AUTHOR=Chen Weicong , Cui Chaohua , Lai Changsheng 

TITLE=Developing a predictive model for lower extremity deep vein thrombosis in acute ischemic stroke using a nomogram

JOURNAL=Frontiers in Neurology

VOLUME=Volume 16 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2025.1506959

DOI=10.3389/fneur.2025.1506959

ISSN=1664-2295

ABSTRACT=BackgroundDeep vein thrombosis (DVT) is a prevalent complication among patients with acute ischemic stroke (AIS). However, there remains a deficiency of patient-specific predictive models. This study aims to develop a nomogram to estimate the risk of lower extremity DVT in AIS patients during the acute phase (within 14 days of onset).MethodsThis retrospective multicenter study analyzed 391 eligible AIS patients from two tertiary hospitals in Guangxi, China. Sixty-three clinical variables encompassing demographic profiles, clinical characteristics, laboratory parameters, and therapeutic interventions were systematically extracted from electronic health records. All participants underwent standardized Doppler ultrasound assessments for bilateral lower extremity DVT within 14 days of symptom onset. Variable selection via backward stepwise logistic regression informed nomogram construction, with model performance evaluated through calibration curves and decision curve analysis.ResultsData from one hospital were used as the modeling cohort, while data from another hospital were used for external validation. Multivariate logistic regression analysis showed that gender, age, diabetes, anemia, bed rest exceeding 3 days, and medium-frequency electrical therapy are independent risk factors for DVT in AIS patients. A nomogram was developed based on these six independent risk factors, with the area under the ROC curve (AUC) for predicting DVT risk within 14 days post-AIS being 0.812 for the modeling cohort and 0.796 for the external validation, indicating good predictive performance. Calibration of the nomogram showed Hosmer-Lemeshow test results with p values of 0.200 for the modeling set and 0.432 for the validation set, indicating good model consistency. In decision curve analysis, the nomogram demonstrated superior net benefit over staging systems across a wide range of threshold probabilities.ConclusionWe developed a nomogram to personalize the prediction of DVT risk in patients with AIS, assisting healthcare professionals in the early identification of high-risk groups for DVT and in implementing appropriate interventions to effectively prevent its occurrence.