Journal Club With Dr. Peter Attia | Metformin for Longevity & the Power of Belief Effects
Dr. Peter Attia, a physician trained at Stanford and Johns Hopkins, shares his insights as the host of The Drive and author of 'Outlive.' He and the host dive deep into the potential of metformin for enhancing longevity and its fascinating ties to diabetes management. They explore the influence of belief on health outcomes, revealing how mindset can alter drug effects. The conversation emphasizes understanding scientific literature, the role of glucose monitoring, and the delicate balance between lifestyle choices and health impacts.
Top snips
Metformin 101 — what it does
13:40 – 18:35
Metformin = long‑used first‑line diabetes drug.
- Main action: reduces hepatic glucose output.
- Mechanism: weak inhibitor of mitochondrial complex I → shifts AMP/ADP ratios.
Used for decades (trade name Glucophage). It’s cheap, generic, and often discussed as a potential geroprotector.
Metformin is a weak inhibitor of mTOR.
— Peter Attia
The 2014 Bannister finding that changed minds
29:33 – 30:10
Bannister et al. (2014) compared metformin‑treated diabetics vs matched non‑diabetics and reported a 15% lower all‑cause mortality in the metformin group.
This paper sparked huge interest in metformin as an anti‑aging drug — but later scrutiny raised questions about study methods (e.g., censoring).
It got the world very excited about the potential for geroprotection.
— Peter Attia
New analysis: Danish cohort + discordant twins 🇩🇰
40:41 – 41:37
Keys et al. re‑ran the question using ~500k from Danish registries and a clever discordant same‑sex twin analysis.
Design highlights:
- singletons matched diabetic vs non‑diabetic
- twin pairs where one twin had diabetes (on metformin) and the co‑twin did not
Goal: reduce confounding (genetics, early environment).
Keyes results: metformin didn’t beat diabetes
1:02:23 – 1:08:10
Keyes et al. found that diabetics on metformin had higher crude mortality vs non‑diabetics. After adjustments the hazard ratios still showed increased risk (e.g., HR ~1.32 in singletons). Censoring changed numbers but didn’t reproduce Bannister’s protective signal.
Bottom line: epidemiology can’t fully settle causation.
Peter’s personal take — stopped metformin (2018)
1:16:03 – 1:17:44
Peter stopped metformin in 2018. Why? He measured elevated resting lactate on metformin → suggests mitochondrial impact that raised baseline anaerobic metabolism.
Clinical view: metformin remains great for insulin resistance/diabetes, but for athletes or highly active people, effects on performance and lactate warrant caution.
I stopped taking metformin.
— Peter Attia
Belief effects — study setup with nicotine MRI
1:51:08 – 1:57:37
New human study tested belief dose vs actual nicotine dose. Smokers vaped the same low nicotine dose but were told they’d received low / medium / high.
While in fMRI they played a reward/market task to engage thalamus → vmPFC → reward circuits. Measures: blood nicotine, subjective belief, BOLD activation.
Mind changes the brain — dose‑dependent belief
1:58:33 – 2:01:15
Key finding: the brain responded in a dose‑dependent way to what subjects were told, not the actual nicotine amount.
- Thalamic activation scaled low→medium→high belief.
- Connectivity thalamus → vmPFC also scaled with belief.
Implication: what you believe about dose can change real neural response.