Objective: To investigate the correlation between electroencephalogram (EEG) complexity and postoperative delirium (POD) in elderly patients undergoing orthopedic surgery.
Methods: Elderly patients, aged ≥ 65 years, ASA physical status Ⅰ-Ⅳ, scheduled for elective orthopedic surgery from January to August 2024 were enrolled. Continuous intraoperative EEG data were collected, and nonlinear EEG features were extracted using custom scripts in MATLAB. Multivariate logistic regression and convolutional neural network (CNN) were used to explore the predictive value of nonlinear EEG features for POD. Receiver operating characteristic (ROC) curve was plotted, and the area under the curve (AUC) was calculated to evaluate predictive performance.
Results: A total of 265 patients were included, among whom 31 (11.7%) developed POD. Multivariate logistic regression analysis showed that decreased permutation Lempel-Ziv complexity (PLZC), a nonlinear EEG metric (OR = 7.39, 95% CI 1.59-34.38, P = 0.011), and advanced age (OR = 1.16, 95% CI 1.08-1.25, P < 0.001) were positively associated with occurrence of POD in elderly patients undergoing orthopedic surgery. The AUC of decreased PLZC for predicting POD was 0.60 (95% CI 0.50-0.70). After incorporating time factors, a reduction in nonlinear EEG characteristic index PLZC achieved moderate predictive performance for POD (AUC = 0.74, 95% CI 0.64-0.84).
Conclusion: Reduced electroencephalographic complexity in elderly patients undergoing orthopedic surgery is correlated with POD and can predict the occurrence of POD. |
