基于实验室常规检验数据建立急性缺血性脑卒中患者溶栓后出血转化预测模型
A prediction model of hemorrhage transformation in AIS patients after thrombolysis based on routine laboratory data
杨华 1董月稳 1张晓霞1
作者信息
摘要
目的 基于实验室常规检验数据建立急性缺血性脑卒中(acute ischemic stroke,AIS)患者阿替普酶溶栓后出血转化(hemorrhagic transformation,HT)的风险可视化模型.方法 选取2021年1月至2024年1月邢台市中心医院收治的AIS患者252例.患者均行阿替普酶溶栓,根据溶栓后24 h内是否发生HT分为发生组52例和未发生组200例.分析AIS患者阿替普酶溶栓后HT的影响因素,建立风险可视化模型并验证其效能.结果 252例AIS患者中52例(20.63％)发生HT.发生组高血压、入院时美国国立卫生研究院卒中量表评分、中性粒细胞、D-二聚体、总胆红素(total bilirubin,TBIL)水平显著高于未发生组,尿酸、血钾水平显著低于未发生组(P＜0.05,P＜0.01).多因素logistic回归分析显示,中性粒细胞、D-二聚体、TBIL是AIS患者阿替普酶溶栓后HT风险的危险因素(OR=2.753,95％CI:1.399～5.417,P=0.003;OR=1.987,95％CI:1.322～2.986,P=0.001;OR=2.121,95％CI:1.392～3.230,P=0.000),尿酸、血钾是保护因素(OR=0.417,95％CI:0.202～0.860,P=0.027;OR=0.160,95％CI:0.028～0.911,P=0.039);风险可视化模型预测AIS患者阿替普酶溶栓后HT风险的曲线下面积、敏感性、特异性分别为0.920(95％CI:0.880～0.950)、96.15％、80.50％.对可视化模型进行Hosmer-Lemeshow拟合优度检验(x2=1.888,P=0.169),说明本模型具有较好的可信度.Bootstrap法内部验证C-index指数为0.921.结论 中性粒细胞、D-二聚体、TBIL是AIS患者阿替普酶溶栓后HT风险的危险因素,尿酸、血钾是保护因素,据此建立的风险可视化模型经验证效能良好.
Abstract
Objective To develop a model for visualizing the risk of hemorrhagic transformation(HT)in acute ischemic stroke(AIS)patients after alteplase thrombolysis based on routine labo-ratory data.Methods A total of 252 AIS patients receiving alteplase thrombolysis in our hospital from January 2021 to January 2024 were enrolled,and then divided into HT group and non-HT group according to developing HT or not within 24 h after thrombolysis.The influencing factors for HT in AIS patients after alteplase thrombolysis were analyzed,and then a model of visualizing the risk was conducted and its predictive performance was verified.Results HT occurred in 52 out of 252 AIS patients(20.63％).The HT group had significantly higher ratio of hypertension,higher NIHSS score at admission,elevated neutrophil count,and increased D-dimer(D-D)and to-tal bilirubin(TBIL)levels,whereas lower uric acid(UA)and serum potassium levels when com-pared with the non-HT group(P＜0.05,P＜0.01).Multivariate logistic regression analysis identi-fied neutrophils,D-D,and TBIL as independent risk factors for HT(OR=2.753,95％CI:1.399-5.417,P=0.003;OR=1.987,95％CI:1.322-2.986,P=0.001;OR=2.121,95％CI:1.392-3.230,P=0.000),while UA and serum potassium were protective factors(OR=0.417,95％CI:0.202-0.860,P=0.027;OR=0.160,95％CI:0.028-0.911,P=0.039).The AUC value of the constructed model in predicting HT in AIS patients after alteplase thrombolysis was 0.920(95％CI:0.880-0.950),with a sensitivity of 96.15％and a specificity of 80.50％.Hosmer-Lemeshow test(x2=1.888,P=0.169)confirmed the model had good fit,and Bootstrap internal validation yielded a C-index of 0.921.Conclusion Neutrophils,D-D,and TBIL are risk factors,while UA and serum potassium are protective factors for HT in AIS patients following alteplase thrombolysis.Our developed model of risk visualization demonstrates robust predictive performance.
关键词
缺血性卒中/组织型纤溶酶原激活物/比例危险度模型Key words
ischemic stroke/tissue plasminogen activator/proportional hazards models引用本文复制引用
出版年
2025
NETL
NSTL
万方数据