Comparative Analysis of the Sensitivity, Specificity, Concordance, and 5-Year Predictive Power of Diabetes-Related Autoantibody Assays

Abstract

Interassay concordance and 5-year diabetes prediction of islet cell autoantibody detection using the radiobinding assay (TrialNet), two independently developed multiplex electrochemiluminescence detection methods, the luciferase immune precipitation system, detection by agglutination-PCR, and truncated GADA, and IA2βA radiobinding assays are reported. There was considerable discordance that varied by type of autoantibody across the assays. Type 1 diabetes prediction was relatively high and uniform, implying confirmation of increased diabetes risk among those who are multiple autoantibody positive, although substantial false positive rates need to be considered when autoantibodies alone are used for screening to identify high diabetes risk.