We appreciate Dr. Gumpel's historical perspective on the preservation of insulin secretion in type 1 diabetes and Dr. Charles-Marcel's suggestion of another potential mechanism to explain our findings. With regard to the historical perspective, J.M. Brush reported the uncontrolled results of a therapeutic regimen that grew out of an "effort to simplify the management of diabetic children" [1]. Comparison of his observations with the controlled experiment of the DCCT reveals the enormous strides made in clinical research, in general, and diabetes clinical research and care, in particular, in the past 50 years. Brush reported experience with a 1-month inpatient regimen in which children with new-onset diabetes were treated with relatively large doses of insulin, administered as regular insulin divided into four doses (including one at 4 a.m.). Total insulin dose was adjusted to render the urine glucose free-routine blood glucose testing was not easily available-and was continued until hypoglycemic "shock" occurred. Insulin doses were subsequently decreased, and patients left the hospital receiving relatively small doses of insulin, having achieved a new "steady state." Brush ascribed the decreased insulin requirements to recovery of pancreatic function, although he recognized that he could not prove his "teleologic imagery."

Brush's work was hampered by the absence of acceptable controls-the first controlled clinical trial in the modern era didn't occur until 1948 [2]-and by the lack of readily available quantitative measures of glycemia, or any direct assay of insulin (which was not developed until 1960 [3]). In contrast to Brush's empirical methods, the DCCT performed a carefully controlled experiment and had the advantage of using blood glucose, glycosylated hemoglobin, and C-peptide assays. Although the stage during which the interventions were applied (all DCCT participants had had diabetes for at least 1 year) and the intensive treatment regimens differed, ß-cell function was preserved in Brush's experience and in the DCCT. That the DCCT was able to prove Dr. Brush's theory, with rigorously applied methods, is a testament to his prescience and to the power of modern experimental methods.

Dr. Charles-Marcel's suggestion that restoration of amylin secretion concurrent with insulin secretion may have contributed to the observed metabolic results is intriguing. The potentially opposing effects of amylin on glucose metabolism-inhibiting insulin secretion [4] compared with slowing gastric emptying and inhibiting glucagon secretion-would need to be weighed carefully to determine, on balance, whether restoration of amylin secretion would benefit or worsen glycemia in the setting of type 1 diabetes with residual insulin secretion.