Resveratrol Honestly: What Grapes Can and Cannot Do

Table of Contents

1. What Resveratrol Actually Is
2. How Much Is in a Grape — and a Glass of Wine
3. The Experiments That Started the Boom
4. The Dose Gap: Doing the Arithmetic
5. The Bioavailability Problem
6. The Sirtuin Question and the Assay Artefact
7. What Happened When Someone Measured Real People
8. The Research-Fraud Case in This Exact Field
9. The French Paradox Is a Hypothesis, Not a Fact
10. What the Human Trials Actually Show
11. What This Means for You
12. Key Research Papers
13. Connections
14. Featured Videos

What Resveratrol Actually Is

Resveratrol is a stilbene — chemically, 3,5,4'-trihydroxystilbene, two phenol rings joined by a short two-carbon bridge. Grapevines do not make it for our benefit. They make it as a phytoalexin: a defensive chemical the plant synthesises in a hurry when it is attacked by a fungus, wounded, or hit with ultraviolet light. That single fact explains most of the confusing variability in the numbers you will read. A vineyard with heavy fungal pressure, thin skins, cool damp weather and minimal fungicide produces grapes with far more resveratrol than a hot, dry, sprayed vineyard does. The compound is a stress response, not a stable nutrient like potassium or vitamin C.

It lives almost entirely in the skin, with a little in the seeds and essentially none in the pulp or juice. This is why the compound is associated with red wine rather than white: red winemaking ferments the juice in contact with the skins for days or weeks, extracting skin compounds into the wine, while white winemaking presses the skins off almost immediately. Table grapes eaten with the skin on deliver some; peeled grapes and clear grape juice deliver very little.

Resveratrol exists in two geometric forms, trans and cis. The trans form is the one that is biologically studied and the one meant whenever a label says "resveratrol". It is also the less stable of the two: light, heat and oxygen convert trans to cis, which is why storage and handling change the content of a bottle over time. Grapes are not even the richest dietary source — Japanese knotweed (Reynoutria japonica, formerly Polygonum cuspidatum) contains far more, which is why almost every resveratrol supplement on the market is extracted from knotweed rather than from grapes, whatever the grapes on the label might suggest.

Back to Table of Contents

How Much Is in a Grape — and a Glass of Wine

This is the number that decides everything else on this page, so it is worth being precise about how uncertain it is. A published review of resveratrol measurements across a large number of red wines found the content varies enormously by grape variety, region, vintage and winemaking method, with typical red wines clustering in the low milligrams per litre and outliers in both directions. Pinot Noir, grown in cool damp climates with thin-skinned fruit, tends to sit at the high end. Wines from hot dry regions sit at the low end. White wines and rosés, fermented with little or no skin contact, contain a small fraction of what reds do.

Fresh table grapes contain less again, on a per-serving basis, than a glass of red wine does — the fermentation and extended skin maceration of red winemaking concentrate skin compounds in a way that eating the fruit does not. Grape juice sits in between, depending on whether it was pressed with skins. Raisins, which are dried grapes with the skins intact, retain some but lose a fraction to the drying process.

The honest bottom line: the dietary intake of resveratrol from any realistic diet is measured in milligrams per day at the very most, and for most people in fractions of a milligram. Hold that figure in mind for the next two sections, because everything the supplement industry has built rests on numbers that are three or four orders of magnitude larger.

Back to Table of Contents

The Experiments That Started the Boom

Three papers built the resveratrol phenomenon, and all three are real, published, and worth understanding on their own terms.

The first was Jang and colleagues, in Science, 1997. They screened plant compounds for cancer-chemopreventive activity and found that resveratrol, isolated from grapes, inhibited events at all three classical stages of carcinogenesis in laboratory models — initiation, promotion and progression. It was an elegant piece of natural-product screening. It was also entirely cell-culture and mouse-skin work. Nobody in that paper claimed a human would get the same result from drinking wine.

The second was Howitz and colleagues, in Nature, 2003. They reported that a set of small molecules, resveratrol prominent among them, activated sirtuins — a family of enzymes then being proposed as the mechanism behind the life-extending effects of calorie restriction — and extended the lifespan of budding yeast. This is the paper that connected resveratrol to ageing, and it is where the phrase "calorie-restriction mimetic" entered popular circulation.

The third, and the one that broke into the newspapers, was Baur and colleagues, in Nature, 2006. Middle-aged mice were put on a high-calorie diet and given resveratrol. The treated mice survived significantly longer than untreated mice on the same diet; they had better insulin sensitivity, lower IGF-1, more mitochondria, better motor function, and gene-expression patterns that opposed the high-calorie diet across 144 of 153 significantly altered pathways. That is a striking result and the paper is not in dispute. Alongside it, Baur and Sinclair's review in Nature Reviews Drug Discovery, published the same year, laid out the case for resveratrol as a therapeutic agent and became one of the most-cited reviews in the field.

Notice what all three have in common. Yeast. Cultured cells. Mouse skin. Mice on a high-calorie diet, dosed by the experimenter. Not one of them is a study of humans eating grapes, and none of them claimed to be.

Back to Table of Contents

The Dose Gap: Doing the Arithmetic

Here is where the story turns, and it turns on arithmetic rather than on any new experiment.

The mice in the 2006 experiment were not sipping wine. They were given resveratrol as a measured supplement, at a dose per kilogram of body weight that is orders of magnitude beyond dietary exposure. When researchers have translated the animal doses used across the resveratrol literature into human-equivalent terms, the answer has repeatedly come back the same way: to reach them from wine you would have to drink a volume that would kill you from the alcohol long before the resveratrol did anything at all. Weiskirchen and Weiskirchen put exactly this question in the title of their 2016 review in Advances in Nutrition — "Resveratrol: How Much Wine Do You Have to Drink to Stay Healthy?" — and concluded that the concentrations shown to be active in laboratory systems are simply not achievable through diet.

There are two ways to scale an animal dose to a human, and both give an unreachable answer here. Straight body-weight scaling multiplies the mouse dose by human body weight and lands in the range of grams per day. Allometric scaling, which corrects for the faster metabolic rate of small animals and gives a much smaller number, still lands in the range of a hundred milligrams or more per day. Now divide either figure by the few milligrams of resveratrol in a litre of red wine. The first calculation asks for hundreds of litres of wine a day. The second, the generous one, still asks for tens of litres a day, every day, for life.

This is not a subtle statistical quibble. It is the difference between a laboratory result and a dietary recommendation, and it applies to a great deal of the "superfood compound" literature, not only to resveratrol. Whenever you read that a compound in a food "has been shown to" do something impressive, the first question worth asking is: at what dose, and could a person eating food ever reach it?

Back to Table of Contents

The Bioavailability Problem

Even if you could swallow enough, your body would not let much of it circulate.

Walle and colleagues settled this in 2004 with a carefully designed human study using radiolabelled resveratrol in six volunteers. Their findings deserve to be quoted precisely, because they are so often summarised backwards. Absorption of a dietary-relevant 25 mg oral dose was at least 70% — resveratrol crosses the gut wall very efficiently. Peak plasma concentration of resveratrol plus its metabolites reached about 491 ng/mL, roughly 2 micromolar. But unchanged resveratrol in plasma was under 5 ng/mL — trace amounts only.

The explanation is metabolic. The intestinal wall and the liver conjugate resveratrol almost instantly, attaching sulfate and glucuronic acid groups to its phenol rings; gut bacteria hydrogenate part of the molecule as well. Sulfate conjugation by the intestine and liver is so rapid that the authors identified it as the rate-limiting step in resveratrol's bioavailability. The compound is absorbed, chemically dismantled and packaged for excretion before it ever reaches the tissues where all those laboratory effects were demonstrated.

This matters because virtually every cell-culture experiment on resveratrol applies free, unconjugated resveratrol directly to cells at micromolar concentrations. Human plasma, after a realistic oral dose, contains free resveratrol at concentrations far below that. A later study using radiolabelled resveratrol in men undergoing prostate surgery examined how much actually reached the target tissue, and the general picture from that line of work has been consistent: what arrives in tissue is mostly conjugated metabolite, not the parent compound.

There is a legitimate counter-argument worth stating fairly. Walle and colleagues themselves suggested that resveratrol may accumulate in the epithelial cells lining the digestive tract — where concentrations are highest and conjugation happens — and that some conjugated metabolites may be biologically active or may be de-conjugated locally in inflamed tissue. That is a real hypothesis and it is still being investigated. It is not, however, the claim on the supplement bottle.

Back to Table of Contents

The Sirtuin Question and the Assay Artefact

The 2003 Nature paper said resveratrol activates SIRT1. That claim underwrote the entire "calorie-restriction mimetic" framing, and by 2009 it was in serious trouble.

The original assays used a fluorescently labelled peptide substrate — a convenience common in enzyme screening. Two independent groups showed that this was the source of the effect. Beher and colleagues reported in 2009 that resveratrol is not a direct activator of SIRT1 enzyme activity, and Pacholec and colleagues at Pfizer confirmed and extended this in 2010 in the Journal of Biological Chemistry, showing that resveratrol and several designed sirtuin-activating compounds appeared to activate SIRT1 only when the fluorophore-tagged peptide was used. Switch to a native substrate without the fluorescent tag and the activation disappeared. The compounds were interacting with the label, not activating the enzyme on its natural target.

The field did not end there — subsequent work has argued that resveratrol can activate SIRT1 allosterically with certain substrates, and that some of its effects in cells run through AMPK and other pathways instead. The mechanism remains genuinely unsettled. But the clean, confident story that resveratrol is a direct sirtuin activator, and that this is how it mimics calorie restriction, does not survive contact with the assay-artefact papers. When you see that story told without qualification, it is a decade out of date.

Back to Table of Contents

What Happened When Someone Measured Real People

Every argument so far has been about laboratory doses and laboratory conditions. In 2014 a group led by Richard Semba did something more direct: they measured resveratrol in people who were eating an ordinary diet, and followed them for nine years.

The setting was ideal. The InCHIANTI study — Invecchiare in Chianti, "Ageing in the Chianti Region" — is a population cohort in two villages in Tuscany, a place where red wine, grapes and olive oil are ordinary daily food. The team measured 24-hour urinary resveratrol metabolites in 783 community-dwelling men and women aged 65 and over, and followed them from 1998 to 2009. Urinary metabolites are a good exposure marker precisely because of the bioavailability problem described above: whatever is absorbed comes out in the urine as conjugates, so this captures real intake rather than a food-frequency questionnaire's guess.

Over nine years, 268 participants (34.3%) died. Sorted from the lowest to the highest quartile of urinary resveratrol metabolites, the proportion who died was 34.4%, 31.6%, 33.5% and 37.4% — a flat line, with a P value of 0.67. In the adjusted model, the lowest quartile had a hazard ratio for death of 0.80 (95% confidence interval 0.54 to 1.17) compared with the highest — a confidence interval that comfortably spans 1, meaning no detectable difference. Resveratrol levels were not associated with C-reactive protein, interleukin-6, interleukin-1β or TNF; nor with prevalent or incident cardiovascular disease; nor with cancer.

The authors' conclusion was blunt: resveratrol levels achieved with a Western diet did not have a substantial influence on health status or mortality risk in this population. This is the single most relevant study for the question most people are actually asking — will the resveratrol in my food or my wine help me live longer? — and its answer, in the best available setting, was no.

Two fair caveats. This is an observational cohort, not a trial, so it cannot rule out an effect of supplemental doses far higher than any diet delivers. And a single cohort in a single population is one line of evidence, not a verdict. But it is a well-designed study using a hard exposure biomarker and a hard endpoint, and nothing since has overturned it.

Back to Table of Contents

The Research-Fraud Case in This Exact Field

Any honest account of resveratrol has to include this, because it is a documented part of the field's history and it explains why some early literature should be discounted entirely.

In January 2012 the University of Connecticut announced the conclusion of a three-year internal investigation into Dipak K. Das, director of its Cardiovascular Research Center and one of the most prolific publishers of resveratrol and red-wine cardioprotection research. The investigation report ran to thousands of pages and the university stated that it had found numerous counts of fabrication and falsification of data, principally the manipulation of Western blot images. The university notified the journals involved and the federal agencies that had funded the work.

A substantial number of papers were subsequently retracted across several journals. The retractions are part of the permanent scientific record and can be looked up: one example that is straightforward to verify is the retraction notice for "Phytochemicals from Plants to Combat Cardiovascular Disease" in Current Medicinal Chemistry in 2012, on which Das was an author. Das died in 2013; he and his institution had disputed aspects of the investigation.

Two things should be said carefully here. First: this does not mean resveratrol research as a whole is fraudulent. The Jang, Howitz, Baur, Walle and Semba papers discussed above are from entirely unrelated groups and are not implicated. Science caught this one, which is the system working, if slowly. Second: it does mean that the sheer volume of positive resveratrol-and-the-heart papers published in the 2000s is a poorer guide to the truth than the count suggests, because a meaningful fraction of that volume came from one laboratory whose data were later found unreliable. If you are counting studies rather than weighing them, this is exactly the kind of thing that will mislead you.

The wider lesson is one worth carrying to any health claim: a field with a strong commercial incentive, a compelling story and a lot of enthusiastic early publications is exactly the environment in which errors and misconduct propagate before anyone checks.

Back to Table of Contents

The French Paradox Is a Hypothesis, Not a Fact

The whole resveratrol enterprise sits on top of an older idea. In 1992 Renaud and de Lorgeril published a paper in The Lancet observing that French populations had lower coronary heart disease mortality than their intake of saturated fat seemed to predict, and proposing that regular moderate wine consumption — specifically its effect on platelet aggregation — might explain the discrepancy. The phrase "French paradox" entered the language, an American television programme ran a segment on it, and red wine sales in the United States rose sharply.

It was a reasonable hypothesis in 1992. It has not held up cleanly since, and several serious alternative explanations exist:

1. The time-lag explanation. Law and Wald argued in the BMJ in 1999 that heart disease mortality in a population reflects saturated fat and cholesterol levels from decades earlier, not current intake. French fat consumption rose later than in Britain and the United States, so French heart disease rates at the time of the "paradox" were tracking a leaner French diet of the 1950s and 60s. On this account the paradox is a statistical artefact of comparing current diet with lagged disease, and it largely disappears when the lag is accounted for.
2. Ascertainment and coding differences. Countries certify and code cause of death differently. Deaths that would be coded as coronary heart disease in one country may be coded otherwise in another, which can shift the apparent national rate without any biological difference.
3. The rest of the diet and the rest of the life. Ferrières, reviewing the question in Heart, set the wine hypothesis alongside the broader French dietary pattern — more vegetables and fruit, more fish, olive oil in the south, smaller portions, meals eaten slowly and at table — and the difficulty of attributing a population difference to any single component.
4. Alcohol itself, not resveratrol. To the extent that moderate alcohol consumption is associated with lower cardiovascular risk in observational data, ethanol's own effects on HDL and on platelet function are a more plausible route than a compound present in milligram quantities. And even that association is now contested, with newer analyses arguing that much of the apparent benefit of moderate drinking is confounded by who ends up in the abstainer group — former heavy drinkers and the already unwell.

The honest position is that the French Paradox is a contested observation with several competing explanations, of which "a stilbene in red wine" is neither the most parsimonious nor the best supported. It is emphatically not a reason to start drinking. Alcohol carries well-established harms of its own — see our page on alcohol — and no health authority recommends taking up drinking for cardiovascular benefit.

Back to Table of Contents

What the Human Trials Actually Show

Resveratrol has now been tested in humans many times, at supplemental doses far above anything food provides. The results are neither nothing nor what was hoped.

Smoliga, Baur and Hausenblas reviewed the human clinical trial literature in 2011 and found that supplementation was generally well tolerated at the doses tested, with scattered signals in metabolic and vascular measures and a striking lack of the dramatic effects the animal work had promised. Berman and colleagues reviewed the clinical trials again in 2017 in npj Precision Oncology, with a similar conclusion: plausible activity, small and inconsistent human effects, and persistent bioavailability limits. Novelle and colleagues, writing in Ageing Research Reviews in 2015 under the pointed title "Resveratrol supplementation: where are we now and where should we go?", catalogued the same pattern of promising preclinical work and underwhelming translation.

The single most instructive human trial is Yoshino and colleagues, in Cell Metabolism, 2012. This was a properly controlled test of the central metabolic claim: non-obese women with normal glucose tolerance, given resveratrol supplementation, assessed with rigorous metabolic measurements. The title states the result — resveratrol supplementation did not improve metabolic function in this group. Some subsequent trials in people who are obese or diabetic have shown modest effects on individual markers, and it is genuinely possible that resveratrol does more in metabolically unhealthy people than in healthy ones. But the marquee claim — that resveratrol reproduces the benefits of calorie restriction in humans — has not been demonstrated.

It is worth adding that high-dose resveratrol supplements are not automatically benign. Gastrointestinal upset is common at gram-level doses. Resveratrol inhibits certain drug-metabolising enzymes and affects platelet function, which matters if you take anticoagulants. And "natural" supplement doses of a hundred to a thousand times any dietary exposure are, by definition, a pharmacological intervention rather than a food.

Back to Table of Contents

What This Means for You

None of this is a reason to stop eating grapes. It is a reason to eat them for the right reasons.

• Eat grapes because they are good fruit — water, fibre, potassium, vitamin K, vitamin C, and a genuinely substantial load of anthocyanins and flavan-3-ols in the skins. Those polyphenols have a better human evidence base than resveratrol does, and we cover it in Polyphenols and Vascular Health.
• Eat the skins. Every interesting compound in a grape is in or near the skin. Peeling grapes throws away the reason to eat them.
• Choose dark grapes when you have the choice. Red and black varieties carry far more anthocyanin than green ones. This is one place where the colour genuinely tracks the chemistry.
• Do not buy resveratrol supplements on the strength of the mouse studies. If you want to try one for reasons of your own, that is your call — but know that you are buying a knotweed extract on the basis of animal evidence that human trials have not confirmed, and that a large cohort study of people with real dietary exposure found no association with inflammation, cardiovascular disease, cancer or death.
• Do not start drinking wine for your health. The French Paradox does not support it, the resveratrol content cannot support it, and alcohol's risks are not hypothetical.
• Use this page as a template. The pattern here — spectacular animal result, unreachable dose, poor bioavailability, no human confirmation, heavy marketing — recurs across the supplement industry. Once you have seen it clearly once, you will recognise it.

Back to Table of Contents

Key Research Papers

1. Semba RD, Ferrucci L, Bartali B, Urpí-Sarda M, et al. Resveratrol levels and all-cause mortality in older community-dwelling adults. JAMA Internal Medicine. 2014;174(7):1077–1084. — doi:10.1001/jamainternmed.2014.1582
2. Walle T, Hsieh F, DeLegge MH, Oatis JE Jr, Walle UK. High absorption but very low bioavailability of oral resveratrol in humans. Drug Metabolism and Disposition. 2004;32(12):1377–1382. — doi:10.1124/dmd.104.000885
3. Baur JA, Pearson KJ, Price NL, Jamieson HA, et al. Resveratrol improves health and survival of mice on a high-calorie diet. Nature. 2006;444(7117):337–342. — doi:10.1038/nature05354
4. Baur JA, Sinclair DA. Therapeutic potential of resveratrol: the in vivo evidence. Nature Reviews Drug Discovery. 2006;5(6):493–506. — doi:10.1038/nrd2060
5. Jang M, Cai L, Udeani GO, Slowing KV, et al. Cancer chemopreventive activity of resveratrol, a natural product derived from grapes. Science. 1997;275(5297):218–220. — doi:10.1126/science.275.5297.218
6. Howitz KT, Bitterman KJ, Cohen HY, Lamming DW, et al. Small molecule activators of sirtuins extend Saccharomyces cerevisiae lifespan. Nature. 2003;425(6954):191–196. — doi:10.1038/nature01960
7. Pacholec M, Bleasdale JE, Chrunyk B, Cunningham D, et al. SRT1720, SRT2183, SRT1460, and resveratrol are not direct activators of SIRT1. Journal of Biological Chemistry. 2010;285(11):8340–8351. — doi:10.1074/jbc.M109.088682
8. Beher D, Wu J, Cumine S, Kim KW, et al. Resveratrol is not a direct activator of SIRT1 enzyme activity. Chemical Biology & Drug Design. 2009;74(6):619–624. — doi:10.1111/j.1747-0285.2009.00901.x
9. Yoshino J, Conte C, Fontana L, et al. Resveratrol supplementation does not improve metabolic function in nonobese women with normal glucose tolerance. Cell Metabolism. 2012;16(5):658–664. — doi:10.1016/j.cmet.2012.09.015
10. Weiskirchen S, Weiskirchen R. Resveratrol: how much wine do you have to drink to stay healthy? Advances in Nutrition. 2016;7(4):706–718. — doi:10.3945/an.115.011627
11. Stervbo U, Vang O, Bonnesen C. A review of the content of the putative chemopreventive phytoalexin resveratrol in red wine. Food Chemistry. 2007;101(2):449–457. — doi:10.1016/j.foodchem.2006.01.047
12. Smoliga JM, Baur JA, Hausenblas HA. Resveratrol and health — a comprehensive review of human clinical trials. Molecular Nutrition & Food Research. 2011;55(8):1129–1141. — doi:10.1002/mnfr.201100143
13. Berman AY, Motechin RA, Wiesenfeld MY, Holz MK. The therapeutic potential of resveratrol: a review of clinical trials. npj Precision Oncology. 2017;1:35. — doi:10.1038/s41698-017-0038-6
14. Novelle MG, Wahl D, Diéguez C, Bernier M, et al. Resveratrol supplementation: where are we now and where should we go? Ageing Research Reviews. 2015;21:1–15. — doi:10.1016/j.arr.2015.01.002
15. Cai H, Scott EN, Britton RG, et al. Distribution and metabolism of [¹⁴C]-resveratrol in human prostate tissue after oral administration. The American Journal of Clinical Nutrition. 2021;113(5):1115–1125. — doi:10.1093/ajcn/nqaa414
16. Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. The Lancet. 1992;339(8808):1523–1526. — doi:10.1016/0140-6736(92)91277-F
17. Law M, Wald N. Why heart disease mortality is low in France: the time lag explanation. BMJ. 1999;318(7196):1471–1480. — doi:10.1136/bmj.318.7196.1471
18. Ferrières J. The French paradox: lessons for other countries. Heart. 2004;90(1):107–111. — doi:10.1136/heart.90.1.107
19. Vasanthi HR, ShriShriMal N, Das DK. Retraction notice: phytochemicals from plants to combat cardiovascular disease. Current Medicinal Chemistry. 2012;19(14):2242–2251. — doi:10.2174/092986712800229078 (one of the retractions arising from the University of Connecticut investigation)
20. Live literature search — the full retraction record for this author: PubMed: Das DK retracted publications
21. Live literature search — ongoing human resveratrol trials: PubMed: resveratrol randomised controlled trials

Back to Table of Contents

Connections

• Grapes — the main topic page: varieties, nutrition, selection and storage.
• Grapes — Benefits Deep Dive — the hub for this set of articles.
• Grapes: History and Origins — domestication in the South Caucasus and the phylloxera catastrophe.
• Resveratrol — the compound's own page.
• Anthocyanins — the pigments that give dark grapes their colour, with a stronger human evidence base.
• Grape Seed Extract — the proanthocyanidin concentrate and its trial record.
• Polyphenols and Vascular Health — what the grape polyphenols other than resveratrol actually do.
• Alcohol — why "drink red wine for your heart" is bad advice.
• Longevity & Healthy Aging — calorie restriction, sirtuins and what the evidence supports.
• Quercetin — another polyphenol with a similar dose-and-bioavailability story.
• All Antioxidants — the complete compound index.
• Cardiology — the conditions this literature claims to affect.

Back to Table of Contents

Featured Videos