A Vegan Diet Cost Older Adults 2.5 lb of Muscle in 12 Weeks. The Protein They Stopped Eating Is Most of the Story.

A trial landed this month that is going to be misread in both directions, so here is what it actually did.

Seventy-two healthy adults aged 65 and over were randomised into three groups for twelve weeks [1]. One group went vegan. One carried on eating as they always had. One went vegan and lifted weights twice a week.

Then researchers put them in an MRI scanner and measured the actual volume of their thigh muscle — before and after. Not a questionnaire. Not a bathroom scale. A picture of the muscle.

The vegan group lost 2.5 lb (1.15 kg) of lean mass from their arms and legs. The omnivore group gained a tenth of a pound, which is another way of saying they did not change [1].

That is a real result, measured properly, and it should not be waved away.

Now the sentence that decides what it means

It was a self-selected vegan diet.

Nobody handed these people a meal plan. Nobody coached them on protein. They were asked to go vegan and they went vegan the way a person actually goes vegan — and their protein intake fell by between 19 and 28 grams a day, every measurement point, for twelve weeks [1].

So what is this trial a test of? Not veganism as a nutritional philosophy. It is a test of switching to a vegan diet without thinking about protein, which is the most common way it is done and therefore the most useful thing to measure. The researchers chose that on purpose.

But it means the headline writes itself wrongly. “Vegan diets shrink your muscles” is not what happened here. What happened is that people ate substantially less protein and lost muscle, which is what the protein literature has predicted for thirty years.

How do we know it is the protein and not the plants?

Because somebody separated those two things, and the answer is fairly clear.

A meta-analysis pooled sixteen randomised trials comparing animal protein against plant protein for lean mass and strength [4]. Across those trials, protein source did not affect absolute lean mass or muscle strength at all.

And the detail that makes it fit: in those trials, total protein intake was at or above the recommended daily allowance from start to finish [4]. Nobody was going short. Change the protein source while keeping the amount, and the muscle does not notice much.

There is a wrinkle, and this page is not going to hide it because it points the awkward way.

The same research group fed older adults a vegan meal matched to an omnivorous meal for both calories and protein — identical numbers on paper — and still measured a lower rate of muscle building afterwards [3]. Plant proteins tend to be lower in leucine, the amino acid that acts as the trigger, and are often digested a little less completely.

So: total protein is most of the story. It may not be all of it. Somebody eating plant protein may need somewhat more of it, not merely the same amount — and that is a live question rather than a settled one.

The third group is the practical one

The group that went vegan and lifted twice a week lost about half as much: 0.35 L of thigh volume against 0.61, and 1.4 lb of lean mass against 2.5 [1]. Their muscle-building rate came back up to almost exactly the omnivores’ figure.

Read that carefully though, because it is a mitigation and not a cure. They still lost muscle. Lifting halved the damage of the protein drop; it did not cancel it.

Which is a familiar shape on this site. It is the same answer we reached this week about intermittent fasting and lean mass: the dietary pattern is not the villain, the protein and the training are the variables, and the people who lift keep more of what they have.

What a reader over 65 should take from this

The expert recommendation for healthy older adults is at least 1.0 to 1.2 grams of protein per kilogram of bodyweight per day [5] — for someone weighing 154 lb (70 kg), that is 70 to 84 grams. Older adults need more than younger ones, not less, because ageing muscle responds less to the same dose.

If you eat plants and you are over 65, the finding here is not that you have made a mistake. It is that the protein does not look after itself, and on a plant diet it is easier to drift down without noticing. Count it for a week. Most people are surprised in one direction or the other.

And lift something twice a week, which the evidence in older adults supports about as strongly as anything on this site.

A conflict worth stating plainly

The senior author of this trial declares research grants, consulting fees and speaking honoraria from companies that make dairy and meat products [1]. On a paper reporting that vegan diets cost older people muscle, you should know that.

Two things are also true. The same declaration lists funding from companies producing plant-derived proteins, so the interest does not point cleanly in one direction. And the finding rests on MRI volumes and heavy-water tracers, with the protocol published in advance [2] — the parts of a study that are hardest for an interested party to nudge.

We are telling you it is there. We are not telling you it invalidates the measurement, because we have no evidence that it did.

What this is rated, and what the rating covers

Preliminary. One randomised trial, 72 people, twelve weeks.

It is rated Preliminary rather than Supported because a single trial is a single trial, however well it was measured — and this one was measured unusually well [1]. Nothing has replicated it yet. If it replicates, this page gets upgraded and the change will be recorded on it.

What is not rated here: whether a vegan diet is good or bad for you generally, which is a much larger question this page does not touch; anything about people under 65, who were not studied; and whether plant protein is inferior at matched intake, where the meta-analysis says no for lean mass [4] and the meal study says maybe for the acute response [3].

Related: how much protein actually builds muscle, whey against plant protein, and the opposite end of the same argument, the carnivore diet. How we read a study, and what each tier means, is set out here.