Name: Resveratrol: Real in a Petri Dish, Unproven in People
Creator: Realised
Published: 2026-06-01

Summary

Resveratrol is sold as the molecule behind the "red wine is good for your heart" story and as a longevity supplement, and that human claim does not hold up — the dose in wine is trivially small, systemic bioavailability is around half a percent, human trials move surrogate markers inconsistently at best with no proven longevity or cardiovascular outcome, a large human cohort found dietary resveratrol unrelated to who lived longer, the flagship commercial program (GSK's $720M Sirtris bet) collapsed, and part of the early cardioprotection literature was retracted for fraud — yet the molecule is

Why Emerging

Emerging because the entry's positive content is a real but early, animal-and-mechanism signal that has not translated to a human outcome. The preclinical replication (yeast/worm/fly lifespan, overfed-mouse survival) is coherent and cross-species, but everything above the animal line weakens: the mouse benefit is diet-conditional, human bioavailability is around half a percent, human trials reach only surrogate markers and inconsistently, the best human outcome cohort is null, the flagship translation program failed, and part of the early literature was fraud. An early mechanism with a failed-to-null human translation is the definition of the Emerging tier for the intervention claim.

NOT Moderate because there is no body of consistent human trials on meaningful endpoints — the human data are surrogate-only and inconsistent, and the single best outcome signal is null. The therapeutic claim does not clear the Moderate bar.

NOT Experimental-and-dismissed because the preclinical signal is genuinely replicated and mechanistically grounded; this is not fringe biology, it is real biology whose human relevance is unproven. The honest label is Emerging with the human longevity claim effectively null.

The per-claim split (read this, not just the headline):
• Preclinical lifespan extension (yeast/worm/fly), overfed-mouse survival: real and replicated, but animal-only (Emerging as applied to humans).
• Diet-conditional cap (no lifespan gain on a normal diet): strong disconfirming receipt from within the animal data.
• Bioavailability around 0.5%; wine dose trivial: well-established pharmacokinetics (high confidence).
• Human surrogate-marker effects: inconsistent, no outcome (surrogate-endpoint trap).
• Human longevity/CVD outcome: unproven-to-null (null cohort; failed commercial program).

Practical takeaway

The framing to hold: resveratrol is a real preclinical mechanism with an unproven human outcome. It is not a longevity supplement you should take to live longer, and "red wine is good for you because of resveratrol" is not true at any dose you would actually drink. Treat it as interesting biology, not a proven intervention.

Don't take it for the reasons it's sold.
• Longevity / anti-aging in a healthy person: skip it. No human trial or cohort shows resveratrol extends life or prevents age-related decline; the best cohort signal is null, and the mouse benefit disappears on a normal diet.
• "Heart health via red wine": the resveratrol dose in wine is trivial and cannot be the mechanism. Do not treat this as a reason to drink — the alcohol question is separate and belongs to alcohol_cross_pillar_effects.

If you're weighing it anyway, know the ceiling.
• The realistic human evidence is limited to inconsistent surrogate-marker effects (a modest blood-pressure nudge in some trials), mostly in metabolically stressed people, with no demonstrated hard outcome.
• Systemic bioavailability of the free compound is around 0.5%, so most of an oral dose never reaches circulation as resveratrol; the marketing rarely mentions this.
• The best-funded serious attempt to turn the mechanism into a therapy (Sirtris/GSK) failed on both efficacy and safety — a real-world signal that translation is hard, not a mere commercial footnote.

Hold the both-ways line. The mechanism is real and worth watching; the human longevity claim is unproven; the wine story is false on the arithmetic. "Resveratrol does nothing" and "resveratrol makes you live longer" are both wrong, in opposite directions.

Evidence detail

Why This Entry Exists

Resveratrol occupies the exact gap this library exists to police: a real molecular mechanism that gets laundered into a human healthspan promise it has never earned. The preclinical work is genuine — sirtuin activation, lifespan extension in yeast and worms, a physiology-shifting effect in overfed mice — and it is then stretched, with the help of a memorable "French paradox" story, into a claim that a supplement (or a glass of red wine) makes people live longer and protects their hearts. It does not, on any evidence we have. So this entry has to hold two opposite lines at once: refuse the lazy "resveratrol does nothing" dismissal that ignores a coherent, replicated preclinical signal, and dismantle the "drink red wine / take resveratrol for longevity" pitch that the same word is used to sell.

The failure modes run in both directions. Over-claim the mechanism and a healthy person spends money on a poorly-absorbed supplement, or worse treats "red wine is heart-healthy" as a licence to drink. Dismiss it entirely and you miss that this is real biology with a plausible pathway that genuinely helps sick, metabolically overloaded animals. The load-bearing distinction is between a mechanism or biomarker (sirtuin activation, a nudge in blood pressure) and a hard human outcome (living longer, fewer cardiovascular events). Resveratrol moves the former inconsistently and has never demonstrated the latter. Getting that ordering right — a real preclinical signal, a failed human translation, and a red-wine story that dies on the dose arithmetic — is the whole point of the entry.

What bad advice this protects against, in all directions:
• "Red wine is good for your heart — that's the resveratrol" → the resveratrol dose in wine is trivial (roughly 1.5 to 3 mg per litre); matching the doses that helped mice would take on the order of hundreds to a thousand glasses a day. The wine-longevity mechanism does not survive the arithmetic, and the alcohol itself is a separate question (see alcohol_cross_pillar_effects).
• "Resveratrol extends lifespan — it's proven in animals" → it extends lifespan in yeast, worms and flies and improved survival of overfed mice, but on a normal diet it did NOT extend mouse lifespan; the effect is a rescue of metabolic overload, not a general longevity switch.
• "Take resveratrol and you'll live longer / avoid heart disease" → no human trial or cohort shows this; a large cohort found urinary resveratrol unrelated to mortality, cardiovascular disease or cancer. The human outcome claim is unproven-to-null.
• "It activates SIRT1, so it works" → sirtuin activation is a mechanism, not an outcome; independent labs argued the original assay was an artifact, and the best-funded attempt to turn it into a drug failed on efficacy and safety.
• "Just take a supplement to get the dose" → systemic bioavailability is around 0.5% because resveratrol is rapidly conjugated on first pass; most of what you swallow never reaches the bloodstream as free compound.
• "Resveratrol does nothing, it's a scam" → the reverse over-correction; the preclinical signal is real and replicated, and some human trials do nudge surrogate markers. The honest message is "real mechanism, unproven human outcome," not a blanket dismissal.

This entry owns the resveratrol preclinical-signal-versus-human-claim verdict and the red-wine dose debunk. It defers the alcohol question entirely to alcohol_cross_pillar_effects (do not let "red wine" become an endorsement or condemnation of drinking here), and the surrogate-versus-outcome logic to surrogate_endpoints_vs_outcomes rather than re-deriving it. It states those boundaries and routes there.

Evidence

Organised from the real preclinical signal up through its collapse in humans, with the tier signal inline. The preclinical replication is the firmest positive part; the diet-conditional cap, the bioavailability barrier, and the null human cohort are the firmest disconfirming parts. Read the tiers, not just the thesis.

The preclinical signal is real and replicated (genuine, but animal-only).

1. Resveratrol extends lifespan in simple organisms and improved survival of overfed mice, shifting their physiology toward lean controls. In yeast, worms and flies it extends lifespan in a Sir2-dependent way; in the flagship mouse study, high-calorie-fed middle-aged mice given resveratrol shifted toward the physiology of lean controls (increased insulin sensitivity, AMPK/PGC-1α activation, more mitochondria) and had significantly improved survival. This is a coherent, replicated preclinical signal, not nothing. (Baur JA, Sinclair DA, et al. "Resveratrol improves health and survival of mice on a high-calorie diet." Nature 2006, 444:337; building on Howitz KT, Sinclair DA, et al., Nature 2003, on yeast Sir2. Emerging as applied to humans — this is animal/preclinical evidence only, but it is replicated across species and mechanistically grounded. Sinclair-lab work; David Sinclair co-founded Sirtris and held a commercial stake in the sirtuin-activator story, a pro-resveratrol conflict of interest to weigh.)

2. The lifespan extension is diet-conditional: on a NORMAL diet, resveratrol did NOT extend mouse lifespan. A follow-up found that resveratrol improved some healthspan markers in mice on a standard diet but did not extend their lifespan. The benefit in the earlier study was a rescue of metabolic overload, not a general longevity effect — a distinction that directly caps the longevity claim even within rodent data. (Pearson KJ, de Cabo R, et al. "Resveratrol Delays Age-Related Deterioration and Mimics Transcriptional Aspects of Dietary Restriction without Extending Life Span." Cell Metabolism 2008, 8:157. A strong disconfirming receipt that limits the claim from inside the animal literature. NIA intramural work led by de Cabo — independent of the supplement industry, which strengthens the null rather than weakening it.)

The red-wine story fails the dose arithmetic (kills the wine-longevity mechanism specifically).

3. Red wine carries only about 1.5 to 3 mg of resveratrol per litre — orders of magnitude below the doses that helped mice. Matching the resveratrol doses used in the mouse studies would take roughly hundreds to a thousand glasses of wine per day. The "French paradox / red wine is protective because of resveratrol" mechanism does not survive this arithmetic; any real effect would require concentrated supplements, not wine. (Linus Pauling Institute resveratrol review; NIH/press commentary on the wine-dose gap, consistent with the Sinclair group's own dose estimates. Kills the wine-longevity mechanism specifically. The wine industry and its trade press benefit from the "paradox" narrative — cui bono for the pro-wine framing. The alcohol question itself is out of scope here — see alcohol_cross_pillar_effects.)

The bioavailability barrier (well-established pharmacokinetics).

4. Systemic bioavailability is around 0.5% — resveratrol is absorbed but almost entirely conjugated on first pass. Oral resveratrol is roughly 70% absorbed but has only about 0.5% systemic bioavailability, because it is rapidly conjugated to sulfates and glucuronides, so almost no free resveratrol reaches the circulation. This is a pharmacokinetic barrier that undercuts any simple translation of animal doses to humans. (Walle T, et al. bioavailability studies on oral resveratrol; clinical pharmacokinetic meta-analyses (2024) confirming roughly 0.5% free-compound systemic exposure. Well-established pharmacokinetics; independent literature with no obvious commercial interest.)

Human trial evidence: surrogate markers only, inconsistent, no outcome (surrogate-endpoint trap).

5. Human RCTs and meta-analyses are inconsistent and confined to surrogate metabolic markers — with no proven hard clinical outcome. Systematic reviews and meta-analyses of resveratrol trials report modest and inconsistent effects on surrogate markers (some blood-pressure reduction, variable metabolic indices), with no hard clinical-outcome data; many trials enrolled metabolically healthy people with little room to improve. Moving a biomarker is not extending healthspan. (Multiple systematic reviews/meta-analyses of resveratrol RCTs, e.g. Frontiers in Physiology 2022 and an umbrella review of obesity indices, 2025 — modest, inconsistent surrogate effects, no outcome data. Emerging-to-Moderate on surrogates, null on outcomes — the surrogate-endpoint trap, see surrogate_endpoints_vs_outcomes. Mixed funding; several resveratrol RCTs are supplement-adjacent, a publication-bias risk toward positive surrogate findings.)

The best human outcome signal is null (anchors the "unproven in people" verdict).

6. In a large older-adult cohort, urinary resveratrol was NOT associated with mortality, cardiovascular disease, cancer, or inflammatory markers. In the InCHIANTI cohort (n=783, aged 65 and over), urinary resveratrol metabolite concentration — a marker of dietary resveratrol intake — showed no association with inflammatory markers, cardiovascular disease, cancer, or all-cause mortality. Dietary resveratrol at real-world intake did not predict who lived longer. (Semba RD, Ferrucci L, et al. "Resveratrol Levels and All-Cause Mortality in Older Community-Dwelling Adults." JAMA Internal Medicine 2014, 174:1077, InCHIANTI cohort. The best available human outcome signal, and it is null — this anchors the verdict. NIA-funded independent cohort; a supplement-company figure publicly attacked the study, which is pro-supplement pushback to weigh, not a flaw in the cohort.)

The commercial flagship failed (strong real-world disconfirmation).

7. GSK's $720M Sirtris bet collapsed, and independent labs argued resveratrol does not directly activate SIRT1. GSK acquired Sirtris for about $720M in 2008 for its resveratrol formulation (SRT501); a Phase IIa myeloma trial was halted in 2010 after cases of renal failure and poor tolerability, and GSK shut Sirtris in 2013. Separately, Pfizer and Amgen groups reported that resveratrol and related "sirtuin activators" were not direct activators of SIRT1 — the apparent activation was a fluorophore assay artifact. The best-funded attempt to translate the mechanism failed on efficacy and safety. (GSK/Sirtris coverage, Nature News 2013 and trade press; Pacholec M, et al. (Pfizer), "SRT1720, SRT2183, SRT1460, and resveratrol are not direct activators of SIRT1." J Biol Chem 2010. Strong real-world disconfirmation. Amgen and Pfizer were commercial competitors to Sirtris — cui bono against it — but the assay-artifact critique was mechanistically substantiated and GSK's own program corroborated poor drug-likeness.)

Part of the early literature was fraud (contaminated evidence base).

8. The University of Connecticut found 145 counts of data fabrication by a leading resveratrol-cardioprotection researcher, leading to 20+ retractions. A 2012 UConn misconduct investigation found 145 counts of data fabrication and falsification by Dipak K. Das, resulting in more than 20 retractions and notification of 11 journals. A chunk of the early "red wine / resveratrol is heart-protective" primary literature therefore cannot be trusted. (University of Connecticut misconduct investigation, 2012; Retraction Watch and mainstream reporting (145 counts, 20+ retractions). Contaminated evidence base. Independent institutional finding; Das disputed it but was removed as a journal co-editor-in-chief. Note: fraud taints that slice of the literature but does not by itself disprove the mechanism — keep the two separate.)

Mechanism

This entry owns the preclinical-signal-versus-human-claim split and the dose/bioavailability reality, not the general surrogate-versus-outcome principle, which is deferred to surrogate_endpoints_vs_outcomes. What follows is only enough mechanism to make the signal and its ceiling intelligible.

Why the preclinical signal is real. Resveratrol is a polyphenol that, in simple organisms, activates the sirtuin Sir2 and triggers a stress-response program overlapping with the effects of caloric restriction — increased insulin sensitivity, activation of AMPK and PGC-1α, and mitochondrial biogenesis. In yeast, worms and flies this lengthens lifespan; in overfed mice it shifts a metabolically overloaded system back toward the physiology of a lean one. This is a coherent pathway, which is exactly why the molecule is worth taking seriously rather than dismissing.

Why the signal caps out at "sick or overloaded systems." The benefit tracks a metabolic deficit or stress, not health. In overfed mice, resveratrol relieved the overload; in mice on a normal diet, it did not extend lifespan. The mechanism is a rescue of a stressed system, not a general accelerant of longevity in an already-healthy one — the same reason its human surrogate effects concentrate in metabolically stressed subgroups and vanish in healthy ones.

Why "activates a pathway" is not "extends human healthspan." Sirtuin activation is a molecular event; living longer is a clinical outcome. The link between them was never demonstrated in people, and independent labs then argued that resveratrol does not even directly activate SIRT1 — the original readout was an artifact of the fluorescent assay used to measure it. A mechanism that may not be the mechanism, and that has never been tied to a human outcome, cannot license a healthspan claim. This is the load-bearing split: mechanism and biomarker real (or at least plausible), human outcome unproven.

Why the dose and absorption bury the wine story. Wine carries only a few milligrams of resveratrol per litre, so reaching the concentrations used in animals would take an impossible volume of wine. And even a concentrated oral dose is largely wasted: resveratrol is rapidly conjugated to sulfates and glucuronides on first pass, leaving systemic bioavailability of the free compound around half a percent. So the two ways people imagine "getting" resveratrol — drinking red wine or swallowing a supplement — are defeated respectively by the dose in wine and by first-pass metabolism.

Risks And Contraindications

• The main risk is opportunity cost and a false longevity promise, not acute toxicity at typical supplement doses. For most people the realistic harm is money spent on a poorly-absorbed compound with no demonstrated outcome, plus the false confidence of "doing something" for aging that the evidence does not support.
• Do not let "red wine is protective" become a reason to drink. This is the most consequential real-world harm around resveratrol. The wine dose cannot be the mechanism, and the alcohol itself is a separate, non-trivial risk — defer entirely to alcohol_cross_pillar_effects. Never let this entry read as an endorsement of drinking.
• High-dose safety is not established — treat megadosing cautiously. As a general principle for a novel longevity compound, unproven long-term safety is itself a caveat: chasing animal-equivalent doses via concentrated supplements moves you into territory human trials have not cleared. The Sirtris SRT501 program was halted after cases of renal failure and poor tolerability at therapeutic dosing — a concrete signal that "more" is not automatically safe. (By analogy with other longevity-marketed compounds, megadosing a supplement in pursuit of a mechanism can carry its own harms — as with vitamin C megadoses and kidney-stone risk — so "the animal dose" is not a target to reverse-engineer at home.)
• Do not overstate the preclinical signal into a human claim. The lifespan data are animal-only and diet-conditional; presenting sirtuin activation or a mouse survival curve as if it were a human outcome is exactly the error this entry exists to prevent.
• Keep the fraud separate from the mechanism. The Das retractions taint a slice of the cardioprotection literature but do not by themselves disprove the biology. Do not use "there was fraud" as a shortcut to "resveratrol does nothing" — the honest disconfirmation is the diet-conditional cap, the bioavailability barrier, the null cohort, and the failed program, not the misconduct alone.
• Avoid the opposite over-correction. "It's a total scam / does nothing" ignores a real, replicated preclinical signal and the modest surrogate effects. The honest message is "real mechanism, unproven human outcome."

Controversy

Nature: a molecule with a genuine, replicated preclinical signal wrapped in a human longevity and "red wine is heart-healthy" pitch that the evidence does not support — with error possible at both poles: over-claiming resveratrol as a proven human intervention on one side, and dismissing the real biology as nothing on the other.

Position A — "Resveratrol has a genuine preclinical signal." The real-biology take.
• Best evidence: Sir2-dependent lifespan extension in yeast, worms and flies; improved survival and a lean-shifted physiology in overfed mice (Baur 2006); a coherent mechanism (insulin sensitivity, AMPK/PGC-1α, mitochondrial biogenesis). This is real biology.
• Where it goes wrong if overstated: it slides into "so a supplement extends human healthspan," ignoring that the mouse benefit is diet-conditional, that human bioavailability is around half a percent, and that no human outcome has ever been shown.

Position B — "The human longevity/cardiovascular claim is unproven-to-null." The debunk take.
• Best evidence: bioavailability around 0.5%; the wine dose is trivial (hundreds to a thousand glasses to reach animal doses); human RCTs move only surrogate markers inconsistently with no hard outcome; a large cohort found urinary resveratrol unrelated to mortality/CVD/cancer; even in mice the lifespan gain vanished on a normal diet; the flagship GSK/Sirtris program failed; and part of the early cardioprotection literature was fraud.
• Where it goes wrong if overstated: it can tip into "resveratrol does nothing," which ignores the replicated preclinical signal and the modest surrogate-marker effects in stressed subgroups.

The funding/bias dimension — cui bono, both ways. Toward over-claiming, the pressure is heavy: the supplement industry (resveratrol and pterostilbene sellers), the wine industry and its trade press (French-paradox marketing), and researchers with equity in sirtuin-activator ventures (Sinclair and Sirtris; GSK's $720M bet) all profited from the longevity narrative, and the null InCHIANTI cohort drew a public attack from a supplement-company figure. Toward the corrective pole, the receipts also carry interest: Amgen and Pfizer, who published the "not a direct SIRT1 activator" assay-artifact critique, were commercial competitors to Sirtris. But the assay critique was mechanistically substantiated, and the strongest disconfirming evidence — the diet-conditional lifespan, the poor bioavailability, the null human cohort, the failed GSK program — comes from independent NIA and academic sources with no seller interest.

Realised Position: Both are true, and that is the point. A real mechanism in simple organisms and overfed mice does not license a human healthspan claim. Resveratrol activates a plausible pathway and helps sick, overloaded mice, but in humans the dose you can realistically get is trivial, what little reaches the bloodstream is mostly conjugated metabolites, RCTs move surrogate markers inconsistently at best, and no trial or cohort shows people live longer or have fewer cardiovascular events. The red-wine "French paradox" version specifically does not survive the dose arithmetic, and the alcohol question is separate. Treat resveratrol as an Emerging mechanism, not a proven intervention — and effectively null on the human longevity claim.

Cross-Pillar Connections

Resveratrol is a diet/supplement topic whose claim sits on longevity and cardiovascular framing, so its connections span the evidence-method and cardiovascular lines.
• Foundations (surrogate_endpoints_vs_outcomes): owns the general principle that moving a biomarker is not the same as improving a hard outcome; this entry holds only the resveratrol-specific instance (surrogate-marker RCTs, no outcome) and defers the framework there rather than re-deriving it.
• Cross-pillar (alcohol_cross_pillar_effects): owns the alcohol question entirely; this entry holds only that the resveratrol dose in wine is trivial and cannot be the mechanism, and explicitly does not treat "red wine" as an endorsement or condemnation of drinking.
• Conditions (cardiovascular_health_management): owns actual cardiovascular management; this entry holds only that resveratrol has shown no proven cardiovascular outcome and that part of the early cardioprotection literature was retracted for fraud.
• Foundations (publication_bias_and_evidence_distortion): the mechanism behind the selective, positive-surrogate publication pattern and the single-sponsored-program inflation that made resveratrol's evidence look stronger than it is.
• Foundations (healthy_user_bias): why observational "wine drinkers / resveratrol users are healthier" impressions do not establish a supplement or wine benefit — directly relevant to reading the French-paradox claim critically.

What would change our mind

Falsifiability: explicit upgrade/downgrade criteria from source

• We'd upgrade toward a real human claim on a randomised controlled trial powered for a HARD clinical outcome — all-cause mortality, major adverse cardiovascular events, or a validated healthspan endpoint — showing benefit from a bioavailable resveratrol formulation. Not another surrogate-marker study (blood pressure, glucose, NAD+): a biomarker win would not move this.
• We'd also revisit on a positive outcome trial of a next-generation, verified SIRT1-selective activator that resolves both the bioavailability barrier and the assay-artifact problem — showing the pathway delivers a human outcome even if resveratrol itself is a poor drug.
• We'd revise the wine debunk only if evidence showed a plausible resveratrol dose is achievable and clinically effective at real-world wine intake — which the arithmetic currently rules out.
• What would NOT move us: more animal lifespan data, another surrogate-marker RCT, higher blood levels from a fancy formulation without an outcome, or sirtuin activation in vitro. The gap is specifically mechanism-to-human-outcome, and only outcome data closes it. Absent that, the entry stays Emerging for the mechanism and effectively Experimental/null for the human longevity claim.

Industry bias note

Structural incentives the evidence base may reflect

Cui bono runs both ways, but the inflating pressure is heavily on the pro-longevity side, and the strongest disconfirming evidence is conflict-clean.
• The supplement and longevity-influencer economy is the primary seller here — name it. Resveratrol and pterostilbene sellers, and the broader "take this to live longer" longevity-supplement market, profit directly from the healthspan narrative. This is the dominant bias vector for the whole claim, and it is exactly the claim the human evidence does not support.
• The wine industry benefits from the "French paradox." The narrative that red wine is heart-protective "because of resveratrol" sells wine and generates favourable trade-press coverage. The dose arithmetic kills the mechanism, so this framing survives on marketing, not evidence.
• Researcher equity in sirtuin-activator ventures. David Sinclair co-founded Sirtris and had a commercial stake in the sirtuin-activator story; GSK then made a $720M bet on it. Pro-resveratrol findings from that lineage carry a structural conflict to weigh — and, tellingly, the program failed anyway.
• The counter-incentive is real but weaker, and substantiated. Amgen and Pfizer, who published the "not a direct SIRT1 activator" assay-artifact critique, were commercial competitors to Sirtris — cui bono against it. But the artifact critique was mechanistically demonstrated (a fluorophore effect), and GSK's own program corroborated poor drug-likeness. Present it as substantiated while acknowledging the competitive context.
• The load-bearing counter-claims serve no seller. The diet-conditional lifespan cap (NIA), the ~0.5% bioavailability (independent PK literature), and the null human cohort (NIA-funded InCHIANTI) come from independent academic and government sources with no product to sell — which is precisely why they are trustworthy. A supplement-company figure publicly attacked the null cohort, which is the pro-supplement pushback pattern, not a flaw in the study. Net: the pro-longevity claim is the one most inflated by commercial incentive (see healthy_user_bias, publication_bias_and_evidence_distortion).

Resveratrol: Real in a Petri Dish, Unproven in People