Predictors of successful extubation in neurosurgical patients

A respiratory therapist-driven weaning protocol incorporating daily screens, spontaneous breathing trials (SBT), and prompts to caregivers has been associated with superior outcomes in mechanically ventilated medical patients. To determine the effectiveness of this approach in neurosurgical (NSY) patients, we conducted a randomized controlled trial involving 100 patients over a 14-mo period. All had daily screens of weaning parameters. If these were passed, a 2-h SBT was performed in the Intervention group. Study physicians communicated positive SBT results, and the decision to extubate was made by the primary NSY team. Patients in the Intervention (n = 49) and Control (n = 51) groups had similar demographic characteristics, illness severity, and neurologic injuries. Among all patients, 87 (45 in the Control and 42 in the Intervention group) passed at least one daily screen. Forty (82%) patients in the Intervention group passed SBT, but a median of 2 d passed before attempted extubation, primarily because of concerns about the patient's sensorium (84%). Of 167 successful SBT, 126 (75%) did not lead to attempted extubation on the same day. The median time of mechanical ventilation was 6 d in both study groups, and there were no differences in outcomes. Overall complications included death (36%), reintubation (16%), and pneumonia (9%). Tracheostomies were created in 29% of patients. Multivariate analysis showed that Glasgow Coma Scale (GCS) score (p < 0.0001) and partial pressure of arterial oxygen/fraction of inspired oxygen ratio (p < 0.0001) were associated with extubation success. The odds of successful extubation increased by 39% with each GCS score increment. A GCS score > or = 8 at extubation was associated with success in 75% of cases, versus 33% for a GCS score < 8 (p < 0.0001). Implementation of a weaning protocol based on traditional respiratory physiologic parameters had practical limitations in NSY patients, owing to concerns about neurologic impairment. Whether protocols combining respiratory parameters with neurologic measures lead to superior outcomes in this population requires further investigation.