Sleep Patterns, Predictors, and Pharmacologic Interventions in the ICU

Many patients in the intensive care unit (ICU) struggle with sleep disturbance, but researchers are still learning how poor sleep affects recovery and outcomes.

At the American Thoracic Society (ATS) 2025 International Conference, 3 posters looked at different aspects of sleep in patients who are critically ill, from how well sleep medications like temazepam work to whether certain sleep patterns can help predict how long a patient will stay in the ICU or how likely they are to recover. Together, these findings suggest that sleep in the ICU is more than just a comfort issue—it may provide important clues about a patient’s overall health and chances of recovery.

Atypical N3 Sleep Patterns Linked to Longer ICU Stays

Researchers of an observational study evaluated the presence of atypical N3 (AN3) sleep, a delta wave pattern with minimal theta and sigma activity, among patients who are critically ill and its association with outcomes.1 AN3 has been previously linked to sepsis-associated encephalopathy and neurodegenerative disorders.

In one study, temazepam was linked to a modest increase in observed sleep duration. | Image credit: Gorodenkoff – stock.adobe.com

Using FDA-approved Sleep Profiler technology to record 24-hour sleep data, the study included 56 patients with sufficient sleep recording time. Twenty-five patients had more than 1.5 hours of AN3 sleep, which, compared with less than 1.5 hours, was significantly associated with:

• Older age (78.2 vs 60.8 years; P < .001)
• Lower Glasgow Coma Scale scores (8.9 vs 12.0; P = .006)
• Longer ICU stays (15.6 vs 7.5 days; P < .001)

Most notably, AN3 sleep was significantly associated with ICU stays longer than 10 days (OR, 1.33; 95% CI, 1.04-1.71; P = .023), suggesting its potential as a biomarker for prolonged ICU admissions. It also showed a near-significant trend toward higher APACHE II scores (P = .052), which had the second strongest correlation with these longer ICU stays (OR, 1.12; 95% CI, 1.03-1.3; P = .048).

While AN3 did not correlate with mortality, disposition, sepsis, or intubation status, the authors said this could be skewed by the small cohort size.

“Though this current study is limited to a single center, our findings suggest that AN3 may have a clinical role in the early recognition of outcomes in ICU patients, which warrants further investigation,” they said.

Hypersomnia May Signal Poorer Outcomes

In a sister study by Mount Sinai, researchers focused on hypersomnia, defined as more than 11 hours of sleep in a 24-hour period, and its association with patient outcomes in the ICU.2

Hypersomnia (n = 34) was significantly associated with worse clinical indicators than getting less than 11 hours of sleep (n = 22). Patients with hypersomnia were older (74 vs 61.3 years; P = .03) and had higher APACHE II scores (20.1 vs 15.1; P = .026). They experienced longer ICU (17.4 vs 10.9 days; P = .001) and hospital stays (26.9 vs 17.8 days; P = .004) and were less likely to be discharged home (14.7% vs 63.2%; P < .001). They also had higher rates of sepsis (67.6% vs 36.4%), intubation (70.6% vs 9.1%), and sedation (38.2% vs 0%).

Importantly, mortality was significantly higher among patients with hypersomnia than those without (50% vs 13.6%; P = .009), and logistic regression showed that ICU admission reason and intubation status were significant predictors of hypersomnia. While hypersomnia may not be an independent predictor due to confounding variables like age and illness severity, it shows promise as a marker of poor outcomes and may help identify high-risk patients early in the ICU stay, warranting further research.

Temazepam May Increase Sleep Duration

The randomized, blinded, placebo-controlled DREAM trial looked at how temazepam could improve patients’ sleep in the ICU.3 Conducted in Australia, the study included 26 patients who were critically ill receiving a nightly dose of 10 mg to 30 mg temazepam and 27 receiving placebo. The study only administered medication at 9:00 PM and measured sleep from 9:00 PM to 7:00 AM, leaving room for future research looking outside of this window of time.

Use of the benzodiazepine was associated with a modest increase in observed sleep duration. Patients taking the treatment slept a mean of 351 minutes vs 287 minutes in the placebo group—a 64-minute difference that narrowly missed statistical significance (P = .07). However, in an adjusted multivariable model accounting for sex, age, illness severity, and ICU stay length, the difference increased to 77 minutes and reached significance (P = .04).

Despite this improvement in objective sleep time, subjective sleep quality showed no significant differences between groups, whether assessed by nurses (57 vs 49; P = .15) or patients themselves (50 vs 51; P = .70). These findings suggest temazepam may modestly increase sleep quantity but may not meaningfully improve perceived sleep quality.

References

1. Choi C, Samuel B, Shah S, et al. Atypical N3 sleep pattern in critically ill patients and its role as a marker of outcome. Presented at: ATS 2025 International Conference; May 18, 2025; San Francisco, CA. https://www.atsjournals.org/doi/10.1164/ajrccm.2025.211.Abstracts.A3095
2. Samuel B, Choi C, Shah S, et al. Hypersomnia and outcomes in critically ill patients. Presented at: ATS 2025 International Conference; May 18, 2025; San Francisco, CA. https://www.atsjournals.org/doi/10.1164/ajrccm.2025.211.Abstracts.A3094
3. Showler L, Ali Abdelhamid Y, Ankravs M, et al. A parallel-group, blinded, placebo-controlled, randomised, pragmatic clinical trial investigating the effect of temazepam on objective and subjective measures of sleep in critically ill patients (the DREAM trial). Presented at: ATS 2025 International Conference; May 18, 2025; San Francisco, CA. https://www.atsjournals.org/doi/10.1164/ajrccm.2025.211.Abstracts.A3093#aff3