Lack of agreement among two commercial enzyme-linked immunosorbent antibody assays and a conventional immunofluorescence-based method for detecting islet cell autoantibodies.

Abstract

Two commercial enzyme-linked immunosorbent assays (ELISAs) for antibodies associated with development of insulin-dependent (type 1) diabetes mellitus (IDDM) were evaluated in conjunction with a conventional indirect immunofluorescent-antibody-islet cell antibody (ICA) test and a radioimmunoprecipitation method for detection of insulin autoantibodies in sera from a selected group of patients. The anti-ICA ELISA was positive for only 1 to 17 serum samples from newly diagnosed IDDM patients but yielded false-positive results with 2 of 6 serum samples containing non-diabetes-related autoantibodies. Although the anti-glutamic acid decarboxylase ELISA did not show positive results for sera with other autoantibodies, it was positive for only 4 of 29 serum samples from recently diagnosed IDDM patients and for 49% of 37 indirect immunofluorescent-antibody-ICA test-positive sera. Until the antibodies associated with the development of diabetes are better characterized, allowing better standards for comparison, it will be difficult to evaluate commercial assays in this field.