Abstract: Objective: To investigate the incidence and risk factors of postoperative delirium (POD) in elderly hip fracture patients and evaluate the predictive efficacy of these risk factors. Method: The subjects who met the inclusion criteria were evaluated by the self-designed postoperative delirium influencing factors questionnaire and the FRAIL scale before the operation, and the full version of Confusion Assessment Method was used to diagnose POD, and the clinical manifestations were recorded after the operation. The risk factors of POD were determined by multivariate Logistic regression analysis, and the receiver operating characteristic (ROC) curve was used to evaluate the predictive efficacy and determine the cut-off value. Results: POD occurred in 63 of 217 patients (29%). Multivariate analysis showed that age, preoperative frailty, stroke, intraoperative hypoxia, and length of ICU stay were risk factors for POD. ROC curve analysis showed that age, preoperative frailty and length of ICU stay were with moderate predictive efficacy, stroke and intraoperative hypoxia were with low predictive efficacy for POD. The cut-off value of age, ICU stay and frailty status were 80 years old, 1 day and pre-frail respectively. Conclusion: The incidence of POD is high in elderly hip fracture patients. Age over 80, preoperative pre-frail or frailty, a history of stroke, intraoperative hypoxia and ICU stay more than 1 day could predict the occurrence of POD.

Key words: aged; hip fracture; postoperative delirium; risk factor; predictive factor; ROC curve