We’ve done quite a few articles and podcasts digging into the possibility that the cheap diabetes drug metformin might have a secret identity as a geroprotector—that is, a drug with broad-based effects on the aging process, beyond its specific effect on blood sugar. This includes many conversations with Dr. Nir Barzilai and a dedicated Premium Article on the subject. Long story short: Much as I respect Nir, I am profoundly unimpressed by the case for metformin combating the aging process.
The one study that did the most to convince people of metformin’s gerotherapeutic potential was an observational study by Bannister et al.,1 which seemed to show that diabetic metformin users not only lived longer than people with Type 2 diabetes who relied on other diabetes medications, but that they actually lived longer than nondiabetic people in the general UK population. People with diabetes have a shorter life expectancy than the general population, and metformin doesn’t cure diabetes. So if metformin had such a potent effect on mortality as to make diabetics longer-lived than typical nondiabetics, it would be quite provocative evidence that metformin must have something much greater up its pharmacological sleeves than suppressing blood sugar output from the liver.
Unfortunately, as discussed in a previous newsletter, Bannister et al. and many similar studies were flawed by substantial selection and time-based bias, often comparing healthier metformin users with sicker diabetic patients receiving additional therapies. Once these biases are addressed, the evidence for broad geroprotective effects becomes far less compelling, if not collapsing entirely.
Unfortunately, it’s not only researchers who study the biological aging process that have been fooled by this kind of error: The epidemiology on metformin is riddled with it.2,3
And that’s how we got to a failed cancer trial.
On the lookout
Animal studies and epidemiological evidence had long suggested that metformin might modulate cancer cell biology to make it less aggressive, and even prevent cancer outright.2,3 This evidence had already seemed promising enough to inspire a previous trial in which metformin was tested as an adjunct to androgen deprivation therapy in 1,874 nondiabetic patients with full-on metastatic hormone-sensitive prostate cancer.4
Metformin failed to improve overall survival in that trial. But maybe it had started therapy too late: the underlying epidemiological studies found metformin to be associated with lower risk of new cancers, not better treatment of existing ones. Conversely, a trial of metformin for the prevention of prostate cancer would take too long, require too many men, and would be even more expensive than a treatment trial—and since metformin is a generic drug, such a trial would have no obvious source of funding. (See Nir Barzilai’s decade-long effort to get the TAME Trial funded).
The sweet spot for such a trial might be men in active surveillance for low-risk prostate cancer. (”Low-risk” here means cancers that are slow-growing entities whose cells did not show signs of aggressiveness and that have not spread beyond the prostate itself (that is, had not metastasized)).
Such cancers are not likely to be an immediate threat to the patient’s life, and might not ever be. In active surveillance, no specific therapy is given in response to the diagnosis alone: instead, patients and physicians closely monitor the cancer with periodic PSA tests, digital rectal exams (DRE), biopsies, and imaging, and only initiate treatment if the cells begin showing signs of becoming more aggressive.5 This gives men more time without treatment and any treatment-associated side effects, even if they eventually turn out to need it, and allows some men to live out their lives without treatment entirely.
However, despite the favorable overall long-term outcomes for active surveillance (low rates of metastasis and mortality in appropriately-selected patients), progression is common, and 24-40% of men eventually require radical treatment.6 Fortunately, in the 35 years since screening with prostate-specific antigen (PSA) began, scientists have developed additional tools to distinguish between cancers that require immediate action and those that can be left to slumber. But it would be nice to have something that men could do to prevent somnolent cancers from waking up.
There are some such “secondary prevention” options available, including the anti-androgen cancer drug enzalutamide/Xtandi®7 and the 5-alpha-reductase inhibitor dutasteride/Avodart®,8 which is used for benign prostatic hyperplasia (BPH) and male pattern baldness. However, not many men opt for or stick with these approaches, due to low efficacy, side effects, and concerns that they might adversely affect the long-term biology of a man’s cancer.
This is where metformin entered the picture.
Tied to the MAST
MAST (the Metformin Active Surveillance Trial) was a Phase III, randomized, double-blind, placebo-controlled trial conducted in 12 cancer centers across Canada.9 It enrolled men with low-risk localized prostate cancer who were on active surveillance—and to avoid conflating a true anticancer effect with an effect on diabetes, they had to be nondiabetic, as assessed on multiple fronts. 408 men were randomized to receive metformin (850 mg twice daily) or placebo, and were followed up to 36 months, with scheduled biopsies as part of active surveillance at the midway point and at the end of the trial. The primary endpoint was to see if metformin would delay the time it took for the men’s cancers to progress, based on either malignant-looking changes on biopsy or treatment for cancer.
After a median of three years, 144 out of the 408 men’s cancers progressed: 70 in the metformin group, and 74 in the men on placebo. If that sounds like no effect, you’re right: there was no significant difference between the two groups in overall progression-free survival (HR 1.09, 95% CI 0.79-1.52, p=0.59) or progression subtypes.9 And there was also no statistically significant difference between the number of men whose final biopsy was negative, which might otherwise have suggested a longer-term benefit: 41.0% on metformin vs 31.1% on placebo; p=0.181.9
Our research team spends hundreds of hours each month vetting studies and distilling dense literature to deliver evidence-informed insights on health and longevity. If you find value in our work, consider becoming a premium member and supporting our mission.
“Here we go again”?
Recalling that the epidemiology on metformin and cancer is entirely in men using metformin for diabetes, you might reasonably wonder if metformin’s observed association with low prostate cancer risk might involve correcting insulin resistance, and thus only be effective in people somewhere along the metabolic syndrome-diabetes continuum. There’s no way to answer that question with certainty, since MAST rigorously excluded diabetic men—but in a preplanned subgroup analysis, the investigators looked separately at the results in obese (BMI ≥30) versus nonobese subjects, since people with obesity are much more likely to be insulin resistant than are leaner people.9
In nonobese patients, there was still no significant difference between the two groups (HR 0.82, 95% CI 0.55 to 1.23, p=0.33). By contrast, in obese patients, metformin was associated with increased pathologic progression (HR 2.36, 95% CI 1.21–4.59, p=.0092)—exactly the opposite of what you’d expect if metformin’s association with low cancer risk in observational studies were because of its effects on glucose metabolism!9 Given the number of subgroup analyses performed, this could easily represent chance rather than biology, so we shouldn’t get too anchored to this as a real biological difference, but there’s certainly no signal here for a benefit driven by the metabolic consequences of obesity.
Alternatively, there is a signal in the observational data10 to suggest that, if there is a causal protective effect of metformin on mortality in prostate cancer, it comes from improving prostate cancer patients’ metabolic health rather than controlling their cancer. An effect specific to non-cancer mortality would not show up on the prostate biopsies in MAST, but would still benefit men.
Consistent with this hypothesis, the earlier STAMPEDE trial did show that metformin significantly ameliorated the adverse metabolic impact of androgen deprivation therapy when used as an adjunct in men with metastatic hormone-sensitive prostate cancer.4
Pointing in the opposite direction, however, one of several preplanned subgroup analyses in the STAMPEDE platform trial seemed to show that the only men in whom metformin improved survival were the ones who had more advanced disease (HR 0.79, 95% CI 0.67-0.94; p=0·0072), with no effect in those who had relatively few and less deadly metastases (HR 0.98 (0.78-1.23; p=0·87)).4
MAST found that metformin is of little use for secondary prevention during active surveillance. However, if (hypothetically) metformin did have a protective effect that was restricted to metastatic disease, or that involved metabolic benefits that would accrue independent of any effect on the cancer itself, such benefits would not have emerged during MAST, and it would still leave some hope for using the drug in the right conditions.
Of babies and bathwater
The widely- (albeit naively-) reported association of metformin use with reduced risk of cancer or cancer mortality in observational studies has sparked quite a few trials, and most of them have failed. A 2022 meta-analysis of 22 randomized controlled clinical trials (RCTs) that tested metformin as a treatment or adjunct against cancer in 5,943 participants found that metformin had no effect on either how long patients survived without their cancer progressing (HR 0.97, 95% CI 0.82–1.15, I2 = 50%) or their overall survival (HR 0.98, 95% CI 0.86–1.13, I2 = 33%).11
By a statistical squeaker, that same study found that metformin may have had a marginal effect on progression-free survival in reproductive system cancers (breast, ovary, endometrium, and prostate combined—HR 0.86, 95% CI 0.74–1.00), but it significantly worsened progression-free survival in digestive system cancers (HR 1.45, 95% CI 1.03–2.04).11 However, trials published after the cutoff for this meta-analysis found that metformin did not improve progression-free survival in women with advanced-stage ovarian cancer,12 high-risk operable breast cancer,13 or men with prostate cancer (the STAMPEDE trial4), which tends to undermine the meta-analysis’ reproductive system cancer finding—especially since the breast cancer trial alone was half again as large as the combined populations of the reproductive system cancer studies included in the meta-analysis.
While this meta-analysis synthesized trials in which metformin was used to treat existing cancers, a separate meta-analysis that looked at new cancer incidence as an outcome in a secondary analysis of 27 metformin RCTs found that metformin had no preventative effect (RR 1.07, 95% CI 0.87–1.31).14 The meta-analysis included over 20,000 people and 378 new cases of cancer, so they had the statistical power to exclude a significant effect of metformin on cancer incidence. And the result was the same when the researchers restricted their analysis to trials that lasted longer than two years, or in which subjects on metformin either gained or lost weight relative to whatever drug (or placebo) against which they were compared.14
That’s a long litany of failures, sprinkled with a couple of trials that are suggestive but not convincing, so you might be about ready to stop reading this newsletter and vow not to read anything about metformin and cancer in the future. If so, hold on for just a little longer. Because despite these multiple failures, a handful of relatively small trials continue to suggest that metformin might be of benefit in specific high-risk precancerous conditions.
Colorectal cancer prevention
In a multicenter Phase III RCT in 151 nondiabetic volunteers who already had colorectal polyps or adenomas (high-risk polyps) removed, subjects were carefully randomized to receive either 250 mg metformin or placebo. After one year of followup, metformin cut the risk of new adenomas by 40% (RR 0.60, 95% CI 0.39-0.92), and reduced the total polyp recurrence by a third (RR 0.67, 95% CI 0.47-0.97). No serious adverse events were observed.15 This trial was preceded by a small, positive pilot study from the same group.16
More recently, Chinese scientists conducted an RCT in an even higher-risk group: 272 nondiabetic volunteers who had already had 3 or more adenomas removed. The investigators randomized volunteers into either a control group with no intervention (not even a placebo) or to receive 500 or 1,000 mg of metformin. After one year, 48.9% of the control subjects had suffered an adenoma recurrence, versus about 30% in each of the two metformin arms—a statistically significant difference.17
However, the picture is not uniformly positive, even for colorectal adenoma risk. In another small trial, Park et al. conducted a placebo-controlled RCT of metformin in 34 patients with familial adenomatous polyposis (FAP), an inherited genetic disease in which mutations in one of several tumor suppressor genes cause high numbers of cancer-prone polyps to form in the large intestine. Neither of two doses of metformin reduced the number or size of polyps relative to placebo.18 However, the trial’s very small size and the fact that the subjects in this trial had FAP mutations from birth argue against weighting it strongly against the positive results of the other trials.
Endometrial Cancer Prevention
Cancer of the endometrium (the inner lining of the uterus) is already one of the most common women’s cancers in the world—and its rates are rising, in part because obesity and insulin resistance are major risk factors for the disease, and the prevalence of those conditions has risen dramatically. The role of insulin resistance in the disease has motivated scientists to test metformin across the spectrum of endometrial cancer.
Endometrial hyperplasia (EH), a condition in which the cells of the endometrium proliferate too rapidly, predisposes women to develop endometrial cancer. Because one driver of EH is estrogen that is not “opposed” by progesterone, doctors often use synthetic progesterone derivatives (progestins) to prevent or treat EH. In a meta-analysis of trials of metformin to treat EH, metformin seemed more effective than one progestin (megestrol) at reversing hyperplasia, and combining metformin with either megestrol or another progestin (levonorgestrel) seemed to work better than either progestin alone, although the investigators expressed high uncertainty about these results because of the very small sample sizes and risks of bias in the underlying trials.19
Another thing that can put women at risk for endometrial cancer is tamoxifen, a drug that acts on estrogen receptors that is used to treat some breast cancers and to reduce the risk of developing breast cancer in some high-risk women. Despite tamoxifen’s benefits, its estrogenic stimulation confers a small risk of endometrial cancer. So researchers tested whether metformin could potentially shield breast cancer survivors who were treated with tamoxifen. In a placebo-controlled RCT with 102 women, metformin significantly reduced the median thickness of the endometrium one year after starting tamoxifen therapy.20
EH is extremely unlikely to progress to endometrial cancer so long as endometrial thickness is kept below 4 millimeters, so this has been used as a safety endpoint in drug trials. In another analysis of the results, only 5.7% of women on metformin went on to develop an endometrial thickness that exceeded the worrisome 4 millimeter threshold, versus 13.3% of women allocated to placebo (p=0.26).20
Asking a specific question
So what have we got here? Failed trials are often most valuable when they tell us which broad hypotheses to abandon. MAST adds to a growing body of evidence suggesting that metformin is not a general-purpose anticancer drug. What remains are a handful of narrower, more interesting questions.
That isn’t a prescription for every aging person to start taking metformin to prevent cancer. First, we already have evidence that that doesn’t work.14 And second, we now have multiple trials confirming that metformin interferes with your ability to reap the adaptive benefits of exercise, which is the most powerful broad-based longevity intervention available to us today (as we discussed in two previous issues of the newsletter—and see also two more recent reports21,22).
Instead, these studies give us increasing levels of precision: not “metformin in the drinking water,” or even “metformin for all cancer patients,” but specific contexts where a cheap, common drug may be able to shield specific people against the scythe of the Second Horseman. Those exceptions should motivate better trials, not broader use.
That’s a lesson we could all learn about our approach to supplements and even drugs: not to fall for hazy promises that a single molecule will deliver vaguely-defined benefits for all aging persons, but to demand rigorous evidence for well-defined benefits for well-defined problems in well-defined populations.
For a list of all previous weekly emails, click here.
References
1. Bannister CA, Holden SE, Jenkins-Jones S, et al. Can people with type 2 diabetes live longer than those without? A comparison of mortality in people initiated with metformin or sulphonylurea monotherapy and matched, non-diabetic controls: Can people with type 2 diabetes live longer than people without diabetes? Diabetes Obes Metab. 2014;16(11):1165-1173.
2. Tsilidis KK, Capothanassi D, Allen NE, et al. Metformin does not affect cancer risk: a cohort study in the U.K. Clinical Practice Research Datalink analyzed like an intention-to-treat trial. Diabetes Care. 2014;37(9):2522-2532.
3. Farmer RE, Ford D, Mathur R, et al. Metformin use and risk of cancer in patients with type 2 diabetes: a cohort study of primary care records using inverse probability weighting of marginal structural models. Int J Epidemiol. 2019;48(2):527-537.
4. Gillessen S, Murphy L, James ND, et al. Metformin for patients with metastatic prostate cancer starting androgen deprivation therapy: a randomised phase 3 trial of the STAMPEDE platform protocol. Lancet Oncol. 2025;26(8):1018-1030.
5. Garisto JD, Klotz L. Active surveillance for prostate cancer: How to do it right. Oncology (Williston Park). 2017;31(5):333-340, 345.
6. Bernardino R, Sayyid RK, Leão R, et al. Using active surveillance for Gleason 7 (3+4) prostate cancer: A narrative review. Can Urol Assoc J. 2024;18(4):135-144.
7. Shore ND, Renzulli J, Fleshner NE, et al. Enzalutamide monotherapy vs active surveillance in patients with low-risk or intermediate-risk localized prostate cancer: The ENACT randomized clinical trial: The ENACT randomized clinical trial. JAMA Oncol. 2022;8(8):1128-1136.
8. Fleshner NE, Lucia MS, Egerdie B, et al. Dutasteride in localised prostate cancer management: the REDEEM randomised, double-blind, placebo-controlled trial. Lancet. 2012;379(9821):1103-1111.
9. Fleshner NE, Bernardino RM, Izawa J, et al. Metformin Active Surveillance Trial in low-risk prostate cancer. J Clin Oncol. 2025;43(34):3662-3671.
10. Stopsack KH, Ziehr DR, Rider JR, Giovannucci EL. Metformin and prostate cancer mortality: a meta-analysis. Cancer Causes Control. 2016;27(1):105-113.
11. Wen J, Yi Z, Chen Y, et al. Efficacy of metformin therapy in patients with cancer: a meta-analysis of 22 randomised controlled trials. BMC Med. 2022;20(1):402.
12. Romero IL, Lengyel E, Wahner Hendrickson AE, et al. Metformin for patients with advanced stage ovarian cancer: A randomized phase II placebo-controlled trial. Gynecol Oncol. 2025;194:18-24.
13. Goodwin PJ, Chen BE, Gelmon KA, et al. Effect of metformin vs placebo on invasive disease-free survival in patients with breast cancer: The MA.32 randomized clinical trial. JAMA. 2022;327(20):1963-1973.
14. Mesquita LA, Spiazzi BF, Piccoli GF, et al. Does metformin reduce the risk of cancer in obesity and diabetes? A systematic review and meta-analysis. Diabetes Obes Metab. 2024;26(5):1929-1940.
15. Higurashi T, Hosono K, Takahashi H, et al. Metformin for chemoprevention of metachronous colorectal adenoma or polyps in post-polypectomy patients without diabetes: a multicentre double-blind, placebo-controlled, randomised phase 3 trial. Lancet Oncol. 2016;17(4):475-483.
16. Hosono K, Endo H, Takahashi H, et al. Metformin suppresses colorectal aberrant crypt foci in a short-term clinical trial. Cancer Prev Res (Phila). 2010;3(9):1077-1083.
17. Wang W, Luo B, Wang Z, et al. Metformin for recurrent colorectal polyp or adenoma prevention after polypectomy in patients without diabetes mellitus: a prospective study. Neoplasma. 2025;72(1-2):144-151.
18. Park JJ, Kim BC, Hong SP, et al. The effect of metformin in treatment of adenomas in patients with familial adenomatous polyposis. Cancer Prev Res (Phila). 2021;14(5):563-572.
19. Shiwani H, Clement NS, Daniels JP, Atiomo W. Metformin for endometrial hyperplasia. Cochrane Database Syst Rev. 2024;5(5):CD012214.
20. Davis SR, Robinson PJ, Jane F, et al. The benefits of adding metformin to tamoxifen to protect the endometrium-A randomized placebo-controlled trial. Clin Endocrinol (Oxf). 2018;89(5):605-612.
21. Malin SK, Heiston EM, Battillo DJ, et al. Metformin blunts vascular insulin sensitivity after exercise training in adults at risk for metabolic syndrome. J Clin Endocrinol Metab. 2026;111(4):e1124-e1135.
22. Kristensen JM, Lillelund C, Kjøbsted R, et al. Metformin does not compromise energy status in human skeletal muscle at rest or during acute exercise: A randomised, crossover trial. Physiol Rep. 2019;7(23):e14307.
