Does acupuncture improve insulin sensitivity more effectively than metformin or sham acupuncture in women with polycystic ovary syndrome (PCOS) and insulin resistance (IR)?Among women with PCOS and IR, acupuncture was not more effective than metformin or sham acupuncture in improving insulin sensitivity.Uncontrolled trials have shown that acupuncture improved insulin sensitivity with fewer side effects compared with metformin in women with PCOS and IR. However, data from randomized trials between acupuncture and metformin or sham acupuncture are lacking.This was a three-armed randomized controlled trial enrolling a total of 342 women with PCOS and IR from three hospitals between November 2015 and February 2018, with a 3-month follow-up until October 2018.Women aged from 18 to 40 years with PCOS and homeostasis model assessment of insulin resistance (HOMA-IR) ≥2.14 were randomly assigned (n = 114 per group) to receive true acupuncture plus placebo (true acupuncture), metformin plus sham acupuncture (metformin, 0.5 g three times daily) or sham acupuncture plus placebo (sham acupuncture) for 4 months, with an additional 3-month follow-up. True or sham acupuncture was given three times per week, and 0.5 g metformin or placebo was given three times daily. The primary outcome was change in HOMA-IR from baseline to 4 months after baseline visit. Secondary outcomes included changes in the glucose AUC during an oral glucose tolerance test, BMI and side effects at 4 months after baseline visit.After 4 months of treatment, the changes of HOMA-IR were -0.5 (decreased 14.7%) in the true acupuncture group, -1.0 (decreased 25.0%) in the metformin group and -0.3 (decreased 8.6%) in the sham acupuncture group, when compared with baseline. True acupuncture is not as effective as metformin in improving HOMA-IR at 4 months after baseline visit (difference, 0.6; 95% CI, 0.1-1.1). No significant difference was found in change in HOMA-IR between true and sham acupuncture groups at 4 months after baseline visit (difference, -0.2; 95% CI, -0.7 to 0.3). During the 4 months of treatment, gastrointestinal side effects were more frequent in the metformin group, including diarrhea, nausea, loss of appetite, fatigue, vomiting and stomach discomfort (31.6%, 13.2%, 11.4%, 8.8%, 14.0% and 8.8%, respectively). Bruising was more common in the true acupuncture group (14.9%).This study might have underestimated the sample size in the true acupuncture group with 4 months of treatment to enable detection of statistically significant changes in HOMA-IR with fixed acupuncture (i.e. a non-personalized protocol). Participants who withdrew because of pregnancy did not have further blood tests and this can introduce bias.True acupuncture did not improve insulin sensitivity as effectively as metformin in women with PCOS and IR, but it is better than metformin in improving glucose metabolism (which might reduce the risk of type 2 diabetes) and has less side effects. Metformin had a higher incidence of gastrointestinal adverse effects than acupuncture groups, and thus acupuncture might be a non-pharmacological treatment with low risk for women with PCOS. Further studies are needed to evaluate the effect of acupuncture combined with metformin on insulin sensitivity in these women.This work was supported by grants 2017A020213004 and 2014A020221060 from the Science and Technology Planning Project of Guangdong Province. The authors have no conflicts of interest.Clinicaltrials.gov number: NCT02491333.8 July 2015.11 November 2015.© The Author(s) 2021. Published by Oxford University Press on behalf of European Society of Human Reproduction and Embryology.