Protein timing: the window that isn’t, and the meal that might be

For about twenty years, the most urgent object in a gym was a shaker bottle. Not the barbell — the bottle, and the belief that it had to be emptied before you reached the car park.

The rule had a name that sounded like physiology: the anabolic window. Train, and a door opens. Get protein through it within thirty minutes and the session counts. Miss it and you have, in some vague accounting, wasted the hour you just spent under a bar. It was a good story. It sold an enormous amount of powder. It has also now been tested properly, more than once, and the results are worth knowing — partly because they are so unexciting, and partly because the honest version of the question turns out to be a different question altogether.

Where the thirty-minute rule came from

The underlying observation was real. Resistance exercise makes muscle more responsive to amino acids: the same protein does more when it lands on trained tissue than on rested tissue. Nobody disputes that. The error was arithmetic rather than biological — a sensitivity that lasts many hours got described as if it lasted half of one.

Once a window is defined in minutes, it becomes a product. Timing is easy to sell, because it converts a slow, boring variable — how much protein you eat, over months — into an urgent one you can act on today.

What happened when the window was actually tested

A 2025 systematic review and meta-analysis in Nutrients by Casuso and Goossens pooled five randomised controlled trials — six reports, 193 participants — that compared protein taken before training with the same protein taken after. Crucially, every trial controlled total daily protein, so the only thing that differed was the clock.

Chest-press strength: a standardised mean difference of 0.07, with a confidence interval running from −0.25 to 0.40. Lean body mass: −0.08, from −0.40 to 0.24. Both intervals sit squarely across zero. In plain terms, the timing did nothing you could detect.

One result did move. Leg-press strength came out at 0.70 in favour of eating before training — but its interval starts at 0.005, which clears zero by roughly the width of a hair, and it rests on a small handful of studies. The authors graded the certainty of the whole body of evidence as low to very low. There is a wry footnote here: the single result that budged points the wrong way for the shaker bottle. It favours the protein you ate before you walked in.

Twelve weeks, forty grams, three ways

The meta-analysis compares two schedules against each other. A 2025 trial in the Journal of the International Society of Sports Nutrition did something more useful: it added a group that took no supplement at all.

Klemp and colleagues put 30 untrained men, average age 66, through twelve weeks of supervised resistance training twice a week. One group drank 40 g of protein immediately after each session. One drank 40 g before bed. The third just trained.

All three grew, and they grew by the same amount. Quadriceps thickness rose by 0.16 cm at the vastus lateralis, 0.13 cm at the rectus femoris and 0.18 cm at the vastus intermedius. Leg-press one-rep max went up 28.3 kg, chest press 21.0 kg. Between the groups: nothing that reached significance.

The caveat is the interesting part rather than a disclaimer. These men were already eating at or above roughly 1.0 g/kg of protein a day before the study began. So the trial says something narrow and useful — extra protein, on either schedule, added nothing detectable to men who were already adequately fed — and it cannot say anything at all about someone who is not. That distinction does most of the work in this article.

The question quietly changes

If the hour around training is not where the action is, the natural next question is the one people were too busy timing their shakes to ask: does it matter how protein is spread across the whole day?

Here the evidence stops being tidy, which is a relief, because tidy evidence in nutrition is usually evidence nobody has looked at hard enough.

Two trials that disagree

Yasuda and colleagues, in the Journal of Nutrition in 2020, followed 26 healthy young men through twelve weeks of resistance training three times a week. Both groups ate roughly the same total, around 1.3 to 1.45 g/kg per day. The only difference was breakfast. One group took 0.33 g/kg in the morning; the other took 0.12 g/kg and pushed the balance to dinner, which reached 0.83 g/kg.

Lean tissue mass rose 2.50 kg in the higher-breakfast group and 1.77 kg in the other. The p-value was 0.056 — which is to say a difference big enough to matter, in a study too small to prove it. One arm finished with twelve of a planned seventeen participants. No difference in strength appeared at all.

Then Justesen and colleagues, in Nutrients in 2022, ran the same idea past 24 healthy adults aged 65 to 80. Even distribution — 30% of the day's protein at each of three meals — against a skewed pattern of 15%, 15% and 60% at dinner, on a generous total of 1.5 g/kg of lean body mass, over ten days. Muscle protein fractional synthesis rate came out at 2.16%/day on the even pattern and 2.23%/day on the skewed one. The p-value was 0.647. Nothing.

The variable hiding inside both

Put the two designs side by side and a reading emerges that fits both without either being wrong.

Yasuda's low-breakfast group was not eating a small protein meal in the morning. They were, for practical purposes, not eating one. Around eight grams is below anything that has ever been shown to move muscle. Justesen's skewed group, by contrast, still had a real breakfast — a smaller share of a larger total, but not an absence.

So the trial that found a difference contained a floor, and the trial that found nothing did not. Both results are consistent with a single idea: the thing worth avoiding is not an uneven day. It is a meal so light in protein that it does not register.

Hold that loosely. It is a reading that fits two small datasets, not a hypothesis anyone has tested directly. Different age groups, different outcomes, different timeframes — twelve weeks of measured tissue against ten days of measured synthesis — can disagree for reasons that have nothing to do with how the day was arranged. What the reading has going for it is that it explains the disagreement rather than picking a side in it.

What this changes on a Tuesday morning

The practical rearrangement is small and slightly boring, which is usually a sign that something is real.

Get the total right first. Daily intake remains the variable with the most evidence behind it; timing only arranges what is already there. If you are unsure whether you are near enough, we went through who extra protein actually helps — the answer is narrower than the marketing suggests.

Then audit your smallest protein meal, not your post-workout one. For most people that meal is breakfast, and it is frequently close to zero: toast, fruit, coffee. That is the only lever in this article the evidence supports touching.

Stop timing the shake. Anywhere in the several hours around training is fine, and if it is more convenient beforehand, the one result that moved in the meta-analysis is quietly on your side.

Protein before bed is not wrong. It simply was not better, in men already eating enough, over twelve weeks. If it helps you reach your total, it is a reasonable way to do it.

One thing to notice this week: write down what you ate at breakfast. Not the whole day — just breakfast, for three days. If the honest answer is under about ten grams of protein, you have found something worth changing, and it has nothing to do with a stopwatch. If you want the wider picture of how meal timing interacts with the rest of your day, when you eat covers the circadian side, and protein for longevity covers the long view. For anyone training seriously past midlife, strength training after 40 is the part that actually built the muscle in all four of these trials.

The bottom line

The anabolic window was never thirty minutes wide, and when researchers held daily protein steady and moved the clock, the differences mostly vanished. What survives is duller and more useful: eat enough protein, spread it so that no meal is effectively empty, and put your attention on the training, which is the variable that made every group in every one of these studies stronger.

Use numbers to correct fantasy, not to replace experience. The stopwatch was the fantasy. Breakfast is the experience.