Search

Background Atopic dermatitis (AD) is associated with immunological dysfunction, which may influence cancer development. Previous studies of AD and cancer demonstrate inconsistent results and few of these studies examined children or AD severity and treatment. Objectives To determine malignancy risk among children and adults with AD. Methods We conducted a cohort study using electronic health records data from UK general practices in The Health Improvement Network between 1994 and 2015. Children (< 18 years old) and adults (≥ 18 years old) with AD were matched on age, practice and index date to patients without AD. AD was categorized as mild, moderate or severe using treatments and dermatology referrals as proxies. The primary outcome was any incident malignancy, including in situ malignancy, identified using diagnosis codes and categorized into haematological, skin and solid organ malignancies. Secondary outcomes included specific malignancies: leukaemia, lymphoma, melanoma, nonmelanoma skin cancer (NMSC) and common solid-organ cancers. Results Among 409 431 children with AD (93.2% mild, 5.5% moderate, 1.3% severe) and 1 809 029 children without AD who had median follow-up of 5–7 years, the incidence rates of malignancy were 1.9–3.4 and 2.0 per 10 000 person-years (PY), respectively. The adjusted risk of malignancy overall did not differ with respect to AD [hazard ratio (HR) 1.02 (95% confidence interval 0.92–1.12)]. Severe AD was associated with increased lymphoma risk [HR 3.18 (1.41–7.16), excluding cutaneous T-cell lymphoma (CTCL)], and mild AD was associated with increased NMSC risk [1.55 (1.06–2.27)]. Among 625 083 adults with AD (65.7% mild, 31.4% moderate, 2.9% severe) and 2 678 888 adults without AD who had median follow-up of 5 years, incidence rates of malignancy were 97.4–125.3 per 10 000 PY and 103.7 per 10 000 PY, respectively. The adjusted risk of any malignancy did not differ with respect to AD [HR 1.00 (0.99–1.02)]. However, adults with severe AD had a twofold higher risk of non-CTCL lymphoma. AD was also associated with slightly higher skin cancer risk [HR 1.06 (1.04–1.08)] and slightly lower solid cancer risk [0.97 (0.96–0.98)] but results varied by specific cancers and AD severity. Conclusions Epidemiological evidence does not support a strong overall malignancy risk in AD but lymphoma risk may be increased with severe AD. [ABSTRACT FROM AUTHOR]