Insulin autoantibodies are of less value compared with islet antibodies in the clinical diagnosis of autoimmune type 1 diabetes in children older than 3 yr of age.

BACKGROUND Insulin autoantibodies (IAA), antibodies against endogenous insulin, may be detected in type 1 diabetic children before the start of insulin treatment. OBJECTIVE To relate IAA to islet antibodies (i.e., islet cell antibodies [ICA], and antibodies against two ICA-related islet antigens, glutamic acid decarboxylase 65 [GADA] and protein tyrosine phosphatase IA-2 [IA-2 A]) at diagnosis, and to endogenous beta-cell function at follow-up after diagnosis in diabetic children. SUBJECTS We investigated 74 children, aged 1-15 yr, at the diagnosis of diabetes and 1-10 yr later. Insulin treatment may induce antibody development against exogenous insulin. Patients with insulin treatment > or = 1 wk (n = 5) were therefore excluded from the final analysis. METHODS Radioligand-binding assays based on human recombinant antigen were used to measure IAA, GADA, and IA-2 A. ICA were determined with indirect immunofluorescence. RESULTS IAA were detected at a significantly lower frequency (43%; p < or = 0.001) than ICA (86%), GADA (72%), and IA-2 A (80%). In agreement, IAA measurements only marginally increased the frequency of positive autoimmune markers at diagnosis of diabetes (from 97 to 99% positive for at least one autoantibody). Preserved beta-cell function (detectable fasting p-C-peptide levels) was found in only nine patients, who were older (13 +/- 3 vs. 7 +/- 6 yr, p = 0.002) and had fewer of the antibodies (IAA, GADA, IA-2 A, ICA) in high titer (> median) compared with 60 patients with undetectable p-C-peptide levels. CONCLUSIONS Insulin autoantibodies are of less clinical value compared with islet antibodies in the diagnosis of autoimmune type 1 diabetes in children.