Epworth sleepiness scale (ESS) is widely used to evaluate degree of somnolence among Chinese patients with sleep-disordered breathing. Yet no Mandarin-Chinese translation has ever reported its validation data. In this study we translated and validated the ESS into Mandarin Chinese (CESS). We found, in 31 bilingual patients' responses to the CESS and the English ESS obtained 1-2 weeks apart did not differ significantly (10.5 +/- 3.7 vs. 9.6 +/- 3.9, p = 0.32, Wilcoxon's signed rank test) and were significantly correlated (Spearman's p = 0.67, p = 0.0004). A total of 359 sleep-disordered breathing subjects were enrolled into the validation study in a prospective manner. The CESS showed acceptable internal consistency (Cronbach's alpha = 0.81). Thirty out of these patients answered CESS twice at an interval of 2-4 weeks, to yield an acceptable level of test-retest reliability (p = 0.74, p = 0.001). The respiratory disturbance indices obtained from 251 out of 359 subjects were significantly correlated with their CESS scores (p = 0.22, p < 0.001). The standard response mean of CESS was 0.86 as obtained from 94 out of 359 subjects who had initial control of their symptoms at 3 months after radio-frequency palate surgery. We conclude that the CESS is reliable in both a linguistic and a test-retest sense, and appears to be valid and sensitive to clinical change. The CESS could be used to advantage among other Mandarin-speaking subjects as a standardised screening test of sleepiness in daily life.