Response to Hinke et al.

We agree with Hinke et al. (1) that metformin acts mainly through the inhibition of hepatic glucose output and the enhancement of peripheral insulin sensitivity, through unidentified molecular mechanisms in the liver and skeletal muscle. However, it has been observed that metformin could enhance glucose-induced insulin secretion in some experimental conditions (2), although the relevance of this effect for the antihyperglycemic action of metformin is questionable. Furthermore, glucagon-like peptide (GLP)-1 has been shown to increase insulin sensitivity and non–insulin-mediated glucose disposal (3,4), suggesting that DPP-IV inhibitors, which increase GLP-1 levels, could be expected to improve insulin sensitivity as well as insulin secretion.

Our study (5) has shown that the increase of GLP-1 levels after an oral glucose load determined by metformin, consistent with previous reports, is not due to drug-induced differences in glycemia or insulinemia; in fact, this effect can also be observed in isoglycemic and isoinsulinemic conditions, i.e., during a hyperinsulinemic-euglycemic clamp. The contribution of enhancement of secretion and inhibition of degradation to the increase of GLP-1 levels during metformin therapy needs to be elucidated through further specifically designed studies, as was clearly stated in our study. The measurement of total GLP-1, as suggested by Hinke et al., would be of little use in this respect; in fact, total GLP-1 should obviously be expected to be increased, even in the case of metformin inhibiting degradation without stimulating secretion. In vitro or ex vivo experimental models, such as isolated intestinal L-cells or perfused ileum, would be more informative for the study of the effects of metformin on GLP-1 secretion.

We also agree with Hinke et al. that, theoretically, the combination of DPP-IV inhibitors (acting mainly via the increase of early postprandial insulin secretion) and metformin (acting mainly through the enhancement of insulin sensitivity and suppression of hepatic glucose output) could be useful in the treatment of type 2 diabetes. However, the choice of therapeutic combinations should be based on evidence derived from clinical studies rather than on theoretical consideration. Demuth et al. (6) reported that the Probiodrug DPP-IV inhibitor P32/98 has a significant hypoglycemic effect in type 2 diabetic patients treated with sulfonylureas, but it does not reduce blood glucose in those already treated with metformin. We agree with Hinke et al. that other DPP-IV inhibitors could have a more favorable profile of action when given in combination with metformin, but we advise greater caution in designing future therapeutic scenarios when so little sound clinical evidence is available.