Predicting Outcome in Aged and Severely Ill Patients with Prolonged Respiratory Failure

Objectives: Consultations for tracheotomy are often sought on aged and severely ill patients with respiratory insufficiency. This patient population has high short-term mortality rates and is difficult to stratify on the basis of expected outcome. We examined whether APACHE III (Acute Physiology and Chronic Health Evaluation III) scores or neurologic status assessment (NSA) scores in sedated individuals are predictive of outcome. Methods: We performed a retrospective study examining aged patients who underwent tracheotomy for respiratory insufficiency and prolonged intubation. The APACHE III scores (n = 30) and NSA (based on a modified Glasgow Coma Scale) scores (n = 37) were calculated before tracheotomy. All patients were mildly sedated. Using APACHE III and NSA scores as predictor variables and using death and ability to be weaned from the ventilator as outcome variables, we performed a Kaplan-Meier survival analysis and a Cox proportional hazard regression. Results: The APACHE III was not significantly predictive of either outcome. Higher NSA scores were associated with increased survival rates (log rank = 19.7, p < .0001) and longer median survival (88 days versus 16 days for lower scorers). Higher NSA scores also predicted a higher rate of ventilator independence. Conclusions: Neurologic function in sedated patients (and not APACHE III scores) can be used to stratify aged individuals with respiratory insufficiency on the basis of expected outcome.