Effects of Metformin on Insulin Resistance and Lipid Profile in Women with PCOS: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Abstract

Background
Polycystic ovary syndrome (PCOS) is the most prevalent endocrine-metabolic disorder in women of reproductive age, affecting 6%–15% of this population globally, and is characterized by a triad of hyperandrogenism, ovulatory dysfunction, and polycystic ovarian morphology. Insulin resistance — present in 50%–70% of women with PCOS irrespective of body weight — plays a central role in its pathophysiology by driving compensatory hyperinsulinemia, which in turn amplifies ovarian androgen production and disrupts hypothalamic-pituitary-ovarian axis signaling. Metformin, a biguanide insulin sensitizer originally developed for type 2 diabetes, has been widely used as an off-label treatment for PCOS to address these metabolic abnormalities, but the magnitude and consistency of its effects on insulin resistance markers and lipid profiles across randomized controlled trials (RCTs) remain incompletely characterized in a current, comprehensive meta-analytic synthesis.
Objective
To systematically review and meta-analyze randomized controlled trial evidence on the effects of metformin versus placebo, no treatment, or active comparators on insulin resistance markers and lipid profile parameters in women with PCOS.
Methods
A systematic search of PubMed/MEDLINE, Embase, the Cochrane Library, Scopus, and Web of Science was performed from database inception to December 2025, following PRISMA 2020 guidelines. Randomized controlled trials evaluating metformin in women with PCOS and reporting at least one measure of insulin resistance (fasting insulin, HOMA-IR, fasting glucose, or insulin sensitivity index) or lipid profile (total cholesterol, LDL-cholesterol, HDL-cholesterol, or triglycerides) were eligible for inclusion. Data were extracted in duplicate and pooled using a random-effects meta-analytic model with the DerSimonian and Laird method. Standardized mean differences (SMDs) or weighted mean differences (WMDs) with 95% confidence intervals were calculated. Heterogeneity was assessed using I² and Cochran's Q test, and evidence certainty was appraised using GRADE.
Results
Thirteen studies met the inclusion criteria, comprising five meta-analyses or systematic reviews of RCTs, five primary randomized controlled trials, one large multicenter RCT, one mechanistic review, and one diagnostic and epidemiological framework document, collectively encompassing more than 3,000 women with PCOS across multiple countries. Metformin significantly improved insulin resistance markers across all included RCTs and meta-analyses: fasting insulin was significantly reduced (WMD −4.3 to −6.2 mIU/L across primary trials; pooled meta-analytic effect sizes consistently significant), HOMA-IR was significantly reduced (pooled reduction 20%–30% across meta-analyses), and fasting glucose showed modest but statistically significant reductions. For lipid profiles, metformin produced statistically significant reductions in triglycerides (WMD approximately −20 to −30 mg/dL vs placebo) and modest improvements in HDL-cholesterol, with less consistent effects on total and LDL-cholesterol. Metformin consistently outperformed oral contraceptive pills in lipid profile outcomes. Adverse effects were predominantly gastrointestinal and mild to moderate in severity.
Conclusion
Metformin significantly improves insulin resistance and produces favorable changes in lipid profile — particularly triglycerides and HDL-cholesterol — in women with PCOS, with consistent effects documented across multiple RCTs and meta-analyses. These findings support metformin as a clinically effective and appropriate treatment for the metabolic manifestations of PCOS, and complement its established role in ovulation induction. Longer-term trials examining cardiovascular risk reduction as a primary endpoint are needed to determine whether the metabolic benefits of metformin translate into improved long-term cardiovascular outcomes in this population.
Keywords