Comparison of NoSAS score with Berlin and STOP-BANG scores for sleep apnea detection in a clinical sample.

Because of their high costs and low availability, sleep recordings cannot be used as routine procedures for sleep apnea screening. Therefore, it is important to have a performant screening tool allowing to select patients at higher risk for sleep apnea who need further investigations. The goal of the study is to compare the performances of the three commonly used sleep disordered breathing (SDB) screening questionnaires in a clinical sample. Clinical data and sleep studies performed in consecutive adult patients referred to the Lausanne University sleep center for sleep recordings between November 2016 and February 2018 were analysed. Berlin, STOP-Bang and NoSAS screening scores were calculated and compared with the sleep studies' results. NoSAS score showed a NPV of 0.88, a PPV of 0.43 and a correctly classified rate of 71% for an AHI >15/h. STOP-Bang score had a slightly higher negative predictive value (0.92) but a very low positive predictive value (0.30) and a poor correctly classified rate (47%). Berlin score showed globally poor results (NPV:0.85, PPV: 0.29 and a correctly classified rate: 56%). NoSAS score had the highest area under ROC curves (0.78) compared to STOP-bang (0.71) and Berlin (0.62). NoSAS score however yielded a slightly higher rate of false negative than STOP-Bang score. NoSAS false negatives were mainly female snorers with a low neck circumference. NoSAS score showed a better discrimination capacity compared to Berlin and STOP Bang scores in an unselected clinical population referred to a sleep center