Abstract
BACKGROUND: Postoperative delirium (POD) is a common and serious complication in elderly surgical patients, particularly among the oldest-old (aged ≥ 75 years). Evidence regarding clinical factors associated with POD in this population remains limited, especially across different healthcare settings. This study examined patterns of clinical characteristics associated with POD in two surgical cohorts analyzed in parallel: a national US surgical database and a single-center Taiwanese cohort. METHODS: We conducted a retrospective cohort study using data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) and Taipei Medical University Hospital (TMUH). Patients aged ≥ 75 years who underwent surgery under general anesthesia and entered a POD assessment process as part of routine clinical care were included. Preoperative clinical variables and laboratory parameters were analyzed. Univariate and multivariable logistic regression analyses were performed separately for each cohort to examine associations with POD among clinically evaluated patients. RESULTS: A total of 2598 patients were included in the analytic population (ACS-NSQIP: n = 1316; TMUH: n = 1282), among whom 212 (8.16%) were identified with POD. The proportion of POD identified on screening was higher in the ACS-NSQIP cohort (11.17%) than in the TMUH cohort (5.07%). In the ACS-NSQIP cohort, higher American Society of Anesthesiologists (ASA) physical status, abnormal preoperative serum albumin, and abnormal white blood cell count were independently associated with POD. In the TMUH cohort, older age and higher ASA class were significantly associated with POD. Associations involving serum albumin in the TMUH cohort should be interpreted cautiously due to substantial missingness, which likely reflects nonrandom laboratory testing practices rather than biological effects. Differences in patient characteristics between cohorts included age, body mass index, operative time, and laboratory profiles. CONCLUSION: Among oldest-old surgical patients who underwent POD assessment in routine clinical practice, patterns of clinical factors associated with POD differed across the two cohorts analyzed. These findings should be interpreted in the context of differences in screening practices, case-mix, and data completeness, and do not represent population-based incidence estimates. Prospective studies incorporating standardized delirium screening protocols are needed to support the development of robust and generalizable risk stratification approaches.