Journal Club With Dr. Peter Attia | Metformin for Longevity & the Power of Belief Effects

Episode Summary

Executive Summary: Andrew Huberman and Peter Attia conduct a journal club on two papers: one re-evaluating metformin’s supposed longevity benefit and another showing that beliefs about nicotine dose can scale brain and behavioral responses. They emphasize how to read papers critically, the limits of observational data, and how placebo/belief effects can alter physiology, not just subjective experience.

Main Topics: Metformin and longevity claims (Priority: 5/5): Attia reviews a retrospective cohort study that re-examines the famous Bannister metformin paper and argues the newer data do not support a clear geroprotective benefit in humans, especially given confounding and censoring issues. How to interpret epidemiology vs randomized trials (Priority: 5/5): The discussion contrasts observational cohort studies with randomized controlled trials, highlighting confounding, informative censoring, hazard ratios, and why large sample sizes do not eliminate bias. Biochemistry and clinical use of metformin (Priority: 4/5): Attia explains metformin’s mitochondrial complex I inhibition, reduced hepatic glucose output, and its role in type 2 diabetes management, while noting side effects and uncertainty about off-target longevity mechanisms. Belief effects and placebo beyond binary placebo (Priority: 5/5): Huberman distinguishes placebo effects from broader belief effects, arguing that expectations can change physiology in graded, dose-like ways rather than simply on/off. Nicotine dose belief study (Priority: 5/5): Huberman presents a preprint showing that telling smokers they received low, medium, or high nicotine altered thalamic and thalamo-prefrontal activation in a dose-dependent manner, despite identical nicotine exposure. Practical paper-reading framework (Priority: 4/5): Both hosts explain how they read papers: identify the question, methods, findings, conclusion, and then test whether the conclusion is actually supported by the data.

Key Arguments: The original metformin longevity excitement came largely from observational data, but the newer Danish cohort/twin analysis suggests the apparent benefit may have been overstated by selection bias and informative censoring. Metformin clearly helps manage type 2 diabetes by lowering hepatic glucose output, but that does not prove it extends lifespan in non-diabetics. Even very large observational datasets cannot fully control for confounders such as medication burden, health status, and lifestyle differences. Randomized trials remain necessary to determine whether metformin is truly geroprotective; the TAME trial is positioned to answer this more directly. Beliefs about a drug’s dose can alter brain activity in a graded fashion, meaning physiology can track expectation, not just pharmacology. The nicotine study suggests the thalamus and thalamo-prefrontal circuitry are especially sensitive to belief about dose, even when actual nicotine exposure is held constant. Placebo/belief effects are not merely subjective reports; they can influence measurable biological signals and task performance.

Data Points: Metformin crude death rate (Bannister study): 14.4 deaths per 1,000 patient-years - Type 2 diabetics on metformin in the earlier UK registry analysis Control crude death rate (Bannister study): 15.2 deaths per 1,000 patient-years - Matched controls without diabetes in the earlier UK registry analysis All-cause mortality reduction (Bannister study): 15% relative reduction over 2.8 years - Metformin group vs matched controls in the earlier observational study Type 2 diabetes lifespan impact: ~6 years shorter life expectancy - Attia’s summary of actuarial impact of type 2 diabetes Singleton crude mortality without diabetes: 16.86 deaths per 1,000 person-years - Keyes et al. Danish registry matched controls without diabetes Singleton crude mortality with metformin/diabetes: 24.93 deaths per 1,000 person-years - Keyes et al. Danish registry metformin-treated diabetics Twin crude mortality without diabetes: 12.94 deaths per 1,000 person-years - Discordant twin co-twins without diabetes Twin crude mortality with metformin/diabetes: 24.73 deaths per 1,000 person-years - Discordant twins with diabetes on metformin Unadjusted hazard ratio, singletons: 1.48 - Metformin/diabetes group vs non-diabetic matched controls in Keyes et al. Adjusted hazard ratio, singletons: 1.32-1.33 - After adjusting for medications, marital status, and education in Keyes et al. Unadjusted hazard ratio, twins: 2.15 - Discordant twin analysis in Keyes et al. Adjusted hazard ratio, twins: 1.70-1.80 - After covariate adjustment in Keyes et al. Censoring effect, singletons: 1.48 to 1.39 - Sensitivity analysis including informative censoring in Keyes et al. Censoring effect, twins: 2.15 to 1.97 - Sensitivity analysis including informative censoring in Keyes et al. Sample size, Keyes study: ~500,000 people sampled - Danish health registry cohort used for the reanalysis Metformin singleton sample: 7,842 - Matched diabetic metformin users in Keyes et al. Matched singleton controls: 7,842 - Non-diabetic matched controls in Keyes et al. Metformin twin sample: 976 - Discordant twin pairs where one twin had diabetes on metformin Belief study sample size estimate: N=20 per belief condition for 90% power - Authors’ power estimate for the nicotine belief experiment Exercise-belief weight loss effect: 12% more weight loss - Huberman cites prior belief-effect work in hotel workers

Pivotal Quotes: "This makes much more sense to me than the Bannister paper, which never really made sense to me." — Peter Atiyah: Attia’s bottom-line interpretation of the newer metformin reanalysis "What we believe about the effects of a drug, presumably, in addition to what we believe about how much we're taking and what those effects ought to be, clearly are impacting at least the way that our brain reacts to those drugs." — Andrew Huberman: Huberman’s synthesis of the nicotine belief-dose study "The thalamus is behaving as if it's a high dose when it's the same dose as the low dose group." — Andrew Huberman: Summary of the main finding from the nicotine preprint

Implications: Listeners should be cautious about assuming observational longevity claims prove causation. The episode reinforces that randomized trials are needed for metformin as an anti-aging drug, and that expectations can measurably shape brain and body responses to medications.