摘要
目的建立自体血管动静脉内瘘(AVF)失功的风险预测列线图模型并进行内部验证。方法回顾性分析2020年1月至10月南京大学医学院附属鼓楼医院和泰州市人民医院AVF失功(失功组,96例)及正常进行血液透析(非失功组,100例)的196例患者的临床资料。基于196例患者的数据集,使用Lasso回归模型筛选预测因素,应用多因素Logistic回归分析方法,结合Lasso回归模型,建立列线图预测模型。使用C指数、受试者工作特征(ROC)曲线、校正图、决策曲线对预测模型的预测能力、区分度、校准度、临床实用性进行临床效能评估。内部验证使用Bootstrap自抽样法进行评估。结果预测列线图中包含的预测因素包括合并糖尿病、C反应蛋白升高、血小板的数量、纤维蛋白原的水平、钙磷乘积升高、固定点穿刺、透析后低血压状态。预测模型C指数为0.768(95%CI:0.703~0.833),校正效果良好,内部验证C指数为0.725。决策曲线显示当在14%~96%的失功可能性阈值下干预时,预测列线图在临床上是获益的。结论构建列线图预测模型可以用于血液透析患者AVF失功的风险预测。
Objective To construct a nomogram model for predicting the risk of autogenous arteriovenous fistula(AVF)dysfunction and to conduct internal validation.Methods The clinical data of 196 patients in the Nanjing Drum Tower Hospital and Jiangsu Taizhou People's Hospital from January 2020 to October 2020 were retrospectively analyzed.Among 196 patients,there were 96 patients of AVF dysfunction(dysfunction group)and 100 patients with normal hemodialysis(non-dysfunction group).Based on the data set of 196 patients,the predictors were screened by the Lasso regression model.Multivariate logistic regression analysis was applied to build a prediction model incorporating the Lasso regression model.The clinical performance of the predictive ability,discrimination,calibration,and clinical practicability of the prediction model was assessed using C-index,receiver operating characteristic(ROC)curve,calibration plot,and decision curve.Internal validation was conducted using bootstrapping techniques.Results The predictors contained in the prediction nomogram included diabetes mellitus,elevated C-reactive protein,platelet count,fibrinogen level,increased calcium-phosphorus product,fixed-point puncture,and dialysis hypotension.The model displayed good prediction ability with a C-index of 0.768(95%CI:0.703-0.833)and good calibration.The C-index was 0.725 according to the internal verification.The decision curve showed that the predictive nomogram was clinically useful when intervention was decided at the dysfunction probability threshold of 14%~96%.Conclusion This nomogram prediction model constructed in this study can be used to predict the risk of autogenous arteriovenous fistula dysfunction in hemodialysis patients.
作者
车星
张明
Che Xing;Zhang Ming(Department of Vascular Surgery,Jiangsu Taizhou People's Hospital,Taizhou 225300,China;Department of Vascular Surgery,Nanjing Drum Tower Hospital,Nanjing 210008,China)
出处
《中华血管外科杂志》
2022年第4期256-260,共5页
Chinese Journal of Vascular Surgery
关键词
自体血管动静脉内瘘
失功
预测因素
列线图
Autogenous arteriovenous fistula
Dysfunction
Predictors
Nomogram
