Abstract
The frequencies of islet cell antibodies (ICA) and insulin autoantibodies (IAA) were studied in three clinically well defined groups, using an aprotinin sensitised indirect immunofluorescence assay for ICA and a direct binding solid ELlSA for IAA, and the association of these two serological markers for insulin dependent diabetes analysed. Frequency of ICA was 10.7% in siblings of diabetics, 15.5% in discordant identical twins and 65.9% in newly diagnosed diabetic patients. Frequency of IAA was 7.1% in siblings, 46.7% in discordant twins and 38.6% in newly diagnosed diabetic patients. No correlation was demonstrated between the two autoantibodies in the siblings. In the newly diagnosed diabetic patients there were sera positive or negative for both, but 22 (50%) of the sera showed dissociation between the two antibodies. The studies of twins showed that IAA and ICA fluctuated independently with time, and demonstrate the inappropriateness of seeking such an association in cross-sectional surveys. An association could not be demonstrated in this group even if data from multiple samples taken at different points in time were pooled, scoring an individual as positive if at any time their sera had been positive for the corresponding antibody. Thus our data showed no correlation between ICA and IAA in any of the groups studied.