Depressive symptoms as a risk factor for postoperative delirium in older adults: A systematic review and meta-analysis.

Abstract

BACKGROUND: Post-operative delirium (POD) in older adults may be linked to preoperative depression via shared biological pathways. This review synthesizes current findings, highlights methodological gaps and suggests possible pathophysiological mechanisms. METHODS: Searches across major databases identified eligible studies using validated tools. Standardized mean differences (SMD) and log odds ratios were pooled using random-effects models. Heterogeneity, moderators, and publication bias were assessed using meta-regression and sensitivity analyses in R (metafor package). RESULTS: This meta-analysis synthesized findings from 42 studies examining the association between preoperative depression and postoperative delirium (POD), analysing categorical (k = 16) and continuous (k = 26) reported depression data from 9701 patients. Preoperative depression was significantly associated with increased POD risk (OR = 2.50, 95% CI: 1.90-3.28; SMD = 0.49, 95% CI: 0.26-0.73 after outlier removal). Surgical cohort type and depression assessment tools significantly moderated this association. Cognition evaluated by the MMSE did not significantly influence outcomes. Geriatric depression scale (GDS-15) emerged as a consistent predictor, prompting further investigation into optimal cutoff scores for POD risk stratification. Publication bias and heterogeneity were low to moderate. DISCUSSION: This meta-analysis confirms preoperative depression, including subclinical symptoms, as a significant predictor of POD. Findings highlight the roles of surgical type and depressive symptoms, suggesting tailored screening and further research into symptom-specific risk pathways. CONCLUSION: Preoperative depression predicts POD. Notably, even GDS-15 scores below the conventional cutoff for depression were associated with increased delirium risk. This finding requires further investigation, including exploration of potential underlying mechanisms such as vascular or other symptom-specific pathways.