Rapamycin Matches Dietary Restriction for Lifespan Extension — Metformin Falls Short

Dietary restriction (DR) is one of the most reproducible interventions for extending lifespan across diverse species, but its long-term adherence in humans is poor. This has driven interest in pharmacological DR mimetics — compounds that activate similar metabolic pathways without requiring caloric reduction. Two leading candidates are rapamycin (an mTOR inhibitor used clinically as an immunosuppressant) and metformin (an AMPK activator used to treat type 2 diabetes). Despite decades of research, the comparative efficacy of these compounds relative to DR in vertebrates has remained unclear.

To resolve this, Ivimey-Cook, Sultanova, and Maklakov performed a systematic review and multi-level meta-analysis following PRISMA guidelines. They searched Scopus and Web of Science (July 2023, updated December 2024), ultimately analyzing 911 effect sizes from 167 papers spanning eight vertebrate species. Effect sizes were calculated as log-response ratios of mean and median lifespan, adjusted for small-sample bias. Multi-level models accounted for non-independence across species, papers, and observations. Moderators tested included treatment type (DR, rapamycin, metformin), sex, and DR methodology (percent caloric reduction vs. fasting).

The headline finding is clear: DR robustly extended lifespan across both mean and median log-response measures. Rapamycin also produced a statistically significant lifespan extension, with an effect size comparable in magnitude to DR. Metformin, by contrast, did not produce a significant lifespan extension across vertebrates. Sex was not a consistent moderator of lifespan outcomes for any treatment. Among DR studies, the specific methodology — whether animals underwent caloric reduction or fasting — did not significantly alter the overall effect, suggesting DR's lifespan benefits are robust to implementation differences.

Importantly, the authors identified substantial heterogeneity across all treatment groups and significant publication bias (both small-study and time-lag effects), which influenced the magnitude of estimated effects. Results were also sensitive to whether lifespan was reported as mean or median, though qualitative conclusions remained consistent. Only 0.5% of effect sizes failed normality testing, supporting the validity of the log-response ratio approach.

These findings have meaningful implications for translational longevity research. Rapamycin emerges as the pharmacological intervention most closely mirroring DR's lifespan-extending effects in vertebrates, strengthening the case for its continued investigation as a human longevity drug. Metformin's lack of significant effect in this vertebrate-focused analysis contrasts with some prior invertebrate findings and underscores the importance of species context. The authors caution that high heterogeneity and publication bias remain important limitations, and that human translation requires further clinical evidence.