Noise-sensitivity during sleep in non-sedated mechanically ventilated patients is associated with weaning duration: a polysomnographic study.

Background: Poor sleep is a major concern in intensive care units (ICUs), particularly in mechanically ventilated patients, because it is associated with longer duration of the weaning phase and higher mortality. High noise levels in ICUs are frequently reported by patients as one of the most disturbing sleep-disrupting factors but would be responsible for less than 20% of arousals. This suggests major inter-individual variability in noise sensitivity. Our objectives were to define and assess noise sensitivity in mechanically ventilated patients and to explore its association with sleep duration, sleep quality and weaning duration.

Methods: We retrospectively re-analyzed polysomnographies (PSGs) recorded in 29 non-sedated patients, mechanically ventilated for at least 24 h and difficult to wean (i.e. ≥ 1 spontaneous breathing trial failure). All the arousals were identified on all the PSGs. We calculated mean noise level and identified all noise peaks (an abrupt increase of noise intensity of more than 10 decibels (dBA)) preceding each arousal. Each 21-second period preceding each arousal was divided into seven 3-second bins. We built a pre-event time histogram for each PSG by counting the total number of noise peaks in each bin. If the total number of noise peaks in one bin exceeded the average of the seven bins plus 2 SD, we considered that there was a significant relationship between the number of noise peaks in this bin and the arousal. The patient was then considered as noise-sensitive. Presence of atypical sleep, proportion of sleep stages, and weaning duration were assessed.

Results: Nineteen out of 29 patients (66%) were noise-sensitive. Duration of weaning from ventilator was significantly longer in noise-insensitive patients (median [interquartile range] 2 [1-2] versus 5 [2-8] days; p < 0.01). Proportion of N1, N2, N3 sleep stages and rapid eye movement sleep were similar in noise-sensitive and noise-insensitive patients. In contrast, the proportion of patients displaying atypical sleep was higher in noise-insensitive patients.

Conclusion: Our results report for the first time that most ICU patients were noise-sensitive. Lower noise sensitivity was associated with atypical sleep and could reflect lower brain reactivity to environment.