1. What polyphenols are, and why the category is a problem
Polyphenol is not an ingredient. It is a structural classification covering thousands of plant compounds, grouped by having multiple phenol units. Under that umbrella sit flavonoids, catechins found in tea, anthocyanins that colour berries, resveratrol from grape skins, curcuminoids from turmeric and many more. They differ from one another in absorption, metabolism and biological activity.
This creates a persistent reasoning failure that readers should learn to spot. Evidence generated on one compound is routinely used to sell another on the basis that both are polyphenols. That is roughly as sound as concluding that because one metal conducts electricity, another metal will cure your headache. The category tells you about chemistry, not about effect.
Extract variability compounds it. A botanical extract's composition depends on the plant variety, the growing conditions, the part of the plant used, the extraction solvent and the processing. Two products naming the same plant can differ substantially. Where a study tests a specific standardised extract, its findings describe that extract and not the category, and frequently not even the same plant sold by a different supplier.
2. The mechanism, described accurately
Ultraviolet radiation damages skin partly through direct effects on DNA and partly by generating reactive oxygen species, which damage cell components and drive inflammation. Antioxidant compounds can neutralise some of those reactive species. On paper, an antioxidant present in skin at the time of exposure could reduce part of the downstream damage.
That reasoning is sound as far as it goes, and it is important to see exactly how far it goes. It describes a possible contribution to the secondary, oxidative branch of ultraviolet damage. It does nothing about the direct photochemical damage to DNA. It requires the compound to be present in skin at a relevant concentration, which for dietary polyphenols means surviving digestion, metabolism and distribution, and which for most of them is poorly characterised. And a mechanism that operates after the exposure has occurred is categorically different from a filter that prevents the exposure.
3. What the human studies show, and what they measure
Human research does exist. The typical design supplements participants with a polyphenol rich food, drink or extract for a period of weeks, then measures how much ultraviolet is required to produce visible redness, sometimes with skin biopsies or biochemical markers alongside.
These studies are legitimate and their limitations are structural. They are small. They run over weeks against a process that unfolds over decades. The primary outcome is a threshold for visible redness, which is a real biological measure but is a long way from what a reader cares about, which is whether their skin ages more slowly or their cancer risk falls. And many are funded by food industry bodies with an interest in the ingredient.
Where an effect is reported, it is generally described as a modest shift in that redness threshold. Translated into practical terms, that is a small change in how quickly skin burns, not a protective shield. Marketing that converts it into a sun protection factor equivalent is doing something the underlying research does not support.
Everything else in this review is interesting. This is the part with consequences. No polyphenol, eaten or applied, has been shown to justify reducing sun protection. Ultraviolet exposure is an established cause of skin cancer. A reader who takes an antioxidant supplement and spends longer in the sun as a result has made themselves worse off.
4. The dietary argument, which is better than the supplement one
There is a version of this that we would defend. Diets high in vegetables, fruit, pulses, wholegrains, tea and olive oil are associated with better health outcomes across many measures, and those diets are naturally high in polyphenols. Eating that way is sensible for reasons that have nothing to do with skin.
What does not follow is that the polyphenols are the operative ingredient, or that extracting them into a capsule reproduces the benefit. Whole diets differ from isolated compounds in dose, in the matrix that affects absorption, in what else is present, and in what they displace from the plate. The repeated failure of isolated nutrient supplements to reproduce the benefits observed from whole diets is one of the most consistent findings in nutrition research, and it should temper enthusiasm for any capsule sold on the strength of a dietary pattern.
5. Topical botanicals
Applied polyphenols avoid the digestion problem and inherit a new set. Penetration through the barrier varies enormously between compounds. Many are unstable in formulation, oxidising in the presence of light and air in ways that are not always visible. And the extract variability problem is at its worst here, because cosmetic formulations rarely disclose which extract, at what standardisation, was used.
The comparison with topical vitamin C is instructive. Vitamin C is a single defined molecule with known chemistry and a modest human trial base, and we grade it Moderate. Botanical antioxidants are an undefined mixture with a thinner human trial base, and we grade them Limited. That gap is a description of the evidence, not a judgement about plants.
6. Our position
Graded Limited. The laboratory work is genuine and the mechanism is real. The human evidence is small, short and measures surrogate outcomes, and the category label conceals more variation than it reveals. Eat the diet, because the diet is a good idea for other reasons. Do not buy the capsule expecting protection, and under no circumstances treat any of this as a reason to use less sun protection.
For the strongest preventative evidence in skin health, and where it actually points, see vitamin D and immunity on the sunscreen argument, and how we grade evidence for what would move this grade.