By Sree Roy

The irony of sleep health, or the lack thereof, in the intensive care unit (ICU) is inescapable.

While healthy people become obsessed with achieving the perfect “sleep score” per their wearables, people who are critically ill—and so presumably the most in need of quality slumber—find their sleep disrupted not only by their disease but also by its monitoring, treatments, and the overall ICU environment.

The field of sleep medicine can improve ICU patients’ outcomes during their hospital stays and even after discharge. But first, we need progress on at least three key fronts:

  1. defining “sleep” in the ICU,
  2. connecting sleep biomarkers to specific outcomes, and
  3. having ICU-feasible sleep study equipment.

Though clinical integrations are years away, researchers report exciting updates. Sleep-related interventions could one day lower the risks of failed ventilator weaning attempts, delirium, and post-intensive care syndrome.

Atypical Sleep in the ICU

Defining sleep in the ICU is a challenge “because there are so many other things happening that cause the brain to have different electrical signals than you might expect in a healthy person, like pharmacology, anesthesia, or their critical illness,” says Makayla Cordoza, PhD, RN, CCRN, at the School of Nursing at Vanderbilt University and an author of an American Thoracic Society research statement on sleep and circadian disruption in the ICU.1

Paula L. Watson, MD, and co-investigators identified six electroencephalogram (EEG) patterns to consider when staging sleep in ICU patients, including pathologic wakefulness (when the patient is behaviorally awake but the EEG does not show alpha or beta waves) and atypical 1 (alpha and/or theta in more than 10% of the epoch but without sleep spindles or K-complexes).2

Add to that unconsolidated sleep, brain waves that transition sleep stages more quickly than 30-second epochs, and rapidly changing patient health status, and it’s clear that sleep study technologies need to move beyond standard sleep lab equipment and onto devices comfortable for 24-hour monitoring with real-time auto-scoring and ICU-specific software algorithms.

Linking Sleep Markers to ICU Outcomes

Ventilation Weaning

An exciting area of research is the potential for sleep biomarkers to predict the outcome of weaning off mechanical ventilation.

For example, one study finds that the higher the fraction of monitoring time spent with the Odds Ratio Product—a proprietary continuous index, available in Cerebra systems, that is used to evaluate sleep depth—greater than 2.2, as well as high correlation between the left and right brain hemispheres, the greater the likelihood of a successful spontaneous breathing trial and extubation.3

In another study, Mount Sinai-based pulmonologist-sleep specialist Raymonde E. Jean, MD, FCCP,  FAASM, and co-investigators compared the sleep quality in critically ill patients sedated by different drugs. Propofol was linked with higher sleep duration and less disrupted sleep architecture compared to fentanyl, propofol and fentanyl, or no sedation. “That study provided important evidence to suggest that the selection of a particular sedative can have an impact on outcome,” Jean says, though more evidence is needed to change the standard of care.4  

Pulmonologist-sleep specialist Hameeda S. Shaikh, MD, of Edward J Hines VA Hospital and co-investigators studied sleep in patients discharged from the ICU to a long-term acute care facility for ventilation weaning. In many cases, brain dysfunction, as evidenced by atypical sleep, followed the patients “far beyond the onset of the initial critical illness in the acute ICU and was associated with weaning failure,” Shaikh says. “This indicates that brain dysfunction, demonstrated by atypical sleep and wakeful EEG, may be an important factor in the pathogenesis of failure to wean from mechanical ventilation.”5

They also determined that REM sleep was less prominent in patients with atypical sleep. “This finding allowed us to speculate as to a possible mechanistic link between atypical sleep and weaning failure, as selective REM sleep deprivation is associated with a hypermetabolic state and total body catabolism,” she adds. 


Detecting delirium and subsyndromal delirium states is another area where sleep markers could prove useful. But “EEG findings cannot supplant the clinical assessment when diagnosing delirium,” Shaikh says, especially since a clinician can quickly diagnose delirium by simply talking with the ICU patient.

Yet, anesthesiologist and intensivist Alawi Luetz, MD, of the Universitatsmedizin Berlin Technical University, says, “For deeply sedated patients where a clinical examination-based delirium diagnosis is not possible, EEG data could help guide therapy in the future.”

Sleep EEG markers for ICU delirium include abnormal/atypical N3 (polymorphic delta or sepsis-associated encephalopathy). A study by Jean and co-investigators found the EEG patterns that comprise atypical N3 sleep were associated with delirium, sepsis, and mortality. “Once this relationship is definitively established, then improved detection of hypoactive delirium could be made in patients who are unresponsive to assessment with the CAM-ICU,” Jean says.6

Luetz’s research team is focused on developing nonpharmacological solutions to prevent delirium. This includes an ICU room concept with a ceiling that provides dynamic light therapy for circadian alignment. “The first published study results indicate that this new ICU room concept could reduce the incidence of delirium and influence circadian melatonin rhythms,” he says. “We are analyzing the data to see if the room modifications and light therapy are also associated with changes in sleep quality.”7 

Post-Intensive Care Syndrome

Luetz’s team is also studying what happens to patients after they survive a critical illness.

“We see an increasing percentage of ICU patients who continue to demonstrate disabilities after discharge, a condition known as post-intensive care syndrome,” he says. For example, patients with ICU delirium have a significantly higher probability of developing cognitive problems later.

Here too, better sleep could be part of the solution. “Sleep monitoring as part of a circadian bundle in intensive care could probably reduce the incidence and duration of delirium, thus increasing quality of life after critical illness,” Luetz says.

Obstacles to Clinical Implementation

Outside of research studies, ICU patients won’t have their sleep objectively monitored anytime soon. Cordoza, assistant professor at the Vanderbilt School of Nursing, thinks subjective screening is feasible now. 

“I could see some sort of addition, as part of our daily assessment as nurses and providers at the bedside, of asking patients about sleep, such as how can we improve their sleep and what do we think their sleep was like,” she says. 

Cordoza and co-investigators also evaluated the feasibility and acceptability of objective ICU sleep monitoring in a 20-patient study that used the Cerebra Sleep System. They now have an about 150-patient, three-year study ongoing with that system to study sleep and delirium. Two advantages of that system in the ICU are that it’s wireless and doesn’t need a WiFi connection to record, Cordoza says. However, their study only used the device at night (so it almost certainly missed much of the patients’ total sleep time), and the frontal-EEG device does not have electrodes specifically for the occipital region.8

Trade-offs are going to happen with any of today’s polysomnography technology.

Shaikh’s team used the Natus Sandman Elite, a traditional wired full polysomnography system, and found “nursing care, with frequent repositioning of the patient, often resulted in lead displacement,” she says. “This issue was best addressed by having the investigator be present in the room to assist with patient repositioning.”

Jean and Luetz used Advanced Brain Monitoring’s Sleep Profiler, a forehead-worn EEG with the advantage of real-time sleep staging. Both continue to use the device for ICU sleep research, but Jean points out that most ICU staff are unfamiliar with EEG signals. “Tools that simplify the characterization and interpretation of normal and abnormal sleep are needed, as well as the development of procedures to help the ICU staff know how to intervene when abnormal EEG/sleep patterns are detected,” she says.

Luetz adds, “Real-time EEG interpretation should not only be about sleep. These devices could help provide treatment algorithms and new ICU environments that help to provide a ‘brain protective state’—a bundle that protects the patient’s brain in a calming and stress-relieving cocoon, improving patients’ quality of life after critical illness.”

Perhaps in the future, the ICU could be one of the most nurturing environments for sleep health.


1. Knauert MP, Ayas NT, Bosma KJ, et al. Causes, consequences, and treatments of sleep and circadian disruption in the ICU: An official American Thoracic Society Research Statement. Am J Respir Crit Care Med. 2023 Apr 1;207(7):e49-68.

2. Watson, PL, Pandharipande P, Gehlbach BK, et al. Atypical sleep in ventilated patients: Empirical electroencephalography findings and the path toward revised ICU sleep scoring criteria. Crit Car Med. 2013 Aug; 41(8):1958-67.

3. Dres M, Younes M, Rittayamai N, et al. Sleep and pathological wakefulness at the time of liberation from mechanical ventilation (SLEEWE). A prospective multicenter physiological study. Am J Respir Crit Care Med. 2019 May 1;199(9):1106-15.

4. Jean R, Shah P, Yudelevich E, et al. Effects of deep sedation on sleep in critically ill medical patients on mechanical ventilation. J Sleep Res. 2020 Jun;29(3):e12894.

5. Shaikh H, Ionita R, Khan U, et al. Effect of atypical-sleep EEG patterns on weaning from prolonged mechanical ventilation. Chest. 2024 Jan 9:S0012-3692(24)00011-4. 

6. Genese F, Martillo M, Ventura I, et al. The influence of sepsis on sleep architecture in

the intensive care unit. Presented at Society of Critical Care Medicine meeting. 2016 Feb.

7. Luetz A, Uhrlau H, Piazena H, et al. Modification of ICU environment is associated with reduced incidence of delirium – Results from the VITALITY study. Intens Care Med Exprtl. 2018; 6(Suppl 2):1305.

8. Cordoza ML, Anderson BJ, Cevasco M, et al. Feasibility and acceptability of using wireless limited polysomnography to capture sleep before, during, and after hospitalization for patients with planned cardiothoracic surgery. J Cardiovasc Nurs. 2024 Mar 21.

Top photo: In Berlin, Germany, a research team led by Alawi Luetz, MD, developed a multimedia light ceiling for the ICU to provide dynamic light therapy for circadian alignment. Photography by Tobias Hein.