Autoantibodies raised against β cell antigens are the most reliable preclinical biomarkers for predicting the imminent onset of type 1 diabetes mellitus (T1DM). The most current detection platforms are technically challenging or are run on clinically esoteric equipment. Here, we present a straightforward approach to detect autoantibody biomarkers that employs highly PEGylated microspheres onto which are mounted various capture agents that include affinity-tagged antigens or small molecule "antigen surrogates." After incubation with small quantities of serum, the bound autoantibodies can be measured using a standard flow cytometer. By multiplexing this assay, we show that a panel of antigen and antigen surrogates reliably predicts hyperglycemia in a mouse model of diabetes without false positives.