The production of insulin from proinsulin involves cleavage of intact proinsulin into proinsulin conversion intermediates by the processing of enzymes PC2 and PC3 before fully processed insulin is produced. Intact proinsulin and these conversion intermediates are measured in many immunoreactive insulin (IRI) assays, and therefore contribute to the absolute IRI measurement. The proportion of basal IRI made up of proinsulin (PI)-like molecules (PI/IRI) is increased in NIDDM. Whether stimulated IRI levels are similarly made up of disproportionately increased PI/IRI or whether the relative proportions of proinsulin and its conversion intermediates are altered has not been evaluated. An index of the efficiency of proinsulin processing within the pancreatic beta-cell can be achieved by measuring PI/IRI immediately following acute stimulation of beta-cell secretion, and then determining the proportion of intact proinsulin and proinsulin conversion intermediates contributing to circulating proinsulin-like molecules. In this study, we determined the PI/IRI levels under basal and arginine-stimulated conditions in 17 healthy and 16 NIDDM subjects; high-performance liquid chromatography (HPLC) was also performed in a subset of these subjects to measure the relative contribution of intact proinsulin and its conversion intermediates to total proinsulin-like molecules. In NIDDM subjects, levels of both basal (44.6 +/- 9.6 vs. 9.3 +/- 1.5 pmol/l; P = 0.0007) and stimulated (64.0 +/- 12.7 vs. 19.8 +/- 2.8 pmol/l; P = 0.001) proinsulin-like molecules were higher than in healthy subjects. Although IRI was higher in NIDDM than in control subjects under basal conditions (106 +/- 19 vs. 65.1 +/- 8.1 pmol/l; P = 0.05), it was lower in NIDDM than in control subjects following stimulation (increment: 257 +/- 46 vs. 416 +/- 51 pmol/l; P = 0.03). PI/IRI ratios were increased in NIDDM subjects under both basal (43.3 +/- 5.0 vs. 14.0 +/- 1.3%; P < 0.0001) and stimulated (increment: 10.1 +/- 2.1 vs. 2.5 +/- 0.2%; P = 0.0006) conditions, compatible with the release of a disproportionately increased amount of proinsulin-like products. HPLC analysis revealed that, in the stimulated state, intact proinsulin made up 40.1 +/- 6.7% of proinsulin-like molecules in NIDDM individuals (n = 9) and 30.1 +/- 5.6% in healthy subjects (n = 7; NS). The remainder of the proinsulin-like molecules comprised the des-31,32-split proinsulin conversion intermediate. The increase in PI/IRI in NIDDM under basal and especially under stimulated conditions suggests that proinsulin conversion is indeed perturbed in this disorder. Because the relative proportions of intact and des-31,32-split proinsulin are similar in both healthy and NIDDM subjects, the orderly cleavage of proinsulin at its two junctions appears preserved. However, at the time of exocytosis, the secretory granule in the islet of NIDDM subjects contains an increased proportion of incompletely processed proinsulin, presumably reflecting a slower rate of conversion or granules' reduced time of residence in beta-cells.