Construction of a predictive model for hemorrhagic transformation after intravenous thrombolysis in elderly patients with acute cerebral infarction based on Lasso-Logistic regression model and analysis of its clinical utility
10.3760/cma.j.cn431274-20241011-01537
- VernacularTitle:基于Lasso-Logistic回归模型构建老年急性脑梗死患者静脉溶栓后出血转化的预测模型及临床效用分析
- Author:
Dan WU
1
;
Lishuang LIU
1
;
Yajing WEI
1
;
Ya GAO
1
Author Information
1. 首都医科大学附属北京康复医院神经康复中心，北京 100144
- Publication Type:Journal Article
- Keywords:
Brain infarction;
Thrombolytic therapy;
Hemorrhagic transformation;
Prediction model
- From:
Journal of Chinese Physician
2025;27(10):1515-1520
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct a predictive model for hemorrhagic transformation (HT) after intravenous thrombolysis in elderly patients with acute cerebral infarction (ACI) using the Lasso-Logistic regression model, and to analyze the clinical utility of this predictive model.Methods:A total of 310 elderly ACI patients who received intravenous thrombolysis with alteplase (rt-PA) at the Beijing Rehabilitation Hospital Affiliated to Capital Medical University from May 2022 to May 2024 were selected. The occurrence of HT within 36 hours after intravenous thrombolysis was recorded, and the patients were divided into the HT group and non-HT group based on the presence or absence of HT. Clinical data were compared between the two groups. Lasso-Logistic regression analysis was used to screen the influencing factors of HT after intravenous thrombolysis in elderly ACI patients. A nomogram predictive model for HT after intravenous thrombolysis in elderly ACI patients was constructed based on these influencing factors, and the clinical value of the nomogram predictive model was analyzed.Results:The incidence of HT within 36 hours after rt-PA intravenous thrombolysis in elderly ACI patients was 29.35%(91/310). The proportions of patients with hypertension, diabetes, anticoagulant use, and atrial fibrillation in the HT group were higher than those in the non-HT group. The onset-to-thrombolysis time (ONT), admission National Institute of Health Stroke Scale (NIHSS) score, pre-thrombolysis peripheral blood platelet count, neutrophil-to-lymphocyte ratio (NLR), and serum levels of high-sensitivity C-reactive protein (hs-CRP), vascular endothelial cadherin (VE-cad), occludin, soluble lectin-like oxidized low-density lipoprotein receptor-1 (sLOX-1), and endothelial cell-specific molecule 1 (ESM-1) in the HT group were higher than those in the non-HT group (all P<0.05). Lasso-Logistic regression analysis showed that atrial fibrillation, ONT, admission NIHSS score, pre-thrombolysis peripheral blood NLR, and serum levels of hs-CRP, VE-cad, occludin, sLOX-1, and ESM-1 were independent risk factors for HT after intravenous thrombolysis in elderly ACI patients (all P<0.05). The area under the curve (AUC) of the constructed nomogram predictive model for predicting HT after intravenous thrombolysis in elderly ACI patients was 0.914(95% CI: 0.879-0.949), indicating high predictive efficiency. When the threshold probability range was 0.05-0.83, the nomogram predictive model showed good net benefit in predicting HT after intravenous thrombolysis in elderly ACI patients and had high clinical utility in predicting the risk of HT. Conclusions:Atrial fibrillation, ONT, admission NIHSS score, pre-thrombolysis peripheral blood NLR, and serum levels of hs-CRP, VE-cad, occludin, sLOX-1, and ESM-1 are independent risk factors for HT after intravenous thrombolysis in elderly ACI patients. The nomogram predictive model constructed based on these factors has high predictive efficiency and clinical utility in predicting the risk of HT.
