Nomogram Prediction for Lower Extremity Deep Vein Thrombosis in Acute Ischemic Stroke Patients Receiving Thrombolytic Therapy

Abstract

This study developed and evaluated a tailored nomogram to predict the potential occurrence of early lower extremity deep vein thrombosis (LDVT) in patients receiving thrombolytic therapy. We performed several logistic analyses on the training cohort and created a corresponding nomogram to forecast early LDVT. The classification accuracy and the accuracy of predicted probabilities of the multiple logistic regression model were evaluated using area under the curve (AUC) and the calibration graph method. According to the multivariate logistic regression model homocysteine, previous history of hypertension and atrial fibrillation, indirect bilirubin, age, and sex was identified as independent determinants of early LDVT. The nomogram was constructed using these variables. The calibration plots showed a good agreement between the predicted and observed LDVT possibilities in the training and validation cohorts with AUCs being 0.833 (95% CI: 0.774-0.892) and 0.907 (95% CI: 0.801-1.000), respectively. Our nomogram offers clinicians a tool for predicting the individual risk of LDVT in the early stage of acute ischemic stroke in patients receiving thrombolytic therapy, which could lead to early intervention.