Liquid- or solid-phase assays have been used for insulin autoantibody (IAA) determination, and the method of IAA measurement has not been standardized.
IAAs were determined by radiobinding assay (RBA) and enzyme-linked immunosorbent assay (ELISA) in two large age-matched groups of nondiabetic and newly diagnosed insulin-dependent (type I) diabetic children.
Positivity for IAA by RBA (≥ to nondiabetic mean + 3SD) was 2 of 178 (1.1%) and 55 of 173 (32%) in nondiabetic and diabetic children, respectively. Prevalence of IAA by RBA was significantly higher in the youngest age-group (63% between 0-4 yr). Positivity for IAA by ELISA was 1 of 178 (0.6%) and 8 of 169 (4.7%) in nondiabetic and diabetic children, respectively. Concordance rates between both assays were 0 of 3 (0%) in control subjects and 5 of 58 (8.6%) in diabetic children.
We conclude that RBA is more appropriate than ELISA for IAA detection at the onset of the disease. In addition, because available data suggest that IAAs detected by RBA only are high-affinity antibodies, it is tempting to speculate that IAAs reflect a mature immune reaction against endogenous insulin.