The results of our nationwide nested case-control study of menopausal hormone therapy and the risk of dementia1 align with those of the largest randomised, double blind, placebo controlled trial on the topic, the Women’s Health Initiative Memory Study (WHIMS), which reported an increased dementia rate among women aged 65 years or older randomised to menopausal hormone therapy.2 Our study, designed to tackle limitations and biases of previous studies, consistently showed increased dementia rates across different durations and ages of hormone therapy use in a dose-response manner.1

By design, our study cannot determine causality, as stated in the linked editorial by Kantarci and Manson.3 Nonetheless, the evidence underlying clinical implications must be valid and solid. Kantarci and Manson cite evidence hampered by substantial limitations that are not discussed.456

Kantarci and Manson cite the WHIMS of Younger Women (WHIMS-Y) as reporting no cognitive effects from hormone therapy, except potential verbal fluency impairment.4 WHIMS-Y was, however, 3.5 times smaller than WHIMS and thus likely underpowered to detect effects of hormone therapy on dementia in this younger population. Of the placebo group, 53.2% had been users of hormone therapy, potentially causing an underestimation of risk estimates. Further, participants were unmasked to their treatment, and cognitive function was tested by telephone, different from the clinical dementia evaluation done in WHIMS. Two other referenced trials had smaller sample sizes with short follow-up and were similarly only designed to detect changes in markers of cognitive function, not clinical outcomes such as dementia.56 Thus, these trials were unable to reach conclusions about the long term safety of menopausal hormone therapy regarding dementia risk.

As future large scale clinical trials are less likely to be initiated because of ethical considerations and costs, observational studies based on valid nationwide, real world data are fundamental. Although the findings of our study are in line with those of WHIMS, the literature on the topic is inconsistent. Therefore, we call for further research to investigate if the observed association is explained by a causal link, which remains uncertain rather than “unlikely.”