Do hypermobile people have less collagen?

• About
• Latest Posts

Adam Foster

Director at The Fibro Guy

Adam Foster is the founder of The Fibro Guy and one of the leading voices in hypermobility, EDS, and chronic pain rehabilitation. A former Reconnaissance Soldier, he developed chronic pain after an IED blast in Afghanistan and built his approach from the ground up when the medical system had no answers. Since 2009 he has worked with thousands of people across eight countries, drawing on pain neuroscience, motor learning, and sensorimotor rehabilitation. He co-authored "No Pain, No Pain" with trauma psychologist Dr Mairi Harper, directed the documentary "Invisible Monster," and has been featured in Forbes, BBC Radio 4, ITV News, LiveScience, and The Telegraph.

Latest posts by Adam Foster (see all)

• Is Fibromyalgia a Disability in the UK, and Can You Claim PIP? - 14 October 2026
• Fibromyalgia and Central Sensitisation, What Is It in Plain English? - 8 October 2026
• Which supplements are worth taking when you’re hypermobile - 4 October 2026

The short answer is no. Across the Ehlers-Danlos syndromes, the total amount of collagen in the body is usually about where you’d expect. What differs is the collagen itself ^[1][2].

Collagen has to be folded inside the cell, trimmed on the way out, and bundled into fibres before it holds anything together. Any of those steps can go wrong, and which collagen is affected changes which one does. Some faults leave the chain folding badly. Some leave the fibres put together unevenly. Some leave the whole structure being broken down faster than it gets laid back down ^[1][2].

That matters, because the two call for different answers. You can top up a shortage. An assembly fault doesn’t work that way.

Collagen Is a Whole Family

Collagen is a family of proteins, and a pretty big one at that. There are something like 28 to 29 recognised types, and different tissues use different mixtures of them, which is why a fault in one collagen can look nothing at all like a fault in another ^[3]. Each molecule starts as three protein chains wound into a triple helix, which then gets modified, secreted out of the cell, and assembled into fibrils that are cross linked into networks ^[4][5]. That’s the bit people picture when they picture collagen: the rope.

Not every collagen is there for bulk though. Type I is the major structural one, and it’s what most of your skin, tendon, bone and a lot of your blood vessel wall is made of ^[6]. Type III matters especially in blood vessels, hollow organs and the stretchier tissues ^[5]. And then there’s type V, which is present in very small amounts and does a completely different job, as it seeds and regulates the assembly of type I fibrils, setting their diameter and their organisation while contributing almost none of the mass ^[7]. The minor collagens are generally like that, they’re the foreman rather than the bricks ^[8].

Which is why counting collagen doesn’t really get you very far. A tissue can have a perfectly ordinary amount of protein sitting in it and still be mechanically wrong, because the type that organises the rest is missing, or because the right type was finished badly on the way out.

What Actually Goes Wrong in Ehlers-Danlos Syndrome

There are thirteen recognised subtypes in the current classification, and most of them come down to a gene for a fibrillar collagen, for one of the enzymes that processes collagen, or for a matrix molecule that controls how collagen gets assembled and maintained ^[9][10]. Sort them by what’s actually broken and the same handful of patterns turns up over and over.

Classical EDS: This is the closest thing to a genuine shortage in the whole family, and even here it isn’t a shortage of collagen in general. Most people with it carry a fault in a type V collagen gene, usually COL5A1, and the commonest effect is simply less usable type V being made, though some faults produce a structurally duff version instead ^[11][12]. Take away the collagen that organises the fibril and the fibril comes out wrong, which has been shown in animals, where deleting type V collagen from tendon and ligament produced the joint laxity you’d expect from the human condition ^[13]. The missing material is specifically type V rather than collagen as a whole, and the tissue problem is failed assembly with type I ^[7].

Vascular EDS: Faults in the type III collagen gene, COL3A1, mostly produce collagen that’s made wrongly rather than not made at all, and the knock on effect is a shifted balance between type III and type I and disordered fibril assembly in the skin and in the vessel wall ^[14]. A good deal of that detail comes from animals carrying the same fault rather than from human tissue though ^[15].

Kyphoscoliotic EDS: The collagen chains here are made perfectly well. What fails is the chemistry afterwards, as an enzyme deficiency stops the chains being hydroxylated and glycosylated properly, so the cross links that lock a fibril together never form correctly and the tissue ends up mechanically unstable ^[16]. A separate form of the same subtype points somewhere else again, at faulty folding and chaperoning, with collagen being retained inside the cell rather than less of it being made ^[17].

Dermatosparaxis EDS: The enzyme that trims procollagen down to its finished form is the problem here, so the molecule arrives in the matrix still wearing part of its packaging ^[18].

The organiser subtypes: Faults in tenascin X, in AEBP1 and in collagen XII leave the collagens themselves broadly normal and break the organisation instead ^[19][20][21]. Tenascin X is the one place where the lay version of the story is nearly right, as losing it does reduce how much collagen gets deposited, as well as changing how the fibrils form ^[19].

So, across the whole family, the recurring answer is which one, or how it was finished, rather than how much ^[1][2].

hEDS Is the One Most People Actually Mean

Most people asking this question have hypermobile EDS or hypermobility spectrum disorder, and that’s precisely where the collagen story stops being tidy. hEDS is the subtype with no confirmed genetic cause behind it, and there’s no collagen gene test that can diagnose it ^[1][22][10]. So a clear genetics panel means a great deal less than it gets treated as meaning in a consulting room, which is how a lot of people end up discharged on the strength of a “normal” result.

What has turned up instead is less satisfying than a single gene, and a lot more interesting. Skin cells from people with hEDS show a disorganised matrix, with collagen types I, III and V, fibronectin, the tenascins, the fibrillins and elastin all sitting in the wrong arrangement, alongside abnormal signalling between the cell and the matrix around it ^[23]. Those same cells behave as though they’re stuck partway into a wound healing state, shifting towards a scar forming cell type, and that shows up in hEDS and in HSD alike ^[24].

Now, a good deal of it looks like a maintenance problem rather than a building problem. Cells from people with hEDS and HSD show excessive breakdown of the matrix, with high levels of the enzymes that chew it up and fragments of type I collagen turning up as a result ^[25][26]. When those enzymes were blocked in cells in a dish, the matrix organised itself better and the scar forming behaviour partly reversed, which makes it a target rather than a treatment ^[25].

The same pattern of matrix fragments shows up in plasma in both hEDS and HSD, which raises the possibility of a blood test where at the moment there’s an examination and a set of clinical criteria ^[27]. It needs replicating and testing for specificity before anybody plans around it though ^[27].

And one more thing from that work, for anyone who’s been handed HSD as the consolation prize: at the level of what skin cells are actually doing, hEDS and HSD look like the same entity ^[26][24].

What a Skin Biopsy Actually Shows

Looking straight at the tissue under a microscope seems like it ought to answer the question outright (it’s usually the first thing people ask us about), and it doesn’t, quite. Some skin samples in hEDS show collagen flowers, which are the irregular flower shaped fibril cross sections that get quoted a lot, along with thinner fibrils, general disarray, or a shifted balance between type I and type III ^[28][29]. Plenty of other samples look entirely normal though, and how abnormal a sample looks doesn’t track cleanly with how hypermobile the person is ^[28][29].

Now, the other problem is that those appearances aren’t specific. Collagen flowers can support a diagnosis, particularly in classical EDS, but they turn up across more than one subtype, so finding them tells you considerably less than it appears to ^[30][7].

Look closer still, at individual fibrils at the nanoscale, and irregular, mechanically weaker fibrils turn up in localised disrupted patches of skin. That was the skin of one person with EDS though, so take it as a lead rather than a finding ^[31].

Not Every Hypermobile Joint Is a Collagen Story

Joint hypermobility is a trait before it’s anything else, and it turns up across a whole range of contexts, only some of which are a defined connective tissue disorder ^[32]. Some symptomatic hypermobile children and adults do show connective tissue differences through the body, including stretchier skin and altered markers of collagen breakdown ^[33]. Plenty of others have no identifiable collagen fault at all, and what’s driving their symptoms sits elsewhere: how well the joint can be sensed, how the muscles around it behave, the shape of the joint surfaces themselves, and how the tissue responds to being loaded ^[34][35][36].

That middle ground is where our own work lives. Position sense, meaning the ability to tell where a limb is without looking at it, comes out reduced on average in hypermobile groups ^[35]. A joint being controlled on a vaguer signal, is a genuinely different problem from a joint built out of weaker material, and in our opinion it’s a considerably more trainable one.

The picture has widened out past collagen as well, into proteomics, which is the whole spread of proteins a tissue is actually making (an ugly word for a sensible idea), and into how the matrix and the immune system talk to each other. That gives a broader model than weak collagen on its own ^[37][38][26].

What This Changes

Two things follow from all of this. The first is about what a negative test means: genetic testing can confirm several of the other subtypes, so a panel is genuinely useful where the picture suggests one of them ^[10][9]. But a clear panel doesn’t rule out hEDS, because there’s nothing on the panel that would pick it up ^[22][1]. If you’ve been sent away on the strength of normal collagen genetics, that was a misreading of what the test does.

The second is about where the effort goes. If the difference were bulk, then bulk would be the target. What gets described instead is a structure built from a slightly different mix, finished less reliably, and taken apart faster than it’s replaced, sitting underneath a joint that’s often being controlled on a poorer signal as well ^{[1][2][27][35]}. Those are separate problems with separate handles on them, and in our experience the last one is the most open to being worked on.

The Fibro Guy

References

[1] Malfait, F., Castori, M., Francomano, C.A., Giunta, C., Kosho, T. and Byers, P.H. (2020) ‘The Ehlers–Danlos syndromes’, Nature Reviews Disease Primers. https://doi.org/10.1038/s41572-020-0194-9

[2] Syx, D. and Malfait, F. (2024) ‘Pathogenic mechanisms in genetically defined Ehlers–Danlos syndromes’, Trends in Molecular Medicine. https://doi.org/10.1016/j.molmed.2024.06.001

[3] Tvaroška, I. (2024) ‘Glycosylation Modulates the Structure and Functions of Collagen: A Review’, Molecules. https://doi.org/10.3390/molecules29071417