Background: We investigated whether metformin prevents tamoxifen-induced endometrial changes and insulin resistance (IR) after a diagnosis of breast cancer. Methods: This was a single-centre, randomized, double-blind, placebo-controlled, parallel group trial. Postmenopausal women with hormone receptor-positive breast cancer taking tamoxifen were randomly allocated to metformin 850 mg or identical placebo, twice daily, for 52 weeks. Outcome measures included double endometrial thickness (ET) measured by transvaginal ultrasound, fasting insulin, glucose and IR estimated by the homeostasis model of assessment (HOMA-IR). Results: A total of 112 women were screened and 102 randomized. Results are presented as median (range). The 101 women who took at least one dose of medication were aged 56 (43-72) years, with 5(0.5-28) years postmenopause, and had taken tamoxifen for 28.9 (0-367.4) weeks. The baseline ET was 2.9 mm (1.4-21.9) for the placebo group (n = 52) and 2.5 mm (1.3-14.8) for the metformin group (n = 50). At 52 weeks, the median ET was statistically significantly lower for the metformin (n = 36) than for the placebo group (n = 45) (2.3 mm (1.4-7.8) vs 3.0 (1.2-11.3); P = 0.05). 13.3% allocated to placebo had an ET greater than 4 mm vs 5.7% for metformin (P = 0.26). There was no endometrial atypia or cancer. Compared with placebo, metformin resulted in significantly greater baseline-adjusted reductions in weight (P < 0.001), waist circumference (0.03) and HOMA-IR (P < 0.001). Conclusions: Metformin appears to inhibit tamoxifen-induced endometrial changes and has favourable metabolic effects. Further research into the adjuvant use of metformin after breast cancer and to prevent EH and cancer is warranted.