Can Genetic Testing Tell Me If Aspirin Could Prevent Dementia?

Low-dose aspirin has been tested for dementia prevention in large randomized trials, and on average it didn't show a significant benefit. A new study in Alzheimer's & Dementia asks a narrower question: could your genes pick out the people it does help? The researchers went back to the ASPREE trial, which randomly assigned older adults to 100 mg of aspirin a day or placebo, and reanalyzed its existing data against thousands of genetic scores. That's what "secondary analysis" means here. There was no new trial and no new clinical genetic test, only a fresh look at old randomized data. They found one genetic group where aspirin looked strongly protective, and also caused more bleeding. It's a credible research lead that can't yet guide your own aspirin decision.

Why a flat average left room for a responder group

The earlier randomized evidence points the same way across trials and populations, and the new analysis doesn't overturn it. What it adds is a possibility the averages couldn't rule out: if a minority responds well and most people don't respond at all, the overall result can look like nothing. ASCEND's estimate was also loose enough that a modest benefit remained possible.

The 71% figure comes from the paper's fully adjusted model. The abstract reports a slightly different number, about 72% lower, from another model. Both describe a relative drop in dementia hazard, which is a different thing from your absolute chance of avoiding dementia.

One score out of 1,848, and a question about what it measures

Several things make this signal worth taking seriously. Aspirin was assigned at random, so within the high-score group the aspirin and placebo arms should be comparable. Dementia and bleeding were confirmed by expert panels who didn't know who took what. And the team applied a strict correction for testing many scores at once. But the scale of the search cuts the other way. They kept 1,848 published scores after quality checks and tested each for an interaction with aspirin. Three platelet-related scores passed the correction, and those three were highly correlated, so they don't amount to three independent confirmations. No independent replication has been published, and the authors reported no comparable prevention trial combining aspirin randomization, confirmed dementia diagnoses, and genotyping. Almost all participants were of European ancestry and 70 or older, so whether the result holds for younger people or other ancestries is unknown.

There's also a basic labeling problem. The paper's abstract describes the key score, PGS003548, as related to platelet count. The public registry that catalogs genetic scores and the paper's own results section identify it as predicting mean platelet volume, which is platelet size. Either way, the study grouped people by a genetic score and didn't measure their platelets, so a high platelet count on a routine blood test is not the selection rule this study used, and no cutoff has been clinically validated for choosing aspirin to prevent dementia.

The biology is plausible. In the Framingham Heart Study, people whose platelets reacted more strongly in midlife were more likely to develop dementia decades later, though that was an observational link. And APOE, the best-known dementia risk gene, wasn't associated with a significant difference in aspirin response, which is consistent with a platelet-related explanation but doesn't establish it. None of this validates using a genetic score to choose aspirin.

In the high-score group, fewer dementia diagnoses and more serious bleeds

Over a median of about 4.6 years, major bleeding occurred in about 4.4% on aspirin versus 2.1% on placebo. In ASPREE, major bleeding included symptomatic bleeding inside the skull or bleeding elsewhere that led to hospital admission, a longer hospital stay, surgery, a transfusion, or death, and across the trial most were gastrointestinal or inside the skull. Deaths from any cause didn't differ significantly between the aspirin and placebo arms in this group.

If aspirin could spare you dementia, that would matter enormously. A stomach bleed treated in hospital and a brain bleed are very different events, and you might reasonably weigh a few years free of dementia above some bleeding risk. Making that personal trade requires numbers you can trust, though, and a subgroup found by screening 1,848 scores, in one trial, with an unresolved score label, can't supply them yet. Lower-scoring groups didn't show a significant benefit, but that doesn't prove aspirin did nothing for them either.

What you can act on for your memory now

Blood pressure is the lever with randomized support. A large pooled analysis of individual patient data from blood pressure trials found that treating hypertension in late midlife and later life lowered the risk of dementia. If your readings run high, home measurements over a few weeks give you a clear picture, and starting or adjusting treatment is a decision for you and a prescribing clinician.

If you already take aspirin, this study doesn't settle whether you should continue. The U.S. Preventive Services Task Force recommends against starting aspirin to prevent a first heart attack or stroke at 60 or older, but that's a cardiovascular recommendation, it doesn't cover people with established heart or vascular disease, and it advises current users to talk with their clinician before stopping. Your reason for taking aspirin and your own bleeding risk should drive that conversation.

For the genetic finding to become a real option, the paper's conflicting platelet-count and platelet-size labels have to be resolved first. Then an independent dataset would need to show the same interaction, and a prospective randomized trial would need to pick people using a prespecified score and cutoff and report absolute dementia and major-bleeding outcomes, how severe the bleeds were, and enough follow-up to judge the trade. Younger and more diverse participants would have to be included before the result could apply to them.