The Problem With Most Collagen Studies - And Why FoYL Was Tested Differently

Walk into any collagen conversation and someone will say "but studies show..." Studies show what, exactly? Once you look at how most collagen research is actually constructed, the phrase starts to mean a lot less than it sounds - and you understand why we had ours analysed differently.

This one is for the readers who do not take a claim at face value: the careful shoppers, and the naturopaths, nutritionists and practitioners who get asked "which collagen should I take?" and want a defensible answer. The short version is that the weakness in most collagen evidence is not the conclusion - it is the input.

The quiet problem with "collagen studies"

A great deal of collagen research is run on generic hydrolysed collagen of mixed, uncharacterised molecular weight. The study will tell you it used "collagen peptides," but not the peptide size distribution, the proportion of small bioactive fragments, or how that specific material behaves. That is a problem, because - as we have argued across this blog - molecular size is the variable that decides how a peptide behaves.

So when a brand waves at "the studies," it is often borrowing the credibility of research conducted on a material that is not the same as the powder in their bag. The conclusion might be real; the relevance to that specific product is unestablished. It is a category-wide sleight of hand, and most of the time nobody is doing it maliciously - it is just how the field grew up.

What "characterised" actually means

To characterise a material is to define precisely what it is before you test what it does: the molecular weight range, the distribution of peptide sizes, the proportion that is in the small, functional form. "Collagen peptides" with no characterisation is like a recipe that just says "flour" - it could be anything from cake flour to wholemeal, and the result will differ accordingly.

FoYL's material was characterised first - confirmed as a 300 to 500 Dalton tripeptide, more than 80% in true tripeptide form - and only then evaluated for its predicted functional behaviour. That sequence (define the molecule, then model what it does) is precisely the step most collagen evidence skips.

Why molecular-weight distribution matters

Two products can share an average molecular weight and still be completely different materials. One could be tightly clustered around a small size; another could be an average of very large and very small fragments that happens to land on the same number. An average hides the distribution. This is why "low molecular weight collagen" as a loose marketing phrase is not the same as a characterised tripeptide with a stated range and a stated proportion in true tripeptide form. The detail is the evidence.

What we did instead

Rather than gesture at someone else's generic peptides, we had our actual material independently analysed by the University of Milan, Department of Pharmaceutical Sciences (2025) - and the methodology is the point:

• In-silico - computational peptide mapping of the actual tripeptide profile, modelling structure and predicted activity.
• In-vitro - laboratory enzyme assays on the real material.
• In-situ - testing against human intestinal cell models (Caco-2) and related human cell lines.

Three complementary angles on the same characterised material, looking for agreement between them - rather than one convenient measurement on an unknown input.

Why this is the more honest position

It would be easier, and more impressive-sounding, to cite a drawer full of human trials. We would rather tell you precisely what was done: independent, third-party, pharmaceutical-sciences analysis of our own characterised material, with the figures framed as modelling rather than medical outcomes. Under-claiming with specificity beats over-claiming with borrowed studies.

For a practitioner, that distinction is the whole game. "Studies show collagen helps" is not a recommendation you can stand behind. "This specific material is a characterised 300 to 500 Dalton tripeptide, independently analysed, with its bioactivity modelled" is information you can actually weigh.

Questions a careful reader will ask

Is in-silico and in-vitro evidence "less than" a clinical trial? It is a different kind of evidence, and we label it as exactly that. It speaks to the characterised material and its predicted behaviour, not to clinical outcomes. We do not inflate it into something it is not.

Why not just run a human trial? Robust human trials are valuable and also slow and expensive. Independent characterisation and laboratory modelling of the actual material is a rigorous, honest step that most of the category skips entirely - and it is the evidence that tells you what is really in the bag.

Can I see the methodology? Yes. We have put it on our science page in plain language, so you can judge the input, not just the headline.

P.S. Next time you hear "studies show," ask the only question that matters: studies on which material, characterised how? The answer separates evidence from atmosphere.