Current data indicate an obvious relation between food allergy and atopic eczema in infants. However, diagnostic methods for food allergy need to be supplemented. The objective was to study the relevance of food patch testing in the detection of food allergy in correlation with oral food challenge and skin prick tests in atopic infants. Infants with atopic eczema (n = 113) aged 2-24 months were studied. Each patient was subjected to double-blind, placebo-controlled, or open cow's milk challenge, and skin prick and patch tests. Polysensitization, as judged from skin test results, was common in patients with atopic eczema (79/113). Cow's milk challenge was positive in 54/113 infants; reactions were immediate in 36/54 and delayed in 18/54. Immediate-type reactions were associated with skin prick test positivity and delayed reactions with patch test positivity. Altogether 26% of the cow's milk-allergic infants were detected by patch testing only. Patch testing improved the accuracy of skin testing in the diagnosis of food allergy in infants with atopic eczema, but it needs to be standardized. Polysensitization appears to be more common than generally believed among infants with atopic eczema.