Resveratrol and sirtuins: what the human evidence shows against the hype

Resveratrol, the polyphenol in red wine marketed for decades as an anti-ageing molecule, does extend lifespan in simple organisms and improves survival in mice fed a high-calorie diet [s1][s2]. In humans the case has not held: the compound does not reliably activate the sirtuin enzyme its whole rationale rests on [s3], and among older adults higher resveratrol intake is associated with no reduction in death, inflammation, cancer or cardiovascular disease [s4].

The idea was genuinely exciting when it appeared. In 2003 a landmark paper reported that resveratrol activates sirtuins — NAD+-dependent enzymes linked to the lifespan-extending effects of calorie restriction — lowering the enzyme's Michaelis constant and, in budding yeast, mimicking calorie restriction to extend lifespan by 70% [s1]. Three years later, resveratrol was shown to shift middle-aged mice on a high-calorie diet toward the physiology of mice on a standard diet and to significantly increase their survival, with lower insulin-like growth factor-1, improved insulin sensitivity and gene-expression changes that opposed the high-calorie diet in 144 of 153 altered pathways [s2]. That is a real result — in mice, on a bad diet.

The mechanism turned out to be an artefact

The trouble is what happened when the founding biochemistry was re-examined. The assay used to show resveratrol "activating" SIRT1 relied on a peptide substrate carrying an attached fluorophore. In 2010 researchers tested the compound against native substrates lacking that fluorophore — including the full-length p53 protein — and found that resveratrol, along with the pharmaceutical activators SRT1720, SRT2183 and SRT1460, did not activate SIRT1 [s3]. The apparent activation appeared only with the fluorophore-tagged peptide, and the compounds were shown to bind that tag directly [s3]. The same paper reported that SRT1720 neither lowered plasma glucose nor improved mitochondrial capacity in mice on a high-fat diet [s3]. The molecule may still do useful things through other routes, but the clean "sirtuin activator" story that launched the field was built on a measurement artefact.

In people, the numbers are flat

The most telling human data come not from a supplement trial but from watching what dietary resveratrol actually does. The InCHIANTI study followed 783 community-dwelling men and women aged 65 or older in two Chianti villages and measured 24-hour urinary resveratrol metabolites — an objective marker of intake from wine, grapes and other sources [s4]. Over nine years, 268 participants (34.3%) died [s4]. Across quartiles of baseline resveratrol, from lowest to highest, the proportion who died was 34.4%, 31.6%, 33.5% and 37.4% — no trend, with P = .67 [s4]. Those in the lowest quartile had a hazard ratio for mortality of 0.80 (95% CI, 0.54–1.17) versus the highest, an estimate that includes no effect in either direction [s4]. Resveratrol levels were likewise not associated with markers of inflammation, or with rates of cancer or cardiovascular disease [s4].

The caveat runs both ways: dietary levels are far below the doses used in animal studies, so InCHIANTI does not test high-dose supplements. But that is precisely the point. The lifespan data everyone cites come from yeast, worms, flies and mice; the human evidence that exists shows no benefit at achievable intakes; and the biochemical bridge between them — direct sirtuin activation — did not survive scrutiny [s1][s2][s3][s4].

The pattern is familiar

This is the same shape as the rest of the anti-ageing supplement shelf: a plausible mechanism from model organisms, a striking mouse result, and a human record that is null or untested at the marketed dose. It mirrors what the trials show for NAD+ precursors such as NMN and NR, which act on the same sirtuin-adjacent pathway, and for spermidine. The intervention that reliably reproduces the underlying biology remains calorie restriction itself — hard to sell in a capsule, and the thing resveratrol was only ever meant to imitate.

This article describes research findings and is not medical advice.