A brand-new analysis just made a shocking claim:
“There is no clinical evidence to support the use of collagen supplements to prevent or treat skin aging” [1].
Table of Contents
The Study
Why is this new claim about collagen supplements shocking?
Well, it directly contradicts the findings of a previous meta-analysis. That one included 26 randomized controlled trials and was published in 2023. It examined the effects of collagen supplements on skin hydration and elasticity—two key aspects of skin health that decline with age [2].
The statistical analysis showed that collagen peptide supplements significantly improved both metrics:
“The overall pooled effect size of 0.63 (95% CI 0.38, 0.88) indicated that HC supplementation significantly improved skin hydration (z = 4.94, p < 0.00001)...
...HC supplementation significantly improved skin elasticity (z = 4.49, p < 0.00001) compared with the placebo group at a pooled effect size of 0.72 (95% CI 0.40, 1.03)” [2].
And this latest meta-analysis found the same thing when taking all included studies together. Collagen peptide supplements significantly improved skin hydration and elasticity. They also improved wrinkles [1].
But then the authors conducted some subgroup analyses. And that’s where they ultimately reached a very different conclusion.
One subgroup they considered was just those studies that weren’t funded by pharmaceutical companies. The logic here is that these would be less likely to give a biased result. So how do the numbers look for this subgroup?
Now there’s no statistically significant impact on hydration, elasticity, or wrinkles [1].
A similar thing happened when they looked at just high-quality studies. And there were also no statistically significant benefits when considering only studies that were both high quality and not funded by pharmaceutical companies [1].
This looks terrible. Because it sounds like this is what the researchers discovered: Low-quality studies funded by pharmaceutical companies find benefits for collagen peptide supplements. But high-quality, non-funded studies don’t. That raises serious doubts about whether collagen peptides really have the benefits we thought.
But the story isn’t quite so simple.
Digging Deeper
Remember, the meta-analysis gives us pooled results when taking all studies together. If we look at individual studies that were identified by this meta-analysis as high quality, we get a surprisingly different picture.
Let me unpack this.
The meta-analysis uses two common measures of study quality. There are 9 studies total that get the highest marks with both measures. We just saw that when the authors looked at just these 9 studies, they concluded there was no benefit overall.
So you might expect that none—or at most 1 or 2—of these 9 studies showed benefits. But here’s the surprise: out of these 9, only 1 found no benefit. Another one found mixed results. The other 7 all showed positive impacts from collagen supplements.
And what about when we zero in on the highest-quality studies that aren’t funded by pharmaceutical companies?
There are just 5 of these. Of those, four found positive results.
In other words, almost all of the best studies found positive results. So how, then, do the authors wind up with such a deflating conclusion?
It has to do with how the statistics behind a meta-analysis work. They can actually lead to a result that’s misleading. Let me explain the two reasons why.
1. Standard Mean Differences
There’s the way researchers pool the impact of collagen peptides. There’s a challenge when trying to add together the results of several different studies. That’s because they don’t measure precisely the same things in the same ways.
For instance, one study might describe improvements in skin moisture by measuring how much water is lost through the skin. Another might use electric current to measure moisture content within the skin. Both of these metrics are relevant to improved moisture. But we can’t directly combine them.
So what researchers do is convert specific measures into what’s called a standard mean difference (SMD). It’s a way to capture the size of an effect on a standard scale. Once that’s done, we can compare the effect sizes of different studies and also pool them to calculate an overall effect size.
The challenge is that when we standardize results this way—especially when the studies themselves are very different—we can lose important context. We can end up with a pooled result that seems to contradict the individual findings.
2. Confidence Intervals
The second part of the process of pooling that can yield a misleading outcome relates to confidence intervals. These are useful in statistics because they tell us how confident we are in the results.
When researchers combine studies where the results are fairly different, the final confidence interval gets larger. It’s possible to have a situation where all studies show a positive result, but the size of those results varies a lot. That can make us say, in effect, “The study results are so different that we don’t have confidence we know what’s going on here.”
And here’s how this issue shows up in the new study.
For skin hydration, the SMD is 0.33, and the confidence interval is -0.06 to 0.73. That confidence interval means we feel pretty sure the true impact is somewhere between these numbers. But the fact that it drops below zero is why the result isn’t statistically significant—it means it’s possible the true impact is no impact [1].
This is technically correct from a statistical point of view. But almost the whole range of possible effect sizes is positive. That means a moderate positive effect is entirely plausible.
It’s the same story for elasticity when looking at just the high-quality studies. The confidence interval just barely crosses 0, so the authors confidently declare no significant effect [1].
This misses the forest for the trees—especially when the vast majority of high-quality studies, before their measurements were converted into standard mean differences, showed a benefit.
We have an additional reason to think positive effects are likely.
That’s because studies give us a mechanism for how collagen peptides work to affect the skin.
For example, one study investigated how fibroblast cells grown in the lab respond to collagen peptides. Fibroblast cells make collagen and elastin. What researchers saw under the microscope confirmed that skin improvements seen with collagen peptide supplements are driven by positive cellular changes.
The fibroblasts exposed to collagen peptides increased production of collagen, elastin, and another molecule called proteoglycan, which is crucial for skin hydration [3].
Okay, but how do we know that when we ingest collagen peptides, they reach our skin?
We have numerous human trials showing that collagen peptide supplements boost the levels of relevant peptides in our blood. As the blood circulates throughout the body, the peptides finally reach the skin.
A study using both mice and humans confirmed this. The collagen peptides we eat do indeed make their way to our skin [4].
But one big limitation of most collagen peptide studies is that they compare collagen peptides to a placebo rather than directly against a protein supplement.
Ideally, we’d want to match protein intake. That way, we know the effects on skin we find are due to collagen peptides—not just more protein intake.
Currently, there’s only one study that directly compares collagen peptides to protein supplements for skin. But the researchers were looking at wound healing, not skin aging. It examined burn patients. One group took protein. The other took a matching amount of collagen peptides.
The collagen peptide group experienced a significantly higher wound healing rate compared to the protein group [5].
While that’s encouraging, it’s high time for a study on skin aging to be done that directly compares collagen peptides to a matched protein supplement.
Implications
So let’s get practical. There are 3 takeaway points from this new meta-analysis.
1. Read the Full Study
It’s critical to read the full study and not just rely on the abstract.
2. Statistical Significance Isn’t Everything
We may need more study data to hit certain thresholds of statistical significance when focused on just the highest-quality studies. But even so, the combined analysis just barely touched the 0 mark. We may be missing the forest for the trees—especially when almost all the individual studies show positive results, including the highest-quality trials not funded by collagen peptide manufacturers.
And new studies continue to emerge. For example, a study published in April looked at the impact of collagen peptides on fibroblasts. Like many others, it found positive impacts on wrinkles, elasticity, and hydration [3].
3. Mechanism of Action
We not only see many high-quality studies showing that collagen peptides help—we also have experiments that help us understand why.
And finally, there are no known health risks associated with taking collagen peptides.
In light of these points, I personally plan to continue taking a collagen peptide supplement.
References
1. https://doi.org/10.1016/j.amjmed.2025.04.034
2. https://www.mdpi.com/2072-6643/15/9/2080
3. https://www.mdpi.com/2079-9284/12/2/79
