Ear oximetry to detect apnea and differentiate rapid eye movement (REM) and non-REM (NREM) sleep. Screening for the sleep apnea syndrome

Abstract

Ear oximetry is commonly employed in screening patients for the sleep apnea syndrome, but the lack of objective information regarding the duration of sleep, including the presence of rapid-eye-movement (REM) sleep, is a major limitation. Based on the premise that both apnea and sleep-state-dependent changes in ventilation give rise to distinctive patterns in the arterial oxygen saturation, we developed a systematic technique to analyze ear oximetric tracings for wakefulness, REM sleep, and non-REM (NREM) sleep. Fifty-four patients were studied by both all-night polysomnography and ear oximetry. A careful comparison of ear oximetric data for sleep states and apnea was then made, using polysomnography as the correct classification to determine sensitivity, specificity, predictive value positive, and predictive value negative of the ear oximetric tracings. When classification of sleep state was compared, ear oximetry correctly classified 280.5 (82 percent) of 340.9 hours of sleep that was either REM or NREM sleep. The sensitivity for classifying NREM sleep was 0.85, for REM sleep was 0.70, and for wakefulness was 0.49. The sensitivity by ear oximetry for apnea was 0.80, with a predictive value negative of 0.87. We conclude that although polysomnography must be performed for definitive evaluation, ear oximetry is a valuable screening test for sleep apnea because the presence or absence of apnea can be determined, total duration of sleep can be estimated, and NREM vs REM sleep can be differentiated.