The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
For Ventilated Patients, Is Sleep Better with PAV Than with Pressure Support?
By Dean R. Hess, PhD, RRT, Respiratory Care, Massachusetts General Hospital, Department of Anesthesiology, Harvard Medical School, Boston, is Associate Editor for Critical Care Alert.
Dr. Hess reports no financial relationship to this field of study.
Synopsis: Patient-ventilator dys-synchrony causes sleep disruption. Proportional assist ventilation may be more efficacious than pressure support ventilation in matching ventilatory requirements with ventilator assistance, resulting in less patient-ventilator dys-synchrony and better quality of sleep.
Source: Bosma K, et al. Crit Care Med. 2007; 35:1048-1054.
The objective of this study was to evaluate the role of patient-ventilator dys-synchrony in the etiology of sleep disruption, and to determine whether optimizing patient-ventilator interactions by using proportional assist ventilation (PAV) improves sleep. It was a randomized crossover clinical trial that enrolled 13 patients during weaning from mechanical ventilation. Patients were randomized to receive pressure-support ventilation (PSV) or PAV on the first night and then crossed over to the alternative mode for the second night. Polysomnography and measurements of light, noise, esophageal pressure, airway pressure, and flow were performed from 10 PM to 8 AM. Ventilator settings (pressure level during PSV, and resistive and elastic proportionality factors during PAV) were set to obtain a 50% reduction of the inspiratory work (pressure time product per minute) performed during a spontaneous breathing trial.
Arousals per hour of sleep time during PSV and PAV were 16 (range 2-74) and 9 (range 1-41), respectively (p = 0.02). Overall sleep quality was significantly improved on PAV (p < 0.05) due to the combined effect of fewer arousals/hour, fewer awakenings/hour (3.5 [range, 0-24] vs 5.5 [1-24]), and greater rapid eye movement (9% [range, 0-31] vs 4% [0-23]), and slow wave sleep (3% [range, 0-16] vs 1% [0-10]). Tidal volume and minute ventilation were lower on PAV, allowing for a greater increase in PaCO2 during the night. Patient-ventilator dys-synchronies/hour were lower with PAV than with PSV (24 ± 15 vs 53 ± 59; p = 0.02) and correlated with the number of arousals/hour (R2 = 0.65, p = 0.0001).
It has been increasingly appreciated in recent years that abnormalities of sleep are common in critically ill patients. Measures to improve the quantity and quality of sleep in critically ill patients include attention to mode of mechanical ventilation, decreasing noise, and sedative agents.1 I first appreciated this from papers published by Meza et al2 and Parthasarathy and Tobin.3 The results of both of these papers showed that PSV induces central apneas during sleep. In a study of 11 critically ill patients during 1 night of sleep, Parthasarathy and Tobin3 observed greater sleep fragmentation during PSV than during assist-control ventilation. Central apneas were more common during PSV than assist-control ventilation. In this study, the most important determinant of apneas was the difference between PCO2 during resting breathing and the apnea threshold. When the resting PCO2 was close to the apnea threshold, central apneas were more likely to develop. In other words, hypocapnea occurs during wakefulness with PSV, and lack of that wakefulness drive with sleep results in apnea in the absence of a back-up rate. Toublanc et al4 recently reported a study of 20 patients randomized to assist-control ventilation or PSV with 6 cm H2O. Assist-control ventilation was significantly associated with a better sleep quality than those recorded during PSV, again suggesting that a back-up rate may improve sleep quality during mechanical ventilation.
PAV is a spontaneous breathing mode in which the ventilator applies pressure in proportion to the inspiratory effort. This differs from PSV, in which the ventilator applies the same pressure regardless of inspiratory effort. During PAV, patient-ventilator synchrony may be optimized since both the amplitude and time course of ventilator assistance are linked to the amplitude and time course of inspiratory effort.
In this study, PSV was set at the level of pressure required to obtain a 50% decrease in pressure-time product (PTP) per minute relative to the values obtained during spontaneous breathing. Values of resistance and elastance obtained during spontaneous breathing were adjusted to obtain a PTP/min equal to 50% of the value obtained during spontaneous breathing. On PSV, 9.2 ± 2.8 cm H2O of ventilator-applied pressure was required to achieve a 54 ± 3% reduction in the inspiratory muscle load relative to spontaneous breathing. On PAV, the 53 ± 5% reduction of the PTP/min was obtained by setting the flow assistance to 5.8 ± 2.9 cm H2O/L/sec and the volume assistance to 8.9 ± 2.6 cm H2O/L. PEEP and FIO2 were set equivalently in both modes at 5.5 ± 0.2 cm H2O and 0.37 ± 0.05, respectively.
The ratio of the pressure applied to airway (airway opening PTP) to the pressure generated by the respiratory muscles (esophageal PTP) correlated significantly with the number of arousals per hour (R2 = 0.71) and the number of patient-ventilator dys-synchronies was regardless of the ventilatory mode (R2 = 0.52). In other words, the greater the proportion of ventilator support relative to patient effort, the more likely was patient dys-synchrony and sleep disordered breathing. It follows that there was as strong association between sleep disordered breathing and dys-synchrony. Morning minute ventilation was higher and PaCO2 was lower with PSV than PAV.
In this study, dys-synchrony and sleep disordered breathing were less with PAV than with PSV. Auto-triggering, ineffective triggering, and delayed cycling were more prevalent during PSV than PAV. It is unclear why auto-triggering should be more prevalent during PSV if the trigger sensitivity is set correctly, and the authors provide no explanation for this finding. Ineffective triggering may be the result of increased support during PSV which lowers respiratory drive, although no index of respiratory drive was reported. Delayed cycling during PSV is most likely due to the fixed 25% flow cycle with this mode on the ventilator used in this study.
An important finding is that setting support based on wakefulness criteria might result in excessive support during sleep. Although this was less likely during PAV, it can occur regardless of mode. Moreover, patient-ventilator dys-synchronies may be reduced during PSV by tailoring the trigger sensitivity, rise time, and cycling-off criteria to suit the respiratory mechanics and breathing pattern of the individual patient and then adjusting these variables as necessary to compensate for changes during sleep and wakefulness.5 Lowering the level of inspiratory assistance may decrease the amount of dys-synchrony and improve sleep quality. Some clinicians increase the level of support at night, which may ironically lead to dys-synchrony and sleep-disordered breathing. Although not evaluated in this study, setting a backup rate (assist-control ventilation) may also decrease the occurrence of patient-ventilator dys-synchrony and sleep disordered breathing.
In conclusion, sleep disordered breathing during mechanical ventilation is related to patient-ventilator dys-synchrony. With spontaneous breathing modes, a higher level of ventilator assistance increases the likelihood of dys-synchrony and arousals during sleep. Approaches to this problem include the use of PAV rather than PSV, appropriately setting rise time and cycle-off criteria during PSV, decreasing the amount of inspiratory support, and using a mode with a back-up rate such as assist-control.