Home / Resources / Clinical Gems / International Textbook of Diabetes Mellitus, 4th Ed., Excerpt #8: Epidemiology and Risk Factors for Type 1 Diabetes Mellitus Part 2 of 5

International Textbook of Diabetes Mellitus, 4th Ed., Excerpt #8: Epidemiology and Risk Factors for Type 1 Diabetes Mellitus Part 2 of 5

Jan 26, 2016


Beta-Cell autoimmunity and risk factors

Type 1A diabetes results from a chronic autoimmune destruction of the pancreatic beta cells, probably initiated by exposure of a genetically susceptible individual to some environmental agent(s). This preclinical period is marked by the presence of autoantibodies to pancreatic beta-cell antigens such as insulin, GAD65 (Glutamic Acid Decarboxylase), ICA512 (called also IA-2) or ZnT8 (Zinc Transporter 8), and precedes the onset of hyperglycemia by a few years. Several prospective studies have reported that these autoantibodies can appear early in childhood and the presence of two or more of these antibodies is highly predictive for the development of diabetes [26]. However, the etiology of the autoimmune process and β-cell destruction is not known. The islet cell antibody (ICA) assay, using immunofluorescence and pancreatic tissue has been notoriously difficult to standardize and has been replaced by a combination of radio assays for autoantibodies to insulin [27], GAD and IA-2. These tests have been shown to be quite sensitive and predictive in relatives of T1DM patients [26] and in the general population. Several prospective cohort studies investigate early genetic and environmental factors contributing to risk of beta-cell autoimmunity and progression to diabetes in relatives of patients with T1DM and in a general newborn population [28,29].


Prediction of T1DM during the preclinical period offers opportunities for prevention of T1DM. Prediction at present is based upon genetic, immunologic, and metabolic information (Figure 2.2). Though genetic markers can identify varying risk, it is only once autoimmunity has begun (marked by the presence of multiple autoantibodies to pancreatic beta-cell antigens such as insulin, GAD65, IA-2 or ZnT8) that a high positive predictive value (>90%) can be achieved, and multiple autoantibodies are present in the great majority of prediabetics [30,31].








Family history of T1DM

Concordance rates for T1DM in monozygous twins (MZ) with long-term follow-up is greater than 50% [32], compared to 6–10% in dizygous (DZ) twins, which is similar to what is found in non-twin siblings. With long-term (>30 years) follow-up, at least 2/3 of initially discordant MZ twins develop persistent beta-cell autoantibodies and/or diabetes.

Among first-degree relatives, siblings are at a higher risk (5–10% risk by age 20) than offspring; offspring of diabetic fathers are at a higher risk (∼12%) than offspring of diabetic mothers (∼6%) [33]. It is not clear why offspring of diabetic mothers are at lower risk compared to offspring of fathers, with one study reporting that presence of transplacental islet autoantibodies may decrease risk [34]. It is very likely that both environmental factors and genetic susceptibility are essential for development of T1DM, but the environmental factors may be ubiquitous and thus not be a major determinant of which individual develops diabetes given genetic susceptibility.

HLA class II susceptibility genotypes

The estimated risk of developing T1DM for general population children is 1/300 while the risk for children who have the HLA-DR3/4,DQB1*0201,DQB1*0302 genotype is approximately 1:15–1:25 (Table 2.1) [35]. Only 2.4% of the general population carries this genotype compared to 25–40% of T1DM patients. Deschamps and coworkers examined the predictive value of HLA typing in a study of 536 siblings of diabetic probands in France [36]. The risk of T1DM after 8 years, estimated by life table analysis, was 10% for siblings who were HLA identical with the probands, 3–4% for siblings with either DR3 or DR4, and 16% for those with HLA-DR3/DR4 genotype. Recently, there has been evidence for additional susceptibility loci within or linked to the MHC independent of HLA-DR/DQ, such as HLA class I alleles [37,38]. In the DAISY (Diabetes Autoimmunity Study in the Young) cohort, siblings of children with T1DM who have HLA-DR3/DR4-DQB1*0302 and are identical by descent for both HLA haplotypes with their diabetic proband sibling had a 65% risk for developing islet autoantibodies by age 7 years and a 50% risk of developing diabetes by age 10 years [39]. This strongly suggests that additional (non-DR,DQ) MHC-linked genes determine T1DM risk. On the other hand, DAISY general population children with DR3/DR4-DQB1*0302 who lack protective alleles DPB1*0402 and DRB1*0403 have a risk of 20% of activating islet autoimmunity [40]. Only 25% of autoantibody positive new onset children presenting to the Barbara Davis Center for Childhood Diabetes in Denver during this past decade have the HLA-DR3/4 genotype [41].










Several HLA haplotypes dominantly protect, including HLA-DQA1*0102,DQB1*0602, DRB1*1401, DQA1*0101, DQB1*0503 and DRB1*0701,DQA1*0201,DQB1*0303 [42]. Too few HLA-DQB1*0602 individuals expressing multiple islet autoantibodies have been studied to know the exact magnitude of protection once autoantibodies are present, but 1% of new-onset patients expressing islet autoantibodies have DQB1*0602 versus 20% of the general US population.

Other genetic factors associated with type 1 diabetes

More than 40 non-HLA susceptibility gene markers have been confirmed [43]. At present, polymorphisms of the insulin gene and PTPN22 gene contribute most to diabetes risk after HLA alleles [44]. Adding high-risk alleles of these genetic markers to HLA Class II genotyping can increase risk, but even for these loci with odds ratios between 1.7 and 2.0, the effect is small. Recent studies indicate that specific single nucleotide polymorphisms (e.g. PTPN22) [45] of non-MHC genes are associated with dramatic differences in T-cell signaling. For instance, the PTPN22 allele associated with diabetes risk is associated with decreased T cell-receptor signaling (gain of function). It is known that a subset of individuals without the specific polymorphism have similar phenotypes, and it is possible that, as diabetes-associated lymphocyte phenotypes are defined, they will contribute to disease prediction [46].

View references.