Relationship of Insulin Autoantibodies to Presentation and Early Course of IDDM in Children

Insulin antibodies have been documented before (insulin autoantibodies [IAAs]) and after (insulin antibodies) insulin administration in children with new-onset insulin-dependent diabetes mellitus (IDDM). Whereas the relationship of IAA to various factors at presentation has been studied in some detail, little is known about their relationship to events during the 1st yr after diagnosis. Furthermore, the course and factors affecting insulin-antibody response to human insulin administration in children with newly diagnosed IDDM remain poorly defined. To study these questions, we measured serum glucose, pH, bicarbonate, C-peptide, and IAA at diagnosis in 84 children between 0.5 and 18 yr of age. Basal and peak C-peptide responses to Sustacal ingestion, glycosylated hemoglobin (HbA_1c), and IAA were then measured in 33 of these patients at 10 days and 1,3,6, and 12 mo after diagnosis. At presentation, lAAs were absent (binding below the mean + 3SD of the binding of control serums) in 51 patients (61%) and present (binding above the mean + 3SD) in 33 patients (39%). Multiple regression analysis showed a significant nonlinear relationship between lAAs and both age (P < .005) and blood glucose (P < .05) at onset. There was a stepwise increase in median insulin-antibody binding throughout the 1st yr. This increase was most marked during the 1st mo of insulin therapy and showed a statistically significant difference between successive measurements only during this period. Analysis of variance showed no effect of the factors measured at diagnosis (age, blood glucose, C-peptide, pH, bicarbonate, IAA) or those measured during the 1st yr (blood glucose, insulin dose, C-peptide, HbA_1c) on the antibody response to insulin. Furthermore, IAA did not correlate with either insulin antibodies or C-peptide at any time during the 1st yr. Because lAAs do not appear to be related to residual insulin secretion and do not predict the insulin-antibody response to human insulin administration, the data failed to define a role for these antibodies in the pathogenesis of this disorder. This suggests that lAAs may be markers of the immune confrontation with the β-cell rather than active participants in its destruction. Furthermore, insulinantibody response after human insulin administration appears to have no effect on residual insulin secretion, insulin requirement, or metabolic control during the 1st yr after diagnosis.