Performance of Epworth Sleepiness Scale and tiredness symptom used with simplified diagnostic tests for the identification of sleep apnea

OBJECTIVE: To compare the performance of Epworth Sleepiness Scale (ESS) and tiredness symptom with the apnea-hypopnea index (AHI) in a population referred to home sleep tests. MATERIALS AND METHODS: This correlational study assessed adult patients through questionnaires and respiratory polygraphy (RP). We estimated sensitivity (S), specificity (Sp), predictive values (PV), odds ratio (OR) adjusted multivariate model and area under ROC curve for each sex and severity. RESULTS: We analyzed 4424 patients, 2761 (62.4%) men and 1663 women, aged 53.6 (42-65 years old) with BMI of 31.3 (25.5-36.1). 78.4% had AHI >5 events/hour. RP (ev/h) indicators were (men vs. women): AHI of 22.8±19.2 vs. 13.2±13.3, ODI of 22.7±19.9 vs. 14.0±13.7, and T <90%: 19.3±26.1 vs. 15.6±25.3. Men presented higher severity levels, night-time hypoxemia and CPAP indications (52.2 vs. 29.2%) p<0.0001. ESS > 10 was found in 25% of population: 8±5.15 in men vs. 7.6±5.1 in women, p<0.001. 12% of men (as compared to 31.5% of women) with ESS > 10 had a normal AHI. 72% of women reported tiredness (vs. 66.1% of men). The R2 between Epworth Scale and AHI showed: 0.022 (CI95%: 0.111-0.185) p<0.0001 in men and 0.0019 (CI95%: -0.004 to 0.092) p>0.074 in women. Logistic regression showed Epworth Sleepiness Scale >10 for each AHI severity category (OR between 1.38 and 1.31 with p<0.05) and tiredness for AHI >30 ev/h only in men (p<0.004). CONCLUSIONS: Epworth Sleepiness Scale >10 demonstrated a low screening performance only when present in male patients. Tiredness performed worse. Due to its limited value in the identification of sleep apnea patients, subjective somnolence should be considered in the context of an objective evaluation.