Received: 13 August 2001 Accepted: 26 June 2002 Published online: 10 August 2002 © Springer-Verlag 2002 Abstract Objective: Laryngeal ede- ma secondary to endotracheal intu- bation may require early re-intuba- tion. Prior to extubation the absence of leak around an endotracheal tube may predict laryngeal edema after extubation. We evaluated the useful- ness of a quantitative assessment of such a leak to identify the patients who will require early re-intubation for laryngeal edema. Methods: This prospective study included 76 pa- tients with endotracheal intubation for more than 12 h. The leak, in per- cent, was defined as the difference between expired tidal volume mea- sured just before extubation, in vol- ume-controlled mode, with the cuff inflated and then deflated. The best cut-off value to predict the need for re-intubation for significant laryn- geal edema was determined and the patients were divided into two groups, according to this cut-off val- ue. Results: Eight of the 76 patients (11%) needed re-intubation for la- ryngeal edema. Patients requiring re- intubation had a smaller leak than the other patients [9 (3–18) vs 35 (13–53)%, p<0.01]. The best cut-off value for gas leak was 15.5%. The high leak group included 51 patients, of whom only two patients (3%) re- quired re-intubation. The low leak group included 25 patients, among whom six patients (24%) required re-intubation (p<0.01). The sensitivi- ty of this test was 75%, the specifici- ty 72.1%, the positive predictive val- ue 25%, the negative predictive val- ue 96.1% and the percent of correct classification 72.4%. Conclusions: A gas leak around the endotracheal tube greater than 15.5% can be used as a screening test to limit the risk of re-intubation for laryngeal edema. Keywords Cuff leak test · Laryngeal edema · Endotracheal intubation · Mechanical ventilation · Extubation failure · Stridor Intensive Care Med (2002) 28:1267–1272 DOI 10.1007/s00134-002-1422-3 ORIGINAL Yann De Bast Daniel De Backer Jean-Jacques Moraine Muriel Lemaire Cécile Vandenborght Jean-Louis Vincent The cuff leak test to predict failure of tracheal extubation for laryngeal edema Introduction Endotracheal intubation may generate local complica- tions, including mechanical lesions like friction, com- pressions between the tube and the anatomic structures [1], and also biochemical reactions at the interaction be- tween the plastic or silicone tube material and the upper airway mucosa. Even though their incidence has de- creased with the use of far more flexible polyvinyl chlo- ride tubes generating less pressure on the anatomical structures and high volume low pressure cuffs inducing fewer mucosal lesions [1, 2, 3, 4, 5, 6], these complica- tions remain frequent. Laryngeal edema manifests itself by respiratory distress and inspiratory whistling called “stridor”. Such edema formation can develop as early as 6 h after intubation [7], so that prolonged intubation is not a prerequisite to its development [8, 9]. By decreas- ing the upper respiratory tract diameter, laryngeal edema increases airway resistance and, consequently, the respi- ratory work. When edema is present, re-intubation may be required if the patient is not able to sustain the in- crease in respiratory work. The development of laryngeal Y. De Bast · D. De Backer · J.-J. Moraine M. Lemaire · C. Vandenborght J.-L. Vincent ( ✉ ) Department of Intensive Care, Erasme University Hospital, Free University of Brussels, 808 route de Lennick, 1070 Brussels, Belgium e-mail: jlvincen@ulb.ac.be Tel.: +32-2-5553380 Fax: +32-2-5554555