Plant vs animal protein: does quality matter in practice?

Barely, and only for muscle mass rather than strength. Plant protein left muscle mass a little lower than animal protein in the trials that compared them, and the whole gap sat in adults under 60.¹ Soy matched milk protein outright.¹ In the cleanest test, 26 grams of soy did what 19 grams of whey did.⁴

What quality actually measures

Protein quality is a claim about amino acids, not about whether the food once had a pulse. The score researchers now favor is the digestible indispensable amino acid score, or DIAAS. It rates one protein source on which essential amino acids it contains and how much of them crosses the gut wall. Ten years after the expert consultation that proposed it, a review of the field called DIAAS the most accurate score currently available for rating single source proteins.⁷

Chemical scores have a built-in limit, and the people who build them say so first. They describe what is in a food and how much of it gets absorbed. They do not describe what the amino acids then do once they are inside.

The practical consequence is stated plainly in a review of the methods. Diet modeling shows essential amino acid density is higher in omnivorous and lacto-ovo-vegetarian diets, and diets built on whole plant foods may need more total protein and more energy to make up for lower quality.⁶ For incomplete plant proteins, it adds, eating complementary proteins may help.⁶

That is the real claim, and it is a claim about amounts. A low score does not mean the protein fails. It means you need more of it.

The trials, and where they disagree

The largest synthesis found 43 eligible trials comparing plant with animal protein for muscle health, and could combine 30 of them.¹ Plant protein came out behind by a standardized mean difference of -0.20.¹ That unit expresses a result in standard deviations: 0.2 counts as small, 0.5 as moderate, 0.8 as large. So this is a small effect sitting at the bottom of its own scale.

The gap did not carry into anything a person does. Muscle strength across 14 trials showed no difference, and physical performance across five showed none either.¹ The age split is the more interesting result. Adults under 60 accounted for the whole gap, while in those 60 and over it shrank to -0.05, with a range spanning zero.¹

An earlier analysis read a similar pile of studies differently. Across 16 trials, the source of the protein did not affect absolute lean mass or muscle strength at all.² Animal protein did favor lean mass as a percentage, and adults under 50 gained 0.41 kg more lean mass on it.² Resistance training made no difference to the result.

One line in that report deserves as much weight as the effect estimates. Total protein intakes were already above the recommended allowance at the start and at the end.² These are not trials of scarcity, and scarcity is where quality is supposed to matter most.

Soy is the exception that explains the rule

If the source mattered as a category, soy would behave like a plant. It does not. Soy and milk protein were indistinguishable for muscle mass across 17 trials, a difference of -0.02.¹ Against non-soy plant proteins, meaning rice, chia, oat and potato, animal protein came out ahead by -0.58 across five trials, and against plant-based diets by -0.51 across seven.¹ The category label is doing no work here. The amino acid profile is.

A synthesis built specifically around soy found the same. Nine studies in 266 participants set soy against whey, beef, milk or dairy protein alongside resistance training.³ Strength rose in every arm, and no comparison separated the groups on strength or on lean body mass.³

What the lower score costs, in grams

The cleanest single trial puts a number on it. Sixty-one untrained young men and women started 12 weeks of supervised resistance training, three times a week.⁴ Half took 19 grams of whey protein isolate and half took 26 grams of soy protein isolate, the amounts chosen so that both delivered 2 grams of leucine.⁴ Forty-eight finished.

Grams needed for the same 2 g dose of leucinegrams of protein isolate

Whey protein isolate19 g

Soy protein isolate26 g

Lean body mass rose 1.54 kg and total body mass 0.68 kg, with leg strength climbing in both groups and nothing to choose between them.⁴ The 7 grams between 19 and 26 is what the lower score costs. Pay it, and the difference disappears.

Blends, and the size of the dose

The newest trial tested the arrangement people actually use. Forty-four untrained young men took three 15-gram doses a day, 45 grams in all, at breakfast, lunch and dinner, through 12 weeks of supervised training.⁵ One group got a soy and pea blend, the other whey.

Both gained, and by the same amount. Whole-body lean mass rose 2.4 kg on the plant blend against 2.5 kg on whey, and leg-press one-rep max 64 kg against 63 kg, with no difference between the groups on any measure.⁵

That is the shape of the whole literature in one trial. Blend two plant proteins so that each covers the other's weak point, take enough of it across the day, and a difference the trials put at a fifth of a standard deviation stops being visible.^1,5

None of this says animal protein is interchangeable with any plant protein at any amount. It says the ranking the chemical scores produce is small, sensitive to how much you eat, and mostly closed by formulation.

What remains unknown

The condition all this is meant to bear on has not been measured. Not one trial in the largest synthesis looked at sarcopenia itself as an outcome, and its authors ask for work on a wider range of plant proteins.¹

The comparisons against non-soy plants rest on five trials, so rice, oat, chia and potato protein are ranked below animal protein on thin evidence.¹

The trials that exist mostly enrolled young adults already eating more protein than recommended. The group where quality should matter most, older adults eating less of it, is exactly where the difference was smallest and least certain.^1,2

And nobody has tested whether matching leucine still works over years rather than 12 weeks, or settled how processing and storage change the scores themselves.^4,7