Pressure Support May Interfere with Sleep

Abstract & Commentary

Synopsis: Sleep was more fragmented during ventilation with PSV than A/C in a small group of critically ill patients studied with polysomnography. In 6 of the 11 patients studied, central apneas associated with arousals developed while receiving ventilation with pressure support but not while receiving A/C with a backup rate.

Source: Parthasarathy S, et al. Effect of ventilator mode on sleep quality in critically ill patients. Am J Respir Crit Care Med. 2002;166:1423-1429.

Eleven male patients requiring mechanical ventilation were studied during sleep in an ICU under 3 different ventilator conditions. Patients ranged from 49 to 90 years old and suffered from a variety of medical conditions including 6 who experienced congestive heart failure (CHF). The patients were monitored with conventional EEG, rib and abdominal motion sensors, capnography, pulse oximetry, and airway pressure tracings, and sleep patterns were manually scored during 2-hour periods of sleep under each ventilatory condition. Assist-control ventilation (A/C) was adjusted to produce a tidal volume of 8 mL/kg, and the backup rate was set at 4 breaths/min less than each patient’s own rate during quiet (awake) breathing. The pressure support (PSV) level was adjusted to produce the same tidal volume (8 mL/kg), and no backup rate was used. In the third ventilatory condition, 100 mL of dead space was added during PSV.

Fragmentation of sleep was common in all patients during all modes of ventilation. Only 4 patients achieved stage 4 sleep (rapid eye movement, [REM]) during any mode, and only 1 achieved REM sleep in all modes. Arousals were the same in PSV and A/C (39 and 35/h) but actually awakenings were more common with PSV (without dead space) than with A/C (39 and 19/h). Adding dead space reduced the number of apneas and awakenings. Five of the 6 patients who developed central apneas on PSV had a diagnosis of CHF. In the 5 patients who did not develop apnea, there was no difference in arousals and awakenings between the different modes of ventilation.

Patients who exhibited the greatest difference in wakefulness during the different modes of ventilation had the lowest ventilatory rates and minute ventilations at rest. Tachypneic patients exhibited little change in wakefulness on any mode. Parthasarathy and colleagues suggest that PSV contributes to wakefulness by producing apneas, which allow more variation in the PCO2 during sleep than with A/C with a backup rate. Since there was a slightly higher end tidal PCO2 in the PSV alone group, it was not simply arousal caused by this high PCO2 that accounted for the differences. Adding the additional dead space returned the patients to the A/C sleep pattern by preventing the apneas.

Comment by Charles G. Durbin, Jr., MD

The greatest value of this study is to emphasize the fact that sleep was not normal for any patient studied. Arousals and awakenings were frequent in all patients. Differences in quality of sleep with different ventilatory modes were most apparent in patients with CHF, a condition known to predispose to periodic breathing patterns.

I have several concerns with this study. Although patients recovering from anesthesia and on vasopressors were specifically excluded, use of sedative drugs was not mentioned in the text. There is a footnote on one of the paper’s tables stating that they were all receiving sedatives, but no information about the specific drug classes or doses. There is additional information in the electronic version of this paper that may answer this question but one must pay an electronic subscription rate to gain access to it, and neither I nor my library have such a subscription. This is a very important issue as sedative drugs and narcotics can profoundly affect ventilation and sleep patterns.

Another concern with this paper is that although Parthasarathy et al state that the level of PSV was selected to produce the same 8 mL/kg tidal volume during quiet breathing as during A/C, this may not have been true. Although not statistically different, it appears that the baseline tidal volume was larger during PSV. This difference became even greater during sleep, increasing to almost twice the desired level, suggesting that more pressure was applied than was necessary initially and patients were actually actively resisting this mode of ventilation. The actual level of pressure support for each patient is not stated but the average can be inferred to be between 17 and 20 cm H2O. It may be that simply less pressure support (ie, a more appropriate level of PSV) would have eliminated any differences seen in wakefulness.

While this paper raises some important and interesting issues related to sleep disturbances in ICU patients, many important details of the study are missing. It would be a mistake to abandon PSV in favor of A/C for presumed sleep benefits in all patients, based on these limited results. However, in those patients with CHF a backup rate may offer some benefit.

Dr. Durbin is Professor of Anesthesiology Medical Director of Respiratory Care University of Virginia.