Response to Comment on: Harrison et al. β-cell function preservation after 3.5 years of intensive diabetes therapy. Diabetes Care 2012;35:1406-1412.

We appreciate the critical review of our article by Aizawa and Yamauchi (1), and we are pleased they found the study well designed and straightforward (2). Their critiques relate to 1) exclusion of reporting insulin sensitivity and 2) lack of a significant drug wash-out period prior to β-cell assessment. We did not perform a direct measure of insulin sensitivity using a hyperinsulinemic-euglyceimc clamp as this did not represent one of our main outcomes and would have added a significant technical and financial burden to the study. We attempted to estimate insulin sensitivity using indirect methods such as homeostasis model assessment of insulin resistance and the Matsuda index, but we found these results uninterpretable because of the presence of insulin antibodies in a large percentage of our insulin-treated patients. Although we could not calculate a “disposition index,” we did adjust the absolute C-peptide response by the postload glycemia. A significant differential treatment effect on insulin sensitivity would have led to either a compensatory change in C-peptide response or a change in postload glycemia. Since neither occurred, we can indirectly assume there were no significant changes in insulin sensitivity over time.

As correctly noted by Aizawa and Yamauchi, the first evaluation of β-cell function was carried out after the 3-month run-in treatment period with insulin and metformin. The purpose of this design was to capture true baseline β-cell function after recovery of any “stunned” β-cells from the acute effects of gluco- and lipotoxicity and to allow both study groups to attain similar glycemic control prior to randomization. All β-cell function testing over the course of the 42-month study was done under the exact same experimental conditions, including the 24-h period withdrawal of hypoglycemic agents. Our study was designed to evaluate whether early insulin treatment followed by either insulin or oral hypoglycemic agents could slow down the inevitable decline in β-cell function, which is well documented to occur in type 2 diabetes. The expectation of washing out all therapies prior to testing is appropriate for diabetes prevention studies, or in more general terms for claims of “disease cure.” Treatment for diabetes is lifelong, and repeated drug withdrawal for the purpose of testing would expose patients to undue risk and is neither clinically recommended nor necessary in our view. Previous studies (A Diabetes Outcome Progression Trial, U.K. Prospective Diabetes Study [3,4]) have clearly documented the progressive decline in β-cell function over time even in the presence of therapy, including sulfonylureas, metformin, and thiazolidinediones.

Furthermore, a close inverse relationship between β-cell function and HbA1c exists, and our study showed that β-cell function can be stabilized for at least 3.5 years as measured by both direct β-cell function testing as well as indirectly by maintaining excellent glycemic control throughout the follow-up period.