Metformin and Non-GLP-1 Diabetes Drugs for Weight Loss: What the Trials Actually Show

Metformin is the single most-searched non-GLP-1 weight-loss drug on the internet, with 10,000 monthly searches on does metformin cause weight loss alone. The published evidence is real but modest. The Diabetes Prevention Program (Knowler NEJM 2002, n=3,234)^[1] reported −2.1 kg with metformin vs −0.1 kg with placebo over 2.8 years. The 15-year DPP Outcomes Study (Apolzan Ann Intern Med 2019)^[3] reported a −6.2% sustained weight loss in metformin responders — durable but small. The Seifarth 2013 trial^[5] in non-diabetic obesity reported −5.6% over 6 months. Compared to semaglutide −14.9% (STEP-1)^[10] and tirzepatide −20.9% (SURMOUNT-1)^[11], metformin produces roughly 1/5 to 1/7 the magnitude. It is cheap, generic, has 60+ years of safety data, and has the best long-term safety profile of any obesity drug — but it is not in the same league as a GLP-1 for absolute weight loss. Here is the verified evidence map.

The DPP — the foundational metformin weight-loss data

The Diabetes Prevention Program^[1] remains the single most important metformin weight study. Knowler and colleagues randomized 3,234 adults with prediabetes (impaired glucose tolerance and elevated fasting glucose) to one of three arms over a mean 2.8 years:

• Lifestyle intervention (low-fat diet, 150 minutes/week of exercise, behavioral modification): −5.6 kg, 58% reduction in incident T2D (95% CI 48-66%)
• Metformin 850 mg twice daily: −2.1 kg, 31% reduction in incident T2D (95% CI 17-43%)
• Placebo: −0.1 kg, reference

Two things to take away. First, the lifestyle arm produced more than twice the weight loss of metformin. Second, metformin's weight effect was real and statistically significant but modest in absolute terms — about 2 kg over nearly 3 years. The diabetes-prevention benefit was also real, but it was smaller than the lifestyle benefit.

DPPOS: 10- and 15-year follow-up

The DPP cohort was followed for an additional decade in the DPP Outcomes Study (DPPOS). The 10-year analysis^[2] reported that metformin sustained approximately −2.5 kg of weight loss out to year 10, while the lifestyle arm regressed back to about −2 kg as behavior change attenuated.

The Apolzan 2019 15-year analysis^[3] in Annals of Internal Medicine looked specifically at the patients who responded to metformin in year 1 with ≥5% weight loss. In those responders, sustained weight loss across years 6-15 was:

• Metformin responders: −6.2% (95% CI 5.2-7.2%)
• Lifestyle responders: −3.7% (95% CI 3.1-4.4%)
• Placebo responders: −2.8% (95% CI 1.3-4.4%)

For the patients who did respond to metformin, the long-term durability was actually greater than for the lifestyle arm. The catch is that fewer patients respond to metformin than respond to lifestyle, and the absolute magnitude is still small. But this is the cleanest published evidence that metformin's weight effect is durable in responders — not a transient effect that washes out.

ADOPT: metformin vs other oral diabetes drugs

The ADOPT trial^[4] (Kahn NEJM 2006, n=4,360) was a head-to-head comparison of three oral T2D monotherapies in newly-diagnosed patients followed for a mean of 4 years. Mean body weight changes at 48 months:

• Metformin: −2.8 kg
• Glyburide (sulfonylurea): +1.6 kg
• Rosiglitazone (TZD): +4.8 kg

ADOPT remains the reference for “does metformin cause weight loss in T2D” — the answer is yes, modestly, and it is one of the only oral diabetes drugs that does. Sulfonylureas (glyburide, glipizide, glimepiride) and thiazolidinediones (rosiglitazone, pioglitazone) cause weight gain. SGLT2 inhibitors and GLP-1 agonists cause weight loss; we cover those classes in detail in our SGLT2 vs GLP-1 article.

Metformin in non-diabetic obesity

For patients without diabetes, the most cited prospective cohort is Seifarth and colleagues^[5] inExperimental and Clinical Endocrinology & Diabetes2013. They followed 154 obese, non-diabetic adults treated with metformin (compared to 45 controls) over 6 months and reported a mean weight loss of −5.8 ± 7.0 kg (about 5.6% body weight reduction). The effect was larger in patients with severe insulin resistance, consistent with metformin's mechanism — the more insulin-resistant the patient, the more metformin moves the needle.

That magnitude is bigger than the DPP weight effect at the same time horizon, and there are a few reasons. The DPP was a placebo-controlled prevention trial in patients with only prediabetes (relatively low baseline metabolic dysfunction). Seifarth was a clinic-based observational cohort in obese patients with greater baseline insulin resistance. Selection effects probably amplify the Seifarth signal somewhat. The honest range is 2-5 kg over 6-12 months for metformin in non-diabetic obesity, with more effect in patients with worse baseline insulin resistance.

The magnitude gap to GLP-1s

Side by side:

• Metformin: −2 to −3 kg over 1-2 years (DPP, ADOPT); −5.8 kg over 6 months in non-diabetic obese (Seifarth); −6.2% durable in long-term responders (DPPOS 15-year).
• Semaglutide 2.4 mg weekly (STEP-1): −14.9% body weight at 68 weeks — approximately −15 kg from a baseline of 100 kg.
• Tirzepatide 15 mg weekly (SURMOUNT-1): −20.9% at 72 weeks — approximately −22 kg.

Metformin produces approximately 1/5 the magnitude of semaglutide and 1/7 of tirzepatide. The comparison is not flattering for metformin in absolute weight terms. But metformin has three things going for it that the GLP-1s do not:

• 60 years of safety data — the longest safety record of any obesity drug
• ~$4/month generic price through US discount pharmacies, vs. $149-$1,400/month for GLP-1s
• Cardiovascular and possibly cancer benefits documented in long-term diabetes cohorts (UK Prospective Diabetes Study, observational cohorts)

For patients who don't want or can't get a GLP-1, and who have insulin resistance or prediabetes, metformin is a defensible option. For patients whose primary goal is weight loss without an underlying metabolic indication, metformin alone is not going to deliver the magnitude they are looking for.

Metformin + GLP-1: combination therapy

The combination of metformin + GLP-1 is the standard of care in T2D and has been studied extensively in the GLP-1 registration trials. STEP-1 was conducted in non-diabetic adults; SURMOUNT-1 was non-diabetic; the diabetes equivalents (STEP-2, SURPASS-2) used metformin background therapy in many patients. The combination is well tolerated with no PK interaction. Adding metformin to a GLP-1 in T2D adds a small additional weight benefit (typically <1 kg) on top of the GLP-1 effect, and adds the cardiovascular and glycemic benefits metformin is known for.

For non-diabetic patients on a GLP-1, metformin is generally not added unless there is an insulin-resistance indication (PCOS, prediabetes). It can be a reasonable bridging strategy before starting a GLP-1, and a reasonable maintenance strategy after stopping a GLP-1, but it is not a substitute.

Trulicity and the GLP-1 dose ladder

Patients searching for “does Trulicity cause weight loss” (1,300/mo) sit at the boundary of this article, because dulaglutide is a GLP-1 receptor agonist — the same class as semaglutide and tirzepatide. The honest answer is yes, but less than semaglutide and dramatically less than tirzepatide.

AWARD-11^[6] tested higher doses of dulaglutide (3.0 mg and 4.5 mg) against the standard 1.5 mg dose in T2D patients on metformin background therapy. The 36-week weight loss results:

• Dulaglutide 1.5 mg: −3.1 kg
• Dulaglutide 3.0 mg: −4.0 kg
• Dulaglutide 4.5 mg: −4.7 kg

These are T2D-population numbers (smaller weight effects than non-diabetic populations) at the highest dulaglutide doses ever tested. Compared with semaglutide 2.4 mg (−15.3 kg) and tirzepatide 15 mg (−22 kg) in non-diabetic obesity, dulaglutide is the lowest-magnitude GLP-1 in current clinical use. It earned its place when it was the only weekly GLP-1 available; semaglutide and tirzepatide have eclipsed it for weight management.

Other non-GLP-1 obesity drugs (briefly)

For completeness, the rest of the non-GLP-1 weight-loss drug landscape:

• Bupropion (Wellbutrin) — covered in our antidepressants article. Anderson 2002^[7] reported −7.2% to −10.1% weight loss with bupropion SR 300-400 mg/day at 24 weeks in obese non-diabetic adults.
• Naltrexone + bupropion (Contrave) — Greenway COR-I 2010^[8] reported −6.1% with NB 32/360 vs −1.3% placebo at 56 weeks in n=1,742. FDA-approved for chronic weight management.
• Topiramate / Qsymia — covered in our dedicated Topamax / Qsymia article. CONQUER showed −9.8% with the 15/92 dose at 56 weeks.
• Vyvanse (lisdexamfetamine) — FDA- approved for binge-eating disorder, NOT weight loss. See our stimulants article.
• HRT — the WHI primary trial^[9] did not report large weight effects. See our HRT and perimenopause article.
• SGLT2 inhibitors (Jardiance, Farxiga) — ~2-3 kg weight loss via urinary glucose excretion; see our SGLT2 vs GLP-1 article.
• Spironolactone, levothyroxine, antidepressants other than bupropion — not weight-loss drugs; some have weight-related side effects but none should be prescribed primarily for weight management.

For a side-by-side lookup of every drug and its expected weight effect, see our Non-GLP-1 Drug Weight Effect Lookup tool (when shipped).

The honest take for patients

• If you have prediabetes or PCOS and want a low-cost, safe, evidence-based first step before considering a GLP-1: metformin is defensible. Expect ~2-3 kg over the first year and possibly more if you respond.
• If your A1c is in the diabetes range and you want both glycemic control and weight loss: a GLP-1 (semaglutide or tirzepatide) is preferred per ADA 2025 Standards of Care. Metformin remains a useful adjunct.
• If your primary goal is significant weight loss (≥10%) and you have no specific metformin-indication, metformin is unlikely to deliver alone. A GLP-1 will be 5-7x more effective in absolute terms.
• If you cannot afford or cannot tolerate a GLP-1 and have insulin resistance or prediabetes, metformin is a reasonable starting point. Combine it with structured lifestyle change (which DPP showed is the highest-leverage intervention) and resistance training.
• If you are already on a GLP-1 + metformin combination (common in T2D), continue both. They have no PK interaction and the combination is well studied.

Bottom line

• DPP (Knowler 2002): metformin −2.1 kg vs placebo −0.1 kg over 2.8 years in 3,234 adults with prediabetes.
• DPPOS 15-year (Apolzan 2019): −6.2% sustained weight loss in metformin responders — durable but small.
• ADOPT (Kahn 2006): −2.8 kg with metformin in newly-diagnosed T2D, vs +1.6 kg glyburide and +4.8 kg rosiglitazone at 4 years.
• Seifarth 2013: −5.8 kg over 6 months in non-diabetic obese adults, larger in those with severe insulin resistance.
• Magnitude gap: metformin produces ~1/5 the weight loss of semaglutide and ~1/7 of tirzepatide.
• But metformin has 60+ years of safety data, ~$4/month generic pricing, and the broadest evidence base of any obesity drug for cardiovascular and possibly cancer benefits.
• For patients who can't access or tolerate GLP-1s, metformin is a defensible first step. For everyone else, it is an adjunct, not a substitute.

Related research and tools

• SGLT2 inhibitors vs GLP-1s — the other major non-GLP-1 diabetes drug class
• Antidepressants and weight on a GLP-1 — bupropion / Wellbutrin in detail
• Berberine vs GLP-1s — the natural-product comparator
• GLP-1 pricing index — the cost half of the metformin-vs-GLP-1 decision
• GLP-1 insurance coverage audit — access patterns
• Foundayo vs Wegovy vs Zepbound — modern GLP-1 head-to-heads

Important disclaimer. This article is educational and does not constitute medical advice. Patients with type 2 diabetes, prediabetes, PCOS, or chronic kidney disease should discuss metformin (initiation, continuation, or combination with a GLP-1) with their prescribing clinician. Metformin is contraindicated in severe renal impairment and during contrast imaging. Every primary source cited here was independently verified against PubMed on 2026-04-08 by a research subagent.