Risk factors for postoperative delirium in elderly patients after major orthopedic surgery: a prediction model

Background: This study mainly aims to assess risk factors of postoperative delirium (POD) in elderly patients after major orthopedic surgery, and construct and validate a simple delirium-predicting model. Methods: The plan was to collect 700 elderly patients admitted to Ningbo No. 6 Hospital for major orthopedic surgery in the analysis cohort. Associated risk factors were first screened by univariate analysis, and then the independent risk factors were screened by binary logistic regression analysis and assigned according to the magnitude of the odds ratio (OR), to establish a risk prediction score model. A cohort of 300 patients admitted to our hospital was included as a prediction cohort. The risk score prediction score model was constructed and validated by assignment according to OR values, i.e.. Results: We developed and verified a practical POD risk prediction score model for elderly patients undergoing major orthopedic surgery, which consists of six independent risk factors: age ≥73 years, cardiac disease, diabetes mellitus, sleep disorders, cortisol ≥411 nmol/L, and CRP ≥14 mg/L. The analysis cohort's results were accurately predicted by this model (AUROC=0.863; 78.4% sensitivity, 78.9% specificity,78.8%accuracy) whereas the prediction cohort's results were also high (AUROC=0.941; 95.2% sensitivity, 89.5% specificity, 90.5% accuracy). Conclusion: The risk prediction model for postoperative delirium, constructed in this study, is highly applicable, and age ≥73 years, cardiac disease, diabetes mellitus, sleep disorders, cortisol ≥411 nmol/L, and CRP ≥14 mg/L are the main risk factors in elderly patients undergoing major orthopedic surgery. Trial registration:Prospectively registered at the China Clinical Trial Registry, registration number: ChiCTR2200056234, date of registration:02/02/2022.