Objective: To describe the incidence of diabetes related autoantibodies in a very large genetically at-risk cohort of children followed from three months of age.
Research Design and Methods: Infants with high risk HLA-DQ haplotypes were enrolled and prospectively followed with standardized autoantibody (IAA, GADA and IA-2A) assessments quarterly through the first 4 years of life and then semiannually thereafter.
Results: 500 of 8502 children (5.9%) had persistent confirmed autoantibodies during 31,801 person-years of follow-up through 72 months of age. The overall incidence was 5.9% (15.7/1000 person-years). The incidence of IAA alone, GADA alone and IAA with GADA was 2.6% (7.1/1000 person-years), 2.0% (5.4/1000 person-years) and 0.9% (2.4/1000 person-years). IA-2A alone (0.1%) appeared only rarely. The incidence of 2 or more autoantibodies was 1.1% (2.9/1000 person years). Autoantibodies occurred before 3 months of age (4.2/1000 person years). The peak incidence of IAA occurred within the first year of life, but not until 5 years of age for GADA. The longer median time (months) to GADA only seroconversion compared to IAA only seroconversion was evident among the children with HLA-DQ2/8 (+16.7 p<0.001) and HLA-DQ4/8 (+35.5. p=0.001) but less significant with HLA-DQ8/8 (+5.3, p=0.04) and not significant with DQ2/2 (+8.4, p=0.32).
Conclusions: Diabetes related autoantibodies first appear singly with IAA appearing at an earlier age than GADA. With this limited follow up, the order of autoantibody appearance was not prognostic for epitope spreading to multiple antibodies or type 1diabetes. The order of appearance is related to the HLADQ genotype type, and to that extent carries different risks for type 1 diabetes.
Funded by NIH, NIDDK, NIAID, NICHD, NIEHS, JDRF and CDC.