We studied environmental risk factors which might contribute to the development of beta-cell autoantibodies in healthy children. Here, we investigated 6000 randomly selected children from the large All Babies in Southeast Sweden (ABIS) cohort, including 17 055 newborns recruited between 1997 and 1999. Questionnaires at birth and at 1 yr of age and the levels of autoantibodies to glutamic acid decarboxylase (GADA) and autoantibodies to tyrosine phosphatase (IA-2A) at 1 yr of age were analyzed. The 99th percentile cutoff for autoantibodies was proposed to identify children at risk of type 1 diabetes (T1D) and the 90th percentile cutoff to identify children in whom beta-cell autoimmunity has been induced. Using the 90th percentile cutoff level, 1156 children had either IA-2A (n = 574) or GADA (n = 582), while 126 children had both GADA and IA-2A. When the 99th percentile cutoff level was used, 114 children had either IA-2A (n = 57) or GADA (n = 57), and six children had both GADA and IA-2A. In logistic regression analysis, celiac disease in grandparents [odds ratio (OR) 2.2] and maternal gastrointestinal infection (OR 1.1) represented a risk for simultaneous occurrence of both IA-2A and GADA above the 90th percentile. Birth in spring (March to May) (OR 1.5) and male gender (OR 1.3) were risk factors for induction of IA-2A. Mother's low education represented a risk for induction of IA-2A (OR 1.5) and GADA (OR 1.4). T1D in first-degree relatives increased the risk for beta-cell autoimmunity above the 99th percentile (OR 2.6), whereas type 2 diabetes in grandparents was associated with GADA (OR 2.1). Exposure to cow's milk formulas <2 months of age implied an OR of 2.9 for IA-2A above the 99th percentile.