Pattern of mortality after menopausal hormone therapy: long-term follow up in a population-based cohort

Objective To investigate long-term pattern of mortality in menopausal women according to different modalities of hormone therapy. Design Population -based prospective cohort study. Setting Denmark 1993-2013. Population A total of 29 243 women aged 50-64 years at entry into the Diet, Cancer and Health Cohort, enrolled 1993-97 and followed through 31 December 2013. Methods Cox' proportional hazards models for increasingly longer periods of follow-up time were used to estimate mortality pattern according to baseline hormone use adjusted for relevant potential confounders. Main outcome(s) All-cause and cause-specific mortality. Outcome information was obtained from the Danish Register of Causes of Death (linkage 99.6%). Results A total of 4098 women died during a median follow up of 17.6 years. After adjustment for relevant lifestyle risk factors, hormone use had no impact on all-cause mortality, regardless of modality. Among baseline users, lower cardiovascluar disease mortality was only evident after 5 years [hazard ratio (HR) 0.54; 95% CI 0.32-0.92], but dissipated with additional follow up. Conversely, lower colorectal cancer mortality (HR 0.64; 95% CI 0.46-0.89) and higher breast cancer mortality (HR 1.34; 95% CI 1.05-1.72) only became evident after 15 years of follow up. There were no significant associations for mortality from other types of cancer or from stroke. Conclusions In this long-term follow-up study, taking hormones during menopause was not associated with overall mortality among middle-aged women. Investigating cause-specific mortality revealed significant, albeit weak, differential associations according to both causes of death and over time, underlining the importance of carefully considering individual risks and duration of treatment when making decisions on hormone therapy.