A familiar, inexpensive pill that has sat in medicine cabinets for more than a century is back in the scientific spotlight — this time for its possible effect on the brain. Researchers report that older adults carrying a specific genetic profile had a roughly 70% lower relative risk of dementia when they took low-dose aspirin, a finding that could help explain one of the most frustrating puzzles in preventive medicine: why a drug that appears to help some people appears to do nothing for others.
The result, which spread quickly across science and general-interest outlets this week, arrives with an equally important caveat. Previous large studies of aspirin and cognition found no overall protective effect, and the new work does not overturn that conclusion. Instead, it suggests the average may have been hiding two very different outcomes inside the same population.
What the researchers found
According to the study coverage, the team examined older adults and compared dementia incidence between low-dose aspirin users and non-users, then stratified the analysis by genetic profile. Among participants with a particular genotype, aspirin use was associated with a 70% lower relative risk of developing dementia. In the broader group, that signal washed out — consistent with earlier randomized evidence showing no benefit across the board.
Older adults with a specific genetic profile had a 70% lower relative risk of dementia when taking low-dose aspirin, even though previous studies found no overall protective effect. The discovery suggests that genetics could help explain why some people benefit while others do not.
The distinction matters. A relative risk reduction of 70% in a subgroup is a striking number, but it describes a comparison within a genetically defined slice of the population, not a promise to the average patient. Absolute risk — how many actual cases were avoided — depends on how common dementia is in that subgroup to begin with.
Why the framing differed so sharply
Coverage of the study split along a familiar fault line. Several outlets led with the headline number: aspirin may slash dementia risk by 70%. Others foregrounded the catch, emphasizing that the benefit was confined to carriers of a specific genetic profile and that no overall effect had been established. One outlet framed the news more skeptically still, describing it as the latest chapter in "aspirin's fall from anti-aging grace."
That framing is not unfair. Aspirin was, for years, marketed in the popular imagination as a near-universal prophylactic — good for the heart, possibly good for the brain, cheap enough to take on a hunch. The evidence has since grown considerably more complicated.
- The 2018 ASPREE trial, a large randomized study of low-dose aspirin in healthy older adults, found no reduction in dementia or disability — and an unexpected increase in deaths, alongside the well-known rise in bleeding.
- Observational studies have repeatedly hinted at cognitive benefit, but such designs struggle to separate the drug's effect from the characteristics of people who take it.
- Subgroup analyses, while useful for generating hypotheses, multiply the chances of finding an apparent effect by chance alone.
The new genetic stratification is best understood as an attempt to resolve that contradiction rather than to ignore it. If a drug helps one genotype and harms or does nothing for another, a trial that lumps everyone together will report a null result — and a genuinely useful therapy can be dismissed.
The bleeding problem does not go away
None of the coverage suggests aspirin is ready to be prescribed for dementia prevention. Low-dose aspirin carries a documented risk of gastrointestinal and intracranial bleeding, a risk that rises with age — precisely the population most concerned about dementia. Any future recommendation would need to show that cognitive benefit outweighs that hazard in an identifiable group, and genetic testing would have to become routine enough to identify them.
There is also the question of mechanism. Aspirin's anti-inflammatory action is the leading hypothesis: chronic neuroinflammation is increasingly implicated in neurodegenerative disease, and a subgroup with a particular inflammatory profile might plausibly respond differently. But association is not causation, and the study's observational design leaves room for confounding by indication, lifestyle, and access to care.
What happens next
The realistic path forward is a genetically stratified randomized trial — expensive, slow, and exactly the kind of study that turns a headline into a clinical guideline. In the meantime, researchers quoted in the coverage stress that patients should not start daily aspirin on the basis of this result, and that anyone already on it for cardiovascular reasons should continue only under medical advice.
Still, the finding has value beyond aspirin itself. It is a proof of concept for a broader idea: that the failures of preventive medicine may be failures of averaging, and that genetics could one day let doctors match cheap, well-understood drugs to the people most likely to benefit. For a drug that costs pennies a dose and has been studied for decades, that would be a genuinely new chapter.
