Islet autoantibody development during follow-up of high-risk children from the general Norwegian population from three months of age: Design and early results from the MIDIA study Lars C. Stene a,b, * , Elisabet Witsø a , Peter A. Torjesen b,c , Trond Rasmussen d , Per Magnus a , Ondrej Cinek e , Turid Wetlesen a , Kjersti S. Rønningen a a Division of Epidemiology, Norwegian Institute of Public Health, NO-0403 Oslo, Norway b Diabetes Research Centre, Aker and Ulleva ˚ l University Hospitals, NO-0407 Oslo, Norway c Hormone Laboratory, Aker University Hospital, Faculty Division Aker University Hospital, University of Oslo, NO-0514 Oslo, Norway d Division of Administration, Department of Information Systems, Norwegian Institute of Public Health, NO-0403 Oslo, Norway e Motol University Hospital, 2nd Medical School, Charles University in Prague, Prague, Czech Republic Received 7 March 2007; revised 13 April 2007; accepted 16 April 2007 Abstract We describe the design of the MIDIA study and present serial islet autoantibody data from 3 months of age in the 526 first enrolled children from the general population carrying the type 1 diabetes high-risk HLA-DRB1*0401-DQA1*03-DQB1*0302/DRB1*0301-DQA1*05-DQB1*02 genotype. Blood samples were obtained from children at ages 3, 6, 9 and 12 months and annually thereafter to a median age of 12 months. Autoantibodies to insulin, glutamic acid decarboxylase and insulinoma-associated antigen-2 were measured with radiobinding assays. About 25,000 general population newborns were genotyped, and among 526 children with the high-risk HLA genotype, 2104 samples were assayed. Fourteen children were positive in at least two consecutive samples, including 12 who were positive for 2 autoantibodies at least once, of which five developed type 1 diabetes at median age 15.3 months. Seven of 14 persistently positive children seroconverted before 9 months, including two before 6 months of age. The estimated cumulative probability of multiple autoantibody positivity at 5 years was 7.3% (95% confidence interval: 3.5e12.4%). Thus, persistent islet autoimmunity is not uncommon in the first year of life in children from the general population car- rying the high-risk HLA genotype, and may develop as early as at 6 months of age. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Type 1 diabetes; Autoimmunity; Prospective studies 1. Introduction The incidence of type 1 diabetes among children [1] and some details of the molecular events during the beta-cell de- struction in type 1 diabetes are relatively well described [2], but the aetiology and the causes of the immune attack remain obscure and it is currently not possible to prevent the disease [2,3]. Three large-scale intervention studies in high-risk first- degree relatives of persons with type 1 diabetes have failed to prevent or delay the onset of type 1 diabetes [4e6]. For several autoimmune and other immune mediated diseases, specific autoantibodies are important diagnostic tools and may predict disease in healthy individuals at risk [7e11], the latter perhaps most convincingly for type 1 diabetes [7,12]. Longitudinal observational studies of high-risk chil- dren have contributed important information, not only on the ability of islet autoantibodies to predict type 1 diabetes, but also on the natural history of the disease, both of which are im- portant for evaluating potential etiologic factors and designing preventive intervention studies [2,3,13]. * Corresponding author. Division of Epidemiology, Norwegian Institute of Public Health, P.O. Box 4404 Nydalen, NO-0403 Oslo, Norway. Tel.: þ47 23 40 81 76; fax: þ47 23 40 82 52. E-mail address: lars.christian.stene@fhi.no (L.C. Stene). 0896-8411/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaut.2007.04.003 Journal of Autoimmunity 29 (2007) 44e51 www.elsevier.com/locate/jautimm