Extubation Failure in Brain-injured Patients Risk Factors and Development of a Prediction Score in a Preliminary Prospective Cohort Study

Background: The decision to extubate brain-injured patients with residual impaired consciousness holds a high degree of uncertainty of success. The authors developed a pragmatic clinical score predictive of extubation failure in brain-injured patients. Methods: One hundred and forty brain-injured patients were prospectively included after the first spontaneous breathing trial success. Assessment of multiparametric hemodynamic, respiratory, and neurologic functions was performed just before extubation. Extubation failure was defined as the need for ventilatory support during intensive care unit stay. Extubation failure within 48 h was also analyzed. Neurologic outcomes were recorded at 6 months. Results: Extubation failure occurred in 43 (31%) patients with 31 (24%) within 48 h. Predictors of extubation failure consisted of upper-airway functions (cough, gag reflex, and deglutition) and neurologic status (Coma Recovery Scale-Revised visual subscale). From the odds ratios, a four-item predictive score was developed (area under the curve, 0.85; 95% CI, 0.77 to 0.92) and internally validated by bootstrap. Cutoff was determined with sensitivity of 92%, specificity of 50%, positive predictive value of 82%, and negative predictive value of 70% for extubation failure. Failure before and beyond 48 h shared similar risk factors. Low consciousness level patients were extubated with 85% probability of success providing the presence of at least two operating airway functions. Conclusions: A simplified clinical pragmatic score assessing cough, deglutition, gag reflex, and neurologic status was developed in a preliminary prospective cohort of brain-injured patients and was internally validated (bootstrapping). Extubation appears possible, providing functioning upper airways and irrespective of neurologic status. Clinical practice generalizability urgently needs external validation.