Performance of Four Screening Tools for Identifying Obstructive Sleep Apnea Among Patients with Insomnia

Co-morbid insomnia and sleep apnea are common in clinical practice. The existing OSA screening tools have not been fully validated in insomnia populations, and items measuring daytime function may be interfered with the presence of insomnia. This study aims to validate the performance of four commonly used OSA screening tools among individuals with and without insomnia.

A cross-sectional survey was conducted in individuals with suspected OSA referred for sleep studies from December 2021 to December 2023. All participants completed the Insomnia Severity Index (ISI) scale, STOP-Bang, Epworth Sleepiness Scale (ESS), Berlin questionnaire, and NoSAS score. Clinical insomnia was defined as an ISI of 15 or more. Performance of screening tools was primarily assessed by sensitivity, specificity, and the receiver operating characteristic (ROC) curve.

A total of 1266 participants (26% females, age 46.4 ± 12.4 years) were included in the study. The prevalence of apnea-hypopnea index (AHI) ≥15/h was 48% and 52% in the insomnia (n=313) and non-insomnia (n=953) group, respectively (P>0.05). In presence of insomnia, the STOP-Bang, ESS, and Berlin questionnaire demonstrated higher sensitivity but lower specificity. Using conventional cutoffs, the STOP-Bang had the highest level of sensitivity (93.2%, 95% CI 87.6-96.5%), while NoSAS had the highest level of specificity (67.7%, 95% CI 59.9-74.6%) for identifying AHI ≥15/h. The STOP-Bang and NoSAS outperformed ESS and Berlin with areas under the ROC curve >0.7 at all levels of OSA severity. The Youden's index was maximized at score 4 for STOP-Bang and score 7 for NoSAS.

The performance of OSA screening tools incorporating evaluation of daytime function is altered in the presence of insomnia. Under conventional cutoffs, STOP-Bang is the preferred screening tool due to its high sensitivity.