Electroencephalography for Detecting and Predicting Postoperative Delirium: A Systematic Review of Clinical Utility

Postoperative delirium (POD) is a major, preventable, neurocognitive complication in geriatric surgical patients, associated with higher complications and prolonged hospitalization. Electroencephalography (EEG) abnormalities, particularly generalized slowing, reduced alpha power, and burst suppression, have long been linked to delirium and may identify brains that are particularly susceptible to POD, or “vulnerable brains”. Recent evidence suggests that intraoperative EEG patterns predict POD risk, while portable and automated EEG systems offer emerging tools for detection and monitoring. This systematic review synthesized findings across the perioperative phases to evaluate the EEG’s role in predicting, diagnosing, and managing POD. Thirty-one studies were analyzed, spanning pre-, intra-, and postoperative EEG applications. Most cohorts included adults aged 60 or older undergoing cardiac, general, orthopedic, or neurovascular procedures. Consistent EEG markers of delirium vulnerability included reduced alpha power and peak frequency, increased delta/theta activity and burst suppression, and decreased spectral edge frequency and entropy. Predictive accuracy ranged from AUC 0.70 to 0.90, with most PODs occurring within 0–48 hours postoperatively. Across studies, EEG signatures: low alpha activity and prolonged burst suppression, preceded clinical symptoms. This supports EEG as an early, objective biomarker of cortical fragility. While single-channel systems improved feasibility, raw and quantitative EEG offered superior sensitivity. Standardized protocols, multicenter validation, and integration with perioperative care systems are needed to translate the EEG-guided monitoring into delirium prevention strategies.