Background: Patients with brain injury are at high risk of extubation failure.
Methods: We conducted a prospective observational cohort study in four intensive care units of three university hospitals. The aim of the study was to create a score that could predict extubation success in patients with brain injury.
Results: A total of 437 consecutive patients with brain injury were included, and 338 patients (77.3%) displayed successful extubation. In the multivariate analysis, four features were associated with success the day of extubation: age less than 40 yr, visual pursuit, swallowing attempts, and a Glasgow coma score greater than 10. In the score, each item counted as one. A score of 3 or greater was associated with 90% extubation success. The area under the receiver-operator curve was 0.75 (95% CI, 0.69 to 0.81). After internal validation by bootstrap, the area under the receiver-operator curve was 0.73 (95% CI, 0.68 to 0.79). Extubation success was significantly associated with shorter duration of mechanical ventilation (11 [95% CI, 5 to 17 days] vs. 22 days [95% CI, 13 to 29 days]; P < 0.0001), shorter intensive care unit length of stay (15 [95% CI, 9 to 23 days] vs. 27 days [95% CI, 21 to 36 days]; P < 0.0001), and lower in-intensive care unit mortality (4 [1.2%] vs. 11 [11.1%]; P < 0.0001).
Conclusions: Our score exploring both airway functions and neurologic status may increase the probability of successful extubation in patients with severe brain injury.