A study of 200 non-diabetic women with breast cancer analysed the effect of the anti-diabetic drug metformin or a placebo given ahead of surgical resection. There were no significant differences overall in Ki-67 staining (a surrogate for levels of proliferation) in the resected tumours. However, in women with insulin resistance (but not clinical diabetes), or a high body mass index, metformin reduced levels of Ki-67 staining and reduced systemic concentrations of insulin (a known mitogen). Thus, the potential effect of metformin on breast cancer may depend on the metabolic status of the patients, and this should be considered in clinical trial design.