Risk Factors for Delirium Superimposed on Dementia in Elderly Patients in Comprehensive Ward.

Abstract

Objective: To investigate the incidence of delirium and its related risk factors in patients with senile dementia during hospitalization.

Methods: A retrospective analysis of clinical data of 157 patients over 65 with cognitive impairment who were hospitalized in the comprehensive ward from October 2019 to February 2023 was conducted. Patients were assigned into delirium and non-delirium groups according to whether they exhibited delirium during hospitalization. General information about the patients and Visual Analogue Scale (VAS) score, blood C-reactive protein level, and blood superoxide dismutase (SOD) level were recorded. Univariate analysis was used to identify potential risk factors for delirium, and factors with statistical significance were subjected to multivariate logistic regression analysis. A prediction line chart for delirium in elderly dementia patients was constructed using R 4.03 software, and the model was validated.

Results: Among the 157 patients with senile dementia, 42 patients exhibited delirium and 115 patients exhibited non-delirium. Multivariate logistic regression analysis showed that diabetes, cerebrovascular disease, VAS score ≥4 points, use of sedative drugs, and blood SOD <129 U/mL were independent risk factors for delirium during hospitalization in elderly dementia patients. A prediction nomogram was plotted based on the five risk factors, and receiver operating characteristic curve analysis presented an area under the curve of .875 (95% CI: .816-.934). The nomogram model was internally validated by the Bootstrap method, and the calibration curve showed good agreement between predicted and actual results. Hosmer-Lemeshow test demonstrated that the model had a good fit and high predictive ability.

Conclusion: Diabetes, cerebrovascular disease, VAS ≥4 points, use of sedative drugs, and blood SOD <129 U/mL were independent risk factors for delirium in patients with senile dementia during hospitalization. The nomogram model had good accuracy and clinical application value for predicting delirium in this study.