To identify specific differences in sleep architecture and circadian rhythms between older hospitalized patients with and without delirium.

In this prospective cohort study, we collected and analyzed 513 single-channel electroencephalogram (EEG) recordings, each approximately 24 hours in duration, from 208 hospitalized patients aged 50 years and older (n = 208; 134 non-delirious, 74 delirious). We defined delirium using the 3-minute diagnostic interview for the Confusion Assessment Method (3D-CAM). Sleep-wake activity was assessed in 30-second epochs using multiple automated sleep staging models. 

Across both automated sleep staging models, patients with delirium exhibited significantly greater total sleep duration compared to patients without delirium (linear mixed effects model (LMEM); U-Sleep: t=6.58, p<0.001; Luna: t=6.04, p<0.001). Patients with delirium also spent significantly more time in N2 sleep (Estimated Marginal Means; U-Sleep: t=8.10, p<0.001; Luna: t=19.80, p<0.001 ), had more N2-to-REM transitions (LMEM U-Sleep: t = 6.03, p < 0.001; LMEM Luna: t = 4.76, p < 0.001), and had greater intervals of uninterrupted sleep (LMEM U-Sleep: t = 4.17, p < 0.001; LMEM Luna: t = 5.95, p < 0.001).

Delirium status is associated with changes in sleep quantity, quality, and structure. Ongoing work aims to examine specific circadian patterns to determine how the timing and distribution of sleep across the day may differ between delirious and non-delirious patients. Identifying specific circadian disruptions in sleep architecture may guide targeted interventions for improving sleep onset and quality, potentially improving outcomes in patients with delirium.