Sleep duration and architecture in non-intubated intensive care unit patients: an observational study

Sleep disturbances are very common and associated with severe complications in patients admitted to intensive care units (ICU). Commonly, sedatives like propofol or benzodiazepines have been demonstrated to increase sleep duration but worsen sleep architecture. Dexmedetomidine seems to positively affect both sleep aspects.

The present study aimed to investigate sleep characteristics in non-intubated patients admitted to intensive care unit. The subgroups consisted of those without sedation (NO-DEX), and those which received dexmedetomidine infusion (DEX), titrated to a Richmond Agitation-Sedation Scale target of -1/-2, were also explored. An auto-staged electroencephalographer was used to measure sleep duration and architecture. The Richard-Campbell-Sleep questionnaire was administered to all patients.

A multivariate analysis conducted in the overall population showed that dexmedetomidine infusion was the only variable independently associated with N2 increase (p < 0.001). DEX (n = 36) versus NO-DEX (n = 36) group showed longer N2 stage [68.9% (57.5-80.9) versus 49.5% [35.7-61.4]; p < 0.0003]; longer Total Sleep Time [6.5 h (5.7-7.7) versus 3.4 h (1.8-4.9); p < 0.0001, and higher Sleep Efficiency [84.2% (71.3-92.6) versus 47.7% (23.4-60.9); p < 0.0001]; shorter N1 (percentage of Total Sleep Time) [10.5% (7.8-20.0) and 38.8% (25.6-50.3); p < 0.0001]; longer N3 stage [13.6% (1.9-23.3) versus 4.3% (0.4-14.0); p = 0.058]; fewer Cortical Arousals [15 episodes/hour (8.1-24.6) versus 48.7 episodes/hour (29.7-80.4); p < 0.0001]. The questionnaire showed better values in DEX-group in all explored items (p < 0.0001).

Abnormal sleep is common in intensive care unit patients who have not received sedation. Dexmedetomidine, titrated to reach an appropriate sedation level, may optimize sleep duration and architecture.