Metformin and Cancer: What Is the Connection?

Metformin, the most widely prescribed diabetes drug in the world, has a complicated and still-unfolding relationship with cancer. Early population studies suggested that people with diabetes who took metformin developed cancer roughly 30% less often than those on other diabetes treatments, sparking enormous research interest. But the story has grown murkier as scientists have dug deeper: some of those early studies carried statistical flaws that inflated the benefit, randomized trials have delivered mixed results, and the drug’s anticancer effects in the lab require concentrations far higher than what patients actually reach in their bloodstream. The connection is real at a biological level, but translating it into a proven cancer therapy has turned out to be far harder than the early headlines suggested.

Where the Excitement Began

The metformin-cancer link entered mainstream research attention around 2010, when a systematic review and meta-analysis found that metformin use in diabetic patients was associated with lower cancer risk across multiple tumor types.¹ A comprehensive review later put the figure at about 30% lower likelihood of developing cancer compared with patients taking other diabetes medications.² These were observational studies, meaning they tracked what happened to patients who were already taking metformin for diabetes rather than randomly assigning it. Still, a 30% risk reduction across large populations was hard to ignore, and it launched hundreds of laboratory studies and clinical trials aimed at understanding how metformin might work against tumors.

What Metformin Does Inside Cancer Cells

In the lab, metformin attacks cancer cells through several routes. One of the best-studied is its ability to activate an enzyme called AMPK, which acts as an energy sensor inside cells. When AMPK switches on, it dials down growth-promoting pathways that tumors rely on to keep dividing. Research in colorectal cancer cells showed that metformin’s activation of AMPK suppressed a downstream signaling hub called mTOR and targeted cancer stem cells, the subpopulation of tumor cells thought to drive relapse and drug resistance.³

Another mechanism operates even more directly on the cell’s energy machinery. Metformin inhibits complex I of the mitochondrial electron transport chain, essentially choking off one of the main ways cells generate energy. In experiments using colon cancer cells, metformin fully shut down complex I-dependent oxygen consumption, an effect that could be traced to a direct interaction with the complex itself rather than some indirect signaling event.⁴ Cancer cells, which often run their metabolism in overdrive, are thought to be more vulnerable to this energy disruption than normal tissue.

Cancer Stem Cells and Drug Resistance

One of the more striking laboratory findings is that metformin selectively kills cancer stem cells. These are the small fraction of tumor cells that can regenerate the entire tumor and are often the reason cancers come back after chemotherapy. In breast cancer experiments, low doses of metformin killed cancer stem cells across four genetically distinct types of breast cancer. When combined with the chemotherapy drug doxorubicin, the pair killed both stem and non-stem cancer cells and shrank tumors in mice far more effectively than either drug alone, while also preventing relapse.⁵

A phase II clinical trial in ovarian cancer provided some human evidence for this idea. Patients treated with metformin showed roughly a 2.4-fold reduction in the percentage of cells carrying cancer stem cell markers compared with controls. Their tumor cells were also more sensitive to platinum chemotherapy and appeared less likely to develop resistance over time.⁶ Similarly, in colorectal cancer cells, combining metformin with a Wnt signaling inhibitor reduced cancer stem cell markers and overcame resistance to chemotherapy in a synergistic way.⁷

Effects Beyond the Tumor Itself

Metformin does not only act on cancer cells directly. It also reshapes several systemic conditions that tumors exploit to grow. One of these is insulin-like growth factor 1, or IGF-1, a hormone that promotes cell growth throughout the body and is elevated in many cancers. A randomized trial in non-diabetic women with metastatic breast cancer found that adding metformin to chemotherapy significantly lowered IGF-1 levels, reduced disease progression, and reduced mortality compared with chemotherapy alone.⁸

Perhaps more intriguingly, metformin appears to boost the immune system’s ability to fight tumors. Inside a tumor, the low-oxygen environment suppresses the immune cells that are supposed to attack cancer. Lab studies showed that metformin rescued CD8 T cells, the immune system’s main tumor-killing cells, from this hypoxia-induced shutdown, keeping them alive, functional, and proliferating while reducing markers of immune exhaustion.⁹ Separate research confirmed that metformin increased the number of CD8 T cells infiltrating tumors and protected them from the functional exhaustion that typically blunts immune responses against cancer.¹⁰ This immune-boosting effect has fueled interest in combining metformin with immunotherapy drugs.

There is also emerging evidence that metformin may influence cancer indirectly through the gut microbiome. In mouse experiments, metformin altered gut bacteria in animals on a high-fat diet, and transplanting the microbiome from metformin-treated mice into untreated mice on the same diet suppressed colorectal tumor growth, even though the recipient mice never received the drug themselves.¹¹ This suggests that some of metformin’s anticancer activity may be routed through the microbial ecosystem of the gut rather than through direct contact with tumor cells.

The Dose Problem

For all the impressive laboratory results, a stubborn pharmacological puzzle hangs over the entire field. Most cell-culture experiments require millimolar concentrations of metformin to see anticancer effects. A study in osteosarcoma cells, for example, found that the concentration needed to kill half the cancer cells was around 3 millimolar.¹² But a diabetic patient taking a typical daily dose reaches plasma levels of only about 10 micromolar, roughly a hundredfold lower.¹³

This gap raises a fair question: can the drug actually reach tumors at concentrations high enough to do anything? Some researchers argue that metformin accumulates in certain tissues at levels above plasma concentration, or that chronic low-level exposure over months and years may produce effects that a short lab experiment at higher doses is modeling in fast-forward. Others point out that the indirect mechanisms, like lowering insulin and IGF-1 or reshaping the immune response, do not require the drug to be present inside the tumor at high concentrations. The honest answer is that nobody has fully resolved this discrepancy, and it remains one of the central uncertainties in the field.

When the Epidemiology Misleads

The 30% risk reduction that launched the field has itself come under serious scrutiny. A detailed review of observational studies on metformin and cancer found that 13 of them suffered from a statistical flaw called immortal time bias. This bias arises because the period between a patient’s diabetes diagnosis and the start of metformin treatment is incorrectly counted as “metformin-exposed” time, during which the patient cannot, by definition, have the outcome (death or cancer diagnosis) attributed to the drug. The studies affected by this bias reported cancer risk reductions ranging from 20% to 94%. The three studies that avoided the bias found no effect of metformin on cancer incidence at all.¹⁴

A focused meta-analysis of metformin and gastric cancer drove this point home even more starkly. Among the six cohort studies that contained immortal time bias, metformin appeared to reduce gastric cancer risk by a third. But among the eight studies that properly accounted for the bias, the association vanished entirely.¹⁵ This does not mean metformin has zero relevance to cancer, but it does mean that the headline-grabbing population-level risk reductions were probably exaggerated, and that the strongest evidence needs to come from randomized trials rather than database studies.

What Randomized Trials Have Found

The largest randomized trial of metformin in cancer to date is MA.32, which enrolled 3,649 patients with high-risk, non-metastatic breast cancer across Canada, the US, the UK, and Switzerland. Patients without diabetes were randomly assigned to take metformin (850 mg twice daily) or a placebo for five years while receiving standard cancer treatment.¹⁶ A secondary analysis of the same trial also examined whether metformin prevented entirely new cancers from developing during follow-up.¹⁷ The results were underwhelming for the overall population: metformin did not significantly improve invasive disease-free survival across the whole group. However, subgroup analyses hinted at possible benefits in certain molecular subtypes, keeping the door open rather than slamming it shut.

In colorectal cancer prevention, the picture is somewhat brighter. A Japanese randomized trial gave low-dose metformin or placebo to patients without diabetes who had already had polyps removed. After a year, about 38% of the metformin group had new polyps compared with roughly 57% in the placebo group, a significant reduction. The rate of new adenomas, the type of polyp most likely to become cancerous, was similarly lower in the metformin arm.¹⁸ This remains one of the cleanest demonstrations of a preventive effect in humans.

In lung cancer, by contrast, the evidence has been disappointing. A meta-analysis of trials adding metformin to standard therapy in non-diabetic patients with advanced non-small cell lung cancer found no meaningful improvement in progression-free survival, overall survival, or tumor response rates.¹⁹ The pattern emerging across tumor types is that metformin’s benefit, if real, is not universal and likely depends on the biology of the specific cancer being treated.

Prostate Cancer and Observational Signals

Prostate cancer occupies an interesting middle ground in the metformin story. Lab work has shown that metformin can reduce the viability of prostate cancer cells by downregulating the androgen receptor, the protein that drives most prostate cancers. It worked in both standard androgen-dependent cells and in a resistant cell line that expresses a truncated version of the receptor, which is associated with treatment failure.²⁰

Large observational studies have supported this: diabetic men on metformin who were receiving androgen deprivation therapy for advanced prostate cancer had about a 30% lower risk of dying from the disease compared with diabetic men not taking metformin.²¹ A systematic review and meta-analysis similarly found that metformin use was linked to better overall survival, cancer-specific survival, and recurrence-free survival in prostate cancer patients.²² These remain observational findings, with all the caveats that apply, but the biological plausibility is stronger here than in some other cancers because the androgen receptor mechanism provides a clear target.

Why Some Tumors Respond and Others Do Not

One of the most important recent developments in this field is the realization that not all tumors can take up metformin equally. The drug enters cells through a family of membrane channels called organic cation transporters, and tumors vary widely in how many of these channels they express. In animal models, mammary tumors that responded to metformin had significantly more cells expressing one of these transporters (OCT2), and the amount of metformin found inside tumors correlated directly with OCT2 levels.²³ A separate study confirmed this by showing that metformin worked better against breast tumors with higher cation transporter expression, and that its anticancer activity inside the tumor tracked with how much drug actually got in.²⁴

On the flip side, when researchers knocked down a related transporter (OCT1) in ovarian cancer cells, those cells became less sensitive to metformin, and primary human tumors showed considerable variation in OCT1 expression.²⁵ This variability helps explain why metformin works against some tumors and fails in others: if a tumor lacks the doorway for the drug to get inside, the drug’s direct effects are simply blocked. It also suggests that testing a tumor’s transporter expression before treatment could identify which patients are most likely to benefit, a step toward precision use of metformin in oncology.

Genetic mutations in the tumor itself may matter too. Preclinical work in non-small cell lung cancer showed that tumors carrying a specific combination of mutations (in the genes KRAS and LKB1) responded well when metformin was combined with chemotherapy or chemo-immunotherapy. These tumors were unable to compensate for the metabolic stress metformin created, while tumors without those mutations handled it just fine.²⁶ This finding is especially relevant because the broad-population lung cancer trials showed no benefit, but a genetically defined subset may still respond.

Metformin as a Partner Drug

Even researchers skeptical about metformin as a standalone anticancer agent are increasingly interested in its potential as a combination therapy. In pancreatic cancer cell lines, metformin made radiation therapy roughly a third to nearly half more effective, and it also enhanced the killing power of gemcitabine, a standard chemotherapy drug for pancreatic cancer, both alone and combined with radiation.²⁷ Across various tumor types, combination strategies pairing metformin with chemotherapy, radiotherapy, targeted therapy, and immunotherapy have shown stronger anticancer effects and longer survival than any single treatment alone in preclinical and early clinical settings.²⁸

The immune-boosting angle is particularly interesting for combination with checkpoint inhibitors, the immunotherapy drugs that have transformed treatment for melanoma, lung cancer, and other malignancies. If metformin can keep CD8 T cells active inside the hostile tumor microenvironment, it could theoretically amplify the effect of drugs that are already designed to release the brakes on immune attack. Several trials exploring this pairing are underway, though definitive results from large randomized studies are still pending.

What This Means If You Take Metformin

If you are taking metformin for diabetes, nothing in the current evidence should change your treatment plan, but the research trajectory is worth understanding. You should not think of your diabetes medication as a proven cancer preventive. The early observational numbers were inflated by statistical biases, and the randomized trials so far have not shown a dramatic benefit in the general cancer population. At the same time, metformin’s biological effects on tumor metabolism, cancer stem cells, the immune system, and growth factor signaling are real and well-documented. The drug is inexpensive, has a long safety record, and carries relatively mild side effects, which is why oncology researchers remain interested in it even after the disappointing headlines.

The field is moving away from asking “does metformin prevent cancer in everyone?” and toward “which patients, with which tumors, might benefit?” Transporter expression, specific tumor mutations, and the metabolic profile of a patient’s cancer are all being investigated as ways to select the right people for metformin-based strategies. That is a slower and less dramatic path than a universal cancer pill, but it is a more honest reflection of what the science actually supports. Meanwhile, dozens of clinical trials continue to test metformin in combination with standard treatments across breast, colorectal, prostate, ovarian, pancreatic, and lung cancers, and a clearer picture of where this cheap old drug fits into modern oncology should emerge over the next several years.