Abstract
BACKGROUND: Brain injuries are frequent causes of intubation and mechanical ventilation. The aim of this study was to investigate the accuracy and sensitivity of clinical parameters in predicting successful extubation in patients with acute brain injury. METHODS: Six hundred and forty-four patients assisted at a high-complexity hospital were recruited. Patients were divided as for successful or failed extubation. The VISAGE score, maximum inspiratory and expiratory pressures, peak cough flow, and airway occlusion pressure at 0.1 s were used as predictors. Logistic regression analyses using ROC-curve identified values of accuracy and sensitivity. The Hosmer-Lemeshow test and the stepwise method calibrated the statistical model. RESULTS: VISAGE score (odds ratio of 1.975), maximum inspiratory pressure (odds ratio of 1.024), and peak cough flow (odds ratio of 0.981) are factors consistent in distinguishing success from failure extubation. The ROC curve presented an accuracy of 79.7% and a sensitivity of 95.8%. CONCLUSIONS: VISAGE score, maximum inspiratory pressure and peak cough flow showed good accuracy and sensitivity in predicting successful extubation in patients with acute brain injury. The greater impact of VISAGE score indicates that patients' neurological profile should be considered in association with ventilatory parameters in the decision of extubation.