The endotracheal tube cuff-leak test as a predictor for postextubation stridor.

BACKGROUND The endotracheal tube (ETT) cuff-leak test (CLT) has been proposed as a relatively simple, noninvasive method for detecting the presence of laryngeal edema prior to tracheal extubation. OBJECTIVE To determine the value of the CLT for predicting postextubation stridor (PES) among medical and surgical patients, and to assess the impact of certain variables on the incidence of PES. METHODS We conducted a prospective, observational study in the intensive care unit at Washington Hospital Center, a 907-bed acute care hospital in Washington DC, with patients who were intubated for > 24 h. As part of respiratory therapy quality assurance, patients intubated for > 24 h are evaluated daily for extubation readiness, and CLT is conducted prior to extubation. The CLT results and the postextubation outcomes were prospectively recorded for 6 months. RESULTS Of the 462 patients studied, 20 (4.3%) developed PES that required treatment; 7 of those 20 (1.5%) required reintubation. With patients who failed the CLT, defined by an absolute leak volume < or = 110 mL, the positive predictive value for PES was 0.12, the negative predictive value was 0.97, the sensitivity was 0.50, and the specificity was 0.84. Using different definitions for CLT failure did not improve the accuracy of CLT for predicting PES. Patients who had PES were more likely to be female (6.5% vs 2.4%, p = 0.04), to have a longer duration of translaryngeal intubation (6.5 + 4 d vs 4.5 + 4 d, p = 0.02), and to have a larger ratio of ETT size to laryngeal size (49.5 + 6% vs 45.5 + 6%, p = 0.01). CONCLUSIONS Failing the CLT was not an accurate predictor of PES and should not be used as an indication for either delaying extubation or initiating other specific therapy. Female patients, those whose ratio of ETT size to laryngeal diameter was > 45%, and patients intubated for > 6 d were more likely to develop PES.