Collagen Supplements: What the Skin and Joint Studies Actually Show

Collagen is the best-known ingredient in this category, and also the clearest example of why “clinically proven” needs a follow-up question: proven by whom, and how well? A 2025 systematic review pooled 23 skin-focused trials and found real, measurable benefits overall — but that same review found the benefit disappeared once industry-funded and lower-quality trials were set aside, and the supplement industry has pushed back hard on how that review was conducted. Joint health tells a different, more reassuring story: a larger, more recent trial-sequential meta-analysis found small-to-moderate but consistent benefits for pain and function in osteoarthritis, less entangled with the funding question. Both things can be true about the same ingredient at once, and a reader deserves to know which claim they’re looking at.

The short version

• Collagen is the most abundant protein in the body, making up over 90% of skin’s structural mass. Natural collagen production declines with age (roughly 1–1.5% per year), which is the biological rationale marketers lean on for both skin and joint claims.
• For skin: a 2025 systematic review and meta-analysis (23 RCTs, 1,474 participants) found that when all trials were pooled, collagen supplementation showed statistically significant improvements in hydration, elasticity, and wrinkles — but when the researchers split trials by funding source, studies not funded by supplement or pharmaceutical companies showed no significant effect, and when they split by study quality, only the lower-quality studies showed a significant benefit for elasticity; the higher-quality studies showed no significant effect in any category. The authors’ own conclusion: “there is currently no clinical evidence to support the use of collagen supplements to prevent or treat skin aging.”
• That 2025 review is genuinely contested, not settled science, and it’s important to say so plainly. Industry groups (the Collagen Stewardship Alliance and the Gelatin Manufacturers of the World, among others) have publicly disputed it, alleging specific dose- and duration-reporting errors in several included trials and misclassification of some trials’ funding source. We have not independently re-verified either side’s claims trial-by-trial; the honest position is that this is an active scientific dispute, not a closed question in either direction.
• For joints, the picture is more consistent: a 2024 trial-sequential meta-analysis (35 RCTs, 3,165 patients, published in Osteoarthritis and Cartilage) found collagen derivatives produced small-to-moderate improvements in pain and function in people with osteoarthritis, with moderate-to-high certainty evidence and a good safety profile — a materially stronger evidence base than the skin claims, though still modest in effect size and mostly studied over less than six months.
• Collagen supplements are generally well tolerated, with no major documented drug interactions. The one real caution: as a protein source, high intakes add to total daily protein load, which matters specifically for people with advanced kidney disease who need to manage protein intake.
• “Collagen” isn’t one thing. Trials use different types (Type I, II, III), different sources (bovine, marine, porcine), different forms (hydrolyzed peptides vs. undenatured), and different doses (commonly 2.5–10 g/day for skin studies, though the 2025 review noted many of its included Asian trials used lower doses around 3 g/day) — a real source of inconsistency across the research base that any single verdict has to gloss over somewhat.

The skin claim: a genuinely contested meta-analysis

Collagen’s marketing centers heavily on skin — hydration, elasticity, wrinkle reduction, “glowing from within.” The strongest available evidence on this specific question is a 2025 systematic review and meta-analysis published in The American Journal of Medicine (Myung & Park), which pooled 23 randomized controlled trials with 1,474 total participants.

The topline, pooled result was positive: collagen supplementation was associated with statistically significant improvements in skin hydration, elasticity, and wrinkle reduction when every trial was averaged together. If the story stopped there, this would be a straightforward “Moderate” or better rating.

It doesn’t stop there. The same authors ran two further breakdowns of their own data:

• By funding source: trials not funded by pharmaceutical or supplement companies showed no statistically significant benefit in hydration, elasticity, or wrinkles. Trials that were industry-funded showed significant positive effects.
• By study quality: when the researchers restricted the analysis to higher-quality trials only, no significant effect remained in any category. Only the lower-quality subset of trials showed a significant benefit (specifically for elasticity).

Based on this, the authors concluded there is “currently no clinical evidence to support the use of collagen supplements to prevent or treat skin aging” — a notably stronger statement than the pooled headline number alone would suggest, and specifically because the pooled number appears to be carried by trials with a financial or methodological reason to find a positive result.

This finding has been publicly disputed, and that dispute deserves to be represented honestly rather than smoothed over. Industry groups responded directly to the published review:

• The Collagen Stewardship Alliance (CSA) alleged specific data-transcription errors in several included trials — for example, a trial’s actual dose or study duration recorded incorrectly in the meta-analysis’s data tables — and argued that some trials the review classified as “independent” actually had commercial ties, which (if true) would undermine the funding-based subgroup comparison.
• The Gelatin Manufacturers of the World (GROW) argued the review’s abstract and conclusion were internally inconsistent (acknowledging a significant pooled benefit, then concluding “no clinical evidence”), that the high/low quality classification method wasn’t disclosed using a named, validated tool, and that other meta-analyses (e.g., De Miranda et al. 2021) have found significant skin benefits at similar doses.
• Individual ingredient suppliers questioned whether pooling many different collagen types, sources, and doses together as one category was methodologically sound in the first place.

We have not independently re-checked the disputed dose/duration figures trial-by-trial, and we are not positioned to referee this dispute definitively. What we can say confidently: the existence of a documented funding-outcome pattern (independently-funded trials showing null results, industry-funded trials showing positive ones) is itself a meaningful, disclosable fact about this evidence base, consistent with a well-established pattern in nutrition research generally — and it’s a good enough reason on its own to rate the skin claim as Limited rather than Moderate, regardless of how the specific data-error allegations eventually resolve.

The joint claim: a stronger, less contested case

Collagen’s joint-health claims rest on different, and currently more consistent, evidence. A 2024 trial-sequential meta-analysis published in Osteoarthritis and Cartilage — a more rigorous statistical method than a standard meta-analysis, designed specifically to check whether the accumulated evidence has reached a reliable conclusion rather than a false-positive from too few trials — pooled 35 randomized controlled trials covering 3,165 patients (main analysis on 25 RCTs, 2,856 patients) with osteoarthritis.

The finding: collagen derivatives produced small-to-moderate effects on pain reduction, with moderate-to-high certainty of evidence, and similarly supported improvements in physical function. Safety was reassuring — collagen derivatives were not associated with an increased risk of adverse events or trial withdrawal compared to control. The main caveat the authors themselves raised: most included trials followed patients for under six months, so longer-term efficacy and durability of benefit remain open questions.

Individual, more recent randomized trials in knee osteoarthritis specifically have reported reductions in pain scores and inflammatory markers (CRP, ESR) alongside functional improvement, consistent with the pooled picture. This is a meaningfully different evidence situation than the skin literature: while individual trials still vary in collagen type, dose, and source, the joint-health signal has not (so far) been shown to collapse when isolating higher-quality or independently funded trials the way the skin signal did in the 2025 review — though we note that a funding-source-specific breakdown for the joint literature specifically, comparable to what Myung & Park ran for skin, is not something we independently verified exists or was checked here.

One compliance-relevant point worth naming directly: osteoarthritis is a diagnosed medical condition, not a cosmetic concern. Any future product content citing this joint-health evidence needs to describe supporting “joint comfort” or “joint function” rather than language that could be read as treating or curing a diagnosed disease — the same disease-claim line this site’s Editorial Policy holds every ingredient to, regardless of how solid the underlying evidence is.

What “collagen” actually means as an ingredient category

Collagen is the most abundant protein in the human body and the primary structural protein in skin, providing its mechanical integrity; skin’s own collagen synthesis is estimated to decline by roughly 1–1.5% per year with age, which is the plausible biological mechanism behind both the skin and joint marketing claims. But “collagen supplement” is not one standardized product: trials studied here used different collagen types (Type I is most common in skin studies; Type I/III and Type II both appear in joint studies), different animal or marine sources, different degrees of processing (hydrolyzed peptides being the most common, broken into smaller, more absorbable pieces, versus undenatured collagen used in some joint-specific trials), and different daily doses — commonly in the 2.5–10 g/day range for skin research, though the 2025 review noted many of its Asian-conducted trials used doses closer to 3 g/day, which industry critics flagged as potentially not representative of typical Western consumption patterns. This variability is a genuine limitation of drawing one clean verdict from “collagen” as a single category, and it cuts both ways: it means a single negative or positive trial doesn’t necessarily generalize to every product on shelves.

What we could not check

• We did not independently re-verify the specific data-transcription allegations (dose, duration, funding classification) that industry groups raised against the 2025 skin meta-analysis — we’re reporting that a documented, public dispute exists and summarizing both sides’ positions, not adjudicating whose figures are correct.
• We did not conduct our own funding-source breakdown of the 2024 joint-health meta-analysis the way Myung & Park did for skin — we don’t know whether the same industry-funding pattern would appear if someone ran that analysis on the joint literature, and we’re not presenting the joint evidence as immune to the same possible bias, only as not yet shown to have it.
• We did not independently assess any single branded collagen product’s type, source, dose, or third-party testing — this article covers collagen as a researched ingredient category, not a specific product.
• We did not evaluate collagen’s other marketed uses (hair, nail, gut, bone density beyond osteoarthritis specifically) in this article.

Our rating, and why

For skin hydration, elasticity, and wrinkles: Limited. The largest available meta-analysis found a positive pooled effect that disappeared under funding-source and study-quality subgroup analysis — a documented pattern, not a hypothetical concern — even though the review’s methodology has been publicly and specifically disputed by industry groups. Limited reflects a real, unresolved dispute over a genuinely thin independent evidence base, not a dismissal of collagen outright.

For joint pain and function in osteoarthritis: Moderate. A larger, more methodologically rigorous meta-analysis found small-to-moderate, statistically credible benefits with a good safety profile, and — as far as we could determine — without the same funding-outcome pattern documented in the skin literature. Still Moderate rather than Strong because effect sizes are modest and most trials are short-term.

Sources

1. Myung SK, Park Y. Effects of Collagen Supplements on Skin Aging: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. The American Journal of Medicine. 2025. doi: 10.1016/j.amjmed.2025.04.034. https://www.amjmed.com/article/S0002-9343(25)00283-9/abstract
2. Stern C. Industry pushes back on meta-analysis concluding collagen supplements show no benefit for skin aging. NutraIngredients. Published 2025-08-26. https://www.nutraingredients.com/Article/2025/08/26/industry-reacts-to-meta-analysis-concluding-collagen-supplements-show-no-proven-benefit-for-skin-aging/ (secondary source — summarizes CSA and GROW industry responses to source 1; underlying trial-level dispute not independently re-verified)
3. Liang CW, et al. Efficacy and safety of collagen derivatives for osteoarthritis: A trial sequential meta-analysis. Osteoarthritis and Cartilage. 2024 May. PMID: 38218227. https://www.oarsijournal.com/article/S1063-4584(24)00004-9/fulltext (read via abstract and secondary summaries)