Abstract
BACKGROUND: Postextubation stridor as a result of laryngeal edema has been reported in up to 5% of patients who have received mechanical ventilation for longer than 24 hours. The cuff-leak test has been proposed as a method of determining whether patients can be extubated without developing upper airway obstruction. This test, however, has not been prospectively studied. METHODS: 100 consecutive patients who had been intubated for longer than 24 hours were studied prior to extubation. The cuff-leak test was performed by two different methods. The endotracheal cuff was totally deflated, the lumen of the endotracheal tube was then occluded, and the presence (small or large) or absence of a peritubular leak during spontaneous ventilation was determined. The test was then repeated with positive pressure ventilation using the patients’ pre-weaning tidal volume. The results were not available to the intensivist, and the patients were extubated regardless of the results of the leak tests. RESULTS: The mean (SD) length of intubation was 3.8 (3.2) days. Six patients had no leak with both tests, 22 patients had a small leak with both tests, and 60 patients had a large air leak with both tests. Twelve patients had a small leak with one test and a large leak with the other. Two patients developed postextubation stridor. Both were successfully treated with aerosolized racemic epinephrine. These two patients had no leak with either test, making the likelihood of developing stridor significantly higher in the group of patients that had no peritubular leak prior to extubation (p = 0.003). The cuff-leak tests had a negative predictive value of 100% but a positive predictive value of only 33%. CONCLUSION: The presence of a cuff leak indicates that postextubation stridor is unlikely. Although the absence of a cuff leak does not guarantee extubation difficulty, such patients should be observed closely during the immediate postextubation period [Respir Care 1996;41(6):509-511].
Get full access to this article
View all access options for this article.
