Combination of symptoms and oxygen desaturation index in predicting childhood obstructive sleep apnea

To develop a screening process of obstructive sleep apnea in children based on a combination of symptoms and oxygen desaturation index (ODI).

We performed a retrospective study of 141 Chinese patients who were referred to a pediatric sleep laboratory for possible obstructive sleep apnea (OSA). The parents of each patient answered a questionnaire before their child underwent polysomnography (PSG) in the laboratory. An apnea-hypopnea index (AHI) greater than five on nocturnal PSG was defined as OSA. The nocturnal PSG was interpreted by a sleep laboratory physician. The ODI and occurrence ratio of sleep problems such as snoring, observable apnea during sleep, mouth breathing, and restless sleep, among others were compared between the OSA and non-OSA groups using the chi-square test. Items that indicated statistically significant differences were tested with non-parametric Spearman correlation tests to determine the correlation between these items and AHI. ODI and the items that indicated a statistically significant difference between the OSA and non-OSA groups were analyzed using binary logistic regression. The ODI cut-off point was determined through ODI receiver operating characteristic analysis to distinguish between OSA and non-OSA. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated to determine the combination of OSA predictors that exhibited the best diagnostic performance.

Among the 141 patients, 78 (55%) were diagnosed with OSA by PSG. The occurrences of observable apnea during sleep, mouth breathing, and restless sleep were significantly different between the OSA and non-OSA groups (20.5% vs. 4.8%, 85.9% vs. 71.4%, 69.2% vs. 52.4%, respectively, with P<0.05). The median of ODI in the OSA group was significantly higher than that in the non-OSA group. The ODI and the occurrences of observable apnea during sleep, mouth breathing, and restless sleep were correlated with AHI and were important diagnostic factors of OSA in children, as determined through binary logistic regression. The presence of observable apnea during sleep had 95% specificity, 84% PPV, and 4.31 positive likelihood ratio (PLR). When score ≥3 (i.e., 3 or 4) was used as the cut-off point, specificity, PLR, and PPV were 0.86, 4.22, and 0.84, respectively. When score ≥2 was used as the as cut-off point, sensitivity, NLR, and NPV were 0.92, 0.2, and 0.80, respectively.

Observable apnea during sleep was an independent positive predictive factor for OSA in children. A child with observable apnea during sleep should be referred to a special sleep laboratory for PSG diagnosis. When the total score is 3 or 4 based on a combination of symptoms and ODI, OSA can be diagnosed and the child should be referred to a sleep pediatrician for appropriate intervention. When the total score is 0 or 1, the child can be considered normal but should be monitored. When the total score is 2, the result cannot be determined and the child should be referred to a special sleep laboratory for PSG diagnosis. Thus, a screening process is developed based on a combination of symptoms and ODI.