ABSTRACT
The detection of brain death (BD) plays a fundamental role in the management of hospital donation. Delayed diagnosis of BD is the main cause of donor loss. A tool for monitoring and prognosis in the neurocritical patient is essential to meet these objectives. The most used prognostic scores in intensive care units are Simplified Acute Physiology Score (SAPS) II and Acute Physiology and Chronic Health Evaluation (APACHE) II. A predictive model of good performance (ModSPN) in predicting BD in neurocritical patients with Glasgow Coma Scale score < 8 was published in 2014. With the objective of analyzing the predictive capacity of ModSPN and comparing it with SAPS II and APACHE II, 2307 patients admitted to the neurocritical patient monitoring (SPN) program of the INDT were analyzed. The predictive capacity for death and brain death of SAPS II, APACHE II, and ModSPN was compared using receiver operating characteristic curves. The area under the curve showed a better APACHE II performance for the prediction of death and the ModSPN being a better predictor of the probability of dying in BD. Therefore, for the prediction of death in the neurocritical patient, APACHE II was superior, but for the prediction of encephalic death, the ModSPN presented the best predictive power for all causes of brain injury.
