Effective strategies for primary prevention are urgently needed to combat the rapidly increasing prevalence of childhood obesity. Evidence accumulates that early nutrition programmes later obesity risk. Breast feeding reduces the odds ratio for obesity at school age, adjusted for biological and sociodemographic confounding variables, by some 20-25%. We propose that the protective effect of breast feeding is related in part by the induction of a lower weight gain in infancy, which is related to differences in substrate intake. Protein intake per kg bodyweight is some 55-80% higher in formula fed than in breast fed infants. We hypothesize that high early protein intakes in excess of metabolic requirements enhance weight gain in infancy and increase later obesity risk (the "early protein hypothesis"). The European Childhood Obesity Programme tests this hypothesis in a randomized double blind intervention trial in 1150 infants in five European centres. Infants that are not breast fed are randomized to formulae with higher or lower protein content and followed up to school age. If an effect of infant feeding habits on later obesity risk should be established, there is great potential for effective preventive intervention with a significant potential health benefit for the child and adult population.