Abstract
Globally, pediatric traumatic brain injury (TBI) is a significant cause of neurodisability, with a disproportionate burden in low-resource countries with limited neuroimaging access. Blood biomarkers may aid in diagnosis and triage. We evaluated glial fibrillary acidic protein (GFAP) and ubiquitin carboxy-terminal hydrolase L1 (UCH-L1) in Ugandan children with TBI to assess their utility in predicting injury severity, computed tomography (CT) abnormalities, and short-term neurodisability. This prospective study was conducted at Mulago National Referral Hospital in Uganda (2022–2024). Children aged 5–15 years with TBI were enrolled within 48 h of injury, along with uninjured controls. TBI severity was classified using the Glasgow Coma Scale as mild (13–15) or moderate/severe (≤12). CT was performed when clinically indicated. Plasma GFAP and UCH-L1 were measured using the Abbott i-STAT® Alinity platform. Neurological assessments were completed at admission, discharge, and 2-week follow-up. Biomarkers were compared across TBI severity, CT status, and neurological outcomes. Diagnostic performance was assessed using receiver operating curve analysis and logistic regression. Sixty-five children with TBI (mean [SD] age 9.13 [3.06] years; 58% male) and 40 controls (10.27 [2.87] years; 60% male) were included. TBI patients had significantly elevated biomarker levels (GFAP: 704 vs. 30 pg/mL; UCH-L1: 80 vs. 40 pg/mL; p ≤ 0.01). GFAP was higher in moderate/severe vs. mild TBI (p = 0.006), showed excellent diagnostic performance (area under the curve [AUC] 0.93), and severity discrimination (AUC 0.74). GFAP predicted CT abnormalities (AUC 0.87) and neurological deficits at discharge and follow-up (OR 3.06; 3.37). UCH-L1 added limited value to combined models. GFAP showed strong diagnostic and prognostic potential for pediatric TBI and may aid in early diagnosis and triage where CT is unavailable.
Introduction
Traumatic brain injury (TBI) is a leading cause of death and acquired neurodisability worldwide, with an exceptionally high burden in low- and middle-income countries (LMICs). 1 In low-resource settings, delays in care, lack of neuroimaging infrastructure, and variability in clinical management contribute to poor outcomes, particularly in children who are at risk of long-term neurodisability due to ongoing brain development and heightened risk of repeated injury.2–5 In sub-Saharan Africa, the incidence of TBI exceeds 800 per 100,000, more than eight times the global average.1,6 Recent regional trauma studies have underscored the absence of standardized diagnostic and treatment protocols as well as wide variation in provider decision-making as key contributors to TBI-related morbidity. 7
Computerized tomography (CT) remains the gold standard for diagnosing structural brain injury. However, its availability is often limited in LMICs due to high costs, logistical barriers, radiation exposure, and the need for trained radiologists. 8 In contrast, high-income countries have increasingly adopted blood-based TBI biomarkers to guide triage and reduce unnecessary CT use in children with mild TBI (mild TBI).9,10 These advances offer a potential paradigm shift for LMICs, where both overuse and underuse of CT are common, potentially delaying timely care.
Glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase L1 (UCH-L1) have emerged as leading candidates for blood-based diagnosis of TBI. GFAP, an astrocyte intermediate filament protein, is released during astrogliosis and remains elevated for up to 24 h postinjury.11–13 UCH-L1 is a neuron-specific enzyme involved in tagging proteins for ubiquitination or degradation in response to cellular damage and maintaining synaptic protein homeostasis. It is the most abundant central nervous system (CNS) protein, released early postinjury but declining rapidly, achieving peak diagnostic performance within the first 8–12 h.11,14 In 2021, the U.S. Food and Drug Administration (FDA) approved the combined use of GFAP and UCH-L1 within 12 h of injury to assist in CT decision-making for adults with mild TBI. 15 The Abbott i-STAT Alinity analyzer enables the rapid, portable testing of these biomarkers from a small plasma sample in approximately 15 min, making it a promising tool for use in LMIC emergency settings. 10
In large adult studies (ALERT-TBI, TRACK-TBI), GFAP and UCH-L1 demonstrated high sensitivity (>95%) but modest specificity (<40%) for detecting CT abnormalities.16–19 Pediatric data remain limited but promising. A recent study from France reported 100% sensitivity and 67% specificity for detecting clinically meaningful TBI using these biomarkers. 20 However, no prospective studies have assessed their performance in African children or in LMIC health systems, where TBI management presents distinct challenges compared to those in high-income countries. 21
LMIC-specific factors, such as delayed presentation, comorbid infections (e.g., malaria), and varying CT utilization patterns, may impact biomarker kinetics and diagnostic utility. For instance, in malaria-endemic areas, comorbid neurological insults may confound biomarker interpretation. Prior studies in Ugandan children with severe malaria found that acute elevations in UCH-L1 and tau, but not GFAP, were associated with poor neurodevelopmental outcomes.22,23 These findings underscore the importance of validating biomarker performance in LMIC populations with overlapping disease burdens and a longer time to presentation.
This study investigated plasma GFAP and UCH-L1 levels in a prospective cohort of Ugandan children with mild TBI, moderate/severe TBI, and uninjured community children as controls. Using the i-STAT Alinity analyzer, we evaluated the diagnostic utility of GFAP and UCH-L1 for distinguishing TBI cases from controls, stratifying TBI severity, and predicting CT abnormalities and short-term neurological deficits. We hypothesized that GFAP and UCH-L1 levels would be significantly elevated in children with TBI, correlate with CT abnormalities, and identify children at risk for postdischarge neurodisability. We also explored the relative and combined diagnostic performance of these biomarkers in an LMIC emergency setting.
Methods
Study overview and informed consent
This prospective cohort pilot study was conducted at the Mulago National Referral Hospital (MNRH) in Kampala, Uganda, between June 2022 and July 2024. MNRH is the country’s primary national referral hospital, located in the capital, Kampala. School-aged children (5–15 years) who presented to the emergency department at MNRH with TBI were enrolled, and blood samples were collected within 48 h of injury. TBI severity was assessed at admission using the Glasgow Coma Scale (GCS) score and categorized as mild TBI (13–15) or moderate/severe TBI (3–12). A comparison group of uninjured children without active illness was recruited from the extended family or neighborhood of the children with TBI. Exclusion criteria for all participants included a history of head injury, chronic disease, or acute illness requiring prescription medication, such as malaria treatment, in the month before enrollment. Among the TBI cases, children with severe trauma requiring complex surgery or prolonged hospitalization were excluded, including those who underwent delayed surgical intervention after enrollment. The parent study was focused on long-term neurocognitive outcomes following mild-to-moderate TBI; therefore, children with severe injuries and a high risk of death or significant disability were not enrolled.
The study was approved by the institutional review boards of Makerere University School of Medicine (MAK-SOMREC-2022-340), Indiana University (#15282), and the Uganda National Council for Science and Technology (HS2628ES). Administrative clearance for the study was obtained from the Mulago National Referral Hospital (MHREC 2385). Written informed consent was obtained from the caregivers of the study participants, and assent from children aged 7 years and older. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Study procedures
The study was conducted in accordance with the standard-of-care protocols established by the MNRH emergency department. Demographic and clinical data were collected upon admission using standardized interviews, physical examinations, and neurological assessments. Neurological evaluations were repeated at discharge and at the 2-week follow-up. A composite score of motor and cranial nerve impairments, ataxia, movement, speech deficits, or visual impairments was used to determine the presence of neurological deficits at each time point. A neurologist independently assessed all positive neurology reports. Patients who met the criteria for CT scanning for clinical management, as outlined in the MNRH guidelines, underwent head CT scans. The study neurosurgeon interpreted the scans as part of routine care, as access to radiology experts is limited in the emergency department. Radiological abnormalities, including depressed or orbital skull fractures, epidural and subdural hematomas, diffuse axonal injury, and contusions, were documented in the clinical charts. CT findings were categorized as “abnormal” based on radiographical evidence of TBI. All available CT scans were independently reviewed by a study radiologist and classified according to the Marshall Classification System to characterize the lesion type and injury severity. 24
Blood sampling and biomarker assessments
Blood samples were obtained from patients with TBI within 48 h of injury and from controls at enrollment. Ethylenediaminetetraacetic acid (EDTA) plasma was stored at −80°C until analysis. Plasma GFAP and UCH-L1 levels were measured using the i-STAT Alinity analyzer (Abbott, NJ, USA), according to the manufacturer’s instructions.18,25 Testers were blinded to clinical data. The handheld analyzer employed TBI plasma cartridges with a built-in sandwich ELISA and electrochemical detection capabilities. Assay runtime was approximately 15 min. The GFAP assay had a calibration range of 0–50,000 pg/mL and a reportable range of 30–10,000 pg/mL. For UCH-L1, the calibration range was 0–20,000 pg/mL, and the reportable range was 200–3200 pg/mL. The limits of quantification were 23 and 70 pg/mL for GFAP and UCH-L1, respectively.
Statistical analysis
This analysis was based on a prospective convenience sample of children with TBI and uninjured controls who were enrolled during the study period. The sample size was determined using power calculations to detect differences in biomarker levels. Final enrollment was based on feasibility. The Shapiro–Wilk test was used to confirm that GFAP and UCH-L1 levels were non-normally distributed, justifying the use of nonparametric tests. Wilcoxon rank-sum tests were used to compare continuous variables, and Pearson’s χ2 tests were used for categorical variables. The Mann–Whitney U test was used to compare the GFAP and UCH-L1 levels between the groups. Biomarker concentrations were reported as medians with interquartile ranges (IQRs). Receiver operating characteristic (ROC) curves were used to assess the biomarker performance in distinguishing TBI cases from controls, differentiating TBI severity, and predicting abnormal CT findings. The area under the curve (AUC) and 95% confidence interval (CI) were calculated. Logistic regression models adjusted for age and sex were used to evaluate the associations between biomarker levels and neurological deficits at admission, discharge, and 2-week follow-up. The results are presented as odds ratios (ORs) with 95% CIs. A two-sided alpha level of 0.05 was used to determine the statistical significance. Analyses were performed using Stata version 18.0 (StataCorp LLC, TX, USA), and graphs were created using Prism version 10.0 (GraphPad Prism, SD, USA).
Results
Baseline demographic and clinical characteristics
Of the 125 children enrolled in the study, 105 had sufficient plasma available for testing TBI biomarkers (Fig. 1; Supplementary Table S1). Among 65 children with TBI, 58% were male, with 48% male patients with mild TBI and a mean (standard deviation, SD) age of 8.92 (2.9) years (N = 51). In contrast, the moderate/severe TBI group consisted of 60% males and had a mean (SD) age of 9.91 (11.3) years (N = 14) (Table 1). The control group included 60% males and had a mean age of 10.27 (2.87) years (N = 40) (Supplementary Table S1).
Flow diagram of study participant enrollment and inclusion in the study. The diagram depicts enrollment and inclusion of participants with traumatic brain injury (TBI) and uninjured community children as controls. Reasons for exclusion and the number of participants with available biomarker and CT data on admission are shown. Additionally, those assessed for neurological deficits on admission, discharge, and at 2-week follow-up are noted, including any missed visits. CT, computed tomography; TBI, traumatic brain injury.
Sociodemographic and Clinical Characteristics of Children with Traumatic Brain Injury and Available Plasma Biomarker Data
Significant p values (<0.05) indicated in bold.
Data presented as median (IQR) unless otherwise stated. “n” is listed in-line when N is lower than listed for each group.
p values were derived from pairwise comparison of children with mild versus moderate/severe TBI using the Wilcoxon rank-sum test for continuous variables and Pearson’s chi-square test for categorical variables.
AUC, area under the curve; CC, community children; GCS, Glasgow coma score; mod/sev, moderate/severe; GFAP, glial fibrillary acidic protein; ROC, receiver operating characteristic curve; UCH-L1, ubiquitin C-terminal hydrolase-L1.
The median time from injury to hospital admission for children with TBI was 22 h (IQR 17.50, 26.67); N = 40 children were enrolled within 24 h, and N = 25 in the 24- to 48-h window (Supplementary Table S2). CT was available for 46 of 65 children with TBI (71%), including 34 of 51 children with mild TBI (67%) and 12 of 14 children with moderate/severe TBI (86%). Abnormal findings were observed in 14 of 34 children with mild TBI (41%) and in 9 of 12 children with moderate/severe TBI (75%) (p = 0.044). CT abnormalities included skull fractures (temporal, parietal, supraorbital, and frontal), epidural and subdural hematomas, cerebral contusions, diffuse axonal injury, intraventricular hemorrhage, subgaleal hematoma, and pneumocephalus. Epidural hematomas and contusions were more common among children with mild TBI, while diffuse axonal injury and multifocal hemorrhages were more prevalent in moderate/severe TBI.
Over 40% of children with TBI presented with neurological deficits at admission that persisted until discharge. Persistent neurological deficits were observed in 12 children, including four (8%) with mild TBI and eight (61.54%) at the 2-week follow-up visit (Table 1).
Admission TBI biomarker levels and diagnostic utility
Biomarker levels (median [IQR] pg/mL) were significantly higher in 65 patients with TBI compared with 40 controls (GFAP: 704 [109, 2598] vs. 30 [30, 30]; UCH-L1: 80 [35, 147] vs. 40 [17, 67.5]; p ≤ 0.011) (Supplementary Table S1 and Supplementary Fig. S2). GFAP levels were higher in 14 children with moderate/severe TBI (3247 [870, 6981]) than in 51 children with mild TBI (602 [96, 1681]; p = 0.006). UCH-L1 levels did not differ by severity (mild TBI: 88.5 [45, 188] vs. moderate/severe TBI: 79 [17, 142]; p = 0.198) (Table 1, Fig. 2). Using the 95th percentile for biomarker levels in community children without active illness (GFAP: 39 pg/mL; UCH-L1: 229 pg/mL)—a threshold consistent with FDA validation studies for GFAP and UCH-L1—we observed elevated GFAP levels in 54 children (83%) with TBI, compared to 2 (5%) among the controls (p < 0.001). In contrast, UCH-L1 levels were elevated in nine children with TBI (13.9%), but this difference was not statistically significant compared with controls (p = 0.151). No differences in biomarker levels were observed across age groups, sexes, or time since injury (under or over 24 h) (Supplementary Tables S2 and S3).
Plasma GFAP and UCH-L1 levels and diagnostic performance by TBI status and severity.
Furthermore, GFAP demonstrated excellent diagnostic utility in differentiating TBI patients from community children (AUC = 0.93 [0.88, 0.98]) and in assessing TBI severity (AUC = 0.74 [0.57, 0.91]) (Fig. 2B). The diagnostic utility of UCH-L1 for differentiating TBI patients (0.65 [0.54, 0.75]) and for assessing TBI severity (0.64 [0.48, 0.79]) was lower (Fig. 2B).
Relationships between TBI biomarkers and CT outcomes
Plasma levels of GFAP and UCH-L1 were significantly higher in children with abnormal CT findings (GFAP: median 3504 pg/mL [IQR 704–6981]; UCH-L1: 124 pg/mL [73–280]) than in those with normal scans (GFAP: 202 [31–744]; UCH-L1: 62 [21–126]; all p ≤ 0.009) (Fig. 3A, B). Among the 19 children without CT scans available during clinical evaluation, median biomarker levels were lower than in children with CT abnormalities. ROC analysis showed that GFAP exhibited stronger diagnostic performance (AUC 0.87 [95% CI: 0.76–0.98]) than UCH-L1 (AUC 0.73 [0.58–0.87]) for predicting CT abnormalities (Fig. 3C). Logistic regression combining GFAP and UCH-L1 yielded a nearly identical AUC of 0.87 (95% CI: 0.76–0.98), suggesting that UCH-L1 added limited additional predictive value beyond GFAP alone.
Plasma GFAP and UCH-L1 levels and predictive utility in identifying CT abnormalities.
Prognostic utility and associations with neurological outcomes
Elevated GFAP levels were significantly associated with increased odds of neurological deficits at discharge (odds ratio, 3.06 [95% CI: 1.47, 6.37]; p = 0.003) and at the 2-week follow-up (3.37 [1.17, 9.72]; p = 0.030). No significant association was observed between UCH-L1 levels and neurological deficits at any time point. We also investigated whether CT abnormalities could predict neurological deficits and found associations with neurological deficits in the hospital (p ≤ 0.030), but not at follow-up (p = 0.160) (Fig. 4A).
Association of TBI biomarker concentrations and comparison to the association of CT findings with neurological deficits at admission, discharge, and 2-week follow-up.
ROC analysis demonstrated the prognostic utility of GFAP in predicting neurological deficits improved after admission (AUCs 0.59 [0.44, 0.73]) for 28 children with deficits at discharge (AUC 0.73 [0.61, 0.86]) and 12 children with deficits persisting at the 2-week follow-up (AUC 0.73 [0.57, 0.87]). UCH-L1 levels showed modest predictive power at discharge (AUC 0.65 [0.51, 0.79]) but not at admission (AUC 0.58 [0.44, 0.73]) or at the 2-week follow-up (AUC 0.57 [0.40, 0.74]) (Fig. 4B).
Discussion
To our knowledge, this is the first study to evaluate the utility of i-STAT plasma GFAP and UCH-L1 for TBI detection in an LMIC setting. We showed that GFAP and UCH-L1 levels are acutely elevated following TBI in school-aged Ugandan children. GFAP demonstrated superior performance in distinguishing TBI cases from controls and in reflecting TBI severity. More than 90% of community children without TBI had GFAP levels below the detection threshold, and the remaining 10% were within 11 pg/mL of the threshold, indicating a low false-positive rate for GFAP. Using 95th-percentile biomarker levels from community children, >80% of children with mild or moderate/severe TBI had elevated GFAP levels, compared with 5% of controls. Both GFAP and UCH-L1 levels were significantly elevated in children with CT abnormalities; however, GFAP showed greater sensitivity for predicting CT abnormalities. Furthermore, the combined predictive value of GFAP and UCH-L1 did not differ significantly from that of GFAP alone, suggesting that UCH-L1 has limited additional predictive value. CT scans were unavailable for 19 children in the study (17 patients with mild TBI). Biomarker levels in this group were comparable to those in children with normal CT findings, underscoring the potential of TBI biomarkers to exclude CT in mild cases. Finally, acutely elevated GFAP levels showed prognostic utility in detecting lingering post-TBI neurological deficits. Multisite studies with long-term follow-up are necessary to establish plasma TBI biomarkers as alternatives in low-resource emergency triage settings, where access to objective diagnostics and imaging is limited.
A decade of research on blood-based TBI biomarker detection has facilitated its integration into clinical practice. Papa et al. 26 analyzed serial blood samples from adult patients in the United States with mild-to-moderate TBI and found that UCH-L1 peaks within the first 8 h, whereas GFAP peaks between 12 and 24 h postinjury. GFAP has consistently demonstrated superior diagnostic accuracy and a stronger correlation with injury severity compared to UCH-L1. 27 Our LMIC-focused pediatric study aligns with data from high-income countries and suggests that GFAP outperforms UCH-L1. A recent U.S.-based study by Muñoz-Pareja et al. evaluated six serum biomarkers, including GFAP and UCH-L1, in 34 children with TBI and 19 controls. 28 Both biomarkers were elevated within 48 h of injury; UCH-L1 measured at 0 h and GFAP measured at 48 h predicted unfavorable outcomes up to 12 months postinjury. Our study enrolled nearly twice as many participants (65 TBI cases and 40 controls) and used a point-of-care platform reflective of low-resource clinical workflows. A potential explanation for the superior diagnostic utility of GFAP in our study is the median time of blood collection, which occurred 22 h postinjury. Delays in accessing medical care in Kampala are multifactorial and well-documented. 29 Our study timeframe likely coincides with the period when GFAP reaches its peak diagnostic performance, thereby enhancing its discriminatory capacity. GFAP remained a strong predictor of CT abnormalities and neurological deficits, while UCH-L1 added limited incremental value, likely due to its earlier peak and shorter half-life. The excellent diagnostic accuracy of GFAP (AUC 0.93) underscores the clinical relevance, and the significant effect size further supports its utility in distinguishing pediatric TBI cases from controls in LMIC settings. Although GFAP levels differed by TBI severity, subgroup analyses were limited by insufficient sample sizes in each subgroup. No differences in UCH-L1 levels were observed between children admitted before and after 24 h, although earlier sampling (<8 h post-injury) may have improved diagnostic performance. These findings add to the pediatric biomarker literature by demonstrating diagnostic and prognostic utility in a low-resource clinical setting.
Growing awareness of concussions has coincided with increased CT use, often driven by caution rather than clinical necessity. 30 Across African countries, imaging guidelines have demonstrated utility in optimizing CT utilization, particularly for mild TBI; however, variability in decision-making persists.8,31 At Mulago Hospital, although all head trauma cases are recommended for CT imaging, the modality is frequently unavailable or expensive. In particular, mild cases are triaged for observation until imaging is available or discharged after clinical and neurological assessments. A shortage of neuroradiology-trained staff compounds this limitation. 31 In high-income countries, GFAP and UCH-L1 have consistently demonstrated negative predictive values exceeding 97% for ruling out intracranial lesions in adults and children with TBI,26,32–34 leading to FDA approval of these biomarkers for the diagnosis of mild TBI in adults and to aid in clinical decision-making regarding the need for CT scans. The findings from our limited dataset align with prior studies, showing elevated GFAP and UCH-L1 levels in patients with CT abnormalities, as well as AUCs of 0.87 and 0.73 for predicting CT abnormalities, respectively.
Although our study evaluated GFAP and UCH-L1 separately to assess their individual diagnostic and prognostic performance, we acknowledge that these biomarkers have been approved by the FDA for combined use on the i-STAT analyzer, given their complementary kinetics and pathophysiological profiles. Prior studies have shown that GFAP and UCH-L1 are highly sensitive for detecting intracranial lesions when used in combination, particularly within the first 12 h postinjury. Owing to sample size and timing limitations in our pilot cohort, we focused on evaluating individual biomarkers. However, future analyses integrating a combinatorial GFAP–UCH-L1 model may enhance diagnostic accuracy and better reflect real-world use of the i-STAT analyzer, particularly for early triage and monitoring in low-resource settings.
Lastly, GFAP levels at admission were significantly elevated and showed acceptable predictive values for identifying neurological deficits at discharge and follow-up. These findings align with previous studies showing that GFAP is highly predictive of mortality and adverse outcomes after severe TBI in U.S. adults. We further found an association between CT abnormalities on admission and neurological deficits at admission and discharge. Neurological examinations on admission and discharge, and daily examinations for hospitalized patients, are routine clinical care at Mulago Hospital and are frequently used to determine the need for repeat CT. 35 Taken together, our findings present a compelling case for additional studies involving larger cohorts of children with mild-to-moderate and severe TBI that incorporate clinical assessments, neurological examinations, and TBI biomarkers to develop clinical care protocols optimized for CT utilization as needed, thereby informing resource allocation in the hospital.
Limitations
This study had limitations. The conclusions are based on pilot study data from a single hospital in Uganda’s urban capital, which limits generalizability. However, MNRH is Uganda’s largest government-run public health facility, to which complex cases are referred from across the country. Patterns associated with pediatric TBI in Kampala mirror trends observed in Uganda and other LMICs.36,37 Nevertheless, the study should be expanded to include additional sites with larger sample sizes to confirm these results. Although this study focused on GFAP and UCH-L1, other promising biomarkers, such as tau, p-tau, and neurofilaments, should be evaluated for compatibility with the i-STAT analyzer. Notably, the i-STAT handheld analyzer is an accessible tool with a history of value-added in LMIC settings. 38 Only a single blood sample was collected at admission, restricting our ability to assess the persistence of biomarker elevation over time. Future prospective studies should include multiple sampling time points during hospitalization and post-discharge, coupled with comprehensive clinical assessments, to better characterize the temporal dynamics of these biomarkers and their association with clinical outcomes.
Conclusions
Our data suggest that blood-based GFAP and UCH-L1 levels, as assessed by the i-STAT Alinity analyzer, were significantly elevated in Ugandan children with TBI compared with uninjured controls, with GFAP showing greater sensitivity and superior diagnostic utility across severity levels. These findings align with studies from high-income countries, supporting the potential of GFAP and UCH-L1 as diagnostic biomarkers for identifying intracranial lesions in pediatric TBI. GFAP additionally demonstrated modest prognostic value for predicting neurological deficits at discharge and 2-week follow-up.
Given the associations between elevated biomarker levels, CT abnormalities, and early neurological outcomes, our findings underscore the importance of integrating rapid point-of-care biomarker testing with comprehensive clinical evaluation, including neurocognitive, mental health, and sensorimotor assessments, in low-resource settings. Such integration may support early triage, guide CT utilization, and optimize care delivery in settings with limited access to neuroimaging. Larger multisite studies with longitudinal follow-up are needed to validate biomarker-informed protocols for pediatric TBI triage and prognosis in LMICs.
Transparency, Rigor, and Reproducibility Statement
This study was conducted and reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for observational cohort studies. The study protocol, eligibility criteria, primary and secondary outcomes, and data collection procedures were prespecified and approved by the institutional review boards of Makerere University College of Health Sciences, Indiana University, and the Uganda National Council for Science and Technology. Written informed consent (and assent where appropriate) was obtained from all participants or their guardians.
All clinical and neurological assessments were conducted in accordance with standardized protocols at MNRH. Plasma biomarker levels were measured using the Abbott i-STAT® Alinity TBI analyzer according to manufacturer instructions. Blood samples were collected within 48 h of injury (median: 22 h), reflecting real-world delays in care often encountered in low-resource settings and aligning with GFAP’s peak detection window.
The sample size was determined using power calculations to detect differences in biomarker levels between groups. The 95th percentile of biomarker values from uninjured community children was used as a threshold for elevation, consistent with FDA-approved validation studies. All analyses were conducted using validated software (Stata v18.0 and GraphPad Prism v10.0). Outliers and missing data were handled in accordance with the prespecified rules detailed in the Methods section.
Authors’ Contributions
D.D. and K.K. conceptualized and designed the study, drafted the initial article, and critically reviewed and revised it. A.M. and D.E. validated the data collection tools, collected clinical data, carried out initial analysis, and critically reviewed and revised the article. E.K. and C.L. were responsible for patient care, collecting clinical data and initial data analysis, and critically reviewing the article. C.S. was responsible for sample processing and laboratory assessments. V.M. critically reviewed the study data and revised the article. G.K. and M.G. developed the study database, supervised data collection, conducted quality checks, and critically reviewed the data and final article. S.P.-B. contributed to the interpretation of the data and critically reviewed and revised the article for accuracy and content. R.N., R.I., and P.B. coordinated and supervised clinical and neurological assessments, data quality checks, analysis, and interpretation, and critically reviewed and revised the article for important intellectual content. All authors approved the final article as submitted and agree to be accountable for all aspects of the work.
Footnotes
Acknowledgments
The authors thank the children and their parents who participated in this study, as well as the study team for their dedicated efforts in treating the children and for collecting the data.
Author Disclosure Statement
Abbott provided the i-STAT Alinity analyzer and testing cartridges for research use in this study. The company played no role in the execution of the study, data analysis, or article drafting. The authors have no conflicts of interest relevant to this article to disclose.
Funding Information
This work was supported by the National Institutes of Health, National Institute of Neurological Disorders and Stroke (R21NS129234), and Fogarty International Center Global Health Program for Fellows and Scholars (D43TW009345). Abbott provided the i-STAT Alinity analyzers and TBI plasma cartridges, and reviewed the article prior to submission. The funders played no role in the study design, data collection, analysis, interpretation, article preparation, or decision to submit the article for publication.
Data Availability Statement
De-identified individual participant data that support the findings of this study are available from the corresponding author upon reasonable request. Data sharing is subject to institutional approvals and ethical guidelines governing research conducted in Uganda.
Supplemental Material
Abbreviations
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
