Hormone therapy for menopause is not associated with higher death rates, Danish study of 876,000 women finds

Two decades ago, the landscape shifted for menopausal hormone therapy. A large randomized trial - the Women's Health Initiative - raised concerns about risks of breast cancer, heart disease, and stroke in women taking combined estrogen-progestogen therapy. Prescribing rates fell sharply, and many women stopped treatment or never started it. Subsequent analysis suggested the original trial had studied an older population and formulations that may not reflect current practice, but uncertainty remained. A study published in The BMJ, drawing on nationwide Danish health registers covering 876,805 women, addresses the most fundamental remaining question: does hormone therapy use alter the risk of dying?

The answer from this large observational dataset, after careful statistical adjustment, is no meaningful difference.

Who was studied and for how long

Danish health registers cover the entire population with administrative linkage between prescriptions, diagnoses, and vital status records. The research team tracked women born between 1950 and 1977 who were alive at age 45, following each from their 45th birthday until death or July 31, 2023. The median follow-up exceeded 14 years. Women were excluded if they had prior blood clots, liver disease, breast, womb, or ovarian cancer, had previously used hormone therapy, or had undergone bilateral oophorectomy (surgical removal of both ovaries).

Of the 876,805 women included in the main analysis, 104,086 (11.9%) filled at least one prescription for menopausal hormone therapy. Among those users, the median treatment duration was 1.7 years. Only 7,337 women (0.8%) used therapy for ten or more years. By the end of follow-up, 47,594 women (5.4%) had died.

The raw numbers and what adjustment revealed

Before accounting for influential factors, hormone therapy users appeared to have higher mortality: 54.9 deaths per 10,000 person-years, compared to 35.5 per 10,000 in women who never used therapy. This apparent excess reflects confounding rather than a drug effect. Women who use hormone therapy are more likely to be receiving it because they have significant menopausal symptoms, and they are more likely to have underlying health conditions that both prompt treatment and influence mortality. They are also older on average within the analysis window.

When the researchers adjusted for age, number of children, education, income, country of birth, diabetes, high blood pressure, heart disease, and other potentially influential factors, the difference disappeared. No meaningful difference in all-cause mortality remained between users and non-users. The finding held across subgroup analyses and was consistent across cause-specific mortality categories including heart disease, stroke, and cancer.

Duration of use: no increased risk at ten or more years

One of the persistent concerns about hormone therapy has been that long-term use might carry greater risk than short-term use, even if brief exposure is safe. The Danish data found no increased mortality risk at any duration, including the small group who used therapy for ten or more years. This is a meaningful finding given the clinical reality that women who need therapy for symptom control often require it for extended periods, and prior uncertainty about long-term use has discouraged continuation even in women who are benefiting.

What this means for clinical practice

Current guidelines - including those from the North American Menopause Society and major European guidelines groups - already recommend hormone therapy for women who have recently entered menopause, have moderate to severe symptoms, and have no contraindications. The Danish findings support those guidelines by providing large-scale real-world evidence that therapy does not increase mortality in this population.

The study cannot address all aspects of hormone therapy safety. Associations between specific preparations, routes of administration, and risks such as breast cancer or venous thromboembolism require separate analysis and are not resolved by all-cause mortality data alone. The population studied - Danish women born 1950 to 1977, predominantly using contemporary hormone formulations - may not generalize fully to other populations or older formulations used in earlier decades.

The observational design, despite extensive adjustment, cannot fully rule out residual confounding. Women who use hormone therapy may differ from non-users in health behaviors or unmeasured clinical characteristics that influence mortality independently of the treatment. The consistency of the null finding across multiple subgroups and analytical approaches, however, makes unmeasured confounding as the sole explanation less likely.