Inability of clinical history to distinguish primary snoring from obstructive sleep apnea syndrome in children.

STUDY OBJECTIVE To determine whether primary snoring (PS) could be distinguished from childhood obstructive sleep apnea syndrome (OSAS) by clinical history. DESIGN Retrospective study of clinical history of 83 children with snoring and/or sleep disordered breathing who were referred for polysomnography. SETTING Tertiary referral center; pediatric pulmonary sleep apnea clinic. MEASUREMENTS We evaluated the ability of a clinical obstructive sleep apnea (OSA) score and other questions about sleep, breathing, and daytime symptoms to distinguish PS from OSAS in children. Parents were asked about the child's snoring, difficulty breathing, observed apnea, cyanosis, struggling to breathe, shaking the child to "make him or her breathe," watching the child sleep, afraid of apnea, the frequency and loudness of snoring, and daytime symptoms such as excessive daytime sleepiness (EDS). RESULTS Based on polysomnography results, 48 patients were classified as PS and 35 as OSAS. Peak endtidal CO2 (49 +/- 3.2 vs 55 +/- 8.2 [SD] mm Hg); lowest arterial oxygen saturation measured by pulse oximetry (95 +/- 1.9 vs 82 +/- 14%); and apnea/hypopnea index (0.27 +/- .3 vs 8.4 +/- 6 events/h) indicated that the diagnostic criteria for PS versus OSA were reasonable. There were no differences between PS and OSA patients with respect to age, sex, race, failure to thrive, obesity, history of EDS, snoring history, history of cyanosis during sleep, or daytime symptoms except for mouth breathing. There were no significant differences in sleep variables between PS patients and those with any severity of OSAS. The OSA score misclassified about one of four patients. Comparing PS and OSA patients, significant findings were daytime mouth breathing (61 vs 85%; p = 0.024); observed apnea (46 vs 74%; p = 0.013); shaking the child (31 vs. 60%; p = 0.01); struggling to breathe (58 vs 89%; p = 0.003); and afraid of apnea (71 vs 91%; p = 0.028). However, none of these were sufficiently discriminatory to predict OSAS. CONCLUSION We conclude that PS in children cannot be reliably distinguished from OSAS by clinical history alone.