Although Drs. Feinglos and Hollis and Dr. Epstein list several reservations about our data, they do not provide any firm evidence for the conclusion that increasing the sulfonylurea dose above 10 mg is associated with improved glycemic control. Such studies are lacking. The studies cited by Drs. Feinglos and Hollis highlight the problem. One study [1] represented an uncontrolled multicenter study that compared glipizide with several other agents, including phenformin, and the other study [2] compared two sulfonylureas that were started at different doses. No study tried to reduce the dose to test whether the dose increase was necessary. For example, in the second study [2], the glipizide dose was increased from a mean dose of 10.9 mg to 29.7 mg without any significant improvement in glucose control. The inverse correlation between glipizide dose and mean blood glucose level also suggests that the cost of a mean reduction in blood glucose level of 10 mg/dL is 30 mg of drug.

The weekdays on which blood glucose was measured at home were fixed for individual patients but could vary among patients. We have successfully used this approach in many studies over the years [3, 4]. Some patients with low baseline HbA_1c concentrations did experience mild hypoglycemia while receiving maximal doses but also when receiving lower doses.

The above findings are not unique to glipizide. Glyburide also shows lack of an increased effect at concentrations over 200 nmol/L, equivalent to doses of less than 10 mg [5]. We agree that the patient group was heterogeneous and that some were more hyperglycemic than others, although they had normal or high C-peptide concentrations. This, however, represents the target group of patients in whom an increase in the sulfonylurea dose is considered. Some persons (especially obese persons) may still benefit from higher doses of sulfonylureas, although the response to treatment could not be predicted. Given the lack of scientific evidence for an improved effect of glipizide or glyburide doses above 10 mg, it may be justified to reconsider 15 mg as a maximum dose, which should be exceeded in only rare cases.

We think that sulfonylurea treatment should be started earlier, when the drugs are still effective, but it should also be stopped earlier, when the drugs can no longer achieve the treatment goals. Postponing the decision to start insulin treatment by further increasing the sulfonylurea dose does not benefit the patient.