Collagen Peptides for Muscle Growth: A Contradiction the Field Has Not Resolved

Stable-isotope work says collagen does not stimulate muscle protein synthesis. A 78-trial network meta-analysis ranks it first for strength. Both are published, and the gap between them is the story.

Leon H · Edited by Caroline S · Published 2026-09-14

Illustration: Weightlifting shoes on a concrete gym floor in early morning light. — Illustration

Two published results on collagen and muscle sit very badly together.

One is a stable-isotope experiment saying collagen does not do the thing muscle growth requires. The other is a network meta-analysis of 78 randomised trials placing collagen at the top of the strength rankings, above whey. Neither is obscure, neither is retracted, and almost every page on this subject cites one and ignores the other.

What the amino acids predict

Collagen is a protein by any chemical definition, and a poor one by every nutritional metric. United States Department of Agriculture composition data for unsweetened dry gelatin powder, the reference form, gives the profile directly:

Per 100 g of gelatin powder
Protein	85.6 g
Glycine	19.049 g
Proline	12.295 g
Lysine	3.46 g
Leucine	2.454 g
Valine	2.081 g
Methionine	0.606 g
Tryptophan	0.0 g

Muscle protein synthesis is triggered largely by leucine and cannot proceed without a full set of essential amino acids. Collagen supplies little of the first and none of tryptophan at all. That is not a marginal deficiency; it is a zero, and it is why collagen scores nothing on the protein digestibility-corrected amino acid score — the arithmetic that United States food-labelling regulation uses to judge whether a protein counts, which we take apart on collagen versus whey protein.

On the mechanism the field uses to rank proteins, collagen should be last.

What the tracer study found

A double-blind trial at McMaster University tested that prediction the expensive way. Twenty-two healthy older women, mean age 69, were assigned to 30 g of whey or 30 g of collagen twice daily for six days. Each performed unilateral resistance exercise, so one leg was fed-only and the other fed-plus-exercised, and muscle protein synthesis was measured by labelled-phenylalanine infusion for the acute response and deuterated water for the longer-term one.

Whey raised acute synthesis in the rested leg by 0.017%/h and in the exercised leg by 0.032%/h, both significant.
Collagen raised it only in the exercised leg, by 0.012%/h.
Synthesis was significantly greater on whey than on collagen in both legs (p = 0.02).

The prediction held. Thirty grams of collagen twice a day — a large intake by any standard — did not move the process that adds muscle protein, except marginally where exercise itself had already moved it.

What the pooled trials say

Then there is the 2026 systematic review and network meta-analysis: 78 randomised trials, 4,755 participants, 13 types of protein supplement plus placebo and control arms, all combined with strength training. Against placebo, for strength:

Supplement	SMD	95% CI	SUCRA
Collagen	0.41	0.09 to 0.73	88.05%
Whey protein	0.15	0.03 to 0.27	—

Collagen came first. That is what the paper reports and it should not be waved away.

The reconciliation, and its limits

The most defensible explanation is not that one result is wrong but that the two are answering different questions, and the giveaway is in the comparator.

A network meta-analysis can only work with the comparisons its source trials actually made. Collagen's body-composition trials have overwhelmingly been run against inert placebos. The most-cited of them randomised 53 sarcopenic men, mean age 72.2, to 15 g a day of collagen peptides or to silica, with 12 weeks of supervised resistance training and fat-free mass measured by DXA — a well-run trial with the right instrument, whose comparator was sand. One of its authors is listed at the Collagen Research Institute GmbH in Kiel, it has attracted a published comment, and an erratum was issued in 2025.

Whey, by contrast, has spent thirty years being tested against casein, soy, pea, beef, different doses of itself, and every timing protocol anyone could devise. Its effect size is measured against a crowded field. Collagen's is measured against silica.

Look at the intervals rather than the ranks and the same thing shows up mechanically. Collagen's estimate spans 0.09 to 0.73; whey's spans 0.03 to 0.27. A range three times wider is the signature of fewer, smaller trials, and SUCRA — the statistic that produces the 88% — rewards a high point estimate without penalising the uncertainty around it.

What is missing, which is the whole answer

There is no published randomised trial that ran for months, put collagen in one arm and an equal dose of whey in the other, trained both, and measured muscle mass by imaging at the end.

Every piece of evidence currently used to argue this question is a substitute for that trial: a six-day tracer study that measured synthesis instead of mass, three-week damage protocols that measured soreness and force, and a pooled ranking assembled from trials that compared collagen with nothing. The two head-to-head comparisons that do exist, both in the recovery literature covered on collagen peptides and muscle recovery, found no advantage for collagen over whey on any outcome.

For where collagen sits on this site's scored evidence ledger against the research compounds, see peptides for muscle growth and the ranked reading on the best peptides for muscle growth.

This is research journalism. Amounts are reported with their sources. Nothing here is a recommendation.

Sources

United States Department of Agriculture, FoodData Central, Gelatins, dry powder, unsweetened, SR Legacy FDC 169599
Oikawa SY, Kamal MJ, Webb EK, et al. Whey protein but not collagen peptides stimulate acute and longer-term muscle protein synthesis with and without resistance exercise in healthy older women. Am J Clin Nutr 2020 — PMID 31919527 (erratum Am J Clin Nutr 2020;112(6):1656)
Zdzieblik D, Oesser S, Baumstark MW, Gollhofer A, König D. Collagen peptide supplementation in combination with resistance training improves body composition and increases muscle strength in elderly sarcopenic men. Br J Nutr 2015 — PMID 26353786 (comment Br J Nutr 2016;116(3):569–70; erratum Br J Nutr 2025;134(5):440)
Drummond MDM, Silva RAD, Junior NC, Melo MS, Ferreira MHL. Which protein-based dietary supplements most effectively enhance fat-free mass and strength gains in healthy adults undergoing resistance training? A systematic review and network meta-analysis. 2026 — PMID 41635649

MuscleLedger facts

United States Department of Agriculture composition data for unsweetened dry gelatin powder records 85.6 g of protein per 100 g, of which 0.0 g is tryptophan and 2.454 g is leucine. Glycine accounts for 19.049 g and proline 12.295 g.
Because it contains no tryptophan, collagen scores zero on the protein digestibility-corrected amino acid score, the measure United States food labelling regulation uses to judge protein quality.
A double-blind trial at McMaster University gave 22 healthy older women 30 g of either whey or collagen protein twice daily for six days, with unilateral resistance exercise, and measured muscle protein synthesis by labelled-phenylalanine infusion and deuterated water.
In that trial whey raised acute muscle protein synthesis in both the fed-only leg (0.017%/h) and the fed-plus-exercise leg (0.032%/h), while collagen raised it only in the exercised leg (0.012%/h). Synthesis was significantly greater on whey than on collagen in both legs (p = 0.02).
The most-cited body-composition trial randomised 53 sarcopenic men with a mean age of 72.2 years to 15 g a day of collagen peptides or to silica as the placebo, alongside 12 weeks of supervised resistance training three times a week, with fat-free mass measured by dual-energy X-ray absorptiometry.
That trial's comparator was silica, not another protein, and one of its authors is listed at the Collagen Research Institute GmbH in Kiel. The paper carries a published comment from 2016 and an erratum issued in 2025.
A 2026 systematic review and network meta-analysis pooled 78 randomised trials and 4,755 participants across 13 protein supplement types and ranked collagen the most effective for strength against placebo (SMD 0.41, 95% CI 0.09 to 0.73, SUCRA 88.05%), ahead of whey protein (SMD 0.15, 95% CI 0.03 to 0.27).
Collagen's confidence interval in that analysis runs from 0.09 to 0.73 while whey's runs from 0.03 to 0.27 — a span roughly three times wider, which is what a ranking built on fewer and smaller trials looks like.

Frequently asked questions

Does collagen build muscle?

The published literature does not currently agree with itself, and no page should pretend otherwise. Stable-isotope tracing at McMaster University found that whey raised muscle protein synthesis and collagen largely did not, with synthesis significantly greater on whey in both the fed and the exercised leg. A 2026 network meta-analysis of 78 randomised trials ranked collagen first for strength gains against placebo. Those two results are hard to hold at once, and the most likely reconciliation is about what each body of evidence compared collagen against.

Why would the amino acid profile predict that it fails?

Because muscle protein synthesis is driven largely by leucine and requires a complete set of essential amino acids. United States Department of Agriculture data for dry gelatin records 85.6 g of protein per 100 g containing 2.454 g of leucine and no tryptophan at all, against 19 g of glycine and 12 g of proline. A protein missing an essential amino acid entirely cannot support synthesis on its own, which is why collagen scores zero on the protein digestibility-corrected amino acid score used by food-labelling regulation.

How can a meta-analysis rank it first, then?

A network meta-analysis inherits the comparisons its source trials made. Collagen's body-composition trials were largely run against inert placebos — the most-cited one used silica — while whey has been tested against every other protein and against itself at different doses for decades. Against an inert comparator a supplement is competing with nothing; against whey it is competing with the best-characterised protein in the field. The wide confidence interval on the collagen estimate, 0.09 to 0.73 against whey's 0.03 to 0.27, is the visible sign of a ranking resting on fewer and smaller trials.

What did the most-cited positive trial actually do?

It randomised 53 men with sarcopenia, mean age 72.2 years, to 15 g a day of collagen peptides or to silica, with 12 weeks of supervised resistance training three times a week, and measured fat-free mass by dual-energy X-ray absorptiometry rather than by impedance — which is the right instrument. Two things qualify it: the comparator was silica rather than protein, and one author is listed at the Collagen Research Institute GmbH in Kiel. The paper has also attracted a published comment and an erratum.

Is there a trial of collagen against whey with a muscle-mass endpoint?

Not a long one. The McMaster trial ran six days and measured synthesis rather than mass, and the head-to-head trials in the recovery literature ran three weeks and measured damage and function. What does not exist in the published record is a months-long randomised trial with collagen in one arm, an equal dose of whey in the other, and muscle mass measured by imaging at the end. Until that exists the question stays open, which is the honest position rather than a hedge.

How much was used in these studies?

Reported with sources, not as guidance: 30 g twice daily for six days in the McMaster synthesis trial, and 15 g a day for 12 weeks in the sarcopenia trial. The network meta-analysis pools a range of intakes across 78 trials. This site reports what trials administered and never tells anyone what to take.

Is this the same question as collagen for recovery?

No, and the two literatures behave differently. Growth is about whether collagen adds contractile tissue over months; recovery is about whether it shortens the soreness and force loss after a damaging session. Our page on collagen and muscle recovery covers the second, where the trials comparing collagen against whey came back null twice. This page is about the first, where the contradiction is live.