Thirteen grams of new muscle protein from 100 grams of milk protein. Four and a half grams from 25 grams. That is the number the viral posts led with, and on its face it looks like a rout: nearly three times the muscle built from four times the protein.
The paper's own title did the rest of the work. "The anabolic response to protein ingestion during recovery from exercise has no upper limit in magnitude and duration in vivo in humans."
FitChef ran with it. "The ceiling vanished." The 30-gram rule, they wrote, "was never a body limit. It was a measurement limit."
It is a seductive frame, and it is not entirely wrong. But it buries the number that actually matters for how you should eat.
Here it is: the 100-gram dose stimulated muscle protein synthesis only 20 to 30 percent more than the 25-gram dose. A four-fold increase in protein bought a fifth to a third more MPS.
Stuart Phillips, who chairs kinesiology at McMaster and has spent a career mapping protein dose-response, read that as the opposite of a vanished ceiling. He called it "a definitive plateau in how much protein can be used."
His arithmetic is blunt: "the small increment in MPS from 25 to 100g that is 20-30% when protein dose increased 4-fold suggests the per meal dose is actually MUCH closer to 25g than 100g."
That is the reframe. Not that the body caps out at 25 grams and torches the rest — the tracer data clearly show it does not — but that the marginal return collapses almost immediately past it. You pay 4x and collect roughly 1.25x.
Not that the body caps out at 25 grams and torches the rest, but that the marginal return collapses almost immediately past it. You pay 4x and collect roughly 1.25x.
So where did the headline "no upper limit" come from, if the magnitude of the response barely moved? It came from the clock.
The old studies — Moore 2009, Witard 2014, Churchward-Venne 2020 — measured for three to five hours and found the response plateaued around 20 to 25 grams. Trommelen's Maastricht team, 36 recreationally active young men, milk protein, a quadruple isotope tracer, watched for twelve.
And over twelve hours the 100-gram dose kept working. At four hours, 26 percent of it had reached the bloodstream; by eight hours, 44 percent; by twelve, 53 percent, with the absorption curve still climbing when the researchers stopped measuring.
The 25-gram dose, by contrast, was "done at 5 hours" — amino acids surged, peaked, and returned to baseline on the schedule the old literature predicted.
That is the real finding. The big bolus does not build proportionally more muscle in the moment; it builds a modestly larger amount slowly, stretching the anabolic response across a long tail the shorter studies never stayed awake to see.
Which is genuinely interesting physiology. It is also, for practical eaters, close to useless.
Phillips said as much: "The extended duration of response is interesting but largely moot as most folks eat more than once daily! Plus it's just milk protein and it's MPS."
Sit with why that "moot" is doing the heavy lifting. The 100-gram dose's advantage lives almost entirely in hours five through twelve — the window in which the 25-gram dose has finished and gone quiet.
FitChef frames that as the large dose's edge growing over time: roughly 20 percent higher MPS in the first four hours, roughly 40 percent higher from hour four to twelve.
But that widening gap only exists because nothing else is being eaten. It is the advantage of a single bolus stretched across an artificial fast.
A person who eats again in three or four hours never enters that window empty. The next meal refills the amino acid pool precisely when the 25-gram dose tapers — capturing the same extended anabolism the megadose achieves by dribbling out one enormous meal over half a day.
So the study does not overturn the per-meal target. It confirms it, then quietly explains why spreading intake works: multiple ~25-gram feedings reconstruct the long tail without needing to swallow 100 grams at once.
The dietary implication is unglamorous. Distribute protein across meals near the 25-gram mark rather than backloading a single megadose and calling the extended absorption a win. On typical human eating patterns, the plateau is the operative fact and the tail is a laboratory artifact.
Now the caveats that keep this honest, because they cut against overclaiming in both directions.
This was a single bolus of milk protein — fast-digesting, leucine-rich, not a mixed meal of steak and rice and fat that empties the stomach at a different rate. It was young men, aged 18 to 40, after a full-body resistance session, which is the population and context most primed to respond.
And the endpoint is muscle protein synthesis, an acute rate, not weeks of measured gains in lean mass. Higher MPS over twelve hours is a reasonable proxy for muscle building; it is not the same as bigger muscles at the end of a training block. Phillips flagged both limits himself: "it's just milk protein and it's MPS."
It is also worth noting the authors and at least one commenter read the same data more expansively — arguing the increase on top of an already exercise- and protein-stimulated rate is substantial, and that the right yardstick for a plateau is the dose-response regression, not the raw percentage over 25 grams. The tracer numbers are not in dispute. What they license is.
Here is the part where the "the body can use more protein than you think" logic actually earns its keep — not in the gym, but in the clinic.
Peter Attia makes a parallel argument about chronic kidney disease, where, per Attia, protein restriction to roughly 0.6 to 0.8 g/kg/day has reportedly been standard advice for stage 3 to 5 CKD for decades.
The mechanism is real. Processing a protein load makes the kidney raise its filtration rate — hyperfiltration — and in kidneys that have already lost nephrons, the survivors are already running under higher pressure, so the added strain could plausibly accelerate decline.
But Attia's point is that a plausible mechanism is not proof a recommendation improves outcomes. Protein restriction is, in his framing, "a legacy recommendation" — it took hold before the current era of kidney-protective drugs, when clinicians had little else to offer a patient whose function was slipping.
The 2024 KDIGO guideline has already softened, suggesting around 0.8 g/kg/day and cautioning against high intake above 1.3 g/kg/day only in patients at risk of progression — not the aggressive 0.3 to 0.4 g/kg/day very-low-protein diets some older recommendations reached for.
His challenge is a burden-of-proof shift: anyone still recommending routine restriction should show it adds benefit on top of modern therapy, not merely that the physiology makes sense.
Notably, Attia grants the flip side the Trommelen tail cannot: in healthy kidneys, the filtration bump appears to be normal adaptation, with no clear evidence that high-protein meals damage them. The concern is specific to already-compromised nephrons — a narrow use case, not a general rule.
That is the through-line. Both stories are corrections to over-extrapolation — one from a percentage that got dressed up as "no limit," the other from a mechanism that got dressed up as a mandate.
The tracer data are honest. The 100-gram dose really does build about 13 grams of muscle protein to the 25-gram dose's 4.5, and it really does keep going past hour twelve.
But if you eat more than once a day — and you do — the usable per-meal number still sits far closer to 25 grams than to the 100 the label wants you to buy.