Objectives: To compare the influence of three ventilatory modes on sleep.
Design: Prospective, comparative, crossover study.
Setting: Medical intensive care unit in a university hospital.
Patients: Fifteen conscious, nonsedated, mechanically ventilated patients.
Interventions: Patients were successively ventilated with assist-control ventilation, clinically adjusted pressure support ventilation (cPSV), and automatically adjusted pressure support ventilation (aPSV). Sleep polysomnography was performed during three consecutive 6-hr periods, one with each mode in random order. Airway pressure and thorax and abdomen plethysmography were used to diagnose central apneas and ineffective efforts.
Measurements and main results: The main abnormalities were a low percentage of rapid eye movement (REM) sleep counting, for a median (25th-75th percentiles) of 10% (3.5-12.5) of total sleep, and a highly fragmented sleep with 29 arousals and awakenings per hour of sleep. REM sleep duration was similar in the three ventilatory modes, 7% in assist-control, 4% in aPSV, and 1% during cPSV (p = .54), as well as in the fragmentation index, 31 arousals and awakenings per hour in assist-control, 32 in aPSV, and 34 during cPSV (p = .62). Ineffective efforts occurred similarly with the three modes (seven per hour of sleep in assist-control, 16 in aPSV, and 12 during cPSV) or central apneas during PSV (five in aPSV, seven during cPSV). Minute ventilation was similar with the three modes.
Conclusions: In conscious, mechanically ventilated patients, sleep architecture was highly abnormal, with a short REM stage and a high degree of fragmentation. The ventilatory mode did not influence sleep pattern, arousals, awakenings, and ineffective efforts.