Predicting functional outcomes in severe traumatic brain injury: Role of S100B along with other clinical and imaging parameters

Judy Ann John; Jane Elizabeth Sajan; Anna Oommen; Mathew Joseph; Suranjan Bhattacharji

doi:10.4103/cmi.cmi_103_21

ORIGINAL ARTICLE

Year : 2022 | Volume : 20 | Issue : 2 | Page : 74-81

Predicting functional outcomes in severe traumatic brain injury: Role of S100B along with other clinical and imaging parameters

Judy Ann John¹, Jane Elizabeth Sajan¹, Anna Oommen², Mathew Joseph³, Suranjan Bhattacharji¹
¹ Department of Physical Medicine and Rehabilitation, Christian Medical College, Vellore, Tamil Nadu, India
² Department of Neurological Science, Neurochemistry Laboratory, Christian Medical College, Vellore, Tamil Nadu, India
³ Department of Neurological Science, Neurocritical Care and Trauma Unit, Christian Medical College, Vellore, Tamil Nadu, India

Date of Submission	19-Nov-2021
Date of Decision	21-Mar-2022
Date of Acceptance	25-Mar-2022
Date of Web Publication	07-May-2022

Correspondence Address:
Dr. Judy Ann John
Department of PMR, Christian Medical College, Vellore - 632 002, Tamil Nadu
India

Source of Support: None, Conflict of Interest: None

Check

DOI: 10.4103/cmi.cmi_103_21

Abstract

Background: The objective of this prospective cohort study was to evaluate the role of serum S100B, along with other clinical and imaging parameters, in predicting functional outcome in severe traumatic brain injury (TBI). Methodology: We included 23 patients with severe TBI admitted within 48 h of injury. The Glasgow Coma Scale (GCS), pupil reactivity, and Marshall's computerized tomography grade were assessed at admission and serum levels of S100B were estimated at 48 h and 21 days post injury. ROC curve was generated to determine the cutoff value for S100B levels. Clinical data were analyzed to study their association in predicting the functional outcome as assessed by the Glasgow coma scale (GOS), Functional Independence Measure (FIM), and Modified Mini-Mental State Examination (3MS) at 6 months. Results: S100B levels above 1.37 μg/L at 48 h significantly predicted poor outcomes at 6 months as assessed by GOS (sensitivity of 64%, specificity of 83%, and likelihood ratio (LR) of 3.76), FIM (sensitivity of 75%, specificity of 85%, and LR of 5.0), and 3MS (sensitivity of 60%, specificity of 83%, and LR of 3.53). On linear regression analyses, GCS motor score at 96 h and S100B levels were independent predictors of GOS, FIM, and 3MS. The positive predictive value for poor outcome (GOS ≤3 or FIM <72 or 3MS <75) was 100% when S100B levels at 48 h ≥1.37 μg/L were combined with GCS motor scores at 96 h ≤3. Conclusion: S100B levels at 48 h post injury and GCS motor score at 96 h were significant predictors of long-term functional outcome in severe TBI.

Keywords: Glasgow Coma Scale, Glasgow Outcome Scale, recovery of function, S100 calcium-binding protein beta subunit, traumatic brain injuries

How to cite this article:
John JA, Sajan JE, Oommen A, Joseph M, Bhattacharji S. Predicting functional outcomes in severe traumatic brain injury: Role of S100B along with other clinical and imaging parameters. Curr Med Issues 2022;20:74-81

How to cite this URL:
John JA, Sajan JE, Oommen A, Joseph M, Bhattacharji S. Predicting functional outcomes in severe traumatic brain injury: Role of S100B along with other clinical and imaging parameters. Curr Med Issues [serial online] 2022 [cited 2023 Mar 22];20:74-81. Available from: https://www.cmijournal.org/text.asp?2022/20/2/74/344923

Introduction

The Glasgow Coma Scale (GCS), since its development by Teasdale and Jennet in 1974, has become the most widely used clinical measure of the severity of injury in patients with traumatic brain injury (TBI).^[1] This score can be easily estimated at the bedside and is universally documented in all cases of head injury. Although a large number of other clinical and radiological prognostic markers have been studied, such as age, findings on computerized tomography (CT) scan, pupillary response, hypotension, and intracranial pressure, the GCS still appears to have withstood the test of time as the most reliable prognostic marker available in TBI.^{[1],[2],[3],[4],[5],[6],[7]}

Nevertheless, there are many important issues to be considered in using GCS in prognosticating functional outcome in TBI. For example, the modern, prehospital treatment of many TBI patients (sedation, pharmacological paralysis, and/or endotracheal intubation) complicates the early determination of accurate GCS scores in nearly half of the patients admitted to trauma centers.^[8] To overcome these limitations, various models, such as the International Mission for Prognosis and Analysis of Clinical Trials in TBI and the Corticosteroid Randomization after Significant Head Injury, have been developed to provide a more comprehensive, reliable, and accurate prognostic indicator. These models incorporate CT findings and other clinical parameters along with GCS to improve the prediction of mortality and functional outcomes in TBI.^{[9],[10],[11]} However, it was suggested that these models may require regular updating according to specific population characteristics.^[11]

There is considerable interest in biochemical markers that can aid prognostication.^[12] A promising biochemical marker is the protein S100B, a part of a family of the calcium-binding proteins, that has been localized in astroglial and Schwann cells.^[13] Serum S100B provides information not only on the extent of primary injury to the brain but also on the ongoing secondary damage, thus aiding prognostication of long-term functional outcome.^[13] A recent review article reported that S100B could be utilized as a predictor of functional outcome in moderate-to-severe TBI.^[14] A study done in an Indian population reported a positive correlation of S100B levels with neurocognitive outcome in mild TBI.^[15] However, the utility of S100B in predicting outcome in severe brain injury in an Indian population is still to be explored.

The objective of this prospective cohort study was to estimate S100B levels at 48-h and 21 days post injury in patients with severe TBI. Cutoff values were determined (using ROC analyses) to calculate sensitivity, specificity, predictive values, and likelihood ratios in order to predict good/poor functional outcome, 6 months later. In addition, we used linear regression models to determine the value of combining S100B levels with GCS, pupil reactivity, and Marshall CT grade to improve prognostication of functional outcome in these patients.

Methodology

Study design, participants, and setting

This is a prospective study involving 23 patients with severe head injury (GCS ≤8) admitted within 48 h of injury to the Neurosurgical Intensive Care Unit at a tertiary care center in South India from August 2000 to January 2003. Ethical approval for the study was obtained from the Institutional Review Board (IRB Min. Number 4530, dated July 18, 2000) and the study design conformed to the guidelines of the Declaration of Helsinki 1975 and subsequent amendments. Informed consent was obtained from the nearest relatives or guardians of the patients involved, since the patients themselves were unable to provide consent.

Variables and data source

All patients were treated according to standard institutional therapy protocol targeted at maintaining cerebral perfusion pressure. Arterial blood pressure, intracranial pressure, and cerebral perfusion pressure were continuously monitored. Clinical variables related to outcome were noted including GCS, pupil reactivity, CT findings, and the presence of extracerebral injury. All of them were on ventilatory support and were gradually weaned off in approximately 48 h. The recording of GCS and motor scores at 96 h was done when there was no effect of sedation or pharmacological paralysis. After initial management under Neurosurgery, patients were transferred to the Department of Physical Medicine and Rehabilitation for further care. Each patient underwent appropriate coma stimulation and therapy according to their neurological status. Rehabilitation goals were set depending on the progress the patient made with therapy. Once they were medically stable, a minimum of 6-week coma stimulation was given to patients who did not show any meaningful response to stimuli. They were then discharged with advice to continue therapy and care by relatives at home. Those who showed progressive improvement, were continued on therapy till maximum functional status was achieved.

Measurement of S100B protein in the serum

Venous blood was drawn at 48 h and 21 days post injury for estimation of serum levels of S100B. Serum was isolated within 2 h of blood collection and stored at −20°C until the time of analysis. Estimation of S100B levels was done using the commercially available SANGTEC^® 100 immunoradiometric assay.

Outcome measures

All patients were assessed for clinical and functional outcome at 6 months post injury using Glasgow Outcome Score (GOS), Functional Independence Measure (FIM), and Modified Mini-Mental State Examination (3MS). Poor functional outcome was defined as GOS ≤3, FIM <72, or 3MS <75, which indicated that the persons with TBI following their rehabilitation would be dependent in their daily activities within their home environment.^{[16],[17],[18],[19]}

Statistical methods

Receiver operator characteristic (ROC) curves were generated for serum S100B levels to determine the cutoff values with the highest sensitivity and specificity for predicting death and poor outcomes (based on GOS, FIM, and 3MS). These cutoff values were used to derive sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratios for the outcome measures. The association between the various prognostic factors of interest with outcome measures was studied. Univariate and multivariate linear regression analyses were carried out to determine the factors that contributed independently to prognostication of outcomes. Statistical analyses were carried out using the Statistical Package for the Social Scientist version 21 (IBM Corp., Armonk, NY, USA).

Results

Demographic, clinical, and radiological data

The demographic and clinical data of the patients are given in [Table 1]. The age of patients ranged from 17 to 54 years with a mean of 29.4 years. The mean delay in admission to hospital post injury was 2 h. The cause of TBI in majority of the patients was road traffic accident; majority of patients did not have polytrauma (83%).

Table 1: Distribution of demographic characteristics

Click here to view

Mean serum S100B levels at 48 h and 21 days post injury

Serum S100B levels tended to be elevated among the survivors (n = 17) compared to nonsurvivors (n = 6); however, this was not statistically significant [Figure 1]a. Serum S100B levels were significantly lower at 21 days compared to levels at 48 h post injury [Figure 1]b. Levels at 48 h were significantly higher in those with poor functional outcome (as assessed by GOS and FIM but not 3MS at 6 months post injury) compared to those who had good outcome [Figure 1]c. Levels at 21 days post injury were not significantly different in those who had good outcome compared to those who did not [Figure 1]d.

Figure 1: Serum S100B levels at 48 h and 21 days post injury. Mean serum S100B levels in survivors and nonsurvivors (a) and at 48 h and 21 days post injury (b). Mean S100B levels measured at 48 h (c) and 21 days (d) post injury in patients with good and poor outcomes (based on Glasgow Outcome Score, Functional Independence Measure, and 3MS) at 6 months. Interquartile range (box), median (black line), and data range (whiskers) are shown. Mann–Whitney test was used to test for statistical significance in all cases

Click here to view

ROC analyses of serum S100B levels for predicting mortality and functional outcomes at 6 months

S100B levels at 48 h with a cutoff value of 1.96 μg/L predicted mortality with a sensitivity of 67% and specificity of 82% (likelihood ratio [LR]: 3.72); however, the area under the curve (AUC) of the ROC curve was not statistically significant (AUC = 0.735, P = 0.093) [Figure 2]a.

Figure 2: ROC analyses of S100B at 48 h and 21 days for predicting poor functional outcome. Prediction of mortality (a) and poor outcome at 6 months based on Glasgow Outcome Score ≤3 (b), Functional Independence Measure ≤72 (c), and 3MS ≤75 (d) by S100B levels at 48 h post injury. Prediction of poor outcomes at 6 months based on GOS ≤3 (e), Functional Independence Measure ≤72 (f), and 3MS ≤75 (g) by S100B levels at 21 days post injury. In cases where the area under the curve of the ROC curve was significant (P < 0.05), optimal cutoff values and their sensitivity, specificity, and likelihood ratio were calculated

Click here to view

S100B at 48 h significantly predicted poor outcome at 6 months as assessed by GOS (AUC = 0.758, P = 0.036) [Figure 2]b and FIM (AUC = 0.846, P = 0.042) [Figure 2]c. The optimal cutoff value of 1.37 μg/L showed a sensitivity of 64%, specificity of 83%, and LR of 3.76 for predicting poor outcomes based on GOS. The same cutoff value had a sensitivity of 75%, specificity of 85%, and LR of 5.0 for predicting poor outcomes based on FIM. However, the AUC was not significant for predicting outcomes based on 3MS (AUC = 0.638, P = 0.246) [Figure 2]d.

On the other hand, the AUC for S100B levels at 21 days was not statistically significant for GOS (AUC = 0.636, P = 0.366) [Figure 2]e, FIM (AUC = 0.583, P = 628) [Figure 2]f, and 3MS (AUC = 0.618, P = 0.462) [Figure 2]g. It was noted that, while higher S100B levels at 48 h post injury were associated with lower GOS and FIM scores, higher levels on the 21^st day were associated with higher GOS and FIM scores.

Linear regression analyses of factors predicting Glasgow Outcome Score, Functional Independence Measure, and 3MS scores at 6 months

Univariate linear regression analyses showed that GCS motor score at 96 h post injury was the single most important predictor of GOS at 6 months (B = 0.771, 95% confidence interval: 0.435–1.107, P < 0.001, adjusted R² = 0.551). A multivariate model incorporating GCS scores, pupil reactivity, Marshall CT grade (at admission), and serum S100B levels showed higher predictive power after adjusting for age (adjusted R² = 0.748). In this model, GCS motor score at 96 h and serum S100B levels (both at 48 h and at 21 days) were independent predictors of GOS [Table 2].

Table 2: Linear regression analysis of factors predicting Glasgow Outcome Score at 6 months post injury

Click here to view

Univariate analysis of factors predicting FIM and 3MS scores at 6 months also showed that higher GCS scores (at admission and 96 h post injury) were significantly associated with better functional outcome. As with GOS, multivariate models incorporating GCS scores, pupil reactivity, Marshall CT grade (at admission), and serum S100B levels (at 48 h and 21 days) showed higher predictive power for FIM and 3MS (adjusted R² = 0.899 and 0.834, respectively). Here again, GCS motor score at 96 h and S100B levels (at 48 h and 21 days) were independent predictors of FIM and 3MS [Table 3].

Table 3: Linear regression analysis of factors predicting Functional Independence Measure and 3MS score at 6 months post injury

Click here to view

Combining prognostic markers to improve prediction

The combination of S100B levels at 48 h ≥1.37 μg/L and GCS motor scores at 96 h ≤3 was able to improve the positive predictive value to 100% with GOS, FIM, and 3MS as outcome measures. The sensitivity of this model was 45% with GOS, 33% with FIM, and 25% with 3MS. When we combined Marshall CT scan grades (III–IV) to the above model, although the specificity and positive predictive value remained 100%, the sensitivity reduced to 27% with GOS and remained the same with FIM and 3MS [Supplementary Table 1].

Discussion

Eighty-five percent of the world's population live in low- and middle-income countries where the pooled mortality for severe brain injury rate has been reported to be 51%, compared to 30% in high-income countries.^[17],[20] In India, the mortality has been reported as high as 73.5% in severe brain injury.^[21] In addition, India reports poor functional outcomes in patients with severe TBI compared to those in high-income countries.^[22],[23] An estimate of prognosis after head injury is central to clinical decisions. Reliable prediction of outcomes allows realistic counseling of relatives and appropriate allocation of scarce resources to those predicted to survive or recover with a good outcome.

GCS is, by far, the most commonly used clinical parameter to predict outcomes. When based on the initial GCS score, only 68.6% of those predicted to have a good outcome and 76% of those predicted to have poor outcomes actually had such outcomes at follow-up after 1 year. If the later GCS (assessment done once the patient is out of sedation) was used for prediction, there was a significant increase in the rate of correct predictions for a good outcome (80.6%), but the rate of correct predictions for a poor outcome remained essentially unchanged (78.6%).^[24] The results of the present study show that a person with brain injury who has a motor score at 96 h >3, is likely to have a good outcome with a LR of 9.12 for GOS, LR of 6.25 for FIM, and LR of 5 for 3MS at 6 months. Motor scores have been reported as better predictors than the summed GCS score due to variability inherent in the verbal (particularly for those on tracheostomy or ventilatory support) and eye-opening response (for sedated patients).^{[4],[18],[25]} To the best of our knowledge, our study is the first to report on a cutoff GCS motor score at 96 h which can discriminate between good and poor outcomes in patients with severe head injury.

S100B as a marker of neuronal injury

Elevated S100B levels have been reported in polytrauma patients with and without head injury; however, these values have been reported to come down significantly by 24–48 h.^[26],[27] Hence, in our study, S100B level at 48 h was taken as the best measure to reflect the extent of brain injury in the acute phase. S100B levels were reassessed at 21 days post injury, during the rehabilitation phase, to assess its correlation to functional outcomes. An added advantage of S100B is that it is not affected by alcohol consumption and hence it can be used during acute phase when clinical parameters such as GCS assessment may be impaired due to sedation or alcohol intoxication.^[28]

The mean serum S100B concentration for a normal healthy population was reported to be 0.03 μg/L (standard deviation: 0.04). The upper 97.5% and 95% reference limits were 0.13 and 0.10 μg/L, respectively. No major age or sex differences were observed.^[26] In the present study, the S100B levels at 48 h ranged from 0.2 to 8 μg/L with a mean of 1.5 μg/L [Figure 2]b. Since S100B is excreted and eliminated fully by the kidneys,^[13] we looked at the creatinine levels for all patients included in the study and found that they were within the normal range. In another review article looking at the clinical utility of S100B levels, sustained or rising value of this protein after 27 h post injury or secondary peaks were reported as indicators of ongoing brain damage or of new cerebral lesions.^[13]

The best cutoff value for S100B in our study as calculated with a receiver operator characteristic curve was 1.96 μg/L for mortality. Our findings are similar to those reported in the systematic review and meta-analysis done by Mercier et al. in which the first measured S100B level after head injury was taken for analysis.^[17] In this review article, mortality prediction in six studies reported a mean specificity of 91% and a sensitivity of 39% for S100B cutoff values that ranged from 2.5 to 3.0 μg/L. In our study, S100B levels above 1.37 μg/L at 48 h significantly predicted poor outcomes at 6 months as assessed by GOS (sensitivity of 64%, specificity of 83%, and LR of 3.76), FIM (sensitivity of 75%, specificity of 85%, and LR of 5.0), and 3MS (sensitivity of 60%, specificity of 83%, and LR of 3.53). With S100B levels above 2.2 μg/L, the specificity for poor outcome increased to 100%, but the sensitivity reduced to 45% for GOS, 50% for FIM, and 40% for 3MS. All the studies in the systematic review done by Mercier et al. only used GOS as the outcome measure. This meta-analysis reported a specificity of 94% and a sensitivity of 38% for poor outcome. The value of S100B ranged from 2.5 to 3.0 μg/L.^[17] The variation in the S100B values in the studies included in this review in comparison to our study can be explained by the difference in the time period in the measurement of S100B levels. Low sensitivity indicating false-negative values of S100B levels may be due to localized contusions in strategic areas like the brainstem when the outcome can be poor despite a small volume of injured tissue.

Is S100B in addition a neuroprotective marker?

Although serum S100B levels at 21 days did not differ significantly between those who had poor and good outcomes, linear regression analysis revealed a significant positive association with good outcomes [Table 2] and [Table 3]. Of note, the levels at 48 h were negatively associated and those at 21 days were positively associated with both GOS and FIM. S100B has a short half-life^[29] and its sustained high serum levels indicate ongoings release from the injured glial and neuronal tissue.^[30] However, the association of higher levels of S100B in the blood at 21 days post injury with better functional outcomes, which is seen in our study, may reflect its role in brain injury recovery.

The neuroprotective role of S100B can be explained by multiple mechanisms. These include enhanced survival of neurons by increasing the stability of the tubulin polymers that form microtubules of the cytoskeleton,^[31] stimulation of nerve regeneration,^[32],[33] glial proliferation,^[34] and facilitation of neurite outgrowth and neuronal maturation.^[30],[35] Although there have been many in vitro and animal studies which have shown the neurotrophic and gliotrophic action of S100B protein, this is the first clinical study that may support this evidence.

Models combining prognosticating factors

In this study, a model incorporating GCS scores at admission and at 96 h, pupil reactivity, Marshall CT grade (at admission), and serum S100B levels (at 48 h and 21 days post injury) was able to significantly predict the outcomes as assessed by GOS, FIM, and 3MS. GCS motor scores, when combined with serum S100B levels, contributed significantly to predicting functional outcomes in severe TBI. The probability of a person with severe head injury with GCS motor score ≤3 at 96 h and S100B at 48 h >1.37 μg/L, having a poor functional outcome (GOS ≤3 or FIM <72 or 3MS <75), is nearly certain [Supplementary Table 1].

In our study, the S100B protein contributed significantly to improving the prediction of outcome when combined with the GCS motor score at 96 h in persons with severe head injury. However, as the sample size studied was small, further larger multicentric studies are needed to corroborate these findings. Another limitation was that the inclusion of admission CT scan did not contribute to improving the predictive value of the outcomes. Hence, we propose that inclusion of the “worst CT scan” during entire course of admission may correlate better with long-term outcomes.^[36]

Conclusion

In severe TBI, serum S100B level at 48 h is a good predictor of long-term functional outcome. With the addition of GCS motor score at 96 h, this model could better inform the treating team and relatives on expected clinically important outcomes and, therefore, optimize the provision of appropriate healthcare.

Research quality and ethics statement

All authors of this manuscript declare that this scientific study is in compliance with standard reporting guidelines set forth by the EQUATOR Network. The authors ratify that this study required Institutional Review Board/Ethics Committee review, and hence prior approval was obtained IRB Min. No. 4530 dated July 18, 2000. We also declare that we did not plagiarize the contents of this manuscript and have performed a plagiarism check.

Acknowledgments

The authors would like to thank Dr. Joe Varghese for help with statistical analyses of the data.

Financial support and sponsorship

This study was funded by an intramural grant from the Christian Medical College, Vellore.

Conflicts of interest

There are no conflicts of interest.

References

1.	Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The Glasgow Coma Scale at 40 years: Standing the test of time. Lancet Neurol 2014;13:844-54.
2.	Marshall LF, Gautille T, Klauber MR, Eisenberg HM, Jane JA, Luerssen TG, et al. The outcome of severe closed head injury. J Neurosurg 1991;75:S28.
3.	Vollmer DG, Torner JC, Jane JA, Sadovnic B, Charlebois D, Eisenberg HM, et al. Age and outcome following traumatic coma: why do older patients fare worse. J Neurosurg 1991;75:S37.
4.	Choi SC, Muizelaar JP, Barnes TY, Marmarou A, Brooks DM, Young HF. Prediction tree for severely head-injured patients. J Neurosurg 1991;75:251.
5.	Chesnut RM, Marshall SB, Piek J, Blunt BA, Klauber MR, Marshall LF. Early and late systemic hypotension as a frequent and fundamental source of cerebral ischemia following severe brain injury in the Traumatic Coma Data Bank. Acta Neurochir Suppl (Wien) 1993;59:121.
6.	Chesnut RM, Marshall LF, Klauber MR, Blunt BA, Baldwin N, Eisenberg HM, et al. The role of secondary brain injury in determining outcome from severe head injury. J Trauma 1993;34:216.
7.	Clifton GL, Miller ER, Choi SC, Levin HS. Fluid thresholds and outcome from severe brain injury. Crit Care Med 2002;30:739.
8.	Marion DW, Carlier PM. Problems with initial Glasgow Coma Scale assessment caused by prehospital treatment of patients with head injuries: results of a national survey. J Trauma 1994;36:89.
9.	Egea-Guerrero JJ, Rodríguez-Rodríguez A, Gordillo-Escobar E, Fernández-Delgado E, Martínez-Roldán Á, Roldán-Reina Á, et al. IMPACT Score for traumatic brain injury: Validation of the prognostic tool in a Spanish cohort. J Head Trauma Rehabil 2018;33:46-52.
10.	MRC CRASH Trial Collaborators; Perel P, Arango M, Clayton T, Edwards P, Komolafe E, Poccock S, et al. Predicting outcome after traumatic brain injury: Practical prognostic models based on large cohort of international patients. BMJ 2008;336:425.
11.	Steyerberg EW, Mushkudiani N, Perel P, Butcher I, Lu J, McHugh GS, et al. Predicting outcome after traumatic brain injury: Development and international validation of prognostic scores based on admission characteristics. PLoS Med 2008;5:e165.
12.	Dadas A, Washington J, Diaz-Arrastia R, Janigro D. Biomarkers in traumatic brain injury (TBI): A review. Neuropsychiatr Dis Treat 2018;14:2989-3000.
13.	Thelin EP, Nelson DW, Bellander BM. A review of the clinical utility of serum S100B protein levels in the assessment of traumatic brain injury. Acta Neurochir (Wien) 2017;159:209-25.
14.	Wang KK, Yang Z, Zhu T, Shi Y, Rubenstein R, Tyndall JA, et al. An update on diagnostic and prognostic biomarkers for traumatic brain injury. Expert Rev Mol Diagn 2018;18:165-80.
15.	Dey S, Gangadharan J, Deepika A, Kumar JK, Christopher R, Ramesh SS, et al. Correlation of ubiquitin C terminal hydrolase and S100β with cognitive deficits in young adults with mild traumatic brain injury. Neurol India 2017;65:761-6. [PUBMED] [Full text]
16.	McDowell I. Canadian study of health and aging: study methods and prevalence of dementia. CMAJ 1994;150:899.
17.	Mercier E, Boutin A, Lauzier F, Fergusson DA, Simard JF, Zarychanski R, et al. Predictive value of S-100β protein for prognosis in patients with moderate and severe traumatic brain injury: Systematic review and meta-analysis. BMJ 2013;346:f1757.
18.	Rothoerl RD, Woertgen C, Brawanski A. S-100 serum levels and outcome after severe head injury. Acta Neurochir Suppl 2000;76:97-100.
19.	Whitlock JA Jr., Hamilton BB. Functional outcome after rehabilitation for severe traumatic brain injury. Arch Phys Med Rehabil 1995;76:1103-12.
20.	De Silva MJ, Roberts I, Perel P, Edwards P, Kenward MG, Fernandes J, et al. Patient outcome after traumatic brain injury in high-, middle- and low-income countries: Analysis of data on 8927 patients in 46 countries. Int J Epidemiol 2009;38:452-8.
21.	Shekhar C, Gupta LN, Premsagar IC, Sinha M, Kishore J. An epidemiological study of traumatic brain injury cases in a trauma centre of New Delhi (India). J Emerg Trauma Shock 2015;8:131-9. [PUBMED] [Full text]
22.	Agrawal A, Savardekar A, Singh M, Pal R, Shukla DP, Rubiano AM, et al. Pattern of reporting and practices for the management of traumatic brain injury: An overview of published literature from India. Neurol India 2018;66:976-1002. [PUBMED] [Full text]
23.	Agrawal D, Ahmed S, Khan S, Gupta D, Sinha S, Satyarthee GD. Outcome in 2068 patients of head injury: Experience at a level 1 trauma centre in India. Asian J Neurosurg 2016;11:143-5. [PUBMED] [Full text]
24.	Thatcher RW, Cantor DS, McAlaster R, Geisler F, Krause P. Comprehensive predictions of outcome in closed head-injured patients. The development of prognostic equations. Ann N Y Acad Sci 1991;620:82-101.
25.	Jagger J, Jane JA, Rimel R. The Glasgow coma scale: To sum or not to sum? Lancet 1983;2:97.
26.	Anderson RE, Hansson LO, Nilsson O, Dijlai-Merzoug R, Settergren G. High serum S100B levels for trauma patients without head injuries. Neurosurgery 2001;48:1255-8.
27.	Pelinka LE, Toegel E, Mauritz W, Redl H. Serum S 100 B: A marker of brain damage in traumatic brain injury with and without multiple trauma. Shock 2003;19:195-200.
28.	Calcagnile O, Holmén A, Chew M, Undén J. S100B levels are affected by older age but not by alcohol intoxication following mild traumatic brain injury. Scand J Trauma Resusc Emerg Med 2013;21:52.
29.	Raabe A, Grolms C, Sorge O, Zimmermann M, Seifert V. Serum S-100B protein in severe head injury. Neurosurgery 1999;45:477-83.
30.	Rezaei O, Pakdaman H, Gharehgozli K, Simani L, Vahedian-Azimi A, Asaadi S, et al. S100 B: A new concept in neurocritical care. Iran J Neurol 2017;16:83-9.
31.	Brewton LS, Haddad L, Azmitia EC. Colchicine-induced cytoskeletal collapse and apoptosis in N-18 neuroblastoma cultures is rapidly reversed by applied S-100beta. Brain Res 2001;912:9-16.
32.	Haglid KG, Yang Q, Hamberger A, Bergman S, Widerberg A, Danielsen N. S-100beta stimulates neurite outgrowth in the rat sciatic nerve grafted with acellular muscle transplants. Brain Res 1997;753:196-201.
33.	Kleindienst A, McGinn MJ, Harvey HB, Colello RJ, Hamm RJ, Bullock MR. Enhanced hippocampal neurogenesis by intraventricular S100B infusion is associated with improved cognitive recovery after traumatic brain injury. J Neurotrauma 2005;22:645-55.
34.	Selinfreund RH, Barger SW, Pledger WJ, Van Eldik LJ. Neurotrophic protein S100 beta stimulates glial cell proliferation. Proc Natl Acad Sci U S A 1991;88:3554-8.
35.	Marshak DR. S100 beta as a neurotrophic factor. Prog Brain Res 1990;86:169-81.
36.	Servadei F, Murray GD, Penny K, Teasdale GM, Dearden M, Iannotti F, et al. The value of the “worst” computed tomographic scan in clinical studies of moderate and severe head injury. European Brain Injury Consortium. Neurosurgery 2000;46:70-5.