RESUMO
OBJECTIVES: This research aims to study the prognostic role of serum S100 as a predictor of mortality in vascular and traumatic brain injuries. METHODS: This prospective cohort study involved 219 patients. In the blood serum, neuron-specific markers (S100, NSE) and glucose, acid-base state and gas composition of arterial blood were obtained at admission, on the 3rd, 5th and 7th days of patients' stay in the intensive care unit. RESULTS: The most significant risk factor for an unfavorable outcome is the marker S100 with a cut-off point of 0.2 mcg/l. The analysis results indicate a statistically significant direct relationship between S100 > 0.2 mcg/l and NSE ≥ 18.9 ng/ml compared to other variables, while the chance ratio (OR) is 11.9 (95%CI:3.2927-1.6693;). With blood sugar increase above 7.4 mmol/l, the OR is 3.82 (95% CI: 2.1289-0.5539;); with a Glasgow scale below 13 points, the OR is 3.69 (95% CI: 2.1316-0.4819;); with an increase in pCO2 < 43.5 mm Hg, the OR was 3.15 (95% CI: 1.8916- 0.4062;). The obtained model certainty measure according to pseudo R2 Nagelkerke criterion is 263.5, showing the excellent quality of the mathematical model's predictive ability. The developed prognostic model, including the dependent variable S100 and independent variables as predictors of a poor outcome of NSE, pCO2, GCS and Hb, reached a cut-off point of 84.51%, AUC - 0.88 with high levels of sensitivity and specificity: 91.89% and 64.14%, respectively. NOVELTY: This model can be used to predict the outcome in patients with acute cerebral pathology.