To assess the accuracy of the respiratory inductive plethysmograph (RIP) during sleep in obese patients with obstructive sleep apnea (OSA), we monitored 13 patients with OSA during wakefulness and nocturnal sleep with simultaneous measurements of tidal volume from RIP and integrated airflow. Patients wore a tightly fitting face mask with pneumotachograph during wakefulness and sleep. Calibrations were performed during wakefulness prior to sleep and compared with subsequent wakeful calibrations at the end of the study. Patients maintained the same posture during sleep (supine, 11; lateral, two) as during calibrations. There were no significant differences in calibrations before sleep and after awakening. The mean error in 13 patients undergoing RIP measurements of tidal volume during wakefulness was -0.7 +/- 3.4 percent while that during sleep was 2.1 +/- 14.9 percent (p < 0.001). The standard deviation (SD) of the differences between individual breaths measured by RIP and integrated airflow was 9.8 +/- 5.5 percent during wakefulness and 25.5 +/- 18.6 percent during sleep (p < 0.001). During both wakefulness and sleep, errors in RIP tidal volume were not significantly correlated with body mass index. In 12 patients with at least 10 percent time in each of stages 1 and 2 sleep, SD was greater in stage 2 sleep compared with wakefulness and stage 1 (p < 0.001). In three patients who manifested all stages of sleep, SD was greater in REM sleep than in wakefulness and all stages of non-REM sleep (p < 0.001). In three patients who manifested all stages of sleep, SD was greater in REM sleep than in wakefulness and all stages of non REM sleep (p < 0.001). This was associated with paradoxic motion of the rib cage in two patients during REM. We conclude that, despite increased errors in individual breath measurements during sleep, more marked during stages 2 and REM sleep, RIP is clinically useful to measure ventilation quantitatively in obese patients with sleep apnea. The criterion of a decrease of 50 percent in tidal volume assessed by RIP is appropriate to define hypopneas in such patients.