Risk factors and prediction of cognitive dysfunction in diabetes-related stroke

Abstract

Introduction: Type 2 diabetes mellitus (T2DM) with secondary cerebral infarction often leads to cognitive dysfunction (CD), impacting patients’ quality of life and prognosis. The aim of the study was to explore factors influencing CD in T2DM patients with cerebral infarction and develop a prediction model. Material and methods: This was a retrospective analysis of 244 T2DM patients with cerebral infarction treated from January 2020 to December 2023. Patients were split into a training set (n = 170) and a test set (n = 74). Logistic regression and random forest models were developed using RStudio. Results: Logistic regression analysis indicated that age, 25-hydroxyvitamin D [25(OH)D], and triglyceride (TG) were independent influencing factors for CD in patients. In the random forest model, the variables were prioritized based on their importance, with 25(OH)D ranked highest, followed by age, TG, National Institute of Health stroke scale (NIHSS) score, duration of T2DM, and diabetic neuropathy. The area under the receiver operating characteristic curve (AUC) for the Logistic model was 0.799 in the training set and 0.793 in the test set, while the AUC values for the random forest model were recorded at 0.875 and 0.804, respectively. The predicted probabilities of both models in the training and test sets aligned well with the actual incidence of CD. Conclusions: Age, 25(OH)D, and TG are key factors for CD in T2DM patients with cerebral infarction. The random forest model showed superior predictive performance, making it a promising tool for clinical use.