Abstract
PURPOSE: To investigate whether the relation between estrogen-progestagen menopausal hormone therapy (EP-MHT) and breast cancer risk varies according to the delay between menopause onset and treatment initiation. PARTICIPANTS AND METHODS: Between 1992 and 2005, 1,726 invasive breast cancers were identified among 53,310 postmenopausal women from the French E3N cohort (mean duration of follow-up, 8.1 years). Hazard ratios (HRs) and CIs were estimated using Cox models, with MHT never users as the reference. RESULTS: Among recent users of EP-MHT, the risk of breast cancer varied according to the timing of treatment initiation. This variation was confined to short durations of use (< or = 2 years): the HR was 1.54 (95% CI, 1.28 to 1.86) for short treatments initiated in the 3-year period following menopause onset and 1.00 (95% CI, 0.68 to 1.47) for short treatments initiated later (P = .04 for homogeneity). However, this pattern of risks was not observed in users of EP-MHT containing progesterone, among whom there was no significantly increased risk associated with short duration of use (HR was 0.87 [95% CI, 0.57 to 1.32] for treatments initiated < or = 3 years after menopause, and HR was 0.90 [95% CI, 0.45 to 1.81] for treatments initiated later). Longer durations of EP-MHT use were generally associated with increases in breast cancer risk, whatever the gap time. CONCLUSION: Our results suggest that, for some EP-MHT, the timing of treatment initiation transiently modulates the risk of breast cancer and that, when initiated close to menopause, even short durations of use are associated with an increased breast cancer risk. Estrogen + progesterone combinations might be an exception in this regard.