三种统计方法建立空腹血清胰岛素参考区间及临床验证  

The Establishment of Fasting Serum Insulin Reference Interval by Three Statistical Methods and Its Clinical Verification

作　　者：成斐 陈永德[1] 李延伟[1] 孙磊 朝浩鹏[1] CHENG Fei;CHEN Yongde;LI Yanwei;SUN Lei;CHAO Haopeng(Laboratory Department, Wangjing Hospital, China academy of Chinese Medical Sciences, Beijing 100102, China)

机构地区：[1]中国中医科学院望京医院检验科,北京100102

出　　处：《标记免疫分析与临床》2020年第11期1857-1861,共5页Labeled Immunoassays and Clinical Medicine

基　　金：中国中医科学院望京医院院级课题基金(编号:WJYY2018-13)。

摘　　要：目的采用三种统计方法建立空腹血清胰岛素的生物参考区间并验证其临床应用效果。方法通过问卷调查及严格标准,选取2018年10月至2019年2月在中国中医科学院望京医院进行体检的表观健康个体,采用Roche Cobas8000免疫分析系统检测其空腹血清中胰岛素水平。依据CLSI C28-A3中的非参法、对数(log)转换后非参数及参数法分别建立空腹血清胰岛素的参考区间。选择初诊糖尿病个体用本研究建立的参考区间、试剂说明书参考区间和全国临床检验操作规程参考区间对其进行验证。结果共入选123例表观健康个体。空腹血清胰岛素水平无性别或年龄差异。直接和对数转换(log)后非参数法得到参考区间为4.11~17.98μIU/mL;对数转换(log)后参数法得到参考区间为3.91~18.59μIU/mL。共入选36例初诊的糖尿病患者,其中1型10例,2型26例。非参法和log转换后非参法、log转换后参数法、试剂说明书和全国临床检验操作规程临床符合率(95%可信区间)分别为52.8%(36.5%~69.1%)、52.8%(36.5%~69.1%)、36.1%(20.4%~51.8%)和36.1%(20.4%~51.8%),差异无统计学意义(P=0.16)。结论三种统计方法建立的空腹血清胰岛素参考区间基本一致,本研究建立的参考区间更有助于糖尿病患者的诊断与及时治疗。Objective To establish the fasting serum insulin(FSI)reference intervals(RI)using different statistical methods and to verify its diagnostic value in diabetes mellitus.Methods The study subjects for establishing RIs were selected based on questionnaires and strict criteria from physical examination center of Wangjing Hospital,China Academy of Traditional Chinese Medicine from October,2018 to February,2019.FSI concentrations were obtained using Roche cobas8000.RIs were established following the EP28-A3 guideline.Patients who were diagnosed as DM were enrolled to verify the diagnostic value of FSI RIs from our study,also the reference intervals based on reagent instruction and the National Clinical Test Regulation of Operation.Results A total of 123 individuals were selected and used to calculate the relevant RIs.There was no significant difference among different ages and between sexes.The RIs for FSI were 4.11-17.98μIU/mL when nonparametric methods were used whether log transformation or not.The RIs for FSI were 3.91-18.59μIU/mL when parametric methods(log transformed)were used.Clinical concordance rates were 52.8%(36.5%-69.1%),52.8%(36.5%-69.1%),36.1%(20.4%-51.8%)and 36.1%(20.4%-51.8%),respectively,for nonparametric methods whether log transformation or not,and parametric methods,using reagent instruction and the National Clinical Test Regulation of Operation in 36 patients.Conclusion The RIs and clinical concordance rates for FSI are similar when 3 different statistical method are used.The FSI RI developed from our study is more suitable for the diagnosis and treatment of DM patients timely.

关 键 词：CLSI C28-A3 空腹胰岛素 参数法 非参数法 参考区间 

分 类 号：R392-33[医药卫生—免疫学]