Which is better, peptides or hyaluronic acid?

They answer different questions, and on the trial record neither dominates the other. In one identical query shape, topical peptides have 8 randomized trials on wrinkles or photoageing and hyaluronic acid has 10 — effectively level, and both small. On hydration, hyaluronic acid leads 11 to 6, which is its own claim territory. No trial has ever compared the two topically on a face, so anyone saying one wins is inferring it.

Do they do the same thing?

No. Hyaluronic acid is a glycosaminoglycan that holds water, and the claim made for it is hydration and plumping — a physical effect. Peptides are sold on signalling claims: fragments intended to prompt the skin to behave differently. One is a humectant, the other is an instruction. That is why they are routinely used together rather than instead of each other, and why a comparison between them is less useful than it sounds.

Why does hyaluronic acid seem so well-proven?

Because of the injections. Searching hyaluronic acid as a filler on wrinkle endpoints returns 127 randomized controlled trials, against 10 for the topical form on the same endpoints. Dermal fillers are one of the most-studied aesthetic interventions there is, and none of that evidence transfers to a serum: a filler places the molecule in the dermis, while a serum leaves it on the surface, which is a different intervention with a different mechanism.

Is the retinoid number the real story?

It is the number that puts the other two in proportion. In the same query shape, retinoids return 32 randomized trials on wrinkles and photoageing — roughly three times either ingredient here. Our main skin entry reaches the same conclusion by a different route. A shopper choosing between a peptide serum and a hyaluronic acid serum is choosing between two small evidence bases while a larger one sits beside them.

Has anyone put both on the same skin problem?

Once, and not in the way a serum shopper would hope. A 2024 study treated 30 patients' stretch marks with microneedling on both sides of the abdomen, then applied hyaluronic acid to one side and topical insulin — a peptide hormone of 51 amino acids — to the other. Excellent improvement was recorded in 23.3% of patients for hyaluronic acid and 30.0% for insulin, with marked improvement in 30.0% and 20.0%. Both went in through needling channels, so neither result speaks to intact skin.

Why do the numbers here differ from the census on your skin page?

Because the query shape differs, and that is the point. This page uses one shape applied identically to three ingredients so that the comparison between them is fair. The census on our skin page used a different shape, dated earlier, and reported 9 trials for topical peptides against 55 for retinoids. Both sets are reproducible from the queries printed beside them; neither set should be mixed with the other. Re-running the older retinoid shape today, without the topical restriction, returns 46.

Peptides vs Hyaluronic Acid: One Query Shape, Three Ingredients, and a Surprise in the Numbers

Run the identical trial search for peptides, hyaluronic acid and retinoids and the two ingredients people compare come out level — 8 trials against 10. Hyaluronic acid's real evidence base is 127 trials of injecting it, which is a different product entirely.

"Peptides or hyaluronic acid" is one of the most-asked questions in a skincare aisle, and it has an unusual answer: on the published trial record, the two are level, and both are smaller than people assume.

The comparison is only worth anything if the search is identical for each ingredient, so that is how we ran it. One shape, three ingredients, one day — 2026-09-15.

The matched census

Query shape: topical in title or abstract, plus the ingredient, plus wrinkle* or photoaging, restricted to the randomized controlled trial publication type.

Ingredient	Randomized trials, wrinkles/photoageing
Retinoids (retinol, retinoid, tretinoin)	32
Hyaluronic acid / hyaluronan	10
Peptides	8

Change only the endpoint terms to hydration or moisturiz*:

Ingredient	Randomized trials, hydration
Hyaluronic acid / hyaluronan	11
Peptides	6

Two readings follow, and they are both modest.

On ageing endpoints, peptides and hyaluronic acid are within two trials of one another. Neither has a commanding record; the ingredient that does is the one neither of them is usually compared against.

On hydration, hyaluronic acid is ahead, which is what you would expect from a humectant tested for the thing it does. This is the cleanest use of the data: each ingredient looks best on the endpoint its mechanism predicts.

The 127 trials that are not about serums

Hyaluronic acid's reputation does not come from that column of 10. Run the same restriction on hyaluronic acid as a filler or injectable, on wrinkle and nasolabial endpoints, and the count is 127 randomized controlled trials.

That is more than twelve times the topical record, and it is a different product. A filler is placed in the dermis by needle, in a volume a shopper can see; a serum leaves the same molecule on the surface of the stratum corneum. The mechanism of the first is physical volume; the mechanism of the second is water-binding at the top of the skin. None of the filler evidence transfers, and a great deal of the confident language around topical hyaluronic acid borrows its authority from trials of injecting it.

The same caution applies to peptides in the other direction, which is why our microneedling entry separates what a needle does from what a molecule does.

Nobody has run the head-to-head

The obvious study — a peptide serum against a hyaluronic acid serum, same base, same faces — does not exist. Six records carry both ingredients and the randomized-trial type, and reading them shows what they are:

Four are oral supplement trials of collagen peptides or a nutricosmeceutical, in which hyaluronic acid is a co-ingredient of the drink or a skin component being measured. Swallowing collagen peptides is a separate subject, covered on our collagen entry.
One is a caviar-extract supplement trial.
One is topical, and it is instructive: a 2020 paper describing a post-injection serum in which peptides are intended to stimulate collagen, elastin and hyaluronic acid. Hyaluronic acid there is not the comparator — it is one of the things the peptides are claimed to raise. The measured endpoint was bruising, where the product showed 73% less bruise colour intensity at day 2–3 than a bland moisturizer, and the authors are associated with the manufacturer. The ingredient itself has its own entry here.

The nearest thing to a head-to-head is not a face trial at all. In 2024, 30 patients had their abdominal stretch marks microneedled on both sides, then hyaluronic acid applied to one side and topical insulin — a 51-amino-acid peptide hormone — to the other. Excellent improvement was recorded in 23.3% of patients for hyaluronic acid and 30.0% for insulin; marked improvement in 30.0% and 20.0%. Both were delivered through needling channels, and the results are close enough that the study separates neither ingredient from the needling. We cover it in full on the stretch-marks entry.

Why the two are usually sold together

The mechanisms do not compete.

Hyaluronic acid is a glycosaminoglycan — a long sugar chain that binds water. Applied topically, its effect is to hold moisture at the surface, which plumps fine lines temporarily and improves the feel and look of skin while it is there. It is a physical effect with an immediate, measurable endpoint: corneometry.

A cosmetic peptide is a short chain of amino acids sold on a signalling claim: that a fragment resembling part of a larger protein prompts the skin to make more collagen, or blocks a neuromuscular step, or carries copper. Those are slower, indirect claims, and they need instruments and months to test — which is why the trials are few and why so many of them test a finished multi-active product rather than the peptide alone.

An ingredient that holds water and an ingredient that claims to send a message are not rivals, and a routine can hold both. What neither deserves is the other's evidence.

A note on comparing counts

The numbers above are not the same as the census on our main skin entry, which reported 9 randomized trials for topical peptides against 55 for retinoids in a search run on 2026-09-06. That shape differed, and re-running its retinoid version today without the topical restriction returns 46.

That is not a contradiction; it is the reason both pages print their queries. A trial count is a function of its query, and the only fair comparison is one shape applied to every ingredient in the same session. Mixing shapes — quoting one ingredient's broad count against another's narrow one — is the single easiest way to make either ingredient look like the winner, and it is what most comparison articles do without saying so.

Sources

Widgerow AD, Jacob C, Palm MD, Garruto JA, Bell M. Developing a topical adjunct to injectable procedures. J Drugs Dermatol 2020;19(4):398–404. PMID 32272517.
Ebrahim MRY, Darwish HMA, Galal SA. Efficacy of dermapen combined with topical insulin versus dermapen combined with topical hyaluronic acid in treatment of stretch marks: comparative study. Arch Dermatol Res 2024;317(1):54. PMID 39601879.
PubMed counts, all run 2026-09-15 via the NCBI E-utilities esearch endpoint. Ageing shape: topical[tiab] AND <ingredient> AND (wrinkle*[tiab] OR photoaging[tiab]) AND "randomized controlled trial"[pt] — peptides 8, hyaluronic acid/hyaluronan 10, retinoids 32. Hydration shape: the same with (hydration[tiab] OR moisturiz*[tiab] OR moisturis*[tiab]) — hyaluronic acid 11, peptides 6. Injectable shape: ("hyaluronic acid"[tiab] OR hyaluronan[tiab]) AND (filler*[tiab] OR inject*[tiab]) AND (wrinkle*[tiab] OR nasolabial[tiab]) AND "randomized controlled trial"[pt] — 127. Both-ingredient set: peptide*[tiab] AND ("hyaluronic acid"[tiab] OR hyaluronan[tiab]) AND skin[tiab] AND "randomized controlled trial"[pt] — 6. PubMed.