Two menopause studies put the dementia question on the calendar, not the prescription pad

The surgical menopause subgroup is where the arithmetic gets interesting. Its risk reduction runs 2.6 times the one measured across the whole cohort ^[11], and Anne-Marie Minihane, the senior author and a professor of nutrigenomics at the University of East Anglia, offers two explanations without choosing between them: those women tend to have shorter lifetime estrogen exposure, and they are usually given estrogen-only HRT rather than the combined form with progesterone ^[12]. Those are not the same claim. One says the deficit is deeper, so replacing it matters more. The other says the drug is different. Work published earlier this month in thousands of donated brains, which tied estrogen-only HRT to lower Alzheimer's risk ^[13], argues for the second.

The two papers also pull against each other on mechanism. In the JAMA Network Open cohort, the cognitive decline and earlier-diagnosis signals were strongest after surgical menopause ^[7], while faster accumulation of white matter hyperintensity volume, a marker of tissue damage, showed up mainly in women whose menopause arrived on its own ^[8]. If small-vessel white matter injury were the route from menopause timing to dementia, the imaging marker and the diagnoses would crowd into the same group. They do not. The UK numbers lean the same way: the Alzheimer's-specific reduction runs six percentage points wider than the all-dementia one ^[17], which is the pattern you get when the effect is not mostly vascular.

Size is the other problem. A relative reduction of that order, drawn from records rather than randomisation, sits inside the range that healthy-user selection alone can generate, and correcting for ethnicity, smoking status and education ^[3] does not touch it. Minihane says plainly that this is not a clinical trial and not enough to conclude that HRT protects against dementia or Alzheimer's ^[5], even while calling it the most comprehensive analysis in this space to date ^[16].

The timing story needs the same discipline. The favourable initiation band is ten years wide ^[14], and its comparison group is described only as later in life. Age at initiation is also the variable an observational cohort handles worst, because it travels with surgical history and with how much medicine a woman was receiving anyway.

Minihane's reading, that prevention may begin long before advanced age and that there is a window to intervene ^[15], is a hypothesis about exactly that variable. Riley Bove, the neurology professor at the University of California, San Francisco who led the second paper, claims something narrower and likelier to survive: a lasting trace of menopause on women's brains decades later ^[10]. A trace is a mark. Whether the mark moves is a different study.