Control of breathing during sleep assessed by proportional assist ventilation.

We used proportional assist ventilation (PAV) to evaluate the sources of respiratory drive during sleep. PAV increases the slope of the relation between tidal volume (VT) and respiratory muscle pressure output (Pmus). We reasoned that if respiratory drive is dominated by chemical factors, progressive increase of PAV gain should result in only a small increase in VT because Pmus would be downregulated substantially as a result of small decreases in PCO2. In the presence of substantial nonchemical sources of drive [believed to be the case in rapid-eye-movement (REM) sleep] PAV should result in a substantial increase in minute ventilation and reduction in PCO2 as the output related to the chemically insensitive drive source is amplified severalfold. Twelve normal subjects underwent polysomnography while connected to a PAV ventilator. Continuous positive air pressure (5.2 +/- 2.0 cmH2O) was administered to stabilize the upper airway. PAV was increased in 2-min steps from 0 to 20, 40, 60, 80, and 90% of the subject's elastance and resistance. VT, respiratory rate, minute ventilation, and end-tidal CO2 pressure were measured at the different levels, and Pmus was calculated. Observations were obtained in stage 2 sleep (n = 12), slow-wave sleep (n = 11), and REM sleep (n = 7). In all cases, Pmus was substantially downregulated with increase in assist so that the increase in VT, although significant (P < 0.05), was small 0.08 liter at the highest assist). There was no difference in response between REM and non-REM sleep. We conclude that respiratory drive during sleep is dominated by chemical control and that there is no fundamental difference between REM and non-REM sleep in this regard. REM sleep appears to simply add bidirectional noise to what is basically a chemically controlled respiratory output.