News · Nutrition & Diet
Plant-based low-carb eating tracked 15% less dementia over nine years, whatever the genetic risk
A low-carbohydrate diet and a low-fat diet look symmetrical on paper. Across 5,301 Americans followed for nine years they were not: the low-carb score tracked less dementia, the low-fat score did not, and the food sources decided both.
- A lower-carbohydrate way of eating tracked about 15% less dementia over nine years.
- A low-fat pattern tracked nothing, so the two are not mirror images.
- Where the food came from decided it: plant-based and unrefined versions did best.
- An unhealthy low-fat pattern, heavy on refined carbohydrate, tracked more dementia.
- The pattern held whatever a person's inherited risk, across 5,301 older Americans.
Low-carb and low-fat are usually argued about as if they were opposites, and as if picking one settled something.
They are not opposites. Eating less of one macronutrient means eating more of the others, and the interesting question was never which lever you pull. It is what you pull it toward.
Writing in the American Journal of Clinical Nutrition, researchers followed 5,301 dementia-free adults aged 55 and over from the Health and Retirement Study for nine years, scoring each person on how low-carbohydrate and how low-fat their reported eating was.
A greater low-carbohydrate score was associated with lower dementia risk. The low-fat score was not associated with anything.
What a diet score actually measures
Nobody in this study was put on a diet.
The scores are rankings. Overall, low-carbohydrate and low-fat indices were constructed based on macronutrient composition rankings assessed using a food frequency questionnaire, which means each person was placed relative to everyone else on how much of their energy came from carbohydrate, fat and protein.
A high low-carbohydrate score in a cohort of older Americans is not a ketogenic diet. It is someone eating rather less bread, rice and sugar than their neighbors, and correspondingly more of everything else. That is the exposure this result is about.
The part that decides the answer
The overall scores are the least interesting numbers in the paper, and the subscores are where the finding lives.
Plant-based and healthy low-carbohydrate diets showed stronger inverse associations, at roughly 15% and 18% lower risk for each standard deviation of score. Meanwhile, higher animal-based and unhealthy low-fat diets were linked to higher dementia risk, the unhealthy version by about 13%.
So the two headline scores point in opposite directions, but the split within each is larger than the gap between them. A low-carbohydrate pattern built on meat and cheese did not carry the benefit. A low-fat pattern built on white bread and sugar carried harm.
The authors’ conclusion follows directly: adherence to low-carbohydrate diets, particularly plant-based and higher-quality ones, is what the association attaches to.
Good nutrition is about healthy eating, which means regularly choosing healthy foods and beverages. That sentence has never sounded like a finding. Here it is roughly what the data say.
The genetics test that matters
A diet finding that only worked in people at low genetic risk would be a curiosity. This one was checked.
Genetic susceptibility was assessed using APOE genotype and an Alzheimer disease polygenic risk score, the two standard measures of inherited risk. The dietary associations were consistent across APOE genotype and polygenic risk strata.
That is a meaningful result in both directions. It means carrying the high-risk gene did not cancel the dietary pattern, and it also means the pattern was not doing anything special for the people who needed it most. Diet and genes appear to be operating separately rather than one modifying the other.
Why nine years is the number to watch
Dementia begins changing behavior long before anyone gets a diagnosis, and eating is one of the first things it changes.
A person in the earliest stage of cognitive decline shops differently, cooks less, and drifts toward food that requires no decisions, which in a Western diet means refined carbohydrate. If that happens for five years before diagnosis, a study will find that people who ate more refined carbohydrate developed more dementia, and the arrow will be pointing backwards.
During the 9-year follow-up, 506 participants developed dementia. Nine years is long enough to push most of that effect out of the window, and not long enough to guarantee it. The diet was measured once, in 2013 to 2014, by questionnaire, and questionnaires record what people believe they eat.
What to do with it
Not a carbohydrate ban, which is what a headline about this would usually produce.
There is no cure for most types of dementia, including Alzheimer’s disease and Lewy body dementia, and the absence of treatment is exactly what makes weak prevention findings so tempting to over-read. This is a single observational cohort with self-reported food.
What it supports is narrow and unexciting: the source of your fat and protein appears to matter more than the ratio. The pattern that did best in this cohort was less refined carbohydrate, with the replacement coming from plants rather than animals, which is the same eating pattern already recommended for the heart, and a healthy eating plan includes eating a variety of foods, including vegetables, fruits, and whole-grain products.
The genuinely new part is the negative half. A low-carbohydrate diet is not protective by virtue of being low-carbohydrate, and a low-fat diet is not protective by virtue of being low-fat. Both can be built badly.
People also ask
What did the study find?
A greater low-carbohydrate diet score was associated with lower dementia risk (HR per SD 0.90; 95% CI 0.82, 0.99), whereas an overall low-fat diet was not (HR per SD 1.05; 0.96, 1.15). Plant-based and healthy low-carbohydrate diets showed stronger inverse associations (0.85; 0.78, 0.94 and 0.82; 0.74, 0.90), whereas higher animal-based (1.10) and unhealthy low-fat diets (1.13; 1.04, 1.24) were linked to higher dementia risk.
What is a low-carbohydrate diet score?
Not a diet anyone was told to follow. It is a ranking built from what people reported eating: how their share of calories from carbohydrate, fat and protein compares with everyone else's. A high score means a relatively low-carbohydrate pattern, not a ketogenic one.
What do plant-based and healthy mean here?
Subscores that separate where the macronutrients came from. A plant-based low-carbohydrate pattern gets its fat and protein from plants rather than animals. A healthy version additionally penalizes refined grains and added sugars. Both scored better than the overall version.
Why does the genetic result matter?
Because a diet effect that only appeared in people without APOE risk alleles would be much less useful. Associations were consistent across APOE genotype and polygenic risk strata, which means the pattern was not confined to the genetically lucky.
Does this mean cutting carbs prevents dementia?
No. This is a cohort study built on a food frequency questionnaire answered once. It shows that people whose reported eating leaned a particular way developed less dementia, which is a different claim from one about what happens if you change.
What is the reverse-causation worry?
Dementia changes what people eat years before diagnosis. Someone in the earliest stages may drift toward convenient, refined, carbohydrate-heavy food, which would produce this association with no dietary effect at all. Nine years of follow-up reduces that risk without eliminating it.
What is the practical version?
It is not a low-carb prescription. The pattern that did best was fat and protein from plants with fewer refined grains and less added sugar, which is close to advice that already exists for the heart. This is general information rather than medical advice.