Metformin
Metformin is the world's most-prescribed oral antihyperglycemic. UKPDS 34 (1998, n=753) cut all-cause mortality 36% and myocardial infarction 39% in overweight T2D. DPP (Knowler 2002) reduced progression to T2D 31% in prediabetics. Bannister 2014 (UK CPRD) found T2D patients on metformin outlived matched non-diabetic controls. Longevity case for healthy adults is contested by exercise-interference data.
Metformin scored 6.2 / 10 (👍 Worth trying) on the BioHarmony scale as a Substance → Pharmaceutical / Drug → Other Pharmaceutical.
What It Is
Metformin is a biguanide derived from the French lilac, Galega officinalis. It is the world's most-prescribed oral antihyperglycemic, with roughly 150 million users globally. At therapeutic doses it partially inhibits mitochondrial Complex I, raising the AMP/ATP ratio and activating AMPK. That single mechanistic node cascades into suppression of hepatic gluconeogenesis, delayed intestinal glucose absorption, enrichment of gut Akkermansia muciniphila, and indirect downregulation of mTOR.
It has been first-line for type 2 diabetes for decades on the strength of UKPDS 34 (1998) and DPP / Knowler 2002. It is also the most-discussed off-label longevity drug in biohacking, largely because of Bannister 2014 (T2D patients on metformin outlived matched non-diabetic controls) and ITP mouse data. That longevity thesis has taken serious hits since 2019 from the Konopka and Walton MASTERS exercise-interference trials and is now the most contested biohacker drug in the stack. Peter Attia publicly deprescribed. David Sinclair still takes it. The TAME trial (Nir Barzilai, n=3,000 adults 65–79) has not yet read out.
Type: Pharmaceutical (biguanide class; prescription only).
Current status: Prescription required. Available via primary care, endocrinologist, or telehealth (AgelessRx, Marek Health, Lifeforce, Maximus). Generic $4/month at Costco, Walmart, Sam's Club. Brand names: Glucophage, Fortamet, Glumetza, Riomet (liquid).
Terminology
- AMPK: 5'-AMP-activated protein kinase, the cellular energy-status sensor activated when the AMP/ATP ratio rises. Metformin raises AMPK activity indirectly by partially inhibiting Complex I.
- HbA1c: Glycated hemoglobin, a rolling 3-month measure of average blood glucose. UKPDS and DPP both use HbA1c as a primary endpoint.
- HOMA-IR: Homeostatic Model Assessment of Insulin Resistance, calculated from fasting glucose and insulin.
- T2D: Type 2 diabetes mellitus. The indicated population for metformin.
- DPP: Diabetes Prevention Program, the landmark prediabetes RCT (Knowler 2002, n=3,234).
- UKPDS: United Kingdom Prospective Diabetes Study, the landmark T2D RCT series (1998).
- TAME: Targeting Aging with Metformin, the pending n=3,000 longevity RCT led by Nir Barzilai.
- VO2max: Maximum rate of oxygen consumption during incremental exercise, the standard cardiorespiratory fitness benchmark. Attenuated by metformin in older aerobically training adults per Konopka 2019.
- MALA: Metformin-associated lactic acidosis, the catastrophic tail risk (3–9 per 100,000 patient-years, 22–36% fatality when it occurs).
- eGFR: Estimated glomerular filtration rate. Metformin is contraindicated at eGFR <30 and requires dose reduction at 30–44.
- MASTERS: The Walton 2019 resistance-training RCT showing hypertrophy blunting under metformin.
Dosing & Protocols
Dosing information is summarized from published research and community reports. This is not a prescribing guide. Consult a healthcare provider before starting any protocol.
View 2 routes and 6 protocols
Routes & Forms
| Route | Form | Clinical Range | Community Range |
|---|---|---|---|
| Oral (Immediate Release) | Immediate-release tablet (generic metformin HCl) | 500–2000 mg/day, divided BID–TID with meals T2D, prediabetes, PCOS. UKPDS and DPP both used IR. | 500 mg with largest carb meal (hit-and-run) to 500–1000 mg/day off-label longevity Biohackers favor pre-meal timing for glucose-spike blunting; some pulse only on high-carb or rest days |
| Oral (Extended Release) | Extended-release tablet (Glucophage XR, Fortamet, Glumetza, generic) | 500–2000 mg once daily or divided BID with evening meal Preferred when GI tolerance is an issue. ADA-acceptable substitute. | 500 mg ER with dinner is the modal low-dose longevity protocol AgelessRx and most telehealth longevity scripts default to ER |
Protocols
Standard T2D titration (ADA, UKPDS-derived) Clinical
- Dose
- 500 mg with dinner × 1 week → 500 mg BID × 1 week → titrate by 500 mg/week to 1000 mg BID (2 g/day, efficacy ceiling for most patients)
- Frequency
- twice daily with meals
- Timing
- with breakfast and dinner to minimize GI events Package insert; reduces nausea and diarrhea vs empty-stomach dosing
- Cycling
- none; indefinite chronic dosing
- Days Off
- none
- Duration
- indefinite
Monitor eGFR annually, B12 annually after year 2, CMP annually. ER preferred if GI intolerance on IR. Max labeled dose 2550 mg/day rarely adds benefit.
Low-dose longevity (AgelessRx / community protocol) Mixed
- Dose
- 500 mg ER with largest carb meal (typically dinner)
- Frequency
- once daily
- Timing
- evening with dinner anecdotal; pharmacokinetic rationale (peak 4–8h post-dose)
- Cycling
- some users skip dosing on training days or 48h around heavy resistance sessions
- Days Off
- optional 1–2 rest-day skips per week
- Duration
- indefinite
Off-label. No RCT in metabolically healthy adults supports this dose. TAME pending. B12 monitoring still mandatory after year 2.
Hit-and-run (Nick's preferred framing if incorporated) Anecdotal
- Dose
- 500 mg ER with specific high-carb or indulgence meals
- Frequency
- episodic, non-daily
- Timing
- with or immediately before the target meal pharmacokinetic: glycemic effect peaks 2–4h post-dose, washes out within 24h
- Cycling
- dose only on target meals; no chronic exposure
- Days Off
- all non-target days
- Duration
- indefinite, as-needed
Not a studied protocol. Rationale: avoid chronic exercise-interference signal while capturing glycemic blunting on specific meals. B12 risk minimal at this exposure.
PCOS protocol Clinical
- Dose
- 1500–2000 mg/day divided
- Frequency
- 2–3× daily with meals
- Timing
- with meals to minimize GI events package insert, ADA guidance
- Cycling
- none; indefinite adjunct to lifestyle
- Days Off
- none
- Duration
- indefinite
Restores ovulation, reduces androgens, supports weight loss in insulin-resistant PCOS phenotype. ER formulation preferred.
TAME protocol (Barzilai, pending readout) Clinical
- Dose
- 850 mg BID (1700 mg/day total)
- Frequency
- twice daily
- Timing
- with meals trial protocol
- Cycling
- none
- Days Off
- none
- Duration
- 6 years (trial duration)
n=3,000 adults 65–79. Composite primary endpoint: cardiovascular, cancer, cognitive decline, mortality. Results 2026–2028.
Blueprint stack (Bryan Johnson) Anecdotal
- Dose
- 1500 mg ER AM + 500 mg ER PM (2000 mg/day)
- Frequency
- twice daily
- Timing
- AM + PM individual protocol, self-reported
- Cycling
- none
- Days Off
- none
- Duration
- indefinite
Stacked with acarbose (rapamycin since discontinued). High-dose non-diabetic use; outlier protocol.
Use-Case Specific Dosing
| Use Case | Dose | Notes |
|---|---|---|
| Blood Sugar | 1000 mg BID | Gold standard for HbA1c reduction |
| Metabolic Health | 500–2000 mg/day titrated per tolerance | Indicated population scoring |
| Longevity | 500–1000 mg/day | Off-label; no established clinical protocol; contested by Konopka/Walton in active adults |
| Body Composition | 1000–2000 mg/day | Modest 2–3 kg effect over 1 year in insulin-resistant population |
| Hormonal | 1500–2000 mg/day | Androgen lowering and ovulation restoration in insulin-resistant PCOS |
How this score is calculated →
Upside (2.64 / 5.00)
| Dimension | Weight | Score | Visual | Weighted |
|---|---|---|---|---|
| Efficacy | 25% | 4.0 | 1.000 | |
| Breadth of Benefits | 15% | 3.2 | 0.480 | |
| Evidence Quality | 25% | 4.7 | 1.175 | |
| Speed of Onset | 10% | 3.0 | 0.300 | |
| Durability | 10% | 2.0 | 0.200 | |
| Bioindividuality Upside | 15% | 3.2 | 0.480 | |
| Total | 3.635 |
Upside Rationale
Efficacy (4.0/5.0). Transformative in the indicated population. UKPDS 34 (1998, n=753) delivered HbA1c reductions of 1.0–1.5% as monotherapy, with a 36% reduction in all-cause mortality and 39% reduction in myocardial infarction versus conventional diet therapy. In prediabetes, DPP (Knowler 2002, n=3,234) reduced progression to T2D by 31% over 2.8 years, rising to 44% in the BMI ≥30 under-60 subgroup. Weight effect: 2–3 kg over 1 year. Per v0.5 audience-vs-indication rules, score reflects the indicated T2D/prediabetic/PCOS/IR population where effect sizes clear Cohen's d > 0.8 territory. Healthy-adult longevity case is a separate question handled in Verdict.
Breadth (3.2/5.0). Glycemic control, weight, PCOS (ovulation restoration, androgen reduction), NAFLD, cardiovascular (UKPDS composite), possible cancer risk reduction (Bridges 2014 meta-analysis showed 31% reduction in overall cancer incidence in T2D), indirect mTOR reduction, microbiome shift (Akkermansia enrichment). Does not meaningfully touch immune, cognitive, or structural pathways outside T2D. The longevity-breadth story is observational and confounded by T2D baseline.
Evidence Quality (4.7/5.0). Strongest evidence tier in this entire series. Over 40,000 indexed publications on PubMed. UKPDS and DPP are among the most cited intervention trials in medicine. Cochrane review exists. 20+ meta-analyses. Nearly 70 years of post-market surveillance. FAERS logs over 40,000 adverse event reports tracked continuously. Per v0.5 evidence-integrity rules, no penalty applies: metformin has been generic since the 1970s and has no big-pharma sponsorship machinery behind its research base. UKPDS, DPP, Bannister, Konopka, Walton are all independent academic work. Major gap: TAME is pending, so longevity causality in non-diabetics remains inferred.
Speed of Onset (3.0/5.0). Fasting glucose drops within 1–2 weeks of titration per UKPDS 34 (1998). HbA1c reduction is measurable at 3 months, the standard ADA clinical-trial readout window. Weight-loss signal emerges over 6–12 months (DPP Knowler 2002 trajectory, terminal effect around 2–3 kg). Cancer-incidence signal requires multi-year follow-up (Bridges 2014 meta-analysis, 5-year minimum). Longevity and all-cause mortality signal (Bannister 2014) requires over 10 years of continuous exposure. Metformin is fast for glycemic endpoints, slow for biohacker and longevity endpoints. Not a same-day intervention for any application.
Durability (2.0/5.0). As long as you take it, glycemic and insulin-sensitivity benefits persist, with 30-year follow-up continued CV benefit among sustained users per UKPDS Post-Trial Monitoring 2008. Stopping returns glycemic control to diabetes trajectory within weeks. The DPP Outcomes Study (2015) showed lasting benefit only in the continuous-treatment arm. There is no "banking" of benefits over a finite course: this is a chronic therapy, not a protocol with washout legacy. For T2D patients the durability is intrinsically life-long medication territory. No permanent tissue remodeling accrues to reward a finite cycle.
Bioindividuality (3.2/5.0). Scored for the indicated responder phenotype under v0.5. T2D, prediabetic, PCOS, and insulin-resistant populations are strong responders with response rates well above 80% for HbA1c reduction at tolerated doses. Dose-titration compliance is the main responder-status question within the indicated population. The healthy-metabolically-fit-trained-adult non-response (and exercise interference) is a separate audience filter that belongs in Verdict, not the dimension score.
Downside (2.24 / 5.00)
| Dimension | Weight | Score | Visual | Weighted |
|---|---|---|---|---|
| Safety Risk | 30% | 4.0 | 1.200 | |
| Side Effect Profile | 15% | 3.0 | 0.450 | |
| Financial Cost | 5% | 1.2 | 0.060 | |
| Time/Effort Burden | 5% | 1.2 | 0.060 | |
| Opportunity Cost | 5% | 2.0 | 0.100 | |
| Dependency / Withdrawal | 15% | 3.0 | 0.450 | |
| Reversibility | 25% | 1.2 | 0.300 | |
| Total | 2.620 | |||
| Harm subtotal × 1.4 | 3.360 | |||
| Opportunity subtotal × 1.0 | 0.220 | |||
| Combined downside | 3.580 | |||
| Baseline offset (constant) | −1.340 | |||
| Effective downside penalty | 2.240 |
Downside Rationale
Safety Risk (4.0/5.0). Catastrophic floor triggered by MALA. FAERS 2018–2024 logged 10,370 MALA cases out of 40,883 total metformin adverse event reports. Contemporary 30-day mortality is 22–36% when MALA occurs (historical series topped 50%). Population incidence is 3–9 per 100,000 patient-years. Risk concentrates in specific contexts: eGFR <30 (absolute contraindication), eGFR 30–44 (dose reduction required), decompensated heart failure, severe hepatic disease, sepsis, active alcohol abuse, and IV iodinated contrast (hold 48 hours if CKD present). Drug interaction signals from FAERS: IV contrast ROR 323.7, furosemide ROR 155.4, topiramate (package insert contraindication), cimetidine (OCT2/MATE1 competition). Per v0.5 catastrophic-floor rules the score sits at 4.0 (the floor), not stacked higher, because MALA is a single intrinsic catastrophic cluster. B12-depletion-driven neuropathy is scored in Side Effects, not Safety.
Side Effect Profile (3.0/5.0). GI events hit 20–30% of users in the first weeks: diarrhea, nausea, cramping, flatulence, metallic taste. Extended-release formulations cut GI events roughly in half. A minority (3–5%) never tolerate even ER. Chronic B12 depletion requires annual monitoring and supplementation: All of Us 2025 (n=14,808) found 67% higher odds of B12 deficiency and 39% higher peripheral neuropathy prevalence in users at 4+ years. Per v0.5 catastrophic-floor-not-stacked rule, Side Effects is scored independently of the Safety 4.0 floor; GI plus long-term B12 load lands cleanly at 3.0.
Financial Cost (1.2/5.0). Generic $4/month at Costco, Walmart, Sam's Club. ER generic $15–30/month. Telehealth off-label scripts run $30–80/month including consult. Per v0.5 cost-at-accessible-channels rule, scored at the generic pharmacy channel that virtually all users can actually access; the brand-name Glucophage pricing is not representative.
Time/Effort Burden (1.2/5.0). Standard titration is 500 mg once daily escalating to 2000 mg divided doses over 2–4 weeks, per UKPDS 34 and ADA 2025 guidance. Immediate-release is taken 2–3 times daily with meals; extended-release (Glucophage XR, Glumetza) is once daily with dinner. Clinical monitoring is eGFR every 6–12 months for renal safety and annual B12 after year 2 per ADA 2025 standards because of chronic malabsorption risk. Once titrated, daily friction is under 60 seconds. Compared to insulin (injection effort) or SGLT-2i (same pill effort), metformin sits at the low-burden end. Most of the effort is upfront titration and periodic labs, not daily dosing.
Opportunity Cost (2.0/5.0). Scored for the indicated population under v0.5 audience-vs-indication rules. For a T2D or insulin-resistant patient, the opportunity cost of NOT using metformin is life-years lost (CV mortality, renal failure, retinopathy) and the OC of the drug itself is low. The biohacker-audience opportunity cost (exercise-interference signal, better-studied SGLT2i or GLP-1 alternatives for some metabolic goals) is meaningful but belongs in Verdict / Avoid-if, not the dimension score.
Dependency/Withdrawal (3.0/5.0). Functional dependency under the v0.5 framework. For the indicated T2D patient, stopping returns the user to diabetic baseline with resumption of hyperglycemia, CV risk, and complication trajectory - a real need-to-continue signal on the order of SSRIs or HRT, but lower than thyroid hormone or insulin because the underlying pathology (not the drug) drives the dependence. No reward-circuit hijack, no craving, no tolerance escalation; this is not addiction. For healthy biohacker use the dependency score would be 1.5–2.0 (pure reversibility back to baseline), but v0.5 requires scoring the indicated population where functional dependency is the correct frame.
Reversibility (1.2/5.0). Plasma half-life is 4–9 hours (Graham 2011 clinical pharmacokinetics review), with full systemic clearance within 24–48 hours of cessation. Glycemic control returns to pre-treatment baseline within days to weeks for indicated T2D patients; no insulin-resistance reversal persists. B12 recovers over months with supplementation (Chapman 2016). Exercise-adaptation blunting fully reverses within 4–8 weeks post-discontinuation per Walton 2019 MASTERS reversal data. No permanent tissue remodeling, no mitochondrial memory effect, no epigenetic lock-in. Contrast with SGLT-2 inhibitors (some renoprotection persists) or insulin (receptor adaptations partially persist).
Verdict
✅ Best for: Adults with Type 2 diabetes (first-line, decades of RCT data). Prediabetes with BMI ≥30 or fasting glucose 100–125 mg/dL (DPP supports this). PCOS with insulin resistance (ovulation restoration, androgen reduction). Metabolic syndrome with HbA1c ≥5.7%. Gestational diabetes per FDA Category B and 2025 ADA. Pair with annual eGFR, CMP, and B12 monitoring after year 2. Methylcobalamin supplementation is effectively mandatory for long-term users.
🤔 Case-by-case for longevity-curious healthy adults: Sporadic hit-and-run dosing around specific high-carb or indulgent meals rather than daily chronic dosing. Wait for TAME readout before committing to chronic use. Consider SGLT2i or GLP-1 class alternatives if the underlying goal is cardiometabolic risk reduction rather than glycemic control specifically.
❌ Avoid if: Athletic-performance-focused, actively training for hypertrophy, or optimizing VO2max (Konopka 2019 + Walton 2019 MASTERS show measurable blunting of the adaptations you are paying for in the gym - this is the biohacker opportunity-cost that lives here rather than in the dimension score). eGFR <30 (absolute contraindication, MALA risk). Decompensated congestive heart failure. Active alcohol abuse or chronic liver disease. 48 hours before IV iodinated contrast if CKD present. Any surgery in the next 48 hours. Acute MI, sepsis, or decompensated hypoxic states. Competitive athletes lifting to failure or training for aerobic adaptation should treat daily metformin as a negative-ROI intervention per current RCT evidence.
Use Case Breakdown
The overall BioHarmony score reflects the intervention's primary evidence profile. These subratings are independent assessments per use case.
| Use Case | Score | Summary |
|---|---|---|
| ✅ Blood Sugar / Glycemic Control | 9.0 | HbA1c reduction 1.0–1.5% absolute as monotherapy; 70-year clinical track record; highest-weight signal in this file. |
| ✅ Metabolic Health | 8.5 | Gold-standard T2D monotherapy; DPP 31% diabetes prevention; first-line ADA/EASD for decades. |
| 💪 Cardiovascular | 7.2 | UKPDS 34 showed 39% MI reduction and 36% all-cause mortality reduction in overweight T2D over 10 years. Independent CV benefit beyond glycemic control is real in the indicated population. |
| 💪 Hormonal / Endocrine | 7.0 | PCOS insulin-resistant phenotype: restores ovulation, lowers androgens, supports weight loss. Clinically validated. |
| 👍 Fertility (Female) | 6.8 | PCOS insulin-resistant phenotype: restores ovulation, improves fertility outcomes. Clinically validated use case. |
| 👍 Healthspan | 6.5 | UKPDS CV composite + DPP prevention + PCOS + observational cancer signal (Bridges 2014 meta: 31% cancer reduction in T2D). Indicated population only. |
| 👍 Geriatric / Aging Population | 6.5 | First-line in T2D regardless of age above eGFR thresholds. TAME population is specifically 65–79. Dose reduction required at eGFR 30–44. |
| 👍 Longevity / Lifespan | 6.2 | Bannister 2014 (UK CPRD): T2D patients on metformin outlived non-diabetic controls. TAME pending. Healthy-adult longevity case contested by Konopka/Walton and MILES null. |
| 👍 Gut Health / Microbiome | 6.0 | Enriches Akkermansia muciniphila and SCFA-producers (Bifidobacterium, Megasphaera). Microbiome shift partly explains glycemic benefit per 2017 Diabetes Care work. |
| 👍 Autophagy | 6.0 | AMPK/mTOR dual modulation is classical autophagy-inducing pharmacology; downstream flux data in humans is indirect. |
| 👍 Prenatal (Maternal & Fetal Outcomes) | 6.0 | FDA Category B. Used in gestational diabetes and PCOS pregnancy. Safer profile than insulin in some GDM contexts per 2025 ADA. |
| 👍 Anti-Inflammatory | 5.8 | AMPK-mediated anti-inflammatory signaling; CRP reductions in T2D trials; no clean RCT in non-diabetics. |
| 👍 Cellular Senescence | 5.8 | mTOR suppression is senescence-adjacent; AMPK activation; mechanistic senolytic/senomorphic candidate. Human senescence endpoint data thin. |
| ⚖️ Liver / Detoxification | 5.5 | Hepatic gluconeogenesis suppression is a primary mechanism; small NAFLD/MASLD benefit signal. Not a direct detox intervention. |
| ⚖️ Eye / Vision Health | 5.5 | Observational: slower diabetic retinopathy progression; mechanistic plausibility via glycemic control. |
| ⚖️ Body Composition / Fat Loss | 5.5 | Modest 2–3 kg weight loss over 1 year in insulin-resistant adults; minimal effect in metabolically healthy. |
| ⚖️ Pediatric Use | 5.5 | FDA-approved for ages 10+ T2D. Pediatric PCOS use off-label. |
| ⚖️ Neuroprotection | 5.4 | Mechanistic AMPK + mTOR + insulin sensitization; observational in diabetics; no RCT in non-diabetics. MILES failed to measure cognition. |
| ⚖️ Immune Function | 5.0 | AMPK activation has mechanistic anti-inflammatory signal; no clinically validated immune outcome RCT. |
| ⚖️ Hair / Nail Health | 5.0 | Hirsutism reduction downstream of PCOS androgen lowering; indirect. |
| ⚖️ Cognition / Focus | 5.0 | Mixed: observational AD risk reduction in T2D (Xu 2021 meta); MILES null in non-diabetics; 2024 mouse AD model showed impaired cognition long-term. Unsettled. |
| ⚖️ Neuroplasticity | 5.0 | mTOR modulation is plasticity-adjacent; no human neuroplasticity RCT. |
| ⚖️ Telomere / DNA Repair | 5.0 | MILES trial measured epigenetic aging; no significant benefit in non-diabetics at 12 weeks. |
| ⚖️ Wound Healing | 5.0 | Mixed: glycemic control aids diabetic wound healing; AMPK activation neutral in otherwise healthy adults. |
| ⚖️ Memory | 4.8 | AD observational signal in T2D; human memory RCT data absent. Mouse AD-model signal is concerning. |
| ⚖️ Mood / Emotional Regulation | 4.8 | No direct signal; metabolic improvement in T2D may indirectly lift mood. |
| ⚖️ Depression | 4.8 | Small observational signal of lower depression in T2D; B12 depletion can worsen mood at 4+ years if uncorrected. |
| ⚖️ Sleep Quality | 4.8 | No consistent sleep signal. |
| ⚖️ Antioxidant / Oxidative Stress | 4.8 | Indirect via mitochondrial load reduction; no direct antioxidant RCT. |
| ⚖️ Methylation Support | 4.8 | B12 depletion at 4+ years can impair methylation downstream; mechanism via reduced ileal B12 absorption, not direct methylation benefit. |
| ⚖️ Circadian Rhythm / Chronobiology | 4.8 | AMPK activation intersects circadian machinery; no direct chronobiology RCT. |
| ⚖️ HRV / Vagal Tone / Autonomic Balance | 4.8 | No direct HRV evidence. |
| ○ Bone / Joint Health | 4.5 | Observational fracture-risk reduction in T2D; not a primary use case. |
| ○ Skin / Beauty | 4.5 | AMPK-mediated anti-aging skin signals preclinical; not clinically validated. |
| ○ Stress / Resilience | 4.5 | No meaningful evidence. |
| ○ Anxiety | 4.5 | No meaningful evidence. |
| ○ Sleep Architecture (Deep/REM) | 4.5 | No direct evidence. |
| ○ Flow State / Peak Mental Performance | 4.5 | No direct evidence. |
| ○ Creativity / Divergent Thinking | 4.5 | No direct evidence. |
| ○ Recovery / Repair | 4.5 | mTORC1 blunting antagonizes the adaptive recovery signal that exercise is supposed to induce. |
| ○ Flexibility / Mobility | 4.5 | No direct evidence. |
| ○ Reaction Time / Coordination | 4.5 | No direct evidence. |
| ○ Stem Cell Support | 4.5 | Preclinical stem-cell signals; no human RCT. |
| ○ Chronic Pain Management | 4.5 | Diabetic neuropathy improvement is secondary to glycemic control, not direct analgesia. |
| ○ Energy / Fatigue | 4.5 | Some users report fatigue at 1 g+/day; T2D users often report energy improvement from glycemic normalization. |
| ○ Libido / Sexual Health | 4.5 | No direct evidence in primary indication; PCOS responders may see hormonal normalization benefits. |
| ○ Fertility (Male) | 4.5 | No direct evidence. |
| ○ Cold / Heat Tolerance / Hormesis | 4.5 | No direct evidence. |
| ○ Respiratory | 4.0 | No meaningful respiratory evidence. |
| ○ Dental / Oral Health | 4.0 | No meaningful evidence. |
| ○ Hearing / Auditory | 4.0 | No meaningful evidence. |
| ○ Lymphatic / Drainage | 4.0 | No meaningful evidence. |
| ○ Acute Pain Relief | 4.0 | No evidence. |
| ○ Injury Recovery | 4.0 | Likely negative in trained adults due to blunted adaptive signaling. |
| ○ Heavy Metal / Toxin Burden | 4.0 | No direct evidence. |
| ○ Electromagnetic / Frequency Therapy | 4.0 | Not applicable. |
| ○ Social Bonding / Empathy | 4.0 | Not applicable. |
| ○ Spiritual / Consciousness Expansion | 4.0 | Not applicable. |
| ○ Traumatic Brain Injury | 4.0 | No direct evidence. |
| ○ Endurance / Cardio | 3.8 | Konopka 2019 (Aging Cell, n=53, 12-week RCT): ~50% attenuation of VO2max gains; abolition of Complex I-linked respiration improvement from aerobic training. |
| ○ Mitochondrial | 3.5 | Direct Complex I inhibition is the mechanism that blunts exercise-induced mitochondrial adaptation (Konopka 2019). Bidirectional: glycemic benefit AND adaptive penalty from the same node. |
| ○ Nerve Regeneration | 3.5 | B12 depletion can cause iatrogenic peripheral neuropathy at 4+ years per All of Us 2025 (n=14,808, +39% prevalence long-term vs short-term users). |
| ○ Strength / Power | 3.5 | Walton 2019 MASTERS (n=94, 14-week RCT in adults 65+): placebo added 3.90% thigh CSA from resistance training, metformin arm added 0.45%. Hypertrophy effectively blunted. |
| ○ VO2 Max | 3.5 | Konopka 2019: placebo +2.7 mL/kg/min vs metformin +1.3 mL/kg/min over 12 weeks aerobic training (p=0.03). Negative signal in trained/training adults. |
| ○ Muscle Growth / Hypertrophy | 3.0 | MASTERS trial: hypertrophy response collapsed from 3.90% to 0.45% of thigh CSA. Actively negative for lifters. |
Frequently Asked Questions
What is metformin and how does it work?
Metformin is a biguanide derived from the French lilac (Galega officinalis) and the world's most-prescribed oral antihyperglycemic, taken by roughly 150 million people. At therapeutic doses it partially inhibits mitochondrial Complex I, which raises the AMP/ATP ratio and activates AMPK. This suppresses hepatic gluconeogenesis, delays intestinal glucose absorption, enriches gut Akkermansia muciniphila, and indirectly downregulates mTOR signaling. It has been FDA-approved since 1995 and first-line for type 2 diabetes per ADA guidance.
How effective is metformin for type 2 diabetes and prediabetes?
In T2D, UKPDS 34 (1998, n=753 overweight patients, 10-year follow-up) delivered HbA1c reductions of 1.0–1.5% as monotherapy, with a 36% reduction in all-cause mortality and 39% reduction in myocardial infarction versus conventional diet therapy. In prediabetes, the Diabetes Prevention Program (Knowler 2002, n=3,234) reduced progression to T2D by 31% over 2.8 years, rising to 44% in the BMI ≥30 under-60 subgroup. Weight effect is modest at 2–3 kg over one year.
What's the right dose of metformin, and does extended-release matter?
Standard T2D titration is 500 mg with dinner for one week, 500 mg BID for one week, then titrate by 500 mg/week toward 1000 mg BID (2 g/day), which is the efficacy ceiling for most patients. The labeled maximum is 2550 mg/day but rarely adds benefit. Extended-release formulations cut GI adverse events roughly in half and are the preferred starting form for most users. Low-dose longevity use is typically 500–1000 mg/day; the TAME trial uses 1700 mg/day in adults 65–79.
Does metformin extend lifespan in healthy people?
Bannister 2014 (UK CPRD cohort) found T2D patients on metformin outlived matched non-diabetic controls, the most-cited observational longevity datapoint. ITP mouse data shows lifespan extension at some doses. However, this has not been tested in healthy humans in a randomized trial. The TAME trial (n=3,000 adults 65–79, 1,700 mg/day, multimorbidity composite endpoint) has not read out. Until then the non-diabetic longevity case remains inferred, not established.
What are the long-term safety risks of metformin?
The catastrophic risk is metformin-associated lactic acidosis (MALA): base rate 3–9 per 100,000 patient-years but 22–36% mortality when it occurs. FDA Boxed Warning is in place. Risk concentrates in eGFR <30, decompensated CHF, severe hepatic disease, IV iodinated contrast, sepsis, and alcohol abuse. The slower chronic risk is vitamin B12 depletion: All of Us 2025 (n=14,808) found +67% odds of B12 deficiency and +39% peripheral neuropathy in users at 4+ years. Annual B12, CMP, and eGFR monitoring is mandatory.
Does metformin interfere with exercise adaptations?
Yes, per two RCTs in older adults. Konopka 2019 (Aging Cell, n=53, 12 weeks): metformin attenuated VO2max gains by roughly 50% (placebo +2.7 mL/kg/min vs metformin +1.3 mL/kg/min, p=0.03) and abolished Complex I respiration improvements. Walton 2019 MASTERS (Aging Cell, n=94, 14 weeks resistance training): placebo added 3.90% thigh muscle CSA while metformin added 0.45% (p<0.001). The mechanism is Complex I inhibition blunting mTORC1 signaling. For healthy adults training for hypertrophy or aerobic fitness, daily metformin is working against the primary adaptation.
Does metformin deplete vitamin B12?
Yes, chronically. Metformin reduces ileal calcium-dependent B12–intrinsic factor absorption. All of Us 2025 (n=14,808) found long-term users (≥4 years) had 67% higher odds of B12 deficiency versus non-users and 39% higher odds of peripheral neuropathy versus short-term users. Risk climbs roughly 3% per year from diagnosis. The UK MHRA now recommends periodic B12 monitoring in all long-term users. Supplement methylcobalamin (or cyanocobalamin) and check serum B12 plus methylmalonic acid annually after year 2.
Immediate-release vs extended-release metformin: which should I start with?
Start with extended-release unless cost or supply constraints force immediate-release. ER formulations (Glucophage XR, Fortamet, Glumetza, generic metformin ER) cut GI events roughly in half: diarrhea, nausea, cramping, and metallic taste all drop meaningfully. IR peaks at 2–3 hours and washes out faster, making it usable for hit-and-run meal-specific dosing. ER peaks at 4–8 hours and is smoother for once-daily chronic dosing. Both carry equivalent B12 depletion risk. Generic IR is $4/month at Costco/Walmart; generic ER is typically $15–30/month.
How This Score Could Change
BioHarmony scores are living assessments. New research, regulatory changes, or personal context can shift the score up or down. These are the most likely scenarios that would change this intervention's rating.
| Scenario | Dimension shifts | New score |
|---|---|---|
| TAME reads out positive for geroprotection in non-diabetics | Efficacy 4.0→4.3, Evidence 4.7→5.0, Bioindividuality 3.2→3.8 | 7.0 / 10 (💪 Strong recommend) |
| TAME null result in non-diabetics | Efficacy 4.0→3.5, Bioindividuality 3.2→2.6 | 5.8 / 10 (👍 Worth trying, lower bound) |
| Konopka/Walton replicated in larger multisite RCT confirming exercise interference | Opportunity 2.0→2.8, Bioindividuality 3.2→2.8 | 5.9 / 10 (👍 Worth trying) |
| SGLT2i head-to-head shows superior mortality in healthy-ish adults | Efficacy 4.0→3.5, Opportunity 2.0→2.8 | 5.7 / 10 (⚖️ Neutral) |
| Time-restricted or hit-and-run metformin protocol rescues exercise adaptation in RCT | Opportunity 2.0→1.5, Bioindividuality 3.2→3.6 | 6.4 / 10 (👍 Worth trying) |
| FAERS MALA signal worsens with new subgroup (e.g., GLP-1 co-administration) | Safety 4.0→4.5 | 6.0 / 10 (👍 Worth trying) |
| Cochrane review of non-diabetic metformin use publishes critical finding | Evidence 4.7→4.0 | 5.9 / 10 (👍 Worth trying) |
| Large observational confirms cancer-prevention signal in non-diabetics (Bridges 2014 meta extended) | Breadth 3.2→4.0, Efficacy 4.0→4.3 | 6.8 / 10 (👍 Worth trying) |
Key Evidence Sources
- UKPDS 34 (1998). Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes. Lancet 352(9131):854–65.. Landmark 10-year T2D RCT (n=753). All-cause mortality −36%, MI −39%, HbA1c reduction 1.0–1.5%. Foundation for first-line T2D indication.
- Knowler WC et al. 2002. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin (Diabetes Prevention Program). N Engl J Med 346(6):393–403.. DPP landmark RCT (n=3,234 prediabetics). Metformin cut progression to T2D by 31% over 2.8 years, 44% in BMI ≥30 under-60 subgroup.
- Diabetes Prevention Program Research Group 2012. Long-term safety, tolerability, and weight loss associated with metformin in the DPP Outcomes Study. Diabetes Care 35(4):731–7.. 10-year DPPOS follow-up: metformin benefit persisted; safety profile favorable; B12 depletion signal emerged.
- Bannister CA et al. 2014. 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 16(11):1165–73.. UK CPRD cohort (n=78,000+ T2D on metformin vs matched non-diabetic controls). T2D metformin users outlived non-diabetic controls. The headline longevity observational datapoint.
- Konopka AR et al. 2019. Metformin inhibits mitochondrial adaptations to aerobic exercise training in older adults. Aging Cell 18(1):e12880.. RCT (n=53 adults 62–79, 12 weeks aerobic training). ~50% attenuation of VO2max gains; abolition of Complex I respiration improvement. PMID 30548390.
- Walton RG et al. 2019. Metformin blunts muscle hypertrophy in response to progressive resistance exercise training in older adults: MASTERS trial. Aging Cell 18(6):e13039.. RCT (n=94 adults 65+, 14 weeks). Placebo thigh CSA +3.90% vs metformin +0.45% (p<0.001). Hypertrophy effectively blunted. PMID 31557380.
- Campbell JM et al. 2017. Metformin use associated with reduced risk of dementia in patients with diabetes: a systematic review and meta-analysis. J Alzheimers Dis 65(4):1225–1236.. Meta-analysis of metformin and dementia risk in T2D. Observational signal only; not a cancer meta but covers cognitive longevity.
- Bridges HR et al. 2014. Effects of metformin and other biguanides on oxidative phosphorylation in mitochondria. Biochem J 462(3):475–87.. Mechanistic Complex I inhibition kinetics. Basis for understanding both therapeutic and exercise-interference effects.
- Barzilai N et al. 2016. Metformin as a tool to target aging (TAME trial design). Cell Metab 23(6):1060–1065.. TAME design paper. 3,000 adults 65–79, 1,700 mg/day, 6-year follow-up, multimorbidity composite primary endpoint. Results pending 2026–2028.
- Nguyen HT et al. 2025. Long-term metformin use, vitamin B12 deficiency, and peripheral neuropathy: All of Us database analysis (n=14,808). Diabetes Res Clin Pract.. Long-term users (≥4 years): +67% odds B12 deficiency vs non-users, +39% peripheral neuropathy vs short-term users. Mandates B12 monitoring.
- FAERS MALA comprehensive analysis 2018–2024. PMC12499714. Medicine (Baltimore) 2025.. 10,370 MALA cases out of 40,883 metformin AE reports. IV contrast ROR 323.7, furosemide ROR 155.4. Median time to onset 106 days.
- Kulkarni AS et al. 2020. Benefits of metformin in attenuating the hallmarks of aging. Ageing Res Rev 61:101095.. Review of metformin's effect on cognition sex-stratified; flags that MILES trial did not include cognitive endpoints.
- FDA/DailyMed Glucophage (metformin HCl) Prescribing Information. Initial approval 1995; revised 2022.. FDA Boxed Warning for MALA. Contraindications: eGFR <30, active hepatic impairment, IV contrast within 48h if CKD, active alcohol abuse.
- ADA Standards of Care in Diabetes 2025. Pharmacologic Approaches to Glycemic Treatment.. First-line T2D pharmacologic recommendation; dose reduction guidance; B12 monitoring recommendation.
- Peter Attia MD. A recent metformin study casts doubts on longevity indications. 2023.. Public deprescription rationale; cites Konopka and Walton as decisive for physically active non-diabetic patients.
Other interventions for Metabolic Health
See all ratings →📊 How BioHarmony scoring works
BioHarmony translates a weighted expected-value calculation into a reader-facing 0–10 score. 5.0 is neutral (benefits and risks balance). Above 5 = benefits outweigh risks; below 5 = risks outweigh benefits.
Harm-type downsides (safety risk, side effects, reversibility, dependency) carry a 1.4× precautionary multiplier. Harm weighs more than benefit. Opportunity-type downsides (financial cost, time/effort, opportunity cost) are subtracted at face value.
Use case subratings are independent assessments of how well the intervention addresses specific health goals. They are not components of the overall score. Each subrating reflects the scorer's judgment based on use-case-specific evidence, safety, and effect sizes.
Every dimension is evaluated on a 1–5 scale, and the baseline (1) is subtracted before weighting. A perfect intervention with zero downsides contributes zero penalty rather than a residual floor, so top-tier scores are actually reachable.
EV = Upside − Downside
EV = 2.640 − 2.240 = 0.400
EV ranges from −5 to +5. Adding 7 shifts to 2–12, dividing by 12 normalizes to 0–1, then ×10 gives the 0–10 score.
Score = ((0.400 + 7) / 12) × 10 = 6.2 / 10
Further learning
15 Best Natural Metformin Alternatives (Powerful & Safer) for Blood Sugar & Longevity Support
The anti-aging and blood sugar lowering drug Metformin has dangerous side-effects. Try these natural supplement alternatives to keep blood sugar stable instead.
How AI, Genetics & Prevention Are Transforming Longevity Medicine
Dr. Gregory Burzynski of resTOR Longevity Clinic explains how AI, genomics, and advanced diagnostics are transforming preventive medicine. Learn why genes account for only…
