children with T1D born in 2000–2004 and participating in DiPiS, HLA-genotyping was done at birth, and for non-participating children born 1999–2005, HLA-genotyping was done at the time of diagnosis. A more detailed description of the study population is illustrated in Fig. 1.
Children with a high risk of disease were followed in Step 2 (see Fig. 1). From the DiPiS Step 2 cohort three children without T1D were randomly selected for each case, matched for HLA genotype and birth year. We performed as exact HLA matching as possible. A match was considered exact if the case child and the control had two of the same DQA1 alleles and the same DQB1 alleles. Three exact matches was available for 130 (38%) of the cases, 148 (43%) had two exact matches, 53 (15%) had one exact match and for 13 (4%) of the cases no exact match was available. Where no exact match was possible a control with random HLA was chosen. For