Abstract
BACKGROUND: Diagnosis of sleep-disordered breathing (SDB) requires overnight polysomnography (PSG). Because of the cost and low availability of these procedures, the NoSAS score was developed to identify subjects at high risk of SDB. To evaluate the clinical utility of the NoSAS score for screening patients with SDB in China and to compare the predictive value of the NoSAS score with the Epworth Sleepiness Scale (ESS), we used the STOP-Bang questionnaire and the Berlin questionnaire. METHODS: In our study, we retrospectively reviewed the existing clinical data of patients who underwent an overnight PSG for suspected SDB from June 2014 to September 2017 at the sleep medical center of Guangdong Medical University Affiliated Second Hospital. The information we collected included all parts of the NoSAS score, the ESS, the STOP-Bang questionnaire and the Berlin questionnaire. Based on the severity of SDB determined by the apnea-hypopnea index (AHI), the patients were classified into four groups of primary snoring (<5 events/h), mild SBD (AHI ≥5 and <15 events/h), moderate SBD (AHI ≥15 and ≤30 events/h) and severe SBD (>30 events/h). We calculated the sensitivity, specificity, positive predictive value, negative predictive value and area under the receiver operating characteristic curve of the five questionnaires to compare their relative efficacy for screening SDB. RESULTS: A total of 479 consecutive patients (374 males and 105 females) ranging in age from 18 to 80 years old (mean ± SD, 48.9±14.4 years old) were recruited into this study. When using the standard of AHI ≥5 for diagnosing SDB, the NoSAS score had the largest area under the curve (AUC) (AUC =0.734), and the Berlin questionnaire (AUC =0.732) came second. Both exhibited a better predictive value than the ESS score and the STOP-Bang questionnaire. Using NoSAS ≥8 to predict AHI ≥5 events/h, AHI ≥15 events/h and AHI >30 events/h, the sensitivity and specificity were 0.590 and 0.707, 0.649 and 0.626, and 0.644 and 0.562, respectively; for the STOP-Bang questionnaire, the values were 0.721 and 0.512, 0.752 and 0.440, and 0.763 and 0.399, respectively; and for the Berlin questionnaire, the values were 0.721 and 0.512, 0.752 and 0.440, and 0.763 and 0.399, respectively. CONCLUSIONS: The NoSAS score and the Berlin questionnaire both exhibited good predictive value for SDB patients. NoSAS is a more suitable questionnaire to use in clinic for the conveniences but the similar performance with another questionnaire.