Does Drinking Collagen Actually Do Anything?

A reader wrote in last week with a question, not a story with an ending yet. She'd seen the same ad on Instagram for two weeks straight, a jar of collagen powder promising firmer, more hydrated skin by the end of the month, and she wanted to know if she should just buy it already or if this was one more thing marketing had oversold. I didn't have a confident answer sitting around, so I went and read the actual research instead of guessing along with her.

This is a different question from the two collagen questions already answered in this journal. It isn't the topical collagen cream question. Cream can't cross into the layer of skin where collagen is made, and that one's settled. It also isn't the "you lose 30 percent of your skin's collagen in five years" number, which turned out to be inflated from the original data. This is about drinking or eating collagen: powders, gummies, the marine and bovine peptide blends sold everywhere from the pharmacy counter to a wellness influencer's link in bio. Does anything happen when it goes in through your mouth instead of onto your skin?

What the researchers actually looked at

Pu and colleagues pulled together every randomized controlled trial they could find on oral collagen and skin aging and ran a systematic review and meta-analysis, published in Nutrients in 2023. Twenty-six trials, 1,721 participants combined, mostly healthy adult women in their thirties to sixties, taking collagen by mouth for anywhere from 2 to 12 weeks. That's enough studies pooled together that one bad trial can't quietly decide the whole answer.

What it found

The pooled results leaned positive on two measures. Skin hydration improved, effect size 0.63 (95% confidence interval 0.38 to 0.88). Elasticity improved more, effect size 0.72 (95% confidence interval 0.40 to 1.03). Trials that ran longer than eight weeks tended to show a bigger effect than the shorter ones. Neither confidence interval crosses zero. Something in that pooled data moved in the direction the collagen industry has been claiming for years.

In the language researchers use for this kind of number, 0.6 to 0.7 sits in the moderate range, not the small range and not the large one. It's not a rounding error. It's also not the kind of result where every woman who drinks the stuff wakes up with visibly different skin. It's the kind of result that shows up when you average over hundreds of people and hundreds of measurements, some of whom likely felt no change at all.

The honest caveats

Here's where I slow down, because the same paper reporting those numbers also reports its own weaknesses, in the same breath. Thirteen of the 26 trials had attrition bias: enough participants dropped out partway through that the final results may not reflect what would have happened if everyone had stayed. Seven of the 26 had blinding problems, which means participants or researchers may have known who was taking the real collagen and who wasn't. That can quietly shape something as soft as "does your skin feel more hydrated." The authors say outright that the field needs larger, better-designed trials before anyone treats this as settled.

There's a second thing worth naming, and the paper doesn't hide it either: a large share of oral collagen research is funded by the companies that manufacture and sell it. That doesn't make every finding wrong, but the incentive to find a positive result runs in one direction, worth knowing before you weigh the numbers, the same way you'd want to know who paid for a taste test before trusting the ranking.

What I told the reader who wrote in

I wrote back with the honest, slightly unsatisfying version, the kind that doesn't fit neatly into an Instagram caption. The data isn't zero. It also isn't the kind I'd build a whole skincare budget around. Two dozen trials pooled together, with real gaps in how many participants stuck around and who funded the work, land somewhere in the middle of what I've been reading this whole month. Weaker than daily sunscreen, which held up in a long-running randomized trial with essentially no downside (I wrote up the full numbers on that one separately). Stronger than the collagen cream sitting in a lot of medicine cabinets, which can't physically reach the layer of skin it's supposed to rebuild.

If she already drinks the stuff and likes it, stirred into hot water or the gummies that taste like dessert, nothing here says it's doing harm, and there's a small positive signal that it might be doing something. If she was hoping it would be the one thing that changes how her skin looks and feels, I'd rather she know now that sunscreen is still standing ahead of it, evidence-wise. I'm not trying to talk her out of the jar. I just want her spending her money with her eyes open about what it can and can't do.

Not a yes, not a no. Just where the thing actually sits on the ladder, so the next ad that promises firmer skin by next month isn't the only voice in the room.

I've been putting this kind of thing, what's solid, what's shaky, what's worth paying for, and what isn't, into one place with the studies linked next to each claim. It's called Skin After 40, a printable workbook, on Etsy for $17. If you'd rather have the sourced version in one document than piece it together from a month of journal entries, it's there.