Search for almost any GLP-1 side effect (loose skin, thinning hair, aching knees) and you will land on a collagen ad within two clicks. An entire genre of "best collagen for GLP-1 users" pages now exists, which tells you something real about demand and nothing at all about whether it works.
Collagen does something, just less than most of those pages claim. The evidence differs for skin, tendons, and joints, and a single verdict for or against gets all three wrong. Here's what you're buying.
First: three different products share one name
This is the distinction that explains most of the confusion, and almost no product page makes it.
| What's on the label | Typical dose | How it's supposed to work |
|---|---|---|
| Gelatin (cooked collagen) | 5–15 g | Dissolves warm, sets cold. Raw material only. |
| Hydrolyzed collagen / collagen peptides | 2.5–15 g per day | Raw material, plus small peptides that may act as signals |
| Undenatured (native) type II collagen — "UC-II" | 40 mg per day | Immune signaling through the gut. Not raw material. |
Look at that dose gap: 10 grams versus 40 milligrams. A 250-fold difference. At 40 mg, nothing can supply building material for a joint, so the two can't be working the same way. Undenatured type II is thought to work by presenting the intact collagen molecule to immune tissue in your gut, calming the inflammatory response aimed at cartilage. Hydrolyzed collagen is doing something else entirely.
So when a page says "collagen helps joints," ask which one. They have separate evidence bases and are not interchangeable.
What the type numbers mean, and where the label story breaks down
The collagen family has 28 members (Ricard-Blum, Cold Spring Harbor Perspectives in Biology, 2011, DOI (external link)). Four matter here:
- Type I: the tensile workhorse. It makes up most of the collagen in your skin's dermis and in tendon and ligament, and it's also in bone.
- Type III: co-assembles with Type I in skin, blood vessels, and gut. Forms finer fibres; the ratio of III to I shifts with age.
- Type II: the collagen of hyaline cartilage, the smooth surface inside your joints. Also spinal discs.
- Type V and X: supporting roles; you'll see these on eggshell-membrane products.
Supplements map onto those fairly cleanly at the source stage. Bovine hide gives you Types I and III. Marine and fish-skin collagen is mostly Type I. Chicken sternal cartilage gives Type II, usually alongside chondroitin sulfate and hyaluronic acid. Laboratory extraction from that source yields collagen chains around 130–270 kDa and chondroitin sulfate around 80 kDa (Ata et al., Food Chemistry, 2024, DOI (external link)).
Here's the part the marketing skips. You do not absorb collagen as collagen. It's a protein; your gut takes it apart. What crosses into your blood is free amino acids plus a small fraction of short fragments containing hydroxyproline, chiefly Pro-Hyp and Hyp-Gly. In a study of 9 people given gelatin, Pro-Hyp peaked at about 15.5 nmol/mL two hours after the dose and Hyp-Gly at about 2.3 nmol/mL after one hour (Iwasaki et al., Food Chemistry, 2021, DOI (external link)).
Those fragments look broadly similar whether they came from a cow, a fish, or a chicken. Which means the confident claim that Type II peptides go to your cartilage while Type I peptides go to your skin has never been demonstrated. There is no adequately powered head-to-head trial comparing collagen types for the same outcome. The type label describes what was extracted, not where the pieces end up.
The exception is real and worth knowing: for undenatured type II, the intact molecule is the mechanism, so the type does matter there. For hydrolyzed powders, treat the type number as a description of the source animal, not a delivery address.
One thing that may differ between products is the peptide profile: how the manufacturer cut the protein up. Processing choices measurably change which fragments you end up with. That's why the trials are run on specific branded peptides rather than "collagen" generically, and why a tub with no trial behind it is a bet on a different product than the one that was studied.
The three promises, graded separately
Skin: probably a real effect on hydration, weaker on everything else
This has the most trials and the shakiest quality. Multiple meta-analyses report improvements: 19 randomized trials with 1,125 participants found favorable pooled effects on hydration, elasticity, and wrinkles over about 90 days (de Miranda et al., International Journal of Dermatology, 2021, DOI (external link)). A 10-trial analysis found a large hydration effect and a moderate elasticity effect, at a typical dose of about 4 g daily (Danessa et al., Indian Journal of Dermatology, Venereology and Leprology, 2025, DOI (external link)). A 26-trial pooling of 1,721 people agreed (Pu et al., Nutrients, 2023, DOI (external link)).
Then a 2025 analysis did something none of the others had: it split 23 trials by who paid for them and how well they were run.
In studies not funded by pharmaceutical companies, there was no effect on hydration, elasticity, or wrinkles. Industry-funded studies showed effects in all three. High-quality studies showed no significant effect in any category (Myung & Park, The American Journal of Medicine, 2025, DOI (external link)).
A 2026 review of 25 trials reached a gentler version of the same conclusion: improvements in most trials, but "the majority of included trials were rated as having a high risk of bias" (Bassila et al., European Journal of Clinical Nutrition, 2026, DOI (external link)).
That doesn't mean collagen does nothing for skin. It means the published effect sizes are an optimistic ceiling. And the most mechanistically convincing trial found something narrower than the ads promise: 99 women taking 1 g or 5 g for 12 weeks had more water in the outer skin layer and less water loss through it, driven by a rise in natural moisturizing factor. But skin elasticity and skin thickness did not change (Miyanaga et al., Skin Pharmacology and Physiology, 2021, DOI (external link)). An earlier trial did find elasticity gains at 2.5 g and 5 g over 8 weeks, which persisted a month after stopping in older women (Proksch et al., Skin Pharmacology and Physiology, 2013, DOI (external link)).
Better-hydrated skin does look better and makes fine lines less obvious. That's a fair claim. "Rebuilds your dermal collagen" is not.
Grade: modest and plausible for hydration. Weak for wrinkles. See our loose skin guide for why none of this reverses hanging excess skin.
Tendons: the best mechanism, the most contradictory results
Tendon is where collagen supplementation has its most interesting idea, and it comes with a condition attached.
Tendon collagen is mostly Type I, and tendon only builds new collagen when it's been loaded. So the strategy is: take the raw material shortly before you train, and let the exercise provide the signal. In the foundational study, 15 g of vitamin C-enriched gelatin taken one hour before six minutes of rope skipping doubled a blood marker of collagen synthesis (Shaw et al., The American Journal of Clinical Nutrition, 2017, DOI (external link)). A follow-up crossover study in 10 men saw the marker trend about 20% higher after gelatin and hydrolyzed collagen, but not after placebo or a gummy version. The variability was large, and none of the results were statistically significant (Lis & Baar, International Journal of Sport Nutrition and Exercise Metabolism, 2019, DOI (external link)).
Vitamin C matters here because it's the required cofactor for the enzymes that stabilize collagen's triple helix.
Over longer training programs, though, good trials disagree outright:
- 50 men taking 5 g daily through 14 weeks of heavy leg training grew significantly more patellar tendon cross-sectional area than placebo (Jerger et al., European Journal of Sport Science, 2023, DOI (external link)).
- 39 men taking 15 g daily through 15 weeks of lower-body training showed no difference from placebo in tendon size, stiffness, or mechanical properties (Balshaw et al., Medicine & Science in Sports & Exercise, 2023, DOI (external link)).
- 22 elite female masters athletes taking 30 g with 500 mg vitamin C before eccentric training for 8 weeks gained more tendon size and rate of force development than placebo (Nulty & Erskine, International Journal of Sport Nutrition and Exercise Metabolism, 2025, DOI (external link)).
Two of three positive, but the one null trial used a dose between the two positive ones, so dose alone doesn't explain it. A systematic review of nutrition and tendon problems found collagen peptides the most-studied option and combined training-plus-supplement approaches promising, while concluding that firm dietary recommendations for tendon problems "remain elusive" (Hijlkema et al., Journal of the International Society of Sports Nutrition, 2022, DOI (external link)).
Grade: promising, unsettled, and completely dependent on training. If you're not loading the tissue, there's no mechanism for it to work. See why strength training matters.
Joints: the most clinically meaningful evidence, and a competitor that beats it
Joint trials measure things patients feel, such as pain and function scored over three to six months, instead of instrument readings of skin. That makes this the most decision-relevant evidence of the three.
Both routes have support. On the hydrolyzed side, 250 people with knee osteoarthritis taking 10 g daily for six months reported significant improvement in joint comfort, with the largest benefit in those whose joints were worst and whose habitual meat intake was lowest (Benito-Ruiz et al., International Journal of Food Sciences and Nutrition, 2009, DOI (external link)). In a different population, 139 athletes with functional knee pain but no arthritis, 5 g daily for 12 weeks significantly reduced activity-related pain (Zdzieblik et al., Applied Physiology, Nutrition, and Metabolism, 2017, DOI (external link)). That second group may be closer to a lot of GLP-1 users: knees that complain during new activity, not diagnosed osteoarthritis.
On the undenatured side, 40 mg of UC-II daily outperformed both placebo and a combined glucosamine-chondroitin regimen over 180 days (Lugo et al., Nutrition Journal, 2016, DOI (external link)). A review of 8 trials covering 243 UC-II patients found better pain and function scores plus improved walking tests, with side effects comparable to other supplements, while noting the literature is still limited (Kumar et al., American Journal of Translational Research, 2023, PubMed (external link)).
The fairest comparison comes from a network meta-analysis that put seven supplements side by side across 39 trials and 4,599 patients. Collagen showed real benefits on some outcomes, but Boswellia ranked highest for pain and stiffness, with curcumin and krill oil also competitive (Zhang et al., Nutrients, 2025, DOI (external link)).
Grade: the strongest of the three, with modest effects. If joint pain is your only goal and you're buying one thing, collagen isn't clearly the best pick.
The one myth that can actually cost you
Most collagen myths just waste money. This one can work against the thing you're on a GLP-1 to protect.
"Collagen counts toward my protein goal." It shouldn't. Collagen contains no tryptophan and is low in the essential amino acids — including leucine, the one that triggers muscle protein synthesis. It is protein by the label's counting rules and not by your muscle's.
That's a general nutrition footnote for most people and a real problem for you. Appetite suppression already makes protein the hardest target to hit on a GLP-1, and rapid weight loss already costs lean mass. Swapping a 20 g whey shake for a 20 g collagen scoop, because a supplement page said protein is protein, means eating the same number on the label and considerably less of what protects muscle. If you want collagen, add it on top of your protein target, never in place of part of it. Our guide on reading protein quality off a label covers how to spot this, and high-protein meals and snacks for low appetite covers hitting the target itself.
Other claims circulating
- "Type II for joints, Type I/III for skin." Overstated for hydrolyzed powders, for the reasons above. True only for undenatured type II.
- "Collagen tightens loose skin." No trial has shown this. Hydration and elasticity measurements are a different outcome from excess skin after major weight loss. Our loose skin guide covers what actually helps.
- "Collagen creams and sheet masks rebuild collagen." Collagen molecules are far too large to cross the skin barrier. Topicals can hydrate the surface; they don't reach the dermis.
- "Marine collagen is absorbed better, so it's superior." Marine peptides do tend to be smaller. No clinical trial has compared marine and bovine collagen head-to-head for the same outcome. The closest thing is indirect: in the 26-trial skin meta-analysis, 14 trials used fish collagen and only one used bovine, and the authors' subgroup analysis favored fish for hydration while finding no difference by source for elasticity (Pu et al., Nutrients, 2023, DOI (external link)). Comparisons across separate trials can't show that "better absorbed" means "works better."
- "Collagen is a natural GLP-1" or helps you lose weight. There is no credible evidence for this. Treat any product making that claim as a signal about the seller.
A note on what's in the tub
Collagen is concentrated from animal hides, bones, and fish scales, tissues that accumulate heavy metals if the animal was raised on contaminated ground. This has reached the courts: a California Proposition 65 case over lead and cadmium in Vital Proteins supplements settled in 2018 with limits on exposure without a warning (California Attorney General (external link)). Labeling is a separate risk. In July 2026 a marine-collagen supplement sold on TikTok was recalled because its label didn't disclose that the collagen came from fish, an allergen (ConsumerLab (external link)). Supplements aren't reviewed by the FDA for safety or efficacy before sale, so third-party testing is the only real check you have. Look for it, and be skeptical of a brand that doesn't publish results.
If you want to try it
Collagen peptides are safe, well tolerated, and cheap enough that trying them is low-stakes. A reasonable approach: 5–10 g daily of a hydrolyzed powder that has actually been in a trial, taken 30–60 minutes before training on days you lift, with vitamin C. That single habit covers the plausible skin benefit and the plausible tendon benefit and lands in the studied joint range. Give it 12 weeks. Expect better-hydrated skin, possibly slightly less knee complaint during activity, and no visible change to hanging skin.
And keep it separate from your protein target. That one distinction matters more than any type number on the front of the tub.
Research findings above are attributed to peer-reviewed articles indexed in PubMed. This is general education, not medical advice; discuss supplements with your prescriber, particularly if you have a fish, shellfish, egg, or beef allergy, or kidney disease.