Sequential screenings improve prediction in obstructive sleep apnoea severity

Abstract

Objectives: Sleep apnoea syndrome (SAS) poses serious health risks and requires Polysomnography (PSG) for diagnosis. Due to PSG’s labour-intensive process and long waiting times, there’s a need for predictive models to prioritize severe SAS cases for early PSG.

Methods: We retrospectively reviewed PSG cases from the University Malaya Medical Centre’s ENT clinic (January-December 2023). Data included demographics, anthropometrics, sleep patterns, STOP-BANG questionnaire, Epworth Sleepiness Scale (ESS), and upper airway assessments, analysed for their relationship with AHI. All variables that exhibited statistical significance were categorised into two sets of combined variables: Clinical Examination Score (CES) and ENT Examination Score (EES).

Results: We studied 201 cases, with an average age of 47.8±15.4 years (range: 16–89). Of these, 125 (62.2%) were male, with a mean AHI of 49.0±33.5. Severe sleep apnoea (AHI≥30) was diagnosed in 127 (63.2%) cases. Significant differences in AHI (p<0.05) were found based on gender, snoring, apnoea, nocturia, drooling, BMI>35kg/m², neck circumference > 40cm, Q1 of ESS (r=0.17), Q3 of ESS (r=0.23), Q6 of ESS (r=0.17), total ESS score (r=0.20), Modified Mallampati (r=0.26), Palatine Tonsil grade (r=0.27), Retropalatal grade (r=0.29), and Retrolingual grade (r=0.29). Hierarchical Multiple Regression analysis explained 33.0% of the variance in AHI, F(12,188)=7.2, p<0.01. CES had the highest correlation with AHI (R=0.475) and AUC (0.735) in ROC analysis, showing high sensitivity (82.7%) and moderate specificity (50.0%); combined with EES, specificity improved to 72.5%.

Conclusion: The CES and EES help prioritize moderate and severe OSA patients for early diagnosis and treatment. This reduces complications, eases healthcare workload, and shortens diagnostic wait times.