Collagen style pills only slightly reduced wrinkles in 19 trials, review finds

Drinks, powders and pills that promise smoother skin are easy to find. Many contain collagen, the protein that keeps skin firm, or other short protein pieces called peptides. A new review asked what the trials behind products like these show when their results are added together [1].

Across 19 trials with 1,341 adults, the products made only a small difference to wrinkles. The added up gap was 0.27 on a wrinkle score. That result only just passed the usual test for ruling out luck [1].

The main caveat comes early. The trials scored wrinkles in different ways, and in many of them, people may have known which group they were in. Both make the added up number less trustworthy [1].

The review was led by Houriah Nukaly of the McGill University Health Centre in Montreal and published in Frontiers in Medicine in March 2026. The team searched three large libraries of research papers and wrote down its plan ahead of time in a public list called PROSPERO. The authors say no one paid for the review, and that they have no ties to companies selling these products [1].

In each trial, chance decided, like a coin flip, who got the product and who got a dummy version with nothing active in it. People were about 50 years old on average. Seventeen of the trials tested something swallowed. Only two tested a cream, with 105 people between them [1].

The reviewers summed up their findings this way:

Peptides, particularly oral formulations, significantly improved hydration and brightness, with a modest pooled effect on wrinkle reduction (MD = 0.27, p = 0.04).

In plain words, skin moisture and brightness moved more clearly, and wrinkles moved only a little. MD stands for the average gap between groups. The true wrinkle gap could sit anywhere from 0.01, barely above zero, to 0.52 [1].

Why is the wrinkle number so hard to read?

The number 0.27 has no unit, like inches or pounds. Some trials measured how deep wrinkles were with a tool. Others had people score photos. Adding those together is a bit like averaging one temperature in Fahrenheit with another in Celsius [1].

The trials also differed a lot from each other. In one comparison, the review found that 96 percent of the spread between studies came from real differences between them rather than chance. In other words, the trials were not really measuring the same thing [1].

Then there is the question of who knew what. In a fair trial, the people taking part and the people scoring the results do not know who got the real product. That matters a lot when the result is how skin looks. A scorer who knows may score a little kinder without meaning to. The review said it was unclear whether several trials kept people from knowing, and one trial had a high risk of this problem [1].

Most of the wrinkle effect came from the products that were swallowed. The two cream trials showed a smaller effect that did not pass the usual test. The paper also does not quite agree with itself on one figure. A summary box gives the swallowed products a gap of 0.35, while the abstract and results section say 1.5 for what seems to be the same comparison [1].

Who runs these trials? Newer studies outside the review show a pattern. A 2024 study followed 135 women aged 45 to 65 for three months on a supplement with collagen peptides. Their wrinkle scores dropped from 5.9 to 5.0. But everyone knew what they were taking, there was no comparison group, and all five authors work for the company that makes the product [2].

A 2025 trial of a product made from pig placenta was built more carefully, with 90 people and a dummy group, and nobody knew who got what. Two of its authors work for the company that sells the ingredient [3]. A 2026 paper says the company that paid for it also helped plan the study, study the data and write the article [4].

None of that makes a result false. A paper that says who paid is doing the right thing. But it does mean the finding needs more checking by people with no stake in it [1][2][3][4].

This review does not show which product does what, and it says almost nothing about creams. It does not show that any one product works. What it does show is that, added together, these trials point to a small change in wrinkles, mostly from swallowed products, resting on studies that need better design. The reviewers themselves call for larger trials that all measure skin the same way [1].