Background: S-1OOB, a protein of astroglial cells, is described as a marker for neuronal damage. Reliable outcome prediction from severe head injury is still unresolved. Clinical scores such as the Glasgow Coma Scale score (GCS) and diagnostic scores such as the Marshall Computed Tomographic Classification are well established and investigated, but there are still some concerns about these tools. The aim of this study was to investigate the predictive value of the initial serum level of S-100B compared with the predictive value of the GCS score and the Marshall Computed Tomographic Classification to outcome after severe head injury.
Methods: Forty-four patients with severe head injury (GCS score < 9) were included. Blood samples were drawn within 1 to 6 hours of injury. After a period of 11 months, their outcome was correlated by using the Glasgow Outcome Scale. Patients with an S-100B serum level above 2 microg/L, a GCS score between 3 and 5, and a computed tomographic scan in the categories 4 to 6 are predicted to have an unfavorable outcome. The predictive values of these tools were calculated according to these definitions.
Results: The protein S-100B had with 17% the lowest total misclassification rate. When compared with the GCS score and Marshall Computed Tomographic Classification the S-100B serum level calculated on admission had the highest positive predictive value (87%) and negative predictive value (77%).
Conclusion: The serum level of S-100B calculated within 1 to 6 hours of a severe head injury is a useful additional outcome predictor.