To investigate the relative contribution of various risk factors to the surgical outcome of adenotonsillectomy for obstructive sleep apnea syndrome in children. Children (n = 110; mean age, 6.4 +/- 3.9 years) underwent two polysomnographic evaluations before and after adenotonsillectomy. In addition, 22 control children were studied. History for allergy and family history of sleep-disordered breathing was taken before each polysomnographic evaluation. Significant changes in sleep stage percentages and sleep fragmentation were found in the postsurgery study compared with the presurgery study; 25% of the children had apnea/hypopnea index (AHI) </=1, 46% had AHI >1 and <5, and 29% had AHI >/=5 in the postsurgery study. The frequency of subjects with AHI </=1 after surgery was significantly lower among obese subjects (P < .05). Comparison between the children who had AHI </=1 after surgery and 22 control children showed complete normalization of sleep architecture after surgery. Adenotonsillectomy yields improvements in respiratory abnormalities in children with obstructive sleep apnea syndrome. Complete normalization occurs in only 25% of the patients. Obesity and AHI at diagnosis are the major determinant for surgical outcome. When normalization of respiratory measures occurs after surgery, normalization of sleep architecture will also ensue.