Preoperative and intraoperative EEG alpha power in relation to postoperative delirium: a systematic review and meta-analysis

Abstract Background Perioperative electroencephalogram (EEG) alpha power has been proposed as a marker of brain vulnerability and a candidate biomarker for postoperative delirium (POD). However, reported associations between preoperative or intraoperative EEG alpha power and POD remain inconsistent. Methods We performed a PRISMA-guided systematic review and meta-analysis of observational studies in adults (≥ 18 years) undergoing non-neurological surgery that reported resting-state preoperative or intraoperative EEG alpha power and POD assessed with validated tools. PubMed, Ovid MEDLINE, Embase, Web of Science, and the Cochrane Library were searched to December 26, 2025. Standardized mean differences ( Hedges's g ) between groups of POD and without POD were pooled using random-effects models. Results Twelve studies (1637 patients) met inclusion criteria. Intraoperative EEG alpha power was lower in patients who developed POD compared to those without POD ( g = −0.57, 95% CI [−0.92, −0.22], P = 0.006). Sensitivity analyses further supported the robustness of this association—after excluding a cardiac cohort with cognitive impairment, heterogeneity decreased ( I 2 reduced from 85.3% to 57.1%) and the result remained robust in an extended analysis ( I 2 = 0, g = −0.44). In contrast, preoperative alpha power showed substantial heterogeneity. Sensitivity analyses identified baseline cognition as an essential source of variability, and after adjustment, a significant association between higher preoperative alpha power and POD was observed ( I 2 = 0, g = 0.27, P = 0.037). Conclusions Lower intraoperative alpha power was consistently associated with subsequent POD across diverse surgical populations, supporting its potential as a marker of intraoperative cerebral vulnerability. In contrast, associations between preoperative alpha power and POD were variable and were influenced by baseline cognitive status. Further studies incorporating cognition alongside more standardized perioperative EEG are needed. Protocol registration Registered at PROSPERO on February 10, 2025 (CRD42024599991).