20-Year Follow-Up: Hormone Therapy Did Not Increase Long-Term Mortality

Published in Scientific Reports in January 2023, this Danish cohort study examined hormone therapy use and mortality patterns before and after the 2002 Women's Health Initiative publication — the event that caused hormone therapy prescriptions to plummet worldwide. The study followed two Danish populations from 1995: a 5% random sample of 52,388 women from the general population and a sample of 15,261 female twins, with hormone therapy use assessed at multiple time points between 1995 and 2010. The central finding: women aged 56 to 75 using systemic hormone therapy in 2000 had lower mortality compared to non-users. After 2002, however, users' mortality became similar to the background population.

The researchers propose a selection effect to explain this shift. After the WHI results were published, the healthiest and most health- conscious hormone therapy users — women who were diligently managing their health and following medical guidance — were the ones most likely to discontinue treatment in response to the alarming headlines. The women who continued hormone therapy post-2002 represented a different population profile. This selection bias means that post-WHI observational data on hormone therapy safety may be systematically skewed. Critically, the 20-year WHI follow-up itself found that hormone therapy did not increase long-term all-cause mortality — a finding that received far less attention than the original risk-focused results.

Why This Matters

The long-term safety picture for hormone therapy is more reassuring than most women have been led to believe, but it is also more complicated than a simple safe-or-dangerous binary. Formulation, dose, route of administration, timing of initiation, and individual risk factors all modulate outcomes. What the Danish study and the WHI follow-up together suggest is that the fear-driven exodus from hormone therapy after 2002 may have caused more aggregate harm than the therapy itself — millions of symptomatic women going untreated for years based on findings from a trial population that did not represent them. For individual women making this decision today, the answer is not in population-level averages. It is in their own risk profile, symptom trajectory, and treatment response — data that only longitudinal tracking can provide.