Factors affecting the insulin autoantibody ELISA

IgG antibodies to insulin are present in insulin-treated patients and are detected in the prodrome of untreated type I diabetes. Sporadic reports of autoantibodies to insulin suggest that they are also present in other disorders. To establish the incidence of insulin autoantibodies in other endocrine and autoimmune diseases an ELISA was used to examine sera from 529 subjects with no prior insulin therapy. These untreated patients included: normal controls (adults and children), newly-diagnosed type I diabetes, first-degree relatives of diabetics, type II diabetes, Graves' hyperthyroidism, and systemic lupus erythematosus. As a positive control group, 280 insulin-treated patients were studied. Measurement of IgG antibodies by direct binding to insulin coated plates was complicated by differences between adult and pediatric populations and by overlap of binding between treated and untreated subjects. Competitive inhibition with excess soluble human insulin overcame these problems and permitted identification of insulin specific binding. Using this approach insulin antibodies were most frequent in insulin-treated diabetics (98%) and in type I diabetics (37%) prior to treatment. The absolute numbers of subjects with insulin autoantibody in the other groups differed depending upon whether a cut-off for binding (mean + 2SD of controls) or for insulin inhibition of binding (45%) was used. Regardless of the criteria used there were subjects (2-24%) in all groups tested with circulating insulin-specific IgG autoantibody detected by ELISA. These low level antibodies detected in solid phase assays may be part of the normal immune repertoire.