Background
Therapies that preserve beta-cell function in type 1 diabetes (T1D) are being evaluated. Beta-cell function is most commonly assessed as two-hour meal-stimulated plasma C-peptide (2hAUC). To simplify this assessment we sought to devise a surrogate marker using measures obtained at a single time point.
Method
Linear models to interpolate 2hAUC from fasting biochemical and clinical measures from people aged less than 21 years with recently diagnosed T1D were developed using placebo-arm data from three TrialNet studies and tested in other trial cohorts for their ability to detect significant loss of beta-cell function (>7.5% of baseline 2hAUC) and effects of immune therapy.
Results
A model based on fasting plasma C-peptide alone (2hAUC = 44.08 + 178.38 x fasting plasma C-peptide in ng/ml) accurately diagnosed loss of beta-cell function (area under ROC 0.83, 95% CI 0.77 to 0.90). Its accuracy was superior to insulin dose-adjusted HbA1c (IDAA1C) and a model based on HbA1c, insulin dose, disease duration and body mass index (BMI), and was comparable to the accuracies of models that included these variables together with fasting C-peptide. When the model was applied to trials that showed treatment benefit, the differences between the placebo and active arms were qualitatively similar to those obtained using 2hAUC.
Conclusion
2hAUC interpolated from fasting plasma C-peptide identifies children and young adults with recently diagnosed T1D who lose significant beta-cell function within the first year. Interpolated 2hAUC is more accurate than IDAA1C and could serve as an outcome measure in future trials of disease-modifying therapy in T1D.