Neurocritical Care

, Volume 14, Issue 2, pp 200–207

Association of CSF Biomarkers and Secondary Insults Following Severe Traumatic Brain Injury

Authors

    • Division of Critical Care/Program in Trauma, R Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Joseph A. Kufera
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Allison Lindell
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Karen R. Murdock
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Jay Menaker
    • Division of Critical Care/Program in Trauma, R Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Grant V. Bochicchio
    • Division of Critical Care/Program in Trauma, R Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
  • Bizhan Aarabi
    • Division of Critical Care/Program in Trauma, R Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine
  • Thomas M. Scalea
    • Division of Critical Care/Program in Trauma, R Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine
    • The Shock, Trauma and Anesthesiology Research Organized Research CenterUniversity of Maryland School of Medicine
Original Article

DOI: 10.1007/s12028-010-9496-1

Cite this article as:
Stein, D.M., Kufera, J.A., Lindell, A. et al. Neurocrit Care (2011) 14: 200. doi:10.1007/s12028-010-9496-1

Abstract

Background

Management of severe traumatic brain injury (TBI) focuses on mitigating secondary insults. There are a number of biomarkers that are thought to play a part in secondary injury following severe TBI. Two of these, S100β and neuron-specific enolase (NSE), have been extensively studied in the setting of neurological injury. This pilot study was undertaken to investigate the relationship of S100β and NSE to clinical markers of severity and poor outcome: intracranial hypertension (ICH), and cerebral hypoperfusion (CH).

Methods

Patients at the R Adams Cowley Shock Trauma Center were prospectively enrolled over an 18-month period. Inclusion criteria were: age > 18, admission within the first 6 h after injury, Glasgow Coma Scale (GCS) < 9 on admission, isolated TBI, and placement of an intraventricular catheter (IVC). Patients were managed according to an institutional protocol based on the Brain Trauma Foundation Guidelines. CSF was collected from the IVC on admission and twice daily for 7 days. S100β and NSE levels were analyzed by ELISA. CSF levels drawn before (PRE) and after (POST) 12-h time periods were compared to percentage time intracranial pressure (ICP) > 20 mmHg (% ICP20) and cerebral perfusion pressure (CPP) < 60 mmHg (% CPP60), and cumulative “Pressure times Time Dose” (PTD) for episodes of ICP > 20 mmHg (PTD ICP20) and CPP < 60 mmHg (PTD CPP60). Statistical analysis was performed using the Student’s t test to compare means and non-parametric Wilcoxon statistic to compare ranked data. Linear regression methods were applied to compare levels of S100β and NSE with ICP and CPP.

Results

Twenty-three patients were enrolled. The cohort of patients was severely injured and neurologically compromised on admission (admission GCS = 5.6 ± 3.1, Injury Severity Score (ISS) = 31.9 ± 10.6, head Abbreviated Injury Scale (AIS) = 4.4 ± 0.7, Marshall score = 2.6 ± 0.9). Elevated levels of S100β and NSE were found in all 223 CSF samples analyzed. ICH was found to be associated with PRE and POST S100β levels when measured as % ICP20 (r = 0.20 and r = 0.23, P < 0.01) and PTD ICP20 (r = 0.35 and r = 0.26, P < 0.001). POST increasing NSE levels were weakly correlated with increasing PTD ICP20 (r = 0.17, P = 0.01). PRE S100β levels were associated with episodes of CH as measured by % CPP60 (r = 0.20, P = 0.002) and both PRE and POST S100β levels were associated with PTD CPP60 (r = 0.24 and r = 0.23, P < 0.001). PRE and POST NSE levels were also associated with episodes of CH as measured by % CPP60 (r = 0.22 and r = 0.18, P < 0.01) and PTD CPP60 (r = 0.20 and r = 0.21, P < 0.01).

Conclusions

In this preliminary analysis, S100β levels were associated with ICH and CH over a full week of ICP monitoring. We also found associations between CH and NSE levels in CSF of patients with severe TBI. Our results suggest that there is an association between levels of ICH and CH and these biomarkers when measured before episodes of clinically significant secondary insults. These markers of neuronal cell death demonstrate promise as both indicators of impending clinical deterioration and targets of future therapeutic interventions.

Keywords

Traumatic brain injuryBiomarkersCerebral perfusion pressureIntracranial pressureSecondary insults

Introduction

Traumatic brain injury (TBI) is the leading cause of death and disability in children and young adults and accounts for over 50,000 deaths each year in the United States [1]. Severe TBI is the leading cause of death following injury and contributes substantially to over 50% of trauma-related deaths [2, 3]. TBI is a major public health problem with estimates of 5.3 million people living in United States who require long-term or lifelong assistance due to the effects of a TBI [4].

Currently, there is little that can be done to treat the primary insult to the brain that occurs at the time of injury. Secondary injuries are those that occur in the minutes to hours and days following the primary injury. The secondary insults that occur following TBI involve an exceptionally complex interplay of numerous factors and substances including alterations in cerebral blood flow, biochemical derangements, edema, oxidative stress, release of excitotoxic mediators, inflammation, apoptosis, and necrosis [5, 6]. It is a well-recognized fact that these secondary insults contribute significantly to outcome [710]. Therefore, the management strategies for patients with severe TBI are targeted toward prevention and mitigation of these secondary insults. Initial management of the patient with severe TBI focuses on prevention of systemic hypotension and hypoxia which have been repeatedly demonstrated to be associated with worse outcome [11, 12]. Management strategies then target prevention and treatment of intracranial hypertension (ICH) and cerebral hypoperfusion (CH) [10, 1317]. It is thought that these episodes of ICH and CH not only cause further secondary insult, but are also the clinical manifestation of additional ongoing neuronal injury.

The Brain Trauma Foundation has made recommendations for treatment thresholds for intracranial pressure (ICP) and cerebral perfusion pressure (CPP) based on the best available evidence [10]. These treatment thresholds guide the need for therapeutic interventions in the patient with severe TBI. The thresholds have been established largely based on clinical observations and trials that have demonstrated worse outcome in patients with ICH and CH [13, 14, 16]. However, few studies have focused on the biochemical and inflammatory processes that occur in the brain above and below these treatment thresholds.

There are a huge number of biomarkers, inflammatory mediators, and proteins that have been implicated in playing a role in the secondary insults following TBI. S100β is a Ca++-binding protein that is found in astrocytes, oligodendrocytes, and Schwann cells [18] and neuron-specific enolase (NSE) is a cytoplasmic glycolytic neuronal enzyme [19]. Each of these has been studied extensively in the setting of TBI as well as other neurological insults [6, 2027]. These substances have been found to correlate with outcome when measured in serum [6, 2025] and cerebrospinal fluid (CSF) [23, 25, 26] of patients with severe TBI. Few studies have attempted to correlate the levels of these substances with clinical measures of severity of TBI; however, none has studied the relationship of ICH and CH to CSF levels of S100β and NSE repeatedly over an extended period of time.

This pilot study was undertaken to determine the feasibility of measurement of CSF levels of S100β and NSE over time and investigate the relationship with ICP and CPP in patients with severe TBI. In addition, in order to learn more about the biochemical and pathophysiologic derangements that occur during episodes of ICH and CH, we conducted this preliminary evaluation to determine whether elevations of these biomarkers were either caused by episodes of ICH and CH (POST measurements) or whether the cellular injury that causes elevations of these markers contributed to these clinically evident derangements (PRE measurements).

Methods

Patient Population

After approval by the University of Maryland School of Medicine Institutional Review Board (IRB), patients at the R Adams Cowley Shock Trauma Center were prospectively enrolled over a 18-month period. Informed consent was obtained from the patient’s Legally Authorized Representative before collection of any data or specimens. Inclusion criteria were age greater than 17 years, admission within the first 6 h after injury, Glasgow Coma Scale (GCS) < 9 on admission, and placement of a clinically indicated intraventricular catheter (IVC). TBI was verified by computerized tomography (CT). Exclusion criteria were any body region other than brain with an Abbreviated Injury Severity (AIS) Score > 3 [28] to exclude multisystem trauma, a non-survivable brain injury, or ICP monitoring initiated > 24 h post-injury.

Data Collection

Baseline demographics and injury-specific data including mechanism of injury, head AIS score, predicted survival (by TRISS methodology), and Injury Severity Score (ISS) [29] were recorded. CT findings, admission vital signs, and GCS were also recorded. Marshall Classification scores were assigned to all admission head CTs by a blinded reviewer [30]. Hourly measurements of ICP and CPP were also recorded starting on admission and continuing for 7 days. Outcome measures included in-hospital mortality, discharge GCS, and length of hospital and intensive care unit (ICU) stay. The Extended Glasgow Outcome Score (GOSE) was used to evaluate long-term functional outcome at 3 months, 6 months, and 1 year following injury [31]. The GOSE of survivors was obtained in structured phone interviews by an experienced trauma clinical research coordinator.

Patient Management Protocol

At the R Adams Cowley Shock Trauma Center, patients with severe TBI are managed according to an institutional protocol based on the Brian Trauma Foundation Guidelines [10]. The protocol targets maintenance of ICP < 20 mmHg and CPP > 60 mmHg utilizing initially adequate sedation and analgesia, mechanical ventilation targeting PaCO2 of 35–40 mmHg, head elevation (30–45°), and maintenance of normovolemia, normotension, and normoxia. First tier therapies for episodes of ICH (ICP > 20 mmHg) are treated with insertion of external ventricular drainage/IVC (Codman, Raynham, MA), increasing doses of sedation, and/or hyperosmolar therapy with mannitol or hypertonic saline. Second tier therapies for intractable ICH include moderate hyperventilation (PaCO2 < 35 mmHg), barbiturate coma induction, decompressive craniectomy, and/or decompressive laparotomy [32, 33]. CH (CPP < 60 mmHg) is managed with aggressive treatment of ICH as above, assuring euvolemia, and use of vasopressors as guided by hemodynamic monitoring.

Specimen Collection and Processing

CSF was collected upon insertion of the IVC or as soon as possible after consent was obtained and twice daily at standard times for 7 days or until the IVC was removed as clinically indicated. Fresh CSF specimens were drawn using sterile technique from the drip chamber of the drainage system and immediately centrifuged to remove any cellular debris. They were then frozen at −80°C and stored until processing.

S100β and NSE levels were analyzed by enzyme-linked immunosorbant assay (ELISA) using commercially available kits (S100β, Biovendor, Candor, NC; NSE, American Research Products, Inc., Belmont, MA). The analytical limits of detections are 5 pg/ml for S100β and 1 mcg/l for NSE. All the samples were run in duplicate.

Statistical Analysis

Hourly values for ICP and CPP were recorded and means, minimum (for CPP) or maximum (for ICP) values, and percentage time ICP > 20 mmHg (% ICP20) and CPP < 60 mmHg (% CPP60) were calculated for 12-h time periods and compared to the CSF S100β and NSE levels that were drawn before the 12-h time period (PRE) and after the 12-h time period (POST). We also calculated ICP and CPP recordings as cumulative “Pressure times Time Dose” (PTD; mmHg h) to describe both the cumulative amplitude and duration of episodes above and below clinically defined thresholds, as a better measure of the scope of ICH and CH than calculations of frequency or duration of episodes of diminished CPP [34, 35]. PTD was calculated for episodes of ICP > 20 mmHg (PTD ICP20) and CPP < 60 mmHg (PTD CPP60) for 12-h time periods and similarly compared to CSF biomarker sample levels drawn before (PRE) and after (POST) the 12-h time period.

The Student’s t test was used to compare means based on continuous data. Owing to the small sample size, analysis of continuous data was also made using the non-parametric Wilcoxon statistic, but findings were similar; thus means and standard deviations were calculated. Linear regression methods were applied to compare PRE and POST levels of S100β and NSE with calculated values of ICP and CPP. Although it is possible that the relationship between biomarkers and treatment processes may be nonlinear, linear regressions, and correlation coefficients were analyzed in this pilot study to ascertain if an association existed, but not to determine the form of that association. Results provided reflected all data combined, which often included multiple observations per patient. Similar analyses using a mixed model approach, taking into account the correlation of measurements within a patient (i.e., repeated measurements), yielded similar results in all cases. However, because no single measurement of r has been developed to analyze correlated data within the mixed model framework [36, 37], reported r-values and P-values were based on all data combined. A probability level of less than 0.05 was considered statistically significant for all tests.

Results

Patient Enrollment

Twenty-three patients were enrolled in this pilot study over 18 months. Demographic and injury data for the study subjects are shown in Table 1. Mean age was 32 with a range of 16–59 years and most study subjects were male. As predicted by the inclusion criteria for this study, this cohort of patients was severely injured and neurologically compromised on admission. Outcome data are shown in Table 2. Median hospital length of stay was approximately 3 weeks and ICU length of stay was just over 2 weeks. There were five in-hospital deaths all from brain injury. Good functional outcome was noted in 10 of the 23 study subjects and 55.6% of survivors.
Table 1

Baseline characteristics of study subjects

 

All (N = 23)

Age (years), mean ± SD

31.6 ± 13.2

Males, n (%)

20 (87.0)

Mechanism of injury, n (%)

 Motor vehicle/motorcycle crash

12 (52.2)

 Fall

5 (21.7)

 Assault

4 (17.4)

 Other

2 (8.7)

Admission RTS, mean ± SD

4.98 ± 1.44

ISS, mean ± SD

31.9 ± 10.6

Predicted survival (TRISS), mean ± SD

0.63 ± 0.27

Admission SBP, mean ± SD

131.0 ± 29.9

Admission GCS, mean ± SD

5.6 ± 3.1

Marshall score, mean ± SD

2.6 ± 0.9

Head AIS, mean ± SD

4.4 ± 0.7

RTS Revised Trauma Score, ISS Injury Severity Score, TRISS Trauma Score-Injury Severity Score, SBP Systolic Blood Pressure, GCS Glasgow Coma Score, AIS Abbreviated Injury Scale

Table 2

Outcomes for study subjects

 

All (N = 23)

Surgical intervention, n (%)

12 (52.2)

 On admission

6 (12.1)

 Decompressive craniectomy

11 (47.8)

LOS (days), median (IQR)

20.0 (12.5–24.0)

ICULOS (days), median (IQR)

16.5 (9.5–23.8)

Good functional outcome*, n (%)

10 (43.4)

In-hospital mortality, n (%)

5 (21.7)

* GOSE > 4 at 6 months; LOS length of stay, ICU intensive care unit

Correlation with Clinical Measures: ICP and CPP

Two hundred and twenty-three CSF samples were taken and analyzed for levels of S100β and NSE. Elevated levels of S100β and NSE were found in all CSF samples (ranges 35.9–14,951.9 pg/ml and 2.2–928.1 mcg/ml, respectively).

Mean ICP was found to be associated with PRE and POST S100β levels (r = 0.22, P < 0.001 and r = 0.21, P = 0.001, respectively) and maximum ICP over the 12-h period of monitoring was correlated with POST S100β levels (r = 0.16, P = 0.02). Both PRE and POST S100β levels were associated with ICP when measured as % ICP20 (r = 0.20, P = 0.002 and r = 0.23, P < 0.001, respectively) and PTD ICP20 (r = 0.35, P < 0.001 and r = 0.26, P < 0.001, respectively). The associations between mean ICP and PTD ICP20 and PRE S100β levels were also seen with repeated measures taking into account individual patient correlations. Only POST NSE levels were weakly correlated with ICH with a correlation coefficient of 0.17 (P = 0.01) for PTD ICP20. See Figs. 1 and 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs12028-010-9496-1/MediaObjects/12028_2010_9496_Fig1_HTML.gif
Fig. 1

Correlation of pressure times time dose of ICP > 20 mmHg versus PRE S100β (r = 0.35, P < 0.001) and NSE (r = 0.12, ns) levels. (S100β trendline = solid line, NSE trendline = dash line)

https://static-content.springer.com/image/art%3A10.1007%2Fs12028-010-9496-1/MediaObjects/12028_2010_9496_Fig2_HTML.gif
Fig. 2

Correlation of pressure times time dose of ICP > 20 mmHg versus POST S100β (r = 0.26, P < 0.001) and NSE (r = 0.17, P = 0.01) levels. (S100β trendline = solid line, NSE trendline = dash line)

Similar associations were found with episodes of CH and S100β and NSE in the CSF. Figures 3 and 4 show the relationship of PRE and POST S100β and NSE levels with CPP, respectively. Mean CPP and was found to be correlated with both PRE and POST S100β levels (r = −0.21, P = 0.002 and r = −0.22, P < 0.001, respectively) and minimum CPP were also, albeit weakly, correlated with both PRE and POST S100β levels (r = −0.13, P < 0.05 and r = −0.15, P = 0.03, respectively). Only PRE S100β levels were associated with CH when measured as % CPP60 (r = 0.20, P = 0.002), while PTD CPP60 was associated with both PRE and POST levels (r = 0.24, P < 0.001 and r = 0.23, P < 0.001, respectively). PRE NSE levels were weakly correlated with mean CPP (r = −0.14, P = 0.04) and failed to correlate with minimum CPP values over the 12-h periods of monitoring. CH, as measured by % CPP60, was associated with both PRE and POST NSE levels (r = 0.22, P < 0.001 and r = 0.18, P = 0.005, respectively) as was PTD CPP60 (r = 0.20, P = 0.002 and r = 0.21, P = 0.002, respectively). These same associations between S100β and NSE levels in the CSF and % time and PTD CPP60 were seen with repeated measures.
https://static-content.springer.com/image/art%3A10.1007%2Fs12028-010-9496-1/MediaObjects/12028_2010_9496_Fig3_HTML.gif
Fig. 3

Correlation of pressure times time dose of CPP < 60 mmHg versus PRE S100β (r = 0.24, P < 0.001) and NSE (r = 0.20, P = 0.002) levels. (S100β trendline = solid line, NSE trendline = dash line)

https://static-content.springer.com/image/art%3A10.1007%2Fs12028-010-9496-1/MediaObjects/12028_2010_9496_Fig4_HTML.gif
Fig. 4

Correlation of pressure times time dose of CPP < 60 mmHg versus POST S100β (r = 0.23, P < 0.001) and NSE (r = 0.21, P = 0.002) levels. (S100β trendline = solid line, NSE trendline = dash line)

Discussion

The secondary injury that occurs following TBI involves a complex cascade of processes that ultimately leads to edema, ischemia, and neuronal cell death. The various proteins and chemicals involved in these processes are the focus of numerous investigations. S100β and NSE are two biomarkers that have been extensively studied in the setting of TBI as markers of neuronal cell injury and death.

Elevation of S100β levels in the setting of TBI has been widely reported. Experimental models suggest that the calcium-binding protein found primarily in astroglial and Schwann cells modulates neuronal–glial interactions and is likely to have both neurotrophic and neurotoxic functions [3840, 42]. It has been described in many studies as a good marker of brain tissue damage and elevated levels in both serum and CSF have been associated with poor outcome [6, 2126, 41]. NSE, a glycolytic enzyme found almost exclusively in the neuronal cytoplasm, is thought to be more specific to neuronal injury than the astrocyte-derived S100β [26]. Many studies have also investigated both CSF and serum levels of the role of NSE in TBI and, like S100β, found elevated levels to be highly correlated with poor outcome [6, 20, 21, 23, 26].

The purpose of this pilot study was to determine whether CSF levels of S100β and NSE, correlated with clinical markers of severity of TBI, namely ICH and CH. Although many studies have examined the role of these biomarkers of cellular injury and death in patients with TBI, few have examined the relationship of serum and/or CSF levels with clinical markers of severity of injury. Hayakata and colleagues measured CSF levels of S100β at 6 times periods (4 days) following severe TBI and found increased levels of S100β in the dichotomized subset of patients with ICH [25]. Pelinka et al. examined serum levels of S100β and glial fibrillary acidic protein (GFAP) and demonstrated that levels were higher in patients with ICP ≥ 25 mmHg and CPP < 60 mmHg when the maximum and minimum daily values were compared to serum levels for that day [22]. Kirchhoff et al. examined levels of S100β in the serum and CSF of patients with severe TBI over a 3-day period and were able to correlate increased levels with patients who develop progressive intracranial hemorrhage [27]. In this study of 21 patients, those with ICH on admission, defined as ICP > 15 mmHg, were found to have higher mean S100β levels than those without ICH or normal controls. No studies examining the relationship of NSE levels with ICP or CPP were found in the literature.

In this preliminary analysis, S100β levels were associated with % time of ICH and CH and with PTD of ICP > 20 mmHg and CPP < 60 mmHg over a full week of ICP monitoring. ICP and CPP recordings as PTD, an “area under the curve” function that describes and defines cumulative depth and duration of ICH and CH, has been demonstrated to be better associated with outcome than the number or duration of events [34]. Although these associations were statistically significant, none of the correlations was particularly strong, which is most certainly due to both the limited number of study subjects in this pilot project and the extreme heterogeneity of this disease process. The weak correlations of NSE with ICH and CH do support the conclusions of other studies that elevated levels of NSE in CSF are indicative of neuronal damage as is known to occur during episodes of ICH and CH [23, 41].

It would be expected that episodes of ICH and CH would lead to cellular injury. As tissue becomes inflamed and ischemic, it would also be expected that edema leading to further ICH and CH would occur. We attempted to determine in this pilot study whether elevated levels of biomarkers occurred before or after episodes of high ICP and low CPP to determine whether ICH and CH were more causative of or resultant from neuronal injury. PRE levels of S100β were more highly associated with ICH than POST levels. This may indicate that the measurement of S100β in the CSF, a byproduct of neuronal injury, may precede clinically evident insults. However, the fact that POST levels were also associated with episodes of high ICP and low CPP implies that there is a component of neuronal injury that may result from these secondary insults, namely ICH and CH.

The potential clinical relevance of this study is the suggestion in this data that we may be able to “predict” impending secondary insults following TBI before the clinical manifestation of these events. Given the known morbidity of ICH and CH, early intervention, or more importantly, prevention may have a significant impact on outcome in this severely injured patient population. This becomes even more important in austere environments in which the availability of aggressive intervention may be limited [42]. In addition, with respect to expeditionary care and aeromedical evacuation of critical injured military casualties, the finding that levels of biomarkers in the CSF increase before episodes of clinical apparent ICH and CH may assist our military medical personnel in developing ideal “time-to-fly” scenarios [43].

This investigation is quite limited by the sample size and was initiated as a preliminary investigation upon which to design future studies. The nature of our findings suggests that S100β may be a better candidate marker for ICH and CH than NSE and is worthy of more research. In this preliminary analysis, CSF S100β levels were associated with ICH and CH over a full week of ICP monitoring. We also found associations, albeit weak ones, between CH and NSE levels in CSF of patients with severe TBI. Our results suggest that there is an association between levels of ICH and CH and these biomarkers when measured before episodes of clinically significant secondary insults. These markers of neuronal cell death demonstrate promise as both indicators of impending clinical deterioration and targets of future therapeutic interventions.

Acknowledgment

This study was funded in part by W81XWH-07-2-0118.

Copyright information

© Springer Science+Business Media, LLC 2011