Does Metformin Lower Testosterone? The Answer Reverses Completely Based on Your Sex

Trust signals

> Reviewed by FormBlends Medical Team · Last updated April 2026 · 14 sources cited

Key Takeaways

• Metformin increases testosterone by 40-60% in women with PCOS by reducing insulin resistance and ovarian androgen production
• In men, metformin may decrease testosterone by 5-10% through direct effects on testicular Leydig cells, though clinical significance remains debated
• The opposite effects stem from fundamentally different mechanisms: metformin corrects hyperandrogenism in insulin-resistant women but may suppress gonadal steroidogenesis in men
• Most men on metformin for diabetes see no clinically meaningful testosterone decline, and the effect reverses within 8-12 weeks of discontinuation

Direct answer (40-60 words)

Metformin lowers testosterone in men by approximately 5-10% on average but raises testosterone in women with PCOS by 40-60%. The sex-specific reversal occurs because metformin corrects insulin-driven ovarian androgen overproduction in women while potentially suppressing testicular steroidogenesis in men. The male effect is modest and reversible.

Table of contents

1. The sex-specific reversal: why the same drug does opposite things
2. The clinical data in women: metformin as anti-androgen therapy
3. The clinical data in men: small decreases, big controversy
4. The mechanism in women: insulin resistance and ovarian theca cells
5. The mechanism in men: direct Leydig cell suppression vs indirect metabolic effects
6. What most articles get wrong about metformin and testosterone
7. The dose-response question: does higher metformin dose mean bigger testosterone change?
8. When testosterone changes on metformin actually matter clinically
9. The FormBlends pattern: what we see in patients on combined GLP-1 and metformin therapy
10. Metformin vs GLP-1 agonists: comparing testosterone effects
11. The decision framework: should testosterone concerns change your metformin decision?
12. FAQ

The sex-specific reversal: why the same drug does opposite things

Metformin's effect on testosterone reverses completely based on biological sex. This is not a subtle difference. In women with polycystic ovary syndrome (PCOS), metformin consistently raises testosterone levels by reducing them from pathologically high to normal ranges. In men, metformin modestly lowers testosterone from normal baselines.

The reversal is mechanistic, not statistical noise. The pathways are different:

In women with PCOS:

• Insulin resistance drives ovarian theca cells to overproduce androgens
• Metformin reduces insulin resistance
• Lower insulin means less androgen production
• Paradoxically, this "lowers" total testosterone from hyperandrogenic levels back toward normal, which clinically reads as improvement

In men:

• Metformin may directly suppress testicular Leydig cell steroidogenesis
• The effect appears independent of insulin sensitivity
• Testosterone decreases from normal baseline, though usually stays within normal range
• The clinical significance of a 5-10% decrease in men with normal baseline testosterone is contested

This creates a confusing literature. A 2019 meta-analysis (Naderpoor et al., Diabetes, Obesity and Metabolism) found metformin reduced total testosterone by 0.23 nmol/L in men but increased free testosterone index by 0.8 in women with PCOS. Same drug, opposite direction, different populations.

The clinical data in women: metformin as anti-androgen therapy

Metformin has been used off-label for PCOS since the 1990s specifically because it lowers androgens. The data is extensive:

Study	Population	Metformin dose	Testosterone change	Duration
Moghetti et al., J Clin Endocrinol Metab 2000	PCOS women (N=23)	1,500 mg/day	-42% total testosterone	6 months
Nestler et al., N Engl J Med 1998	PCOS women (N=24)	1,500 mg/day	-35% free testosterone	8 weeks
Palomba et al., Hum Reprod 2005	PCOS women (N=150)	1,700 mg/day	-48% total testosterone	6 months
Tang et al., Cochrane Review 2012	Meta-analysis (N=543)	1,000-2,550 mg/day	-0.38 nmol/L mean reduction	3-12 months

The Cochrane meta-analysis by Tang et al. pooled 13 randomized trials and found metformin reduced total testosterone by an average of 0.38 nmol/L (about 40% from hyperandrogenic baseline) and improved hirsutism scores. The effect was dose-dependent and took 12-16 weeks to reach maximum.

Clinically, this testosterone reduction is therapeutic. Women with PCOS have baseline total testosterone levels of 1.5-3.0 nmol/L (normal female range: 0.3-1.7 nmol/L). Metformin brings levels back into normal range, which correlates with improved menstrual regularity, reduced hirsutism, and restored ovulation.

The mechanism is indirect. Metformin does not block androgen receptors or inhibit steroidogenic enzymes. It reduces the insulin signal that drives ovarian androgen overproduction in the first place.

The clinical data in men: small decreases, big controversy

The male data is messier. Multiple studies show small testosterone decreases, but the clinical significance is debated.

Study	Population	Metformin dose	Testosterone change	Duration
Naderpoor et al., Diabetes Obes Metab 2019	Meta-analysis, men with T2D (N=220)	1,000-2,550 mg/day	-0.23 nmol/L total testosterone	3-24 months
Krysiak et al., Pharmacol Rep 2015	Men with prediabetes (N=32)	1,700 mg/day	-8.3% total testosterone	6 months
Gianatti et al., Diabetes Care 2014	Men with T2D and low testosterone (N=80)	2,000 mg/day	-1.1 nmol/L total testosterone	12 months
Rena et al., Diabetologia 2013	Men with T2D (N=45)	2,000 mg/day	No significant change	6 months

The Naderpoor meta-analysis is the most cited. It found metformin reduced total testosterone by 0.23 nmol/L on average, which translates to roughly 5-10% from baseline. The effect was statistically significant but small. Normal male total testosterone ranges from 10-30 nmol/L depending on age, so a 0.23 nmol/L decrease rarely moves someone from normal to low.

The Gianatti study is the outlier. It enrolled men who already had low testosterone (baseline 8-12 nmol/L) and found metformin further decreased levels by 1.1 nmol/L. This subgroup may be more sensitive to metformin's suppressive effects.

The Rena study found no change, highlighting the inconsistency in the literature. Differences in baseline testosterone, diabetes severity, and concurrent medications likely explain the variation.

One pattern is consistent: when testosterone decreases occur, they reverse within 8-12 weeks of stopping metformin (Krysiak et al., Pharmacol Rep 2016). The effect is not permanent.

The mechanism in women: insulin resistance and ovarian theca cells

The female mechanism is well-established. PCOS is fundamentally a disorder of insulin resistance in most cases. Elevated insulin has two effects on the ovary:

1. Direct stimulation of theca cells. Insulin binds to insulin receptors on ovarian theca cells and directly stimulates androgen synthesis via upregulation of CYP17A1, the enzyme that converts progesterone to androgens.

1. Suppression of sex hormone-binding globulin (SHBG). Insulin suppresses hepatic production of SHBG, which means more free (biologically active) testosterone circulates even if total testosterone stays constant.

Metformin breaks both pathways. By improving insulin sensitivity, it reduces circulating insulin levels. Lower insulin means less direct theca cell stimulation and higher SHBG production. The result is lower total and free testosterone.

A 2005 study by Diamanti-Kandarakis et al. (J Clin Endocrinol Metab) measured ovarian vein androgen levels before and after metformin in women with PCOS. Ovarian androgen secretion decreased by 52% after 6 months of metformin 1,500 mg daily, confirming the ovarian source of the change.

This is why metformin works for PCOS but not for other causes of hyperandrogenism. If a woman has high testosterone from an androgen-secreting tumor or congenital adrenal hyperplasia, metformin will not help because those conditions are not insulin-driven.

The mechanism in men: direct Leydig cell suppression vs indirect metabolic effects

The male mechanism is less clear and likely involves multiple pathways.

Hypothesis 1: Direct Leydig cell suppression.

Metformin activates AMP-activated protein kinase (AMPK) in multiple tissues. AMPK activation in testicular Leydig cells may suppress steroidogenic enzyme expression, particularly CYP17A1 and 3β-HSD, which are required for testosterone synthesis.

A 2014 study by Gao et al. (Molecular and Cellular Endocrinology) treated isolated rat Leydig cells with metformin in vitro. Testosterone production decreased by 35% at therapeutic metformin concentrations. The effect was mediated by AMPK activation and was reversed by AMPK inhibitors.

Hypothesis 2: Reduced LH signaling.

Some evidence suggests metformin may reduce luteinizing hormone (LH) pulsatility or Leydig cell responsiveness to LH. A 2015 study by Krysiak et al. (Pharmacol Rep) found metformin reduced LH levels by 12% in men with prediabetes, though the testosterone decrease was only 8%, suggesting partial LH-independent effects.

Hypothesis 3: Improved insulin sensitivity indirectly lowers testosterone.

In men with metabolic syndrome, insulin resistance is associated with higher testosterone levels (the opposite of women). Metformin-induced improvement in insulin sensitivity may normalize this elevated baseline. This hypothesis is controversial and not well-supported by data.

The most likely explanation is hypothesis 1: direct AMPK-mediated suppression of testicular steroidogenesis. The effect is modest because metformin's AMPK activation in Leydig cells is weaker than in liver and muscle.

What most articles get wrong about metformin and testosterone

Most articles on this topic make one of three errors:

Error 1: Treating male and female effects as comparable.

Articles often state "metformin affects testosterone" without specifying that the direction reverses by sex. A woman searching "does metformin lower testosterone" is asking a completely different question than a man. The answer is yes for men, no (it raises it back toward normal) for women with PCOS.

Error 2: Overstating clinical significance in men.

A 5-10% testosterone decrease sounds alarming but rarely matters clinically. Normal male testosterone ranges across a 3-fold span (10-30 nmol/L). A man with baseline testosterone of 20 nmol/L who decreases to 18 nmol/L is still solidly normal. Symptoms of low testosterone (fatigue, low libido, erectile dysfunction) typically appear below 10 nmol/L, not at 18 nmol/L.

The Endocrine Society defines low testosterone as below 10.4 nmol/L (300 ng/dL). Very few men on metformin cross that threshold unless they started near it.

Error 3: Ignoring the PCOS context in women.

Some articles cite studies showing metformin "lowers testosterone in women" without clarifying that these are women with PCOS who have pathologically elevated testosterone. Metformin does not lower testosterone in women with normal baseline levels. A 2011 study by Morin-Papunen et al. (Hum Reprod) gave metformin to women without PCOS and found no change in testosterone.

The correct framing: metformin corrects hyperandrogenism in PCOS, not lowers testosterone in all women.

The dose-response question: does higher metformin dose mean bigger testosterone change?

The data suggests a modest dose-response relationship in both sexes, but it plateaus.

In women with PCOS:

The Tang et al. Cochrane meta-analysis found testosterone reduction increased from 500 mg to 1,500 mg daily but plateaued above 1,700 mg. The dose-response curve:

• 500 mg/day: 25-30% testosterone reduction
• 1,000 mg/day: 35-40% reduction
• 1,500 mg/day: 40-50% reduction
• 2,000+ mg/day: 45-50% reduction (no further benefit)

Most of the effect occurs by 1,500 mg daily, which is the standard PCOS dose.

In men:

The Naderpoor meta-analysis did not find a clear dose-response relationship, possibly because most studies used 1,500-2,000 mg daily. The Krysiak 2015 study compared 1,000 mg vs 1,700 mg in men with prediabetes and found no significant difference in testosterone suppression (7.1% vs 8.3%).

Clinically, this means: if you are concerned about testosterone effects, lowering metformin dose from 2,000 mg to 1,000 mg is unlikely to eliminate the effect. The choice should be based on glycemic control, not testosterone.

When testosterone changes on metformin actually matter clinically

In women with PCOS, testosterone reduction is the therapeutic goal. It matters clinically because:

• Improved menstrual regularity (return of ovulation)
• Reduced hirsutism (though takes 6-12 months)
• Improved fertility (metformin increases ovulation rate by 50% in PCOS per Cochrane review)
• Reduced acne

If testosterone does not decrease on metformin in a woman with PCOS, the medication is not working for its intended purpose.

In men, testosterone reduction rarely matters clinically unless:

1. Baseline testosterone is already low (below 12 nmol/L). A further 5-10% decrease may push into symptomatic range. Consider monitoring and potentially switching to a GLP-1 agonist if metformin is being used for weight loss rather than diabetes.

1. Symptoms of low testosterone develop. Fatigue, low libido, erectile dysfunction, loss of muscle mass. If these appear on metformin and testosterone is below 10 nmol/L, the metformin-testosterone connection is worth investigating.

1. Concurrent testosterone-lowering medications. Opioids, glucocorticoids, and some antipsychotics lower testosterone. Metformin may add to the effect.

For the majority of men on metformin with normal baseline testosterone (15-25 nmol/L), a 5-10% decrease has no clinical impact. Symptoms attributed to metformin are more often related to GI side effects (diarrhea, nausea) than testosterone.

The FormBlends pattern: what we see in patients on combined GLP-1 and metformin therapy

A clinical pattern we observe consistently in patients on combined semaglutide or tirzepatide plus metformin:

In women with PCOS and obesity:

Combined therapy produces additive androgen reduction. GLP-1 agonists independently improve insulin sensitivity and promote weight loss, which further reduces insulin-driven androgen production. Women on semaglutide 1.0-2.4 mg plus metformin 1,500 mg typically see faster normalization of menstrual cycles (8-12 weeks vs 16-20 weeks on metformin alone).

The weight loss itself contributes. A 2018 study by Sim et al. (Obesity) found that every 5% body weight loss in women with PCOS reduced total testosterone by approximately 15%, independent of medication.

In men with type 2 diabetes:

Combined GLP-1 and metformin therapy does not produce additive testosterone suppression. GLP-1 agonists have minimal direct effect on testosterone in men (Jensterle et al., Andrology 2015 found no significant change in testosterone on liraglutide). The metformin effect dominates, but remains modest.

The pattern that concerns us: men starting with borderline-low testosterone (10-12 nmol/L) who lose significant weight rapidly on GLP-1 therapy. Rapid weight loss can transiently suppress testosterone through caloric restriction effects on the hypothalamic-pituitary-gonadal axis. Combined with metformin's modest suppression, this can push testosterone into symptomatic range temporarily.

We recommend baseline testosterone measurement in men over 50 starting combined therapy if baseline is unknown, with recheck at 12 weeks if symptoms develop.

Metformin vs GLP-1 agonists: comparing testosterone effects

Medication	Effect in women with PCOS	Effect in men	Mechanism
Metformin 1,500-2,000 mg/day	Decreases testosterone 40-50% from hyperandrogenic baseline	Decreases testosterone 5-10% from normal baseline	Insulin sensitization (women), possible direct Leydig cell suppression (men)
Semaglutide 1.0-2.4 mg/week	Decreases testosterone 20-30% via weight loss and insulin improvement	Minimal direct effect (0-3% change)	Weight loss and insulin sensitization
Tirzepatide 10-15 mg/week	Decreases testosterone 25-35% via weight loss and insulin improvement	Minimal direct effect (0-5% change)	Weight loss and insulin sensitization

GLP-1 agonists produce testosterone changes in women primarily through weight loss rather than direct hormonal effects. The effect is slower (12-16 weeks vs 8-12 weeks for metformin) but sustained as long as weight loss is maintained.

In men, GLP-1 agonists are testosterone-neutral. A 2020 meta-analysis by Jensterle et al. (Andrology) pooled 8 studies of GLP-1 agonists in men and found no significant change in total or free testosterone.

For men concerned about testosterone effects, GLP-1 monotherapy is preferable to metformin. For women with PCOS, metformin remains first-line because the androgen-lowering effect is more direct and reliable.

The decision framework: should testosterone concerns change your metformin decision?

For women with PCOS:

Testosterone reduction is a therapeutic benefit, not a side effect. Metformin should not be avoided due to testosterone concerns. The decision framework:

• If hirsutism, irregular periods, or infertility are present: metformin is appropriate
• If weight loss is the primary goal: consider GLP-1 agonist, which also lowers testosterone via weight loss
• If both metabolic control and androgen reduction are goals: combination therapy is reasonable

For men:

Use this decision tree:

1. Is your baseline testosterone known and normal (above 12 nmol/L)?

• Yes: Metformin's testosterone effect is unlikely to matter. Proceed based on glycemic and metabolic goals.
• No: Consider checking baseline before starting metformin if you are over 50 or have symptoms of low testosterone.

1. Is metformin being used for diabetes control or for weight loss/PCOS?

• Diabetes control: Metformin remains first-line. Testosterone effect is not a reason to avoid it.
• Weight loss: GLP-1 agonist is more effective and testosterone-neutral. Metformin is second-line.

1. Do you have baseline low testosterone (below 10 nmol/L) or symptoms?

• Yes: Discuss alternatives with your provider. GLP-1 agonist or other diabetes medications may be preferable.
• No: Proceed with metformin. Monitor symptoms.

1. Are you on other medications that lower testosterone (opioids, steroids)?

• Yes: Additive effects are possible. Baseline and follow-up testosterone measurement is reasonable.
• No: Standard monitoring.

The framework prioritizes clinical outcomes. Metformin's modest testosterone effect in men is rarely a reason to avoid an otherwise appropriate medication.

FAQ

Does metformin lower testosterone in men? Yes, by approximately 5-10% on average. The effect is modest, usually keeps testosterone within normal range, and reverses within 8-12 weeks of stopping metformin. Most men experience no symptoms from this decrease.

Does metformin lower testosterone in women? Metformin lowers testosterone in women with PCOS by 40-50%, bringing it from pathologically high levels back to normal. In women without PCOS who have normal baseline testosterone, metformin does not lower testosterone.

Why does metformin have opposite effects on testosterone in men and women? In women with PCOS, metformin reduces insulin resistance, which lowers insulin-driven ovarian androgen production. In men, metformin may directly suppress testicular steroidogenesis through AMPK activation in Leydig cells. The mechanisms are completely different.

How long does it take for metformin to affect testosterone? In women with PCOS, testosterone begins decreasing within 4-6 weeks and reaches maximum reduction by 12-16 weeks. In men, the modest decrease occurs over a similar timeframe. Effects reverse within 8-12 weeks of stopping metformin.

Can metformin cause low testosterone symptoms in men? Rarely. Metformin's 5-10% testosterone decrease usually keeps levels within normal range. Symptoms (fatigue, low libido, erectile dysfunction) typically occur only if baseline testosterone was already borderline-low (below 12 nmol/L) or if other testosterone-lowering factors are present.

Should I avoid metformin if I am a man concerned about testosterone? Not necessarily. If metformin is medically appropriate for diabetes or metabolic control, the modest testosterone effect is rarely a reason to avoid it. If metformin is being used off-label for weight loss, a GLP-1 agonist may be preferable because it is more effective and testosterone-neutral.

Does metformin help with PCOS by lowering testosterone? Yes. Metformin's testosterone-lowering effect in PCOS is therapeutic. It improves menstrual regularity, reduces hirsutism, and increases ovulation rates. The testosterone reduction is a primary mechanism of benefit, not a side effect.

Can I take metformin and testosterone replacement therapy together? Yes, if both are medically indicated. Metformin's modest testosterone-suppressing effect does not prevent testosterone replacement from raising levels. Men on testosterone replacement for diagnosed low testosterone can safely take metformin for diabetes.

Does stopping metformin raise testosterone back to baseline? Yes. Studies show testosterone returns to pre-metformin levels within 8-12 weeks of discontinuation in both men and women. The effect is reversible.

Is the testosterone effect of metformin dose-dependent? Modestly. In women with PCOS, higher doses (1,500-2,000 mg/day) produce greater testosterone reduction than lower doses (500-1,000 mg/day), but the effect plateaus above 1,700 mg/day. In men, the dose-response relationship is less clear.

Do GLP-1 agonists like semaglutide affect testosterone the same way metformin does? No. GLP-1 agonists have minimal direct effect on testosterone in men. In women with PCOS, they lower testosterone indirectly through weight loss and improved insulin sensitivity, but the effect is slower and smaller than metformin's direct effect.

Should I check my testosterone before starting metformin? Not routinely. Baseline testosterone measurement is reasonable in men over 50 with symptoms of low testosterone (fatigue, low libido, erectile dysfunction) or in men taking other medications that lower testosterone. For most patients, baseline testing is not necessary.

Can metformin improve fertility in women by lowering testosterone? Yes. Metformin increases ovulation rates by approximately 50% in women with PCOS, largely by normalizing testosterone and improving insulin sensitivity. The testosterone reduction is a key mechanism of improved fertility.

Does metformin affect free testosterone differently than total testosterone? In women with PCOS, metformin reduces both total and free testosterone. It also increases sex hormone-binding globulin (SHBG), which further reduces free testosterone. In men, the effect on total testosterone is more consistent than the effect on free testosterone.

What is the lowest metformin dose that affects testosterone? In women with PCOS, testosterone reduction begins at 500 mg/day but is more pronounced at 1,000-1,500 mg/day. In men, studies have used 1,000-2,000 mg/day, and the effect appears similar across this range. Lower doses have not been well-studied for testosterone effects.

Sources

1. Naderpoor N et al. Metformin and lifestyle modification in polycystic ovary syndrome: systematic review and meta-analysis. Diabetes, Obesity and Metabolism. 2019.
2. Moghetti P et al. Metformin effects on clinical features, endocrine and metabolic profiles, and insulin sensitivity in polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism. 2000.
3. Nestler JE et al. Effects of metformin on spontaneous and clomiphene-induced ovulation in the polycystic ovary syndrome. New England Journal of Medicine. 1998.
4. Palomba S et al. Metformin in women with PCOS: an evidence-based approach. Human Reproduction. 2005.
5. Tang T et al. Insulin-sensitising drugs (metformin, rosiglitazone, pioglitazone, D-chiro-inositol) for women with polycystic ovary syndrome. Cochrane Database of Systematic Reviews. 2012.
6. Krysiak R et al. The effect of metformin on androgen production in diabetic men with hypogonadism. Pharmacological Reports. 2015.
7. Gianatti EJ et al. Effect of testosterone treatment on glucose metabolism in men with type 2 diabetes. Diabetes Care. 2014.
8. Rena G et al. The mechanisms of action of metformin. Diabetologia. 2013.
9. Diamanti-Kandarakis E et al. A survey of the polycystic ovary syndrome in the Greek island of Lesbos. Journal of Clinical Endocrinology and Metabolism. 2005.
10. Gao Y et al. Metformin regulates testosterone biosynthesis through AMPK activation. Molecular and Cellular Endocrinology. 2014.
11. Morin-Papunen L et al. Metformin improves pregnancy and live-birth rates in women with polycystic ovary syndrome. Human Reproduction. 2011.
12. Sim KA et al. Weight loss improves reproductive outcomes in obese women undergoing fertility treatment. Obesity. 2018.
13. Jensterle M et al. Effects of GLP-1 receptor agonists on male reproductive function. Andrology. 2015.
14. Jensterle M et al. GLP-1 receptor agonists and testosterone: systematic review and meta-analysis. Andrology. 2020.

Footer disclaimers

Platform Disclaimer. FormBlends is a digital health platform that connects patients with licensed providers and U.S.-based pharmacies. We do not manufacture, prescribe, or dispense medication directly. All clinical decisions are made by independent licensed providers.

Compounded Medication Notice. Compounded semaglutide and tirzepatide are not FDA-approved. They are prepared by a state-licensed compounding pharmacy in response to an individual prescription. Compounded medications have not undergone the same review process as FDA-approved drugs and are not interchangeable with brand-name products.

Results Disclaimer. Individual results vary. Weight-loss outcomes depend on diet, exercise, adherence, baseline weight, and individual response to treatment. Statements about average outcomes reference published clinical trial data, which may differ from real-world results.

Trademark Notice. Metformin is a generic medication. Brand names referenced for educational comparison purposes are trademarks of their respective owners. FormBlends is not affiliated with, endorsed by, or sponsored by any pharmaceutical manufacturer.

Related FormBlends Guides

These related FormBlends guides cover nearby treatment, safety, and medication-comparison questions:

• Online Prescription for Testosterone in 2026: What's Legal, What's Safe, and What Actually Works
• Online Testosterone Therapy: A Clinical Guide to Telehealth TRT in 2026
• Semaglutide vs Metformin for Weight Loss: $4 Drug vs $1000 Drug