OBJECTIVEPostextubation distress after a successful spontaneous breathing trial is associated with increased morbidity and mortality. Predicting postextubation distress is therefore a major issue in critically ill patients. To assess whether lung derecruitment during spontaneous breathing trial assessed by lung ultrasound is predictive of postextubation distress.DESIGN AND SETTINGProspective study in two multidisciplinary intensive care units within University Hospital.PATIENTS AND METHODSOne hundred patients were included in the study. Lung ultrasound, echocardiography, and plasma B-type natriuretic peptide levels were determined before and at the end of a 60-min spontaneous breathing trial and 4 hrs after extubation. To quantify lung aeration, a lung ultrasound score was calculated. Patients were followed up to hospital discharge.MEASUREMENTS AND MAIN RESULTSFourteen patients failed the spontaneous breathing trial, 86 were extubated, 57 were definitively weaned (group 1), and 29 suffered from postextubation distress (group 2). Loss of lung aeration during the successful spontaneous breathing trial was observed only in group 2 patients: lung ultrasound scores increased from 15 [13;17] to 19 [16; 21] (p < .01). End-spontaneous breathing trial lung ultrasound scores were significantly higher in group 2 than in group 1 patients: 19 [16;21] vs. 10 [7;13], respectively (p < .001) and predicted postextubation distress with an area under the receiver operating characteristic curve of 0.86. Although significantly higher in group 2, B-type natriuretic peptide and echocardiography cardiac filling pressures were not clinically helpful in predicting postextubation distress.CONCLUSIONLung ultrasound determination of aeration changes during a successful spontaneous breathing trial may accurately predict postextubation distress.