Objectives: Several studies have established the relevance of S- 100 in blood as a marker of brain damage after traumatic brain injury. However, a more specific marker is required and glial fibrillary acidic protein (GFAP) is considered to be a good candidate. Methods: In order to assess the increase of GFAP in serum (s- GFAP) after a severe traumatic brain injury (TBI) we collected daily serum samples from 59 patients with severe TBI starting on the day of the trauma. S- GFAP was measured using a sandwich ELISA. The Glasgow outcome scale (GOS) assessed outcome after 1 year. Results: All but one patient had maximal s-GFAP values above the laboratory reference value (median increased 10- fold). The highest detected levels were seen during the first days after TBI and then decreased gradually. Patients with unfavourable outcome had significantly (p < 0.001) higher maximal s-GFAP values in the acute phase compared with patients with favourable outcome. All patients (n = 5) with s-GFAP > 15.04 μ g /L died (reference level < 0.15 μ g/L). We found no significant difference in the maximal s-GFAP levels of patients with isolated brain injury in comparison with patients with multiple traumas. Conclusion: Serum-GFAP is increased during the first days after a severe traumatic brain injury and re-lated to clinical outcome. Objectives: Several studies have established the relevance of S- 100 in blood as a marker of brain damage after traumatic brain injury. However, a more specific marker is required and glial fibrillary acidic protein (GFAP) is considered to be a good candidate. Methods : In order to assess the increase of GFAP in serum (s-GFAP) after a severe traumatic brain injury (TBI) we collected daily serum samples from 59 patients with severe TBI starting on the day of the trauma. Results: All but one patient had maximal s-GFAP values ​​above the laboratory reference value (median increased 10-fold). The highest detected levels were seen during the First days after TBI and then decreased gradually. Patients with unfavorable outcome had significantly (p <0.001) higher maximal s-GFAP values ​​in the acute phase compared with patients with favourable outcome. All patients (n = 5) with s-GFAP> 1 5.04 μg / L died (reference level <0.15 μg / L). We found no significant difference in the maximal s-GFAP levels of patients with isolated brain injury in comparison with patients with multiple traumas. Conclusion: Serum-GFAP is increased during the first days after a severe traumatic brain injury and re-lated to clinical outcome.