"The anabolic response to protein ingestion during recovery from exercise has no upper limit in magnitude and duration in vivo in humans."
That is the actual title of Trommelen et al., published in Cell Reports Medicine in December 2023, corresponding author Luc van Loon, whose citation count runs to five figures. It is not a headline written by a supplement company. It is a claim the authors stand behind, built on a quadruple isotope tracer approach that is about as rigorous as human feeding studies get.
And the data underneath it are real. Per Trommelen et al., feed a person 100 grams of protein after exercise instead of 25, and the anabolic response is reportedly both greater and longer, stretching past twelve hours. Whole-body net protein balance rises. Myofibrillar, connective, and plasma protein synthesis all climb.
The paper's own summary line is blunt: the belief that excess amino acids are simply oxidized reportedly "lacks scientific proof." Protein ingestion, they report, had a negligible impact on amino acid oxidation rates. Nothing, in their window, went to waste.
You can feel why this persuades. If the big bolus keeps synthesis elevated for half a day, and the surplus isn't being burned off, then the old "20 grams per meal and the rest is wasted" rule looks like a cartoon. The gut becomes a time-release capsule. Eat big, absorb slow, build all day.
Last week in this space we left an open thread: will a multi-week lean-mass trial confirm that spread ~25 gram feedings match a single 100 gram bolus? This week we can sharpen it, because two things have arrived — twenty-four-hour tracer data, an eight-week metabolic ward trial — and because the researchers themselves have now drawn the line in public.
Start with the line Trommelen drew. Confronted directly, he reportedly did not defend the muscle-mass version of his claim, framing it instead as a mechanistic point.
"Our study challenges a MECHANISTIC argument commonly used in favor of protein distribution: consuming more than 20 g of protein in a single meal is wasted," he wrote. "A full discussion about protein distribution or muscle mass is a much bigger topic."
That is the whole game, stated by the author of the "no upper limit" paper. His result is a mechanism, not an outcome. An elevated synthesis curve over twelve hours tells you the machinery kept running. It does not tell you what the machinery built over a month.
His result is a mechanism, not an outcome. An elevated synthesis curve over twelve hours tells you the machinery kept running. It does not tell you what it built over a month.
Stuart Phillips put the boundary the same way, if less gently: "I enjoy a good MPS study as much as you do, but there are tight limits on what it can show. It is, at best, insight into a possible mechanism. But as you know, it's short. Even your 'long' study is a tiny fraction of what protein could do."
So the question is not whether 100 grams outsynthesizes 25 grams over twelve hours. It does. The question is whether that acute advantage survives into lean mass and function over months. On that, the acute studies are silent by construction.
Here is the tension the big-bolus camp has to answer. The mechanistic literature is not uniformly on its side once you extend the clock.
Symons fed subjects either 30 grams of protein as lean beef or 90 grams. Same muscle protein synthesis over the measurement period. Three times the steak, not a hair more synthesis — and if digestion rate were the true limiter, the big meal is exactly where you'd expect it to pull ahead.
Trommelen's rebuttal is fair and worth stating: the Symons window was only five hours. What would have happened over a longer period? The main lesson of his own paper, he argues, is that the assessment window must cover the full response, or it isn't a fair comparison.
Fine. So look longer. Mallinson measured twenty-four hours with deuterated water after two lifts and two feeds. Per Phillips, sixty grams didn't beat thirty. Numerically, it lost.
That is the crack in the "saved in the gut" story. Nitrogen not oxidized inside a twelve-hour window doesn't stay banked — it gets oxidized in the next one. The longer window doesn't rescue the bolus; it exposes it.
Because there is no amino acid storage pool. Every species from fish to mammals evolved a way to dump excess nitrogen — ammonia, uric acid, urea. The body will grant a temporary breathing space. It does not keep a long-term reservoir of spare protein waiting to be assembled into muscle.
Which brings us to a piece of evidence that does what no acute tracer study can: it feeds people more protein for eight weeks and then weighs what actually changed.
Bray 2012. A metabolic ward — subjects living inside the study, intake controlled, not self-reported. About 950 kilocalories per day over maintenance, sustained for eight weeks, at 5, 15, or 25 percent of energy from protein.
This is, per Phillips, more or less that thought experiment: the NG tube running blended complete food all day, no stomach limit, nothing ever "wasted." If gut capacity is the ceiling, the high-protein overfed group should be building muscle around the clock.
The result. Per Phillips, lean mass rose 2.9 kilograms at 15 percent protein and 3.2 kilograms at 25 percent. Two-thirds more protein every day for two months reportedly bought 0.3 kilograms of extra lean mass.
And lean mass isn't muscle, as Phillips notes. So even that 0.3 kg may overstate the muscle-specific payoff.
Where did the surplus protein go? Into urea and heat. Per Phillips, resting energy expenditure rose 227 kilocalories per day versus 160 — on his reading, the body spending the extra nitrogen, not stockpiling it.
And the calories that came along for the ride became fat. Fat gain was identical across all three groups. The extra protein didn't partition the surplus toward muscle and away from adipose; it was oxidized and deferred, exactly as the urea cycle predicts and exactly opposite to what time-release logic requires.
That is the lean-mass answer. Not the twelve-hour synthesis curve — the eight-week body composition.
So the two datasets are not in conflict; they answer different questions. Trommelen measured whether synthesis stays elevated after a large bolus. Yes. Bray measured whether feeding more protein for weeks builds more lean tissue. Per Phillips's framing, barely.
The practical translation is unglamorous. Past roughly 1.6 g/kg/day, more protein buys nothing for lean mass and function over months, however you split it. Somewhere between 25 and 40 grams per meal gets you there.
"How much fits in your stomach," as Phillips puts it, isn't a recommendation — it's a shrug. It answers a mechanistic question nobody trains for.
And it leaves the acute science intact. Trommelen's tracer work is honestly measured and, on its own terms, correct.
But the resolution to the protein-ceiling argument was never going to come from another acute MPS study, no matter how long the window. What remains open is the multi-week lean-mass-and-function RCT this debate keeps deferring — big bolus versus spread feeding, matched protein and calories, measured in kilograms of muscle and units of strength, not in synthesis rates.
Until that trial runs, Bray is the best long-horizon evidence we have. And Bray says the surplus left as urea, heat, and fat.
The null hasn't moved.