What a 15-Year Swedish Study Found About Red Meat and the 'Alzheimer's Gene'

When a patient learns they carry a copy of APOE ε4, the version of the apolipoprotein E gene that raises Alzheimer's risk, the conversation that follows tends to be about acceptance. We talk about screening, about planning for the family, about the handful of drugs that slow the decline a little. We almost never talk about what the patient ate yesterday, or what they plan to eat tomorrow. A large Swedish study published in early 2026 suggests that silence is a mistake, because among the people carrying that high-risk gene, diet tracked with whether they kept their memory or lost it. So what did the study find, and why might a plate of meat matter more to one person's brain than to another's?

What the Swedish study found

Norgren and colleagues, writing in JAMA Network Open, followed 2,157 dementia-free adults aged 60 and older from the Swedish National Study on Aging and Care for up to 15 years, tracking their diet, their APOE genotype, and the path their thinking took over time [1]. About a quarter of them, 569 people or 26.4%, carried at least one ε4 allele, close to the rate you'd expect in a population of Northern European descent.

The headline finding sits entirely inside that ε4 group. Carriers who ate the most unprocessed meat, the top fifth of intake, had roughly 55% lower dementia risk than carriers who ate the least, a subhazard ratio of 0.45 where 1.0 would mean no difference (95% confidence interval 0.21 to 0.95), and their cognition held up measurably better over the years that followed. In people who didn't carry ε4, eating more or less meat made no real difference in either direction.

The most striking part is what happened to the gap between the genotypes. At the lowest meat intake, ε4 carriers carried more than double the dementia risk of non-carriers, the inherited penalty everyone expects. At the highest intake, that gap all but vanished, and cognitively, ε4 carriers eating plenty of unprocessed animal food looked much like people who never carried the risky gene at all.

Honesty about the statistics matters here. The formal test of whether genotype changed the diet effect was strong for cognitive trajectory (p = 0.004) but only a trend for dementia itself (p = 0.10), so the cognition result is the firmer of the two. The whole thing is also observational, which means it can show that meat and a better-aging brain travel together in ε4 carriers, not that the meat caused it.

Why this finding is hard to wave away

There's a reason this particular result is hard to explain away, and it has to do with the kind of people who eat the food. In observational nutrition research, the people who eat the most vegetables also tend to exercise more, smoke less, drink less, and sleep better, so when vegetables line up with good health you can never fully separate the broccoli from the lifestyle around it. That clustering is the healthy-user effect, the tendency for health-conscious people to pile every good habit together. Red meat runs the opposite way. In most Western datasets, heavy red-meat eaters don't score better on those other habits, and if anything score a little worse, so when red meat lines up with a 55% drop in dementia risk, the easy confounders are pushing against the finding rather than propping it up. That asymmetry doesn't prove cause, yet it makes the result much harder to dismiss as an artifact of who happens to eat what.

The gene isn't broken, it's ancient

APOE ε4 is the ancestral allele, the original version of the gene that every human once carried before the others existed. The ε3 and ε2 variants, the ones tied to lower Alzheimer's risk, are evolutionary newcomers that spread later, alongside agriculture and the dietary shift that came with it. For most of the long stretch that built the human brain, ε4 was the human gene, carried by people who hunted large game, ate it nose to tail, and leaned on dense animal food at nearly every meal.

That history reframes what the "Alzheimer's gene" is. It isn't a defect that crept into the genome. It's old code, calibrated for a particular diet, now running in an environment that no longer feeds it the way it expects. Trumble and colleagues, in a 2017 study of Amazonian forager-horticulturalists living with heavy parasite exposure and a traditional diet, found that older ε4 carriers there held their cognition steady or even edged ahead of non-carriers, the reverse of the pattern we see in modern cities [2]. The same gene behaves differently depending on the world you run it in. Feed ε4 something close to the diet it evolved alongside, and the disadvantage we treat as fixed starts to look conditional.

What the ε4 brain is asking for

If meat helps the ε4 brain, the question is what in it does the work, and the best answer isn't "meat" as such but the cluster of nutrients packed into unprocessed animal food.

Vitamin B12 is the obvious one. The brain needs it to build myelin, the fatty insulation around nerve fibers, and to run methylation, the everyday chemistry that, among other jobs, keeps homocysteine in check. Homocysteine is an amino acid that climbs when this machinery falters, and high levels are toxic to neurons and tied to faster brain shrinkage. B12 lives in animal foods and is essentially absent from unfortified plants. In a randomized trial, Smith and colleagues gave older adults with mild cognitive impairment B vitamins to lower homocysteine and slowed the rate of brain atrophy compared with placebo, with the benefit concentrated in those who started out with high homocysteine [3]. The ε4 carriers in the Swedish study who ate more meat also carried more B12 in their blood, which is unlikely to be a coincidence.

Creatine is the underrated one. Most people know it as a powder gym-goers stir into water, but the brain is one of its biggest customers, because creatine helps regenerate ATP, the molecule cells spend for energy. This matters for Alzheimer's in particular, since the Alzheimer's brain shows a broken energy metabolism, struggling to fuel itself long before symptoms appear. Ostojic and colleagues, analyzing US national survey data, found that older adults eating more than about 0.95 grams of creatine a day scored better on cognitive testing than those eating less [4], and roughly a gram of dietary creatine maps to two or three servings of red meat. We've written more about creatine in older adults, where the evidence for both muscle and memory is some of the strongest creatine has.

Beyond those two, unprocessed animal food supplies carnitine, heme iron, zinc, taurine, choline, and anserine, each tied in some way to brain energy and upkeep, and each more concentrated in organs than in muscle meat. Liver, heart, and kidney are the densest packages of brain-relevant nutrients on the plate, which is part of why organ meats keep earning their place in an ancestral diet.

This isn't a case for eating only meat

None of this argues that everyone should eat nothing but steak. The high-intake group in the study wasn't carnivore, they ate a median of about 870 grams of meat a week, standardized to a 2,000-calorie day, which works out to roughly 4 ounces a day, two modest servings rather than a mountain. That's a normal eating pattern with meat near the center of it, not the elimination of every plant on the plate.

Processed meat pulled in the other direction. The more of a person's meat came from processed sources, the bacon and deli slices and sausage, the worse the outcomes, and a lower processed share lined up with lower dementia risk across every genotype, not only in ε4 carriers. Unprocessed meat helped, processed meat hurt, and the genotype decided who stood to gain the most.

Why your neurologist isn't having this conversation

So why isn't this the discussion in the exam room? The Alzheimer's conversation in conventional medicine tends to orbit three things, the genetic risk you can't change, the medications that work poorly, and the lifestyle factors that get named in passing and then dropped. I trained inside that system, and I don't remember anyone raising the possibility that we might be running ancestral hardware on industrial fuel, and that the people hit hardest might be the ones least suited to the modern diet. The conversation after a positive test is about acceptance and planning. This study suggests it should also be about the next meal.

What the data points to

If you carry an ε4 allele, or Alzheimer's runs in your family, or you want to feed your brain what it was built for, the data points in a fairly clear direction.

• Build the plate on unprocessed animal food. Make ruminant meat, beef, lamb, or bison, the foundation, a couple of palm-sized servings across the day, with eggs and, if you tolerate it, dairy. This is moderate intake, not a contest to eat as much as possible.
• Eat organs. Liver, heart, and kidney carry more B12, heme iron, zinc, carnitine, taurine, and, in heart especially, creatine than muscle meat does. If cooking liver isn't for you, a desiccated organ supplement is a reasonable substitute.
• Cut the processed meat. The cured, smoked, and packaged versions tracked with worse outcomes for everyone in the study, so let your meat come from a cut, not a packet.
• Avoid the modern diet wholesale. Ultra-processed food, refined carbohydrate, and industrial seed oils are the fuel ε4 never adapted to, and they're worth cutting whether or not you carry it.
• Know your status if it would change what you do. APOE genotyping comes with most consumer genetic tests, and if you sit in a higher-risk group, Northern European ancestry, a family history, female and past menopause, knowing can be the thing that makes you take the food seriously.

The reflex worry here is that loading up on red meat trades your brain for your heart. I've argued at length that this fear is largely misplaced, that saturated fat is a fairly neutral player rather than the driver of heart disease, and that the real damage to arteries comes from insulin resistance, high blood sugar, chronic inflammation, and oxidized seed-oil fats. Your blood ApoB and LDL particle burden still belong in the picture, so track them with your physician, and never stop a prescribed medication on the strength of a blog post. The point isn't to ignore your heart, it's that whole, unprocessed meat is not the threat to it we were taught it was.

Change the environment, change the outcome

The genetic disadvantage of carrying the "Alzheimer's gene" didn't hold when carriers ate close to the diet their genes evolved with. This means a vulnerability the medical system describes as fixed may instead be a vulnerability to a modern environment that no longer matches our biology, and an environment is something you can change. None of this is settled, because the study is observational and needs confirmation from trials that assign people to diets, so hold it as a strong lead rather than a closed case. The move it points to, though, costs little and risks less. Eat real food, mostly the animal foods we evolved on, eat the organs, and skip the processed shortcuts. Do that for the decades that matter, and the brain you bring to 80 may not be the one your APOE result told you to expect.

See how our metabolic program works

References

1. Norgren J, Carballo-Casla A, Grande G, et al. Meat Consumption and Cognitive Health by APOE Genotype. JAMA Netw Open. 2026. PMID: 41854609
2. Trumble BC, Stieglitz J, Blackwell AD, et al. Apolipoprotein E4 is associated with improved cognitive function in Amazonian forager-horticulturalists with a high parasite burden. FASEB J. 2017. PMID: 28031319
3. Smith AD, Smith SM, de Jager CA, et al. Homocysteine-lowering by B vitamins slows the rate of accelerated brain atrophy in mild cognitive impairment: a randomized controlled trial. PLoS One. 2010. PMID: 20838622
4. Ostojic SM, Korovljev D, Stajer V. Dietary creatine and cognitive function in U.S. adults aged 60 years and over. Aging Clin Exp Res. 2021. PMID: 33866527