Skip to main content
Journal of Neurotrauma logoLink to Journal of Neurotrauma
. 2023 Dec 29;41(1-2):106–122. doi: 10.1089/neu.2023.0039

Prognostic and Diagnostic Utility of Serum Biomarkers in Pediatric Traumatic Brain Injury

Jennifer C Munoz Pareja 1,*, Juan Pablo de Rivero Vaccari 2, Maria Mateo Chavez 3, Maria Kerrigan 4, Charlene Pringle 5, Kourtney Guthrie 5, Kathryn Swaby 1, Jennifer Coto 6, Firas Kobeissy 7,8, K Leslie Avery 5, Suman Ghosh 9, Rajderkar Dhanashree 10, Prashanth Shanmugham 11, Lauren A Lautenslager 12, Shannon Faulkenberry 13, Maria C Pareja Zabala 14, Nora Al Fakhri 15, Ricardo Loor-Torres 3, Lance S Governale 16, Jason E Blatt 16, Joslyn Gober 17, Paula Karina Perez 18, Juan Solano 1, Heather McCrea 19, Chad Thorson 20, Kristine H O'Phelan 21, Robert W Keane 2,22, W Dalton Dietrich 2, Kevin K Wang 7,8
PMCID: PMC11071081  PMID: 37646421

Abstract

Traumatic brain injury (TBI) remains a major cause of morbidity and death among the pediatric population. Timely diagnosis, however, remains a complex task because of the lack of standardized methods that permit its accurate identification. The aim of this study was to determine whether serum levels of brain injury biomarkers can be used as a diagnostic and prognostic tool in this pathology. This prospective, observational study collected and analyzed the serum concentration of neuronal injury biomarkers at enrollment, 24h and 48h post-injury, in 34 children ages 0–18 with pTBI and 19 healthy controls (HC). Biomarkers included glial fibrillary acidic protein (GFAP), neurofilament protein L (NfL), ubiquitin-C-terminal hydrolase (UCH-L1), S-100B, tau and tau phosphorylated at threonine 181 (p-tau181). Subjects were stratified by admission Glasgow Coma Scale score into two categories: a combined mild/moderate (GCS 9–15) and severe (GCS 3–8). Glasgow Outcome Scale-Extended (GOS-E) Peds was dichotomized into favorable (≤4) and unfavorable (≥5) and outcomes. Data were analyzed utilizing Prism 9 and R statistical software. The findings were as follows: 15 patients were stratified as severe TBI and 19 as mild/moderate per GCS. All biomarkers measured at enrollment were elevated compared with HC. Serum levels for all biomarkers were significantly higher in the severe TBI group compared with HC at 0, 24, and 48h. The GFAP, tau S100B, and p-tau181 had the ability to differentiate TBI severity in the mild/moderate group when measured at 0h post-injury. Tau serum levels were increased in the mild/moderate group at 24h. In addition, NfL and p-tau181 showed increased serum levels at 48h in the aforementioned GCS category. Individual biomarker performance on predicting unfavorable outcomes was measured at 0, 24, and 48h across different GOS-E Peds time points, which was significant for p-tau181 at 0h at all time points, UCH-L1 at 0h at 6–9 months and 12 months, GFAP at 48h at 12 months, NfL at 0h at 12 months, tau at 0h at 12 months and S100B at 0h at 12 months. We concluded that TBI leads to increased serum neuronal injury biomarkers during the first 0–48h post-injury. A biomarker panel measuring these proteins could aid in the early diagnosis of mild to moderate pTBI and may predict neurological outcomes across the injury spectrum.

Keywords: biomarkers, outcomes, pediatrics, prognosis, traumatic brain injury


graphic file with name neu.2023.0039_figure8.jpg

Introduction

Injury is the leading cause of death among children in the United States and other developed nations.1 Among the American population, one in 14 children is afflicted by head injuries. As a result, each year approximately 475,000 patients aged 0–14 years visit emergency departments secondary to traumatic brain injury (TBI).2 A TBI is also a significant contributor to death and disability in children.3

The direct and indirect costs associated with TBI incidence in the United States were estimated to be $76.5 billion per year in 2010.4 Further, long-term disability resulting from TBI is estimated to impact approximately 3.2 to 5.3 million individuals, equivalent to approximately 1–2% of the U.S. population.5 Such disabilities have significant economic and emotional impacts on families and society.6

After a TBI, both neuronal and glial cells in the central nervous system (CNS) are compromised, releasing various structural components that act as biomarkers. These biomarkers include microtubule-associated protein 2 from dendrites, neuron-specific enolase (NSE), and ubiquitin C-terminal hydrolase L1 (UCH-L1) from the cell body, and light-chain neurofilament (NfL), medium-chain neurofilament (Nf-M), and tau protein from axons. Damage to oligodendrocytes results in the release of myelin basic protein (MBP), whereas astrocyte injury produces glial fibrillary acidic protein (GFAP) and S100B.

In addition, traumatic forces can cause modifications in the vasculature, the release of molecular mediators, and the activation of the immune system to provide protection against perceived harmful signals, potential pathogens, and damaged nervous system cells.7

Accurate classification of TBI ensures appropriate treatment and effective follow-up. The Glasgow Coma Scale (GCS), a 15-point neurologic injury severity scale, is widely used in adults and children because of its high interobserver reliability and good prognostic capability.8,9 Special populations, however, such as children with pre-existing neurological conditions or those who are mechanically ventilated and sedated, may encounter utilization challenges, necessitating additional tools to increase the usefulness and accuracy of GCS in diagnosing brain injury.

While imaging studies can provide a clear visual assessment of the severity of TBI, the use of computed tomography (CT) in childhood poses a risk of elevated cancer because of radiation exposure.10,11 In addition, magnetic resonance imaging (MRI), which can detect more subtle injuries, is often costly and not readily available, especially in resource-limited settings where access to brain imaging diagnostic modalities may not be feasible.

Prognostic assessment in pTBI patients is an essential parameter that necessitates careful consideration not only for clinicians to decide on appropriate therapy but also for patients and their families to understand their life expectancy and the potential need for further specialty care.

The Glasgow Coma Scale (GCS) is commonly used to evaluate prognosis in pTBI patients, because it is the most reliable and feasible outcome predictor. Nonetheless, a few shortcomings should be kept in mind regarding this scoring system. Although GCS is generally considered reliable in adults, potential issues arise in pediatric and neonate patients. The reason is that the scoring system is based on consciousness and the ability to understand and perform orders, which can be challenging to assess in these populations.12

In addition, there is concern that the threshold for severe classification of GCS may need to be lowered, particularly in patients with suspected hypoxic-ischemic injury, because such injuries can act as confounders of poor outcomes, potentially leading to an overestimation of neurophysiological dysfunction.13 Further, conventional prognostic assessment measurements, such as cerebral metabolism and hemodynamic monitoring, may lack the required sensitivity or specificity to accurately predict outcomes in ventilated and sedated patients.14,15

Finally, predicting recovery from TBI in children is more complicated than in adults because of various factors such as brain development and changes in anatomical structure, which make clinical presentations and complexities more intricate.16

In addition to the previously described challenges, the management of concussion injuries is a crucial topic that needs to be considered, especially when there is an absence of objective measures to ensure the safe resumption of activities. Existing guidelines for return-to-play and return-to-learn recommendations do not include interventions that can be objectively measured. Instead, they rely on assessments based on patient-reported symptoms, which are subjective and can lead to a higher risk of secondary injuries caused by an early resumption of daily activities, leading to increased symptoms and delayed healing.17,18

To mitigate these risks and promote optimal recovery, objective measures need to be developed to assist clinicians in making informed decisions about the safe resumption of activities. These measures help reduce subjectivity in the decision-making process and provide more accurate and objective assessments of an individual's readiness to return to normal activities after a concussion injury.

There is a pressing need for alternative diagnostic tools that are safe, cost-effective, and easily accessible, especially in resource-limited settings. Recent evidence indicates that blood-based TBI biomarker tests show promise in evaluating the severity of TBI and determining patient prognosis, particularly in cases where other neurological measures may not be informative.19-21

The Scandinavian Neurotrauma Committee published evidence-based guidelines in 2013 for the initial management of TBI in adults, incorporating S100B into diagnosis of TBI associated with CT abnormalities.22 The U.S. Food and Drug Administration (FDA) has recently approved using UCH-L1 and GFAP as diagnostic TBI biomarkers, and a reliable point-of-care blood test for adults is being developed.23 Despite recent advancements in the field, however, there is a lack of literature evaluating the simultaneous use of multiple biomarkers for pediatric TBI, and further research is needed to validate the effectiveness of these biomarkers compared with those used in adults.19,24,25

The primary purpose of this study was to evaluate the diagnostic and prognostic properties of a TBI biomarker panel in the pediatric population. To our knowledge, this is the first report on this specific combination of biomarkers in a cohort of critically ill children with TBI.

Methods

Study population

The study enrolled eligible patients based on specific criteria, including age between 0–18 years, admission to the University of Florida (U.F.) or transfer from an outside hospital to the Pediatric Intensive Care Unit (PICU) between February 2017 and 2020, and a clinical diagnosis of TBI (with a GCS score of 3–15) that necessitated brain imaging studies within 24h of injury. Patients with evidence of severe psychiatric disorders as per the Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition (e.g., schizophrenia spectrum and other psychotic disorders, bipolar and related disorders, depressive disorders) were excluded.26

The study collected common data elements for pediatric TBI, demographic information, and detailed clinical assessments, including seven days of relevant critical care data points.27 Long-term outcomes were measured using the Glasgow Outcome Scale-Extended (GOS-E) Peds at three intervals: 2–6 weeks, 6–9 months, and 12 months post-injury. The U.F. Institutional Review Board 01 (IRB201600237) approved the study with a modification of informed consent, allowing participants or their authorized representatives to provide written informed consent within 72h of admission.

Healthy controls (HC) were recruited from a cohort of apparently healthy pediatric patients who presented at the Emergency Department for minor complaints unrelated to TBI. The HCs with a history of TBI or an acute neurological process were ineligible for enrollment. Blood samples were collected from all participants at enrollment, 24h and 48h post-injury, for genetic and proteomic analysis.

Serum biomarkers

The TBI-Common Data Elements (TBI-CDE) Biospecimens and Biomarkers Working Group Consensus guidelines for plasma preparation were followed.28 Blood sampling took place by venipuncture if required for routine clinical care; otherwise, blood was collected via a catheter that was already in place as part of clinical care. A total of 2 cc/kg, up to max 5 mL (age 0–4 years); 10 mL (age 5–18 years).

Blood samples were drawn into red top SST BD Vacutainer® Plus tubes. Samples sat upright for 30 min at room temperature and were centrifuged at 4000 rpm for 10 min. The cleared serum (supernatant) was pipetted and stored in aliquots of 0.5 mL. Samples were stored within 2h of the blood draw in freezers at or below −80˚C in a biorepository at the UF McKnight Brain Institute. All samples were stored in a de-identified manner with a unique study number specific to the participant.

Biomarker analysis

Biomarker analysis was performed at the Program for Neurotrauma, Neuroproteomics & Biomarker Research at U.F. Serum samples were thawed in room temperature batches and centrifuged at 10,000 rpm for 10 min at 4˚C before testing. Levels of GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 concentrations were measured using the same batch of reagents by investigators blinded to clinical data using SIMOA neuro 4 plex (N4PB) kit (item 103345), and p-tau (phospho-Theronine181 epitope) Discovery kit (item 103377) in SR-X immunoassay analyzer (Quanterix Corp, Boston, MA).

S100B concentrations were measured using S100β enzyme-linked immunoassay (ELISA, ca t#EZHs100b-33K, EMD Millipore). A signal was detected with a chromogenic substrate (3-3-5-5-tetramethylbenzidine), and the intensity was measured using a colorimetric plate reader (450 nm). The assay methodology is described in detail previously.29

The assay performance was as follows: S100B assay - range of assay 0-2,000 pg/mL, lowest limited of quantification (LLOQ) 2.7 pg/mL, the lowest limit of detection (LLOD), 1.4 pg/mL, intraassay CV 4.8 %, interassay CV 4.4 %; UCH-L1 assay - range 0–40,000 pg/mL, LLOQ 30 pg/mL, LLOD 60 pg/mL, intraassay CV 11.7%, interassay CV 9.3%; GFAP assay – range 0-40,000 pg/mL, LLOQ 9.38 pg/mL, LLOD 1.32 pg/mL, intraassay CV 10.5%, interassay CV 9.5%; NfL assay – range 0-2,000 pg/mL, LLOQ 0.625 pg/mL, LLOD 0.0971 pg/mL, intraassay CV 7.9%, interassay CV 6.9 %; tau assay – range 0-400 pg/mL, LLOQ 0.100 pg/mL, LLOD 0.0236 pg/mL, intraassay CV 11.7%, interassay CV 9.3%; and p-tau (T181) assay -–range 0-1,012 pg/mL, LLOQ 5.16 pg/mL, LLOD 0.233 pg/mL, intraassay CV 8.7%, interassay CV 5.9 %. For samples exceeding the assay range, they were diluted three-fold to four-fold and reassayed.

Outcome measures

The GOS-E Peds was used as the primary tool to quantify long-term outcomes after TBI. The GOS-E Peds is a validated clinical scale used to quantify rehabilitative success and identify permanent disability in a patient recovering from TBI. In addition, measurements are designed to screen for the onset of new cognitive or behavioral changes, including sleep disturbances, social disruptions, and learning impairments.30 The GOS-E Peds was conducted at subsequent pediatric neurology follow-up visits or by phone if the patient was not scheduled for a visit or was lost to follow-up at the TBI clinic. GOS-E Peds score ranges from 1 (indicating death) to 8 (indicating complete recovery).

To simplify the analysis, all measurements were recoded into two binary categories: favorable outcome (GOS-E Peds ≤4 = 1) and unfavorable outcome (GOS-E Peds ≥5 = 0). All surveys were administered by a trained research associate and completed directly by the patient's caregiver.

Clinical variables and scores

Clinically relevant variables commonly assessed as prognostic factors in the intensive care unit (ICU) were measured and analyzed in correlation with biomarker serum levels at enrollment, 24h and 48h. These included: intracranial pressure (ICP) monitoring, cerebrospinal fluid (CSF) diversion, utilization of hypertonic solution, seizures, hypoxic events, hyperventilation, cardiac arrest, invasive mechanical ventilation, hypothermia, hospital length of stay, ICU length of stay, as well as the results of Injury Severity Score (ISS), and Pediatric Risk of Mortality (PRISM) score and Marshall admission scores.

Statistical analysis

Frequencies and percentages were calculated to analyze categorical variables pertaining to demographic information. Chi-square and/or Fisher exact test were employed to evaluate associations between categorical variables, such as the characteristics of the patients with TBI and HC. Biomarker values were logarithmically transformed to achieve a normal distribution.

Descriptive statistics, including mean, standard deviation, and interquartile range (IQR), were computed for the continuous variables. Mann–Whitney U and Kruskal–Wallis tests were used to compare differences between groups for continuous variables. Receiver operating characteristic (ROC) analysis was performed to determine the accuracy of biomarker levels in differentiating favorable versus unfavorable outcomes.

Correlations between binary critical care variables and biomarker levels were analyzed by simple logistic regression. Correlations between non-binary critical care variables or clinical scores and biomarker levels were analyzed by calculating the Spearman correlation coefficient and subsequent analysis of the corresponding correlation matrix. All statistical tests were two-sided with a significance level of 0.05.

Consistent with the current statistical consensus, we considered an area under the curve (AUC) of less than 0.7 as poor, 0.7–0.8 as fair, 0.8–0.9 as good, and greater than 0.9 as excellent. We analyzed the data using Prism 9 software and R (2022).

Results

Thirty-four children with TBI and 19 HC were included in this analysis. Of those, 15 patients had severe TBI (GCS 3–8), six moderate TBI (GCS 9–12), and 13 mild TBI (GCS 13–15), the mean age for HC was 10.5 years and 5.6 years for the patients with TBI across all categories. The control group comprised nine males and 10 females, whereas the TBI group consisted of 21 males and 13 females.

Causes of hospitalization for apparently HCs were mainly inflammatory or infectious diseases (n = 11), followed by orthopedic injuries or minor trauma (excluding TBI and spinal cord injury, n = 5) and other causes (n = 3). Additional characteristics of this cohort of patients and HC are described in Table 1.

Table 1.

Demographics and Injury Characteristics of Pediatric Traumatic Brain Injury Subjects as per Glasgow Comma Scale Classification at Enrollment and Healthy Controls

 
Control
Eligible patients by GCS Category
 
Mild (GCS = 13–15)
Moderate (GCS = 9–12)
Mild/Moderate (GCS = 9-15)
Severe (GCS = 3–8)
Characteristics (n = 19) (n = 13) (n = 6) (n = 19) (n = 15) p
Age in years, mean (SD)a 10.5 (5.76) 5.1 (5.3) 7.7 (7.9) 5.9 (5.8) 5.6 (5.2)  
Sex, n (%)a            
Female 10 (52.63%) 5 (38.46 %) 2 (33.33 %) 7 (36.84 %) 6 (40 %)  
Male 9 (47.37%) 8 (61.54 %) 4 (66.67%) 12 (63.16 %) 9 (60 %)  
BMI, mean (SD)a 21.32 (6.79) 25.83 (26.72) 18.27 (4.53) 24.62 (22.82) 18.50 (5.56)  
Race, n (%)a            
White 9 (47.23%) 7 (53.84%) 3 (50%) 10 (52.63 %) 8 (53.33 %)  
African American 1 (5.26%) 3 (23.07 %) 2 (33.33 %) 5 (26.31 %) 4 (26.66 %)  
Asian 0 1 (7.69 %) 0 1 (5.26 %) 1 (6.66 %)  
Unknown 4 (21.05%) 0 1 (16.67 %) 1 (5.26 %) 1 (6.66 %)  
Unreported 5 (26.31%) 2 (15.38%) 0 2 (10.53 %) 1 (6.66 %)  
Payer status, n (%)a            
Medicaid 6 (31.58%) 3 (23.07 %) 4 (66.67%) 7 (36.84 %) 10 (66.67 %)  
Other 13 (68.42%) 10 (76.92 %) 2 (33.33 %) 12 (63.16 %) 5 (33.33 %)  
Mechanism of TBI, n (%)b,c            
Acceleration/Deceleration NA 3 (23.07 %) 3 (50 %) 6 (31.57 %) 10 (66.67 %)  
Ground level fall 1 (7.69 %) 1 (16.67 %) 2 (10.53 %) 0  
Blow to head 4 (30.76 %) 2 (33.33 %) 6 (31.57 %) 7 (46.67 %)  
Fall from height >1 meter (3 ft) 4 (30.76 %) 2 (33.33 %) 6 (31.57 %) 2 (13.33 %)  
Direct impact: Head against object 3 (23.07 %) 2 (33.33 %) 5 (26.31 %) 2 (13.33 %)  
ISS Score, n (%)b            
Minor to Serious (1–24) NA 6 (46.15 %) 3 (50 %) 9 (47.37 %) 1 (6.66 %) 0.0348
Severe to Maximum (25–75) 7 (53.84 %) 3 (50 %) 10 (52.63 %) 14 (93.33 %)
PRISM Score, n (%)b,c            
5–9 NA 7 (53.84 %) 1 (16.66 %) 8 (42.11 %) 0  
10–14 4 (30.76 %) 3 (50 %) 7 (36.84 %) 3 (20 %)  
15–19 1 (7.69 %) 0 1 (5.26 %) 5 (26.66 %)  
20–24 0 0 0 1 (6.66 %)  
25–29 0 0 0 2 (13.33 %)  
30–34 0 0 0 0  
≥35 0 0 0 1 (6.66 %)  
GOS-E Peds Average Outcome, n (%)b            
Favorable (≤4) NA 12 (92.30 %) 5 (83.33 %) 17 (89.47 %) 7 (46.67 %) 0.0228
Unfavorable (≥5) 1 (7.69 %) 1 (16.67 %) 2 (10.53 %) 8 (53.33 %)
Abnormal Findings at Enrollment CTs, n (%)b            
CT Positive NA 11 (84.62 %) 5 (83.33 %) 16 (84.21 %) 15 (100 %)  
SF 2 (18.18 %) 0 2 (12.50 %) 1 (6.66 %)  
SAH 2 (18.18 %) 0 2 (12.50 %) 2 (13.33 %)  
SDH 2 (18.18 %) 0 2 (12.50 %) 1 (6.66 %)  
EH 0 0 0 1 (6.66 %)  
IPH 0 0 0 1 (6.66 %)  
DAI 0 1 (20 %) 1 (6.25 %) 0  
SAH + SDH 0 2 (40 %) 2 (12.50 %) 2 (13.33 %)  
SAH + SDH + SF 2 (18.18 %) 0 2 (12.50 %) 0  
SAH + SDH + IVH 0 1 (20 %) 1 (6.25 %) 1 (6.66 %)  
IVH + IPH 0 0 0 0  
AI + TH 0 0 0 1 (6.66 %)  
Not specified 3 (27.27 %) 1 (20 %) 4 (25 %) 5 (33.33 %)  
CT Negative 2 (15.38 %) 1 (16.67 %) 3 (15.78 %) 0  

AI, anoxic injury; BMI, body mass index; CT, computed tomography; DAI, diffuse axonal injury; EH, epidural hematoma; GCS, Glasgow Coma Scale; GOS- E Peds, Glasgow Outcome Scale Extended Pediatrics; ISS, Injury Severity Score; IVH, intraventricular hemorrhage; IPH, intraparenchymal hemorrhage; PRISM, Pediatric Risk of Mortality Score; SF, Skull Fracture; SD, standard deviation; SAH, subarachnoid hemorrhage; SDH, subdural hemorrhage; TH, tonsillar herniation; TBI, traumatic brain injury.

a

p values demonstrate comparison of frequency distributions by control and Glasgow Comma Scale classification score on enrollment.

b

p values for comparison of frequency distributions by Glasgow Comma Scale classification score on enrollment.

c

Analysis performed of cohort of patients with available data.

Patients classified as severe according to their admission GCS score were associated with worse outcomes, as measured by the GOS-E Peds score average and were more likely to have a higher ISS compared with patients in other GCS classifications. There was no significant difference, however, in demographic characteristics, such as sex or race, and clinical features, such as body mass index (BMI) at admission, between patients and healthy controls.

GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 serum profiles in pTBI patients with favorable versus unfavorable outcomes

Serum biomarker levels were measured at three post-injury time points (0h-enrollment, 24h, and 48h) and further associated with GOS-E Peds scores, which were assessed at 2–6 weeks, 6–9 months, and 12 months post-injury. Data were dichotomized into two GOS-E Peds categories: favorable (GOS-E Peds ≤4) and unfavorable (GOS-E Peds ≥5) (Table 2).

Table 2.

Receiver Operating Characteristic With Area Under the Curve (95% Confidence Interval) of Binary Glasgow Outcome Scale Extended Pediatrics Categorization Predicting Unfavorable Outcomes in Each Biomarker at 0h, 24h, and 48h Post-Injury

  GOS-E Peds intervals Post-injury time points AUC SE 95% CI p  
GFAP 12 mo 48h 1 0 1.000 - 1.000 0.0286 Cutoff >207.4 (pg/mL)
Sensitivity 100%
Specificity 75%
NfL 12 mo 0h 0.84 0.1164 0.6164 - 1.000 0.0390 Cutoff >4.364 (pg/mL)
Sensitivity 100%
Specificity 44.44%
tau 12 mo 0h 1 0 1.000 - 1.000 0.0043 Cutoff >1.812 (pg/mL)
Sensitivity 100%
Specificity 83.33%
UCH-L1 6–9 mo 0h 0.81 0.1191 0.5832 - 1.000 0.0420 Cutoff >30.41 (pg/mL)
Sensitivity 100%
Specificity 30%
12 mo 0h 0.98 0.03443 0.9103 - 1.000 0.0020 Cutoff >98.55 (pg/mL)
Sensitivity 100%
Specificity 88.89%
S100B 12 mo 0h 0.96 0.06192 0.8370 - 1.000 0.0190 Cutoff >356.3 (pg/mL)
Sensitivity 100%
Specificity 83.33%
p-tau181 2–6 wk 0h 0.90 0.0958 0.7170 - 1.000 0.0140 Cutoff >0.2600 (pg/mL)
Sensitivity 100,00%
Specificity 33.33%
6–9 mo 0h 0.8 0.13 0.5453 - 1.000 0.0216 Cutoff >0.345 (pg/mL)
Sensitivity 100%
Specificity 40%

GOS-E Peds, Glasgow Outcome Scale Extended Pediatrics; standard error (SE); confidence interval (CI).

a

Only biomarkers with statistically significant p values are shown in this table.

Serum levels of GFAP measured at 48h were increased in patients with unfavorable outcomes at 12 months post-injury (p value 0.029; median 3120 pg/mL; IQR 988.4-9837 pg/mL) compared with TBI subjects of the same cohort with favorable outcomes (median, 99.03 pg/mL; IQR 7.87-220.3 pg/mL). For NfL, serum levels were higher in TBI patients with unfavorable outcomes than those of the favorable group when measured at 0h post-injury (p value 0.0390; median 61.59 pg/mL; IQR, 21.26-85.02 pg/mL).

Tau was significantly elevated when measured at 0h post-injury in TBI patients with unfavorable outcomes at 12 months (p value 0.004) compared with the favorable group. At 0h post-injury, UCH-L1 was increased in TBI patients with unfavorable outcomes at 6–9 months (p value 0.042; median 196.3 pg/mL; IQR 82.20-1264 pg/mL), and at 12 months (p value 0.002; median 989.3 pg/mL; IQR 407.3-1422 pg/mL).

The S100B elevations were also statistically relevant when measured at enrollment post-injury in TBI patients with unfavorable outcomes at 12 months (p value 0.019; median 1714 pg/mL; IQR 773.3-2615 pg/mL), compared with patients who had favorable outcomes (Fig. 1).

FIG. 1.

FIG. 1.

Serum levels of brain biomarkers as predictors of traumatic brain injury (TBI) outcome. Box and whisker plots showing statistically significant biomarker levels in pg/mL after TBI associated with unfavorable outcomes: p-tau181 0h at 2–6 weeks, 6–9 months, and 12 months; UCH-L1 0h at 6–9 months and 12 months; GFAP 48h at 12 months; and NfL, tau, and S100B 0h at 12 months. Box and whiskers are shown for the 5th and 95th percentile.

Interestingly, our preliminary data suggest that p-tau181 measured at 0h could potentially predict unfavorable outcomes at all three time points: 2–6 weeks, 6–9 months, and 12 months post-injury (p value 0.014, 0.021, 0.004, accordingly) with an AUC of 0.90 at 2–6 weeks, 0.80 at 6–9 months, and 0.95 at 12 months (Fig. 2). These findings indicate that GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 are useful prognostic biomarkers in pTBI patients.

FIG. 2.

FIG. 2.

Receiver operating characteristic (ROC) of biomarkers associated with unfavorable outcome. ROC curves indicate the area under the curve (sensitivity vs. 1-specificity) for statistically significant biomarker levels after traumatic brain injury associated with unfavorable outcomes: p-tau181 0h at 2–6 weeks, 6–9 months, and 12 months; UCH-L1 0h at 6–9 months and 12 months; GFAP 48h at 12 months; and NfL, tau, and S100B 0h at 12 months.

GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 serum profiles at different post-injury time intervals across injury severity in pTBI patients and HC

Each biomarker concentration was measured at three post-injury time intervals (0h-enrollment, 24h, and 48h) for patients with TBI and HC (Table 3) . All six biomarkers (GFAP, NfL, UCH-L1, S100B, tau, and p-tau181) measured at enrollment across injury severity were elevated when compared with HC (p < 0.05) (Fig. 3), with an AUC of 0.82 for GFAP, 0.74 for NfL, 0.78 for UCH-L1, 0.82 for S100B, 0.83 for tau, and 0.91 for p-tau181. In addition, GFAP, NfL, UCH-L1, and tau levels were elevated compared with HC at 24h, with an AUC of 0.83 for GFAP, 0.99 for NfL, 0.83 for UCH-L1, and 0.87 for tau.

Table 3.

Receiver Operating Characteristic-Derived Cutoffs for 100 Percent Sensitivity and Specificity Across Injury Severity

  Post-injury time intervals AUC SE 95% CI p  
GFAP 0h 0.83 0.06413 0.7018 - 0.9532 0.0012 Cutoff >29.21(pg/mL)
Sensitivity 96,00%
Specificity 25,00%
24h 0.84 0.06856 0.7049 - 0.9737 0.0005 Cutoff >41.44 (pg/mL)
Sensitivity 95.24%
Specificity 25,00%
48h 0.72 0.1071 0.5089 - 0.9286 0.3821 Cutoff >15.35 (pg/mL)
Sensitivity 91.67%
Specificity 18.75%
NfL 0h 0.75 0.07576 0.5997 - 0.8966 0.0475 Cutoff >1.176 (pg/mL)
Sensitivity 96.3%
Specificity 20,00%
24h 0.99 0.006904 0.9823 - 1.000 <0.0001 Cutoff >7.541 (pg/mL)
Sensitivity 100%
Specificity 93.33%
48h 0.90 0.06231 0.7742 - 1.000 <0.0001 Cutoff >1.719 (pg/mL)
Sensitivity 100%
Specificity 20,00%
Tau 0h 0.84 0.06559 0.7067 - 0.9638 0.0008 Cutoff >0.3242 (pg/mL)
Sensitivity 95.45%
Specificity 25,00%
24h 0.88 0.05666 0.7669 - 0.9890 0.0002 Cutoff >0.4417 (pg/mL)
Sensitivity 95.24%
Specificity 31.25%
48h 0.73 0.1052 0.5281 - 0.9406 0.2752 Cutoff >0.3171(pg/mL
Sensitivity 91.67%
Specificity 25,00%
UCH-L1 0h 0.78 0.08361 0.6187 - 0.9465 0.0433 Cutoff >9.011 (pg/mL)
Sensitivity 100%
Specificity 30.77%
24h 0.83 0.06966 0.6942 - 0.9673 0.0007 Cutoff >6.147 (pg/mL)
Sensitivity 100%
Specificity 23.08%
48h 0.78 0.09001 0.6046 - 0.9575 0.0584 Cutoff >8.903 (pg/mL)
Sensitivity 100%
Specificity 30.77%
S100b 0h 0.88 0.08025 0.7260 - 1.000 0.0239 Cutoff >16.99 (pg/mL)
Sensitivity 93.33%
Specificity 25,00%
24h 0.98 0.02958 0.9228 - 1.000 0.0530 Cutoff >29.30 (pg/mL)
Sensitivity 100%
Specificity 75%
48h 0.97 0.04218 0.8896 - 1.000 0.0118 Cutoff >29.35 (pg/mL)
Sensitivity 100%
Specificity 75,00%
P-Tau181 0h 0.91 0.04466 0.8237 - 0.9988 <0.0001 Cutoff >0.1800 (pg/mL)
Sensitivity 100%
Specificity 30.77%
24h 0.81 0.07758 0.6578 - 0.9619 0.0537 Cutoff >0.1350 (pg/mL)
Sensitivity 100%
Specificity 23.08%
48h 0.92 0.05161 0.8158 - 1.000 0.0008 Cutoff >0.1100 (pg/mL)
Sensitivity 100%
Specificity 15.38%

Area Under The Curve (AUC); standard error (SE); confidence interval (CI).

A

Injury severity assessed by Glasgow Coma Scale Scores.

a

In each biomarker at enrollment, 24h and 48h post-injury.

FIG. 3.

FIG. 3.

Serum levels of brain biomarkers versus healthy controls (HCs) across injury severity. Box and whisker plots show mean quartiles and outliers for each serum biomarker of interest at 0h, 24h, and 48h post-injury, with respective p values listed above. Dots correspond to data points outside the 5th and 95th percent confidence interval. All six biomarkers measured at 0h across injury severity were elevated when compared with HC. GFAP, NfL, UCH-L1, and tau levels were elevated compared with HC at 24h. Finally, NfL, S100B, and p-tau demonstrated an increased serum value compared with HC at 48h.

Finally, NfL, S100B, and p-tau181 demonstrated an increased serum value compared with HC at 48h, with an AUC of 0.89 for NfL, 0.97 for S100B, and 0.91 for p-tau181 (Fig. 4). These findings suggest that each biomarker could distinguish injury versus non-injury in patients with pTBI at enrollment.

FIG. 4.

FIG. 4.

Receiver operating characteristic (ROC) of statistically significant brain biomarkers across injury severity compared with healthy controls. Combined graphic with multiple ROC curves that indicate the area under the curve (sensitivity vs. 1-specificity) for statistically significant biomarker levels compared with healthy controls across injury severity.

GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 serum profiles in pTBI patients with different severity categorized by initial GCS

Concentrations of each biomarker level at three post-injury time intervals (0h or enrollment, 24h, and 48h) were measured and associated with TBI severity categorized by initial GCS. Subjects were subsequently stratified into two categories: mild/moderate (GCS 9–15) and severe (GCS 3–8). The mild and moderate categories were combined because of the small sample size in the moderate category across time intervals (Table 4).

Table 4.

Receiver Operating Characteristic-Derived Cutoffs for 100 Percent Sensitivity and Specificity of Individual Biomarkers Categorized by Initial Glasgow Coma Scale Severity A at Enrollment, 24h and 48h Post-Injury

  GCS severity Post-injury time intervals AUC SE 95% CI p  
GFAP Mild/Moderate 0h 0.83 0.06772 0.6976 - 0.9631 0.008 Cutoff >18.26 (pg/mL)
Sensitivity 100%
Specificity 18.75%
Severe 0h 0.92 0.0727 0.7825 - 1.000 0.0006 Cutoff >29.21(pg/mL)
Sensitivity 100%
Specificity 25%
24h 1 0 1.000 - 1.000 0.0002 Cutoff >425.1 (pg/mL)
Sensitivity 100%
Specificity 93.75%
48h 1 0 1.000 - 1.000 0.0002 Cutoff >425.1(pg/mL)
Sensitivity 100%
Specificity 93.75%
NfL Mild/Moderate 48h 0.81 0.104 0.6078 - 1.000 0.0499 Cutoff > 2.827
Sensitivity 90%
Specificity 20%
Severe 0h 0.88 0.077 0.7322-1.000 0.0108 Cutoff > 5.232
Sensitivity 90.91%
Specificity 60%
24h 1 0 1.000 - 1.000 0.0021 Cutoff > 8.558
Sensitivity 100%
Specificity 93.33%
48h 1 0 1.000 - 1.000 <0.0001 Cutoff > 8.558
Sensitivity 100%
Specificity 93.33%
tau Mild/Moderate 0h 0.84 0.06726 0.7119 - 0.9756 0.0170 Cutoff >0.3242 (pg/mL)
Sensitivity 100%
Specificity 25%
24h 0.82 0.0801 0.6602 - 0.9744 0.0738 Cutoff >0.2863 (pg/mL)
Sensitivity 100%
Specificity 25%
Severe 0h 0.90 0.07778 0.7510 - 1.000 <0.0001 Cutoff >0.1725 (pg/mL)
Sensitivity 100%
Specificity 12.5%
24h 0.93 0.06715 0.7996 - 1.000 0.0003 Cutoff >0.4417 (pg/mL)
Sensitivity 100%
Specificity 31.25%
48h 0.99 0.01664 0.9570 - 1.000 0.0008 Cutoff >2.253 (pg/mL)
Sensitivity 100%
Specificity 93.75%
UCH-L1 Severe 0h 0.90 0.0593 0.7929 - 1.000 0.0013 Cutoff >28.60 (pg/mL)
Sensitivity 100%
Specificity 61.54%
24h 0.95 0.05452 0.8393 - 1.000 0.1051 Cutoff >14.39 (pg/mL)
Sensitivity 100%
Specificity 46.15%
48h 0.96 0.04772 0.8625 - 1.000 0.0004 Cutoff >47.18 (pg/mL)
Sensitivity 100%
Specificity 69.23%
S100B Mild/Moderate 0h 0.87 0.069 0.7382 - 1.000 0.0223 Cutoff > 16.99 (pg/mL)
Sensitivity 95%
Specificity 25%
Severe 0h 1 0 1.000 - 1.000 <0.0001 Cutoff > 29.80 (pg/mL)
Sensitivity 100%
Specificity 75%
24h 1 0 1.000 - 1.000 <0.0001 Cutoff > 29.80 (pg/mL)
Sensitivity 100%
Specificity 75%
48h 1 0 1.000 - 1.000 0.0004 Cutoff > 29.80 (pg/mL)
Sensitivity 100%
Specificity 75%
p-tau181 Mild/Moderate 0h 0.87 0.06213 0.7500 - 0.9936 0.0018 Cutoff >0.1800 (pg/mL)
Sensitivity 100%
Specificity 30.77%
48h 0.93 0.04888 0.8378 - 1.000 0.0311 Cutoff >0.2350 (pg/mL)
Sensitivity 100%
Specificity 53.85%
Severe 0h 1 0 1.000 - 1.000 <0.0001 Cutoff >0.3400 (pg/mL)
Sensitivity 100%
Specificity 92.31%
24h 0.84 0.09189 0.6584 - 1.000 0.0681 Cutoff >0.1350 (pg/mL)
Sensitivity 100%
Specificity 2308,00%
48h 0.91 0.08145 0.7557 - 1.000 0.0006 Cutoff >0.1100 (pg/mL)
Sensitivity 100%
Specificity 15.38%

Glasgow Coma Scale (GCS); Area Under the Curve (AUC); standard error (SE); confidence knterval (CI).a GCS Severity was clustered into two categories: Mild-Moderate (GCS 9–15) and Severe (3–8).

b

Only biomarkers with statistically significant p values are shown in this table.

We considered the mild GCS category a critical subgroup of patients for our study, however, which is the reason an additional independent analysis of biomarkers in the mild GCS category was performed to determine their diagnostic capability in this critical subgroup of patients. In the stand-alone mild GCS category, GFAP, tau, S100B, and p-tau181 were statistically significant at 0h (p value 0.0151, 0.0113, 0.0081, and 0.0015, respectively), while NfL, UCH-L1, and p-tau181 were elevated at 48h measurements (p value 0.0122, 0.0062, and 0.0042) (Supplementary Table 1).

Of the four biomarkers significant at enrollment, S100B, tau, and p-tau181 obtained AUCs >0.7 (Supplementary Fig. S1). In the combined mild/moderate GCS category, serum levels for all biomarkers, except for p-tau181 at 24h, were significantly higher in the severe TBI group compared with HC at all three-time intervals (p < 0.05) (Fig. 5), with AUCs >0.9 for all measured biomarkers (Fig. 6).

FIG. 5.

FIG. 5.

Serum biomarker profiles in pediatric patients with traumatic brain injury (pTBI) who have different severity were categorized by initial Glasgow Coma Scale (GCS) score. Box and whisker plots showing all six biomarker levels in pg/mL after TBI categorized by initial severity score (Mild/Moderate GCS = 9–15, Severe GCS = 3–8) at 0h, 24h, and 48h post-injury with respective p values listed above. Box and whisker plots are shown for the 5th and 95th percentile. Dots correspond to data points outside the 5th and 95th percent confidence interval.

FIG. 6.

FIG. 6.

Receiver operating characteristic (ROC) of biomarkers of severe traumatic brain injury by initial Glasgow Coma Scale (GCS) score. Combined graphic with multiple ROC curves that indicate the area under the curve (sensitivity vs. 1-specificity) of GFAP, NfL, tau, UCH-L1, S100b, and p-tau181 measured at 0h, 24h, and 48h as biomarkers of severe injury as categorized per initial GCS score.

The GFAP, tau, and S100B levels demonstrated the ability to differentiate TBI in the mild/moderate group in contrast to HCs when measured at 0h post-injury (p value 0.008, 0.017, and 0.0223, respectively) with an AUC of 0.83 for GFAP, 0.84 for tau, and 0.87 for S100B. Further, tau showed increased concentrations in the mild/moderate group versus HC at 24h (p value 0.0172). NfL could also differentiate severity in the mild/moderate category at 48 h versus the H.C. group (p value 0.0499) with an AUC of 0.81 (Fig. 7).

FIG. 7.

FIG. 7.

Receiver operating characteristic (ROC) of biomarkers of mild to moderate traumatic brain injury by initial Glasgow Coma Scale (GCS) score. (A) ROC curves that indicate the area under the curve (AUC) (sensitivity vs. 1-specificity) of GFAP at 0h, NfL at 48h, and S100B at 0h as biomarkers of mild to moderate traumatic brain injury as categorized per initial GCS score. (B) Combined graphic with multiple ROC curves that indicate the AUCs (sensitivity vs. 1-specificity) of tau at 0h and 24h; and p-tau181 at 0h and 48h as biomarkers of mild to moderate injury as categorized per initial GCS score.

Serum p-tau181 levels were increased at enrollment and 48h post-injury in both the mild/moderate and severe categories, demonstrating its potential utility to differentiate TBI categories based on GCS. These results indicate a diagnostic role of injury severity (mild, mild/moderate, and severe TBI) in the pediatric population for GFAP, NfL, UCH-L1, S100B, tau, and p-tau181.

GFAP, NfL, UCH-L1, S100B, tau, and p-tau181 serum levels in correlation with critical care variables and clinical scores

Values of all six biomarkers were analyzed at three post-injury intervals—enrollment, 24h, and 48h—and correlated with intensive care variables essential for evaluating the severity of illness and predicting the outcome of patients with pTBI, as described in the Methods section. Logistic regression was used to analyze the relationship between binary critical care variables and serum biomarker levels.

Positive correlations were found between NfL at 0h and ICP monitoring (odds ratio [OR]: 1.04, p value 0.03347) and GFAP at 0h and cardiac arrest (OR: 1.00, p value 0.04727). Spearman rank correlation was computed to assess the relationship between ISS, PRISM, and Marshall scores, and serum biomarker levels; however, none yielded statistically relevant results.

Discussion

This study analyzed six serum biomarkers levels (GFAP, NfL, UCH-L1, S100B, tau, and p-tau181) and investigated their potential diagnostic and prognostic value in patients with pTBI. At 0h post-TBI, p-tau181 was the most promising biomarker for revealing neurological injury, strongly associated with poor GOS-E Peds outcomes (≥5) as early as 2–6 weeks post-initial injury. These findings were consistent with the poor prognostic assessment at 6–9 months and 12 months (Fig. 1, Fig. 2). In addition, NfL, UCH-L1, S100B, and tau all demonstrated a significant relationship with unfavorable outcomes at 12 months when measured at 0h after TBI (Fig. 2).

Although the sample size was small, our data suggest with 95% confidence that there is enough statistical evidence to demonstrate a strong association between p-tau181 levels sampled at enrollment and an unfavorable outcome assessed by GOS-Peds after a TBI event.

It is important to note, however, that the validation of cutoff values was beyond the scope of this study, and normalized values for serum concentrations of these biomarkers in both healthy pediatric patients and children with TBI remain largely unknown.

These findings are consistent with previous studies in adults that have demonstrated the clinical utility of these biomarkers as adjunctive diagnostic tools to traditional diagnostic modalities. For example, a 2014 study by Papa and associates31 compared GFAP and S100B measurements in adult patients with TBI and found that GFAP outperformed S100B in diagnosing intracranial injuries and was more specific as a neurological biomarker.

Another study by Papa and colleagues32 in 2016 concluded that serum GFAP in children and young adult TBI cohorts could classify injury severity into mild and moderate categories and predict clinical outcomes. Bazarian and coworkers also found that UCH-L1 and GFAP had high sensitivity and negative predictive value, suggesting their potential clinical utility in ruling out the need for a CT scan in adult patients with TBI.20

Our data aligns with these findings, because GFAP differentiated between TBI and HCs, distinguished between mild/moderate and severe injuries, and predicted long-term clinical outcomes in our pediatric cohort. It is important to note that further studies are needed to assess the utility of other relevant biomarkers, such as p-tau181, in pediatric TBI. In addition, more research is necessary to determine whether GFAP and UCH-L1 perform similarly in the pediatric population as they do in adult studies and to assess whether the possible clinical applications of these biomarkers can be extrapolated to this patient population.33

The biomarkers analyzed in the present study represent most of the pathomechanistic responses encountered in TBI. For instance, GFAP's efficacy as a diagnostic marker and a predictor of outcomes in adults with TBI has increased research interest within the pTBI population.20,32,34,35 NfL is a structural scaffolding protein exclusively expressed in neurons; it is found in elevated levels after axonal damage in adults.36,37

Tau is a microtubule-binding protein that regulates and stabilizes axonal microtubule assembly and disassembly. When tau undergoes post-translational modifications in the form of hyperphosphorylation, it yields p-tau, an abnormal insoluble tau protein. Adult studies have shown elevations in t-tau 1h after TBI, with the persistence of elevated levels up to six days after injury.38 In addition, initial cleaved tau protein in patients with severe TBI was found to have a 92% sensitivity and a 94% specificity in predicting outcomes.39 Moreover, Stukas and colleagues40 concluded that tau could be used as a biomarker for mild pTBI, in concordance with the findings of this study.

UCH-L1 is almost exclusively localized to the cytoplasm of neurons and plays a significant role in neuronal development. After TBI, UCH-L1 concentrations rapidly increase with quick release into various biofluids. This strong temporal relationship enables UCH-L1 to be an early diagnostic biomarker of TBI in adults.19,20

S100B is a calcium-binding protein found in astrocytes, glial cells, adipose tissue, cardiac cells, and skeletal muscles. It binds and promotes neurite extension leading to increased cell survival. S100B demonstrates a strong negative predictive value in detecting intracranial lesions, making it useful in the early detection and prognostication of TBI.41

Our study has various strengths. First, it demonstrates an association between increased serum levels of biomarkers and early GCS assessment, not only in the mild/moderate GCS category but also in a separate study conducted solely in the mild GCS category, indicating that these biomarkers may be valuable clinical tools on admission for promptly assessing TBI severity. The results showed elevated serum levels of GFAP, tau, S100B, and p-tau181, measured at enrollment, in the combined mild/moderate GCS category.

In the stand-alone mild GCS category, tau, S100B, and p-tau181 were statistically significant at 0hours, with AUC values >0.7 (Supplementary Fig. S1). Because the current therapeutic trend aims to diagnose mild pTBI as accurately and timely as possible, these findings offer promise for future use of these biomarkers in acute, low-resource settings where mild brain injuries are frequently sustained but not well diagnosed.

In a study conducted in a lower-resource emergency department by Sawaya and associates,42 the implementation of the Pediatric Emergency Care Applied Research Network (PECARN) prediction rules for the appropriate use of CT in clinically important traumatic brain injury (ciTBI) did not significantly change the CT rate nor increase the number of missed ciTBI.42 In such instances, a serum brain biomarker panel could be potentially valuable, as well as in settings where imaging studies are not readily available, particularly for high-risk mild TBIs that may require imaging studies.

Second, this study sheds light on the potential utility of biomarkers as prognostic tools for predicting pTBI outcomes based on the GOS-E Peds scores. In our study, GFAP demonstrated a statistically significant ability to differentiate favorable versus unfavorable outcomes at 48h post-injury, which suggests it may serve as a useful TBI monitoring tool, particularly in cases where imaging is unremarkable or limited, or when GCS accuracy is unreliable, such as in mechanically ventilated patients or those with pre-existing cerebral damage.

In addition, p-tau181 measured at 0h was also significant for distinguishing unfavorable versus favorable neurological outcomes at our earliest time point (2–6 weeks), as well as in subsequent assessments at 6–9 and 12 months, indicating it may be an optimal biomarker for identifying and categorizing pTBI severity, as well as predicting both short- and long-term neurological outcomes. Similar results were reported by Rubenstein and coworkers21 where a variant of p-tau, known as PT231 tau, outperformed total tau levels in discriminating TBI severity, changes in imaging studies, and outcome categories.

Finally, this study highlights the significance of brain injury biomarkers as novel tools with potential diagnostic and prognostic value in critically ill patients. These biomarkers may serve as stand-alone diagnostic tools or be used with other essential clinical variables to evaluate patient prognosis. Importantly, it should be emphasized that the utility of blood-based brain injury biomarkers extends beyond the realm of patients with TBI alone, encompassing a broader range of critically ill individuals.

In a multi-center prospective cohort study by Fink and associates,43 brain injury biomarkers were measured in serum samples from post-cardiac arrest patients, where NfL was strongly associated with unfavorable outcomes, categorized as death or unfavorable adaptive behavior one year post-event. In our analysis, we observed a significant statistical association between cardiac arrest and elevated serum levels of GFAP at enrollment. In addition, ICP monitoring demonstrated a significant relationship with increased levels of NfL at 0h.

Although the remaining variables examined in this study did not yield statistically significant findings, it is important to note that this could be attributed to the limitations imposed by the sample size.

Nevertheless, previous research provides substantial evidence supporting the correlation between elevated serum brain biomarker levels and certain critical care variables investigated in our study, stressing their potential for predicting patient outcomes. For instance, Galuee and coworkers44 conducted a prospective study that found increased serum levels of GFAP, NfL, and tau in neonates with low cord pH and moderate to severe hypoxia compared to healthy controls.44 Similarly, Ennen and colleagues45 reported an association between increased serum GFAP levels and neonates with hypoxic-ischemic encephalopathy during their first week of life and positive brain injury prediction in MRI studies. These findings emphasize the importance of further research in this area, because it holds the potential to yield critical results.

Notwithstanding the promising outcomes presented in this work, our results should be analyzed considering the study's limitations. First, GOS-E Peds outcomes obtained at six-, nine-, and 12-months post-injury were analyzed to produce a binary favorable versus unfavorable outcome results because of the limited subject numbers in each category. Therefore, the results of this analysis should be interpreted with these restrictions in mind.

In addition, despite our efforts, because of the nature of the study and the limited sample size, we cannot account for selection bias, which can lead to overrepresentation of favorable results.

Finally, none of the biomarkers analyzed in this study are readily available for clinical practice within a pediatric patient cohort. Despite these limitations, we believe our findings constitute a fair and rigorous addition to the current literature on this novel topic.

Conclusions

All six biomarkers included in this research— GFAP, NfL, UCH-L1, S100B, tau, and p-tau181— provided significant diagnostic information when obtained 0–48h post-injury. Our data suggest that GFAP, p-tau181, and tau could possibly serve as a diagnostic tool in mild to moderate pTBI when measured at 0h post-injury. In addition, p-tau181 measured at 0h could potentially be associated with short- and long-term unfavorable outcomes in this pediatric population.

Based on these findings, it may be possible to use an assessment of individual neuronal injury biomarker levels to diagnose and manage TBI in children. These results pave the way for future studies to further evaluate these biomarkers' potential for use in clinical settings.

Transparency, Rigor, and Reproducibility Summary

The sample size was 20 per group based on an alpha (α) of 0.05 (two-tailed) and power (1 –β) of 0.80. For protein biomarkers, we projected at least a 1.4-fold increase in the unfavorable outcome group compared to the favorable outcome group of mild/moderate TBI (GCS 9-15) versus severe (GCS 3-8). Our lab rigorously followed the common data elements biomarker guidelines for collecting, processing, and storing the studied biomarkers. (26) This study followed the Strengthening the Reporting of Observational Studies in Epidemiology STROBE.(43)

Supplementary Material

Supplemental data
Suppl_TableS1.docx (16.5KB, docx)
Supplemental data
Suppl_FigureS1.docx (121.4KB, docx)

Acknowledgments

We acknowledge logistical support from The Miami Clinical and Translational Science Institute (CTSI), The University of Florida College of Medicine Medical Student Research Program/MSRP (L.L.), The McKnight Brain Institute, and The University of Florida Pediatric Critical Care Medicine Student Investigators (PACISI), Department of Pediatric Critical Care and Department of Emergency Medicine, University of Florida College of Medicine.

Contributor Information

Collaborators: the Florida pTBI Consortium

Authors' Contributions

Jennifer C. Munoz Pareja contributed to the conception, design, acquisition, analysis, and interpretation of data and drafted the manuscript. Juan Pablo de Rivero Vaccari contributed to the analysis and interpretation of data and substantively revised the manuscript. Maria Mateo Chavez contributed to the analysis and interpretation of data and substantively revised the manuscript. Maria Kerrigan contributed to the acquisition, analysis, and interpretation of data and substantively revised the manuscript. Charlene Pringle contributed to the acquisition, analysis, and interpretation of data and substantively revised the manuscript. Kourtney Snodgrass contributed to the acquisition, analysis, and interpretation of data. Kathryn Swaby contributed to the acquisition, analysis, and interpretation of data and substantively revised the manuscript. Firas Kobeissy contributed to the acquisition, analysis, and interpretation of data. K Leslie Avery contributed to the conception and design and substantively revised the manuscript. Suman Ghosh, MD, contributed to the conception, design, and interpretation of data. Rajderkar, Dhanashree, contributed to the conception, design, and acquisition of data. Prashanth Shanmugham contributed to the acquisition, analysis, and interpretation of data. Lauren A. Lautenslager contributed to the acquisition, analysis, and interpretation of data and substantively revised the manuscript. Shannon Faulkenberry contributed to the acquisition, analysis, and interpretation of data and substantively revised the manuscript. Maria C. Pareja Zabala contributed to the acquisition, analysis, interpretation of data and substantively revised the manuscript. Lance S. Governale, contributed to the conception, design, and interpretation of data. Jason E. Blatt contributed to the conception, design, and interpretation of data. Nora AlFakhri contributed to the acquisition, analysis, interpretation of data. Jennifer Coto contributed to the acquisition, analysis, interpretation of data. Joslyn Gober contributed to the acquisition, analysis, interpretation of data. Paula Karina Perez contributed to the acquisition, analysis, interpretation of data. Heather McCrea contributed to the acquisition, analysis, interpretation of data. Chad Thorson contributed to the acquisition, analysis, interpretation of data. Kristine H. O'Phelan contributed to the acquisition, analysis, interpretation of data. Juan Solano contributed to the acquisition, analysis, interpretation of data. Ricardo Loor Torres contributed to the analysis, interpretation of data and substantively revised the manuscript. Robert W. Keane contributed to the interpretation of data and substantively revised the manuscript. W. Dalton Dietrich contributed to the interpretation of data and substantively revised the manuscript. Kevin K. Wang contributed to the conception, design, acquisition, analysis, interpretation of data and drafted the manuscript.

Funding Information

The project described was supported by Grant Number UL1TR002736, Miami Clinical and Translational Science Institute, from the National Center for Advancing Translational Sciences and the National Institute on Minority Health and Health Disparities. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.

Author Disclosure Statement

KKW is a shareholder of Gryphon Bio Inc. and Banyan Biomarkers Inc. For the remaining authors, no competing financial interests exist.

Supplementary Material

Supplementary Figure S1

Supplementary Table S1

References

  • 1. Krug EG, Sharma GK, Lozano R. The global burden of injuries. Am J Public Health 2000;90(4):523–526; doi: 10.2105/ajph.90.4.523 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Faul M, Xu L, Wald MM, et al. Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations and Deaths 2002–2006. Centers for Disease Control and Prevention: Atlanta, GA; 2010. [Google Scholar]
  • 3. Araki T, Yokota H, Morita A. Pediatric traumatic brain injury: characteristic features, diagnosis, and management. Neurol Med Chir (Tokyo) 2017;57(2):82–93; doi: 10.2176/nmc.ra.2016-0191 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Centers for Disease Control and Prevention: Moderate and Severe TBI. November 2022. Available from: www.cdc.gov/traumaticbraininjury/moderate-severe/index.html [Last accessed: 10/23/22]
  • 5. Thurman DJ, Alverson C, Dunn KA, et al. Traumatic brain injury in the United States: a public health perspective. J Head Trauma Rehabil 1999;14(6):602–615; doi: 10.1097/00001199-199912000-00009 [DOI] [PubMed] [Google Scholar]
  • 6. Popernack ML, Gray N, Reuter-Rice K. Moderate-to-severe traumatic brain injury in children: complications and rehabilitation strategies. J Pediatr Health Care 2015;29(3):e1–7; doi: 10.1016/j.pedhc.2014.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Pareja JCM, Li X, Gandham N, et al. Biomarkers in moderate to severe pediatric traumatic brain injury: a review of the literature. Ped Neurol 2022;130(60–68); doi: 10.1016/j.pediatrneurol.2022.03.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Narayan RK, Michel ME, Ansell B, et al. Clinical trials in head injury. J Neurotrauma 2002;19(5):503–557; doi: 10.1089/089771502753754037 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974;2(7872):81–84; doi: 10.1016/s0140-6736(74)91639-0 [DOI] [PubMed] [Google Scholar]
  • 10. Miglioretti DL, Johnson E, Williams A, et al. The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr 2013;167(8):700–707; doi: 10.1001/jamapediatrics.2013.311 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Brenner DJ, Hall EJ. Computed tomography — an increasing source of radiation exposure. N Engl J Med 2007;357(22):2277–2284; doi: 10.1056/NEJMra072149 [DOI] [PubMed] [Google Scholar]
  • 12. Ghaffarpasand F, Razmkon A, Dehghankhalili M. Glasgow Coma Scale score in pediatric patients with traumatic brain injury; limitations and reliability. Bull Emerg Trauma 2013;1(4):135–136 [PMC free article] [PubMed] [Google Scholar]
  • 13. Chung CY, Chen CL, Cheng PT, et al. Critical score of Glasgow Coma Scale for pediatric traumatic brain injury. Pediatr Neurol 2006;34(5):379–387; doi: 10.1016/j.pediatrneurol.2005.10.012 [DOI] [PubMed] [Google Scholar]
  • 14. Hillered L, Vespa PM, Hovda DA. Translational neurochemical research in acute human brain injury: the current status and potential future for cerebral microdialysis. J Neurotrauma 2005;22(1):3–41; doi: 10.1089/neu.2005.22.3 [DOI] [PubMed] [Google Scholar]
  • 15. Figaji AA, Zwane E, Fieggen AG, et al. Acute clinical grading in pediatric severe traumatic brain injury and its association with subsequent intracranial pressure, cerebral perfusion pressure, and brain oxygenation. Neurosurg Focus 2008;25(4):E4; doi: 10.3171/foc.2008.25.10.E4 [DOI] [PubMed] [Google Scholar]
  • 16. Mayer AR, Kaushal M, Dodd AB, et al. Advanced biomarkers of pediatric mild traumatic brain injury: Progress and perils. Neurosci Biobehav Rev 2018;94(149–165; doi: 10.1016/j.neubiorev.2018.08.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. McAbee GN. Are we permitting pediatric athletes with sports-related concussion to return to play too soon after concussion? J Child Neurol 2018;33(12):759–761; doi: 10.1177/0883073818790169 [DOI] [PubMed] [Google Scholar]
  • 18. May KH, Marshall DL, Burns TG, et al. Pediatric sports specific return to play guidelines following concussion. Int J Sports Phys Ther 2014;9(2):242–255 [PMC free article] [PubMed] [Google Scholar]
  • 19. Wang KK, Yang Z, Zhu T, et al. An update on diagnostic and prognostic biomarkers for traumatic brain injury. Expert Rev Mol Diagn 2018;18(2):165–180; doi: 10.1080/14737159.2018.1428089 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Bazarian JJ, Biberthaler P, Welch RD, et al. Serum GFAP and UCH-L1 for prediction of absence of intracranial injuries on head CT (ALERT-TBI): a multicentre observational study. Lancet Neurol 2018;17(9):782–789; doi: 10.1016/s1474-4422(18)30231-x [DOI] [PubMed] [Google Scholar]
  • 21. Rubenstein R, Chang B, Yue JK, et al. Comparing Plasma Phospho tau, total tau, and phospho tau-total tau ratio as acute and chronic traumatic brain injury biomarkers. JAMA Neurol 2017;74(9):1063–1072; doi: 10.1001/jamaneurol.2017.0655 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Undén J, Ingebrigtsen T, Romner B. Scandinavian guidelines for initial management of minimal, mild and moderate head injuries in adults: an evidence and consensus-based update. BMC Med 2013;11(50); doi: 10.1186/1741-7015-11-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Food and Drug Administration. FDA authorizes marketing of first blood test to aid in the evaluation of concussion in adults. February 2018.Available from: www.fda.gov/news-events/press-announcements/fda-authorizes-marketing-first-blood-test-aid-evaluation-concussion-adults [Last accessed: 09/17/22]
  • 24. Agoston DV, Shutes-David A, Peskind ER. Biofluid biomarkers of traumatic brain injury. Brain Inj 2017;31(9):1195–1203; doi: 10.1080/02699052.2017.1357836 [DOI] [PubMed] [Google Scholar]
  • 25. Papa L, Ramia MM, Kelly JM, et al. Systematic review of clinical research on biomarkers for pediatric traumatic brain injury. J Neurotrauma 2013;30(5):324–338; doi: 10.1089/neu.2012.2545 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th Ed. American Psychiatric Publishing: Washington, DC; 2013 [Google Scholar]
  • 27. Adelson PD, Pineda J, Bell MJ, et al. Common data elements for pediatric traumatic brain injury: recommendations from the working group on demographics and clinical assessment. J Neurotrauma 2012;29(4):639–653; doi: 10.1089/neu.2011.1952 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Manley GT, Diaz-Arrastia R, Brophy M, et al. Common data elements for traumatic brain injury: recommendations from the biospecimens and biomarkers working group. Arch Phys Med Rehabil 2010;91(11):1667–1672; doi: 10.1016/j.apmr.2010.05.018 [DOI] [PubMed] [Google Scholar]
  • 29. Czeiter E, Amrein K, Gravesteijn BY, et al. Blood biomarkers on admission in acute traumatic brain injury: Relations to severity, CT findings and care path in the CENTER-TBI study. EBioMedicine 2020;56(102785); doi: 10.1016/j.ebiom.2020.102785 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. McCauley SR, Wilde EA, Anderson VA, et al. Recommendations for the use of common outcome measures in pediatric traumatic brain injury research. J Neurotrauma 2012;29(4):678–705; doi: 10.1089/neu.2011.1838 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Papa L, Silvestri S, Brophy GM, et al. GFAP out-performs S100β in detecting traumatic intracranial lesions on computed tomography in trauma patients with mild traumatic brain injury and those with extracranial lesions. J Neurotrauma 2014;31(22):1815–1822; doi: 10.1089/neu.2013.3245 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Papa L, Mittal MK, Ramirez J, et al. In children and youth with mild and moderate traumatic brain injury, glial fibrillary acidic protein out-performs S100β in detecting traumatic intracranial lesions on computed tomography. J Neurotrauma 2016;33(1):58–64; doi: 10.1089/neu.2015.3869 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Korley FK, Jain S, Sun X, et al. Prognostic value of day-of-injury plasma GFAP and UCH-L1 concentrations for predicting functional recovery after traumatic brain injury in patients from the US TRACK-TBI cohort: an observational cohort study. Lancet Neurol 2022;21(9):803–813; doi: 10.1016/s1474-4422(22)00256-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Mondello S, Kobeissy F, Vestri A, et al. Serum concentrations of ubiquitin C-terminal Hydrolase-L1 and glial fibrillary acidic protein after pediatric traumatic brain injury. Sci Rep 2016;6(28203); doi: 10.1038/srep28203 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Wang KK, Yang Z, Yue JK, et al. Plasma anti-glial fibrillary acidic protein autoantibody levels during the acute and chronic phases of traumatic brain injury: a transforming research and clinical knowledge in traumatic brain injury pilot study. J Neurotrauma 2016;33(13):1270–1277; doi: 10.1089/neu.2015.3881 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Khalil M, Teunissen CE, Otto M, et al. Neurofilaments as biomarkers in neurological disorders. Nat Rev Neurol 2018;14(10):577–589; doi: 10.1038/s41582-018-0058-z [DOI] [PubMed] [Google Scholar]
  • 37. Oliver JM, Jones MT, Kirk KM, et al. Serum neurofilament light in American football athletes over the course of a season. J Neurotrauma 2016;33(19):1784–1789; doi: 10.1089/neu.2015.4295 [DOI] [PubMed] [Google Scholar]
  • 38. Ling H, Hardy J, Zetterberg H. Neurological consequences of traumatic brain injuries in sports. Mol Cell Neurosci 2015;66(Pt B):114–122; doi: 10.1016/j.mcn.2015.03.012 [DOI] [PubMed] [Google Scholar]
  • 39. Zemlan FP, Jauch EC, Mulchahey JJ, et al. C-tau biomarker of neuronal damage in severe brain injured patients: association with elevated intracranial pressure and clinical outcome. Brain Res 2002;947(1):131–139; doi: 10.1016/s0006-8993(02)02920-7 [DOI] [PubMed] [Google Scholar]
  • 40. Stukas S, Higgins V, Frndova H, et al. Characterisation of serum total tau following paediatric traumatic brain injury: a case-control study. Lancet Child Adolesc Health 2019;3(8):558–567; doi: 10.1016/s2352-4642(19)30194-4 [DOI] [PubMed] [Google Scholar]
  • 41. Oris C, Pereira B, Durif J, et al. The biomarker s100b and mild traumatic brain injury: a meta-analysis. Pediatrics 2018;141(6); doi: 10.1542/peds.2018-0037 [DOI] [PubMed] [Google Scholar]
  • 42. Sawaya RD, Wakil C, Wazir A, et al. Does implementation of the PECARN rules for minor head trauma improve patient-centered outcomes in a lower resource emergency department: a retrospective cohort study. BMC Pediatr 2020;20(1):439; doi: 10.1186/s12887-020-02328-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Fink EL, Kochanek PM, Panigrahy A, et al. Association of blood-based brain injury biomarker concentrations with outcomes after pediatric cardiac arrest. JAMA Netw Open. 2022. Sep 1;5(9):e2230518; doi: 10.1001/jamanetworkopen.2022.30518 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Gaulee P, Yang Z, Sura L, et al. Concentration of serum biomarkers of brain injury in neonates with a low cord pH with or without mild hypoxic-ischemic encephalopathy. Front Neurol 2022;13:934755; doi: 10.3389/fneur.2022.934755 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Ennen CS, Huisman TA, Savage WJ, et al. Glial fibrillary acidic protein as a biomarker for neonatal hypoxic-ischemic encephalopathy treated with whole-body cooling. Am J Obstet Gynecol 2011;205(3):251.e1-7; doi: 10.1016/j.ajog.2011.06.025 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental data
Suppl_TableS1.docx (16.5KB, docx)
Supplemental data
Suppl_FigureS1.docx (121.4KB, docx)

Articles from Journal of Neurotrauma are provided here courtesy of Mary Ann Liebert, Inc.

RESOURCES