Metformin use in women with polycystic ovary syndrome (PCOS): Opportunities, benefits, and clinical challenges

Abstract

Metformin, a synthetic biguanide, is widely used to manage type 2 diabetes, and is commonly prescribed in polycystic ovary syndrome (PCOS) to address insulin resistance and associated metabolic and reproductive disturbances. PCOS is characterised by hormonal imbalances such as hyperandrogenism and anovulation, metabolic abnormalities including insulin resistance and increased cardiometabolic risk, and higher rates of pregnancy complications. However, the role of metformin in the multifaceted nature of PCOS remains debated. This review synthesises the mechanisms of action of metformin and its effects on metabolic, hormonal, reproductive, and pregnancy‐related outcomes in PCOS. In non‐pregnant women, metformin improves insulin resistance, menstrual regularity, and androgen levels, particularly in those with obesity or insulin resistance, and may enhance fertility when combined with other treatments. However, it is not effective as a first‐line therapy for weight loss, ovulation induction, or treatment of clinical hyperandrogenic features, including hirsutism or acne. In pregnancy, metformin may reduce early pregnancy loss, miscarriage, and preterm birth, though findings for gestational diabetes and preeclampsia are inconsistent. Evidence is limited by study heterogeneity, varying diagnostic criteria, and the use of aggregate data in meta‐analyses, all of which make interpretation challenging. Future research should prioritise well‐powered clinical trials, individual patient data meta‐analyses, and longer‐term follow‐up studies, particularly in pregnancy, to better define the populations most likely to benefit from metformin use across the PCOS spectrum.

Plain Language Summary

• Polycystic ovary syndrome (PCOS) is a common condition that affects up to 1 in 10 women of reproductive age. It is characterised by irregular or absent periods, signs of elevated male hormones (high androgens or excess hair growth), and/or polycystic ovaries seen on ultrasound. These features can lead to fertility problems, acne, psychological distress, and an increased risk of various disorders such as depression, type 2 diabetes and heart disease. Many women with PCOS also experience challenges during pregnancy, including a higher risk of miscarriage, preterm birth, and gestational diabetes.

• Metformin is a medication most often used to manage diabetes. In women with PCOS, it can help improve how the body responds to insulin, which may also reduce male hormone levels, improve menstrual cycles, and support fertility. This review examines the role of metformin in treating PCOS—both before and during pregnancy—by summarising key findings from the available evidence.

• In women who are not pregnant, metformin can help improve insulin resistance, hormone levels, and menstrual regularity, particularly among those who are overweight or have signs of insulin resistance. However, metformin alone is not a first‐choice treatment for weight loss, ovulation problems, or symptoms such as acne and unwanted hair growth. When combined with other treatments, such as hormone therapy or fertility medications, it may offer additional benefits.

• During pregnancy, metformin is considered safe for use in women with PCOS and may lower the risk of early pregnancy loss and preterm birth. However, its effects on preventing gestational diabetes or high blood pressure are less clear, with mixed results across studies. Some research suggests that babies exposed to metformin in the womb may have slightly larger head sizes or a higher risk of being overweight in early childhood, but the long‐term health effects remain unknown.

• Overall, metformin can be a helpful part of treatment for some women with PCOS, especially those with insulin resistance or certain pregnancy risks. Still, it is not a one‐size‐fits‐all solution. More high‐quality research is needed to better understand which women benefit most and to assess any long‐term effects on children exposed to metformin during pregnancy.

1. INTRODUCTION

Polycystic ovary syndrome (PCOS) is the most common endocrine condition among reproductive‐aged women, affecting 10%–13% of this population. ¹ , ² , ³ Underpinned by endocrine abnormalities, the multifaceted nature of PCOS includes diverse metabolic, reproductive, dermatological, and psychological sequelae, which vary depending on lifestyle, environmental, ethnic, and genetic factors. ⁴ , ⁵ The heterogenous and evolving nature of symptoms across the lifespan in PCOS contributes to its complex presentation and has historically made diagnosis challenging. ⁶ Based on the International Evidence‐based Guideline PCOS diagnostic criteria, ³ which builds on the earlier consensus‐based Rotterdam criteria, ⁷ PCOS is diagnosed in adults on the basis of two of three components: (i) clinical and/or biochemical hyperandrogenism, (ii) ovulatory dysfunction, including anovulation or chronic oligo‐ovulation, and (iii) polycystic ovary morphology or elevated anti‐Müllerian hormone (AMH) levels, ⁸ after exclusion of other possible aetiologies. ⁹ Testing for the latter of these criteria (ultrasound or AMH) is only required where the first two criteria are not met. ³ In adolescents, both hyperandrogenism and ovulatory dysfunction are required, and ovarian features are not included as they are non‐specific for PCOS at this life stage.

There is now well‐documented evidence that the health impacts associated with PCOS extend beyond infertility, ³ , ¹⁰ , ¹¹ , ¹² with reproductive features including endometrial hyperplasia and cancer, due to lack of progesterone exposure. ¹⁰ Metabolic complications in PCOS include increased obesity and a 4–7‐fold higher lifetime risk of type 2 diabetes (T2D), ¹¹ as well as increased hypertension, hyperlipidemia, subclinical atherosclerosis, and elevated risk of cardiovascular diseases. ¹² Psychological comorbidities are very common, including depression, anxiety, and a reduced quality of life, further exacerbated by social stigma and body image concerns. Dermatological features include hirsutism, acne, and alopecia, which can exacerbate psychological issues and impact body image. ³

The cardiometabolic features of PCOS become particularly critical during pregnancy, a period of heightened physiological demand. ¹³ Insulin resistance and hyperinsulinemia not only persist but often intensify during pregnancy, amplifying the risk of adverse maternal and foetal outcomes. ¹³ Indeed, meta‐analyses have consistently associated maternal PCOS with pregnancy complications including but not limited to early miscarriage, pregnancy‐induced hypertension (PIH), gestational weight gain (GWG), preeclampsia, gestational diabetes mellitus (GDM) and preterm birth. ¹⁴ , ¹⁵ Neonatal complications such as low birthweight, foetal growth restriction, and caesarean sections are also increased. ¹⁴ , ¹⁵ , ¹⁶ Identification, screening, and treatment of PCOS both preconception and in pregnancy is, therefore, strongly recommended in the International PCOS Guideline. ³

Currently, treatment of PCOS in general populations is often symptom‐ and context‐dependent, but generally comprises lifestyle management as first‐line therapy, including diet modification and physical activity. ¹⁷ Education and counselling around the chronic nature of the condition and psychological support are also vital aspects of management for women with PCOS. ³ Pharmacotherapies are commonly used to manage PCOS, depending on individual needs and symptoms. ¹⁸ Cosmetic light therapy, along with the combined oral contraceptive pill (COCP) is the recommended treatment for management of menstrual irregularities and hyperandrogenic symptoms such as hirsutism and acne, ¹⁸ with anti‐androgens as second‐line pharmacological treatments when COCPs are contraindicated or poorly tolerated. ¹⁹ For individuals seeking fertility treatment, ovulation‐induction agents such as letrozole ²⁰ are recommended as first‐line therapy. For metabolic features of PCOS including obesity, insulin resistance, hyperinsulinemia, and dysglycemia, alongside lifestyle modification, metformin is recommended first line, and in some cases, glucagon‐like peptide 1 (GLP‐1) receptor agonists or metabolic surgery. ¹⁷ Given the broad spectrum of symptoms and outcomes associated with PCOS, a tailored combination of lifestyle and pharmacological therapies is essential to improve quality of life and mitigate comorbidities in women with PCOS.

In pregnancy, however, the treatment of PCOS can present unique challenges. While lifestyle and/or weight management remain important and can improve pregnancy outcomes, ²¹ the physiological demands of pregnancy, combined with the metabolic and endocrine disturbances characteristic of PCOS, often necessitate additional pharmacological support. However, pharmacological options used to manage PCOS in non‐pregnant individuals are unsuitable during pregnancy including the COCP and GLP‐1 receptor agonists (e.g., exenatide), due to contraindications and safety concerns. Anti‐androgens, including spironolactone and finasteride, are contraindicated due to teratogenicity. ¹⁹ Similarly, thiazolidinediones, which can improve insulin sensitivity and metabolic features pre‐pregnancy, have raised concerns around growth restriction and teratogenicity, precluding their use in pregnancy. ²⁰ While metformin has been used in pregnancy for decades, and offers a potentially safe pharmacological option for managing metabolic disturbances in PCOS pregnancies, its efficacy and long‐term safety have been questioned. Potential benefits include preventing early pregnancy loss and limiting excess GWG in PCOS pregnancies, ²² but there are conflicting reports regarding its efficacy for other outcomes such as GDM, preterm birth, and preeclampsia. ²³ Furthermore, concerns regarding its long‐term safety, particularly for offspring metabolic health, continue to fuel debate and contribute to inconsistent clinical recommendations around the use of metformin during pregnancy.

In this review, we explore the functions of metformin and provide a comprehensive synthesis of current evidence on metformin use in women with PCOS, including during pregnancy, examining its efficacy for a broad range of metabolic, hormonal, reproductive, and pregnancy outcomes. By critically evaluating the available literature, this review seeks to clarify the therapeutic utility of metformin as a management option in PCOS and to identify key knowledge gaps to inform future research.

2. METHODS

We searched electronic databases including PubMed and Google Scholar to identify studies examining the use of metformin in women with PCOS. In addition, we manually checked the included studies in the reviews to ensure no relevant studies were missed. We included randomised controlled trials, systematic reviews, and meta‐analyses examining the use of metformin in both pregnant and non‐pregnant women with PCOS. Studies with observational designs or without original data, and those not available in English, were excluded, as were studies in women without PCOS at study entry or who did not receive metformin as an intervention. The search was not systematic and is not intended to introduce new findings, but to provide a broad overview of the existing literature on metformin use in PCOS.

3. METFORMIN: PHARMACOLOGY AND MECHANISMS

Metformin is a first‐line insulin sensitiser indicated for the management of T2D. As a synthetic biguanide chemically related to galegine (originating from the plant Galega officinalis), ²⁴ metformin is rapidly absorbed, with a half‐life of approximately 5–6 h, depending on the formulation (immediate or extended release). ²⁵ Unlike many pharmacotherapies, metformin is not bound to plasma proteins, ²⁶ bypasses hepatic metabolism, and is excreted intact in urine. ²⁵ While these properties minimise drug‐to‐drug interactions, metformin can accumulate in those with renal impairment, with a very rare complication being lactic acidosis. ²⁵

The mechanisms of metformin action in PCOS are outlined in Figure 1. ²⁷ , ²⁸ The key therapeutic effects of metformin involve reducing hepatic gluconeogenesis, enhancing peripheral glucose utilization, and increasing insulin sensitivity in skeletal muscle or adipose tissue, without causing weight gain or hypoglycaemia. ²⁹ Metformin has also been recently shown to enhance the release of glucose from the blood into the intestinal tract. ²⁶ , ³⁰ In line with this, high concentrations of metformin accumulate in the intestine — increasingly considered a primary site of metformin action — where it inhibits mitochondrial complex I, reduces NAD+ availability, and promotes the conversion of pyruvate to lactate in enterocytes. ³¹ , ³² This intestinal lactate is then transported to the liver where it contributes to gluconeogenesis in a form of gut‐liver cross‐talk, promoting a splanchnic glucose–lactate–glucose cycle that may help regulate systemic glucose levels, despite its energetic inefficiency. ³¹ , ³² These multifaceted properties have made metformin a cornerstone of therapy in the management of T2D, and increasingly in PCOS, particularly in women with metabolic features of PCOS. At a molecular level, metformin exerts its effects primarily via activation of the 5′ adenosine monophosphate‐activated protein kinase (AMPK) signalling pathway, but this does not account for all the reported actions of metformin as detailed elsewhere. ³³ Activation of AMPK suppresses hepatic gluconeogenesis, enhances glucose uptake via glucose transporter‐4 (GLUT4) translocation, and promotes fatty acid oxidation by inhibiting acetyl‐CoA carboxylase (ACC). Additionally, activated AMPK inhibits nuclear factor kappa B signalling and inflammasome activation, thus improving vascular and endothelial function. ³⁴ AMPK‐independent mechanisms include inhibition of mitochondrial complex I, as noted above in the context of intestinal action. This inhibition also occurs in other tissues, particularly in the liver, where it reduces ATP production and alters cellular redox potential, thereby suppressing gluconeogenesis and improving glucose homeostasis. ³³ Metformin also modulates gut microbiota composition and promotes gut hormones including GLP‐1 and peptide YY (PYY) which mediate hunger and satiety. These actions can mitigate obesity and insulin resistance, features which are central to PCOS, and can reduce hyperinsulinemia‐driven hyperandrogenism, which contributes to ovulatory dysfunction and infertility. The complex, multiorgan effects of metformin are also potentially useful during pregnancy, a period marked by naturally heightened insulin resistance and altered pre‐prandial and post‐prandial glucose metabolism. ³³ The insulin‐sensitising, immunomodulating, and gut hormone‐regulating effects of metformin, ³⁴ , ³⁵ and its protective effects against endothelial dysfunction, ³⁶ suggest a role in preventing adverse outcomes including excess GWG, GDM, macrosomia, and preeclampsia, often exacerbated in PCOS pregnancies. ¹⁴ , ¹⁵

FIGURE 1.

Molecular mechanisms of metformin and its role in PCOS. Dashed lines indicate indirect effects while solid lines suggest direct effects. ¹ Metformin is transported into the cytoplasm via the OCT1 channel across different tissues. Its primary action occurs in mitochondria, where it increases NADH:NAD+ and AMP levels while reducing ATP, leading to enhanced glycolysis and suppressed gluconeogenesis in the liver. Elevated AMP activates AMPK, which also undergoes direct activation by metformin through PEN2 binding and ATP6AP1 interaction. AMPK enhances glycolysis by reducing PEP levels, mediates inflammatory responses by inhibiting NF‐κB signalling, and lowers ROS levels, which play a key role in maintaining cellular health. Through its widespread effects, AMPK improves insulin sensitivity and increases glucose uptake systemically. ² In the liver, AMPK inhibits ACC, thereby suppressing de novo lipogenesis and promoting fatty acid oxidation (acetyl‐CoA carboxylase). ³ In the intestine, metformin‐activated AMPK stimulates bile acid, mucin glycoprotein, and antimicrobial peptide production either directly or by suppressing mTORC1 activity. These changes influence gut microbiota composition and gut hormones (GLP‐1 and PYY), affecting the gut‐brain axis and glucose homeostasis. Additionally, microbiota alterations increase SCFAs, which further regulate metabolic functions. Metformin also inhibits mitochondrial complex I in enterocytes, increasing lactate production, which supports hepatic gluconeogenesis via gut–liver cross‐talk as part of a splanchnic glucose–lactate–glucose cycle. ⁴ In the pancreas, elevated AMPK and GLP‐1 enhance insulin secretion. ⁵ In adipose tissue, AMPK reduces lipolysis and lipogenesis by inhibiting HSL, ATGL, and SREBP1c activity. ⁶ In skeletal muscle, metformin‐activated AMPK increases glucose uptake by promoting GLUT4 translocation to the membrane. In cardiomyocytes, AMPK facilitates FFA uptake by translocating the CD36 receptor. ⁷ In the ovary, metformin directly modulates FSHR activity, reducing FSH levels. It also lowers androgen levels by inhibiting the HSD3B2 and CYP17‐lyase enzymes through mitochondrial‐mediated pathways. This reduction in androgens may alleviate symptoms of PCOS driven by hormonal imbalances. ACC, acetyl‐CoA carboxylase; AMPK, AMP‐activated protein kinase; ATGL, adipose triglyceride lipase; ATP6AP1, ATPase H+ transporting accessory protein 1; CD36, cluster of differentiation 36; CYP17, Cytochrome P450 17; FFA, free fatty acid; FSH, follicle‐stimulating hormone; FSHR, follicle‐stimulating hormone receptor; GLP‐1, glucagon‐like peptide‐1; GLUT4, glucose transporter type 4; HSD3B2, Hydroxy‐Delta‐5‐Steroid Dehydrogenase, 3 Beta‐ And Steroid Delta‐Isomerase 2; HSL, hormone‐sensitive lipase; mTORC1, mechanistic target of rapamycin complex 1; NADH:NAD+, nicotinamide adenine dinucleotide + hydrogen; NF‐κB, nuclear factor κ‐light‐chain‐enhancer of activated B cells; OCT1, Organic Cation Transporter 1; PEN2, presenilin enhancer 2; PEP, phosphoenolpyruvate; PYY, peptide tyrosine–tyrosine; ROS, reactive oxygen species; SCFAs, short‐chain fatty acids; SREBP1c, sterol regulatory element‐binding protein 1c.

4. EVIDENCE SYNTHESIS: METFORMIN EFFICACY IN PCOS

4.1. Randomised controlled trials in non‐pregnant PCOS populations

Several randomised controlled trials (RCTs) have assessed the efficacy of metformin (either alone or in combination with other treatments) in managing hormonal, reproductive, and metabolic features in adolescents and adults with PCOS (Table S1). However, its effects vary by treatment duration, metabolic profile, and individual characteristics. Below is a summary of findings from trials in general/mixed and selected PCOS populations.

4.1.1. General and mixed PCOS populations

Among general PCOS populations, metformin has demonstrated variable but frequently beneficial effects across metabolic, hormonal, and reproductive outcomes when compared with placebo, other pharmacological agents, or lifestyle and alternative treatments. Compared with placebo, short‐term trials (3–4 months; n = 29–40) using 1500 mg/day metformin reported reductions in luteinising hormone (LH), follicle‐stimulating hormone (FSH), blood pressure, ³⁷ fasting insulin, homeostatic model assessment of insulin resistance (HOMA‐IR), waist circumference, and testosterone, ³⁸ but no effects on other androgens or lipids. A three‐month trial (n = 118) examining bone parameters also found that using 1500–2000 mg/day metformin may reduce bone turnover by lowering markers of formation and resorption compared with placebo. ³⁹ Longer trials (6 months; n = 27–100) using metformin doses of 1500–1700 mg/day reported similar metabolic improvements, including improved lipid profile, insulin sensitivity, blood pressure, and testosterone compared with placebo, while AMH and adiponectin remained unchanged. ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ In a 14‐month cross‐over study of women with PCOS and hirsutism (n = 16), metformin reduced weight and hair growth and improved cycle frequency compared with placebo, with no change in sex hormone binding globulin (SHBG) and free androgen index (FAI). ⁴⁵ One stratified trial (n = 116) reported phenotype‐specific effects, whereby 1700–2550 mg/day of metformin for 6 months reduced dehydroepiandrosterone‐sulphate (DHEA‐S) levels and improved hirsutism and ovulation in lean hyper‐ or normoinsulinemic women with PCOS; and reduced waist‐to‐hip ratio (WHR) and cycle duration in overweight or obese normoinsulinemic women, respectively. ⁴⁶

In head‐to‐head comparisons with other pharmacological agents, six‐month trials (n = 35–96) showed that 1000–2000 mg/day of metformin was more effective than COCPs — including ethinyl estradiol‐cyproterone acetate (EE‐CA)— in improving asymmetric dimethylarginine, HOMA‐IR, homocysteine levels, blood pressure and lipid profiles, with greater reductions in C‐reactive protein (CRP) in women with obesity. ⁴⁷ , ⁴⁸ , ⁴⁹

Combination therapy with COCPs has shown further benefits. Three‐month trials (n = 43–48) combining 1700 mg/day metformin with EE‐CA improved endothelial function, insulin resistance, lipid profile, free testosterone, and DHEA‐S, compared with EE‐CA alone. ⁵⁰ , ⁵¹ Similarly, a six‐month trial (n = 76) in a hyperinsulinemic PCOS population found superior metabolic and vascular outcomes with metformin (1500 mg/day) plus drospirenone/ ethinylestradiol (DRSP/EE) versus either alone, ⁵² while a three‐month trial (n = 53) showed suppression of hyperandrogenism with metformin plus COCPs. ⁵³ In longer studies of 6–12 months (n = 28–87), combining 500–2000 mg/day metformin with various COCPs led to greater reductions in FAI, free testosterone, and inflammatory markers, and improvements in body composition, facial hirsutism, vascular function, and aortic elasticity than either treatment alone, although no additional benefits were seen for lipids, insulin sensitivity, SHBG, or testosterone beyond those provided by COCPs alone. ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² Other studies (n = 34–65) found that metformin (1500–2000 mg/day) improved insulin sensitivity with or without COCPs after 6 months, ⁶² , ⁶³ but had no consistent impact on body mass index (BMI), inflammatory markers, or androgens, ⁶⁰ , ⁶⁴ and was linked to hypoglycaemia in some cases. ⁶³ In a 12‐month study (n = 99), EE‐CA and ethinylestradiol/ desogestrel (EE/DES) alone or combined with metformin were equally effective for improving hirsutism, with no added benefits from low‐dose metformin. ⁶⁵ Conversely, two studies (n = 60, 87) combining metformin (500–1700 mg/day) with Diane‐35 for 3 months reported greater reductions in androgens and body fat and better metabolic profiles with the combination treatment, compared with Diane‐35 alone. ⁶⁶ , ⁶⁷

Metformin has also been evaluated alongside or against non‐COCP pharmacological agents. A three‐month trial (n = 97) showed that 1000 mg/day metformin plus pioglitazone had superior effects on insulin resistance and inflammatory markers than metformin monotherapy. ⁶⁸ , ⁶⁹ In 34 women with PCOS and prediabetes, adding the dipeptidyl peptidase‐4 (DPP‐4) inhibitor, saxagliptin, to metformin improved the insulin secretion‐sensitivity index and menstrual regularity over 4 months, compared with metformin alone. ⁷⁰ Among 75 women with PCOS and hyperprolactinemia, 1000 mg/day metformin combined with cabergoline for 3 months reduced BMI, testosterone, and LH; increased FSH; and improved endometrial blood flow and ovulation. ⁷¹

In comparison with alternative treatments, three‐month treatment with metformin (1000–1500 mg/day) showed similar effects to lavender (n = 62) for increasing progesterone and cycle regulation, ⁷² and to berberine (n = 89) for improving insulin sensitivity and hyperandrogenism. ⁷³ A two‐month trial (n = 110) reported that 1500 mg/day metformin was superior to fenugreek for improving BMI, WHR, insulin resistance, fasting blood glucose, and menstrual regularity. ⁷⁴ Combination treatment with metformin (1500 mg/day) and curcumin for 3 months (n = 195) improved lipid profile, glucose metabolism, LH/FSH balance, body weight, and BMI compared with either agent alone or placebo. ⁷⁵ Similarly, 1700 mg/day metformin combined with a structured exercise regimen improved menstrual regularity, hirsutism, body composition, and testosterone compared with placebo and exercise. ⁷⁶

Overall, metformin shows consistent metabolic benefits in general PCOS populations, particularly for improving insulin sensitivity, glucose metabolism, and lipid profiles. Effects on androgens and hirsutism are inconsistent and appear more pronounced with longer treatment or combination therapies, namely with COCPs. Benefits for ovulation and menstrual regularity are greatest in women with insulin resistance or metabolic dysfunction, with limited effects in normoinsulinemic phenotypes.

4.1.2. Overweight and Obese PCOS populations

In PCOS populations with overweight or obesity, metformin has demonstrated variable effects on metabolic and reproductive outcomes. In short‐term trials (3–6 months; n = 18–122), metformin at doses of 1000–2000 mg/day improved insulin resistance and reduced WHR, testosterone, free fatty acids, and total cholesterol when compared with EE‐CA or placebo. ⁷⁷ , ⁷⁸ However, its effects on BMI, body composition, insulin, metabolic markers, and menstrual frequency were inconsistent. Some trials (n = 76–105) reported modest improvements over 6–12 months when metformin was combined with a hypocaloric diet, with or without flutamide, leading to reductions in BMI, waist and hip circumference, fasting glucose, insulin, and triglycerides. ⁷⁹ , ⁸⁰ Yet, others (n = 38–122) found no significant changes in these outcomes after 6 months despite improved insulin resistance. ⁸¹ , ⁸² , ⁸³ , ⁸⁴

Combination therapies with pharmacological agents, namely GLP‐1 receptor agonists, have shown the most consistent additive effects among populations with overweight or obesity. In two 3‐month trials (n = 36–52), 2000 mg/day of metformin plus liraglutide improved weight loss, waist circumference, and reproductive hormones (testosterone, SHBG, FAI, LH, progesterone) versus metformin alone. ⁸⁵ , ⁸⁶ Similar benefits were observed for exenatide or beinaglutide combined with metformin (500–2000 mg/day) over 3–6 months (n = 40–60), demonstrating superior effects on weight, BMI, abdominal fat, waist circumference, insulin sensitivity, ovulation, menstrual regularity, FAI, and androgen levels compared with either agent alone, with generally good tolerability. ⁸⁷ , ⁸⁸ , ⁸⁹ For DPP‐4 inhibitors, sitagliptin (n = 24) added to metformin (2000 mg/day for 3 months) prevented weight regain in 24 women previously treated with liraglutide, compared with metformin alone. ⁹⁰ In a six‐month study (n = 204), rosiglitazone—alone or with metformin (1000–1500 mg/day) and lifestyle modification—reduced cholesterol and triglycerides more than metformin alone, although metformin led to greater weight loss. ⁹¹ In contrast, adding the sodium‐glucose co‐transporter 2 (SGLT2) inhibitor, canagliflozin, to metformin (1000–2000 mg/day) for 3 months reduced testosterone and glucose–insulin responses in PCOS (n = 41), but did not improve weight, menstrual frequency, or hormonal profiles. ⁹² Anti‐obesity medications were explored in a three‐month study (n = 240), whereby adding orlistat to metformin (500–1500 mg/day) and Diane‐35 resulted in greater reductions in weight, BMI, and blood pressure, and increased high‐density lipoprotein cholesterol (HDL‐C) compared with Diane‐35 alone. ⁹³ Metformin (1500 mg/day) combined with Diane‐35 over 3 months (n = 19) also improved vascular endothelial function compared with Diane‐35 alone. ⁹⁴

Lifestyle and nutraceutical approaches have also been studied in this population. In a twelve‐month pilot trial (n = 23), combining 1700 mg/day metformin with a lifestyle intervention reduced androgens and weight compared with lifestyle and placebo, though ovulation outcomes were unchanged. ⁹⁵ A three‐month hypocaloric diet alone outperformed metformin (1000–2000 mg/day; n = 30) in reducing weight and hs‐CRP. ⁹⁶ Among nutraceuticals, metformin (1000 mg/day) was compared with myoinositol for 6 months (n = 53), the latter showing greater improvements in menstrual regularity and quality of life. ⁹⁷

In summary, metformin improves insulin resistance in women with PCOS and overweight or obesity, although effects on weight, hormones, and menstrual function are inconsistent. Greater benefits are seen when combined with GLP‐1 receptor agonists, while other agents and lifestyle changes may offer additional improvements, especially in relation to anthropometry and glucose and lipid metabolism.

4.1.3. Non‐obese PCOS populations

Few trials have explored the use of metformin in non‐obese populations with PCOS. In 2 six‐month trials (n = 17 and 29), 1000–2000 mg/day of metformin reduced fasting insulin and androgens compared with placebo, ⁹⁸ and improved fasting insulin, WHR, testosterone, and menstrual regularity, but not glucose tolerance or insulin sensitivity, compared with EE‐CA. ⁹⁹ In two trials among lean women (n = 94 and 109), 850–2000 mg/day metformin lowered fasting insulin more than estradiol–cyproterone acetate after 4 months ¹⁰⁰ and reduced advanced glycation end‐products and CRP levels compared with COCPs after 6 months. ¹⁰¹ In two other six‐month trials (n = 87 and 100), using 1700 mg/day of metformin improved insulin resistance, adiponectin, CRP, apolipoprotein B, plasminogen activator inhibitor‐1, and homocysteine, compared with EE ¹⁰² ; and improved ovulation, systolic blood pressure, and insulin sensitivity indices more effectively than rosiglitazone. ¹⁰³ However, adding metformin to either treatment did not yield further improvement. ¹⁰² , ¹⁰³

Other studies have reported additional benefits with combined therapies. In a four‐month trial (n = 40), metformin (1500 mg/day) with EE‐CA improved insulin sensitivity, reduced BMI, WHR, and androstenedione, and increased SHBG compared with EE‐CA alone. ¹⁰⁴ Similar effects were seen in a six‐month trial (n = 50) where combining low‐dose metformin (500 mg/day) with cyproterone acetate (CPA) improved insulin sensitivity and suppressed hyperandrogenism more effectively than CPA alone. ¹⁰⁵

Despite the limited number of studies, metformin has shown consistent effects in non‐obese women with PCOS, whereby it improves insulin resistance and select metabolic markers, with modest and varying effects on hyperandrogenism and ovulation. Benefits appear more pronounced when combined with hormonal agents, particularly in increasing SHBG and improving insulin‐related parameters.

4.1.4. Anovulatory and infertile PCOS populations

In women with PCOS and oligo−/anovulation or infertility, metformin has shown promising effects on menstrual regularity, ovulation, and metabolic parameters, particularly among those with insulin resistance or obesity. In short‐term placebo and head‐to‐head comparisons (3–3.5 months; n = 26–320), metformin (500–2000 mg/day) improved ovulation and pregnancy rates compared with placebo, EE‐CA, and myo‐D‐chiro‐inositol, while also reducing testosterone, LH, FSH, BMI, insulin resistance, and lipids, and increasing SHBG. ¹⁰⁶ , ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ A longer 12‐month trial (n = 27) also demonstrated improved insulin sensitivity with 1700 mg/day metformin compared with placebo, ¹¹¹ and one study (n = 320) found improved live birth rates when metformin was continued for 3 months into early pregnancy. ¹¹² However, in a PCOS population with infertility and overweight/obesity (n = 160), exenatide for 3 months led to a higher spontaneous pregnancy rate compared with 1000–2000 mg/day metformin, although overall pregnancy rates were similar by month 16 (exenatide 79.2%, metformin 76%) with no differences in outcomes. ¹¹³ Some studies suggest that the effects of metformin on reproductive and fertility outcomes are more pronounced among those with insulin resistance ¹⁰⁹ , ¹¹⁰ or obesity, ¹¹² while others have reported improved menstrual regularity in normoinsulinemic anovulatory women (n = 23) after 6 months of treatment. ¹¹⁴ By contrast, metformin has shown limited benefit in clomiphene citrate (CC)‐resistant PCOS populations. In two small trials (n = 20 and n = 32), 1500 mg/day of metformin for 3 months improved BMI, testosterone, leptin, and lipids, but had no effects on androgens, insulin resistance, or ovulation or pregnancy rates compared with placebo. ¹¹⁵ , ¹¹⁶ Similarly, in 150 women with infertility and letrozole resistance, three‐month treatment with either metformin (1500 mg/day) or inositol, both with folic acid, showed no significant effect on ovarian function or pregnancy rate compared with folic acid alone, though inositol was more effective than metformin in women with normal BMI. ¹¹⁷

Combination treatments are thought to be more effective than metformin alone for a range of outcomes in women with PCOS and infertility. In two trials (n = 36–97) metformin (1000 mg/day metformin) combined with pioglitazone for 3 months improved SHBG, AMH, inflammatory markers (IL‐6 and IL‐8) and postprandial glucose levels, and reduced insulin resistance compared with metformin alone, while maintaining stable weight. ⁶⁸ , ⁶⁹ Similarly, in women with obesity and infertility (n = 28), adding low‐dose liraglutide to metformin for 3 months improved pregnancy rates per embryo transfer and cumulative pregnancy rates compared with metformin alone, despite similar weight loss between groups. ¹¹⁸ Metformin has also been combined with ovulation induction agents. In treatment‐naïve women (n = 105), combining 1700 mg/day metformin with CC for 6 months resulted in the highest ovulation, pregnancy, and live birth rates than either agent alone. ¹¹⁹ Conversely, in CC‐resistant women (n = 59), combining metformin (1500 mg/day) and letrozole produced similar ovulation rates to metformin and CC after 6–8 weeks, but improved endometrial thickness and full‐term pregnancy rates, suggesting an advantage of letrozole in this subgroup. ¹²⁰

However, not all combination therapies show clear benefit. In a six‐month trial (n = 114), metformin (500–2000 mg/day) with lifestyle changes did not improve ovulation or weight loss compared with placebo plus lifestyle, though bone mineral density increased. ¹²¹ The PCOSMIC trial found that six‐month treatment with metformin (1500 mg/day), alone or with CC, did not improve pregnancy or live birth rates in 117 infertile women with PCOS and BMI ≤32 kg/m². ¹²² Similarly, a trial of 107 women found similar ovulation rates in those receiving 1700–2000 mg/day metformin plus CC versus CC alone (with both groups receiving ≥6 weeks of metformin pre‐treatment). ¹²³ Another trial (n = 116) comparing metformin (1500 mg/day) plus myoinositol for 6 months versus myoinositol alone showed similar clinical pregnancy rates and metabolic and hormonal parameters, though the combination group had more gastrointestinal side effects, suggesting myoinositol alone may be preferable. ¹²⁴ In the landmark Pregnancy in Polycystic Ovary Syndrome I (PPCOS I) trial (n = 626), CC alone or combined with metformin was superior to metformin alone (500–2000 mg/day for 6 months) in achieving live births among infertile women with PCOS. ¹²⁵ A subsequent secondary analysis of this trial (n = 323) demonstrated that higher baseline AMH levels were associated with reduced ovulation consistently across all treatment groups, suggesting that metformin did not mitigate the negative impact of elevated AMH on ovulation in women with PCOS. ¹²⁶

Taken together, current evidence suggests that metformin monotherapy in PCOS populations with infertility seems beneficial for ovulation and pregnancy in some subgroups, especially in treatment‐naïve women, but is less effective in CC‐ or letrozole‐resistant subgroups. Combination therapies, especially with ovulation induction or adjunct metabolic agents, tend to yield better fertility outcomes than metformin monotherapy; however, letrozole remains first‐line therapy for ovulation induction in PCOS. Individual factors such as baseline BMI, insulin sensitivity, and AMH levels appear to influence responses to treatment.

4.1.5. Adolescent PCOS populations

In adolescents with PCOS, the effects of metformin have varied depending on treatment durations and combinations with other interventions. In a three‐month trial among 21 adolescents with hyperinsulinemia, 1500 mg/day of metformin reduced total testosterone, improved menstrual regularity, and increased HDL‐C, with no impact on insulin sensitivity or body weight compared with placebo. ¹²⁷ Extending treatment to 6 months with a higher dose (1700 mg/day) in 90 adolescents led to significant weight loss and lower fasting insulin levels compared with COCP alone. ¹²⁸ However, other six‐month trials comparing metformin and COCP, with or without lifestyle, have yielded mixed findings. One trial (n = 22) found no added benefit with 1000–2000 mg/day metformin, with COCPs being more effective in reducing free testosterone and improving menstrual regularity, ¹²⁹ while another (n = 31) in adolescents with obesity reported similar effects of both treatments on androgens, weight loss, and insulin sensitivity. ¹³⁰ Combining metformin (1700 mg/day) with lifestyle modifications and COCPs for 6 months reduced central adiposity and total testosterone, and increased HDL‐C, but did not improve weight loss in 43 adolescents with obesity. ¹³¹ Similarly, adding metformin to lifestyle changes (n = 22) over 6 months did not improve hyperandrogenism compared with lifestyle alone, and was associated with increased gastrointestinal side effects. ¹³² Data on metformin use in adolescents with PCOS are limited and variable, precluding definitive conclusions regarding its efficacy in this population.

4.2. Systematic reviews and meta‐analyses in non‐pregnant PCOS populations

Several systematic reviews and meta‐analyses have evaluated the efficacy of metformin in PCOS compared with placebo or lifestyle, pharmacological agents, or alternative treatments (Table S1). A meta‐analysis of 32 RCTs found that metformin, alone or with lifestyle intervention, led to greater reductions in BMI and improvements in menstrual cycle duration and frequency compared with placebo, although no additional benefits were observed over lifestyle alone. ¹³³ Metformin also improved a range of hormonal and metabolic outcomes including total testosterone, fasting glucose, and lipid profiles, with greater benefit in those with a BMI ≥25 kg/m². ¹³³ Two additional meta‐analyses examining metformin monotherapy (n = 18 and n = 14 RCTs) showed significant reductions in total testosterone, FAI, and AMH levels compared with placebo, with stronger effects observed in younger women or those with baseline AMH >4.7 ng/mL. ¹³⁴ , ¹³⁵ Cardiovascular improvements have also been reported, with a systematic review of 12 studies finding that metformin significantly improved endothelial function and carotid intima‐media thickness, suggesting a role in reducing early cardiovascular risk markers. ¹³⁶

Consistent with primary data, pooled analyses have demonstrated further benefits with combination treatments. A systematic review of 16 RCTs showed that adding metformin to DPP‐4 inhibitors significantly improved glycemic control, while adding it to thiazolidinediones improved lipid profiles and adding it to GLP‐1 receptor agonists enhanced anthropometric outcomes, compared with metformin alone; however, hormonal effects were inconclusive. ¹³⁷ Moreover, a meta‐analysis of 8 RCTs found that GLP‐1 receptor agonists combined with metformin improved weight, BMI, waist circumference, HOMA‐IR, fasting glucose, and SHBG, without increasing adverse events. ¹³⁸ Building on this, two other meta‐analyses, each comprising 9 RCTs, compared exenatide (alone or with metformin) to metformin alone and reported greater reductions in weight, waist circumference, and FAI, along with improvements in SHBG and the Matsuda index, again with no increase in adverse events. ¹³⁹ , ¹⁴⁰

Combinations with COCPs have shown mixed results. A large meta‐analysis of 36 RCTs comparing metformin versus COCP, and metformin + COCP versus COCP alone, reported that combination therapy improved biochemical hyperandrogenism, insulin resistance, and insulin levels more effectively than COCP alone, while COCP monotherapy was more effective for regulating cycles and reducing hirsutism. ¹⁴¹ Earlier meta‐analyses– one in non‐obese women (14 RCTs) ¹⁴² and another in the broader PCOS population (33 RCTs) ¹⁴³ – found that adding metformin to COCPs or anti‐androgens improved BMI and glucose tolerance compared with hormonal therapy alone.

For comparisons with nutraceuticals, two systematic reviews and meta‐analyses (each including 8 RCTs) found no significant differences between metformin and myo‐inositol in metabolic or hormonal outcomes, although myo‐inositol showed a better safety profile with fewer side effects. ¹⁴⁴ , ¹⁴⁵ Another meta‐analysis of 9 RCTs (n = 638) reported that myo‐inositol improved fertility outcomes by modulating hyperandrogenism more effectively than metformin, although ovarian function and BMI were not significantly different between the groups. ¹⁴⁶

Other extensive meta‐analyses have been conducted examining fertility and ovulation outcomes in PCOS. The most recent is a technical report informing the 2023 International PCOS Guideline ³ which included meta‐analyses of 36 RCTs comparing metformin with placebo, CC, letrozole, and FSH. ¹⁴⁷ Metformin improved live birth, clinical pregnancy, and ovulation rates versus placebo. While metformin and CC were comparable for live birth, their combination outperformed CC alone in improving ovulation and live birth rates. Adding metformin to letrozole was superior to metformin plus CC, but not superior to letrozole alone, for clinical pregnancy and ovulation rates, although these findings were of low certainty due to high risk of bias. No benefit was seen for metformin plus FSH over FSH alone. Earlier meta‐analyses of 35 RCTs and 21 RCTs found similar results; one showed that metformin improved clinical pregnancy rates and reduced the risk of ovarian hyperstimulation syndrome in women undergoing assisted reproductive technology compared with placebo or no treatment ¹⁴⁸ ; and the other reported that metformin modestly improved clinical pregnancy rates compared with placebo (47.7% vs. 42.9%) in non‐obese women with PCOS, but had comparable efficacy to CC. ¹⁴⁹ While adding metformin to CC reduced miscarriage risk compared with placebo, it was associated with a higher miscarriage rate than CC alone and lower clinical and multiple pregnancy rates than letrozole. ¹⁴⁹ Additionally, a broader systematic review of 38 studies encompassing both pregnant and non‐pregnant women highlighted the potential role of metformin in restoring ovulatory cycles and enhancing pregnancy rates and outcomes compared with placebo or no treatment. ¹⁵⁰

4.3. Randomised controlled trials in pregnant PCOS populations

Multiple RCTs have examined the efficacy and safety of metformin in pregnancies complicated by PCOS (Table S1). Some trials ¹⁵¹ , ¹⁵² have demonstrated that metformin use during the first trimester significantly reduced early pregnancy loss, with rates of 10.8% versus 42.2%, ¹⁵¹ and 10% versus 26%, ¹⁵² compared with control treatments (e.g., folic acid or no intervention). Benefits were also seen among women with PCOS using metformin, including reductions in miscarriage, preeclampsia, GDM, caesarean section rates, and preterm labour, along with an increase in live birth rates compared with placebo or no intervention. ¹¹² , ¹⁵³ , ¹⁵⁴ An early pilot study (n = 40) in Norway by Vanky et al. ¹⁵⁵ showed a reduction in severe pregnancy complications, with no change in androgens, among women with PCOS receiving 850 mg of metformin twice daily compared with those receiving placebo. ¹⁵⁵ In the landmark RCT by the same group, the PregMet trial (n = 273), metformin during pregnancy in women with PCOS did not impact pregnancy complications such as GDM, preeclampsia, or preterm labor, nor did it influence neonatal outcomes, including birth weight, length, or Apgar scores compared with placebo. ¹⁵⁶ The subsequent PregMet2 study conducted in Norway, Sweden, and Iceland is the largest placebo‐controlled, double‐blind, multicenter RCT to date examining metformin use in women with PCOS in pregnancy (n = 487). ¹⁵⁷ It showed that metformin from late first trimester until delivery reduced the rates of late miscarriage and preterm birth but had no effect on the development of GDM or preeclampsia compared with placebo. Combined analysis of data from the three trials (n = 772) did not alter these results, with no effects on GDM prevention, irrespective of diagnostic criteria. ¹⁵⁸ The lack of GDM prevention in this high‐risk population is surprising, given the well‐established efficacy of metformin in T2D prevention and treatment, potentially related to different pathophysiological mechanisms during pregnancy compared with the non‐pregnant state. ¹⁵⁷

Several sub‐studies from the above RCTs have explored outcomes of maternal metformin use in PCOS. ¹⁵⁹ , ¹⁶⁰ , ¹⁶¹ , ¹⁶² , ¹⁶³ , ¹⁶⁴ In one sub‐study (n = 73), the metformin‐treated group had reduced excess GWG compared with placebo, associated with improved leptin sensitivity. ¹⁶² Post‐hoc analyses from the PregMet trial explored metformin, insulin, and androgen concentrations and foetal growth parameters. ¹⁶¹ , ¹⁶³ Metformin from the first trimester until delivery reduced insulin levels at delivery (n = 236 mothers), ¹⁶¹ with no effect on foetal insulin concentrations. There was no effect on androgen levels in the overall population of women with PCOS (n = 262), but metformin significantly reduced androstenedione in non‐obese women with PCOS and lowered both androstenedione and testosterone in those carrying a male foetus. However, another sub‐study (n = 258 infants) using gestational age‐ and sex‐adjusted z‐scores reported that offspring of overweight mothers exposed to metformin had larger head circumferences compared with those receiving placebo, but birth weight and length were unaffected. ¹⁶⁰ Combined analysis from the pilot, PregMet, and PregMet2 trials (n = 779 infants) also found increased head circumference primarily in offspring of overweight/obese hyperandrogenic mothers with PCOS, but no effect on birth weight or length. ¹⁵⁹ Maternal diabetes and/or overweight and obesity are known to enhance foetal growth by increasing foetal glucose transfer, which stimulates foetal insulin production. ¹⁶⁰ This typically promotes disproportionate foetal growth, characterised by increased abdominal and body size with relatively smaller head circumference. In contrast, metformin‐exposed pregnancies in women with PCOS have shown increased foetal head circumference, suggesting more symmetric growth, possibly due to reduced foetal hyperinsulinemia, though the exact mechanism of action and long‐term clinical implications are yet to be determined. Offspring cognitive function was also explored in the CogMet study (n = 93 children) ¹⁶⁴ based on follow‐up data from the pilot and PregMet studies. They found that in utero metformin exposure did not influence cognitive function of children (mean age: 7.7 years) born to women with PCOS. ¹⁶⁴ However, a higher incidence of borderline intellectual function was observed in metformin‐exposed children, though this finding should be interpreted cautiously due to the small sample size. ¹⁶⁴

Other follow‐up studies explored maternal and offspring anthropometric and cardiometabolic outcomes at different timepoints. In utero metformin‐exposed children had higher BMI z‐scores and a greater prevalence of overweight/obesity than placebo‐exposed ones when examined at up to 4 years (n = 160 children) using combined pilot and PregMet data. ¹⁶⁵ However, a follow‐up study of pilot data (n = 25 children) did not find differences in child anthropometric measures at 8 years of age, ¹⁶⁶ noting a small sample size; but it did report that in utero metformin exposure was associated with increased fasting glucose, elevated systolic blood pressure, and a more favourable lipid profile. The PedMet study (n = 141 children) followed up children from the Pilot and PregMet study at 8 years of age and found that in utero metformin‐exposed children had higher BMI, WHR, waist circumference, and a greater prevalence of obesity compared with placebo‐exposed children. ¹⁶⁷ Using PedMet data, metformin was associated with increased 11‐deoxycortisol (a precursor of cortisol and androstenedione) in boys and elevated 17‐hydroxyprogesterone (a precursor of 11‐deoxycortisol and androstenedione) in metformin‐exposed children. ¹⁶⁸ Moreover, at 5–10 year follow‐up, metformin did not alter the metabolic profile of 131 mothers with PCOS (BMI, weight gain, blood pressure, lipids, glucose, insulin or metabolic syndrome). ¹⁶⁹

The evidence presented suggests that metformin use in pregnancy among women with PCOS appears generally safe and may reduce pregnancy loss and preterm birth. However, its impact on other outcomes such as GDM and preeclampsia is inconsistent, reflecting underlying study heterogeneity. Metformin may promote more symmetric foetal growth; however, the clinical significance of these findings remains uncertain. Emerging follow‐up data raise questions about potential long‐term effects on offspring anthropometry, urging caution around the routine use of metformin in pregnancy, pending further research.

4.4. Systematic reviews and meta‐analyses in pregnant PCOS populations

Several systematic reviews and meta‐analyses have examined the use of metformin in PCOS during pregnancy (Table S1), with most (many with overlapping data) reporting that metformin was effective in reducing pregnancy complications. A systematic review and meta‐analysis of 9 RCTs and 8 cohort studies found that metformin reduced maternal complications (early pregnancy loss, GDM incidence, the need for insulin treatment, excess GWG, and pre‐eclampsia) but had no effect on birth weight, length, or Apgar scores, except for an increase in neonatal head circumference. ²² Similarly, pooled analysis of 18 RCTs showed reduced PIH, macrosomia, preterm delivery, and maternal insulin resistance, as well as increased SHBG levels in infants, but suggested potential risks including larger head circumference in infants and higher long‐term BMI in offspring. ¹⁷⁰ , ¹⁷¹ Metformin was also associated with reduced maternal pregnancy complications (e.g., PIH, GDM, macrosomia, and preterm birth), increased head circumference at birth, and a higher risk of obesity in offspring, in an umbrella systematic review that included 13 different types of reviews. ¹⁷⁰ Smaller systematic reviews and meta‐analyses based on fewer RCTs, reviews, and observational studies also reported that metformin reduced early pregnancy loss, ²² , ¹⁷² , ¹⁷³ late miscarriage, ¹⁷⁴ , ¹⁷⁵ GDM, ²² , ¹⁷⁰ , ¹⁷² , ¹⁷³ , ¹⁷⁶ pre‐eclampsia, ²² , ¹⁷³ PIH, ¹⁷⁰ , ¹⁷¹ , ¹⁷² macrosomia, ¹⁷⁰ and preterm delivery ¹⁷⁰ , ¹⁷² , ¹⁷³ , ¹⁷⁴ , ¹⁷⁷ , ¹⁷⁸ in metformin‐exposed women with PCOS, with increased head circumference in offspring. ¹⁷⁷ Conversely, an earlier systematic review of 5 RCTs and 8 observational studies found that metformin did not prevent pregnancy complications, such as GDM. ¹⁷⁹ Importantly, metformin use in pregnant women with PCOS was not associated with an increased risk of major congenital anomalies in reviews of RCTs and observational studies, ¹⁸⁰ or with birth defects ¹⁸¹ or other foetal abnormalities ¹⁸² in reviews of RCTs only.

5. SUMMARY OF FINDINGS AND LIMITATIONS

Based on the available evidence, metformin in PCOS is effective for improving insulin resistance, regulating menstrual cycles, and reducing androgen levels, with additional benefits for ovulation and some metabolic outcomes—particularly in women with elevated BMI or glucose intolerance. However, current data do not support the use of metformin as a first‐line therapy for weight loss, ovulation induction, or treatment of hyperandrogenic symptoms such as hirsutism or acne. These findings are broadly consistent with those of the 2023 International PCOS Guideline, ³ which recommends metformin as an adjunct to lifestyle intervention in women with metabolic risk, and as a second‐line treatment in selected clinical scenarios including anovulatory infertility and impaired glucose tolerance.

For metformin use in PCOS pregnancies, both primary and pooled data suggest that metformin is generally safe and may reduce early pregnancy loss, late miscarriage, preterm delivery, excess GWG, and the need for insulin treatment in women with PCOS. However, findings related to GDM, PIH, and preeclampsia are inconsistent, likely due to heterogeneity in study designs, diagnostic criteria, and the inclusion of varying populations and treatments. Similar inconsistencies are seen across neonatal and offspring outcomes. Many meta‐analyses combine data from women with PCOS and metabolic comorbidities, such as GDM, which may confound the results and reduce applicability specifically to the PCOS population. Limited long‐term follow up and high attrition rates in long‐term studies further complicate interpretation. These limitations suggest caution when drawing conclusions and highlight the need for more rigorous, stratified analyses of integrated primary data (vs. aggregate‐level data) to clarify the efficacy of metformin in PCOS during pregnancy.

Key limitations across the PCOS literature include the notable variations in study populations, outcomes, diagnostic criteria, treatment targets and regimens, as well as methodological limitations, which collectively hinder interpretation and generalizability of findings. Baseline risk factors such as BMI, insulin resistance, ethnicity, family history, comorbidities, and the use of concomitant medications (e.g., anti‐hypertensives, thyroid hormones, antidepressants), as well as behavioural factors such as diet, physical activity, and breastfeeding, vary widely across studies and are often poorly controlled. Study design limitations—such as inadequate randomisation, blinding, dose inconsistency, poor adherence, and delayed metformin initiation—further affect the reliability of results. While systematic reviews have sought to integrate existing data generating larger cohorts, this inherent heterogeneity and the inability to explore confounders with aggregate data preclude meta‐analysis for certain outcomes and raise concerns around the accuracy of pooled estimates for others.

To address these challenges and advance the field, high‐quality, well‐controlled, and adequately powered clinical trials conducted with long‐term follow‐up are now needed. Additionally, integrated analyses of individual patient data from RCTs are essential but, as yet, have not been conducted in the context of PCOS. In pregnancy, the Metformin in Pregnancy Study (MiPS) ¹⁸³ represents a key step toward outcome harmonization to evaluate maternal and offspring health, including in PCOS; however, broader consortia are needed to clarify the effects of metformin across different PCOS phenotypes, populations and life stages.

6. CONCLUSIONS

Metformin improves insulin resistance, menstrual regularity, and androgen levels in women with PCOS, particularly among those with obesity or insulin resistance. It may also support fertility and pregnancy outcomes when combined with other therapies. However, current evidence does not support the use of metformin as a first‐line therapy for weight loss, ovulation induction, or treatment of hyperandrogenic symptoms such as hirsutism or acne. Metformin use in pregnancy is generally safe and may reduce pregnancy loss and preterm birth, but long‐term offspring impacts remain unclear, suggesting targeted rather than routine use. Meta‐analyses using harmonized individual‐level data from RCTs are an important next step to advancing research in this area, with the potential to uncover critical insights and identify PCOS subgroups who may benefit from metformin use both in and outside of pregnancy.

AUTHOR CONTRIBUTIONS

S.S., T.M., and A.M. contributed equally to researching data, discussion of content, and drafting, reviewing, and editing the manuscript. R.G. made a substantial intellectual contribution to the content, created figures, and reviewed the manuscript. E.V. and H.T. made a substantial intellectual contribution to the content and reviewed the manuscript. All authors contributed to the manuscript in line with ICJME criteria and approved the final version for publication.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interests.

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/dom.16422.

Supporting information

Data S1. Supporting Information.

Saadati S, Mason T, Godini R, Vanky E, Teede H, Mousa A. Metformin use in women with polycystic ovary syndrome (PCOS): Opportunities, benefits, and clinical challenges. Diabetes Obes Metab. 2025;27(Suppl. 3):31‐47. doi: 10.1111/dom.16422

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.