2型糖尿病微血管病变的代谢组学标志物和预测模型研究  

Metabolomic biomarkers and predictive models for microangiopathy of type 2 diabetes mellitus

作　　者：张益红[1] 林云志 魏永越[2] 叶小龙[1] ZHANG Yi-hong;LIN Yun-zhi;WEI Yong-yue;YE Xiao-long(The Second Affiliated Hospital of Nanjing Medical University,Jiangsu Nanjing 211166,China;不详)

机构地区：[1]南京医科大学第二附属医院,江苏南京211166 [2]南京医科大学公共卫生学院

出　　处：《江苏预防医学》2022年第6期627-631,654,共6页Jiangsu Journal of Preventive Medicine

基　　金：国家自然科学基金面上项目(81973142)。

摘　　要：目的运用胆汁酸-脂肪酸代谢组学标志物建立2型糖尿病微血管病变预测模型。方法回顾性分析2019年6—10月某医院72例2型糖尿病住院患者,以是否发生糖尿病微血管病变分为病变组(20例)和无病变组(52例),利用患者临床资料,通过弹性网络回归筛选与微血管病变相关危险因素,构建风险得分,将风险得分和微血管病变结局建立logistic回归模型,通过ROC曲线下面积(AUC)评估模型效果。结果弹性网络回归筛选十八碳二烯酸(C18∶2)、γ-十八碳三烯酸(γ-C18∶3)和AA/DHA比值3种代谢组学标志物,及糖化血红蛋白(HbA1c)、总胆红素(tbil)、间接胆红素(ibil)、谷丙转氨酶(ALT)、高密度脂蛋白(HDL)等5种生化指标,分别构建生化指标风险、代谢标志物风险、代谢标志物-生化指标综合模型,AUC分别为0.879(95%CI:0.791~0.967)、0.834(95%CI:0.724~0.943)、0.894(95%CI:0.812~0.971)。结论建立的代谢组学标志物-生化指标糖尿病微血管病变预测模型,预测效果较好,可为糖尿病患者预后评估提供参考。Objective To establish a predictive model for microangiopathy of type 2 diabetes mellitus using bile acids and fatty acids targeted metabolomics markers.Methods The medical records of 72 patients with type 2 diabetes mellitus admitted to a hospital during the period between Jun and Oct 2019 were retrospectively reviewed.All patients were classified into the microangiopathy(n=20)and non-microangiopathy group(n=52)according to the development of diabetic microangiopathy.Based on patients’medical records,the risk factors associated with diabetic microangiopathy were screened with elastic net regression and employed for risk factors.In addition,the risk scores and microangiopathy outcomes were included to create a logistic regression model,and the effectiveness of this model was evaluated using the area under the receiver operating characteristic(ROC)curve.Results Three metabolomic markers(octadecadiynoic acid,γ-Octadecatrienoic acid and arachidonic acid/docosahexaenoic acid ratio)and 5 biochemical parameters(glycosylated hemoglobin,total bilirubin,indirect bilirubin,alanine aminotransferase and high-density lipoprotein)were screened with elastic net regression to create the biochemical parameter,metabolic marker risk models and metabolic marker-biochemical parameter integrated model,with AUC values of 0.879[95%CI:(0.791,0.967)],0.834[95%CI:(0.724,0.943)]and 0.894[95%CI:(0.812,0.971)],respectively.Conclusion The metabolic marker-biochemical parameter integrated model has a high efficiency for prediction of diabetic microangiopathy,which may provide insights into the prognostic evaluation among diabetic patients.

关 键 词：2型糖尿病 微血管病变 预测模型 代谢组学标志物 弹性网络回归模型 

分 类 号：R587.1[医药卫生—内分泌]