Ability of S100B to predict severity and cranial CT results in children with TBI.

OBJECTIVES: To evaluate the ability of S100B to predict severity of TBI and abnormal cranial CT results for children with TBI.

METHODS: This is a secondary analysis of a previously established cohort of consecutive patients presenting to the emergency department with TBI limited to children <19 years of age, who arrived within 6 hours of injury, received a cranial CT scan and consented to blood drawn for S100B.

RESULTS: A total of 109 children were included in this cohort. The mean S100B levels were higher in children with moderate/severe TBI as compared to children with mild TBI based GCS score (0.281 µg L(-1), 95%CI = 0.101, 0.461 vs 0.053, 95%CI = 0.010, 0.095). S100B levels were significantly elevated in children following TBI with abnormal cranial CT as compared to children with a normal cranial CT (0.210 µg L(-1), SD = 0.313 vs 0.036 µg L(-1), SD = 0.046, p = 0.03). Area under the curve for S100B was also significant (0.72, 95%CI = 0.58, 0.86) for prediction of abnormal cranial CT for children with TBI. S100B did not predict abnormal cranial CT for children following TBI with a GCS of 15 (AUC = 0.53, 95%CI = 0.36, 0.71).

CONCLUSIONS: For children following TBI, S100B appears to predict severity of TBI; however, it may not be clinically useful as an independent screening test to select children with mild TBI who need a cranial CT.