Despite the numerous studies on the possible protective effect of breast feeding against the onset of atopic manifestations during childhood, this issue remains controversial. As part of an environmental epidemiological study, we investigated whether different blood concentrations of dichlorodiphenyl-dichloroethylene (DDE) modified the protective effect of breast feeding against atopic manifestations in 338 children. DDE concentration, duration of breast feeding and manifestation of atopic disorders were measured in 1994-95 at age 7-8 years. Information gathered on asthma, atopic eczema and hay fever was based on questionnaire data. We measured the total serum concentration of immunoglobulin E (IgE) and specific IgE levels against inhalant allergens. In 1997, we also determined bronchial hyper-reactivity with a hypertonic saline challenge test. To estimate odds ratios from our cross-sectional analysis, we applied logistic regressions, controlling for confounders. Breast feeding had a protective effect on the two asthma variables (e.g.> 12 weeks breast feeding for doctor-diagnosed asthma, OR = 0.32 [95% CI 0.11, 0.87]; for 'ever' asthma, OR = 0.13 [95% CI 0.02, 0.68]), but not on bronchial hyper-reactivity, hay fever, atopic eczema or the two IgE variables. The protective effect became stronger in children with DDE blood levels below the median of 0.29 micro g/L (e.g. doctor-diagnosed asthma,> 12 weeks breast feeding, OR = 0.24 [95% CI 0.06, 0.95]). Also, for specific IgE against inhalant allergens, the association gained statistical significance. For children with a DDE concentration of 0.29 micro g/L and higher, breast feeding did not show a significant protective effect. Our results suggest that contaminants such as DDE may modify the protective effect and may have contributed to inconsistent findings on the protective effect of breast feeding in previous studies. We recommend determining levels of breast milk contaminants in children when assessing the impact of breast feeding on atopic manifestations.