Study
Comprehensive investigation of resveratrol as an anti-aging supplement in randomised clinical trials
The largest human evidence base of any longevity molecule: 92 of 137 RCTs positive, 42 null, bioavailability the constant limit.
Summary Does resveratrol improve health or slow aging in people? Show / hide ↓
Researchers reviewed 137 published randomized controlled trials, meaning studies that assigned people to resveratrol or a comparison treatment by chance. The trials included 104 separate studies and more than 4,800 adults. About 92 reports found an improvement in at least one main health measure, while 42 found no meaningful benefit. Resveratrol was absorbed and changed quickly by the body, so blood levels were usually much lower than the levels that worked in laboratory animals and cells.
What this means for you: Resveratrol may help some heart or metabolism measures, but this review does not show that it slows aging or extends life. The mixed results and poor absorption make it hard to recommend buying it for general longevity.
conflicting evidenceA systematic review pulled every published randomised controlled trial on pure resveratrol and found 137 studies covering 104 unique RCTs and more than 4,800 people. 92 studies (67 percent) reported a positive effect on their primary endpoint. 42 studies (31 percent) reported no effect. The recurring limit in almost every trial is oral bioavailability. Resveratrol is conjugated so fast by phase-II liver enzymes that plasma levels rarely reach the concentrations that drove the mouse and yeast findings. This is the single largest human RCT count for any molecule in our database. It does not settle whether resveratrol slows aging in people. It does show the molecule has real, dose-dependent effects on some cardiovascular and metabolic endpoints, and near-zero effect on others.
Of 137 RCTs identified, 92 (67 percent) showed a positive effect on the primary endpoint, and 42 (31 percent) showed none; poor bioavailability is the consistent limitation.
Positive-endpoint counts are not effect sizes. Publication bias may inflate the 67 percent figure. Endpoints varied from lipid panels to endothelial function to inflammatory markers.
Cited by
1 encyclopedia entries reference this study.