Sleep in critically ill, mechanically ventilated patients with severe sepsis or COPD

The standard method for scoring polysomnographic (PSG) sleep is insufficient in the intensive care unit (ICU). A modified classification has been proposed, but has not been tested in specific groups of ICU patients. We aimed firstly to (1) use the modified classification to describe sleep in two groups of ICU patients: a severe sepsis group and a chronic obstructive pulmonary disease (COPD) group, and (2) to compare sleep stage distribution in the groups; secondly to compare the PSG findings with nurses' sleep evaluation.

Non-sedated mechanically ventilated patients with severe sepsis or COPD completed up to 20-hours PSG recording in each patient. A modified classification for scoring sleep in ICU was used for scoring the PSGs. Sleep assessment by nurses was done at 15 minutes intervals.

We included 16 patients with severe sepsis and 17 patients with COPD. Half of the patients in the severe sepsis group and 59% in the COPD group had atypical sleep. We found significantly different sleep stage distribution in the two groups, with the COPD group having a higher proportion of atypical sleep (54.4% vs 48.7%, P < .0001). No correlation between nurse sleep assessment and PSG was found in cases of atypical sleep (P < .0001).

Normal PSG sleep characteristics as defined by standard classification are absent in many conscious, non-sedated critically ill patients on mechanical ventilation. Nurse sleep evaluation does not correlate with PSG if atypical sleep is present in the PSG, which limits the reliability of subjective sleep assessment in this patient population.