Polyphenols (resveratrol, quercetin, curcumin): what the evidence really shows

Polyphenols are a huge family of plant compounds, including flavonoids, the stilbene resveratrol and the curcuminoids in turmeric, and diets rich in them track with better heart and metabolic health in large populations. The honest catch is that this evidence is strongest for whole foods: berries, cocoa, tea, coffee, olive oil and herbs. Once you move to isolated supplements, the picture gets thin, because most polyphenols are poorly absorbed and the dazzling test-tube data have largely failed to translate into human benefit. The bottom line is food first, supplements a distant and selective second.

What polyphenols actually are

Polyphenols are not one substance but a sprawling category of several thousand compounds that plants make, largely to defend themselves against sunlight, pests and microbes. They share a common chemical feature, multiple phenol rings, and that is roughly where the family resemblance ends. The colour of a blueberry, the bitterness of green tea, the pungency of extra-virgin olive oil and the gold of turmeric are all polyphenols at work.

It helps to know the main branches, because marketing tends to talk about one compound as if it stood in for the whole family:

• Flavonoids, by far the largest group. This includes the flavanols in cocoa and tea (such as epicatechin), the flavonols like quercetin in onions and apples, the anthocyanins that make berries dark, and the isoflavones in soya.
• Phenolic acids, such as the chlorogenic acid that makes coffee the single biggest polyphenol source in many Western diets.
• Stilbenes, a small group whose famous member is resveratrol, found in grape skins, red wine and peanuts.
• Curcuminoids, the yellow pigments in turmeric, of which curcumin is the most studied.
• Lignans and others, including the compounds in flaxseed and the secoiridoids unique to olive oil.

The richest dietary sources, mapped systematically by Joanne Pérez-Jiménez, Augustin Scalbert and colleagues using the Phenol-Explorer database in 2010, turned out to be spices and dried herbs, cocoa, dark berries, some seeds and nuts, and a handful of vegetables.¹⁰ That single finding quietly anticipates the rest of this article: the action is in food.

The test-tube versus human gap

Polyphenols are one of the clearest examples in nutrition of a gap between laboratory excitement and human reality. In a dish of cells or in a mouse, many look spectacular: they mop up free radicals, dampen inflammatory signalling, switch on protective genes and slow cancer cells. Tens of thousands of papers describe these effects, and they are real, on the bench.

The problem is that a cell in a dish is bathed directly in a high concentration of the pure compound; a person who eats turmeric or swallows a capsule is not. By the time a polyphenol has survived your stomach, gut wall and liver, the amount reaching your tissues is often a tiny fraction of what worked in the lab, and frequently below the level at which anything happened at all. This is why so many human supplement trials disappoint after the preclinical fanfare. The mechanism was not fake; the dose simply never arrives.

The bioavailability problem

If one technical fact explains the whole supplement story, it is bioavailability: how much of a swallowed compound actually reaches your bloodstream intact. For the headline polyphenols, the answer is dispiriting.

Curcumin is the textbook case. On its own it is poorly soluble, badly absorbed, and what does get in is rapidly attached to other molecules (glucuronidated) by the gut and liver and flushed out, so ordinary turmeric leaves almost no free curcumin circulating. The famous workaround came from a 1998 study by Shoba and colleagues in Planta Medica: adding just 20 mg of piperine, the active alkaloid in black pepper, raised curcumin blood levels by about 2,000% by blocking the enzymes that clear it.⁵ This is why supplements pair the two, and why others use liposomal, nanoparticle or phospholipid (such as Meriva) formulations that lift absorption several-fold.

Resveratrol has the same flaw: absorbed reasonably well, then almost instantly metabolised, so free-compound levels are very low and short-lived. The doses that extended life in yeast and mice would, scaled to a human, mean grams per day or implausible quantities of red wine.

The formulation cleverness is double-edged. It can get more compound into you, but it also means the cheap capsule and the expensive "high-absorption" version are not interchangeable, trials run on specific branded formulations may not match what is on the shelf, and, as we will see, more absorption is not automatically safer.

Where the human evidence is least bad

Strip away the hype and a few areas hold up to scrutiny in actual people. None is miraculous, but they are honest.

Curcumin for osteoarthritis. This is the supplement's best card. A 2016 systematic review by James Daily, Sunmin Park and colleagues in the Journal of Medicinal Food, pooling 16 randomised trials in around 1,810 adults, found turmeric extracts reduced knee pain (effect size about -0.82) and improved function comparably with anti-inflammatory drugs; later meta-analyses of curcuminoids alone agree.³ The caveats are large: trials are small, follow-up short, blinding hard with a yellow pigment, and many are funded by supplement makers. Promising and worth a careful try for some, but not settled.

Cocoa flavanols and blood pressure. The 2017 Cochrane review led by Karin Ried, covering 35 trials and about 1,800 mostly healthy adults, found a small but significant fall: around 1.8 mmHg systolic on average, rising to roughly 4 mmHg in people already hypertensive, with no real effect in those with normal pressure.¹ Real, but modest, and it comes with the sugar and calories of chocolate.

Quercetin and blood pressure. A 2016 meta-analysis by Maria-Corina Serban and colleagues in the Journal of the American Heart Association (7 trials, 587 people) found quercetin cut systolic pressure by about 3 mmHg, but only clearly at doses of 500 mg per day or more, well above what food provides.⁴

Dietary polyphenols overall. In observational cohorts, eating more polyphenols tracks with fewer heart attacks and lower mortality; in the Spanish PREDIMED study, higher total intake was linked to reduced cardiovascular events and death.² This is the most encouraging evidence, but it is observational: people who eat more berries, olive oil and vegetables differ in many ways, so causation is not proven. For how to read this kind of nutrition evidence, see our insights on weighing health claims.

The resveratrol overreach

No polyphenol story has been oversold quite like resveratrol. In the mid-2000s, work suggesting it activated sirtuins (the SIRT1 enzyme linked to the lifespan benefits of calorie restriction) and extended life in yeast, worms and obese mice lit up the longevity field and the "French paradox" red-wine narrative, and a biotech company was bought for hundreds of millions of dollars on the promise.

The translation to humans simply did not happen. Trials have measured intermediate markers (blood sugar, inflammation, blood vessel function) with inconsistent, mostly unimpressive results, and no human trial has ever used lifespan as an endpoint.⁷ The mechanism wobbled too: it was seriously debated whether resveratrol activates SIRT1 directly or only through an artefact of how it was tested. Then in 2012 the University of Connecticut found one of its prominent resveratrol researchers, Dipak Das, responsible for 145 counts of data fabrication and falsification, leading to more than 20 retractions.⁸ His work was not the whole field, but the episode crystallised how far the hype had outrun the data.

The calm conclusion: resveratrol is an interesting molecule that did not deliver. Drinking red wine "for the resveratrol" makes no sense, since you would need dangerous amounts to approach studied doses, and alcohol itself raises cancer and cardiovascular risk. This is a clean illustration of why we keep returning to evidence over mechanism.

Your gut microbiome decides

One of the more genuinely exciting, and humbling, recent findings is that whether polyphenols help you may depend less on the dose and more on which bacteria live in your gut. Many polyphenols are barely absorbed in their parent form; instead they travel to the colon, where resident microbes break them into smaller compounds that may be the truly active agents. Crucially, not everyone carries the right microbes.

Two examples make the point. The ellagitannins in pomegranate, walnuts and berries are converted by gut bacteria such as Gordonibacter into urolithins, with urolithin A attracting interest for mitochondrial and muscle health. But people fall into distinct "metabotypes": some produce urolithin A well, some make other forms, and a sizeable group produce essentially none.⁹ Likewise the soya isoflavone daidzein is converted to equol, thought to drive several soya benefits, only in people (more common in parts of Asia) who carry equol-producing bacteria.⁹

This variation may explain much of the maddening inconsistency in polyphenol trials: a study can look positive or null depending purely on how many "producers" were enrolled. It is also a reminder that your gut microbiome is a hidden variable in nutrition, and that a supplement that works for a friend may do nothing for you.

Safety and interactions

Polyphenols from food are about as safe as eating gets. Concentrated supplements are a different matter, and a few signals deserve respect.

Quercetin. Generally well tolerated up to around 1,000 mg a day for a few months in trials, with mostly minor effects such as headache or tingling.⁴ Long-term high-dose safety is not well characterised, and it can theoretically interact with some medicines (it inhibits drug-metabolising enzymes), so it is not a free pass.

Interactions more broadly. Because piperine and high-dose polyphenols can inhibit the liver and gut enzymes that process medicines, concentrated supplements can in principle raise or lower blood levels of prescription drugs, and curcumin can have mild blood-thinning effects that matter if you take anticoagulants or are due surgery. If you take regular medication, treat any polyphenol supplement as something to clear with a pharmacist, not a harmless food. Our stack builder flags this kind of interaction risk before you commit.

The food-first bottom line

Pulling it together: the population evidence that polyphenols are good for you is reasonably strong, but it points at foods, not pills. Whole foods deliver polyphenols in a natural matrix, in sensible amounts, alongside fibre and other nutrients, and without the bioavailability and liver-safety headaches of concentrated extracts. For almost everyone, eating the rainbow beats buying the bottle.

The reframe worth keeping is simple. Polyphenols are a genuine reason to eat more plants, tea, coffee and good olive oil. They are a poor reason to spend money on capsules that the science, once it reaches actual humans, mostly cannot support.

This article is educational and does not constitute medical advice, diagnosis, or a treatment recommendation. Medication uses described as “off-label” are not licensed for that purpose in the UK and should only be considered under qualified clinical supervision. Always speak to your GP, pharmacist, or a registered specialist before starting, stopping, or changing any treatment. If you have severe or alarm symptoms - unintentional weight loss, blood in your stool, difficulty swallowing, persistent vomiting, a fever, or severe pain - seek urgent medical care.