Predictive value of B erlin Q uestionnaire and E pworth S leepiness S cale for obstructive sleep apnea in a sleep clinic population

INTRODUCTION: The Berlin Questionnaire (BQ) and Epworth Sleepiness Scale (ESS) are two of the widely used screening instruments for subjects suffering from sleep disorders. OBJECTIVES: To determine the predictive accuracy of the BQ and ESS in the subjects who were admitted to our sleep clinic, using overnight polysomnography as the 'gold standard'. METHODS: The BQ and ESS were performed to 1450 subjects hospitalized in our sleep clinic for polysomnographic evaluation. RESULTS: One thousand two hundred thirty of 1450 subjects (84.8%) had apnea-hypopnea index (AHI) ≥ 5, and 70.5% of them were classified as being at high risk of sleep apnea with BQ and 45.9% of them were classified as having excessive daytime sleepiness with ESS. Being in the high-risk group with the self-reported, BQ predicted an AHI ≥ 5 with a sensitivity of 0.73 and a specificity of 0.44, vs a sensitivity of 0.46 and a specificity of 0.60 with the self-reported ESS. Sensitivity of BQ (0.80) was highest when we selected the cut-off value of AHI as 30. When we evaluated the predictive accuracy of BQ and ESS together, specificity of both instruments as 0.72 at AHI cut-off value of 5 was found. AHIs were significantly correlated with the scores in each category and the total scores of BQ (P < 0.0001). CONCLUSIONS: The BQ is a poor predictor of obstructive sleep apnea (OSA) in subjects admitted to our sleep clinic. BQ and ESS together may be used to detect patients with low risk for OSA.