Background Epidemiological studies have provided growing evidence of a link between atopy and cancer risk.

Objectives To review the evidence from case–control studies and cohort studies on a possible association between atopic dermatitis (AD) and cancer risk, with particular attention to the case definition of AD.

Methods Studies with quantitative data on the association between AD (eczematous disease) and cancer risk were obtained from MEDLINE in combination with a review of cited references.

Results In 23 publications, AD was implicated in the risk of haematological [childhood leukaemia (n = 3), adult leukaemia (n = 3), non‐Hodgkin lymphoma (NHL; n = 4) and different haematological cancers (n = 1)], pancreatic (n = 5), skin (n = 2) and brain malignancies (n = 5). The overall picture of the results of these studies shows that a history of AD may be associated with a decreased risk of pancreatic cancer, brain tumour and childhood leukaemia, although in most instances the findings were not statistically significant. No consistent associations were observed for skin cancer or NHL. The definition of AD had varying quality, and was imprecise in the majority of publications.

Conclusions The findings of the epidemiological studies tend to support a lower risk of cancer among persons with a history of AD. Although a more careful definition of AD is needed, these epidemiological studies could provide an estimate of the background cancer risk in patients with AD when the long‐term effects of treatments for AD are assessed.