Meat, APOE4, and Dementia Risk: What the Study Found

Most dietary advice treats meat, and especially red and processed meat, as something to limit. So a 2026 study in JAMA Network Open drew attention when it reported the opposite pattern in one specific group: older adults who carry the APOE4 gene variant, the strongest common genetic risk factor for Alzheimer's disease. In that group, eating more meat was linked to better long-term thinking scores and fewer dementia diagnoses (Norgren et al., 2026).

That sounds like a headline in search of a diet, so it is worth slowing down. The study was observational, it followed a mixed-diet Swedish population rather than carnivore dieters, and its authors are careful about what it can and cannot show. Here is what the researchers did, what they found, and why the results are interesting without being proof that meat protects the brain.

What is APOE4?

APOE, short for apolipoprotein E, is a gene carrying the instructions for a protein your body uses to move cholesterol and other fats through the blood (National Institute on Aging). It comes in three common forms, called e2, e3, and e4. Because you inherit one copy from each parent, people end up with pairs such as e3/e3, e3/e4, or e4/e4.

The e4 form matters here because it is the strongest common genetic risk factor for Alzheimer's disease. Carrying one copy raises the odds compared with the most common e3/e3 pairing, and carrying two raises them further. About a quarter of people carry at least one e4 copy, and a smaller share, roughly 2 to 3 percent, carry two.

That is risk, not fate. A gene variant that raises probability is not a diagnosis. Many people who carry e4 never develop dementia, and many people who develop dementia carry no e4 at all. In this study, participants with e3/e4 and e4/e4 were grouped together and labeled APOE34/44, then compared with everyone else.

Why researchers looked at meat and genotype together

Studies asking whether meat helps or harms the aging brain have reached mixed, sometimes contradictory conclusions. One explanation for the muddy picture is that people may not all respond to the same food the same way, and genetics could be part of the reason. This team had a hypothesis set out in advance: that higher meat intake might be more beneficial, or at least less harmful, for people carrying e4 than for everyone else (Norgren et al., 2026).

Part of their reasoning drew on an evolutionary argument that e4 is the oldest human form of the gene and may have suited diets heavier in animal foods, an idea that connects to long-running debates about how much meat early humans ate. Those claims are hypotheses, not established fact, but they were why the researchers looked for a gene-by-diet interaction rather than one answer for everyone.

How the study was conducted

The data came from the Swedish National Study on Aging and Care in Kungsholmen, a long-running population study in central Stockholm. Between 2001 and 2004, researchers enrolled adults aged 60 and older, and this analysis included 2,157 of them who did not have dementia at the start. Their average age was about 71, roughly 62 percent were women, and about 569, or 26.4 percent, carried the e3/e4 or e4/e4 genotype (Norgren et al., 2026).

Participants were followed for up to about 15 years. Diet was measured with validated food-frequency questionnaires at the start and at some later visits, with meat intake expressed relative to each person's total calories so large and small eaters could be compared fairly. Thinking and memory were tested repeatedly across four areas: episodic memory, semantic memory, verbal fluency, and perceptual speed. Dementia was diagnosed through a multi-step clinical process involving more than one physician.

Of the group, 1,680 people had enough repeat cognitive testing to track their trajectory over time, and 296 developed dementia during follow-up. Because many older adults die before a dementia diagnosis can occur, the researchers treated death without dementia as a competing event rather than ignoring it, which keeps the dementia estimates more honest.

What the researchers found

Among participants with the e3/e4 or e4/e4 genotype, those in the highest fifth of meat intake had a more favorable cognitive trajectory than those in the lowest fifth (β = 0.32; 95% CI, 0.07 to 0.56; P = .01), and a lower rate of dementia (subdistribution hazard ratio 0.45; 95% CI, 0.21 to 0.95; P = .04). In everyone else, neither association was statistically significant (Norgren et al., 2026).

A hazard ratio of 0.45 can be read as roughly a 55 percent lower dementia rate in the model, comparing the highest with the lowest meat group. That is easy to misread, so precision matters: it is a modelled association within one population, not a proven 55 percent reduction any individual can expect. The confidence interval (0.21 to 0.95) is wide, so the true figure is uncertain even inside the study.

The most striking part is not that meat looked protective on its own, but where the usual genetic penalty went. At the lowest meat intake, the e4 group carried roughly two and a half times the dementia risk of other genotypes, the disadvantage you would expect. At the highest meat intake, that gap essentially disappeared and was no longer statistically distinguishable.

Cognitive decline versus dementia risk

The evidence was not equally strong for both outcomes, and this is the single most important caveat to hold onto. The test for whether the meat effect truly differed by genotype was convincing for cognitive decline (P for interaction = .004) but did not reach the usual threshold for diagnosed dementia (P for interaction = .10) (Norgren et al., 2026). In plain terms, the case that genotype changes the meat-and-brain relationship is firmer for the gradual slope of memory and thinking scores than for the harder endpoint of a dementia diagnosis. The study did not prove a genotype-specific dementia-prevention effect.

It also helps to keep two terms straight. Dementia is an umbrella term for a loss of memory, reasoning, and other thinking skills serious enough to interfere with daily life, and it has several causes (National Institute on Aging). Alzheimer's disease is the most common of those causes, but not the only one. This study counted all-cause dementia and tracked cognition; it did not confirm Alzheimer's specifically. APOE4 is an Alzheimer's risk factor, and early episodic-memory decline is a classic feature of Alzheimer's, so the outcome is relevant to it, but the two are not interchangeable.

Processed and unprocessed meat were not the same

The study also pushed back on the habit of lumping all meat together. When the researchers looked at how much of a person's meat was processed, a higher processed share was linked to higher dementia risk (subdistribution hazard ratio 1.14; 95% CI, 1.01 to 1.29; P = .04), and that association did not depend on APOE genotype (Norgren et al., 2026). The favorable signals were concentrated in unprocessed meat, and the balance between unprocessed red meat and poultry did not seem to matter for cognition.

Processed meat generally means meat preserved by salting, curing, smoking, or fermentation, such as bacon, sausages, ham, and deli meats, though definitions differ between studies, which makes comparisons rough. The honest summary is narrow: in this cohort, unprocessed meat tended to look more favorable than processed meat. That is not the same as saying every fresh cut is protective, or that any processed product single-handedly causes dementia.

How much meat did participants eat?

It is natural to ask how much meat the top group was eating. Standardized to a 2,000-calorie day, the lowest fifth averaged roughly 200 grams of total meat per week, while the highest fifth averaged around 930 grams per week, or about 130 grams a day. The researchers noted that intake in the higher groups clearly exceeded Nordic dietary recommendations, with the top group eating more than twice the suggested level (Norgren et al., 2026).

Two things keep this from being a meal plan. The amounts are energy-standardized comparisons designed to rank people within this study, not a therapeutic dose, and even the highest group was eating an ordinary mixed diet alongside that meat. These quintiles describe how this Swedish cohort happened to eat; they are not a target, and specifically not a recommendation for how much meat someone with APOE4 should eat.

Possible explanations

Why might the same food track with different outcomes depending on a gene? The researchers offered several possibilities and were careful to frame them as exploratory rather than settled.

One hypothesis involves nutrient handling. In a post hoc analysis, the team used the link between dietary and blood vitamin B12 as a rough proxy for absorption, and saw a pattern hinting that people with e3/e4 or e4/e4 drew more B12 from meat as intake rose. This is a speculative signal, not proof of superior absorption, and B12 is only one of many nutrients meat supplies (see the nutrient content of meat). Other ideas include the food matrix, meaning the whole package a nutrient arrives in, and the chance that foods replacing meat, chiefly cereals and dairy, carry compounds that interact with genotype.

There is also a lipid angle: in the e4 group, higher meat intake went along with lower blood cholesterol and a more favorable fat profile in this sample, which runs against the usual assumption. And there is the evolutionary hypothesis mentioned earlier. None of these mechanisms has been established. The authors present them as directions for future research, not conclusions, and the measured findings matter more than the proposed reasons for them.

Strengths of the study

The study has real strengths. It followed a population-based group for up to 15 years, tested cognition repeatedly rather than once, used repeated dietary data for some analyses, and diagnosed dementia clinically rather than from records alone. It treated death as a competing risk and adjusted for many potential confounders, including age, sex, education, physical activity, smoking, alcohol, overall diet quality, calories, chronic disease, living arrangements, occupation, and baseline cognition (Norgren et al., 2026). It ran sensitivity and triangulation analyses to check whether early, undiagnosed disease could be driving the results, and the genotype hypothesis was specified before the data were examined, which guards against inventing a story after the fact. None of this, though, turns an observed association into proof of cause.

Important limitations

The limitations deserve equal billing. The biggest is built into the design: this is an observational study, so it can show that meat intake and brain outcomes moved together, but not that one caused the other. Statistical adjustment reduces the influence of confounders but cannot remove every difference between people who eat more or less meat, and some residual confounding almost certainly remains (Norgren et al., 2026).

Several other cautions stack up. Diet was self-reported, which introduces measurement error. The genotype-specific findings rest on smaller subgroups and relatively few dementia events, which makes them less stable, and the gene-by-diet interaction for dementia did not reach significance, as noted above. The study ran many secondary and post hoc analyses without adjusting for multiple comparisons, so some results could reflect chance. Participants were older and predominantly of Northern European ancestry, a homogeneous group, which limits how far the findings extend to younger adults or other populations. The authors also flag possible survival bias, since people who reach old age without dementia may be unusually resilient. Because no proposed mechanism was tested directly, the biology remains uncertain. And, centrally, it did not test a carnivore diet or give any basis for prescribing a particular amount of meat.

Does this study support the carnivore diet?

Not in the way the headline might suggest. The study is genuinely relevant to conversations about meat and brain health, and it challenges the blanket assumption that more meat must always be worse. Those are fair points to take from it.

But it is not a carnivore study. Every participant ate a mixed diet, and even the highest-meat group also ate carbohydrates, dairy, and fiber; this was not an all-meat protocol. Much of the public conversation about carnivore eating still rests on self-reports and short-term experiences, such as the survey of more than 2,000 self-described carnivore dieters, rather than long-term outcomes, and this study does not fill that gap. It did not measure carnivore dieters, cannot tell us what strict carnivore eating does over years, and should not be read as evidence that carnivore prevents or treats dementia. What it adds is a data point to a narrower question: whether the meat-and-brain relationship might differ by genotype.

What should readers take from it?

The most useful takeaway is the least dramatic: this is a reason to study personalized nutrition and APOE more carefully, not a set of instructions. It suggests that the relationship between meat and cognitive health may depend on both processing and genetics, a genuinely interesting possibility that deserves more rigorous testing.

It is not enough, on its own, to change dementia-prevention guidance or to recommend a high-meat diet. If you carry APOE4 or suspect you might, this study does not tell you to eat more meat, and it does not tell you to stop managing blood pressure, cholesterol, or other cardiovascular risks, which matter for brain health regardless of genotype. Nobody should discontinue treatment or overhaul their diet on one observational study.

If you have real concerns, about memory, dementia risk, cholesterol, cardiovascular disease, or genetic testing, the right next step is a conversation with a qualified healthcare professional who knows your history, not a change made from one paper.

Bottom line

Among older adults with the APOE e3/e4 or e4/e4 genotype, eating more meat was associated with slower cognitive decline and less dementia, and the usual genetic disadvantage faded at the highest intakes. The pattern was clearer for cognitive decline than for dementia, applied mainly to unprocessed meat, and came from an observational study that did not test a carnivore diet and cannot prove cause. It is a reason to take gene-and-diet questions seriously and study them properly, not a reason to change how you eat today.