文献类型：期刊文献

中文题名：1型糖尿病患者左心室舒张功能障碍影响因素及预测模型建立的研究

英文题名：Influencing factors for left ventricular diastolic dysfunction in patients with type 1 diabetes mellitus and development of a prediction model

作者：张乐媛[1];杜邹玺[2];刘珊珊[1];姚娟娟[1];马俊[1];田利民[3]

第一作者：张乐媛

机构：[1]甘肃中医药大学第一临床医学院,兰州730000;[2]兰州大学第一临床医学院;[3]甘肃省人民医院内分泌科

第一机构：甘肃中医药大学临床医学院

年份：2026

卷号：34

期号：2

起止页码：105

中文期刊名：中国糖尿病杂志

外文期刊名：Chinese Journal of Diabetes

收录：;北大核心:【北大核心2023】；

基金：甘肃省重大科技专项(22ZD6FA033)。

语种：中文

中文关键词：糖尿病,1型;左心室功能;预测模型

外文关键词：Diabetes mellitus,type 1;Left ventricular function;Predictive modeling

摘要：目的探讨T1DM患者发生左心室舒张功能障碍(LVDD)的影响因素,建立并验证预测模型。方法选取2016年1月至2024年2月于甘肃省人民医院内分泌科住院治疗的T1DM患者528例。通过超声心动图测量二尖瓣血流频谱舒张早、晚期血流峰值速度比值(E/A)、二尖瓣口舒张早期血流峰值速度与二尖瓣环舒张早期峰值速度比值(E/e’),根据E/A和E/e’分为单纯T1DM组(n=405)和合并LVDD(E/A<1和E/e’>14,n=123)组。LASSO回归筛选LVDD的影响因素并进行Logistic回归分析,建立T1DM患者发生LVDD的列线图模型。分别通过受试者工作特征(ROC)曲线、Hosmer-Lemeshow拟合优度检验、临床决策(DCA)和临床影响(CIC)曲线评估预测模型的预测能力、校准度和临床有效性。结果LVDD组年龄、DM病程、吸烟率、DKD发生率、DR发生率、DPN发生率、SBP、FPG、谷丙转氨酶、左心房内径、左心室舒张末期室间隔厚度、左心室收缩末期室间隔厚度、左心室舒张末期后壁厚度、左心室舒张末期后壁厚度高于T1DM组(P<0.05),FC-P、2 hC-P、总蛋白、白蛋白、eGFR、左心室射血分数、左心室短轴速率低于T1DM组(P<0.05)。Logistic回归分析显示,DM病程、FC-P、FPG、eGFR和DR是T1DM合并LVDD的影响因素。建立列线图预测模型的ROC曲线下面积为0.797,Hosmer-Lemeshow拟合优度检验、DCA曲线和CIC曲线结果表明,该预测模型拟合度良好且临床有效性较好。结论DM病程、FC-P、FPG、eGFR和DR是T1DM患者发生LVDD的影响因素,基于此构建的模型预测能力较强,可用于临床识别高危患者。
Objective To investigate the factors influencing the occurrence of left ventricular diastolic dysfunction(LVDD)in patients with type 1 diabetes mellitus(T1DM),and to develop and validate the predictive model.Methods A total of 528 patients with T1DM hospitalized in the Department of Endocrinology of Gansu Provincial People's Hospital were enrolled in this study between January 2016 and February 2024.The ratio of early to late diastolic peak velocity of mitral flow spectrum(E/A)and the ratio of early diastolic peak velocity of mitral orifice to early diastolic peak velocity of mitral annulus(E/e')were measured by echocardiography.All the participants were divided into the simple T1DM group(n=405)and the combined LVDD(E/A<1 and E/e'>14,n=123)group according to E/A and E/e'.LASSO regression was used to screen influencing factors for LVDD,followed by Logistic regression analysis to constructa column-line graphical model for the occurrence of LVDD in T1DM.The predictive ability,calibration and clinical validity of the prediction model were assessed by receiver operating characteristic(ROC)curve,Hosmer-Lemeshow goodness-of-fit test,clinical decision-making(DCA)and clinical impact(CIC)curves,respectively.Results Age,DM duration,smoking rate,incidence of diabetic kidney disease,incidence of diabetic retinopathy(DR),systolic blood pressure,fasting plasma glucose(FPG),ghrelin,left atrial internal diameter,left ventricular end-diastolic septal thickness,left ventricular end-systolic septal thickness,left ventricular end-diastolic posterior wall thickness,and left ventricular end-diastolic posterior wall thickness were higher in LVDD group than in the T1DM group(P<0.05),fasting C-peptide(FC-P),two-hour postprandial C-peptide,total protein,albumin,estimated glomerular filtration rate(eGFR),left ventricular ejection fraction,and left ventricular short-axis rate(LVFS)were lower in LVDD group than in the T1DM group(P<0.05).Logistic regression result showed that DM duration,FC-P,FPG,eGFR,and DR were the influencing factors for T1DM combined with LVDD.The area under the ROC curve was 0.797,indicating that the predictive model had moderate predictive ability.The results of the Hosmer-Lemeshow goodness-of-fit test,DCA curve,and CIC curve indicated that the predictive model was well-fitted and had good clinical validity.Conclusions DM duration,FC-P,FPG,eGFR and DR are the influencing factors for the occurrence of LVDD in T1DM patients.The prediction model incorporating these factors demonstrates strong discriminative performance,offering a valuable tool for clinical identification of high-risk patients.