Evidence from case-control studies for the diabetogenicity of introduction of cow's milk-based formulas at early age in infancy is inconclusive. We followed siblings of children with Type 1 diabetes mellitus (Type 1 DM) to investigate a possible relationship between cow's milk consumption during infancy or later in childhood and the emergence of diabetes-associated autoantibodies and progression to clinical Type 1 DM. A cohort of 725 initially unaffected 0 to 25-year-old siblings of 801 index children with Type 1 DM diagnosed in 1986-1989 participated in the study (82 % of those invited). The siblings were observed for Type 1 DM associated autoantibodies at intervals of 3-12 months for 4 years, starting from the diagnosis of Type 1 DM in the index child. The follow-up for Type 1 DM started at the same time and ended on 31 October 1995. The combined prevalence of Type 1 DM associated autoantibodies (islet cell antibodies (ICA), insulin autoantibodies (IAA), CAD autoantibodies (GADA), and/or antibodies to the insulinoma associated cDNA2 protein (IA-2A)) was 13.6 % (95/697) at the beginning of the study. Of the initially seronegative siblings, 7.5 % (45/602) converted to antibody positivity during 4 years, and of all siblings 4.6 % (33/725) developed Type 1 DM during the total follow-up time. The age at introduction of supplementary milk feeding was not significantly related to seroconversion to positivity for Type 1 DM associated autoantibodies or to the development of Type 1 DM in the siblings. When adjusted for age, sex, infant feeding patterns, and maternal age and education, high milk consumption in childhood (greater than or equal to 3 glasses daily) was associated with more frequent emergence of Type 1 DM-associated autoantibodies than low consumption (