Vazquez-Martin et al have raised important issues relating to the potential activity of metformin in cancer clinical trials. However, we note that the metformin concentrations used in their study far exceeded the therapeutic concentrations for this drug. The therapeutic plasma levels of metformin in clinical medicine are 0.465 to 2.5 mg/L, whereas Vazquez-Martin et al evaluated cells selected for 4 months in 1,650 mg/mL (10 mmol/L) metformin, a minimum 660-fold excess over the recommended therapeutic levels. Use of metformin at such high doses likely affects a number of molecular targets not affected at the therapeutic dose, raising the possibility that the reported insulinlike growth factor 1 receptor (IGF-1R) hyperphosphorylation may not relate to the primary therapeutic effect of the drug. Considering that amplification or overexpression of IGF-1R elicits a robust antiapoptotic response against a broad range of cell death–inducing stimuli, its activation after selection with 10 mmol/L metformin may reflect adaptation to general metformin toxicity. To probe whether or not prolonged metformin exposure at therapeutic levels can lead to IGF-1R activation, we suggest systematic evaluation in various cell types obtained from diabetics taking this drug. Until studies of this type are conducted, we are concerned that suggestions that anti– IGF-1R modalities may be synergistic with the antitumor activity of metformin are premature. Nonetheless, we agree that it is important to investigate the mechanism of potential metformin resistance in human cancer. We, and others, have ongoing neoadjuvant studies of metformin in which the potential mechanism of resistance highlighted by Vazquez-Martin et al will be explored. This mechanism relates exclusively to the direct (insulin-independent) effects of metformin on cancer growth. Clinical and epidemiologic evidence, as outlined in our editorial, provides strong support for an indirect, insulin-mediated mode of action that may represent an equally important mechanism by which metformin influences clinical breast cancer. The report by Jiralerspong et al provides evidence that the use of metformin in standard clinical doses may be associated with clinical benefit irrespective of the resistance pathway that has been highlighted by Vazquez-Martin et al Because of this clinical and epidemiologic evidence, we believe that planned adjuvant trials of metformin in breast cancer should proceed. The results of ongoing neoadjuvant studies will provide additional information on the relative importance of direct and indirect effects of metformin, and on potential mechanisms of resistance at clinically relevant exposure levels.