A new study of two of the world’s largest health databases found that people taking statins, the common cholesterol-lowering pills, were diagnosed with dementia about twice as often as people who did not take them, even after the researchers adjusted for differences between the groups1. The study was published in Frontiers in Pharmacology1. It looked back at health records rather than testing the drug in a trial, and the authors themselves say the findings should not yet change how anyone thinks about statin therapy1.
What the study found
The team, led by researchers at Ewha Womans University and Chungbuk National University in South Korea, first studied 83,355 adults from the All of Us Research Program, a US National Institutes of Health project. Of these, 31,513 took statins1. They then checked the result in a second group: 173,282 people from the UK Biobank, including 43,976 statin users1.
In the US group, there were 5,968 new dementia cases. Statin users had about 2.2 times the risk of non-users after adjustment1. Put in numbers people can picture, there were about 65 new dementia cases for every 10,000 statin users followed for a year, against about 31 among non-users1. The link looked similar for Alzheimer’s disease, for vascular dementia (dementia linked to problems with blood flow in the brain) and for other types1. It showed up in every age group and regardless of whether people carried APOE4, a well-known gene variant that raises Alzheimer’s risk1.
In the UK group, dementia was rarer overall: about 12 new cases per 10,000 statin users per year versus about 6 among non-users, which worked out to about 2.5 times the risk after adjustment1. The authors partly put the lower UK rates down to shorter follow-up1.
The researchers also reported a dose pattern. People on high-intensity statins had about 2.7 times the risk of non-users, while people on moderate or low intensity had about 2.1 times. Compared directly with each other, high-intensity users had 14% higher risk than lower-intensity users1.
How the comparison was made
The study included adults aged 40 or older who described themselves as White1. People who already had dementia, a stroke, Parkinson’s disease or a brain injury at the start were left out, and new statin users were compared with non-users1.
The two groups were very different to begin with. In the US data, statin users were more often male and older, and far more of them had other illnesses: high blood pressure in 49.5% of users versus 16.1% of non-users, diabetes in 18.4% versus 3.7%, and heart artery disease in 17.5% versus 3.6%1. To deal with this, the team used a statistical weighting method that tries to make the groups look alike on measured traits, and then adjusted for sex, age, body weight, the APOE gene, other illnesses and other medicines1. Three extra checks with different methods left the results in place in both groups1.
The team also ran an exploratory analysis in about 38,553 UK Biobank participants with blood protein data, 6,108 of them statin users1. They looked at COQ7, a protein involved in making a molecule called coenzyme Q10, as a stand-in because coenzyme Q10 itself is not measured in these datasets1. Statin users had somewhat lower COQ7 levels, and this path accounted for 8.2% of the overall link; a path through LDL (“bad”) cholesterol was not statistically significant1. This is a statistical pattern, not a measured mechanism.
What it does not show
This is an observational study: it watched what happened to people who chose or were prescribed statins, rather than assigning the drug at random. Such a design cannot show cause1. Weighting can only balance the traits that were measured. Statin users started out much sicker, and anything unmeasured that led doctors to prescribe statins, and that also raises dementia risk, could produce a link like this one1.
That point has been made by other experts about earlier statin studies. Commenting in 2016 on a study that linked statins to lower Alzheimer’s risk, Prof Jane Armitage of the University of Oxford said2:
The people prescribed statins are likely to be different in many ways from those not prescribed statins. As a result, conclusions about cause and effect cannot be made despite the authors trying to take account of such differences.
Armitage declared that she has been involved in running large randomized trials of statins and works at Oxford’s Clinical Trial Service Unit, which has received research grants from companies including statin makers Pfizer, Merck and AstraZeneca, under university contracts meant to protect its independence; the unit has a policy of staff not taking personal payments from industry2.
The new study also covers only White adults aged 40 or older, so it may not apply to other groups1. The coverage Biohack read does not say who funded the work or whether the authors reported any conflicts of interest.
How it fits with earlier research
Statins and dementia have been studied for decades, with results pointing in different directions. One early study using UK family doctor records, published in The Lancet in 2000, found that people aged 50 and older who were prescribed statins had a substantially lower risk of dementia34.
Randomized trials, where people are assigned a statin or a placebo (a dummy pill) by chance, have not backed up a protective effect. A summary by McMaster University of a systematic review described two such trials with 26,340 people aged 40 to 82, followed for three to five years, and found no evidence that statins prevented cognitive decline or dementia4. Armitage said in 2016 that trials of statins lasting five to six years did not suggest protection against cognitive decline2. Those trials, as summarized, also did not show statins causing dementia4.
More recent record studies have again pointed toward lower risk in some groups:
- A Danish study of 132,585 people newly diagnosed with type 2 diabetes, presented at the ESC Congress 2026, found that starting a statin within a year of diagnosis was linked to a 15% lower relative risk of dementia over 10 years5. In absolute terms, estimated 10-year risk was 3.3% with an early start versus 3.9% with no statin6. An ESC spokesperson noted it was observational and cannot prove that statins prevent dementia5.
- A Swedish registry study of 15,586 people who already had Alzheimer’s or mixed dementia found statin users had slightly slower decline on a memory and thinking test; the authors said they could not determine cause7.
A review of 560 articles on statins and dementia published between 2007 and 2023 concluded that statins’ role remains uncertain, citing reports of short-term thinking problems and questionable long-term protection8. A health information site notes that the US Food and Drug Administration updated statin labels in 2012 to mention rare reports of short-term memory loss and confusion, which it describes as uncommon and usually resolving after stopping the drug9.
What to watch next
The new study’s authors stress that the observational design and the conflicting state of earlier research mean their results are not a basis for changing statin use1. Experts have long said that only randomized trials can settle whether statins affect dementia risk, though long trials of this kind are hard to run2. Things to watch for include whether other teams can repeat the doubled-risk finding in different populations, whether the full paper reports funding and conflicts, and whether the coenzyme Q10 idea is tested directly. Biohack will update this thread as new studies appear.
