Model predicting survival/exitus after traumatic brain injury: biomarker S100B 24h.

The enigma of Traumatic Brain Injury (TBI), reflected in recent scientific literature, is its uncertain consequences, variability of the final prognosis with apparently similar TBI, necessity for peripheral biomarkers, and more specific predictive models. To study the relationship between serum S100B and survival in TBI patients in various serious situations; the S100B level in patients without traumatic pathology or associated tumour, subjected to stressful situations such as neurological intensive care unit (NICU) stay; the possible overestimation caused by extracerebral liberation in TBI patients and associated polytraumatism; the predictive cutoffs to determine the most sensitive and specific chronology; and achieve a predictive prognostic model. Patients admitted to the NICU within 6 hours after TBI were selected. We measured: a) clinical: exitus yes/no; age and gender, traumatic mechanism, polytraumatism yes/no, GCS score, unconsciousness duration, amnesia duration, neurological focality, and surgical interventions; b) radiological: CT scan for radiological lesions; c) biochemical: serum SB100B at 6, 24, 48 and 72 hours after TBI and drug abuse detected in the urine; d) GOS on hospital discharge. N: 149 TBI patients, independent of polytraumatism, mean serum S100B at 6, 24, 48, and 72 hours: 2.1, 1.3, 1.2, and 0.6 microg/L, respectively; N: 124 without associated polytraumatism, S100B at 6, 24, 48, and 72 hours: 2.0, 1.4, 1.3, and 0.6 microg/L; N: 50 control I S100B 24 hours: 0.17 microg/L (0.04 - 0.56) and 25 healthy subjects S100B 0.057 microg/L (0.02-0.094). Significantly higher S100B levels are observed on exitus, with excellent TBI prognosis and evolution performance. Hospital stay in the NICU produces significant increases in S100B compared to healthy subjects, without invalidating it as a biomarker. Polytraumatism associated to TBI does not significantly alter S100B levels. S100B at 24 hours > or = 0.90 microg/L appears to predict unfavourable TBI evolution with a NPV: 94.2% and PPV: 54.9%. We propose a predictive model when we associate S100B at 24 hours with amnesia duration over 30 minutes with a NPV of 85.5% and a PPV of 83.3%.