Impact of C-peptide preservation on metabolic and clinical outcomes in the Diabetes Control and Complications Trial

Abstract

The Diabetes Control and Complications Trial established that a stimulated C-peptide concentration ≥0.2 nmol/L at study entry among subjects with up to a 5-year diabetes duration is associated with favorable metabolic and clinical outcomes over the subsequent 7 years of follow-up. Herein we further examine the association of both fasting and stimulated C-peptide numerical values with outcomes. In the intensive treatment group, for a 50% higher stimulated C-peptide on entry, such as from 0.10 to 0.15 nmol/L, HbA1c decreased by 0.07% (0.8 mmol/mol; P = 0.0003), insulin dose decreased by 0.0276 units/kg/day (P < 0.0001), hypoglycemia risk decreased by 8.2% (P < 0.0001), and the risk of sustained retinopathy was reduced by 25% (P = 0.0010), all in unadjusted analyses. Other than HbA1c, these effects remained significant after adjusting for the HbA1c on entry. While C-peptide was not significantly associated with the incidence of nephropathy, it was strongly associated with the albumin excretion rate. The fasting C-peptide had weaker associations with outcomes. As C-peptide decreased to nonmeasurable concentrations, the outcomes changed in a nearly linear manner, with no threshold or breakpoint. While preservation of stimulated C-peptide at ≥0.2 nmol/L has clinically beneficial outcomes, so also does an increase in the concentration of C-peptide across the range of values.

Figure 1

Model-free LOESS estimates of the association of the log of the stimulated C-peptide on study entry with (A) the HbA_1c value at year 1 of follow-up, (B) the total insulin dose at year 1, (C) the incidence of retinopathy progression, and (D) the incidence of hypoglycemia. For C and D, the log of the rate per 100 patient-years of the event is plotted within percentiles of the C-peptide distribution.

Figure 2

Mean HbA_1c over follow-up with 95% confidence limits separately for stimulated C-peptide responders vs. nonresponders in the (A) intensive and (B) conventional treatment groups.

Figure 3

Association of stimulated C-peptide at study entry, qualitatively and quantitatively, with the total insulin dose (units per kilogram per day) at each successive year of follow-up, from longitudinal regression models. A: Mean insulin dose among responder vs. nonresponder. B: Change in insulin dose units per kilogram per day per 50% lower stimulated C-peptide, each with 95% confidence limits.