← Longevity & Supplement Guides
The case for eating less protein — and where it stops applying
We publish a guide on this site that tells you to eat more protein. I still think it is broadly right, and I am not taking it down. But on 31 July a review of more than 350 studies landed that makes the most serious case I have read for the opposite — and quietly ignoring it would be exactly the kind of thing Agen exists to not do.
The two positions look like a flat contradiction. They are not. They are careful answers to two different questions, asked of two different bodies. Working out which question is yours turns out to be more interesting than the protein.
Two papers, one season apart, pointing opposite ways
The first is a review by Bailey Knopf and Dudley Lamming at the University of Wisconsin–Madison, published in Cell Press Blue on 31 July 2026: The hallmarks of protein and amino acid restriction in aging and longevity. Its argument, assembled from over 350 papers: across a striking range of organisms, eating less protein tracks with better metabolic health, and in animals with longer life.
The second is a 2026 analysis in npj Aging by Hélio Coelho-Júnior and Emanuele Marzetti, following 532 adults aged 65 and over in the English Longitudinal Study of Ageing across six years. Its finding: the ones eating more protein fell less, lost less walking speed, became frail less often, and died less often.
Same nutrient. Same year. Opposite arrows. Read only the headlines and nutrition science looks like a coin toss — which is roughly what most people conclude, and then go back to whatever they were doing.
What restriction actually does, and who its own authors exclude
The mechanism is not vague hand-waving about "growth signals." When protein intake drops, a cellular sensor called GCN2 registers the shortfall and raises fibroblast growth factor 21, a hormone linked to higher energy expenditure, better fasting glucose and lower inflammatory signalling. At the same time mTORC1 — the pathway that tells a cell to build rather than tidy — quiets down, and autophagy, the cell's recycling programme, picks up.
What makes the review genuinely surprising is that it may not be protein as a whole doing this. Restricting three specific amino acids — methionine, isoleucine and valine — reproduces much of the effect in animal models. That reframes the question from how much to which, and it is a far more interesting question. In human trials the authors cite, people on protein-restricted diets lost weight and fat mass and improved fasting glucose while eating more total calories. That is a strange and real result.
Now the part the coverage tends to bury. The authors themselves write that needs are highly individual, that older adults facing age-related muscle loss and pregnant women often require more protein, and that physically active people appear to tolerate higher intakes without the metabolic penalty, because exercise seems protective. That is not a disclaimer bolted on at the end. It is the shape of the finding.
The study that drew the line at 65
The age-dependence has been visible for over a decade. In 2014, Morgan Levine and colleagues published a Cell Metabolism analysis of 6,381 US adults aged 50 and over in NHANES III. Among people aged 50 to 65, a high-protein diet — over 20% of calories — carried a cancer-mortality hazard ratio of 4.33 compared with a low-protein diet. A frightening number, widely repeated.
Far less repeated: in the same dataset, among people aged 66 and over, the association ran the other way. High protein there tracked with lower all-cause and cancer mortality. And the midlife signal was largely accounted for by animal protein specifically; plant protein did not carry it.
Two cautions. This is observational, so the arrows are associations and nothing more. And 65 is not a biological threshold — it is where the researchers cut the cohort. Nobody's amino acid handling changes on a birthday. What the number marks is a slow crossover in what limits you.
Older bodies are worse at hearing the signal
That crossover has a name: anabolic resistance. The same dose of protein produces a smaller muscle-building response in an older body than a younger one. The signal is being sent; less of it is received. So the intake that was comfortably sufficient at 40 can sit under the threshold at 70 without anything about the meals having changed.
This is what the English Longitudinal Study of Ageing data describes. Among those 532 adults over 65, intakes at or above roughly 0.8–1.0 g per kg of body weight per day — or at or above 18% of total energy — were consistently associated with fewer falls, less mobility limitation, less disability in daily activities, less frailty, slower decline in walking speed, and lower mortality. The PROT-AGE expert group reached a similar place from a different direction, recommending 1.0–1.2 g/kg/day for healthy older adults rather than the 0.8 g/kg RDA — a figure originally set as the floor that avoids deficiency, not a target for thriving.
The honest caveat is the obvious one: frail people eat less. Some of that association is almost certainly decline causing low intake rather than the reverse. It is a real confound and it does not vanish with statistical adjustment.
Training is the variable that settles most of the argument
Here is where the two camps stop arguing, if you let them. Protein is a substrate, and substrate without demand is just surplus. The thing that creates demand is loading your muscles.
Robert Morton's 2018 meta-analysis of 49 studies and 1,863 participants found that protein supplementation added to resistance training increased fat-free mass and strength, with the dose-response curve flattening near 1.62 g/kg/day. Above that, in the pooled data, more protein bought nothing. Note the condition attached to the whole finding: with resistance training. That is the sink. Without it, extra protein is a metabolic signal with nowhere useful to go — which is more or less the population Lamming's review is describing.
It also explains his own carve-out for active people. If you train, the growth signalling that looks costly in a sedentary body is being spent on tissue that behaves like an endocrine organ. We covered the same logic from the exercise side recently: daily movement and structured training are not the same input, and only one of them appears to change how muscle ages.
Illustrative — pick your row
Which question is your body currently asking?
The restriction argument applies most here
0.8–1.2 g per kg per day
No training means no sink for the surplus. This is the group the midlife signal came from, and where the amino acid biology is most likely to matter.
The higher-leverage change is not the number. It is shifting some of it plant-side, and giving your muscles something to do.
Training changes the calculus
1.4–1.8 g per kg per day
The pooled dose-response for resistance training flattens near 1.62 g/kg/day. Above that the data stops paying you.
Exercise appears to be what makes higher intakes metabolically unremarkable — a point the restriction review makes itself.
Anabolic resistance, nothing offsetting it
1.0–1.2 g per kg per day
The older-adult expert consensus sits above the 0.8 g/kg RDA, because the same dose now produces a smaller response.
Protein alone does much less here than protein plus resistance work. If you change one thing, change that one.
The combination the evidence likes best
1.2–1.6 g per kg per day
Higher intake with a real training stimulus is the only pairing where both bodies of evidence point the same way.
Spread it across meals rather than stacking it at dinner; an older body responds better to repeated adequate doses.
What I would actually change
Three things, in descending order of how much they matter.
Load your muscles before you adjust your intake. Almost every disagreement in this literature dissolves once resistance training is in the picture. It converts protein from a signal into a material. If you are choosing between adding a protein shake and adding two sessions a week, the sessions win, and it is not close. Where to start after 40 is its own subject.
Look at the source, not only the total. The midlife mortality association in the NHANES data was carried by animal protein; plant protein did not show it. You do not have to become vegetarian to act on that — moving one or two meals a week toward legumes, lentils and whole grains shifts the amino acid profile without any accounting.
Distribute it. An older body responds better to several adequate doses than one large evening one. Protein contributes to the growth and maintenance of muscle mass and to the maintenance of normal bones — that is what it is authorized to say, and it happens to be exactly the thing at stake as you age.
And one thing to stop worrying about: whether you are eating "too much" protein at 1.4 g/kg while training hard. Nothing in the restriction literature suggests that combination is where the risk lives.
What none of this can tell you yet
The lifespan extension from protein restriction is overwhelmingly animal data. Mice are not small humans, and the amino-acid-restriction results are cleanest in exactly the models that translate least reliably. The human evidence is metabolic markers over weeks and months, not lifespans.
The cohort studies run the other way — real people, real endpoints, but no randomisation. Diet is measured by asking people what they ate, which is a famously lossy instrument, and the frailest participants both eat least and die soonest. Reverse causation is baked in.
Nobody has run the trial that would settle it: decades long, randomised to protein level, in humans, with mortality as the endpoint. Nobody will. So we are left reasoning across two incomplete literatures — which is the normal condition of nutrition science, and the reason confident single-number advice should make you suspicious. Our longevity protocol hub takes the same approach to everything else.
The bottom line
Protein restriction and protein sufficiency are not rival philosophies. They are the correct answers at different points on one curve. In midlife, in a body that is not being loaded, the growth signalling has a plausible cost and the review is worth taking seriously. Later, in a body that no longer hears the signal well, the risk flips to not getting enough. And at every point on the curve, resistance training is what decides which of those two problems you have.
Our practical protein guide still stands. This is the argument it was answering, given its full weight. Muscle you can measure — grip strength and walking speed are cheap, honest signals — and the Agen system exists to help you watch the trend rather than trust a number someone put in a headline.
