Effect of ventilatory variability on occurrence of central apneas.

OBJECTIVE: To compare the influence of 2 ventilation strategies on the occurrence of central apneas. METHODS: This was a prospective, comparative, crossover study with 14 unsedated subjects undergoing weaning from mechanical ventilation in the medical ICU of Hopital du Sacre-Gem, Montreal, Quebec, Canada. The subjects were ventilated alternately in neurally adjusted ventilatory assist (NAVA) and pressure support ventilation (PSV) modes. Inspiratory flow/time and pressure/time waveforms and diaphragmatic electrical activity were used to detect central apneas. Ventilatory variability and breathing pattern were evaluated in both modes. Breathing patterns just before central apneas, and associations between apneas and sleep patterns (electroencephalogram) were studied. RESULTS: Switching from PSV to NAVA did not change mean minute ventilation, tidal volume, or breathing frequency. However, tidal volume variability, defined as the coefficient of variability (standard error/mean), was significantly greater with NAVA than with PSV (17.2 +/- 2.8 vs 10.3 +/- 4, P = .045). NAVA induced a greater decrease in central apneas, compared to PSV (to 0 with NAVA vs 10.5 +/- 11 with PSV, P = .005). Central apneas during PSV were detected only during non-rapid-eye-movement sleep. CONCLUSIONS: NAVA was associated with increased ventilatory variability, compared to constant-level PSV. With NAVA the absence of over-assistance during sleep coincided with absence of central apneas, suggesting that load capacity and/or neuromechanical coupling were improved by NAVA and that this improvement decreased or abolished central apneas.