Previous work from our laboratory has indicated that mechanoreceptor feedback from the respiratory muscles may play an important role in arousal and apnea termination in obstructive sleep apnea (OSA). Other studies have pointed to a prominent role for chemoreceptor stimuli. We postulated that mechanoreceptor stimuli from the respiratory system are the primary determinant of apnea termination, and that chemoreceptor stimuli exert their effect indirectly through stimulation of ventilation and thus proprioceptive feedback. To test this, we measured the diaphragmatic tension-time index (TTdi) during obstructive sleep apneas in seven male subjects with severe untreated OSA. We compared the maximal TTdi values at end-apnea during administration of air, O2, and CO2. We reasoned that if mechanoreceptor stimuli mediate apnea termination, changing the degree of chemoreceptor stimulation during apneas should not alter the level of respiratory effort at end-apnea. O2 administration produced a significant increase in end-apneic arterial oxygen saturation (SaO2) and increased apnea duration. CO2 administration led to an increase in pre- and postapneic end-tidal carbon dioxide pressure (PETCO2), and tended to shorten apneas. However, the mean value for maximal end-apneic TTdi was 0.12 +/- 0.01 (SEM) during room air breathing and was unaltered by O2 (0.12 +/- 0.01) or CO2 (0.11 +/- 0.01) administration. The consistency of end-apneic TTdi values despite the varying chemical drive supports the hypothesis that apnea termination in OSA is mediated by mechanoreceptor feedback from the respiratory system, most likely from the respiratory muscles. The influence of chemoreceptor information may be mediated indirectly through an effect on ventilatory effort.