Does Collagen Work? What the Evidence Says

You’ve probably seen collagen sold as the fix for your skin, your joints and your hair, all from one scoop stirred into your coffee. It’s in creamers now. It’s in gummies. It’s in a lot of very confident before-and-after photos.

So does it do anything? For skin, the honest answer changed in 2025, and most of the marketing hasn’t caught up.

Short answer: Maybe a little, and less than the tubs suggest. Pooled trials show small improvements in skin hydration and elasticity, but a 2025 meta-analysis in the American Journal of Medicine found those improvements disappeared when it looked only at trials without company funding, or only at high-quality trials. The joint, bone and muscle studies are small, and most were paid for or run by collagen makers.

What happens to collagen after you swallow it

First, the part the labels skip. Your body doesn’t absorb collagen as collagen.

Digestion breaks it down into amino acids and short fragments of two or three amino acids. When Japanese researchers gave people gelatin hydrolysate, they found those fragments in the blood one to two hours later, mostly one called Pro-Hyp (a two-amino-acid piece that’s typical of collagen). That’s real, and it’s the basis for the theory that the fragments act as a signal to the cells that make collagen.

What the fragments can’t do is pick a destination. Nothing sends the peptides in your morning coffee to your cheekbones rather than your knees, or your hair, or the protein your body happens to need that day.

There’s a second catch if you think of it as a protein supplement. Collagen contains no tryptophan, which makes it an incomplete protein. A 2019 review in Nutrients worked out that collagen peptides can still make up to 36% of daily protein without throwing the amino acid balance off. Worth knowing: two of its three authors work for Gelita, one of the largest collagen suppliers. So if you’re counting a scoop toward your protein for the day, it isn’t doing the same job as a scoop of whey or a complete plant blend.

The skin research, and the 2025 split

This one needs a little patience, because the headline number and the real finding point in opposite directions.

The earlier meta-analyses looked good for collagen. A 2021 review in the International Journal of Dermatology pooled 19 randomized trials with 1,125 participants, almost all women aged 20 to 70, and found hydrolyzed collagen beat placebo on hydration, elasticity and wrinkles. A 2023 review in Nutrients pooled 26 trials and 1,721 people and found better hydration and elasticity, with stronger results past eight weeks.

Then Myung and Park did something those reviews hadn’t. Their 2025 meta-analysis in the American Journal of Medicine pooled 23 trials with 1,474 participants, got the same positive result overall, and then split the trials by who funded them.

The trials without company funding showed no effect on hydration, elasticity or wrinkles. None! The high-quality trials showed no significant effect in any of the three either. The authors’ conclusion was that there’s no clinical evidence collagen supplements prevent or treat skin aging.

The industry pushed back, and fairly enough it should be heard. The Collagen Stewardship Alliance told NutraIngredients that some trials were put in the wrong funding group and some doses and durations were recorded wrong. Maybe so. But a June 2026 review of 25 trials in the European Journal of Clinical Nutrition rated most of them at high risk of bias anyway, which is the same weakness from a different angle.

Here’s how I read it. When the strongest evidence for a product mostly comes from the people selling it, and the independent studies come back flat, the fair summary isn’t “it works.” It’s “it hasn’t been shown to.”

The rest of the claims, study by study

Claim	Best evidence we found	Who paid or ran it	Our read
Skin hydration and elasticity	Meta-analyses of 19 to 26 trials	Mostly industry-funded; independent trials showed no effect	Not shown
Joint pain in athletes	Penn State trial, 147 randomized, 97 analyzed	Funding not disclosed in the abstract	EFSA rejected the claim in 2011
Knee osteoarthritis (UC-II, 40 mg)	One trial, 191 people	First two authors worked for the maker	Promising, one maker-run trial
Bone density after menopause	One trial, 131 women, 102 finished	Partly funded by Gelita	Promising, one trial
Muscle with strength training	One trial, 53 older men	Gelita supplied the product and paid part of the costs	Small, one trial

Joints, bones and muscle: small trials, familiar funders

The best-known joint study got turned down by a regulator. In 2008, a Penn State trial randomized 147 athletes to 10 grams of collagen hydrolysate or placebo for 24 weeks. Only 97 could be analyzed at the end, and the collagen group reported less joint pain on six measures, most clearly among the ones with knee pain. When the European Food Safety Authority reviewed that same trial for Gelita’s joint health claim, it found no significant differences on any endpoint once the results were adjusted for multiple comparisons, and rejected the claim in 2011.

You may have seen UC-II on a joint supplement. It’s a different product altogether: 40 milligrams of undenatured type II collagen from chicken cartilage, not a heaping scoop of peptides. In a 2016 trial of 191 people with knee osteoarthritis, it beat both placebo and glucosamine with chondroitin on a standard pain-and-stiffness score over 180 days. The first two authors worked for the company that makes it.

Bone has one decent trial. König and colleagues gave 131 postmenopausal women 5 grams a day of a specific Gelita peptide or placebo for a year. 102 finished, and bone density rose at the spine and the hip. Part of the study was funded by Gelita.

Muscle is the smallest of the lot. In a 2015 trial, 53 men averaging 72 years old, all with age-related muscle loss, did 12 weeks of strength training with either 15 grams of collagen or placebo. The collagen group gained 4.2 kilograms of fat-free mass against 2.9 for placebo. Gelita supplied the collagen and paid part of the costs.

None of these results are made up. What they have in common is size, a single trial each, and a collagen supplier somewhere in the acknowledgments. They also tested specific branded peptides at specific doses, so the generic tub on the shelf isn’t necessarily the thing that was studied.

Why the label can say it anyway

The claims on a collagen tub (“supports skin elasticity”, “promotes joint comfort”) are what the FDA calls structure/function claims. A company doesn’t need approval to make one. It notifies the FDA within 30 days of selling the product and prints the line you’ve seen a thousand times: this statement has not been evaluated by the Food and Drug Administration.

Europe runs the other way. Health claims need to clear EFSA first, and collagen has struggled there, including a 2013 opinion that Gelita’s VeriSol P hadn’t been shown to improve skin elasticity. Same ingredient, same studies. Different rules about what you’re allowed to print. It’s the same regulatory gap we cover in the red flags in the supplement industry.

Is it safe to take?

For most people, it looks that way. The 2026 review found adverse events were low across the trials it examined.

Two exceptions deserve a sentence each. Collagen is itself a recognized fish allergen, and heating didn’t stop patients’ antibodies reacting to it, so marine collagen is worth skipping if you’re allergic to fish (more on that in marine vs bovine collagen). And collagen powders have tested positive for lead, which we go through in heavy metals in collagen.

Who this is actually for

If you like collagen in your coffee, the price doesn’t sting, and you think of it as a modest protein top-up, there’s no strong reason to stop. Just don’t count it as a complete protein.

If you’re buying it to undo wrinkles, I’d put the money toward sunscreen instead. An Australian randomized trial of 903 adults found daily sunscreen users showed no detectable increase in skin aging over four and a half years. And if it’s your knees, UC-II is a different product with its own maker-run trial, so talk it through with your doctor rather than grabbing the biggest tub.

Whatever you decide, the product choice still matters. A tested powder with a short label beats an untested one with a longer list of promises, and our vetted collagen picks are the ones that publish their testing. If the “grass-fed” line on the front is what sold you, read which collagen sourcing claims are actually verified first.

Frequently asked questions

Does collagen really work for skin?

It hasn’t been shown to in independent research. Pooled trials show small gains in hydration and elasticity, but a 2025 meta-analysis of 23 trials found no effect in the trials without company funding and no significant effect in the high-quality trials. The industry disputes how some trials were classified.

How long does collagen take to work?

The trials that reported skin improvements usually ran eight to 12 weeks, and a 2023 review found stronger results past eight weeks. Keep in mind those are the same trials whose results didn’t hold up when independent studies were analyzed on their own. Anyone promising visible results in two weeks is going well beyond the research.

Does collagen help with joint pain?

The evidence is thin. The best-known trial in athletes reported less joint pain, but Europe’s food safety regulator found no significant differences once the analysis was corrected, and rejected the claim. UC-II, a small daily dose of undenatured type II collagen, did better than placebo in one knee osteoarthritis trial run by its maker.

Is collagen a complete protein?

No. Collagen has no tryptophan, one of the essential amino acids, so it can’t replace a complete protein like whey, soy or a mixed plant blend. It can make up part of your daily protein without causing a problem, but it shouldn’t be the whole plan.

Can collagen supplements be harmful?

Side effects in trials were uncommon. The real risks are specific: marine collagen can trigger reactions in people with fish allergies, and independent testing has found lead in some collagen powders, so a brand that publishes heavy-metal results is worth the extra few dollars.