Assessment of Screening Scores and Upper Airway Metrics as Diagnostic Alternatives to Polysomnography in Obstructive Sleep Apnea

Background: Obstructive sleep apnea (OSA) is a widespread but frequently overlooked condition because polysomnography (PSG), the reference standard for diagnosis, is expensive and not easily accessible.This study explored whether combining clinical questionnaires, basic anthropometric indices, and acoustic pharyngometry (AP) enhances the detection of moderate-to-severe OSA compared with PSG-derived apnea-hypopnea index (AHI).Materials and methods: We performed a cross-sectional study including 72 adults with confirmed OSA who were referred for oral appliance therapy.Each participant completed the Snoring, Tiredness, Observed apnea, high blood Pressure, Body mass index (BMI), Age, Neck circumference, and Gender (STOP-BANG) Questionnaire, the Epworth Sleepiness Scale (EPSS), and the Berlin Questionnaire.Demographic and anthropometric measures were collected, and airway dimensions were assessed using AP.Associations with AHI were examined using Spearman's rank correlation, and diagnostic validity was tested with receiver operating characteristic (ROC) analyses.Results: The AHI correlated significantly with STOP-BANG (ρ = 0.380, p = 0.001), EPSS (ρ = 0.363, p = 0.002), and Berlin Questionnaire scores (ρ = 0.295, p = 0.012).Acoustic pharyngometry values showed inverse associations with age, body weight, and STOP-BANG but did not correlate directly with AHI.Receiver operating characteristic analyses demonstrated that STOP-BANG [cutoff ≥6.5; sensitivity: 79.6%, specificity: 66.7%, area under the curve (AUC): 0.732, 95% confidence interval (CI): 0.615-0.849]and EPSS (cutoff ≥10.5; sensitivity: 75%, specificity: 75%, AUC: 0.726, 95% CI: 0.611-0.841)had the strongest predictive power.The Berlin Questionnaire showed moderate accuracy (AUC: 0.679), while AP alone had limited discriminative ability (AUC: 0.566).Conclusion: Among the evaluated screening strategies, STOP-BANG and EPSS were the most reliable for predicting moderate-to-severe OSA and may be prioritized for clinical use.Although AP on its own was not diagnostically robust, its anatomical insights suggest value as a supportive measure when integrated into multimodal screening protocols, particularly in healthcare systems with limited PSG access.