The short version
- Pooling 26 randomised controlled trials in 1,721 participants, hydrolysed collagen improved skin hydration with a standardised mean difference of 0.63 and elasticity with 0.72 — moderate effects, not dramatic ones.
- The same review reported bias from missing outcome data in 13 of its 26 included trials, which is half the evidence base, and called for further large-scale trials.
- Skin outcomes in these trials are instrument readings from a corneometer or a cutometer, which is not the same measurement as noticing a difference in the mirror.
- The joint evidence is thinner: one 2008 trial in 147 college athletes taking 10g daily for 24 weeks improved walking pain by 1.11 points against 0.46 on placebo, with 97 producing evaluable data.
- Reviews put the effective duration at around 90 days at doses between 2.5g and 10g daily, so four weeks is not a fair trial of collagen.
Somebody emails us about this every few weeks: why doesn't LUVO make a collagen powder? It is the single most requested product we don't sell, and the honest answer takes about eight hundred words, so here they are.
Let's start where the evidence is strongest, because there genuinely is some.
What the skin meta-analyses found
A 2023 systematic review and meta-analysis in Nutrients pooled 26 randomised controlled trials covering 1,721 participants. Hydrolysed collagen significantly improved skin hydration, with a pooled standardised mean difference of 0.63 (95% CI 0.38 to 0.88), and skin elasticity, at 0.72 (95% CI 0.40 to 1.03). Those are moderate effect sizes with confidence intervals that stay comfortably clear of zero. It is not nothing.
A 2021 meta-analysis in the International Journal of Dermatology reached a similar place from a different set of papers: 19 randomised, double-blind trials, 1,125 participants aged 20 to 70, 95% of them women. Pooled results favoured hydrolysed collagen over placebo for hydration, elasticity and wrinkles, and the authors put the effective duration at around 90 days.
So the direction of travel is consistent across independent reviews. That is more than most ingredients in this category can claim, and we are not going to pretend otherwise just because we don't stock it.
Now the parts the marketing leaves out
The Nutrients authors ran a formal risk-of-bias assessment on their own included trials and reported bias from missing outcome data in thirteen of them — half the evidence base. Their own conclusion is that collagen "can have positive effects" but that further large-scale randomised trials are necessary to confirm the findings. That is a review being straight about its own limits, and it deserves to be quoted as carefully as its effect sizes.
Their subgroup analysis also found the hydration effect varied by the source of the collagen and by how long people took it. Which means a pooled effect size across 26 trials is not a promise about any particular tub.
Then there is what the outcomes actually are. "Hydration improved" is usually a corneometer reading. "Elasticity improved" is usually a cutometer reading. Those are legitimate instruments and they measure real properties of skin. They are not the same measurement as looking in a mirror in April and thinking you look different from January, and the gap between those two things is where a lot of category marketing lives.
A 2019 systematic review in the Journal of Drugs in Dermatology — 11 randomised placebo-controlled trials, 805 patients, doses from 2.5g to 10g daily over 8 to 24 weeks — opens by noting that for these products "regulations are lacking on quality, absorption, and efficacy." Its verdict was that preliminary results are promising and further study is needed. Promising is a real word. It is also not the word on most labels.
The joint evidence is thinner than the skin evidence
The trial everyone cites is Clark and colleagues, 2008, in Current Medical Research and Opinion. One hundred and forty-seven college athletes with activity-related joint pain and no evidence of joint disease took 10g of collagen hydrolysate daily or placebo for 24 weeks. Six pain parameters reached statistical significance; joint pain when walking improved by 1.11 points versus 0.46 on placebo (p=0.007).
Statistically significant, and worth knowing. Also: those are differences of well under a point on a visual analogue scale, only 97 of the 147 subjects produced evaluable data, and it is a single trial from 2008. The authors themselves close by saying future studies are needed to support the findings. Eighteen years later, that sentence is doing more work than a lot of product pages admit.
On safety, briefly
Collagen comes out well here. The 2019 dermatology review reports that oral collagen supplementation was generally safe across the trials it examined, with no reported adverse events. If you are already taking it, nothing in this article is a reason to stop.
So what should you take from this?
If collagen is already part of your routine and you like it, keep going — but judge it on the timeline the trials used. Ninety days at a dose in the range the studies actually tested, then decide honestly whether anything changed. Not four weeks, and not on the strength of a before-and-after photo taken in different lighting.
And if you are deciding whether to start: the evidence supports a modest, measurable effect on skin hydration and elasticity, weaker support for joint comfort under load, and a good safety record. That is the whole claim. Anything beyond it is somebody's copywriter.
As for why we don't make one — we would have to write a product page for an ingredient whose honest summary is "moderate effect, imperfect trials, give it three months." We haven't yet found a way to do that which is both truthful and worth your money. If that changes, we'll tell you. Until then we would rather point you at the meta-analyses and let you make the call yourself.
Good questions
Does collagen actually work?
For skin, modestly, and the evidence is more consistent than most of this category. A 2023 meta-analysis of 26 trials in 1,721 people found improved hydration and elasticity with moderate effect sizes, and a 2021 meta-analysis of 19 trials reached the same place from different papers. We do not sell collagen and we are still not going to pretend the evidence is absent. It is real and it is modest.
How long does collagen take to work?
About 90 days. The 2021 meta-analysis put the effective duration at around three months, and trials in this literature ran 8 to 24 weeks. Judge it on that timeline at a dose in the range the studies used, then decide honestly whether anything changed. Not four weeks, and not on the strength of a before-and-after photo taken in different lighting.
How much collagen should I take?
The trials used between 2.5g and 10g daily. The joint-pain study used 10g. There is no strong basis in this evidence for going above that range, and the subgroup analysis in the 2023 review found the hydration effect varied by the source of the collagen and by duration, which means a pooled effect size across 26 trials is not a promise about any particular tub.
Does collagen help joint pain?
The support is weaker than for skin. The trial everyone cites gave 147 college athletes with activity-related joint pain 10g daily for 24 weeks; walking pain improved 1.11 points against 0.46 on placebo. Statistically significant and worth knowing, but those are differences of well under a point on a visual analogue scale, only 97 subjects produced evaluable data, and it remains a single trial from 2008.
Is collagen safe to take?
It comes out well on safety. The 2019 dermatology review of 11 randomised placebo-controlled trials reported that oral collagen was generally safe with no reported adverse events. If you are already taking it and you like it, nothing in this article is a reason to stop. The same review does note that regulations are lacking on quality, absorption and efficacy across these products.
Why doesn't LUVO sell a collagen powder?
Because we would have to write a product page for an ingredient whose honest summary is moderate effect, imperfect trials, give it three months. It is the single most requested product we do not stock, and we have not yet found a way to sell it that is both truthful and worth your money. If that changes we will tell you. Until then we would rather point you at the meta-analyses.
Sources
- Pu SY, Huang YL, Pu CM, Kang YN, Hoang KD, Chen KH, Chen C. Effects of Oral Collagen for Skin Anti-Aging: A Systematic Review and Meta-Analysis. Nutrients, 2023. View study
- de Miranda RB, Weimer P, Rossi RC. Effects of hydrolyzed collagen supplementation on skin aging: a systematic review and meta-analysis. Int J Dermatol, 2021. View study
- Choi FD, Sung CT, Juhasz ML, et al. Oral Collagen Supplementation: A Systematic Review of Dermatological Applications. J Drugs Dermatol, 2019. View study
- Clark KL, Sebastianelli W, Flechsenhar KR, et al. 24-Week study on the use of collagen hydrolysate as a dietary supplement in athletes with activity-related joint pain. Curr Med Res Opin, 2008. View study
These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.
