Pancreatic insulin secretion rates can be accurately derived by mathematical deconvolution of peripheral C-peptide concentrations either by using individual C-peptide kinetic parameters obtained by analysis of the decay curve of biosynthetic human C-peptide or by using published group parameters with appropriate adjustments for age and degree of obesity. Since the cross-reactivity of proinsulin and related peptides is low (< 10%) in many C-peptide assays, this experimental approach avoids the spurious increase in insulin immunoreactivity resulting from cross-reactivity with proinsulin and related peptides in the insulin assay. Application of this technique has demonstrated that the phenotypic expression of beta-cell dysfunction differs in subjects with different genetic mechanisms of non-insulin-dependent diabetes mellitus (NIDDM). Subjects who have maturity-onset diabetes of the young (MODY) due to mutations in the glucokinase gene demonstrate different patterns of altered insulin secretion when compared with subjects who have mutations in the MODY1 gene on chromosome 20. Glucokinase mutations affect the ability of the beta-cell to detect and respond to small increases in glucose above the basal level. However, compensatory mechanisms operative in vivo, which include a priming effect of glucose on insulin secretion, limit the severity of the observed insulin secretory defect, resulting in a generally mild clinical course in these subjects. In contrast, mutations in the MODY1 gene are associated with an inability to increase insulin secretion as the plasma glucose concentration increases above 7-8 mmol/l and the normal priming effect of glucose on insulin secretion is lost. These characteristics of the dose-response relationships between glucose and insulin secretion result in a more severe degree of hyperglycemia than observed in subjects with glucokinase mutations, and these subjects more frequently need insulin treatment. These alterations are evident in prediabetic subjects with normal glucose levels who carry the MODY1 mutation, suggesting that defective beta-cell function is the primary pathogenetic defect in the diabetic syndrome in these subjects. Studies performed in the classic form of NIDDM demonstrate that subjects with mild glucose intolerance and normal fasting glucose concentrations and glycosylated hemoglobin levels consistently demonstrate defective beta-cell function. These results are consistent with studies in the Zucker diabetic fatty rat, an animal model of NIDDM in which prediabetic animals demonstrate extensive alterations in expression of multiple genes involved in the regulation of insulin secretion. It thus appears that abnormal beta-cell function is present at a relatively early stage in the evolution of NIDDM, even before the onset of overt hyperglycemia.