预测心衰患者并发心肾综合征的列线图模型构建研究

Construction of nomogram model for predicting cardiorenal syndrome in patients with heart failure

摘要:

目的： 构建并验证个体化预测心衰患者并发心肾综合征风险的列线图模型。方法： 纳入2010年1月至2019年12月重庆市中医院心血管内科明确诊断为急性心力衰竭的621名患者作为研究对象，将研究对象随机分为建模组（70%）与验证组（30%）。应用单因素及多因素logistic回归分析建模组数据，分析心衰患者并发心肾综合征的独立危险因素。应用R软件构建心衰患者并发心肾综合征风险的列线图模型，并对该模型进行验证。结果： 多因素logistic回归分析显示，年龄、糖尿病、NYHA分级、肾小球滤过率（estimated glomerular filtration rate，eGFR）均为心衰患者并发心肾综合征的独立危险因素。对该模型进行内部和外部验证，建模组AUC值为0.807（95%CI=0.771~0.843）；验证组AUC值为0.798（95%CI=0.757~0.839）。无论是建模组还是验证组，其校准曲线均提示该预测模型具有良好稳定性。结论： 该列线图能准确预测心衰患者并发心肾综合征的个体化风险，具有较高的潜在临床应用价值。

Abstract:

Objective: To construct and validate the individualized nomogram model for predicting the risk of cardiorenal syndrome in patients with heart failure, by analyzing the independent risk factors of heart failure patients complicated with cardiorenal syndrome. Methods: A total of 621 patients diagnosed with acute heart failure in our hospital from January 2010 to December 2019 were included in the study, and they were randomized into modeling group (70%) and validation group (30%). Univariate and multivariate logistic regression were used to analyze the independent risk factors of cardiorenal syndrome in patients with heart failure. R software was used to construct a nomogram model of the risk of cardiorenal syndrome in patients with heart failure. Results: Multivariate logistic regression analysis showed that age, diabetes mellitus, NYHA grade and estimated glomerular filtration rate (eGFR) were independent risk factors of heart failure complicated with cardiorenal syndrome. Through the internal and external validation, the AUC value was 0.807 (95%CI=0.771-0.843) in the modeling group and 0.798 (95%CI=0.757-0.839) in the validation group. The calibration curves of both the modeling group and the validation group have indicated that the prediction model has good stability. Conclusion: This nomogram can accurately predict the individualized risk of cardiorenal syndrome in patients with heart failure, and has high potential clinical application value.