Metabolic & Cardiometabolic
metformin decreases cancer incidence
In plain terms: Does metformin reduce cancer risk?
Part of: 💊 metformin
Contested — observational cohorts and their meta-analyses suggest lower cancer incidence, but those associations are heavily confounded by time-related (immortal-time) bias, and bias-corrected analyses, target-trial emulations, and the one cancer-endpoint RCT show little or no true effect.
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: All trials, pooled (Meta-analysis)
How the studies fall
The evidence (12)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Liu 2016 · Pharm Biol | meta-analysis | supports | low | Meta of observational studies: metformin associated with reduced colorectal cancer incidence in T2D; observational-only, susceptible to bias. |
| Col 2012 · Breast Cancer Res Treat | meta-analysis | mixed | moderate | Breast cancer meta + critical review: apparent risk reduction but authors explicitly flag methodological bias undermining the association. |
| Suissa 2023 · Diabetes Care | observational | contradicts | high | Real-world evidence failure": re-analysis shows metformin's reported cancer benefit is an artifact of biased designs; properly designed studies show no effect. |
| Campbell 2017 · Ageing Res Rev | meta-analysis | mixed | low | Meta reports lower all-cause mortality and some cancers in metformin users vs non-diabetics, but design cannot separate drug effect from healthy-user/survivor bias. |
| Hernan 2018 · Cancer Epidemiol Biomarkers Prev | observational | contradicts | high | Methodological critique: action-focused/target-trial framing eliminates much of the spurious metformin-cancer signal seen in etiologic observational analyses. |
| Powell 2025 · Am J Epidemiol | observational | contradicts | high | Target-trial emulation vs sulfonylureas accounting for biases: metformin did NOT reduce cancer incidence — well-designed null undercutting the protective claim. |
| Lei 2026 · Urol Oncol | meta-analysis | tested-null | moderate | Systematic review of 85 studies found no significant association between metformin and bladder cancer risk, a directional null against a protective claim. |
| Szymczak-Pajor 2026 · Int J Mol Sci | meta-analysis | supports | moderate | Meta-analysis of 16 observational studies found metformin use associated with reduced pancreatic cancer risk versus other antidiabetic drugs. |
| Petkovska 2024 · J Natl Cancer Inst | meta-analysis | mixed | moderate | Updated systematic review/meta: pooled observational data show lower cancer incidence with metformin, but authors stress residual confounding and bias limit causal inference. |
| Suissa 2012 · Diabetes Care | observational | contradicts | high | Demonstrates time-related (immortal-time, time-window) biases inflate apparent metformin anticancer benefit; correcting them largely abolishes the association. |
| Tsai 2023 · Viruses | observational | supports | low | Cohort in diabetic chronic-HCV patients: metformin use associated with reduced hepatocellular carcinoma; observational, confounding-prone. |
| Higurashi 2016 · Lancet Oncol | RCT | mixed | moderate | Only cancer-endpoint RCT: low-dose metformin modestly reduced metachronous colorectal adenoma recurrence in non-diabetics — small, single-center, surrogate (adenoma not cancer) endpoint. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.