For debate: Starling's curve of the pancreas--overuse of a concept?

It is commonly accepted that insulin secretion follows the pattern of an inverted U, also termed 'Starling's curve of the pancreas' during the natural history of hyperglycemia in glucose intolerance and type 2 diabetes. This concept is based on the cross-sectional observation that insulin concentrations initially increase when insulin sensitivity declines (as a consequence of obesity, for example) and decrease when glucose tolerance deteriorates (impaired glucose tolerance or overt type 2 diabetes). The initial increase in insulin concentrations has been viewed as 'hypersecretion' of insulin, thought to indicate that beta cell dysfunction is not etiological but secondary in nature. However, this view is oblivious to the now well-established fact that assessment of insulin secretion must account for individual insulin sensitivity. Here, we revisit the concept of Starling's curve of the pancreas based on first-phase C-peptide concentrations (hyperglycemic clamp) from subjects with normal glucose tolerance (n=66), impaired glucose tolerance (n=19) and mild type 2 diabetes (n=9). In absolute terms, first-phase C-peptide concentrations plotted against increasing fasting glucose concentrations indeed followed an inverted U. However, adjusted for direct and indirect measures of insulin sensitivity (insulin sensitivity index from the hyperglycemic clamp, body mass index, age and sex), first-phase C-peptide concentrations of the same individuals tended to decrease steadily. In conclusion, while the Starling curve exists for insulin concentrations, and perhaps also for insulin secretion, it does not hold for beta-cell function if that term were to imply appropriateness of insulin secretion (based on a formal test of glucose-stimulated insulin secretion) for the degree of insulin resistance, as it should.