Alzheimer’s disease is often spoken about as an inevitable part of ageing, something that slowly emerges from a complex mix of lifestyle, environment, and chance. Over the past three decades, however, scientists have known that our genes play a major role in shaping who is most at risk. Among these genes, one stands out more than any other: APOE. A large new analysis brings this long-known fact into sharper focus and asks a simple but powerful question: how much of Alzheimer’s disease and dementia can actually be traced back to common variations in this single gene?
APOE is a gene involved in how the body transports fats and cholesterol, including in the brain. Humans commonly carry one of three versions, called ε2, ε3, and ε4. Almost everyone has two copies in total, one inherited from each parent. The ε3 version is by far the most common worldwide, ε4 is less common but well known for increasing Alzheimer’s risk, and ε2 is relatively rare and appears to be protective. For years, ε4 has been described as the “bad” Alzheimer’s gene, ε2 as the “good” one, and ε3 as neutral. This study challenges that comforting simplification.
Researchers analysed genetic and health data from nearly half a million people drawn from several large studies in the UK, Finland, the US, and elsewhere. Some participants were followed through electronic health records to see who developed Alzheimer’s disease or other forms of dementia. Others had brain scans that detect amyloid, a sticky protein that accumulates early in Alzheimer’s. A smaller but crucial group had brain tissue examined after death, allowing researchers to confirm Alzheimer’s pathology directly. By combining these different sources, the authors aimed to reduce the uncertainty that comes from relying on just one method of diagnosis.
A key strength of this work lies in how the comparisons were made. Instead of treating the common ε3 version as the “normal” baseline, the researchers compared everyone to the lowest-risk group: people with two copies of ε2. This matters because ε3 is not truly neutral. When risk is measured against ε2, it becomes clear that ε3 itself raises the likelihood of Alzheimer’s, just not as dramatically as ε4.
The results are striking. Depending on the dataset and how Alzheimer’s was defined, between roughly 70% and more than 90% of Alzheimer’s disease could be statistically attributed to carrying ε3 or ε4 rather than ε2. Even when looking more broadly at all-cause dementia, not just Alzheimer’s, about half of cases were linked to these two common gene variants. In brain imaging studies, over 80% of amyloid build-up could be attributed to ε3 and ε4. In simple terms, without the risk conferred by these very common forms of APOE, most Alzheimer’s disease would probably not occur.
Perhaps the most surprising insight is the role of ε3. While ε4 still accounts for the largest share of risk, ε3 contributes substantially simply because almost everyone carries it. When the researchers separated the effects, ε3 alone may account for roughly a third of Alzheimer’s burden in some analyses. This does not mean ε3 is “bad” in an absolute sense, but it is far from harmless when compared to ε2.
It is important to be clear about what these findings do and do not mean. Alzheimer’s disease is not caused by a single gene in the way that some rare childhood disorders are. Lifestyle, education, vascular health, inflammation, and many other biological pathways still matter. What this study shows is that APOE sits at the centre of this web. Many other risk factors seem to exert their effects on a background shaped by APOE. If that background risk were lowered, the impact of many other factors might shrink dramatically.
The authors also compared APOE with other genetic risk factors discovered through large genome-wide studies. No other single gene comes close to APOE in terms of the proportion of disease it helps explain. This makes Alzheimer’s somewhat unusual among common diseases and highlights why targeting APOE could be so powerful.
Therapeutically, this has major implications. Most current Alzheimer’s drugs focus on removing amyloid from the brain. Their benefits so far have been modest and uncertain. APOE, by contrast, influences amyloid build-up, clearance, inflammation, and brain cell health more broadly. In theory, modifying APOE function could reduce risk earlier and more fundamentally. Approaches being explored include antibodies, small molecules, and even gene-based strategies, although only a handful have reached human trials so far.
This brings us to a broader question that extends beyond this study. As personal whole genome sequencing becomes more accessible, many people will learn which APOE variants they carry long before any symptoms appear. Could this knowledge guide prevention, early intervention, or access to new therapies as they emerge? And at a population level, could tailoring treatments to genetic risk fundamentally change how we approach Alzheimer’s disease, shifting the focus from late-stage care to early, personalised prevention? These are questions that society, clinicians, and researchers will need to confront sooner than we might expect.
References:
Williams, D.M., Heikkinen, S., Hiltunen, M. et al. The proportion of Alzheimer’s disease attributable to apolipoprotein E. npj Dement. 2, 1 (2026). https://doi.org/10.1038/s44400-025-00045-9
