Utility of Serum Biomarkers in the Diagnosis and Stratification of Mild Traumatic Brain Injury

<jats:title>Abstract</jats:title><jats:sec><jats:title>Objective</jats:title><jats:p>The objective was to compare test characteristics of a single serum concentration of glial fibrillary acidic protein (<jats:styled-content style="fixed-case">GFAP</jats:styled-content>), S‐100β, and ubiquitin carboxyl terminal hydrolase L1 (<jats:styled-content style="fixed-case">UCH</jats:styled-content>‐L1), obtained within 6 hours of head injury, to diagnose mild traumatic brain injury (<jats:styled-content style="fixed-case">mTBI</jats:styled-content>) in head‐injured subjects.</jats:p></jats:sec><jats:sec><jats:title>Methods</jats:title><jats:p>Adults aged 18 to 80 years who presented to one of seven <jats:styled-content style="fixed-case">ED</jats:styled-content>s with a blunt closed head injury underwent head <jats:styled-content style="fixed-case">CT</jats:styled-content> within 4 hours of injury and had blood drawn for biomarker analysis within 6 hours of injury were eligible. Subjects were considered to have <jats:styled-content style="fixed-case">mTBI</jats:styled-content> if they had an initial Glasgow Coma Scale (<jats:styled-content style="fixed-case">GCS</jats:styled-content>) > 13 and met one or more of the following criteria: loss of consciousness (<jats:styled-content style="fixed-case">LOC</jats:styled-content>), posttraumatic amnesia, or confusion. Subjects with <jats:styled-content style="fixed-case">mTBI</jats:styled-content> and an abnormal head computed tomography (<jats:styled-content style="fixed-case">CT</jats:styled-content>) scan were categorized as complicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content>; those with a normal head <jats:styled-content style="fixed-case">CT</jats:styled-content> were categorized as uncomplicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content>; and subjects with a <jats:styled-content style="fixed-case">GCS</jats:styled-content> = 15, no <jats:styled-content style="fixed-case">LOC</jats:styled-content>, no posttraumatic amnesia, and no confusion were considered to not have a <jats:styled-content style="fixed-case">mTBI</jats:styled-content>. Biomarker concentration measurements for <jats:styled-content style="fixed-case">GFAP</jats:styled-content> and <jats:styled-content style="fixed-case">UCH</jats:styled-content>‐L1 were performed using an enzyme‐linked immunosorbent assay. S‐100β concentration was determined using an electrochemiluminescence immunoassay. Median biomarker concentration for each group was compared using the Kruskal‐Wallis test. Logistic regression was used to determine area under the receiver operating curve (<jats:styled-content style="fixed-case">AUC</jats:styled-content>) for each of the three biomarkers. Sensitivity, specificity, negative predictive value (<jats:styled-content style="fixed-case">NPV</jats:styled-content>), positive predictive value (<jats:styled-content style="fixed-case">PPV</jats:styled-content>), and negative and positive likelihood ratios (<jats:styled-content style="fixed-case">LR</jats:styled-content>s) for the three biomarkers to differentiate between complicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content>, uncomplicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content>, and no <jats:styled-content style="fixed-case">mTBI</jats:styled-content> were calculated.</jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p>A total of 247 subjects were enrolled and had adequate clinical and biomarker information for analysis. A total of 188 met criteria for <jats:styled-content style="fixed-case">mTBI</jats:styled-content>, with 34 (18.1%) having an acute abnormality on <jats:styled-content style="fixed-case">CT</jats:styled-content> (complicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content>). The mean (±<jats:styled-content style="fixed-case">SD</jats:styled-content>) age of the study population was 45.8 (±17.3) years, and 59.9% were male. Median serum concentrations for all biomarkers were significantly different between groups, lowest in the no <jats:styled-content style="fixed-case">mTBI</jats:styled-content> group, and progressively increasing in the uncomplicated and complicated <jats:styled-content style="fixed-case">mTBI</jats:styled-content> groups (p < 0.0001). All three biomarkers were significant classifiers of <jats:styled-content style="fixed-case">mTBI</jats:styled-content> versus no <jats:styled-content style="fixed-case">mTBI</jats:styled-content>, with the following <jats:styled-content style="fixed-case">AUC</jats:styled-content>s: <jats:styled-content style="fixed-case">GFAP</jats:styled-content>, 0.70; S‐100β, 0.69; and <jats:styled-content style="fixed-case">UCH</jats:styled-content>‐L1, 0.65 (p = 0.17). Sensitivity for <jats:styled-content style="fixed-case">mTBI</jats:styled-content> was highest for S‐100β (96.5%). <jats:styled-content style="fixed-case">NPV</jats:styled-content>s ranged fr ...