Background The use of hormone therapy for the menopause (HT) has been shown to affect the sensitivity and specificity of mammographic screening. However there is little evidence on how the association between HT use and risk of screen detected breast cancer compares with that of interval breast cancer. We examined these associations in a large UK prospective study.

Methods We used Cox proportional hazard models to estimate the relative risk (RR) of screen detected and of interval cancer in relation to HT use among post-menopausal women who attended for routine mammographic screening. Analyses were stratified by year of birth and year of recruitment and adjusted for relevant confounders including socio-economic status, reproductive history, anthropometric and other lifestyle factors.

Results Of the 1,076,203 eligible women in the cohort, 14,730 were diagnosed with a screen-detected cancer and 8,659 with an interval breast cancer. When compared to non-users of HT, current-or-recent users were at a much higher risk of an interval cancer (RR=2.18, 95%CI 2.07–2.30) than of a screen detected cancer (RR=1.44, 95%CI 1.38–1.51). For oestrogen only HT, the corresponding RRs and 95% CI for interval and screen-detected cancers were 1.60(1.48–1.73) and 1.11(1.05–1.19); and for oestrogen and progestogen HT, the corresponding values were 2.75(2.58–2.92) and 1.79(1.70–1.88).

Conclusion In this large cohort of UK women, current-or-recent users of HT were at a substantially higher risk of being diagnosed with an interval cancer than with a screen detected cancer. The difference in risk between screen-detected and interval cancer was apparent for HT preparations containing oestrogen only and combinations of oestrogen and progestogen.