Collagen peptides are broken down in the gut like any other protein. Some small fragments do reach the bloodstream intact, which is the basis of the proposed mechanism, but nothing about swallowing collagen delivers collagen to your skin. The randomized evidence on skin hydration and elasticity is real, mostly positive, mostly small, and mostly funded by companies that sell collagen. That last fact turns out to matter a great deal.
The version of this article that stood here argued a greens powder was a healthier alternative to collagen, on the strength of “DNA regeneration at the cellular level” and a “mild daily liver detox.” Those claims are retracted, and the defensible version of the argument is set out below.
What happens to collagen when you swallow it
Collagen is a structural protein made mostly of glycine, proline and hydroxyproline. Hydrolyzed collagen — collagen peptides — has already been cut into shorter chains, which is why it dissolves in a drink where gelatin would set.
Once swallowed, gastric and pancreatic proteases and brush-border peptidases keep cutting. Most of it arrives in circulation as free amino acids, indistinguishable from the amino acids in any other protein. The exception is a small fraction of hydroxyproline-containing di- and tripeptides that survive. Work in the Journal of Agricultural and Food Chemistry found Gly-Pro-Hyp and Pro-Hyp transported into blood and skin after collagen hydrolysate ingestion, and a 2022 paper in npj Science of Food identified a stable hydroxyproline-containing tripeptide in human blood after ingestion.
The proposed mechanism is that these peptides act as signals to dermal fibroblasts rather than as building material. That is a plausible hypothesis with in vitro support. It is not a demonstrated pathway in humans, and it should be described as a hypothesis wherever it appears.
What the skin trials found
A systematic review and meta-analysis in the Indian Journal of Dermatology, Venereology and Leprology pooled 10 randomized controlled trials with 646 participants, almost all women aged 31 to 62. Doses ranged from 246 mg to 10 g daily, with a median of 3.5 g and 4 g the most common, over one to three months. Pooled standardized mean differences were 1.25 (95% CI 0.77 to 1.74) for skin hydration and 0.61 (95% CI 0.21 to 1.02) for elasticity. The authors rated two studies at low risk of bias and eight as unclear, with heterogeneity around 55 percent.
A larger analysis by Myung and Park in the American Journal of Medicine in 2025 pooled 23 randomized trials and 1,474 participants. Overall, collagen supplementation was associated with improvements in hydration, elasticity and wrinkles. The subgroup analysis is the part worth reading: trials that were not funded by pharmaceutical or supplement companies showed no significant benefit, and neither did the subset judged higher quality. Significant improvement appeared in the low-quality studies.
The funding problem, stated plainly
That subgroup finding is the most important thing on this page. When an effect is present in industry-funded and low-quality trials and absent in independent, higher-quality ones, the most parsimonious explanation is that it is an artifact of how the trials were run and reported rather than a property of the supplement. That is not a reason to conclude collagen does nothing. It is a reason to withhold confidence until independent replication exists.
The analysis has been contested. A published correspondence in the same journal raised methodological objections, and industry groups argued it conflated different collagen types and misclassified some funding sources. Those objections deserve a hearing. They do not, on their own, restore the missing independent evidence.
The trials are also small — typically 40 to 120 participants over 8 to 12 weeks — and measure instrument readings of corneometry and cutometry rather than anything a person would notice in a mirror. A statistically significant change in a cutometer reading is not the same as visibly firmer skin.
What the joint evidence shows
The most cited joint trial is Clark and colleagues’ 24-week study in Current Medical Research and Opinion in 2008. It enrolled 147 athletes with activity-related joint pain, of whom 97 provided evaluable data, and gave 10 g of liquid collagen hydrolysate or placebo daily. Six visual analogue scale parameters improved significantly against placebo, including pain when walking, standing and at rest. The study was sponsored by GELITA Health GmbH, which supplied the product.
That trial then went to a regulator. In its 2011 scientific opinion, the European Food Safety Authority’s NDA Panel assessed collagen hydrolysate and maintenance of joints and concluded a cause-and-effect relationship had not been established. The panel noted that differences between groups did not hold up after adjustment for multiple comparisons, that dropouts compromised the analysis, and that subgroup analyses were not admissible. The health claim was refused.
Later work on undenatured type II collagen at much lower doses has produced positive results in knee osteoarthritis, again in small, largely sponsor-funded trials. The honest summary is that the joint evidence is weaker than the skin evidence, and the skin evidence is not strong.
Dose, and whether vitamin C matters
The doses used across the skin trials cluster between 2.5 and 10 g a day of hydrolyzed collagen, with 2.5 g and 5 g the most common single doses and treatment periods of 8 to 12 weeks. The joint trials used around 10 g. Products sold at a few hundred milligrams per serving are below anything meaningfully tested.
Vitamin C is a real requirement, not an upsell. Prolyl and lysyl hydroxylases cannot hydroxylate procollagen without ascorbate, which is why the deficiency state produces defective collagen — the biochemistry is set out by the Linus Pauling Institute. What follows from that is that adequacy matters, not that extra helps.
The most-quoted study on the combination is Shaw and colleagues in the American Journal of Clinical Nutrition in 2017. Eight healthy men took 5 g or 15 g of vitamin C-enriched gelatin an hour before intermittent exercise; the 15 g dose roughly doubled a blood marker of collagen synthesis, and serum from those subjects increased collagen content in engineered ligaments. Eight men, a biomarker and a tissue model: a reason to run bigger trials, not a reason to make claims.
| Claim on the label | What the evidence supports |
|---|---|
| 2.5–10 g hydrolyzed collagen peptides | The range used in the skin and joint trials |
| Under 1 g per serving | Below the tested range |
| “With vitamin C for absorption” | Vitamin C is a cofactor for collagen synthesis, not an absorption aid |
| “Marine collagen absorbs better” | No adequately powered head-to-head human trial supports this |
| “Plant-based collagen” | Plants do not make collagen. These are amino acid and vitamin blends |
Marine versus bovine
Marine collagen comes from fish skin and scales and is predominantly type I. Bovine collagen supplies types I and III. Both are hydrolyzed into the same classes of peptides before they reach you.
The marketing case for marine rests on lower molecular weight and better absorption. No adequately powered head-to-head randomized trial has tested that against bovine collagen for any clinical endpoint, and the meta-analyses above pooled both sources without separating them. Choosing between them on dietary, religious or sustainability grounds is reasonable. Choosing on absorption is choosing on an undemonstrated claim.
The honest alternative
Collagen is a poor-quality protein by conventional scoring. It contains no tryptophan and is low in the essential amino acids generally, including leucine. That is not a marketing point against it; it is why it should not be counted toward a protein target.
Oikawa and colleagues demonstrated the consequence in the American Journal of Clinical Nutrition in 2020. Twenty-two healthy older women took 30 g of whey protein or 30 g of collagen peptides twice daily for six days. Whey raised myofibrillar protein synthesis both acutely and over the longer measurement period, with and without resistance exercise. Collagen did not raise it above baseline in either condition.
So the defensible position is unglamorous. Eat enough total protein from sources with a complete amino acid profile, and get enough vitamin C. Those are the two inputs collagen synthesis requires, both are cheap, and plenty of adults fall short on both. Protein dose-response is covered in the piece on grass-fed whey and protein dosing, requirements in the article on vitamin C, and the broader picture in eating for healthy aging.
Assessing the claim this post used to make
The old article argued that a greens powder was a healthier alternative to collagen. Part of that survives scrutiny and part does not.
What survives: plants contain no collagen, because collagen is an animal structural protein. What a plant-rich diet can supply is the amino acids and the vitamin C that collagen synthesis requires, which is a genuine nutritional point — and the point every “vegan collagen” product is really making, whether or not the label says so.
What does not survive: the specific claims. “DNA regeneration at the cellular level” is not a meaningful statement about anything. “Alkaline, living water” and “mild daily liver detox” are not real physiological processes and we should not have printed them. And no evidence supports the claim that consuming a greens blend promotes collagen renewal, firms skin tone, or produces shiny hair and strong nails. We withdraw all of it. Where nutrient form and source genuinely do differ is set out in natural versus synthetic nutrients.
Greens Plus blends are formulated to contribute to a diet built on plants, with full disclosure of what each serving contains. We do not claim they improve skin, hair or nails, and no serving of any blended powder delivers the collagen doses used in the trials above.
Common questions
Do collagen supplements actually work for skin?
Pooled across all trials, the answer looks like a modest yes for measured hydration and elasticity. Restricted to trials that were independently funded or judged higher quality, the 2025 American Journal of Medicine analysis found no significant effect. That is an unresolved question, not a settled one in either direction.
How much should I take, and for how long?
Trials that reported effects used 2.5 to 10 g a day of hydrolyzed collagen peptides for 8 to 12 weeks. Products delivering a few hundred milligrams per serving are below anything that has been tested at all.
Does collagen count toward my daily protein?
Not usefully. It lacks tryptophan and is low in essential amino acids, and in the Oikawa trial 30 g twice daily did not raise muscle protein synthesis above baseline while the same dose of whey did. Count it as an addition, not as part of a protein target.
Do I need to take vitamin C with it?
Vitamin C is required for collagen synthesis, so adequacy matters. There is no good human evidence that extra vitamin C alongside a collagen supplement improves any outcome, and the best-known study on the combination had eight participants and measured a blood marker.
Is marine collagen better than bovine?
No adequately powered head-to-head randomized trial has compared them for a clinical outcome. Dietary preference, allergy and sustainability are reasonable grounds to prefer one. Absorption claims are not.
The bottom line
Collagen supplements are not nonsense, and they are not proven. The trials are small, the positive results cluster in industry-funded work, and the doses that were tested are higher than many products supply. Adequate total protein plus adequate vitamin C is the part of this that is not in dispute.
Greens Plus makes whole-food blends for people who want more plants in their day, with labels that state exactly what is in a serving. The range is at greensplus.com.
About the Author
