Hypoglycemia risk is probably the most important limiting factor when attempting to treat to target diabetic subjects. Therefore, one needs always to consider how much a given treatment is likely to induce iatrogenic hypoglycemia when choosing a therapeutic strategy for type 2 diabetes. At present, a surprisingly scant amount of data is available about how frequent hypoglycemia is relative to the use of a particular drug. Furthermore, these data are not easy to interpret, having been collected in different ways and being often related to different ways of defining and detecting hypoglycemia. The available literature does suggest, however, that metformin and thiazolidinedione treatments are associated with a negligible hypoglycemia risk. On the other hand, sulphonylurea use (in particular glybenclamide use) is associated with a frequency of hypoglycemia far greater than commonly thought. Newer therapies (like incretin-based therapies) are instead associated with a very low hypoglycemia risk. With these agents, it appears that a significant hypoglycemia risk is detected only when they are used as add-on therapy to insulin or sulphonylureas.