Protein Timing: Does When You Eat It Matter?

The idea is biologically reasonable. Resistance exercise sensitises muscle to amino acids; feed protein into that sensitised state and myofibrillar protein synthesis rises further than either stimulus alone produces. By the mid-2000s this had hardened into a rule: protein had to be consumed within roughly an hour of training, and some writers went further, proposing that when nutrients were eaten might matter more than how much.^[1]

In 2013 Alan Aragon and Brad Schoenfeld went back through that literature and found the rule had outrun its evidence.^[1] The acute studies that established it had almost all been done in fasted subjects — people who trained first thing in the morning on an empty stomach, where a post-workout feed is the first protein of the day and the contrast with 'nothing' is enormous. In anyone who had eaten a normal meal in the hours before training, amino acids were still appearing in the circulation during and after the session. The authors' conclusion was that the window, if it exists, is probably four to six hours wide on either side of training, scaled to the size and composition of the pre-workout meal.

That is a different claim from the one the supplement aisle makes, and it is the claim the long-term trials went on to test.

The decisive analysis came the same year. Schoenfeld, Aragon and Krieger ran a multi-level meta-regression of randomised controlled trials in which at least one group took protein within an hour of training and a control group did not take protein within two hours.^[2] Trials had to run at least six weeks and measure strength or size directly. Twenty-three studies survived screening, with an average PEDro methodological quality score of 8.7 — a high-quality sample. The strength analysis covered 478 participants and 96 effect sizes; the hypertrophy analysis covered 525 participants and 132 effect sizes.

Pooled crudely, timing looked real. The hypertrophy effect size favoured the timed groups by 0.24 (95% CI 0.04 to 0.44, p = 0.02) — small, but statistically significant. Strength showed nothing (0.38, 95% CI −0.34 to 1.10, p = 0.30).

Then the authors added the covariates. In the full model, controlling for protein matching, training status, blinding, sex, age, body mass and study duration, the hypertrophy advantage fell to 0.16 (95% CI −0.07 to 0.38, p = 0.18) and the strength advantage to 0.28 (95% CI −0.52 to 1.07, p = 0.49). Neither was significant. And when the analysis was reduced to its best-fitting form, one covariate dominated everything else: total daily protein intake (estimate 0.41 per g/kg/day, 95% CI 0.14 to 0.69, p = 0.004).

The reason is visible in the raw diets. In the trials that did not match protein between groups, the control groups averaged 1.33 g/kg/day and the supplemented groups 1.66 g/kg/day. The 'timing' group was simply eating a third of a gram per kilogram more protein every day for three months. Only three included trials matched total intake properly — and two of those three found no timing benefit at all.^[2]

A separate randomised trial then tested the window directly. Twenty-one resistance-trained men took the same 25 g protein dose either immediately before or immediately after training, three days a week for ten weeks. Strength, hypertrophy and body composition changed identically; no measured outcome differed between the groups.^[3] (A correction to that paper was subsequently published; its conclusions were unchanged.)

The case for spreading protein evenly rests on a per-meal ceiling. That ceiling has a specific origin. In 2014 Oliver Witard's group gave 48 resistance-trained men a standardised high-protein breakfast, had them perform unilateral leg work three hours later, then fed them 0, 10, 20 or 40 g of whey isolate.^[4] Myofibrillar protein synthesis rose 49% above the no-protein condition on 20 g and 56% on 40 g — a difference that did not reach significance — while 10 g did nothing measurable. The 40 g dose did raise phenylalanine oxidation and urea production. The authors' conclusion was careful and reasonable: 20 g is sufficient to maximally stimulate synthesis in roughly 80 kg trained men. It was repeated thereafter as a hard limit, and 'your body can only use 20 to 25 g of protein at a time' entered general circulation.

Two things have since complicated it. First, that experiment trained one leg. When the same group repeated the design with a whole-body session — far more muscle sensitised, far more tissue competing for amino acids — 40 g beat 20 g outright (0.059 vs 0.049 %/h, p = 0.005), and lean body mass did not predict the response.^[5] The ceiling moved when the stimulus got bigger.

Second, and more fundamentally, the four-hour measurement window was hiding most of the answer. In 2023 a Maastricht group fed 36 men 0, 25 or 100 g of intrinsically labelled milk protein after resistance exercise and followed them with a quadruple isotope tracer for twelve hours.^[6] Myofibrillar synthesis on 100 g was about 20% higher than on 25 g in the first four hours and about 40% higher from four to twelve hours. Amino acid oxidation rose only negligibly. Meal-derived amino acids supplied 9% of the amino acids incorporated into muscle after 25 g and 27% after 100 g. The authors' stated conclusion is that the anabolic response has 'no apparent upper limit in magnitude or duration.'

This is not a licence to eat all your protein at dinner — a bigger response to 100 g than to 25 g is not the same as four times the response, and the per-gram return clearly falls. But the specific claim that surplus protein above 20–25 g is simply burned off does not survive a twelve-hour measurement. Notably, the researchers who built the original dose-response work have pushed back, pointing out that the finding was obtained in recreationally active rather than resistance-trained men, and that emerging data suggest it may not extend to resistance-trained young women.^[7] The question is live, not closed.

Three lines of evidence, and they do not all point the same way.

The acute work is encouraging. In a crossover study, trained men took 80 g of whey over twelve hours of recovery in one of three patterns: eight 10 g doses every 90 minutes, four 20 g doses every three hours, or two 40 g doses every six hours. Every pattern raised protein synthesis above rest, but the four-by-twenty pattern beat both the others by 31 to 48%.^[8] A separate seven-day crossover in eight healthy adults fed identical total protein either evenly (about 30 g at each of three meals) or skewed toward dinner (11 g, 16 g, 63 g). Twenty-four-hour muscle protein synthesis was 25% higher on the even pattern — 0.075 versus 0.056 %/h on day one, and 0.077 versus 0.056 %/h after a week of habituation.^[9]

So: eat it evenly, build more muscle? The trouble is that a higher fractional synthesis rate measured over 24 hours is a surrogate, not an outcome, and the gap between the two is where most nutrition myths live. Only one randomised trial has followed the question out to actual tissue. Twenty-six young Japanese men trained three times a week for twelve weeks on an identical 1.30 g/kg/day protein intake, differing only in whether breakfast carried 0.33 or 0.12 g/kg.^[10] Both groups gained lean soft tissue. The protein-at-breakfast group gained more — 2.5 ± 0.3 kg versus 1.8 ± 0.3 kg — but the comparison landed at p = 0.06, just the wrong side of conventional significance, with a large effect size (Cohen's d = 0.795) in a very small sample.

That is the honest state of the distribution question: a consistent acute signal, one near-miss chronic trial in 26 people, and no confirmation. The practical recommendation that follows — roughly 0.4 g/kg at each of at least four meals to reach 1.6 g/kg/day, or 0.55 g/kg/meal to reach 2.2^[11] — is sensible, costs nothing and is probably worth doing. It is not established fact, and anyone presenting it as such is ahead of the data. If you want to know how your protein is actually distributed across your day rather than how you imagine it is, logging a few ordinary days in a tracker such as Peak Nutrition answers the question in about a week.

Overnight is the longest fasted stretch most people have, and for a decade the suggestion has been to bridge it with slow-digesting casein. The mechanism checks out. Sixteen young men trained in the evening, ate a standard recovery meal, then took 40 g of labelled casein 30 minutes before sleep. The protein was digested and absorbed normally during sleep, amino acid concentrations stayed elevated all night, whole-body protein synthesis rose (311 ± 8 versus 246 ± 9 µmol/kg over 7.5 h) and net protein balance went from negative to positive (+61 ± 5 versus −11 ± 6 µmol/kg). Mixed muscle protein synthesis was about 22% higher, at borderline significance (p = 0.05).^[12]

The long-term trial that followed is the one usually quoted. Forty-four young men trained for twelve weeks; one group took 27.5 g of protein before bed every night and the other took a placebo. The protein group gained more strength and more quadriceps cross-sectional area (+8.4 ± 1.1 versus +4.8 ± 0.8 cm²), with greater type II fibre growth.^[13] It is a clean result for protein. It is not a result about timing — the placebo was non-caloric, so the pre-sleep group simply ate 27.5 g more protein per day for three months, which is exactly the confound the 2013 meta-analysis identified.

The trial that isolates timing is smaller and less famous. Thirteen men trained for ten weeks on isocaloric diets providing 1.8 g/kg/day of protein, with 35 g of casein taken either earlier in the day or shortly before bed. Lean soft tissue, muscle cross-sectional area, leg press and bench press all improved — with no difference between the groups on any measure.^[14]

Pre-sleep casein is a good way to add protein to a day that is short of it, and a pleasant one. Taken at a fixed total intake, it has not been shown to beat eating the same protein at lunch.

The 2017 International Society of Sports Nutrition position stand on nutrient timing states the practical version plainly: meeting the total daily protein target, preferably in evenly spaced feedings roughly every three hours, should be the primary area of emphasis, with 20–40 g of a high-quality source per dose.^[15] That is not a timing recommendation dressed up — it is a total-intake recommendation with a reasonable default schedule attached.

Two things make the schedule worth following anyway, neither of which is the anabolic window. The first is arithmetic: if you are aiming for 1.6 to 2.2 g/kg/day, as the dose evidence supports,^[16] you are very unlikely to reach it in two meals without unpleasantly large ones. Four servings of 0.4 g/kg is simply the easiest route to the number that actually matters. The second is age: the per-meal dose needed to maximally stimulate synthesis roughly doubles with age, from about 0.24 g/kg of body mass in young men to about 0.40 g/kg in men around 71, even though resting synthesis rates are identical.^[17] For an older adult the breakfast question is not an optimisation — a 10 g breakfast may genuinely fail to cross the threshold at all.

For everyone else, the order of priorities is settled. Hit your daily total. Spread it across three or four meals because that makes the total achievable. Put protein in breakfast, because that is the meal where most people have almost none. Then stop thinking about it and go and train, which is the variable that actually moves the outcome.