Beta-cell autoimmunity and assessment of the risk of progression to type 1 diabetes

Abstract The purpose of this work was to assess the value of humoral and genetic risk markers in the prediction of type 1 diabetes in siblings of children with type 1 diabetes, to characterise preclinical course of beta-cell autoimmunity in siblings, and to investigate the frequency of autoantibodies and their relations to genetic markers, beta-cell function and progression to type 1 diabetes in a schoolchild population. The prevalence and predictive value of autoantibodies was studied in 755 initially unaffected siblings, and the combination of genetic markers and autoantibodies in 701 of these siblings. Islet cell autoantibodies (ICA), insulin autoantibodies (IAA), glutamic acid decarboxylase antibodies (GADA) and IA-2 antibodies (IA-2A) were all shown to be of value in the prediction of type 1 diabetes in siblings initially tested at or close to the diagnosis of type 1 diabetes in the index case in the family. The risk of progression to type 1 diabetes was related to the number of autoantibodies detected, and the PPV of multiple autoantibodies was 55% over a period of 8 years. Autoantibodies were closely associated with HLA risk markers. A combination of the genetic markers and autoantibodies increased the PPVs of all autoantibodies substantially but also markedly reduced the sensitivity. The preclinical course of type 1 diabetes was investigated in 39 initially unaffected siblings who progressed to clinical disease during the follow-up. These individuals were characterised by the high-risk genetic markers, decreased beta-cell function and humoral autoimmunity against multiple beta-cell targets. However, all measures implied a remarkable individual variation in the rate of the disease process and the pattern of humoral beta-cell autoimmunity. Furthermore, the autoimmune process resulting in clinical presentation of type 1 diabetes could not be unambiguously distinguished from autoimmunity not leading to clinical disease within almost 10 years of follow-up. The frequencies of ICA, IA-2A, GADA and IAA in 3652 healthy Finnish schoolchildren were 2.8%, 0.6%, 0.5% and 0.9%, respectively, and multiple antibodies were detected in 0.6% of these children. GADA and multiple antibodies were related to the DQB1*0302 allele and the DQB1*02/0302 genotype. A reduced first-phase insulin reponse (FPIR) was associated with IA-2A, GADA, IAA and multiple antibodies, but not with ICA or any specific DQB1 allele or genotype. Four subjects progressed to type 1 diabetes, all of them having multiple autoantibodies and those two who underwent an intravenous glucose tolerance test had also a reduced FPIR. None of the progressors carried the high risk DQB1*0302 allele and two of them even carried the protective DQB1*0602 or *0603 allele. In conclusion, autoantibodies alone are recommended as first-line screening in siblings, whereas subsequent determination of HLA-DQB1 markers and their combination with autoantibodies provides a valuable tool for more precise risk assessment. Wide heterogeneity in the course of preclinical type 1 diabetes complicates an accurate estimation of the individual risk of progression to type 1 diabetes among siblings of children with type 1 diabetes. Combined screening for autoantibodies is recommended for the assessment of the risk of progression to type 1 diabetes in schoolchild populations, whereas the present observations challenge the value of current genetic risk markers in predictive strategies targeting schoolchildren