Metformin for Obesity Treatment

Summary: Metformin produces modest, durable weight loss of roughly 2 to 3 kg long-term and prevents 31% of diabetes cases in the DPP trial, making it the cheapest legitimate obesity-adjacent medication even though it is not FDA-approved for weight loss.

This content is for informational purposes only and is not medical advice. Always consult a qualified healthcare provider before starting, changing, or stopping any medication.

The short version: metformin is FDA-approved for type 2 diabetes, not for weight loss. Used off-label for obesity it produces a modest but durable 2 to 3 kg of weight loss, costs about $4 a month at Walmart, and has the longest safety track record of any oral metabolic drug. It is nowhere near a GLP-1 in magnitude. It is also not nothing, and for the right patient it is the right first move.

This page is the realistic, evidence-graded answer to a question telehealth marketing keeps fuzzing up: should you take metformin for obesity, what should you expect, and when does it make sense versus a GLP-1.

What the trials actually show

The cleanest data on metformin and weight come from the Diabetes Prevention Program, which is not even an obesity trial. The DPP randomized 3,234 adults with prediabetes (mean BMI 34) to placebo, metformin 1,700 mg/day, or intensive lifestyle ^[1]. Lifestyle won at year three with 5.6 kg of weight loss. Metformin produced 2.1 kg. Placebo produced 0.1 kg.

The interesting part comes later. In the Diabetes Prevention Program Outcomes Study, the long-term follow-up, the lifestyle group regained most of its weight (final loss around 2.0 kg at ten years). The metformin group kept losing slowly and held steady at about 2.5 kg below baseline ^[2]. Roughly 30% of metformin participants lost more than 5% of body weight in year one, and that subgroup averaged 6.2% loss at fifteen years.

So the headline number people quote, "5 to 10 pounds on metformin," is honest. The trial average is closer to 4 to 6 lb. Responders see 10 to 15 lb. Non-responders see nothing.

The comparison is the whole story. Metformin gets you the same order of magnitude as bupropion-naltrexone at one-fiftieth the price, and one-fifth to one-seventh the result of a GLP-1.

How it works: AMPK, the gut, and a small GLP-1 nudge

Metformin's mechanism is still partially mapped, which is unusual for a drug that has been prescribed since 1957. The current best understanding stacks three things.

AMPK activation. Metformin enters hepatocytes via the OCT1 transporter and partially inhibits complex I of the mitochondrial electron transport chain. The resulting drop in cellular ATP activates AMP-activated protein kinase, which suppresses hepatic gluconeogenesis and shifts the liver toward fatty acid oxidation. This is the classic glycemic mechanism. Weight effects flow from it indirectly.

Gut microbiome shifts. Metformin reaches concentrations in the small intestine that are 30 to 300 times higher than in blood. It alters the composition of the gut flora, increasing short-chain fatty acid producers like Akkermansia muciniphila and reducing certain inflammatory species. Stool transplant from metformin-treated donors to untreated mice transmits part of the metabolic benefit. That is strong evidence the gut is not a side stage.

Modest GLP-1 increase. Metformin raises endogenous GLP-1 secretion by intestinal L-cells, roughly 1.5 to 2 fold in fasted and postprandial states. It is a small fraction of what an injected GLP-1 agonist delivers (which sits at pharmacologic, not physiologic, concentrations), but it is the same hormone, and it explains some of the appetite-suppressing effect ^[3].

A fourth pathway, induction of growth-differentiation factor 15, has been documented in both animal and human studies. GDF15 acts on hindbrain receptors to reduce food intake and has been proposed as the main driver of metformin's weight effect. The full causal chain is still under active debate.

The practical takeaway: metformin's appetite suppression is real but quiet. Patients describe feeling slightly less hungry between meals, not the food-noise silence that GLP-1 users describe.

Realistic expectations on a dose-by-dose basis

Weight loss with metformin starts around week 4 and is mostly complete by month 6 to 12 ^[3]. There is no continuing slope after year one in most patients. If you have lost nothing at six months on a therapeutic dose, you are unlikely to lose anything at twelve.

Dose matters. The DPP used 1,700 mg/day. Meta-analyses show the weight effect is strongest at daily doses above 1,500 mg, sustained for at least six months. Starting doses of 500 mg twice daily are appropriate for tolerance but will not produce the trial-level result. The typical titration path is:

• Week 1 to 2: 500 mg once daily with the largest meal
• Week 3 to 4: 500 mg twice daily
• Week 5 to 6: 1,000 mg in the morning and 500 mg in the evening
• Week 7 onward: 1,000 mg twice daily (the standard therapeutic dose)

Extended-release metformin is dosed once daily, usually 1,500 to 2,000 mg with the evening meal, and is much better tolerated. There is no meaningful difference between IR and ER in weight-loss magnitude.

Cost: the actual reason metformin still matters in 2026

Walmart, Costco, and most major US pharmacies sell generic metformin 500 mg or 1,000 mg tablets on $4-per-month / $10-per-90-day cash programs. With a GoodRx coupon the same prescription often runs $3 to $8 for ninety days. Mark Cuban's Cost Plus Drugs sells metformin 500 mg at $3.60 for 30 tablets and 1,000 mg at $4.50 for 60 tablets.

Compare to GLP-1s. Wegovy and Zepbound list at roughly $1,000 to $1,350 per month. Insurance coverage for obesity indication is patchy, and Medicare still does not cover GLP-1s for weight loss as of 2026. Compounded tirzepatide cleared the FDA shortage list in 2025, removing the legal pathway for most compounders. Many cash-pay GLP-1 patients are now paying $400 to $700 a month for the brand pens at best.

For a patient with prediabetes who cannot get GLP-1 coverage and cannot pay several hundred dollars a month, metformin is not a backup. It is the actual evidence-based answer.

Side effects, real ones

The label side effect list for metformin is dominated by gastrointestinal symptoms: diarrhea, nausea, flatulence, abdominal discomfort ^[5]. Roughly 20 to 30% of patients have some GI symptoms at initiation. About 5% cannot tolerate the drug at all. Three things mitigate this:

1. Start low and titrate. Going straight to 1,000 mg twice daily on day one causes most of the dropouts.
2. Take with the largest meal of the day. Empty-stomach dosing is the most common avoidable cause of nausea.
3. Use extended-release. ER metformin has roughly half the GI side effect rate of IR, with no efficacy penalty.

The longer-term issue is vitamin B12 deficiency. Metformin interferes with B12 absorption in the terminal ileum, and roughly 7% of long-term users develop low B12 levels ^[3]. Most are subclinical; a smaller fraction develop megaloblastic anemia or peripheral neuropathy. Standard practice: check serum B12 annually after two years on therapy, supplement with 1,000 mcg/day oral cyanocobalamin if levels drop below 400 pg/mL.

Lactic acidosis is the legacy fear from phenformin, metformin's withdrawn cousin. The actual risk on metformin is about 3 cases per 100,000 patient-years, almost all in patients with severe renal impairment, acute illness, or contrast exposure. The FDA contraindication is eGFR below 30 mL/min/1.73 m². Above 45, no dose adjustment. Between 30 and 45, halve the dose and monitor ^[3].

How metformin stacks with GLP-1 drugs

There is no contraindication to combining metformin with semaglutide, tirzepatide, liraglutide, or any other GLP-1 receptor agonist. The combination is routine in type 2 diabetes care and is increasingly used for weight management in patients with prediabetes or insulin resistance.

The pharmacologic logic is clean. Metformin works mostly hepatic and gut. GLP-1 agonists work mostly brain (appetite) and pancreas. Mechanistic overlap is small. Trials of tirzepatide and semaglutide enrolled large fractions of patients already on metformin, and weight-loss results were not blunted by it. Some endocrinologists keep patients on metformin during GLP-1 titration because it stabilizes glucose during the slow GLP-1 ramp, then continue both indefinitely. Others taper metformin off if HbA1c falls below 5.7% on the GLP-1 alone.

Hypoglycemia is not a meaningful risk on metformin + GLP-1 unless insulin or a sulfonylurea is also onboard. Both drugs are glucose-dependent in their insulinotropic effect.

Who metformin is right for

The drug pulls its weight in five populations:

Prediabetes (HbA1c 5.7 to 6.4%). This is metformin's strongest evidence base. The DPP showed a 31% reduction in progression to type 2 diabetes versus placebo over 2.8 years ^[1]. The benefit is durable at 15 years. ADA guidelines explicitly endorse metformin for prediabetes prevention in adults under 60 with BMI ≥ 35 or a history of gestational diabetes ^[4].

Polycystic ovary syndrome. Metformin improves ovulation, restores menstrual regularity, and produces BMI decreases of 3.4 to 4.55 units in randomized trials. For PCOS patients trying to conceive, metformin is often preferred over GLP-1s (which are contraindicated in pregnancy and require a two-month washout). PCOS is the single most common indication for off-label metformin in the US.

Mild obesity with insulin resistance. BMI in the 30 to 35 range with markers of insulin resistance (elevated fasting insulin, HOMA-IR above 3, acanthosis nigricans, fatty liver) responds reasonably to metformin. The 5 to 10 lb result is achievable and the cost is irrelevant.

Antipsychotic-induced weight gain. Patients on olanzapine, clozapine, risperidone, or quetiapine often gain 7 to 15% of body weight. Metformin attenuates and partially reverses this gain, averaging 3.27 kg of weight reduction in meta-analysis ^[3]. Some psychiatrists co-prescribe it prophylactically when starting clozapine or olanzapine.

Cost-sensitive patients. Anyone without GLP-1 insurance coverage who cannot pay $400 to $1,000 a month. Metformin is the best available answer for this group. The expected magnitude is modest. The cost makes it free.

When to consider a GLP-1 instead

The honest answer: most patients with BMI ≥ 30 and adequate insurance or budget access will lose substantially more on a GLP-1. The ADA's 2025 obesity standards of care list GLP-1 and dual-agonist drugs (semaglutide 2.4 mg, tirzepatide) as first-line pharmacotherapy for obesity ^[4]. Metformin is not on the obesity pharmacotherapy table. It appears only as a recommended adjunct for patients with concurrent type 2 diabetes or prediabetes.

Go to a GLP-1 when:

• BMI is 30 or higher with weight-related comorbidities, or BMI 35+, and access exists
• Goal weight loss exceeds 10% of body weight
• A six-month trial of lifestyle plus metformin produced less than 3% weight reduction
• Metabolic syndrome is severe (HbA1c climbing, fatty liver progressing, blood pressure rising)

Stay on metformin when:

• Prediabetes is the primary problem and lifestyle plus metformin is holding
• PCOS is the indication and pregnancy is planned within 12 months
• Cost is the binding constraint and there is no path to GLP-1 affordability
• The patient already lost what they wanted to lose and is in maintenance

The two are not mutually exclusive. Combining them is normal, well tolerated, and supported by clinical practice.

Where metformin sits in the broader anti-obesity landscape

A quick map of what else is in the room.

A few notes from the secondary keyword list worth touching on.

Phentermine-topiramate (Qsymia) produces 8 to 10 kg average weight loss but is a controlled substance with teratogenic risk; not commonly first-line. Lorcaserin was withdrawn from the US market in 2020 after the CAMELLIA trial showed an increased cancer signal; it is no longer available. Bupropion-naltrexone (Contrave) offers about 4 to 5 kg and is mostly used in patients with concurrent depression or alcohol use disorder. Pure amphetamines for obesity are no longer recommended; the era ended with the 1960s. Melanocortin-4 receptor agonists like setmelanotide treat rare monogenic obesity syndromes, not common polygenic obesity. The Pfizer oral GLP-1 candidate danuglipron was discontinued in 2025 after a hepatotoxicity signal; orforglipron from Eli Lilly remains in late-stage development as the leading oral GLP-1.

Pipeline drugs targeting maridebart cafraglutide (a GLP-1/amylin combo) and the asprosin-blocking antibody program are still in trials. Neither has a 2026 availability timeline. Aldosterone and leptin-based therapies for obesity have not panned out beyond their original research programs; the history of leptin obesity treatment is mostly a story of how monogenic leptin deficiency responds dramatically and common obesity does not. Translational models for obesity treatment continue to use diet-induced obese mice and the Zucker rat as the workhorse preclinical systems.

Frequently asked questions

Is metformin FDA-approved for weight loss?

No. Metformin is FDA-approved only for type 2 diabetes. Use for weight loss, prediabetes prevention, and PCOS is off-label but supported by guidelines and trial data.

How much weight can I expect to lose on metformin?

Trial average is 2 to 3 kg (4 to 6 lb). Roughly 30% of patients lose more than 5% of body weight in the first year. The other 70% see modest or no loss.

How long does it take to see weight loss on metformin?

Weight loss usually begins around week 4 and is mostly complete by month 6 to 12. If you have lost nothing at six months on 1,500 to 2,000 mg/day, you are unlikely to respond.

What is the best dose for weight loss?

1,500 to 2,000 mg/day is the dose used in the DPP and the dose with the strongest weight evidence. Below 1,500 mg/day the effect is much weaker.

Can I take metformin with Ozempic, Wegovy, or Zepbound?

Yes. The combination is routine, has no contraindication, and is commonly used for patients with both type 2 diabetes and obesity. The GLP-1 does not require a metformin dose change.

Does metformin cause hypoglycemia?

Not on its own. Metformin alone almost never causes hypoglycemia. The risk rises only when combined with insulin or a sulfonylurea.

How much does metformin cost?

About $4 a month at Walmart on the generic cash program, or under $5 a month with most GoodRx coupons. Cost Plus Drugs sells it for similar prices. Insurance copays are usually $0 to $10.

Will metformin cause vitamin B12 deficiency?

Long-term use lowers B12 in about 7% of patients. Annual B12 monitoring is standard after two years on therapy. Oral supplementation reverses it.

Is metformin safe for PCOS?

Yes. Metformin is one of the most common PCOS treatments, improving ovulation, menstrual regularity, and fertility, with BMI reductions of 3 to 4 units in trials.

Should I take metformin if I have prediabetes?

ADA guidelines recommend it for adults under 60 with prediabetes who have BMI ≥ 35 or a history of gestational diabetes. It reduces progression to type 2 diabetes by 31% over three years.

What about brain fog from obesity? Does metformin help?

Some patients report cognitive benefits from improved insulin sensitivity, but the data are mixed. Metformin is not prescribed for cognitive symptoms specifically.

Is metformin used for obesity in animals or horses?

Veterinary use exists for insulin resistance in horses and obese dogs, but dosing, safety, and efficacy are species-specific. This article covers human use only.

Bottom line

Metformin is not a weight-loss drug in the GLP-1 sense, and pretending it is sets patients up for disappointment. It is a cheap, safe, well-studied metabolic drug that produces modest weight loss in the right patient, prevents diabetes in the right patient, and stacks cleanly with newer therapies. For prediabetes, PCOS, antipsychotic weight gain, and cost-constrained obesity, it is still a first move. For 50-pound weight loss goals, it is not.