By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Peter Attia MD270 ‒ Journal club with Andrew Huberman: metformin, power of belief, & how to read scientific papers
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
At a glance
WHAT IT’S REALLY ABOUT
Attia and Huberman dissect metformin longevity claims and belief effects in nicotine
- Attia presents a 2022 Danish registry study that redoes the 2014 Bannister analysis and finds metformin-treated type 2 diabetics still die at higher rates than non-diabetic controls, undercutting the idea that metformin clearly offers a survival advantage.
- The discussion explains how epidemiology works, covering hazard ratios, informative censoring, person-years, Kaplan-Meier curves, confidence intervals and statistical power, and why unmatched medications and other confounders limit any observational study.
- Attia explains why he stopped taking metformin for geroprotection, citing elevated resting lactate, the drug's failure in the Interventions Testing Program, and the lack of biomarkers for aging, and says he would reconsider if the TAME trial shows benefit.
- Huberman presents a preprint in which smokers given identical low-nicotine vapes showed brain responses that scaled with the dose they were told they received, suggesting belief effects act on physiology and not just subjective experience.
- Both describe their paper-reading routines, which start with the question, the approach, the findings and whether the conclusions are actually supported, and they urge reading papers repeatedly and checking for missing controls and low power.
IDEAS WORTH REMEMBERING
5 ideasThe Danish replication did not reproduce the survival benefit seen in the 2014 Bannister study.
Metformin-treated diabetics still had hazard ratios above 1 versus non-diabetic controls, even with censoring, and Attia says this fits what one would have expected before the Bannister paper.
Observational studies can never fully remove confounding.
The groups were closely matched on most traits but not on medications, such as 45.6% versus 15.4% on lipid-lowering drugs, and Attia notes that no model can correct for that fully.
The study cannot say whether metformin helps diabetics relative to no treatment.
Attia stresses there is no group of diabetics on nothing, so metformin might still lower their risk, and only a randomized trial such as TAME can settle the geroprotection question.
A non-significant result means little without knowing what the study was powered to detect.
Attia explains that sample size is set from the expected effect size, and this later helps him question the nicotine paper's small sample of roughly 20 per group and its weak correlation of 0.27 at the thalamus.
Attia stopped metformin for personal and evidence-based reasons.
He found a resting lactate around 1.6 mmol, saw metformin fail in the ITP, and noted possible attenuation of strength and hypertrophy, though he still prescribes it for insulin-resistant patients.
WORDS WORTH SAVING
5 quotesyou can never remove all the confounders.
— Peter Attia
in general, it just tends to be a signal-to-noise enhancer.
— Andrew Huberman
if I tell you, "This is a high dose of nicotine," your brain will react as if it's a high dose of nicotine.
— Andrew Huberman
we just have to, I think, all walk around with a, an appropriate degree of humility around what we know and what we don't know.
— Peter Attia
People aren't deluding themselves. The thalamus is behaving as if it's a high dose when it's the same dose as the low dose group. It's wild.
— Andrew Huberman
High quality AI-generated summary created from speaker-labeled transcript.
