Resveratrol
trans-resveratrol

Resveratrol is a polyphenol found in the skin and seeds of grapes, and therefore in red wine. It is also found in the root of a plant, Polygonum cuspidatum, from which it is extracted for some supplements.

Its fame comes from red wine and from a 2003 study in which it extended the life of yeast in the laboratory. That gave rise to the idea of an anti-aging supplement.

In grapes and in the root of Polygonum, resveratrol occurs both free and bound to a sugar; in the root, that bound form, polydatin, is about ten times more abundant than the free one.

Food provides little. EFSA estimates that the people who get the most resveratrol from their diet reach about 2.93 mg a day, and places the 150 mg of a supplement about 50 times above that.

The molecule is a stilbene, C₁₄H₁₂O₃, of 228.25 daltons. The form authorized in the EU is the E isomer, which is the trans one.

Each result is easier to understand with a figure in front of it: 150 mg a day, the maximum the European Union authorizes in supplements for synthetic or yeast-derived resveratrol. Many of the trials used considerably more.

Sugar, cholesterol and blood pressure. At the authorized dose, most of the trials we have gathered find no difference compared with placebo. The exception is a very small, one-month trial linked to the manufacturer of the ingredient.

Inflammation. At that same dose, the trials that measure C-reactive protein, a blood marker of inflammation, also find no difference from placebo. The clearest favorable result, in healthy smokers, was obtained with more than three times that dose.

Sport. The trials do not show a consistent improvement in performance. The only favorable one, in older people and with more than three times the authorized dose, is limited to a leg strength test. In two others, those who took resveratrol improved less with training than those who took placebo on some measures, and one of them used the authorized dose.

Memory. In postmenopausal women, two trials with 150 mg a day found better cognitive performance than with placebo. Both are from the same research group, received product and funding from manufacturers of the ingredient, and we have not located any replication at another center. In healthy adults, single doses did not change cognitive performance, and in Alzheimer’s disease a trial with doses of grams only moved biological markers, with no demonstrated clinical benefit.

Aging and cancer. We have not located any controlled trial in people that measures whether it extends life, or whether it prevents or treats cancer.

In men with metabolic syndrome, 150 mg daily for 16 weeks did not change inflammation, glucose, lipids, blood pressure or liver fat compared with placebo. The same trial had a third group on 1,000 mg daily, and in it total cholesterol, LDL and a glucose marker rose compared with placebo.

In people with blood lipid disorders, 100 mg daily did not change lipids; with 300 and 600 mg, uric acid fell, which was a secondary outcome. In type 2 diabetes, neither 40 nor 500 mg daily for six months improved the metabolic profile.

The cards in §08 cover a six-month trial in overweight people at the authorized dose and the only favorable exception.

The favorable trial in healthy smokers gave 500 mg daily for 30 days and placebo for another 30, to the same people and in random order, a crossover trial. With resveratrol, C-reactive protein ended at 0.47 times that with placebo and triglycerides at 0.71, and the blood’s antioxidant capacity rose. The manufacturer donated the product, with no other declared involvement.

At 150 mg or less, the trials that measure C-reactive protein or interleukin 6 find no differences, and only Timmers’ trial, on its card, saw two other inflammation signals fall. The other signals at low doses come from a root extract, which does not isolate the effect of resveratrol, or from a study in firefighters in which there is no record that allocation was random.

The closest is a study of 20 patients with colorectal cancer who took 0.5 or 1 g daily for the eight days before surgery. A marker of tumor proliferation went from 88.0% to 83.2%, with no comparison group.

Blood pressure. A meta-analysis of 17 trials found no significant overall effect on systolic or diastolic pressure, with high heterogeneity. Its subgroups with doses of 300 mg or more, or with diabetes, did come out significant; we do not give those figures because it is not on record how many trials enter each subgroup. Another meta-analysis describes a drop in systolic pressure of 11.9 mmHg with 150 mg or more, but in a subgroup with mixed populations and with no dose effect in its meta-regression. In general, the pooled figures of the meta-analyses we have gathered say little: their heterogeneity ranges from 50% to 96%.

Fatty liver. There are trials with favorable results and others with null ones, and a meta-analysis found no effect on liver enzymes. A systematic comparison is lacking to explain whether the difference is due to the dose, the population, the quality of the trial or the product.

The same group, in type 2 diabetes. Timmers’ group repeated 150 mg daily for 30 days in 17 people with well-controlled type 2 diabetes who continued their treatment, and insulin sensitivity did not change compared with placebo.

The only favorable trial brought together 30 people aged 65 to 80 who did aerobic and strength exercise, with 500 mg daily for 12 weeks. Compared with placebo, only power and strength when extending the knee at a specific speed improved. Maximal oxygen uptake and mitochondrial density did not improve more than with placebo. In addition, it replaced participants who dropped out after randomization, something that compromises the comparison.

It has barely been studied. The only double-blind placebo-controlled trial we have gathered is a pilot with seven marathon runners who took 600 mg daily the week before: after the race there were no differences in inflammation markers or in muscle soreness.

In the young men doing intervals, maximal oxygen uptake and power in the Wingate test improved in both groups with no statistical difference between them; what separated the groups was maximal aerobic power and the activation of the PGC-1α, SIRT1 and SOD2 genes. Its funding is not accessible, so its independence cannot be determined.

In the favorable trial in older people, the interaction vs. placebo was significant for mean power when extending the knee at 60 degrees per second (p = 0.004) and for its peak torque (p = 0.038), not for maximal oxygen uptake (p = 0.09).

The trial on Gliemann’s card received critical letters arguing that the data were overinterpreted, and its authors replied.

In healthy adults, a single dose of 250 or 500 mg increased blood flow in the brain compared with placebo in one trial; in another, 250 mg did not change it.

In mild or moderate Alzheimer’s disease, a trial with 119 patients and up to 2 g daily for a year was designed to measure biological markers, not symptoms. An amyloid marker remained more stable than with placebo and the brain lost more volume with resveratrol, a finding whose significance, according to the authors, is unclear. They also acknowledge that the trial was not powered to detect a clinical benefit.

In the 14-week trial, the response of the brain’s vessels improved vs. placebo when breathing air with more carbon dioxide (p = 0.010) and during cognitive tasks (p = 0.002). Overall performance (p = 0.020) and verbal memory (p = 0.041) improved after adjusting for variables chosen post hoc, and so did anxiety (p = 0.025). Eighteen participants had no measurable signal in the vessels and contributed only cognitive data. The manufacturer of the ingredient provided the product and a small financial grant.

With 200 mg of resveratrol and 320 of quercetin a day for 26 weeks, word retention improved compared with placebo in healthy overweight adults, but the control group was formed by matching, not by random allocation. A later trial by the same group, with 200 mg daily and probably also quercetin, did not repeat the effect on verbal memory. Neither of the two isolates the effect of resveratrol.

The trials have been done in adults with overweight, metabolic syndrome, type 2 diabetes or blood lipid disorders, in postmenopausal women and in people who were training. None outlines a profile of a healthy person for whom, at the authorized dose, it has worked consistently. The memory signals in postmenopausal women come from a single research group.

The European authorization of synthetic or yeast-derived resveratrol as a novel food is for adults only and sets the maximum at 150 mg a day. The trials have used anything from 40 mg to several grams daily. Those figures describe what has been studied; they are not intake guidelines or recommendations from Suplepedia.

The applicant initially proposed 450 mg daily; some Member States objected and, at EFSA’s request, 150 was set. Root extracts have a history of consumption in the EU before 1997, and we have not been able to verify what limit applies to them.

The trials gave it with breakfast, in the morning on an empty stomach, before breakfast and at bedtime, or split between lunch and dinner; none compared timings. In the United States, a food ingredient notification to the FDA (a GRAS notification), for bottled water, was closed in 2007 without evaluation, at the request of the party that had submitted it.

It is well absorbed, but barely reaches the blood unchanged: the gut and the liver quickly transform it into other molecules. Whether those molecules have any effect is, for now, a hypothesis.

With a 25 mg dose radioactively labeled to track it, at least 70% was absorbed, but intact resveratrol in the blood stayed below 5 nanograms per milliliter: its bioavailability is very low. What circulates is mainly its transformed forms: with single doses of 0.5 to 5 g, their exposure reached up to 23 times that of free resveratrol.

In the laboratory, according to the authors of one of those studies, its effects require at least 5 micromoles per liter, about 1,141 ng/ml, our own calculation. With 5 g, the highest single dose they tested, the peak in the blood was 539 ng/ml.

The two studies do not contradict each other, but they cannot be added together: they compare different things. Neither tested 150 mg, neither compared a standard breakfast with fasting, and the single-dose one was authored by staff of the company developing that ingredient. No recommendation on when to take it follows from them.

With repeated doses of 150 mg six times a day, 900 mg daily, the mean peak of free resveratrol was 63.8 ng/ml. Its half-life in the blood ranges from 2.9 to 8.9 hours depending on the dose, and that of its transformed forms is similar.

No. The resveratrol in supplements can be made by chemical synthesis, produced with a genetically modified yeast or extracted from the root of Polygonum. The first two are the ones authorized in the European Union as a novel food. A root extract is not pure resveratrol: other plant compounds coexist in that root.

The molecule exists in two forms, trans and cis, which are two different arrangements of the same formula. The one authorized and assessed in the EU is the trans form. Foods contain mixtures of both and of their sugar-bound forms.

A 2021 European regulation bans emodin and any preparation containing it, because of the genotoxicity EFSA observed in the laboratory. The ban is defined by substance, not by plant: a root extract falls within its scope if it contains emodin. We have not located published analyses of emodin in commercial resveratrol extracts, so it cannot be stated that a specific product, or root extracts in general, breach this rule.

Many products combine resveratrol with other substances. With piperine, in one trial, the amount reaching the blood did not change; with quercetin, it did not either, in a study at gram doses. The combination with NAD⁺ precursors rests on a laboratory idea: that resveratrol activates the protein SIRT1, which needs NAD⁺ to work. That direct activation is still under debate (§09).

To get around its poor solubility, a micronized form was developed, made of very fine particles: the one in the myeloma trial in §11, created by Sirtris, then part of the GSK group. In the opposite direction, in 2022 the EU removed the maximum particle size requirement for yeast-derived resveratrol, considering its profile comparable to that of the assessed synthetic one. The emodin rule itself notes that hydroxyanthracene derivatives, the family emodin belongs to, can be removed from preparations by filtration, down to traces.

In the trial that tested the combination with piperine, blood flow in the brain rose compared with resveratrol alone; in a separate cohort of six people, what reached the blood did not change. The combination with quercetin rests on the hypothesis, born in the laboratory, that quercetin slows the transformation of resveratrol in the body.

What the label says and what the capsule contains may not match. In 20 products analyzed in Slovenia in 2023, the measured content ranged from 5% to 234% of the declared amount. In 14 brands analyzed in Italy in 2012, two were below the limit of detection.

In the Italian analysis, only 5 of the 14 brands came close to the declared amount within the good manufacturing practice margin, between 95% and 105%; three were between 8% and 64%.

In the Italian analysis, another four brands fell slightly outside the margin, between 83% and 111%. The Slovenian one compared per capsule, tablet or sachet, not per daily dose, and its abstract and results do not match on two percentages; that is why we only give the range here.

Does resveratrol activate the SIRT1 protein directly in the body, or through indirect pathways? Do the transformed forms circulating in the blood do anything, or are they just a reserve?

Is the blunting of training adaptations confirmed at the authorized dose? Today it rests on a single small trial. One hypothesis is that the reactive oxygen species produced by exercise serve as a signal to adapt, and that an antioxidant could dampen it.

Does the memory signal hold up outside the group that found it? Its effect was already smaller in the long trial than in the short one.

Some specific gaps: we have not located trials in people with resveratrol and NAD⁺ precursors, peer-reviewed pharmacokinetic studies of the formulations advertised as better absorbed, analyses of how much trans turns into cis with light or inside the capsule, verifiable quality seals for resveratrol, or a systematic review of adverse effects by dose.

Nor is it clear what the inhibition of cytochrome P450 enzymes means at 150 mg daily, when the study that measured it used one gram.

A reading of ours, not the authors’: when effects appear, they seem to depend on the dose and can go in opposite directions depending on what is measured. Cholesterol rose with 1,000 mg and not with 150; uric acid only fell from 300 upward.

People taking medicines, only under medical supervision. In the EU, the labeling of these supplements must state that “people using medicines should only consume the product under medical supervision”. EFSA, the European Food Safety Authority, noted that resveratrol may interact with specific medicines.

Pregnancy, breastfeeding and childhood. The European authorization of synthetic or yeast-derived resveratrol is for adults only, and we have not located data in people in these situations. The absence of data does not mean the absence of risk.

High doses. From 1 gram a day, digestive discomfort, mainly diarrhea, has been described in studies without a placebo group. In patients and with doses of several grams, elevated liver enzymes have also been described and, with a special formulation, severe kidney failure.

EFSA considered that a metabolite of resveratrol could inhibit cytochrome P450 enzymes and affect medicines that depend mainly on the one called CYP2C9. In 42 healthy volunteers, 1 g daily for four weeks, without placebo, altered the activity of four of those enzymes; that is a dose almost seven times the authorized one. No studies have been located with anticoagulants or antiplatelet drugs.

The rise in liver enzymes was seen with 3 g daily in men with fatty liver, compared with placebo, up to the sixth week; its magnitude is not settled. The kidney failures occurred with 5 g daily of a micronized formulation in patients with myeloma, several with already impaired kidneys, and two needed temporary dialysis.

Based on rat studies, EFSA put the margin between its reference dose and 150 mg daily at 172 times, and below 62 in pregnant rats, whose weight fell already at the lowest dose tested; in that study there were no effects on the development of the offspring. In studies without a placebo group, with 1 to 2 g daily, isolated cases of hot flashes have been described; with the animal data available, EFSA sees no cause for concern about estrogenic activity. In the myeloma trial, the authors attributed the kidney failures to the patients’ disease, without demonstrating it.

This page describes what the studies have measured. It is not medical advice and does not replace consultation with a health professional.