Abstract
Background: Postoperative delirium (POD) is a common surgical complication that increases hospital stay duration, hospitalization costs, readmission rates and mortality. This study aims to describe the incidence of POD in an elderly patient population and to investigate pain assessment as a risk factor for postoperative confusion. Additionally, we aim to determine a predictive model for POD. Methods: We conducted a retrospective, observational, single-center study at La Timone Hospital between September 2020 and September 2021. We included patients aged 65 or older, admitted for scheduled or emergency surgical care, with an expected postoperative stay of three days or more and no history of dementia. Data were collected in three stages of hospitalization: preoperative, perioperative, and postoperative. Preoperative data included medical history and lifestyle; perioperative data included surgical and anesthesia characteristics, and postoperative data were collected. Results: Of the 531 patients admitted for surgical care, we analyzed 109 patients. Among these, 24 (22%) experienced a POD episode within the first three postoperative days. Age, preoperative cognitive impairments, preoperative sensory deficits, and long-term benzodiazepine use were identified as risk factors for POD. A significant difference in pain levels was also observed for all NRS scores during the first three postoperative days. After multivariate analysis, we retained two predictive models for POD. Conclusions: This study identified risk factors for POD and proposed predictive models based on these factors. Two models were particularly notable for their potential use in anesthesia consultations and patient follow-up services to quickly detect patients at risk of POD.