Associations of proinsulin-to-insulin ratios with incident type 2 diabetes have been inconsistent. The use of C-peptide as the denominator in the ratio may allow for better prediction because C-peptide concentration is not affected by hepatic insulin clearance. The objective of this paper was to compare fasting intact and split proinsulin-to-insulin ratios (PI/I, SPI/I) with intact and split proinsulin-to-C-peptide ratios (PI/C-pep, SPI/C-pep) in the prediction of type 2 diabetes.

Prospective data on 818 multi-ethnic adults without diabetes at baseline from the Insulin Resistance Atherosclerosis Study (IRAS) were used. Insulin sensitivity (S_I) and acute insulin response (AIR) were determined from frequently sampled intravenous glucose tolerance tests, and fasting intact and split proinsulin were measured using specific two-site monoclonal antibody-based immunoradiometric assays. Associations of proinsulin ratios with type 2 diabetes were determined using logistic regression and differences in prediction were assessed by comparing areas under the receiver operating characteristic curve (AROCs).

In logistic regression analyses, PI/C-pep and SPI/C-pep were more strongly associated with incident type 2 diabetes (n = 128) than PI/I and SPI/I, and were significantly better predictors of diabetes in AROC analyses (PI/C-pep = 0.662 vs PI/I = 0.603, p = 0.02; SPI/C-pep = 0.690 vs SPI/I = 0.631, p = 0.01). Both PI/C-pep and SPI/C-pep were associated with type 2 diabetes after adjustment for age, sex, ethnicity, waist circumference, impaired glucose tolerance, lipids and S_I. Both PI/C-pep and SPI/C-pep were significantly associated with incident type 2 diabetes in models that included AIR.

Proinsulin-to-C-peptide ratios were stronger predictors of diabetes in comparison with proinsulin-to-insulin ratios. These findings support the use of C-peptide as the denominator for proinsulin ratios, to more accurately reflect the degree of disproportional hyperproinsulinaemia.