Risk Factors of Extubation Failure and Analysis of Cuff Leak Test as a Predictor for Postextubation Stridor

Background: Extubation failure was associated with poor prognosis and high hospital mortality. Cuff leak test (CLT) has been proposed as a relatively simple method for detecting laryngeal obstruction that predispose toward postextubation stridor (PES) and reintubation. We examined the risk factors of extubation failure and evaluated the usefulness and limitation of CLT for predicting PES and reintubation. Methods: Thirty-four consecutive patients intubated more than 24 hours were examined. The subjects were evaluated daily for extubation readiness, and CLT was performed prior to extubation. Several parameters in the extubation success and failure group were compared. The accuracy and limitation of CLT were evaluated after choosing the thresholds values of the cuff leak volume (CLV) and percentage (CLP). Results: Of the 34 patients studied, 6 (17.6%) developed extubation failure and 3 (8.8%) were accompanied by PES. The patients who had extubation failure were more likely to have a longer duration of intubation and more severe illness. The patients who developed PES had a smaller cuff leak than the others: according to the CLV (22.5±23.8 vs 233.3±147.1ml, p=0.020) or CLP (6.2±7.3 vs 44.3±24.7%, p=0.013). The best cut off values for the CLV and CLP were 50ml and 14.7%, respectively. The sensitivity, negative predictive value, and specificity of CLT were relatively high, but the positive predictive value was low. Conclusion: The likelihood of developing extubation failure increases with increasing severity of illness and duration of intubation. A low CLV or CLP (