BACKGROUND: Some studies report that women with anorexia nervosa (AN) have lower risk than others of breast cancer, but increased risk of cancers of other sites. No work has been done to quantify the risk in the English population. METHODS: Retrospective cohort study using a national linked dataset of Hospital Episode Statistics for 1999-2021. We selected individuals with a hospital admission for AN, and compared their relative risk (RR) of developing site-specific cancers, with that in a reference cohort. RESULTS: We identified 75 cancers in 15,029 women hospitalised with AN. There was a low RR of all cancers combined at 0.75 (95%CI 0.59-0.94), and, notably, low RR for breast cancer 0.43 (0.20-0.81), cancers of secondary and ill-defined sites 0.52 (0.26-0.93). The RR for parotid gland cancer was 4.4 (1.4-10.6) within a year of first recorded diagnosis of AN. In men, we found 12 cancers in 1413 individuals hospitalised with AN, but no increased risks beyond the first year of diagnosis of AN. CONCLUSIONS: This is the first report on the association between AN and cancers in the all-England population. The study showed low rates of breast cancer, and of all cancers combined, in women hospitalised with AN. It is possible that some of the metabolic or hormonal changes observed in AN could work as a protective factor for breast cancer. More experimental work is needed to identify and explain these factors. The new finding on the higher risk of salivary gland tumours could inform clinicians caring for patients with AN.
Acta Psychiatr Scand
anorexia nervosa, caloric restriction, cancer, epidemiology, risk