Severe traumatic brain injury in children elevates glial fibrillary acidic protein in cerebrospinal fluid and serum*

OBJECTIVES: 1) To determine the levels of glial fibrillary acidic protein (GFAP) in both cerebrospinal fluid and serum; 2) to determine whether serum GFAP levels correlate with functional outcome; and 3) to determine whether therapeutic hypothermia, as compared with normothermia, alters serum GFAP levels in children with severe traumatic brain injury (TBI). DESIGN: Laboratory-based analyses; postrandomized, controlled trial. SETTING: Four Canadian pediatric intensive care units and a university-affiliated laboratory. PATIENTS: Twenty-seven children, aged 2-17 yrs, with severe TBI (Glasgow Coma Scale score of ≤ 8). INTERVENTIONS: Hypothermia therapy (32.5°C) for 24 hrs with cooling started within 8 hrs of injury and rewarming at a rate of 0.5°C every 2 hrs or normothermia (37.0°C). MEASUREMENTS AND MAIN RESULTS: GFAP was measured in cerebrospinal fluid and serum, using enzyme-linked immunosorbent assay. Levels of GFAP were maximal on day 1 post-TBI, with cerebrospinal fluid GFAP (15.5 ± 6.1 ng/mL) 25-fold higher than serum GFAP (0.6 ± 0.2 ng/mL). Cerebrospinal fluid GFAP normalized by day 7, whereas serum GFAP decreased gradually to reach a steady state by day 10. Serum GFAP measured on day 1 correlated with Pediatric Cerebral Performance Category scores determined at 6 months post-TBI (ρ = 0.527; p = .008) but failed to correlate with the injury scoring on admission, physiologic variables, or indices of injury measured on computerized tomography imaging. The areas under the receiver operating characteristic curves for pediatric intensive care unit day 1 serum GFAP in determining good outcome were 0.80 (pediatric cerebral performance category, 1-2; normal-mild disability) and 0.91 (pediatric cerebral performance category, 1-3; normal-moderate disability). For a serum GFAP cutoff level of 0.6 ng/mL, sensitivity and specificity were 88% to 90% and 43% to 71%, respectively. Serum GFAP levels were similar among children randomized to either therapeutic hypothermia or normothermia. CONCLUSIONS: GFAP was markedly elevated in cerebrospinal fluid and serum in children after severe TBI and serum GFAP measured on pediatric intensive care unit day 1 correlated with functional outcome at 6 months. Hypothermia therapy did not alter serum GFAP levels compared with normothermia after severe TBI in children. Serum GFAP concentration, together with other biomarkers, may have prognostic value after TBI in children.