Metformin for Cancer Prevention

The Bottom Line

Metformin is a first-line medication for Type 2 diabetes mellitus, prescribed to hundreds of millions of people worldwide. Its mechanism (AMPK activation via mitochondrial complex I inhibition) is shared with the longevity benefits of caloric restriction, which generated enormous interest in repurposing it for cancer prevention. Observational studies consistently showed that diabetic patients taking metformin had significantly lower cancer incidence and mortality than those on other diabetes medications. However, the MA.32 trial — a large, well-conducted, placebo-controlled randomized controlled trial in women with early-stage breast cancer published in JAMA (2022) — found that metformin had no effect on invasive disease-free survival (HR 1.01). The primary endpoint was definitively negative. Subgroup analyses hint at benefit in triple-negative breast cancer (HR 0.79), but this is hypothesis-generating, not practice-changing. The leading explanation for the observational data is "healthy user bias" — people who take metformin as prescribed also tend to have healthier behaviors overall.

What Metformin Is

Metformin (dimethylbiguanide) is derived from galegine, a compound found in French lilac (Galega officinalis), a plant used in medieval European folk medicine for symptom relief of diabetes. It was approved for Type 2 diabetes in 1957 (France) and 1994 (US) and has since become the most prescribed diabetes medication worldwide, with over 150 million annual prescriptions. Its patent expired long ago, making it extremely cheap (approximately $5/month for generic metformin in the US). It is on the WHO List of Essential Medicines.

The AMPK/Longevity Hypothesis

Metformin's anticancer interest stems from its mechanism overlapping with the pathways activated by caloric restriction — the most robust longevity intervention in laboratory animals:

• AMPK activation: Metformin inhibits mitochondrial complex I, reducing ATP production and activating AMPK (AMP-activated protein kinase). AMPK is a cellular energy sensor that puts cells in "energy conservation mode."
• mTOR inhibition: AMPK activation leads to inhibition of mTOR (mechanistic target of rapamycin), a key driver of cell growth and protein synthesis. mTOR inhibition is strongly associated with lifespan extension.
• IGF-1 reduction: Metformin reduces circulating IGF-1 (insulin-like growth factor 1), a hormone that promotes cell proliferation and is linked to cancer risk.
• Insulin reduction: Reduces hyperinsulinemia — a risk factor for several cancers (breast, colon, endometrial).
• Inflammation reduction: Modestly reduces CRP, IL-6, and other inflammatory markers.

The Observational Evidence

Dozens of observational studies over two decades showed compelling data:

• A 2005 retrospective cohort study by Evans et al. found diabetic patients on metformin had a 23% lower cancer incidence than those on other medications.
• Multiple meta-analyses found 20–40% reductions in cancer risk for metformin users vs. non-users.
• Retrospective data suggested improved survival in breast, colorectal, pancreatic, and prostate cancer patients taking metformin.

The consistency of this observational signal is what drove multiple large RCTs, including MA.32.

The MA.32 Trial — The Definitive Test

The MA.32 trial (NCT01101438) was a large, placebo-controlled, double-blind, randomized phase III trial conducted by the NCIC Clinical Trials Group at 366 sites in four countries. Published in JAMA (December 13, 2022):

• Population: 3,629 women with early-stage breast cancer (stage I–IIIA), without diabetes, BMI ≥23
• Intervention: Metformin 850mg twice daily (vs. placebo) for 5 years
• Primary endpoint: Invasive disease-free survival (IDFS)
• Result: HR 1.01 (95% CI 0.84–1.22); no difference between groups
• Secondary endpoints: Overall survival, distant recurrence — all essentially null

Why the Observational Data Was Likely Wrong

The healthy user bias hypothesis is the most compelling explanation:

• Metformin requires reasonably good compliance (taken daily, with food). Patients who take metformin reliably tend to be more engaged with their health in general.
• Diabetic patients on metformin tend to have better access to healthcare, more screening, and closer monitoring than those on other regimens.
• Many diabetic patients are on multiple medications — comparing metformin users to "other diabetes drugs" may be comparing more health-conscious patients to less health-conscious ones.
• The effect size in observational studies (20–40% risk reduction) is unusually large for any single intervention, which is a red flag for confounding.

The TNBC Subgroup — A Glimmer of Signal

In the MA.32 trial, women with triple-negative breast cancer (TNBC) showed a trend toward benefit with metformin:

• TNBC subgroup HR: 0.79 (95% CI 0.52–1.19) — not statistically significant but numerically suggestive
• This is biologically plausible: TNBC lacks targeted therapy options, has higher metabolic activity, and may be more responsive to metabolic interventions
• Three other smaller trials have also hinted at TNBC benefit

However, this subgroup analysis is hypothesis-generating only. It does not support prescribing metformin for TNBC patients outside of clinical trials. Several ongoing trials are specifically enrolling TNBC patients, which may clarify this signal.

Other Ongoing Trials

• Several trials specifically in TNBC are ongoing or recently completed
• Colorectal cancer chemoprevention trials are also in progress
• Pre-diabetes/diabetes prevention trials examining cancer as secondary endpoint

Protocol Considerations

• Dose for diabetes: 500mg–2000mg/day, typically started at 500mg and titrated up
• Dose in trials: 850mg twice daily (MA.32 protocol)
• Administration: With meals to minimize GI side effects
• Duration: Long-term for diabetes; 5 years in the MA.32 trial

Safety Profile

• GI side effects: Nausea, diarrhea, abdominal discomfort — common at initiation, usually resolve
• Vitamin B12 deficiency: Long-term use can reduce B12 absorption; monitor levels
• Lactic acidosis: Rare in patients with normal kidney function. Contraindicated in severe renal impairment (eGFR <30).
• No significant hypoglycemia when used alone (without insulin secretagogues)
• Generally well-tolerated and one of the safest diabetes medications

Our Assessment

Metformin is a fascinating case study in the gap between observational evidence and randomized trial results. The observational data was consistent and the biological mechanism was plausible — but when rigorously tested, the hypothesis failed. The MA.32 trial is high-quality, large, and definitive for its primary endpoint. The TNBC signal is worth watching, but it is not sufficient to recommend metformin outside of trials. The healthy user bias explanation is the most likely reason for the observational findings. If you're already taking metformin for diabetes or pre-diabetes, continue — it has cardiovascular benefits. If you're considering it specifically for cancer prevention with no metabolic indication, the evidence does not support it. Watch the TNBC trials for updates.

Sources

• JAMA 2022;328(22):2204-2216: MA.32 trial — metformin in breast cancer
• BMJ 2014;348:g3377: Meta-analysis of metformin and cancer risk (observational)
• Diabetes Care 2006;29(3):554-559: Evans et al. — original observational study
• PMC4720208: AMPK/mTOR mechanism and cancer
• PMC6947533: Healthy user bias in metformin observational studies
• ClinicalTrials.gov NCT01101438: MA.32 registration
• PMC10175789: TNBC subgroup analysis (2023)