Prediction of Postoperative Delirium in the Elderly Using Imaging Markers of Cerebral Small Vessel Disease

Background Postoperative delirium (POD) is common in elderly surgical patients and has significant implications for patient recovery. This study aimed to investigate the association between cerebral small vessel disease (CSVD) imaging markers and the occurrence of POD in elderly surgical patients. Methods Eighty-two elderly patients who had undergone surgery were retrospectively examined. The case group comprised 41 patients who developed POD, while the control group consisted of 41 who did not. The groups were matched 1:1 with respect to age, sex, surgery type, and operation duration. Preoperative brain magnetic resonance imaging was analyzed for CSVD markers, including white matter hyperintensities (WMHs), perivascular spaces (PVSs), lacunar infarcts, and global cortical atrophy. Logistic regression was used to assess the relationship between CSVD markers and POD. Results WMHs, PVSs, and lacunar infarcts significantly differed between the groups. Univariate analysis showed no significant effect for individual CSVD markers on POD. However, deep WMHs and white matter PVSs emerged as the most important independent predictors after stepwise multivariate regression. The final model showed that deep WMHs (odds ratio, 2.34; 95% confidence interval, 1.45–3.77) and white matter PVSs (odds ratio, 2.05; 95% confidence interval, 1.30–3.24) were independent predictors of POD, with a combined area under the receiver operating characteristic curve of 0.724. Conclusions Preoperative assessment of CSVD markers, particularly deep WMHs, and white matter PVSs, is valuable for predicting POD risk in elderly patients. These findings highlight the value of integrating neuroimaging markers into perioperative risk stratification and call for validation in larger prospective studies.