Pilot prospective study of post-surgery sleep and EEG predictors of post-operative delirium.

Objective: Delirium is a common post-operative complication associated with significant costs, morbidity, and mortality. We sought sleep/EEG predictors of delirium present prior to delirium symptoms to facilitate developing and targeting therapies.

Methods: Continuous EEG data were obtained in 12 patients post-orthopedic surgery from the day of surgery until delirium assessment on post-operative day 2 (POD2).

Results: Diminished total sleep time (r=-0.68; p<0.05) and longer latency to sleep onset (r=0.67; p<0.05) on the first night in the hospital were associated with greater POD2 delirium severity. Patients experiencing delirium slept 2.4h less and took 2h longer to fall asleep. Greater waking EEG delta power (r=0.84; p<0.05) on POD1 and less non-REM sleep EEG delta power (r=-0.72; p<0.05) on night 2 also predicted POD2 delirium severity.

Conclusions: Loss of sleep on night1 post-surgery is an early predictor of subsequent delirium. EEG Delta Power alterations in waking and sleep appear to be later indicators of impending delirium. Further work is needed to evaluate reproducibility/generalizability and assess whether sleep loss contributes to causing delirium.

Significance: This first study to prospectively collect continuous EEG data for an extended period prior to delirium onset identified EEG-derived indices that predict subsequent delirium that could aid in developing and targeting therapies.