Objective To investigate if exclusive breast-feeding for 4 months is associated with atopic dermatitis during the first 3 years of life. Study design Data on 3903 children were taken from yearly parental-administered questionnaires from a birth cohort study in Germany (recruited 1995-1998) comprised of a noninterventional (NI) and an interventional (I) subgroup. Outcomes were physician-diagnosed atopic dermatitis (AD) and itchy rash. Multiple logistic regression was performed for the entire cohort and stratified by family history of allergy and by study group adjusting for a fixed set of risk factors for allergies. Results Exclusive breast-feeding (52 % of children) was not associated with higher risk for AD either in the entire cohort (ORadj, 0.95; 95% CI, 0.79-1.14) or if stratified by family history of AD. In the I subgroup, but not in the NI subgroup, exclusive breast-feeding showed a significant protective effect on AD if compared with conventional cow's milk formula (ORadj, 0.64; 95% CI, 0.45-0.90). Conclusion These findings do not support the hypothesis that exclusive breast-feeding is a risk factor for development of atopic dermatitis but is protective if compared with conventional cow's milk. Observational studies might not be able to effectively control for selection bias and reverse causation.