Conventional immunoassays to quantify insulin concentration do not differentiate between insulin and proinsulin. Thus, previous conclusions as to the relationship between the development of hyperglycemia in patients with noninsulin-dependent diabetes mellitus (NIDDM) and pancreatic insulin secretory function may have been confounded by not being able to determine the contribution made by plasma proinsulin to the putative measurements of plasma insulin concentration in these patients. The current study was initiated to address this issue by making specific measurements of plasma insulin, proinsulin, and C-peptide concentrations in 42 individuals: 14 with normal glucose tolerance, 12 with impaired glucose tolerance (IGT), and 16 with NIDDM. The study population was further subdivided into a nonobese (body mass index, < 30 kg/m2) and an obese (body mass index, > 30 kg/m2) group. Mixed meals were given at 0800, 1200, and 1800 h, and blood was removed at 0800 h (before the meal) and at hourly intervals from then until 1600 h. Plasma glucose concentrations throughout the sampling period were slightly, but significantly (P < 0.01), greater in patients with IGT than in normal individuals. Patients with NIDDM had markedly elevated glycemic excursions, greater than either of the other two groups (P < 0.002). Both plasma immunoreactive insulin and C-peptide concentrations from 0800-1600 h were higher (P < 0.002-0.001) in patients with either IGT or NIDDM than in the group with normal glucose tolerance. Although day-long plasma immunoreactive insulin and C-peptide concentrations were higher, on the average, in patients with IGT compared to those with NIDDM, the difference was not statistically significant. Plasma proinsulin concentrations were highest in patients with NIDDM (P < 0.002), lower in those with normal glucose tolerance (P < 0.002), and intermediate in patients with IGT. When the calculated "true" insulin concentration was determined by taking the proinsulin content into consideration, patients with IGT had the highest day-long levels, with the lowest values found in the control population (P < 0.002). Although absolute values varied as a function of obesity, the generalizations outlined above were found in both weight groups. These results show that ambient plasma proinsulin concentrations increase as glucose tolerance declines. However, true plasma insulin concentrations in response to mixed meals remain highest in patients with IGT, lowest in normal individuals, and intermediate in patients with NIDDM. Thus, previous conclusions that absolute day-long plasma insulin concentrations are not lower than normal in patients with NIDDM do not appear to result from an inability to differentiate true insulin from proinsulin.