Investigators from The Netherlands have validated the predictive worth of two biochemical markers for traumatic brain injury (TBI). “We showed that glial fibrillary acid protein (GFAP) and S100B levels in serum are adjuncts to the assessment of brain damage following TBI and, when combined with clinical variables, may enhance our ability to provide accurate prognoses for patients,” says the study's lead researcher Pieter Vos, of the Radboud University Nijmegen Medical Centre.

In patients with TBI, treatment decisions may be guided by neurological prognoses. Current prognostic models for TBI, which include various patient characteristics and brain damage indices, do not exhibit sufficiently high sensitivity or specificity to accurately predict outcomes in individual patients. “It is broadly felt among experts in the field that the need exists to enlarge knowledge on biochemical markers of injury,” says Vos.

In a previous study, Vos and his colleagues showed that serum levels of GFAP (≥1.5 μg/l) and S100B (≥1.13 μg/l) in patients with severe TBI could predict death and unfavorable outcomes with good sensitivity but poor specificity. The new investigation aimed to validate the prognostic capabilities of these serum biomarkers in people with either moderate (Glasgow Coma Scale [GCS] 9–12) or severe (GCS 3–8) TBI.

The researchers recorded various patient characteristics and collected serum samples on hospital admission. GFAP and S100B levels were subsequently measured, and patient outcomes were then determined at 6 months after injury.

Among the 79 patients who met the study's inclusion criteria, those who died within the follow-up period exhibited 33.4-fold and 2.1-fold higher median levels of serum GFAP and S100B, respectively, than those who were still alive at 6 months. Median serum levels of GFAP and S100B were also increased (19.8-fold and 2.1-fold, respectively) in patients exhibiting unfavorable outcomes (Glasgow Outcome Scale Extended [GOSE] score 1–4) compared with individuals showing favorable outcomes (GOSE score 2–8). Logistical regression analysis showed that models including the serum biomarkers (at cut-off values determined in the previous study), mass lesion and pupillary reaction most accurately predicted patient outcomes.

Vos and his colleagues hypothesize that in the case of TBI, cut-off values for GFAP and S100B can be determined that are associated with a small false-positive rate of death or unfavorable outcomes. “To achieve this,” says Vos, “we would need to conduct a larger, multicenter study, involving hundreds rather than tens of patients with this serious and heterogeneous disorder.”