Stability of autoantibodies and their relation to genetic and metabolic markers of Type I diabetes in initially unaffected schoolchildren

Aims/hypothesis. To study temporal changes in positivity for autoantibodies associated with Type I (insulin-dependent) diabetes mellitus and the relat ions between these antibodies, HLA-DQB1-risk markers and first-phase insuli n response (FPIR) in non-diabetic schoolchildren. Methods. The stability of the antibody status over 2 years was assessed in 104 schoolchildren initially positive for islet cell antibodies (ICA) or an tibodies to the 65 000 M-r isoform of the glutamic acid decarboxylase (GADA ) or both and in 104 antibody-negative control children matched for sex, ag e and place of residence. All children were also studied for their first-ph ase insulin response and HLA-DQB1 alleles on the second occasion. Results. On the second occasion 3 of the 98 initially ICA-positive children , 3/13 of those positive for antibodies to the IA-2 protein (IA-2A), 1/17 G ADA-positive and 2/7 of those positive for insulin autoantibodies (IAA) tes ted negative for these antibodies. Children with IA-2A, GADA, IAA and multi ple (2 2) antibodies had significantly lower first-phase insulin responses than the control children. In contrast, these responses did not differ betw een subjects with and without specific HLA-DQB1-risk alleles or genotypes. Of the six subjects with a considerably reduced first-phase insulin respons e three had multiple antibodies on both occasions but none of them had a DQ B1 genotype conferring increased diabetes risk. Two subjects progressed to Type I diabetes within 3.4 years of follow-up, both of them having multiple antibodies and a considerably reduced first-phase insulin response but nei ther of them having a DQB1-risk genotype. Conclusions/interpretation. Positivity for diabetes-associated autoantibodi es is a relatively stable phenomenon in unaffected schoolchildren, although conversion to seronegativity can occur occasionally. Our observations also indicate that DQB1 alleles associated with decreased susceptibility to Typ e I diabetes do not protect from impaired beta-cell function or from progre ssion to overt disease in initially unaffected schoolchildren.