Optimal Cut-Point of Glutamic Acid Decarboxylase Antibody (Gad-Ab) for Differentiating Two Subtypes of Latent Autoimmune Diabetes in Adults

李霞,周智广,黄干,杨琳,颜湘,王建平

DOI: https://doi.org/10.3321/j.issn:1006-6187.2005.01.012

2005-01-01

Abstract:Objective To investigate the optimal cut-point of glutamate decarboxylase antibody (GAD-Ab) for differentiating two subtypes of latent autoimmune diabetes in adults (LADA). (Methods) The frequency distribution of GAD-Ab index in LADA was described, and the estimated curve was made. The clinical features were compared among 145 LADA, 145 type 2 and 130 type 1 diabetic patients. Receiver-operating characteristic (ROC) curve was used to evaluate the diagnostic value of GAD-Ab in insulin-deficiency (representing LADA-1 subtype) and metabolic syndrome (representing LADA-2 subtype), and the optimal cut- point of GAD-Ab was defined. GAD-Ab was (measured) with radioligand assay. Results (1) The frequency distribution of GAD-Ab index in (LADA) was bimodal. (2) The number of clinical features which had significant differences in the patients of two subgroups was increased when cut-point moved from index of 0.1 to 0.3; however, there was a contrary tendency when the cut-point moved from index of 0.4 to 0.7. (3) ROC curve demonstrated that the optimal cut-point for the diagnosis of two subtypes of LADA was at 0.3, with sensitivity and specificity of 54.5% and 92.1% respectively for insulin-deficiency and metabolic syndrome. The area under ROC curve was 0.79 (compared with 0.5, P<0.01). Conclusions The optimal cut-point for the (differentiation) of two subtypes of LADA was at the GAD-Ab index of 0.3.