Is the OSA-18 predictive of obstructive sleep apnea: Comparison to polysomnography

Objectives/Hypothesis To examine the ability of the OSA-18 to predict Obstructive Sleep Apnea (OSA) in a racially diverse population when compared to overnight polysomnography (PSG). Study Design Cross-sectional retrospective. Methods Children 2 to 12 years of age diagnosed with OSA who were treated at a tertiary care institution between 2008 and 2013 and had complete PSG and OSA-18 data were included. We performed logistic regression with OSA as the dependent variable and the OSA-18 total symptom score (TSS), age, gender, race, asthma, and body mass index (BMI) as independent variables. Results Seventy-nine children (32 females) were included (mean age 5.2 ± 2.4 years). The positive predictive value (PPV) was greater than 90 for an obstructive apnea-hypopnea index (oAHI) ≥ 1. The PPV and specificity were higher for white than for nonwhite children; however, sensitivity and negative predictive value (NPV) of OSA-18 TSS were low for mild, moderate, and severe OSA regardless of race. Age, race, and BMI were not significantly associated with oAHI. Conclusions This study, conducted in a racially diverse cohort, examined the ability of the OSA-18 to predict OSA when compared to PSG - the gold standard - and found that sensitivity and NPV were extremely low for both white and nonwhite children. This suggests that the OSA-18 is not sufficiently sensitive to detect OSA nor sufficiently specific to determine the absence of OSA. The OSA-18 should be used as a quality-of-life indicator and is not a reliable substitute for PSG.