A retrospective study of pediatric patients with obstructive sleep apnea who underwent adenotonsillectomy between 1987 and 1990 was undertaken to determine the frequency of postoperative respiratory compromise and to determine if risk factors for its development could be identified. Sixty-nine patients less than 18 years old had polysomnographically documented obstructive sleep apnea and were observed postoperatively in the pediatric intensive care unit. Of these, 16 (23%) had severe respiratory compromise, defined as intermittent or continuous oxygen saturation of 70% or less, and/or hypercapnia, requiring intervention. Compared with patients without respiratory compromise, these patients were younger (3.4 +/- 4 vs 6.1 +/- 4 years) and had more obstructive events per hour of sleep on the polysomnogram (49 +/- 41 vs 19 +/- 30). They were more likely to weight less than the fifth percentile for age (odds ratio [OR], 5.1; 95% confidence interval [CI], 1.4 to 18.7), to have an abnormal electrocardiogram and/or echocardiogram (OR, 4.5; 95% CI, 1.3 to 15.1), and to have a craniofacial abnormality (OR, 6.2; 95% CI, 1.5 to 26). Multiple logistic regression analysis revealed the most significant risk factors were age below 3 years and an obstructive event index greater than 10. Children with obstructive sleep apnea are at risk for respiratory compromise following adenotonsillectomy; young age and severe sleep-related upper airway obstruction significantly increase this risk. We recommend in-hospital postoperative monitoring for children undergoing adenotonsillectomy for obstructive sleep apnea.