Abstract
Blood-based biomarkers have shown utility in discriminating adult patients with and without traumatic brain injury (TBI). Biomarker levels vary with severity, time since injury, and imaging findings (computed tomography [CT] and magnetic resonance imaging [MRI]). However, the association of specific biomarkers with clinical and imaging findings in children across the age spectrum and with different injury severities and presentations is not fully understood. To better characterize biomarker and clinical associations, we studied pediatric blood biomarker patterns within the Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI) study (ClinicalTrials.gov #NCT02119182). TRACK-TBI is an observational multisite study that prospectively enrolled TBI patients across the lifespan and injury spectrum, from 2014 through 2018. Two centers enrolled children (ages 0–17 years). All participants underwent a head CT or MRI at the discretion of treating clinicians as part of acute clinical care, and the majority underwent a study MRI 2 weeks after injury. For this TRACK-TBI pediatric cohort, blood sampling was optional with informed consent provided by guardians. For the purposes of this study, only pediatric subjects with study-specific neuroimaging and plasma biomarkers were included. Plasma biomarkers assessed included glial fibrillary acidic protein (GFAP), ubiquitin carboxy-terminal hydrolase L1 (UCH-L1), S100 calcium-binding protein B (S100B), neuron-specific enolase (NSE), and high-sensitivity C-reactive protein (hsCRP). Biomarker levels were compared between participants with and without radiographical traumatic intracranial findings (day-of-injury CT-positive versus CT-negative and 2-week study MRI-positive versus MRI-negative) and between participants with different clinical indices of injury. Of 158 pediatric participants, 75 consented to blood collection and underwent biomarker analysis. Of these, 65 had acute CT scans (within 24 h of injury) and 43 had standardized study-related MRI scans at 2 weeks (±3 days) from injury; thus, 70 subjects had biomarker levels and study-reviewed CTs and/or MRIs and are included in the present analysis. Univariate analyses showed significant differences in plasma GFAP, UCHL1, S100B, and hsCRP with respect to time since injury and CT findings. GFAP levels differed by Glasgow Coma Scale (GCS) severities and MRI findings; S100B levels differed by GCS and loss of consciousness. Multivariable linear regression models confirmed a significant association of GFAP, UCH-L1, and S100B levels with CT-positive findings and a significant association of GFAP with 2-week MRI-positive findings. Among the biomarkers evaluated in our study, GFAP was the strongest predictor of imaging findings using receiver operating characteristic analysis. Significant age effects were seen for S100B and NSE, with higher values in younger children. Multivariable associations were observed for specific day-of-injury plasma biomarkers and radiographical traumatic intracranial findings on CT and MRI in children with acute TBI. These biomarkers have potential utility to aid in TBI screening by helping to triage which patients are likely to have intracranial findings on neuroimaging.