Should Healthy People Take Metformin for Longevity?

I recently analyzed all the available RCTs on metformin use in healthy individuals and uncovered the following facts: 1) It shows zero improvement in insulin sensitivity; 2) It actually increases fasting blood glucose; 3) It blunts the health benefits of exercise. Based on these findings, it is reasonable to infer that metformin might actually increase the risk of new-onset diabetes in healthy people.

Furthermore, a post-hoc analysis of a large-scale diabetes prevention RCT confirms this suspicion. Even among non-diabetic individuals with impaired glucose tolerance, the lowest-risk subgroup on metformin saw a higher diabetes incidence rate (9.6%) compared to the placebo group (8.3%). While this specific trend did not reach statistical significance, the negative trajectory is clear. Conversely, the same low-risk subgroup assigned to lifestyle intervention achieved a drastically lower incidence rate (3.4%) than the placebo group (8.3%), with strong statistical significance.

Based on this, the debate over whether healthy people should take metformin is officially over.

Risk Stratification Range of Predicted Probability Observed Rate in Placebo Group (Events) Observed Rate in Metformin Group (HR) Observed Rate in Lifestyle Group (HR)
Quarter 1 (Lowest Risk) 1.1% - 9.5% 8.3% (19) 9.6% (HR: 1.07, 95% CI: 0.57 to 2.01) 3.4% (HR: 0.30, 95% CI: 0.12 to 0.75)
Quarter 2 (Medium-Low Risk) 9.5% - 15.1% 17.8% (37) 14.9% (HR: 0.79, 95% CI: 0.49 to 1.28) 8.6% (HR: 0.45, 95% CI: 0.26 to 0.79)
Quarter 3 (Medium-High Risk) 15.1% - 27.0% 29.1% (73) 24.4% (HR: 0.82, 95% CI: 0.57 to 1.18) 15.5% (HR: 0.43, 95% CI: 0.28 to 0.67)
Quarter 4 (Highest Risk) 27.0% - 99.8% 59.6% (140) 38.2% (HR: 0.44, 95% CI: 0.33 to 0.59) 31.3% (HR: 0.34, 95% CI: 0.25 to 0.46)

Improving diabetes prevention with benefit based tailored treatment: risk based reanalysis of Diabetes Prevention Program

1 Like

To be honest, conducting this kind of analysis is quite tedious. You have to include as many RCTs as possible, including non-English papers and paywalled studies. Although I started this analysis hoping to find that metformin would benefit healthy people, unfortunately, the data clearly shows it is not suitable for individuals with completely normal metabolism.

Of course, the diabetes-related metrics I mentioned are just from one subgroup, which I used simply as a straightforward example to illustrate why it shouldn’t be taken. After comprehensively weighing all subgroups across various biological markers, I firmly believe there is absolutely no necessity for healthy individuals to take metformin. While I did uncover some fascinating insights during the process, they unfortunately didn’t change the ultimate conclusion—and this holds true for intermittent dosing as well.

A summary of the paper:

Metformin Efficacy Evaporates in Low-Risk Pre-Diabetes: The Case for Precision Prevention

A post-hoc risk-stratified reanalysis of the landmark Diabetes Prevention Program clinical trial demonstrates that the preventative benefits of metformin are highly non-linear and concentrated almost exclusively within individuals at the highest quadrant of baseline diabetes risk. By applying an internal multivariable risk model to the trial data, researchers found that the bottom 75 percent of participants derived minimal to zero statistical benefit from metformin, with the lowest-risk quarter even showing a minor trend toward accelerated progression. In contrast, intensive lifestyle interventions yielded a powerful, consistent relative risk reduction across all risk tiers, signaling that universal pharmaceutical adoption for longevity or mild metabolic adjustments may be misguided.

For years, the longevity and biohacking communities have viewed metformin as a foundational therapeutic for metabolic optimization. This reputation was largely built on major clinical trials like the Diabetes Prevention Program, which reported that the drug slashed the average risk of progressing to type 2 diabetes by 31 percent. However, a critical reanalysis of the trial data published in The BMJ reveals that average statistics can mask a starkly different reality for the individual. The big idea driving this study is benefit-based tailored treatment, a paradigm shift showing that a patient’s baseline risk dictates how much absolute benefit they will actually receive from an intervention.

To uncover this hidden variation, researchers built a multivariable risk prediction model using baseline physiological data from over 3,000 participants. They factored in seventeen variables, including fasting blood sugar, hemoglobin A1c, family history, and waist measurements, to divide the cohort into four distinct quarters of ascending baseline risk.

The findings upend the conventional approach to preventative medicine. For participants placed in the highest-risk quarter, metformin was profoundly effective, delivering a dramatic drop in diabetes incidence over the 2.8-year tracking period. Yet, for the remaining 75 percent of the cohort, the clinical utility of the drug dropped off a cliff. In the second and third risk quarters, the benefits were marginal and statistically uncertain. Most revealingly, the individuals in the lowest-risk quarter experienced zero benefit from metformin. In fact, those taking the drug in this sub-group had a slightly higher rate of developing diabetes than those taking a placebo, though this trend did not achieve statistical significance.

Conversely, the study brought highly positive news for non-pharmaceutical interventions. The structured lifestyle modification program, which focused on modest weight loss and routine physical activity, achieved a uniform 58 percent relative risk reduction across all four quarters. While the absolute number of prevented cases was naturally higher in the top risk tier because they had more risk to mitigate, even the lowest-risk individuals gained substantial protection from lifestyle changes. This dichotomy indicates that while the human body universally responds to physical optimization, its response to targeted pharmaceutical enzymes depends entirely on the severity of the underlying pathology. For optimal health and longevity, automatic drug prescriptions based on a single borderline lab value should be retired in favor of sophisticated risk stratifications.

Context/Source

  • Paper Title: Improving diabetes prevention with benefit based tailored treatment: risk based reanalysis of Diabetes Prevention Program
  • Access Status: Open Access (Distributed under CC BY-NC 4.0)
  • Lead Institutions: Department of Veterans Affairs Center for Clinical Management Research, University of Michigan, and Tufts Medical Center
  • Country: United States
  • Journal Name: BMJ (The BMJ)
  • Impact Evaluation: The impact score of this journal is 93.6, evaluated against a typical high-end range of 0–60+ for top general science, therefore this is an Elite impact journal.
1 Like

Furthermore, I want to avoid over-extrapolating on whether combining metformin with other drugs would alter its trajectory. Studies on multi-drug combinations in completely healthy individuals are incredibly scarce, let alone for biohackers who typically run complex stacks of prescription drugs and supplements.

This is exactly why frequent testing is so critical. In my opinion, the most vital investment for a biohacker isn’t the drugs, supplements, or health gadgets—it’s frequent, comprehensive lab work. Even then, it’s far from perfect. For instance, something like cancer risk is notoriously difficult to spot through routine blood markers.

Anyway, just a bit of rambling on my end. But as it stands, if you don’t overcomplicate things, the simple conclusion on metformin is just as I described.

All I can say is I find myself hypoglycemic on both Telmisartan and Empagliflozin. I am unsure if most of the longevity community are just older folks in poor health that can actually take these.

Why did I bring up cancer risk? Because I’ve seen multiple cases where drugs that reduce cancer risk individually end up drastically increasing it when combined, especially when accounting for variables like sex, age, and ethnicity. Many of these findings are buried in obscure, paywalled papers—and the authors don’t even mention it in the main text; you have to run your own analysis to spot it.

To be honest, it wouldn’t surprise me at all if many biohackers fail to outlive the average US life expectancy. The field is just riddled with hidden pitfalls. I’d suggest not overthinking it, as it only breeds unnecessary anxiety. At some point, you just have to leave it up to fate.

Anti-aging is incredibly difficult for the youth, though seniors might have it a bit easier. The evidence we rely on is just too imperfect. If you blindly mess around with these interventions, you might honestly end up worse off than someone who simply exercises regularly and eats a clean diet.

My best advice for biohackers is to streamline your stack as much as possible—keep it as simple as possible, based strictly on RCTs that offer definitive conclusions. There’s no need to gamble with a biased mindset. If blindly stacking interventions actually worked, think about all the people out there with poor health who are already taking massive cocktails of different prescription drugs. Over time, that population must have exhausted almost every possible drug combination, much like a brute-force attack cracking a password. Yet, we haven’t seen a sudden surge of people breaking the 120-year age barrier. And that’s that.

1 Like

Thank you for posting this. I am (I think) right on the borderline re: should I be taking metformin. I have been taking metformin for about 6 years. Fasting blood glucose ranged from 90 to 100 before metformin and now is still about the same – but – it had gone up when I started Repatha, at which time I increased from 500 to 1000 mg metformin. HA1C has ranged from about 5.6 to 5.9. I believe I am metabolically healthy, with low fasting insulin, BMI of 19.

So I am just below the official level of “prediabetes” . I mean to stay on Repatha, so if I quit Metformin, glucose will rise. But, I will not suffer the blunting-of-exercise effects. As a twee 76 year old with severe osteoporosis, I need all the muscle I can pack on (and am working hard on this and all the other lifestyle stuff).

Would appreciate views, suggestions.

Should also note:

Metformin helps prevent or less the risk of fracture caused by osteoporosis. (My osteoporosis is severe and is my greatest health risk.) So: helps prevent fracture but as a mitochondrial toxin may contribute to sarcopenia.

Metformin helps prevent the progression of cancer because it reduces the availability of glucose to cancer cells.