Brain injury resulting from traumatic, ischemic and/or haemorrhagic etiology is a significant international health concern, representing a potentially catastrophic debilitating medical emergency with poor prognosis for long-term disability .
Ranking of head injuries by severity is an essential part of clinical management, prognosis, treatment trials, and legal assessment. Various methods for assessing severity are reviewed. Clinical assessment of brain function and neuro-imaging techniques constitute the common ways for diagnosis and assessment of brain damage. Despite the substantial progress in clinical and neuro-monitoring it remains difficult to quantify the extent of primary brain injury as well as the ongoing secondary damage and hence they cannot guide efficient therapeutic measures and prognosticate effectively the final outcome [2,3].
Therefore, the development and the use of biochemical markers for assessment of brain injury severity have been of considerable interest. Biomarkers would have important applications in diagnosis, prognosis and clinical research of brain injuries. Biochemical markers are considered to be a simple and a rapid diagnostic tool that will immensely facilitate allocation of the major medical resources required to treat brain injuries .
The bulk of research in brain injury has focused on neuron-specific enolase (NSE) and S-100β. Currently, the most promising marker of brain damage in the literature is considered to be NSE .
Neuron Specific Enolase (NSE) is a glycolytic enzyme family (enolases) predominantly located in neurons and neuroectodermal cells and serves as a marker of neuronal damage. Increased concentration of NSE can be measured in the cerebrospinal fluid (CSF) and in peripheral blood after neuronal damage and provides a reliable laboratory indicator of the degree of brain cell damage, and may allow for early prediction of outcome [6,7].
Aim of the work
Investigate the serum levels of NSE in patients with head injury admitted to ICU and to correlate them with their outcome and other clinical parameters and to assess the validity of outcome prediction with this serum marker in patients with head injury.
Patients and methods
The study protocol was approved by the local Ethics. We prospectively enrolled twenty patients with head injury, admitted to Critical care unit of the “MISR university for science and technology hospital”; from January 2010 to December 2010. Informed consent was given by the patient or immediate relative (first degree). This study did not interfere with the current medical practice of the investigator. No invasive medical procedure is required by the protocol. The investigator can decide for any treatment that is for the best interest of his patients.
- Age ≥ 18 years.
- Patients with head injury, including traumatic brain injury.
- Patients with stroke either ischaemic or hemorrhagic stroke.
Patients were selected randomly and prospectively and included into the study on the day of ICU admission and they were followed up till the day of discharge or demise.
Evaluation of patients
All included patients were subjected to the following:
- Full clinical evaluation: Including history, obtained usually from family members, police officers, paramedics or witnesses, and appropriate physical and neurological examination with special emphasis on vital signs and GCS; which were evaluated at the day of admission and 48 h later.The patients were classified according to GCS as: mild head injury (GCS 13–15), moderate head injury (GCS 9–12), and severe head injury (GCS ≤ 8).
- Laboratory investigations:
Imaging studies: Including chest X-ray, echocardiography and CT brain; was performed in all patients on admission and was initially interpreted by a radiologist consultant. CT scans were repeated at several follow up examinations depending on clinical course.
Clinical data: Length of ICU stay, final outcome of survival/mortality rates were reported for all patients until ICU discharge.
- Routine labs: CBC (complete blood count): Hemoglobin, Hematocrit, White blood cells and platelet count, Coagulation profile: PT (prothrombin time), PC (prothrombin concentration), INR and PTT (partial thromboplastin time), ABGs (arterial blood gases), Liver function tests: ALT (Alanine aminotransferase), AST (Aspartate aminotransferase), BIL (bilirubin) and albumin, Kidney function tests: Na, K, Creatinine and Urea, and random blood suger.
- These routine labs were withdrawn on study day 1 and then 48 h later.
- Labs specific for our study: Venous blood samples for NSE were drawn during the first 24 h of admission and after 48 h of their admission. Blood samples were allowed to clot and then centrifuged to separate the serum according to common procedures. The serum was separated from the clot within 60 min of collection to avoid leaking of NSE from blood cells. Plasma was not used as it is not recommended since significant amounts of NSE can be released from platelets. Haemolysed samples had not been used since erythrocytes contain significant amounts of NSE. Samples preserved at −70 °C and stored for analysis. Frozen samples allowed thawing slowly at 2–8 °C over night and then the samples brought to room temperature before analysis.
- The serum concentration of NSE was measured by using solid-phase enzyme immunoassay. The essay used a highly specific monoclonal antibody to NSE. During the incubation, the serum NSE reacted with the antibody, which was immobilized on polystyrene beads, and then with a rabbit polyclonal antibody to form a ‘sandwich’. The beads were washed to remove any unbound rabbit antibody and incubated with a highly purified goat antibody against rabbit immunoglobulin, which was conjugated to horseradish peroxidase. The beads were rewashed to remove any unbound antibody enzyme conjugate and incubated with an enzyme substrate solution. The change in substrate color was measured with a Cobas photometer and was directly proportional to the amount of the NSE present .
Data were statistically described in terms of range, mean ± standard deviation (±SD), median, frequencies (number of cases) and relative frequencies (percentages) when appropriate. Comparison of quantitative variables between the study groups was done using simple t-test (for two variables) or ANOVA test (for more than two variables). For comparing categorical data, Chi square (χ2) test was performed. Exact test was used instead when the expected frequency is <5. Receiver operator characteristic (ROC) analysis was used to determine the optimum cut off value for the studied diagnostic markers. Pearson correlation coefficient was used for analysis of relation of bivaried. Linear regression was used to estimate the coefficient of linear equation involving one or more independent variable that best predict the value of dependant variable. A probability value (p value) less than 0.05 was considered statistically significant. All statistical calculations were done using computer programs Microsoft Excel 2003 (Microsoft Corporation, NY, USA) and SPSS (Statistical Package for the Social Science; SPSS Inc., Chicago, IL, USA) version 15 for Microsoft Windows.
- Demographic and baseline clinical data at ICU admission (Table 1).Patients demographic and clinical data are shown in Table 1.
- NSE measurement: The mean NSE level on admission was 9.9 ± 7.8 and after 48 h of admission the mean was (12.1 ± 11.1).
- Correlative analysis:
- Correlation between NSE and age and gender: There was no significant correlation between NSE with age of patients, (p = 0.15, r = 0.5 on admission and p = 0.33 and r = 0.1 48 h after admission). On the other hand, NSE was significantly higher in males (10.4 ± 8.1) than in females (5.5 ± 0.7) (p = 0.02).
- Correlation between NSE with traumatic and non traumatic patients: There was no significant difference of NSE level in traumatic (mean NSE on admission was 8.8 ± 8.1 and NSE after 48 h was 9.4 ± 11.6) or non traumatic patients (mean NSE on admission was 10.9 ± 7.9 and NSE after 48 h was 14.4 ± 10.8) (p value >0.05).
- Correlation between NSE level and severity of brain injury: NSE on admission was higher in the group with moderate to severe brain injury (mean = 13.1 ± 10.2) than in mild head injury (mean = 6.8 ± 2) but it was not significant (p = 0.07). NSE after 48 h of admission was significantly higher in the group with moderate to severe brain injury (mean = 17.2 ± 13.5) than in mild brain injury (mean = 7.1 ± 4.8) (p = 0.04).
- Correlation between NSE level and GCS: There was a significant negative correlation between NSE and GCS on admission (p = 0.01 & r = −0.564), also NSE on admission is negatively correlated with GCS 48 h after admission (p = 0.008 & r = −0.57). NSE 48 h after admission had significant negative correlation with GCS 48 h after admission (p = 0.001 & r = −0.675) (Fig. 1)
- Correlation between NSE level and the change in GCS: There was a significant increase in NSE 48 h after admission in patients with documented deterioration of GCS (mean NSE 48 h after admission in those who improved, the same and deteriorating GCS was 9.8, 10.9 and 30, respectively with p-value was 0.04) while NSE on admission showed no statistically significant difference between these groups (mean NSE on admission was 8.7, 9.1 and 20, respectively with p-value = 0.16). (Table 2)
- Correlation between NSE level and the change in the CT: Although NSE on admission was higher in patients who were deteriorated according to CT findings (mean; 7.9, 9.6 and 20 in patients who improved, the same and with worse CT findings respectively), it was not significant (p = 0.15). NSE 48 h after admission was significantly higher with the worsening of CT finding (mean; 9, 11 and 30, respectively with p = 0.04) (Table 3).
- Correlation between NSE level and outcome: NSE on admission was significantly higher in non survivors (mean = 20 ± 14.1) than in survivors (mean = 8.8 ± 6.6) with p = 0.05. Also NSE 48 h after admission was more significantly higher in non survivors (mean = 30 ± 14.1) than in survivors (mean = 10.2 ± 9.3) with p = 0.01. Regarding the survivors, NSE on admission showed no significant difference among survivors (p = 0.25) but NSE 48 h after admission showed significant difference among survivors; mean NSE 48 h after admission was 5.9, 13.2 and 11.1 in recovered pts, who had moderate disability & those with severe disability respectively with p = 0.05 (Table 4).
- Prognostic ability of NSE: Receiver operator characteristic (ROC) curve was calculated for the use of NSE level as a predictor for mortality. The area under curve (AUC) for NSE to predict ICU mortality was 0.93 (95% confidence interval, 0.809–1.052). The optimal cutoff value of NSE level to predict ICU mortality was 18 μg/L with a sensitivity of 100% and specificity of 90% (Fig. 2a).
Also a cutoff value of NSE level to predict bad outcome was 7.9 μg/L with a sensitivity 85% and specificity 77%. The area under curve for NSE to predict bad outcome was 0.76 (95% confidence interval, 0.53–0.99) (Fig. 2b).
Assessment of severity of brain damage is an essential part of clinical management, prognosis and treatment trials. Clinical examination, neuro-monitoring methods and the modern neuro-imaging techniques are quite often not sufficient to evaluate and quantify the severity of the brain injury and hence they cannot guide efficient therapeutic measures and prognosticate effectively the final outcome. The problem becomes even more prominent with comatosed patients with facial trauma and/or pre-existing pupillary abnormalities .
The measurement of NSE concentrations in serum and cerebrospinal fluid (CSF) following cerebral ischemia and traumatic head injury provides a reliable laboratory indicator of the degree of brain cell damage, and may allow for early prediction of outcome. NSE is useful for both diagnostic and prognostic purposes. It can predict infarct volume, hemorrhagic transformation and other clinical variables .
Therefore, we tried in this study to investigate the prognostic value the serum level of neuron specific enolase (NSE) in patients with head injury, with particular regard to its role in predicting illness severity, clinical course and mortality.
Our study included patients with age ranged from 20 to 80 years with mean age (50.8 ± 16.5). We compared NSE in different ages and found that there is no significant correlation between NSE and age (p = 0.15 and r = 0.5).
This is in agreement with Casmiro et al.  who studied 108 patients (68 men and 40 females) without neurological diseases to analyze the NSE relationship with sex and age. This study showed that NSE level not significantly increasing with advancing age (p = 0.16, r = 0.027).
Also it goes with Vos et al.  who conducted a study on 85 patients with severe brain injury (admission GCS ≤ 8) found that NSE not significantly correlated with age. Also Nygaard et al.  found that serum NSE not correlated with age and only CSF-NSE correlated significantly with age.
Our study included 18 males (90%) and 2 females (10%); when we compared NSE level in both sexes we found that NSE on admission was significantly higher in males (10.8 ± 8.1) than in females (5.5 ± 0.7) (p = 0.02) in spite of there is no significant difference in NSE 48 h after admission between males and females.
This is in agreement with de Kruijk et al.  who studied serum NSE of 91 controls 14% of them was female and 86% were male. The median NSE concentration in serum of male controls was higher than in serum of female controls (9.7 versus 7.6 mg/L; p = 0.037).
Also it goes with Nygaard et al.  who collected serum and CSF-NSE from 63 males and 24 females. They found that NSE concentration was significantly higher in males than in females. However, both Casmiro et al.  and Vos et al.  found that NSE not significantly correlating to gender.
In our study, patients were grouped according to the cause of admission and CT findings into traumatic brain injury (TBI) which accounted for 45% (n = 9) and non traumatic which include those with ischaemic stroke (IS) which accounted for 25% (n = 5) and hemorrhagic stroke (ICH) which accounted for 30% (n = 6). We compared NSE in these groups and found that there was no significant difference between NSE level in traumatic (8.8 ± 8.1) or non traumatic patients (10.9 ± 7.9) (p value = 0.5).
As NSE was not affected by different etiology, we compared the enrolled patients despite the variation in the cause of brain injury. We found no study comparing NSE level in traumatic and non traumatic brain injury; some studies investigate serum NSE in those with traumatic brain injury and others investigate NSE in those with ischemic and hemorrhagic stroke. Our study may be the only one to do this comparison.
Regarding to the degree of severity of head injury, we demonstrated that NSE on admission was insignificantly higher in the group of moderate to severe brain injury (13.1 ± 10.2) than in mild brain injury (6.8 ± 2) with a p-value of 0.07. On the other hand; NSE 48 h after admission was significantly higher in the group of moderate to severe brain injury (17.2 ± 13.5) than in mild brain injury (7.1 ± 4.8) with a p-value of 0.04. Moreover, we found that there was a significant negative correlation between NSE on admission and on admission GCS (p = 0.01 & r = −0.564) and NSE 48 h after admission is more significantly correlated with GCS 48 h after admission (p = 0.001 & r = −0.675).
This is in agreement with Zahra et al.  who studied 45 patients; 14 of them with moderate to severe head injury (GCS < 12) and the rest of them with minor head injuries (GCS score >12) and the results showed serum levels of NSE in those with moderate to severe head injury (22.8 ± 13.3) was significantly higher than in mild (9.8 ± 7.7) with p < 0.01. But he found no significant statistical correlation between NSE concentration and GCS (p = 0.23).
Also it goes with Guezel et al. , Rothoeral et al.  and Herrmann et al.  who found that the levels of neuron-specific enolase were significantly higher in the patients with severe brain injury (GCS ≤ 8) and there was a significant negative correlation between the serum NSE levels and GCS.
This is in contrast with Vos et al.  who conducted the study only on patients with severe brain injury and did not found that significant correlation between NSE and GCS.
In our study we assessed improvement of patients by comparing their GCS on admission and after 48 h. We found that 12 patients (60%) improved, 6 patients (30%) were with the same GCS and 2 patients (10%) were deteriorating. When comparing change of NSE level in these groups we found that NSE level was 9.8, 10.9 and 30 in the 3 groups, respectively and NSE level was significantly higher in those with deteriorating GCS (p = 0.04).
This is in agreement with Meric et al.  who found that NSE levels rose as GCS fell. Also Herrmann et al.  stated that patients with moderate to severe brain injury (GCS ≤ 12) exhibited significantly higher NSE concentrations and a significantly longer release compared to patients with minor head injury (GCS = 13–15). This explains why in our study NSE 48 h after admission was more significant in detecting severity of head injury.
In our study, follow up CT brain was repeated after 48 h of admission. It was observed that the patients fell into three broad categories. The first was the resolving category. In these patients, the contusions and/or infarctions were resolving. This category accounted for 40% (n = 8) of patients. The second was the static category which accounted for 50% (n = 10). In these patients, the CT findings were with the same finding as admission CT. The third category (n = 2) accounted for 10% was the progressive category. In these patients there was extension of the initial infarction/hemorrhage or hemorrhagic transformation of the infarction. NSE levels were compared in all of the three categories and revealed that NSE value get higher with deterioration of head injury. NSE 48 h after admission was more significant than NSE on admission in detecting the progression of brain injury. p Value for NSE 48 h after admission was 0.04 versus 0.15 or NSE on admission.
This is in agreement with Manfred et al.  who found a high significant correlation between NSE and lesion in CT and this correlation reached maximum value at second day of brain insult (r = 0.83, p < 0.0001). He also found that serum NSE did not differ with the type of lesion whether hemorrhagic or ischemic stroke but to the lesion size.
Also it agrees with Zahra et al. , Vos et al.  and Naeimi et al.  who observed a statistically significant correlation between increased levels of NSE and CT changes (p < 0.04).
Again in agreement with our study, Samit , found a strong correlation between serum levels of NSE with the progression of head injury (p < 0.001). In the resolving group, the marker levels showed a decreasing trend.
In our study survivors accounted for 90% (n = 18 patients), and including those with good recovery 6 patients (30%), 7 patients (35%) with moderate disability and 5 patients (25%) with severe disability. While the non survivors accounted for 10% (n = 2). When we compared NSE in these groups, we found that NSE on admission was significantly higher in non survivors (20 ± 14) than in survivors (8.9 ± 6.7) (p-value 0.05). NSE 48 h after admission was more significantly higher in non survivors (mean = 30 ± 14) than in survivors (mean = 10.2 ± 9.3) (p = 0.01). Also we found that NSE 48 h after admission was significantly higher with bad outcome (mean = 5.9, 13.2 and 11.1 in patients with good recovery, moderate and severe disability respectively with a p value of 0.05).
This is in agreement with Samit , who found that NSE in those with good outcome (17 ± 5.1) and in those with poor outcome (25.4 ± 5.1). A strong statistical significance exists between serum NSE level and outcome assessment (p value <0.001).
Also David et al. , Bandyopadhyay et al.  and Vos et al.  found that NSE level correlate with outcome. Patients with bad outcome (death or severe disability) had a higher concentration of NSE than in patients with good outcomes (survivors or with minimal disability) (r = 0.319, p < 0.0001) and the enzyme levels after 48 h were more significant than their level on admission in detecting outcome. p Value for NSE2 was <0.001 versus 0.065 for NSE1.
This also agrees with Olivecrona et al. , Guzel et al. , Sawauchi et al. , and Sogut et al.  who found that NSE was significantly higher in patients who died compared to the survivors. Increased concentration of NSE even precede changes in clinical or other diagnostic parameter.
This is in contrast with Raabe et al.  who studied 82 patients with head injury, 49 showed a favorable outcome (60%) and 33 showed unfavorable outcome (40%) after 6 months with a mortality rate of 38% (31 patients). There was no significant difference between the maximum levels of NSE of both outcome groups (NSE was 26.7 μg/L in favorable outcome versus 12 μg/L in unfavorable outcome with a p value of 0.09).
In our study a receiver operating characteristic (ROC) curve was calculated for the use of NSE level as a predictor of mortality and bad outcome. The area under the ROC curve for NSE to predict mortality was 0.93. The optimal cutoff value for NSE levels to predict mortality was 18 μg/L. This cutoff value gave a sensitivity of 100% and specificity of 90%. The area under the ROC curve for NSE to predict bad outcome was 0.76. The optimal cutoff value for NSE levels to predict bad outcome was 7.9 μg/L. This cutoff value gave a sensitivity of 85% and specificity of 77%.
Vos et al.  found that levels >21.7 μg/L strongly could predict death and that of 3.95 μg/L strongly predict bad outcome. Meric et al.  found that cut-off level of NSE at 20 ng/mL could detect poor neurological outcome by a sensitivity 87% and specificity 82.1%. The area under curve was 0.93. Bandyopadhyay et al.  found that NSE at level 21 ng/dL could detect bad outcome with a sensitivity 86% and specificity 47%. The area under curve was 0.83. Samit , calculated cutoff values for poor outcome and found that NSE > 28.2 μg/L can detect poor outcome with specificity 96.6% and sensitivity 80%. ROC analysis revealed that area under curve was 0.86.
The difference obtained in cut-off point could be explained by that Serum and CSF-NSE values vary widely among different studies on normal populations because of different determination methods; therefore, each laboratory should obtain its own reference values.
Neuron specific enclose can be used as a potentially useful marker for brain damage and can be used as a simple, rapid and easy to perform and interpret test for early prognosis and prediction of adverse outcome.
- Application of NSE as a routine scoring tool on admission and follow up for all patients admitted with head injury.
- Serial measurement of NSE as it will help to monitor the efficacy of therapeutic interventions delivered.
- Future large scale studies comparing the release pattern of NSE between patients admitted with traumatic brain injury and those with non traumatic brain injury.
 Persson L, Hardemark HG, Gustafsson J, et al. S-100 protein and neuron-specific enolase in cerebrospinal fluid and serum: markers of cell damage in human central nervous system. Stroke 1987;18:911-918.
 Berg J, Tymoczko J, Stryer L. Biochemistry. New York: W.H. Freeman & Company; 2006.
 Ingebrigten T, Rommer B. Biochemical serum markers of traumatic brain. J Trauma 2002;52:798-808.
 Schmechel D, Marangos PJ, Brightman M. Neuron-specific enolase is a molecular marker for peripheral and central neuroendocrine cells. Nature 1978;276:834-836.
 Johnsson P. Markers of cerebral ischemia after cardiac surgery. J Cardiothorac Vasc Anesth 1996;10:120-126.
 Ergun R, Bostanci U, Akdemir G, et al. Prognostic value of serum neuron-specific enolase levels after head injury. Neurol Res 1998;20:418-420.
 Ross SA, Cunningham RT, Johnston CF, et al. Neuron-specific enolase as an aid to outcome prediction in head injury. Br J Neurosurg 1996;10:471-476.
 de Boussard N, Lundin A, Karlstedt D, et al. S100B and cognitive impairment after mild traumatic brain injury
. J Rehabil Med 2005;53:37-7.
 Reilly P, Graham D, Adams J, Jennett B. Patients with head injury who talk and die. Lancet 1975;2:375-377.
 Pleines UE, Morganti-Kossmann MC, Rancan M, et al. S-100 beta reflects the extent of injury and outcome, whereas neuronal specific enolase is a better indicator of neuroinflammation in patients with severe traumatic brain injury
. J Neurotrauma 2001;18:491-498.
 Casmiro M, Maitan S, De Pasquale F, et al. Cerebrospinal fluid and serum neuron-specific enolase concentrations in a normal population. Eur J Neurol 2005;5:369-374.
 Vos PE, Lamers KJ, Hendriks JC, et al. Glial and neuronal proteins in serum predict outcome after severe traumatic brain injury
. Neurology 2004;62(8):1303-1310.
 Nygaard O, Langbakk B, Romner B. Neuron-specific enolase concentrations in serum and cerebrospinal fluid in patients with no previous history of neurological disorder. Scand J Clin Lab Invest 1998;58(3):183-186.
 de Kruijk JR, Leffers P, Menheere P, et al. S-100B and neuron-specific enolase in serum of mild traumatic brain injury
patients. A comparison with healthy controls. Acta Neurol Scand 2001;103:175-179.
 Zahra S, Alexandra W, Kambiz S, et al. Predictive value of S-100B protein and neuron specific-enolase as markers of traumatic brain damage in clinical use. Brain Inj 2006;20(5):463-468.
 Guzel A, Er U, Tatli M, et al. Serum neuron-specific enolase as a predictor of short-term outcome and its correlation with Glasgow Coma Scale in traumatic brain injury
. Neurosurg Rev 2008;31(4):439-444.
 Bandyopadhyay S, Hennes H, Gorelick MH, et al. Serum neuron-specific enolase as a predictor of short-term outcome in children with closed traumatic brain injury
. Acad Emerg Med 2005;12(8):732-738.
 Herrmann M, Jost S, Kutz S, et al. Temporal profile of release of neurobiochemical markers of brain damage after traumatic brain injury
is associated with intracranial pathology as demonstrated in cranial computerized tomography. J Neurotrauma 2000;17(2):113-122.
 Meric E, Gunduz A, Turedi S, et al. The prognostic value of neuron-specific enolase in head trauma patients. J Emerg Med 2010;38(3):297-301.
 Manfred H, Hannelore E. Brain derived proteins as markers of acute stroke: their relation to pathophysiology, outcome prediction and neuroprotective drug monitoring. Restor Neurol Neurosci 2003;21:177-190.
 Naeimi ZS, Weinhofer A, Sarahrudi K, et al. Predictive value of S-100B protein and neuron specific-enolase as markers of traumatic brain damage in clinical use. Brain Inj 2006;20(5):463-468.
 Samit Satish Mehta. Biochemical serum markers in head injury: an emphasis on clinical utility. Clin Neurosurg 2010;57:134-140.
 David B, Tomás S, Miguel B, et al. Temporal profile and clinical significance of serum neuron-specific enolase and S100 in ischemic and hemorrhagic stroke. Clin Chem Lab Med 2009;47(12):1513-1518.
 Olivecrona M, Rodling-Wahlström M, Naredi S, Koskinen LO. S-100B and neuron specific enolase are poor outcome predictors in severe traumatic brain injury
treated by an intracranial pressure targeted therapy. J Neurol Neurosurg Psychiatry 2009;80(11):1241-1247.
 Sawauchi S, Taya K, Murakami S, et al. Serum S-100B protein and neuron-specific enolase after traumatic brain injury
. No Shinkei Geka 2005;33(11):1073-1080.
 Sogut O, Guloglu C, Orak M, et al. Trauma scores and neuron-specific enolase, cytokine and C-reactive protein levels as predictors of mortality
in patients with blunt head trauma. J Int Med Res 2010;38(5):1708-1720.
 Raabe A, Grolms C, Seifert V. Serum markers of brain damage and outcome prediction in patients after severe head injury. Br J Neurosurg 1999;13(1):56-59.
 Rothoeral RD, Woertgen C, Brawanski A. S-100 serum levels and outcome after severe head injury. Acta Neurochir Suppl 2000;76:97-100.