冠心病患者新发心房颤动风险的预测模型建立和验证

Establishment and verification of a model for predicting the risk of new-onset atrial fibrillation in patients with coronary heart disease

目的:建立与验证可识别冠心病患者发生心房颤动风险的预测模型,提早识别心房颤动发生高危人群。方法:回顾性配对研究,选取于2018年1月1日至2019年12月31日,在新疆维吾尔自治区人民医院就诊,患有冠心病合并心房颤动的86例患者,同时按照年龄、性别、冠状动脉病变程度相近的原则进行1:1配对,选取不患有心房颤动的冠心病患者86例,并将两组数据作为建模集;选取2020年1月1日至2020年12月31日间患有冠心病合并心房颤动的患者24例,不合并心房颤动的冠心病患者26例,作为验证集,运用Logistic回归模型分析冠心病患者发生心房颤动风险的危险因素,通过Logistic回归筛选变量建立简化的模型,并对模型进行内部验证。结果:二分类Logistic回归分析显示:CAR、AST、K、LVEDD、LVP为冠心病发生心房颤动危险因素,对该模型进行内部验证,建模集中ACU值为0.876(95%CI:0.814~0.938),验证集中ACU值为0.825(95%CI:0.703~0.948)。Hosmer—Lemeshow检验结果提示该模型有良好的校准度,通过DCA曲线检测该模型临床有效性,当建模集与验证集阈概率分别在12%~100%及10%~72%范围内时,该预测模型具有良好的临床有效性。结论:该预测模型具有良好的区分度、校准度,净收益率较好,可用于冠心病患者发生心房颤动风险的预测。

Objective:To establish and verify a predictive model to identify the risk of atrial fibrillation in patients with coronary heart disease and to identify the high risk population of atrial fibrillation in advance.Methods:This study is a retrospective paired study.86 patients with coronary heart disease complicated with atrial fibrillation were selected in Xinjiang Uygur Autonomous region people's Hospital from January 12018 to December 312019.At the same time,according to the principle of similar age,sex and degree of coronary artery disease,86 patients with coronary heart disease without atrial fibrillation were selected.And two sets of data are used as modeling sets.From January 12020 to December 31202024 patients with coronary heart disease with atrial fibrillation and 26 patients with coronary heart disease without atrial fibrillation were selected as the verification set.Logistic regression model was used to analyze the risk factors of atrial fibrillation in patients with coronary heart disease.Logistic regression was used to screen variables to establish a simplified model,and the model was internally verified.Results:binary Logistic regression analysis showed that CAR,AST,serum potassium and LVP were risk factors for coronary heart disease with atrial fibrillation.Internal verification of the model showed that the ACU value was 0.876(95%CI:0.814-0.938)in the modeling set and 0.825(95%CI:0.703-0.948)in the verification set.The results of Hosmer-Lemeshow test indicate that the model has good calibration.Conclusions:in this study,the DCA curve can be used to predict the threshold probability range of atrial fibrillation in patients with coronary heart disease and can safely receive prophylactic treatment.