Metformin From goat’s rue to first-line diabetes, AMPK lore, and the still-pending TAME bet

04Complex I, the AMP/ATP ratio, and hepatic glucose output

The most durable mechanistic account of metformin’s clinical action centers on a mild, dose-dependent inhibition of mitochondrial complex I (NADH:ubiquinone oxidoreductase) in hepatocytes. That inhibition raises the cellular AMP-to-ATP ratio, which in turn suppresses gluconeogenic flux and lowers hepatic glucose output, the effect most directly responsible for metformin’s clinical glucose-lowering action in type 2 diabetes (Kulkarni et al., 2020, PMID 32333835). This account explains a large share of metformin’s established anti-hyperglycemic behavior and is consistent with decades of clinical pharmacology. It does not, by itself, establish a geroprotective mechanism in non-diabetic adults; the same biochemical event that lowers hepatic glucose output in a person with insulin resistance may have a different, unmeasured, or negligible net effect in someone without metabolic disease. Tissue-specific drug concentration is a further complication: metformin depends on organic cation transporters, particularly OCT1, for hepatocyte uptake, and gut tissue is exposed to substantially higher local concentrations than plasma pharmacokinetics alone would predict, a detail that matters for interpreting both its efficacy and its most common side effects (Kulkarni et al., 2020, PMID 32333835).

05AMPK activation, mTOR cross-talk, and the limits of the classroom diagram

The rise in AMP relative to ATP activates AMP-activated protein kinase (AMPK), a cellular energy sensor that, once active, favors catabolic, energy-conserving processes over anabolic growth and restrains signaling through the mTOR pathway discussed at length in the companion rapamycin monograph in this series. That adjacency — a widely prescribed diabetes drug appearing to dampen the same growth-signaling axis that rapamycin inhibits directly — is a substantial part of why metformin drew serious geroscience interest independent of any diabetes-specific rationale. Mechanistic reviews caution, however, that metformin’s signaling footprint is broader than a single AMPK switch: reported effects vary by tissue, dose, and duration of exposure, and some downstream effects persist in AMPK-knockout models, indicating AMPK-independent routes that remain incompletely mapped (Kulkarni et al., 2020, PMID 32333835). Popular geroscience communication tends to compress this into a single tidy arrow from "metformin" to "AMPK on" to "aging off." This monograph keeps that diagram provisional and multi-layered rather than settled.

06Invertebrate and rodent lifespan signals

Onken and Driscoll reported, in a 2010 study in Caenorhabditis elegans, that metformin induced a dietary-restriction-like physiological state and extended nematode healthspan through a pathway involving AMPK, the upstream kinase LKB1, and the stress-response transcription factor SKN-1, providing early invertebrate mechanistic support for a link between metformin exposure and pathways already associated with lifespan extension in that model organism (Onken and Driscoll, 2010, PMID 20090912). In mice, Martin-Montalvo, de Cabo and colleagues reported in 2013 that long-term dietary metformin at a specific, carefully titrated dose improved several healthspan measures and extended median lifespan in male mice, while explicitly noting that a higher tested dose failed to extend lifespan and produced signs of toxicity, a dose-response relationship that is far less forgiving than the pattern seen across the ITP’s rapamycin cohorts (Martin-Montalvo et al., 2013, PMID 23900241). Taken together, these invertebrate and rodent signals are real, peer-reviewed, and mechanistically coherent with the AMPK/complex-I story; they are also narrower, more dose-sensitive, and less consistently replicated across independent multi-site cohorts than the rapamycin animal package summarized in the companion monograph, and this document does not rank metformin above rapamycin in strength of animal geroprotective evidence merely because metformin is the more familiar clinical drug.

07The exercise-interference literature

A separate, less publicized strand of human research complicates any simple "AMPK activation is always good for healthy aging" narrative. Konopka and colleagues reported, in a randomized trial in older adults, that metformin blunted specific mitochondrial adaptations to supervised aerobic exercise training relative to placebo, despite exercise training itself producing the expected mitochondrial improvements in the placebo arm (Konopka et al., 2019, PMID 30548390). Walton and colleagues, in the MASTERS trial, reported a randomized, double-blind, placebo-controlled, multicenter finding that metformin blunted skeletal muscle hypertrophy in response to progressive resistance exercise training in older adults, again relative to a placebo-plus-exercise comparison arm (Walton et al., 2019, PMID 31557380). Both trials measured specific physiological adaptations to supervised exercise programmes in older adults over a period of months; neither is a trial of unsupervised metformin use, nor a trial of metformin without concurrent structured exercise, and this monograph does not extrapolate them beyond what they measured. Their relevance to public geroscience discourse is specific: metformin and structured exercise are frequently recommended together in longevity-oriented wellness content on the premise that their effects are simply additive, and this human randomized-trial literature indicates that, for at least two distinct exercise-adaptation phenotypes, that premise does not hold and the two interventions can interact antagonistically rather than synergistically under supervised trial conditions.

08The Diabetes Prevention Program: metformin as a delaying agent, not a cure

Before UKPDS settled metformin’s place in established diabetes, the Diabetes Prevention Program (DPP) tested whether it could keep people with elevated glucose from developing diabetes at all. In a large randomized trial reported in 2002, the DPP Research Group compared intensive lifestyle intervention, metformin, and placebo in adults with impaired glucose tolerance, finding that lifestyle intervention reduced the incidence of type 2 diabetes by 58 percent relative to placebo over an average follow-up of just under three years, while metformin reduced incidence by 31 percent relative to placebo — a real, statistically significant, but substantially smaller effect than structured diet and exercise achieved in the same trial (Diabetes Prevention Program Research Group, 2002, PMID 11832527). The DPP result is instructive for this monograph in two directions at once. It is genuine randomized human evidence that metformin measurably delays a specific age-related metabolic disease process, which is part of the mechanistic case for broader geroprotective interest. It is also a reminder, from within metformin’s own best human trial data, that lifestyle intervention outperformed the drug on the same endpoint in the same population, a comparison this monograph does not consider small print.

09UKPDS and the first-line settlement

Metformin’s single largest body of human evidence comes from its ordinary use as a diabetes drug, not from any geroscience trial. The United Kingdom Prospective Diabetes Study (UKPDS), whose overweight-cohort metformin substudy reported results in 1998, found that metformin reduced diabetes-related endpoints and all-cause mortality relative to conventional dietary treatment in overweight patients with newly diagnosed type 2 diabetes, a result that, along with metformin’s favorable tolerability and low cost relative to sulfonylureas and insulin, anchored its subsequent elevation to first-line status in essentially every major diabetes treatment guideline worldwide (Bailey, 2017, PMID 28776081, reviewing this history). That evidentiary base is substantial, but its endpoint was diabetes-related morbidity and mortality in a diabetic population, not biological aging or lifespan in metabolically healthy adults, and this monograph treats the two questions as related but distinct.

10Observational survival comparisons: metformin users versus non-diabetic controls

A frequently cited and genuinely striking observational finding comes from Bannister and colleagues, who in 2014 compared mortality among people with type 2 diabetes newly started on metformin or sulfonylurea monotherapy against a matched, non-diabetic control population drawn from the same UK primary-care database. Metformin-treated patients showed mortality that was not only lower than that of sulfonylurea-treated diabetic patients but statistically indistinguishable from, and in some analyses modestly better than, that of the matched non-diabetic controls, whereas sulfonylurea-treated patients showed higher mortality than both groups (Bannister et al., 2014, PMID 25041462). This finding is real, peer-reviewed, and is one of the most-cited pieces of evidence in popular geroscience discussion of metformin. It is also an observational cohort comparison, not a randomized trial, and it carries the interpretive limitations inherent to that design: channeling bias in which physicians preferentially prescribe metformin to healthier patients and reserve sulfonylureas or insulin for those with more advanced disease, unmeasured confounding by diet, adherence, and comorbidity, and the specific comparator population being other diabetics rather than a randomized metformin-versus-placebo contrast in non-diabetic adults. Bannister and colleagues themselves frame the result as hypothesis-generating rather than as proof that metformin extends life beyond what avoiding diabetes-related mortality would predict, and this monograph adopts that same framing rather than presenting the finding as a demonstrated non-diabetic longevity benefit.

11What the labeled safety record establishes, and its limits

Metformin’s decades of use in a very large diabetic population have generated an unusually mature labeled safety picture for a drug of its class: common gastrointestinal intolerance, particularly early in treatment; a well-characterized risk of vitamin B12 depletion with long-term use, now routinely monitored in clinical guidelines; and a rare but seriously regarded risk of lactic acidosis, concentrated in patients with significant renal impairment, severe hepatic disease, or acute hypoxic states, which is why metformin carries renal-function-based dosing restrictions in its approved labeling. That labeled record describes safety within diabetes pharmacology, in a population selected for glucose intolerance and monitored under routine clinical care; it is not, by itself, a safety demonstration for continuous, multi-year, off-label use in metabolically healthy adults pursuing a longevity indication, a use case the label does not address and this monograph does not evaluate as safe or unsafe.

12What TAME was designed to test

Targeting Aging with Metformin (TAME), championed principally by Nir Barzilai and colleagues at the American Federation for Aging Research, was conceived as a direct answer to a specific regulatory problem: no agency recognizes "aging" itself as an approvable disease indication, so no drug can be approved for treating it. Justice and colleagues, writing in 2018 on the development of clinical trials to extend healthy lifespan, laid out the TAME design logic explicitly: rather than propose aging as an endpoint, enroll older adults without a single index disease and track time to onset of a composite of major age-related conditions — cardiovascular disease, cancer, dementia, and mortality among them — using metformin, an inexpensive, generically available drug with an extensive existing safety record, as the test intervention (Justice et al., 2018, PMID 30906924). A companion methodological paper from the TAME Biomarkers Workgroup addressed a related problem: even a successful composite-endpoint trial would benefit from validated blood-based biomarkers of biological aging that could shorten future trials and provide mechanistic insight alongside the primary clinical endpoint, and the workgroup proposed a framework for selecting and validating such biomarkers specifically in the geroscience-trial context (TAME Biomarkers Workgroup, 2018, PMID 30151729). This design strategy — a composite disease-onset endpoint in a drug-repurposing trial, rather than a single-disease label expansion — has become a template cited across the geroscience field, including in subsequent proposals to prioritize other FDA-approved drugs for aging-focused repurposing (Kulkarni et al., 2022, PMID 35343051).

13Status at this compilation

As of this document’s 8 August 2026 compilation, no peer-reviewed TAME primary efficacy result has been published, and secondary public communications about the trial’s funding status, start date, and projected timeline have varied across public statements from the American Federation for Aging Research and related commentary, requiring re-verification against AFAR’s own trial page and ClinicalTrials.gov at the time of any future reissue of this monograph (American Federation for Aging Research, TAME trial programme page, secondary status communication). This monograph does not report a topline TAME result because none has been published in the source material reviewed for this compilation, and it treats the trial’s continued design-stage status as a fact about where metformin geroscience currently stands rather than as either a failure or an imminent breakthrough.

14Scale, and the comparison to rapamycin’s trial infrastructure

TAME’s ambition illustrates, by contrast, just how much infrastructure a genuine human aging-outcome trial requires, even for a drug as inexpensive and thoroughly characterized as metformin: a large, multi-year, multi-site cohort of older adults, a composite endpoint requiring extensive adjudication, and years of follow-up before a first readout is possible. No comparably resourced trial has been completed for rapamycin, the compound with the strongest animal lifespan package in this monograph series, as discussed in the companion rapamycin monograph; the PEARL trial reviewed there is smaller, shorter, and differently designed. That comparison is offered here as context for the scale of the remaining evidentiary gap across the geroscience field generally, not as a claim that either compound’s human trial programme is inadequate for its stated purpose.

15Chemical identity and regulatory status

Metformin is registered in PubChem (CID 4091) as N,N-dimethylimidodicarbonimidic diamide, a biguanide with molecular formula C₄H₁₁N₅ and a molecular weight of approximately 129.2 g/mol, small and hydrophilic relative to most of the other compounds discussed in this monograph series, a property directly relevant to its dependence on organic cation transporters for cellular uptake rather than passive membrane diffusion (National Center for Biotechnology Information, PubChem Compound Summary for CID 4091). It is typically dispensed as metformin hydrochloride. Regulatory status mirrors the pattern described for rapamycin in the companion monograph: metformin holds FDA approval for glycemic control in type 2 diabetes mellitus, in immediate-release and extended-release formulations and in fixed-dose combination products with other antihyperglycemic agents, and that approval, together with its designation as an essential medicine on the World Health Organization’s model list, reflects its diabetes indication specifically (U.S. Food and Drug Administration, metformin labeling context for type 2 diabetes). No regulatory authority has approved metformin for an anti-aging, longevity, or healthspan-extension indication in any jurisdiction as of this compilation. Two related biguanides, phenformin and buformin, were withdrawn or restricted in most markets after the 1970s because of an unacceptably high rate of lactic acidosis relative to their glucose-lowering benefit; metformin survived that class-wide scrutiny because its lactic-acidosis risk, while real and still respected in current prescribing guidance around renal function, proved substantially lower across large-scale clinical experience (Bailey, 2017, PMID 28776081). That comparative safety history, rather than any aging-specific property, is part of why metformin rather than its biguanide relatives became the geroscience field’s repurposing candidate of choice.

16Berberine, rapamycin, and the metformin "stack"

Longevity-community discussion frequently pairs metformin with two other compounds on the strength of a shared or adjacent mechanistic story. Berberine, a plant alkaloid sold as an over-the-counter supplement, is promoted in wellness marketing as a natural AMPK activator with metformin-like glucose-lowering claims; some small human trials have reported glucose-lowering effects for berberine in diabetic populations, but the supplement-grade evidence base is substantially smaller, less standardized in dose and preparation, and less rigorously regulated than metformin’s pharmaceutical record, and no trial reviewed for this monograph directly tested a metformin-plus-berberine combination against either agent alone for a geroscience endpoint. Rapamycin, the subject of the companion monograph in this series, is proposed in the same community discussion as a complementary partner because it inhibits mTOR directly while metformin acts further upstream through AMPK and complex I; some preclinical rodent factorial studies have tested rapamycin and metformin in combination and reported combination-specific effects that differ from either agent alone under specific dietary and genetic conditions, but no adequately powered human factorial trial comparing metformin plus rapamycin against either agent alone and against placebo has been completed and published as of this compilation. Under the evidence standard this monograph applies throughout, a claim that combining metformin with berberine, rapamycin, or any other candidate geroprotector produces additive or synergistic human benefit is unsupported, regardless of mechanistic plausibility. This monograph recommends no stacking, dosing, or combination protocol for any reader.

17The Sinclair lane, and why it is thinner here than for NMN

David Sinclair, the Harvard geneticist whose public advocacy is treated at length in this series’ companion NMN monograph, has spoken publicly, including in his 2019 popular book Lifespan and in numerous podcast and press appearances, about including metformin in his own personal longevity regimen and about his broader view that metabolic drugs originally developed for diabetes deserve serious study as geroprotectors in non-diabetic adults. That personal-use statement and general advocacy position is treated in this monograph as exactly that: a named public figure’s stated practice and opinion, not a citable scientific finding. Unlike the NMN monograph, where Sinclair’s own peer-reviewed sirtuin and NAD+ laboratory work sits directly inside the compound’s primary scientific literature, no metformin-specific primary research paper from the Sinclair laboratory was identified in the source material assembled for this monograph; his metformin comments belong to the same public-advocacy register as his rapamycin comments discussed in the companion rapamycin monograph, resting on general geroscience reasoning about metabolic drugs rather than on his own metformin-focused experimental programme. The TAME trial and the observational and mechanistic literature summarized in Parts Two through Four constitute metformin’s actual peer-reviewed evidence base and stand independently of any individual advocate’s personal choices.

18Evidence gaps and unresolved questions

Several gaps in the metformin-as-geroprotector record remain material to an accurate account of where the evidence currently stands. No completed, published, adequately powered randomized trial has tested metformin against placebo for an all-cause mortality or composite age-related disease endpoint in metabolically healthy, non-diabetic adults; TAME, the trial designed to answer exactly that question, remains in a design and funding status that has not, in the sources reviewed for this compilation, produced a published primary result. The Bannister observational finding, while striking, cannot by itself substitute for that trial because of its inherent confounding and its comparison population of other diabetics rather than a randomized non-diabetic cohort. The mouse lifespan evidence is real but considerably narrower and more dose-sensitive than the rapamycin ITP package, with Martin-Montalvo and colleagues themselves reporting that a higher tested dose produced toxicity rather than benefit. The exercise-interference trials indicate that at least two specific, well-studied combinations — metformin plus aerobic training, and metformin plus resistance training — can be antagonistic rather than additive under supervised randomized conditions, directly complicating a common longevity-culture assumption. No adequately powered human factorial trial has tested metformin combined with rapamycin, NAD+ precursors, or other candidate geroprotectors against each agent alone and against placebo. Each of these gaps is a reason for continued, well-designed research rather than a basis for either dismissing metformin’s geroscience candidacy or treating it as an established longevity intervention.

19Conclusions

Metformin enters this monograph series carrying, by a wide margin, the most extensive human safety and outcomes record of any compound discussed here — a genuine strength, earned through decades of ordinary diabetes pharmacology rather than through any purpose-built aging trial. That record establishes metformin as a first-line, well-tolerated, extensively characterized diabetes drug with a favorable mortality profile in diabetic populations relative to older comparators. It does not establish, and this monograph does not claim, a demonstrated lifespan or healthspan benefit in metabolically healthy, non-diabetic adults. The TAME trial represents the most concrete regulatory and methodological strategy yet devised for testing that separate question, and its continued unpublished status, as of this compilation, is the single most consequential open item in metformin geroscience. Public advocacy for personal, non-diabetic metformin use by researchers and wellness commentators runs ahead of that unpublished trial record, a gap this document treats as a fact about the current state of evidence. This document is Research Use Only. It recommends no human use, dose, route, schedule, or combination protocol for metformin or any related biguanide for any reader.

AMethods of evidence assembly

Source project: 05 therapeutic-peptide research library and the Adjacent Compounds / Longevity harvest (projects/adjacent_compounds_research_2026/deep_harvest/longevity_v2/). Compound key P188. As with rapamycin, metformin is a small molecule rather than a peptide and is included in the Adjacent Compounds longevity commission for its mechanistic and public-discourse proximity to peptide-adjacent geroscience compounds covered elsewhere in this library.

Ten subject-focused PMIDs were resolved against NCBI E-utilities for the reference list, covering the historical and identity literature (Bailey, 2017; Triggle, 2022), the AMPK/complex-I mechanism synthesis (Kulkarni et al., 2020), the TAME trial design and biomarker framework papers (Justice et al., 2018; TAME Biomarkers Workgroup, 2018), the Bannister observational mortality comparison (2014), invertebrate and rodent lifespan work (Onken and Driscoll, 2010; Martin-Montalvo et al., 2013), and the two human exercise-interference randomized trials (Konopka et al., 2019; Walton et al., 2019). Several PMIDs carried over from an earlier draft of this monograph were independently re-verified against NCBI during this revision and replaced where the identifier resolved to an unrelated paper; no unverified identifier was retained in the published reference list. Eight non-PubMed provenance sources were added for regulatory, registry, and identity context: AFAR TAME trial programme page communications, FDA labeling context for metformin’s diabetes indication, ClinicalTrials.gov protocol inventories, the NIA ITP publications index (cited for cross-reference to the rapamycin animal-evidence comparison in Part Four), and PubChem compound summary pages. Page equivalents use 500 words per printed page for XML/HTML.

Identity gate: documents discussing phenformin or buformin without explicit metformin context were excluded from primary evidentiary use and appear only as named historical comparators in the biguanide-class history recounted in Part One. Figure artwork is authored SVG using theme tokens only (no raw hex in presentation attributes).

BReferences

1. Bailey CJ. Metformin: historical overview. Diabetologia. 2017;60(9):1566-1576.
   PMID 28776081 · doi:10.1007/s00125-017-4318-z
2. Bannister CA, Holden SE, Jenkins-Jones S, Morgan CL, Halcox JP, Schernthaner G, 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. Diabetes Obes Metab. 2014;16(11):1165-73.
   PMID 25041462 · doi:10.1111/dom.12354
3. Justice JN, Niedernhofer L, Robbins PD, Aroda VR, Espeland MA, Kritchevsky SB, et al.. Development of Clinical Trials to Extend Healthy Lifespan. Cardiovasc Endocrinol Metab. 2018;7(4):80-83.
   PMID 30906924 · doi:10.1097/XCE.0000000000000159 · PMC6428447
4. Justice JN, Ferrucci L, Newman AB, Aroda VR, Bahnson JL, Divers J, et al.. A framework for selection of blood-based biomarkers for geroscience-guided clinical trials: report from the TAME Biomarkers Workgroup. Geroscience. 2018;40(5-6):419-436.
   PMID 30151729 · doi:10.1007/s11357-018-0042-y · PMC6294728
5. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, et al.. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346(6):393-403.
   PMID 11832527 · doi:10.1056/NEJMoa012512 · PMC1370926
6. Konopka AR, Laurin JL, Schoenberg HM, Reid JJ, Castor WM, Wolff CA, et al.. Metformin inhibits mitochondrial adaptations to aerobic exercise training in older adults. Aging Cell. 2019;18(1):e12880.
   PMID 30548390 · doi:10.1111/acel.12880 · PMC6351883
7. Kulkarni AS, Gubbi S, Barzilai N. Benefits of Metformin in Attenuating the Hallmarks of Aging. Cell Metab. 2020;32(1):15-30.
   PMID 32333835 · doi:10.1016/j.cmet.2020.04.001 · PMC7347426
8. Martin-Montalvo A, Mercken EM, Mitchell SJ, Palacios HH, Mote PL, Scheibye-Knudsen M, et al.. Metformin improves healthspan and lifespan in mice. Nat Commun. 2013;4:2192.
   PMID 23900241 · doi:10.1038/ncomms3192 · PMC3736576
9. Onken B, Driscoll M. Metformin induces a dietary restriction-like state and the oxidative stress response to extend C. elegans Healthspan via AMPK, LKB1, and SKN-1. PLoS One. 2010;5(1):e8758.
   PMID 20090912 · doi:10.1371/journal.pone.0008758 · PMC2807458
10. Triggle CR, Mohammed I, Bshesh K, Marei I, Ye K, Ding H, et al.. Metformin: Is it a drug for all reasons and diseases?. Metabolism. 2022;133:155223.
    PMID 35640743 · doi:10.1016/j.metabol.2022.155223
11. Walton RG, Dungan CM, Long DE, Tuggle SC, Kosmac K, Peck BD, et al.. Metformin blunts muscle hypertrophy in response to progressive resistance exercise training in older adults: A randomized, double-blind, placebo-controlled, multicenter trial: The MASTERS trial. Aging Cell. 2019;18(6):e13039.
    PMID 31557380 · doi:10.1111/acel.13039 · PMC6826125
12. American Federation for Aging Research. TAME Trial programme page (secondary; efficacy unpublished at 8 Aug 2026 harvest — re-verify).
    https://www.afar.org/tame-trial
13. FDA. Metformin labeling context for type 2 diabetes — not an aging approval.
    https://www.fda.gov/
14. ClinicalTrials.gov. TAME-related and metformin aging protocol inventory.
    https://clinicaltrials.gov/
15. ClinicalTrials.gov. Protocol inventory via API v2; not peer-reviewed results.
    https://clinicaltrials.gov/
16. U.S. Food and Drug Administration. Approved labeling context for primary indications — not aging approvals.
    https://www.fda.gov/
17. American Federation for Aging Research. TAME Trial programme page (secondary status communications; verify before reissue).
    https://www.afar.org/tame-trial
18. NIA Interventions Testing Program publications index.
    https://www.nia.nih.gov/research/dab/interventions-testing-program-itp/publications
19. PubChem compound summary pages used for identity cross-checks.
    https://pubchem.ncbi.nlm.nih.gov/

CFigure list

1. Timeline from Galega officinalis folk use through guanidine toxicity screening, Sterne’s Glucophage introduction, UKPDS first-line status, and the still-pending TAME trial.
2. Mechanism layers: mitochondrial complex I inhibition, AMPK signalling, and open debates including gut microbiome effects and exercise-adaptation interference.
3. TAME as trial design versus completed result at this compilation.

No commissioned photographic plates were admitted for this build; all three figures are vector diagrams generated for the monograph.