Insulin autoantibodies at the clinical manifestation of Type 1 (insulin-dependent) diabetes — a poor predictor of clinical course and antibody response to exogenous insulin

Summary To study the possible clinical significance of the appearance of insulin autoantibodies prior to the diagnosis of Type 1 (insulin-dependent) diabetes, and their value in predicting the antibody response to exogenous insulin, we observed 46 newly diagnosed diabetic children and adolescents over the year following diagnosis for the occurrence and duration of clinical remission, daily insulin dose, metabolic control, residual B-cell function, insulin-binding antibodies and conventional as well as complement-fixing islet cell antibodies. Insulin-binding antibodies were determined using both monoiodinated human and porcine insulin. Sixteen children (34.7%) were positive for insulin autoantibodies upon diagnosis of Type 1 diabetes. These subjects were significantly younger (6.2±1.0 versus 10.8±0.8 years; mean±SEM, p<0.001), and their haemoglobin A₁ levels were lower (14.1±0.6 versus 16.0±0.8%, p<0.05) at diagnosis than in the insulin autoantibody negative group. There were no significant differences in the occurrence and duration of clinical remission between insulin autoantibody-positive and -negative test groups. Daily insulin dose, haemoglobin A₁ and serum C-peptide concentrations were of the same magnitude in both groups after the diagnosis, and no association could be found between the presence of insulin autoantibodies at diagnosis and persistently positive islet cell antibodies. In tests conducted 3 months after diagnosis, the group of patients with insulin autoantibodies showed significantly higher levels (p<0.05) of antibodies binding human insulin than the group negative for insulin autoantibodies, but no significant differences could be found between the insulin binding titres of the two groups in subsequent analyses. Those who were still positive for conventional islet cell antibodies one year after diagnosis had significantly higher levels of antibodies binding human insulin (34.6±6.1 versus 12.9±1.7%, p<0.05) as well as antibodies binding porcine insulin (33.0±5.9 versus 12.7±2.9%, p<0.05) than the other subjects. Our observations suggest that insulin autoantibodies developing before the diagnosis of Type 1 diabetes have no influence on the clinical course of the disease over the first year following diagnosis, and they appear to serve as a poor predictor of the antibody response to insulin treatment.