Overnight oximetry is widely used for screening for the sleep apnea hypopnea syndrome (SAHS). The degree of desaturation at an apnea event is known to be affected by the degree of obesity. We hypothesized that the diagnostic ability of oximetry for SAHS is affected by the degree of obesity. A total of 424 consecutive patients referred for possible SAHS were studied. The subjects were classified into three groups of normal-weight, overweight and obese based on the body mass index (BMI). The apnea-hypopnea index (> or = 15 h(-1)) by polysomnography was used as the diagnostic gold standard. Oximetry data were automatically analyzed to calculate the oxygen desaturation index (ODI2/3/4:at 2%/3%/4% threshold). The diagnostic abilities of the ODI were different in the three BMI-groups at a given cutoff value, e.g. the sensitivity/specificity of ODI4 (cutoff = 15) were 54%/100%, 83%/ 97%, and 98%/78% for the normal-weight, overweight and obese groups, respectively (P < 0.0001). The gender and the age had no significant effect on the ability. We demonstrated the diagnostic sensitivity and specificity of the ODI for SAHS depended on the BMI. Oximetry as a screening tool for SAHS may become more useful by selection of a cutoff value appropriate for the BMI of each subject.