Peptides, vitamin C and botanicals rest on small, short trials, often with no inert arm

The active ingredients on a skincare label — peptides, vitamin C, botanical extracts, niacinamide — are not tested the way drugs are tested. The trials behind them are typically small, run for four to twelve weeks, measure instrument readings rather than anything a person would notice, and frequently compare one active against another rather than against something inert. That design cannot separate a working ingredient from a moisturiser, and in several cases the published reviews say so.

Peptides: a review that contradicts itself

A systematic review and meta-analysis published in Frontiers in Medicine pooled 19 randomised controlled trials of oral and topical peptides covering 1,341 participants [s1]. That is a reasonable evidence base by cosmetic-science standards and a thin one by any other.

Its findings, as stated in its conclusion, were that peptides significantly improved hydration and brightness, with a modest pooled effect on wrinkle reduction — a mean difference of 0.27, p = 0.04 — and that the benefit was largely driven by oral polypeptides at a mean difference of 1.5, p = 0.01 [s1]. Effects on elasticity and density were inconsistent, and the products were well tolerated with minimal adverse events [s1].

A discrepancy in the source: the paper's own highlights report the wrinkle-reduction effect as a mean difference of 0.35 with p = 0.05, against the 0.27 and p = 0.04 given in the conclusion. Both numbers appear in the same abstract, describing the same outcome. The paper does not reconcile them. Readers should treat the wrinkle estimate as approximately a quarter to a third of a point on whichever scale was used, and note that a discrepancy of this kind in the most quoted number is itself a comment on the field's editorial standards.

The stronger, more consistent findings were on hydration (mean difference 5.80, p < 0.01) and brightness (2.40, p < 0.01) [s1]. Hydration is the outcome a plain moisturiser also improves, which is exactly why a vehicle-controlled design matters and why its frequent absence matters more.

Botanical extracts: significant, and then not

A meta-analysis of topical plant-based anti-ageing products searched four databases across January 2000 to December 2024 and found eight randomised controlled trials meeting inclusion criteria [s2]. Eight.

Those eight showed statistically significant improvements in skin hydration and skin elasticity and reductions in melanin and erythema against comparator products, with no significant difference in transepidermal water loss [s2].

The subgroup analysis is where the result comes apart. Improvement in overall skin elasticity was significant in short-term treatment of eight weeks or less, and not significant in treatment longer than eight weeks [s2]. Two of the more specific elasticity measures — net elasticity and the ratio of elastic recovery to total deformation — showed no significant difference at all [s2].

An effect that is present at eight weeks and absent thereafter, and that appears on a composite measure but not on its components, is the signature of a measurement artefact rather than a biological change. The review reports both without choosing, which is the correct thing to do and also the thing product marketing never does.

On safety the picture was unremarkable: six of the eight studies reported no adverse events, and the remaining two did not say [s2].

Vitamin C: a trial with no inert arm

A double-blind, split-face randomised trial compared L-ascorbic acid 15% serum with retinol 0.1% serum for skin ageing, assessing participants' cheeks with a modified dermoscopy photoaging scale and an imaging system [s3]. Both treatments produced significant improvement at weeks 2 and 4, and there was no statistically significant difference between them at either time point [s3]. Safety profiles were similar, and the authors conclude that L-ascorbic acid is as effective and safe as retinol for anti-ageing therapy [s3].

Two features of that design limit what it can establish. There is no vehicle or no-treatment arm, so the trial cannot distinguish a shared active effect from a shared effect of applying a serum to one cheek twice a day for four weeks. And four weeks is short even by the standards of this literature — the randomised tretinoin trials that anchor the photoaging field ran from three to 24 months.

The trial is not badly conducted. It is answering a narrower question than its conclusion implies: these two products performed similarly over a month, against each other.

What is known about topical antioxidants generally

The most useful independent assessment comes from an unlikely place — a CMAJ review of sunscreen. Discussing topical antioxidants including vitamin C, vitamin E, selenium and green tea polyphenols, it describes them as emerging potential photoprotective agents rather than established ones, and states that antioxidants cannot yet be stabilised within sunscreen formulations so as to remain biologically active [s4].

It goes further, and this is the sentence worth carrying out of the whole subject: studies have established that sunscreens claiming antioxidant activity have little to no actual antioxidant activity [s4]. That is a finding about product chemistry rather than about biology. An ingredient can be genuinely active in a laboratory and inert in the bottle it is sold in, and almost nothing on a cosmetic label distinguishes the two cases.

How to read a skincare study

Four questions cover most of it, and they are the questions these three reviews keep raising about their own included trials. Was there an inert comparator, or only another active? How long did it run, and did the effect survive past the first eight weeks? Was the outcome something a person experiences, or an instrument reading on a composite index? And is the effect size large enough to mean anything — a mean difference of 0.27 on a wrinkle scale is a real statistical finding and an open question as a visible one.

None of this establishes that these ingredients do nothing. It establishes that the studies capable of showing what they do have mostly not been run, and that the ones that have been run are small, short and often structurally unable to answer the question.

This article is informational and is not medical or cosmetic advice.

Sources

• [s1] Oral and topical peptides for skin aging: systematic review and meta-analysis of randomized controlled trials — Frontiers in Medicine, 17 March 2026. https://doi.org/10.3389/fmed.2026.1618306
• [s2] Efficacy and Safety of Topical Application of Plant-Based Products on Skin Aging in Healthy Individuals: A Systematic Review and Meta-Analysis of Randomized Controlled Trials — Journal of Cosmetic Dermatology, published online 9 December 2024. https://doi.org/10.1111/jocd.16710
• [s3] Effectiveness and safety of topical L-ascorbic acid 15% serum compared to retinol 0.1% serum for skin aging: a double-blind randomized clinical trial — Dermatology Reports, 28 May 2026. https://doi.org/10.4081/dr.2026.10293
• [s4] The efficacy and safety of sunscreen use for the prevention of skin cancer — CMAJ, 13 December 2020. https://doi.org/10.1503/cmaj.201085