Associations of CGM Metrics With Stimulated C-Peptide Measures in Youth With Recent-Onset Type 1 Diabetes

OBJECTIVE Continuous glucose monitoring (CGM) measures could be a surrogate for stimulated C-peptide outcomes in type 1 diabetes trials. RESEARCH DESIGN AND METHODS CGM and mixed-meal tolerance test–derived C-peptide measures at time points out to 52 weeks after diagnosis were compared in 103 children. RESULTS At 52 weeks, CGM metrics moderately correlated with C-peptide area under the curve. The highest Spearman correlations were for time-in-range 70–180 mg/dL, time <70 mg/dL, and glucose coefficient of variation (0.45,-0.33, and-0.58, respectively; the multivariate model using these three metrics had a slightly higher correlation of 0.63). For predicting peak C-peptide concentrations ≥0.2 pmol/mL, this combination had a sensitivity of 68.4% and specificity of 75%. CONCLUSIONS CGM measures correlated with stimulated C-peptide measures; however, the strength of the correlations and sensitivity and specificity of CGM-derived measures were not great enough to replace C-peptide measures in clinical trials.