The evidence for the putative role of cow's milk in the development of type 1 diabetes is controversial. We studied infant feeding patterns and childhood diet by structured questionnaire (n = 725) and HLA-DQB1 genotype by a polymerase chain reaction-based method (n = 556) in siblings of affected children and followed them for clinical type 1 diabetes. In a nested case-control design in a population who had both dietary and genetic data available, we selected as cases those siblings who progressed to clinical diabetes during the follow-up period (n = 33). For each case, we chose as matched control subjects siblings who fulfilled the following criteria: same sex, age within 1 year, not from the same family, the start of the follow-up within 6 months of that of the respective case, and being at risk for type 1 diabetes at the time the case presented with that disease (n = 254). The median follow-up time was 9.7 years (range 0.2-11.3). Early age at introduction of cow's milk supplements was not significantly associated with progression to clinical type 1 diabetes (relative risk adjusted for matching factors, maternal education, maternal and child's ages, childhood milk consumption, and genetic susceptibility markers was 1.60 [95% CI 0.5-5.1]). The estimated relative risk of childhood milk consumption for progression to type 1 diabetes was 5.37 (1.6-18.4) when adjusted for the matching and aforementioned sociodemographic factors, age at introduction of supplementary milk feeding, as well as for genetic susceptibility markers. In conclusion, our results provide support for the hypothesis that high consumption of cow's milk during childhood can be diabetogenic in siblings of children with type 1 diabetes. However, further studies are needed to assess the possible interaction between genetic disease susceptibility and dietary exposures in the development of this disease.