Questionnaire OSA-18 has poor validity compared to polysomnography in pediatric obstructive sleep apnea

• Published in: International journal of pediatric otorhinolaryngology 2013-11, Vol.77 (11), p.1864-1868
• Main Authors: Borgström, Anna, Nerfeldt, Pia, Friberg, Danielle
• Format: Article
• Language: English
• Subjects: Child; Child, Preschool; Children; Cohort Studies; Cross-Sectional Studies; Female; Humans; Male; Medicin och hälsovetenskap; Obstructive sleep apnea; OSA-18; Otolaryngology; Pediatrics; Polysomnography; Polysomnography - methods; Quality of Life; ROC Curve; Sensitivity and Specificity; Severity of Illness Index; Sleep Apnea, Obstructive - diagnosis; Sleep Apnea, Obstructive - psychology; Surveys and Questionnaires
• ISSN: 0165-5876; 1872-8464; 1872-8464
• DOI: 10.1016/j.ijporl.2013.08.030

Abstract Objective To evaluate the diagnostic value of the quality-of-life instrument OSA-18 by comparing it with objective data from polysomnography in children with sleep-disordered breathing. Study Design Cross-sectional. Patients and Methods Full-night polysomnographic data were obtained from 225 subjects, 139 boys and 86 girls, median age 4.5 years (1–12) in our sleep laboratory. Their caregivers answered the OSA-18 quality-of-life instrument (range 18–126). The polysomnographic parameter, the apnea-hypopnea index (AHI) was compared with the total symptom score (TSS) and with the subscale of sleep disturbance (SD) from the OSA-18 questionnaire. Receiver operating characteristic (ROC) curves were created to test the predictive value of OSA-18. Results With the TSS of the OSA-18 at ≥60, compared with AHI levels of >1 and ≥5, the sensitivity was 55.2% and 59.3% respectively, and the specificity 40.9% and 48.4%, respectively. With the TSS > 80 and AHI levels of ≥5 and ≥10, the sensitivity was 24.6% and 32.1%, respectively. For the subscale of SD, the majority of the subjects showed poor correlation with the AHI values. The ROC area under the curve for different levels of the AHI (>1, ≥5, and ≥10) was 0.49, 0.57, and 0.56, respectively. Conclusions The OSA-18 questionnaire showed poor validity in detecting and predicting pediatric OSA. The majority of the children with severe OSA would not be correctly diagnosed if the OSA-18 were used as a dominant diagnostic tool.