Objective: To determine which factors are the best predictors of postextubation stridor in pediatric trauma patients. Design: Prospective cohort study. Setting: The Burn and Trauma ICUs at Harborview Medical Center from March to September 1989. Patients: Children were eligible for the study if they were < 15 yr old, were intubated for > 12 hr, and did not have underlying cardiopulmonary disease. The study included 25 patients with 30 extubations. Risk Factors Assessed: Age, type of injury (burn vs. trauma), location of intubation ("field" vs. hospital), endotracheal tube size, length of intubation, and presence of an airleak around the tube at the time of extubation at 30 cm H2O pressure. Main Outcome Measure: Moderate to severe postextubation stridor requiring treatment with racemic epinephrine, helium-oxygen, reintubation, or tracheostomy. Results: Treatment for postextubation stridor was required after 11 (37%) of 30 extubations, with five reintubations and one tracheostomy. The best predictor of postextubation stridor was absence of an airleak at the time of extubation (sensitivity 100%, positive predictive value 79%, p < .001), followed by type of injury (facial burn vs. all others; sensitivity 64%, positive predictive value 88%, p < .001). After controlling for these two factors, no other factor studied was a significant predictor of postextubation stridor. Conclusion: In pediatric trauma patients, mechanism of injury (facial burn vs. other) and absence of an airleak at the time of extubation are the strongest factors predicting postextubation stridor. Patients with one or both risk factors require special attention to airway management. (Crit Care Med 1991; 19:352)