Correspondence (reply): In Reply

Our correspondents’ comments relate to individual therapeutic variants. It is worth re-stating that diabetology has a substantial problem with evidence, or rather, the lack thereof, because the effectiveness of blood glucose lowering vis-à-vis mortality and complications has been confirmed to a very limited extent in appropriate randomized prospective studies (1). The threshold above which raised blood glucose concentrations will trigger the complications that are typical for diabetes is unclear, and this is also the case for the question regarding which individual factors may pose a particular risk. By contrast, high-level evidence supports the fact that therapy induced hypoglycemia constitutes a relevant risk factor (1). This is reflected in the “relaxed” recommendations of the current guidelines, as we showed in our article. Metformin is recommended by all guidelines, although the evidence—from a small group treated with metformin in the UKPDS Study (3)—is weak (2). It was also in the UKPDS Study that mortality was significantly raised for the combination of sulfonylureas with metformin (3); the national disease management guideline explicitly makes reference to this finding.

With regard to sulfonylureas, our colleagues from DEGAM/AkdÄ focus only on the evidence that supports their argument. The risk of hypoglycemia and weight gain affects patients to very different degrees under treatment with sulfonylureas. The DPP-4 and SGLT-2 inhibitors are clearly superior in this respect, which is relevant for some patients and therefore has to be explained in a CME article.

In the Look AHEAD Study, lifestyle did not affect cardiovascular end points, as Meyer reminds us. But the control group had displayed extremely healthy behaviors and lost 4.5 kg, compared with 6 kg in the intervention group—possibly thanks to the comprehensive press coverage—so that the weight difference was only 1.5 kg. Physical activity was evidently increased only slightly—something that needs to be considered in clinical practice. However, one would hardly advise against healthy behavior—even if this might seem justified on the basis of evidence criteria alone. By contrast, dietary factors—the Mediterranean diet, to be more specific—were successful in significantly lowering cardiovascular mortality; this also applied to the 3700 diabetes patients in this randomized controlled study (4). According to subgroup analyses of the ADVANCE and ONTARGET studies (5, 6), moderate consumption of alcohol, defined as up to 36 g/day, in diabetes patients resulted in a significant reduction in mortality and microvascular/macrovascular complications; however, neuropathy was insufficiently evaluated. With our colleague Roth, we are looking forward to study results that are imminently expected from Israel.