J Neurosurg 121:1232–1238, 2014 1232 J Neurosurg / Volume 121 / November 2014 ©AANS, 2014 E ach year in the United States alone, approximately 1.8 million people experience a traumatic brain in- jury (TBI) event that requires hospitalization. Of these TBIs, about 75% are mild traumatic brain injuries (mTBIs), such as sports-related concussions, motor ve- hicle collisions, and falls. 4 After the initial blow to the head, a secondary injury in the brain may persist for longer periods of time. 3,5,15,19,30 This secondary brain injury consists of excitotoxicity, oxidant injury, mitochondrial dysfunction, infammation, and subsequent cell death that begins almost immediately after the primary injury. 6,10,13,32 With respect to mTBI, to predict neurological outcome after mild, moderate, and severe TBI, the detection of fuid-based neural biomark - ers such as neuron-specifc enolase (NSE), tau, ubiquitin C-terminal hydrolase, spectrin, glial fbrillary acidic pro- tein (GFAP), and S100B in serum have been shown to correlate with a number of outcomes. 7,8,11,12,16–18,21,23,24,27,31,34 Another protein, neuroflament, is detectable in se- Detection of neuroflament-H in serum as a diagnostic tool to predict injury severity in patients who have suffered mild traumatic brain injury Clinical article Joshua W. Gatson, Ph.D., 1 Jennifer Barillas, B.s., 1 linDa s. hynan, Ph.D., 2 ramon Diaz-arrastia, m.D., Ph.D., 3 steven e. Wolf , m.D., 1 anD JosePh P. minei, m.D., m.B.a. 1 Departments of 1 Surgery and 2 Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, Texas; and 3 Center for Neuroscience and Regenerative Medicine, Uniformed Services University of the Health Sciences, Rockville, Maryland Object. In previous studies of traumatic brain injury (TBI), neural biomarkers of injury correlate with injury severity and predict neurological outcome. The object of this paper was to characterize neuroflament-H (NFL-H) as a predictor of injury severity in patients who have suffered mild TBI (mTBI). Thus, the authors hypothesized that phosphorylated NFL-H (pNFL-H) levels are higher in mTBI patients than in healthy controls and identify which subjects experienced a more severe injury such as skull fractures, intracranial hemorrhaging, and/or contusions as detected by CT scans. Methods. In this prospective clinical study, blood (8 ml) was collected from subjects (n = 34) suffering from mTBI (as defned by the American Congress of Rehabilitation and Glasgow Coma Scale scores between 13 and 15) at Parkland Hospital, Dallas, Texas, on Days 1 and 3 after injury). Additional clinical fndings from the CT scans were also used to categorize the TBI patients into those with and those without clinical fndings on the scans (CT+ and CT- groups, respectively). The serum levels of pNFL-H were measured using the enzyme-linked immunosorbent assay. Results. Compared with healthy controls, the mTBI patients exhibited a signifcant increase in the serum levels of pNFL-H on Days 1 (p = 0.00001) and 3 (p = 0.0001) after TBI. An inverse correlation was observed between pNFL-H serum levels and Glasgow Coma Scale scores, which was signifcant. Additionally, using receiver operating characteristic curve analysis to compare the mTBI cases with controls to determine sensitivity and specifcity, an area under the curve of 100% was achieved for both (p = 0.0001 for both). pNFL-H serum levels were only signifcantly higher on Day 1 in mTBI patients in the CT+ group (p < 0.008) compared with the CT- group. The area under the curve (82.5%) for the CT+ group versus the CT- group was signifcant (p = 0.021) with a sensitivity of 87.5% and a specifcity of 70%, using a cutoff of 1071 pg/ml of pNFL-H in serum. Conclusions. This study describes the serum profle of pNFL-H in patients suffering from mTBI with and with- out CT fndings on Days 1 and 3 after injury. These results suggest that detection of pNFL-H may be useful in deter- mining which individuals require CT imaging to assess the severity of their injury. (http://thejns.org/doi/abs/10.3171/2014.7.JNS132474) Key WorDs • mild traumatic brain injury • neuroflament-H • serum Abbreviations used in this paper: AUC = area under the curve; ELISA = enzyme-linked immunosorbent assay; GCS = Glasgow Coma Scale; GFAP = glial fibrillary acidic protein; MAMBA = Mild and Moderate TBI Biomarker (study); mTBI = mild TBI; NFL-H = neurofilament-H; NSE = neuron-specific enolase; pNFL- H = phosphorylated NFL-H; ROC = receiver operating characteris- tic; TBI = traumatic brain injury.