以全天指尖血糖谱及糖化血红蛋白简化肾糖阈计算公式的可行性分析  

Feasibility of simplifying the calculation of renal threshold for glucose excretion with fingertip blood glucose profile and glycated hemoglobin A1c

作　　者：石珍珍 常宝成[1] 单春艳[1] 岳晓丹 王靖宇[1] 张立弋 杨菊红[1] Shi Zhenzhen;Chang Baocheng;Shan Chunyan;Yue Xiaodan;Wang Jingyu;Zhang Liyi;Yang Juhong(Department of Endocrinology,Tianjin Medical University Chu Hsien-I Memorial Hospital,Tianjin 300134,China;Department of Endocrinology,Tianjin Medical University General Hospital Airport Hospital Site,Tianjin 300308,China)

机构地区：[1]天津医科大学朱宪彝纪念医院内分泌科天津市内分泌研究所,300134 [2]天津医科大学总医院空港医院内分泌科,300308

出　　处：《中华糖尿病杂志》2020年第5期305-311,共7页CHINESE JOURNAL OF DIABETES MELLITUS

基　　金：国家重点研发计划(2018YFC1314000);国家自然科学基金(81603461,81774043);天津市科技计划项目(17ZXMFSY00140);天津市卫计委重点攻关项目(16KG167);天津市自然科学基金重点项目(17JCZDJC34700);解放军总医院肾脏疾病国家重点实验室第一批开放课题基金(KF-01-133)。

摘　　要：目的探讨应用糖化血红蛋白(HbA1c)、指尖血糖谱分别计算平均血糖(MBG)以简化肾糖阈(RTG)计算公式的可行性。方法采用分层随机抽样法入选2018年1月至209年1月在天津医科大学朱宪彝纪念医院住院的168例2型糖尿病(T2DM)患者,计算预估肾小球滤过率(eGFR)、检测24 h尿糖,使用持续葡萄糖监测仪监测血糖,全天8次指尖血糖测量及静脉HbA1c来反映MBG计算RTG。采用Pearson相关分析法分析三种RTG的相关性,使用多元线性回归方程建立由HbA1c和指尖血糖计算得RTG的数学模型。重新入组450例T2DM患者作为新数学模型的验证人群,采用Homser和Lemeshow检验来验证上述三种RTG的一致性。结果三种血糖检测方式计算的RTG显著相关。多元线性回归方程可得:RTG(指尖血糖)=-24.572+18.385×指尖MBG(mg/dl)+0.211×eGFR[ml·min-1·(1.73 m2)-1]-0.914×24 h尿糖(g/24 h)(P<0.01),RTG(HbA1c)=-52.334+28.359×HbA1c(%)+0.189×eGFR[ml·min-1·(1.73 m2)-1]-0.616×24 h尿糖(g/24 h)(P<0.01)。Homser-Lemeshow检验两种数学模型的预测值与实际观测值拟合程度高,数据分布一致性好(χ22=9.809,P2=0.679;χ23=6.832,P3=0.555)。由HbA1c计算RTG的受试者工作特征(ROC)曲线下面积(AUC)为0.744(P<0.01),由指尖血糖计算RTG的AUC为0.892(P<0.01),诊断的灵敏度为65.53%,特异度为90.95%,约登指数0.663。多因素logistic回归分析显示,年龄、糖尿病病程、体质指数(BMI)、肾脏体积与RTG独立相关(OR为1.038~2.849,均P<0.05),进一步分层显示T2DM患者的RTG随着年龄、糖尿病病程、BMI、肾脏体积的增加而增加。结论3种检测方式所得MBG计算的RTG高度相关,以指尖血糖、HbA1c估算MBG可简化RTG计算公式,临床简易便行。高RTG的危险因素有年龄增大、糖尿病病程延长、BMI升高、肾脏体积增加。Objective To explore the feasibility of simplifying the calculation of renal threshold for glucose excretion(RTG)by calculating the mean blood glucose(MBG)with venous glycated hemoglobin A1c(HbA1c)and fingertip blood glucose spectrum.Methods A total of 168 hospitalized patients with type 2 diabetes mellitus(T2DM)who were admitted to Tianjin Medical University Chu Hsien-I Memorial Hospital from January 2018 to January 2019 were selected by stratified random sampling method.We estimated the glomerular filtration rate(eGFR)and detected the 24-hour urine sugar,the continuous glucose monitoring(CGM),finger glucose performed 8 times a day and HbA1c to reflect the MBG level,which were used to calculate RTG.Pearson correlation analysis was used to analyze the correlation of three RTG,and multiple linear regression equation was used to establish the mathematical model of RTG calculated by HbA1c and fingertip blood glucose.A total of 450 patients with T2DM were reenrolled as the verification population of the new mathematical model.Homser and Lemeshow tests were used to verify the consistency of three RTG calculated by fingertip blood glucose,HbA1c and dynamic glucose monitoring.Results There was strong correlation between RTG using three different methods to calculate RTG with the mean blood glucose(P<0.01);for fingertip blood glucose profile,the multiple linear regression equation was RTG=-24.572+18.385×fingertip MBG(mg/dl)+0.211×eGFR[ml·min-1·(1.73 m2)-1]-0.914×24 h GLU(g/24 h),and for HbA1c,the equation was:RTG=-52.334+28.359×HbA1c(%)+0.189×eGFR[ml·min-1·(1.73 m2)-1]-0.616×24 h GLU(g/24 h).Homser and Lemeshow test showed a high degree of fitting between the predicted value of those mathematical models with the observed value,and the data distribution was consistent(χ22=9.809,P2=0.679;χ23=6.832,P3=0.555).According to receiver operating characteristic(ROC)curve,the area under curve(AUC)of RTG calculated by HbA1c was 0.744(P<0.01),and the AUC of RTG calculated by fingertip blood glucose was 0.892(P<0.01),the s

关 键 词：糖尿病 2型 血红蛋白 糖基化A 肾糖阈 持续葡萄糖监测 

分 类 号：R692[医药卫生—泌尿科学] R587[医药卫生—外科学]