patients with intracerebral hemorrhage grouped by hematoma volume, i.e., mild bleeding (<20 cc), moderate bleeding (20-40 cc) and severe bleeding (>40 cc). According to this study, the larger hematoma volume was accompanied by a larger serum S100β protein (1.03±0.12, 1.79±0.20, 2.85±0.30 pg/ mL). 17 In Junttila et al. S100β protein concentration in serum at baseline was associated with hematoma volume (rho=0.50; p<0.001). 25 In this study, the results were similar to the aforementioned studies, that serum S100β protein levels at baseline correlated with hematoma volume (r=0.678; p<0.001).
Given the result of our study and supported with many previous finding, we think that S100β protein level is an important marker for predicting neurological function and mortality within 1 week because of the relationship between S100β protein levels and NIHSS, GCS, and hematoma volume.

CONCLUSION
In this study, high levels of serum S100β protein were present in patients with ICH. It also shows a significant correlation with neurological deficits, mortality within 1 week and hematoma volume. These findings suggest that serum S100β protein has potential prognostic utility for acute ICH.