HPR: sold as a 'gentle retinoid', and we could not find a human trial of it alone

A 2026 study in J Cosmet Dermatol.

What was used is a serum containing retinol plus HPR plus peptides plus silybin. The clinical part is an 8-week study in middle-aged Chinese women with mild photoageing.

Reported result: the serum significantly improved wrinkles, elasticity, hydration, barrier function and pigmentation, with excellent tolerability.

Four things about how to read it.

First, HPR was not measured alone. A four-active serum was used. And one of those four, retinol, is among the better-evidenced ingredients on this site. If this serum improved wrinkles, retinol having done the work is the first possibility that comes to mind. This design cannot separate them.

The paper itself states its aim as characterising a combination — that HPR's combined use with retinol remains poorly characterised. So what this study asked about is the combination, not HPR.

Second, the abstract gives no participant count. This site does not invent numbers it does not have, which is why the participant field on the badge above is empty.

Third, the abstract does not mention a control. It says only that efficacy was assessed in an 8-week study. Whether it was compared against a placebo serum, whether people were randomised, or whether before was compared with after, we cannot tell. So the comparator field on the badge reads 'not stated'.

As we write repeatedly on other pages, comparing before with after and no control means time, expectation and the season are all counted as the product's work.

Fourth, every author works at one cosmetics company's research and innovation centre. And the paper declares no conflicts of interest. We mark the affiliation itself on the badge as 'industry funding'.

Hence 'not established'. Not that the ingredient does nothing — that this material lets us say nothing about HPR.

The mechanistic part of the same study. It did not come from people, so we give it its own entry.

• transcriptomic profiling in a keratinocyte-fibroblast co-culture
• protein expression analysis in UV-exposed ex vivo skin models (pieces of skin removed from the body)

Reported results:

• retinol plus HPR activated TGF-β/Smad signalling more than retinol alone, and raised extracellular matrix gene expression
• adding silybin further promoted collagen and elastin synthesis

Why this has its own entry.

This is the only part that directly compares 'retinol plus HPR' against 'retinol alone'. The clinical part contains no such comparison.

But this is not people.

Signalling being more active in cells in a dish and in a removed piece of skin sits several steps away from fewer wrinkles on the face of someone who applied it. In between lie getting through skin, the concentration actually used, individual variation, and time.

What this site guards against most is using laboratory results as if they were human effects. So we write the kind of material onto the badge as it is — the badge above reads 'ex vivo'.

This entry is 'not established' too. Readable as mechanism; not readable as evidence about people.