Spectral analysis of ventilation in elderly subjects awake and asleep.

We studied the periodicities of ventilation in elderly subjects using digital comb filtering. Two groups of subjects were studied, those with and without sleep apnea. Measurements were made in wakefulness, stage 1-2 sleep, and where possible in stage 3-4 sleep. For each of the digital filters we calculated the average power of the oscillatory output. To compare subject groups we first specifically determined the average power in the filter with the maximum output. The mean of this measurement was greater in elderly subjects with apnea compared with those without apnea, both during wakefulness and stage 1-2 sleep. In both groups of subjects the cycle time of the major ventilatory oscillations was on the order of 40-60 s. There was no difference in this cycle time between the two groups of subjects in wakefulness or stage 1-2 sleep. Thus, whereas similar oscillatory processes occur in subjects with and without apnea, it is the magnitude of the oscillation that differs between the two groups. These conclusions are supported by analysis of the output of individual filters of the digital comb filter. In both groups, stage 1-2 sleep produced significantly increased oscillations in ventilation. Both in wakefulness and stage 1-2 sleep, significantly greater periodicities occurred in the apneic compared with the nonapneic group. In the few subjects who had sufficient data in stage 3-4 sleep for spectral analysis, ventilatory oscillations were virtually absent in this state. Our data suggest that subjects who develop apnea during sleep have an increased propensity for periodic breathing even while awake.