Diagnostic meta-analysis of the Pediatric Sleep Questionnaire, OSA-18, and pulse oximetry in detecting pediatric obstructive sleep apnea syndrome

Abstract

The aim of this meta-analysis was to compare the pooled sensitivity and specificity of the Pediatric Sleep Questionnaire (PSQ), Obstructive Sleep Apnea Questionnaire (OSA-18), and pulse oximetry (PO) according to OSAS severity. Three electronic databases were searched for studies evaluating sensitivity and specificity of the three tools against the apnea-hypopnea index measured using overnight in-laboratory or in-home polysomnography in children and adolescents from inception until January 11, 2020. A random-effects bivariate model was used to estimate the summary sensitivity and specificity of the tools. We identified 39 studies involving 6131 clinical and community children (aged 2.9-16.7 y). The PSQ exhibited the highest sensitivity (74%) for detecting symptoms of mild pediatric OSAS. The PSQ and PO had comparable sensitivity in screening moderate and severe pediatric OSAS (0.82 and 0.89 vs 0.83 and 0.83, respectively). PO yielded superior specificity in detecting mild, moderate, and severe pediatric OSAS (86%, 75%, and 83%, respectively) than did the PSQ and OSA-18 (all p < 0.05). Age, percentage of girls, index test criteria, methodology quality, and sample size significantly moderated sensitivity and specificity. For early detection of pediatric OSAS, the combined use of PSQ and PO is recommended when polysomnography is not available. PROSPERO REGISTRATION NUMBER: CRD42018090571.

Keywords: Diagnostic meta-analysis; Obstructive sleep apnea syndrome; Pediatric.