Serum neuron-specific enolase concentrations as a predictor of mortality in children with traumatic brain injury

Summary

Objective:
The aims of this study were the correlations of neuron-specific enolase (NSE) serum levels and mortality, Glasgow Outcome Scale (GOS) and lenght of hospitalization in children with traumatic brain injury (TBI).

Material and method:
The study protocol and informed consent were approved by the ethics committee of University Hospital Brno. During the period from May 2007 to October 2009 sixty-three patients with TBI were enrolled into the prospective study. TBI was verified by computerized tomography. Venous blood samples were taken after admission and every 24 h for a maximum of 6 consecutive days. Serum concentrations of neuron-specific enolase were quantified immuno-luminometrically. Six months after the primary injury, the study protocol documented outcome according to the Glasgow Outcome Score.

Results:
NSE values is significantly higher in patients who died (median 61.84; min 25.55 – max 271.3 µg/l; p=<0.02) compared to patients who survived (median 33.51; min 12.51 – max 97.9 µg/l). Significantly different dynamics of this protein was only in patients who died while the increase is still noticeable during the first two days. For survivors, the values of this protein decreases from the beginning. At D4 there is a few patients with mild elevation, which, however, does not show as a “second peak” effect in the overall average. In relation to the length of hospitalization does not show any diference.

Conclusion:
Biomarker neuron-specific enolase may have the potential to be used as quantitative measures of the success of therapy for TBI and might serve as an objective marker of the predictor mortality and outcome.

Key words:
neuron-specific enolase, traumatic brain injury, biomarker, children

Introduction

Traumatic brain injury (TBI) is a leading cause of death and disability worldwide. In the European Union, brain injury accounts for one million hospital admissions per year (International Brain Injury Association). In Czech Republic in 2006 there were 445 424 cases of injury of children between 0–14 years treated in the surgical ambulances. About 39.5 thousands cases of injury of persons aged 0–19 years were treated in hospital, in 277 cases injury caused by external cause resulted in death [1]. Traumatic head injury in infants, children and adolescents are caused by different mechanisms - from frequent falls, collisions motorized with non-motorized vehicles to child abuse or gunshot wounds. Traffic accidents are the leading cause of severe TBI [2].

Management is challenging – both minor and severe TBI create significant difficulties for clinicians. Several typical questions confront the clinician regarding extent of injury and prognosis. Because there is such marked heterogeneity in TBI, predicting outcome is difficult, as is deciding on optimal treatment. In children, uncertainties about predicting outcome are even greater than in adults and are multifactorial [3]. There are increased difficulties in assessing both initial injury severity and outcome. Objective tools to better  understand the degree of injury and to prognosticate are urgently needed. 

Biomarkers of TBI

A biomarker is an indicator of a specific biological or disease state that can be measured using samples taken from either the affected tissue or peripheral body fluids. These markers can be altered enzymatic activity, changes in protein expression, or post-translational modification, altered gene expression, protein, or lipid metabolites, or a combination of these changes [4].

The identification of pathology-specific biomarkers can assist in the diagnosis and estimation of prognosis, and can serve as surrogate markers for monitoring the effectiveness of a treatment.

In medicine today, several specialties employ biomarker blood tests to diagnose, direct treatment, and prognosticate. Commonly used biomarkers include troponin T/CK-MB in cardiology, procalcitonin and ESR (erythrocyte sedimentation rate) in sepsis, amylase and lipase in pancreatic disease, etc. Many of these markers of organ function and injury have long been used to guide management, yet similar biomarkers have not existed in routine care of TBI. In comparison with the coronary care unit, neurointensive care units have fewer markers or monitors of organ function and fewer therapies for intervention.

In 1983, the properties for an optimal brain biomarker were theorized by Bakay and Ward [5].

The authors suggested that such a biomarker should

1. show high specificity and sensitivity for brain
2.  have a rapid appearance in serum
3. be released only after irreversible destruction of brain tissue
4. be released in a time-locked sequence with the injury
5.  show low age- and sex-related variability
6.  have reliable assays for immediate analysis available

For patients with severe TBI, a biomarker may be helpful to predict which patients are likely to experience se­­con­dary insults, such as raised ICP and help prognosticate. For these reasons, the last decade and, in particular, the last few years have witnessed an increa­sing interest in biomarkers for TBI [6]. 

Neuron-Specific Enolase (NSE)

NSE was originally described by Moore and McGregor in 1965 [7]. NSE initially held promise as a brain injury biomarker, as it was originally believed to be strictly neuronal.

Enolases are glycolytic enzymes occurring as a series of dimeric isoenzymes made of three immunologically distinct subunits, the α, β, and γ chains. The isoforms γ γ and α γ are restricted to neurons, peripheral neuroendocrine tissue, and tumors of the amine precursor uptake and degradation system, and referred to as NSE [8].

Structural damage of neuronal cells causes leakage of NSE into the extracellular compartment and the bloodstream. Thus, it can be detected in the serum after neuronal death secondary to traumatic injury or a cerebrovascular accident [9].

Berger et al. examined NSE concentrations in the CSF of infants and children after inflicted and noninflicted TBI and showed that NSE was increased following TBI and that levels were higher than in adult TBI. They hypothesized that this may be related to an increased susceptibility of the developing brain to cellular death after TBI. The study also showed later peak levels of NSE in patients with inflicted versus noninflicted TBI, thought to be due to delayed neuronal death in abuse victims se­condary to an imbalance between pro- and antiapoptotic factors. This latent peak may allow for some discrimination between inflicted versus noninflicted TBI. The sensitivity and specificity of serum NSE after pediatric TBI as determined by ROC curves was found to be 71% and 64%, respectively [10].

However, like all biomarkers, NSE has limitations. Similar to S100B, however, there have been reports of false positive values in the setting of combined CNS injury plus shock [11]. A second limitation of NSE is the occurrence of false positive values in the setting of hemolysis [12]. Nevertheless, NSE appears to have value as a brain injury biomarker, with potential for use in diagnosis, prognosis, and therapeutic monitoring in neurointensive care.

The aims of this study were the correlations between neuron-specific enolase and mortality, Glasgow Outcome Scale (GOS) and neurological findings. 

Methods

This is a prospective observational study during the period from May 2007 to October 2009. The study protocol and informed consent approach were approved by the ethics committee of University Hospital Brno. The pa­rents provided written informed consent for their child­ren to participate in this trial.

We prospectively enrolled 63 pediatric patients aged  0–19 years. The criteria required for inclusion were TBI with or without multiple trauma 12 h before admission. TBI was verified by computerized tomography.

We recorded sex, age, type of accident (road traffic accident, bike accident, fall, assault), Glasgow Coma Scale (GCS), Pediatric Trauma Score (PTS).

The Glasgow Coma Scale was based on evaluation of the emergency measures taken, although without consideration of pharmacological interventions [13].

Trauma Score (PTS) has been selected as the trauma scoring tool for use in evaluating the severity of injury in the pediatric patient [14]. The PTS adjusts it’s scoring areas to account for the physiological and anatomical differences unique to the pediatric patient in turn more accurately identifying the critical patient.

Brain CT scans were rated by radiologists, blinded to study. CT was repeated after neurosurgical intervention and whenever the clinical course required.

After diagnostic assessment and/or surgery, all patients were transferred to the pediatric intensive care unit and received standard neurointensive care, including intubation and mechanical ventilation, haemodynamic monitoring, and intracranial pressure monitoring based on TBI therapeutic protocol. Intracranial hypertension was treated progressively by use of a standard step-wise protocol that included sedation, paralysis, mild hyperventilation (target paCO₂ 32–35 mmHg), osmotherapy with mannitol, and eventually use of barbiturates during refractory intracranial hypertension.

Cerebral perfusion pressure was maintained at 60 mmHg (50 mmHg in infants) by lowering intracranial pressure to 20 mmHg (15 mmHg in infants) and by maintaining mean arterial blood pressure at 80 mmHg (70 mmHg in infants).

ICP was monitored in the 15–20° position through an intraparenchymal pressure monitor (Codman Microsensor, Randolph, MA). ICP was noted hourly on the subject’s data sheet. The 24-hr mean ICP value was used for statistical analysis.

Lastly, six months after the primary injury, the study protocol documented outcome according to the Glasgow Outcome Score on a basis of 1–5 (1, dead; 2, persistent vegetative state; 3, severely disabled; 4, moderately disabled; 5, good recovery), and the cause of death [15]. Clinical outcome (any neurological deficit) was measured six months after the primary injury, based on neurolo­gical examination (include cranial nerve assessment 1–12, deep tendon reflexes including pathologic reflexes, sensory examination (including visual fields, audition and nociceptive/tactile responsiveness), cerebellar assessment and motor function testing, other areas assessed) and electro-encephalography. The person who decided upon outcome was completely blinded to the clinical information.

Venous blood for proteins measurement was sampled at admission and at the same time every morning, for a maximum of 6 days. Since the duration of time between trauma and admission varies measurement was not always possible at exactly the same number of hours after TBI. Therefore, we established time intervals for sampling of proteins and documented the time interval of mea­su­rement accordingly. The first time interval was 12 h after TBI (sample at admission, <12 h after TBI). Samples were centrifuged, and serum was stored at -72°C for analysis.

NSE was measured immuno-luminometrically with a commercially available kit (Elecsys analyzer, Roche Diagnostics). The person who carried out the assays was completely blinded to the clinical information. 

Statistical analysis

Descriptive characteristics of the input values of NSE, demographic parameters and characteristics of hospita­lized patients were summarized using basic descriptive statistics, statistical parameters (N, mean, standard deviation, median, minimum, maximum).

Dynamics (kinetics) protein values during the 6 days was modeled by analysis of variance (ANOVA). Correlation values in the patients was addressed in the model using the covariance matrix of symmetric (type “compound symmetrical”). The model was evaluated the impact factor (exitus, GOS, length of hospitalization), eva­­lua­ted both as a main effect (for an overall difference of values) and at baseline (D0) and on average during the 6-day period, and as the interaction factor with time, which is ev. difference in the dynamics. Values of statistical significance were presented in tabular form. Beha­vior of proteins (model form) was graphically illustrated by graphs showing the profiles of the values of individual patients (case profile plot) together with modeled curves for different categories of factors evaluated. ROC analysis was performed to determine the discriminatory cha­racteristics of the protein value. ROC curves are shown together with the respective diagnostic characteristics (sensitivity, specificity, positive and negative predictive value), the AUC and the optimal cutoff value for the best diagnostic ability of the protein.

All tests were mathematically evaluated by a  statistical software tool STATISTICA ver. 8.0, Copyright © SatSoft, Inc.

Results

Mortality and GOS were taken as principal end-point for all predictive analyses. The table 1 present basic descriptive characteristics of the file – the input parameters. Reported data are separately for the group of survivors patients and non-survivors.

NSE values is significantly higher in patients who died or have generally worse outcome (GOS). Significantly different dynamics of this protein was only in patients who died while the increase is still noticeable during the first two days. For survivors, the values of this protein decreases from the beginning. At D4 there is a few patients (see gray profiles) with mild elevation, which, however, does not show as a “second peak” effect in the overall average. In relation to the length of hospitalization does not show any diference (figure 1).

Difference in D0 is the difference that can be detected on the time of the first sample collection from the patient (shortly after the accident). The value of NSE is the first indicator. The average variation was eventual difference in the values of NSE within 6 days – that is, if we did not know how much time passed from the accident, or if a patient was transferred from another hospital (NSE va­lues in the D0 were not available). According to our results for the next days from D0 the NSE levels correlate with the mortality and GOS. This means that we can determine the levels of NSE at any time within 6 days, not only in D0.

Six  months after the primary injury neurological assessment was performed,  based on neurological examination. Statistical significance was demonstrated for NSE levels in admission and neurological follow-up at intervals of six months. Patients who had an input value above too high, within six months appeared neurological deficits (figure 2). The most common findings were hemiparesis, specific behavior, spastic quadriparesis, convergent strabismus.

Receiver-operating characteristic curve (ROC) analysis

The prognostic value of serum levels in predicting mortality was evaluated by ROC analysis. The figure 3 shows the receiver operating characteristic curves of the NSE to predict mortality 6 months post injury.

The area under the curve (AUC) is a measure of predictive discrimination: 50% is equivalent to random guessing and 100% is perfect prediction. Thus a higher AUC indicates a better ability to discriminate between poor and good outcome.

NSE cut-off values were determined, namely at day 0 and at day 2 as a predictors of mortality.

The figure 3 clearly shows higher AUC for mortality at day 2 (0.978), high specifity (100%) and sensitivity (87.5%). The results show the importance of dynamics, increase/decrease of protein in time.  

Discussion

During the last decade there was an increasing inte­rest in biochemical markers of brain damage. Glial and neuronal tissue derived proteins seemed to be easily detectable in peripheral blood and were supposed to serve as surrogate markers of brain damage after a variety of clinical conditions such as traumatic brain injury, stroke, heart surgery, cardiac arrest, or psychiatric, inflammatory or degenerative brain diseases. Due to the development of complete commercial ready-torun assays and the easy handling of serum probes an increasing number of studies on neuron specific enolase (NSE) and/or protein S100B release in acute or chronic central nervous system disorders were published. Most of those studies aimed at analyzing the association between serum concentrations and/or release patterns of mole­cular markers of brain damage and the clinical condition of the respective patients [16]. Biomarker studies in pediatric patients with brain injuries is very low.

The concentrations of NSE after severe TBI are several times higher  in children and more consistently increased versus the comparison group than reported in adults [17]. This may reflect increased susceptibility of the developing brain to cell death after traumatic injury. This is supported by Bittigau et al. who demonstrated increased apoptotic neuronal death after experimental TBI in immature rats [18]. The increase in our patients may also reflect greater injury severity in this patient population, children with TBI. It is unlikely to be attributable to an age-dependent difference in the concentration of NSE in neurons in children versus adults; previous studies of CSF NSE concentrations in patients without neurologic disease suggest that NSE concentrations increase, rather than decrease, with age [19]. The second peak in NSE concentration in patiens with TBI is remarkable, and may reflect delayed neuronal death. This finding is consistent with previous research in both experimental animal models of TBI and clinical studies showing increases in markers of delayed neuronal death in abuse victims [20].

Specifically, the increased delayed neuronal death may be related to a relative lack of anti-apoptotic neuroprotectants such as Bcl-2 (B-cell lymphoma 2), in combination with a relative excess of apoptosis triggers such as cytochrome-C [21, 22]. This lack of balance between proapoptotic and anti-apoptotic factors would favor delayed neuronal death. In our work, we found no second peak increase of NSE. It may be also due to short-term monitoring of blood levels.

In addition, acute measurement of NSE serum concentrations may provide a quantitative predictor of the outcome after TBI and might help identify those patients who are at risk of long-term neuropsychological dysfunction [23]. In our work was also shown that NSE le­vels were significantly higher in patients who died or had a poor outcome after 6 months post-injury than those who survived or had a better outcome. When NSE level is higher than 62.5 μg/l in D2, it strongly predicts death (AUC 0.978; SE 87.5%; SP 100%).

However, in some studies in the literature, it has been reported that there is no significant difference between NSE levels of patients with favorable and unfavorable outcomes [24]. This controversy in the literature mainly arises from using different schedules to take blood samples for analyzing the enzyme by diffe­rent investigators. At first glance, timing in obtaining blood sample after TBI seems to be one of the key points. Duration and intervals for obtaining blood samples after trauma could be helpful for clinical management and schedule due to practical objectives. The study groups that revealed no association between NSE levels and outcomes used relatively long time course. Nevertheless, biomarkers helping to estimate the prognosis of patients of TBI should give information about acute brain damage and facilitate the design of early management. In addition to these comments, there is short biological half-life of NSE (approximately 48 h) [25]. Delayed measurements may be associated with secondary brain injury, even though a correlation is detected in late phases of trauma. As NSE is not actively secreted into the bloodstream, it is passively released by cell destruction only; its elevation in the late phases of injury may be explained by delayed onset cell deaths as seen in secondary events. 

Conclusion

This study demonstrates that NSE concentrations are markedly increased in the serum of children after TBI. This protein may have the potential to be used as quantitative measures of the success of therapy for TBI. Si­milarly, NSE quantified early after injury might serve as an objective marker of the predictor mortality.

Future studies will be required to not only identify this biomarker, but to develop a clinically useful test in which a panel of TBI-associated biomarkers (a biomarker signature) are evaluated and used to direct treatment.    

Došlo: 14. 1. 2011

Přijato: 1. 2. 2011

MUDr. Jiří Žurek, Ph.D.

Klinika dětské anesteziologie a resuscitaceLF MU

Fakultní nemocnice Brno

Černopolní 9

625 00 Brno

e-mail: zurekj@post.cz