临床和生化指标评估胰岛素抵抗的研究

The study of clinical and biochemical parameters to estimate insulin resistance

目的 探讨临床和生化指标评估胰岛素抵抗（IR）的可能性.方法 以体检的1 295例人群为研究对象,根据Logistic回归分析结果,对与IR相关的临床和生化指标进行评分,然后计算分值.结果 （1）Logistic回归分析显示,反映IR的指标是空腹血糖、腰围、三酰甘油、2型糖尿病家族史、非酒精性脂肪肝,而血压、血尿酸和高密度脂蛋白胆固醇未进入回归方程;（2）根据是否有糖代谢异常、2型糖尿病家族史、非酒精性脂肪肝和高腰围、高三酰甘油血症的分值评估IR,受试者工作特征（ROC）曲线下面积为0.794,即使增加了血压和（或）血尿酸和（或）高密度脂蛋白胆固醇也没有增加ROC曲线下面积;（3）分值≥2,临床和生化指标诊断IR的敏感性是59.2％,特异性是88.2％,准确性最高为78.5％.临床、生化指标与稳态模式胰岛素抵抗指数（HOMA-IR）判断IR的Kappa值为0.50.结论 临床上可根据是否有糖代谢异常、2型糖尿病家族史、非酒精性脂肪肝和高腰围、高三酰甘油血症的不同组合判断IR.

Objective To investigate clinical and biochemical indicators to identify individuals who were insulin resistance （ IR）. Methods According to the multivariate results of logistic regression analysis,the clinical and biochemical parameters related to IR were determined in 1 295 health examina- tion subjects. The score for each of the clinical and biochemical parameters were summed. Results （1） Multivariate logistic regression analysis showed that fasting blood glucose（FBG） ,waist circumference （ WC ）, triglyceride （ TG）, family history of type 2 diabetes, and nonalcoholic fatty liver disease （ NAFLD ） were independently associated with IR. But blood pressure, blood uric acid and high density lipoprotein cholesterol （ HDL-C ） were not. （ 2 ） FBG, WC, TG, family history of type 2 diabetes and NAFLD were clini- cal and biochemical parameters used to define the score. The area under receiver operating characteristic （ROC） curves that the score diagnosed IR was 0.794. If the score blood pressure, or/and blood uric acid, or/and HDL-C were added to, the area under ROC curves was not increased. （ 3 ） When the score was greater than or equal to 2, the sensitivity for diagnosis of IR by the clinical and biochemical parameters was 59.2% , the specificity was 88.2% and the accuracy was 78.5%. The Kappa value was 0.50 between the score and HOMA-IR index in estimating IR. Conclusion The elevated FBG, increased WC, hyper- triglyceridemia, family history of type 2 diabetes, and NAFLD, alone or in combination could identify sub- jects who were IR.