Abstract
OBJECTIVE: Identification of prognostic biomarkers in pediatric diabetes is important for precision medicine. We assessed whether C-peptide and islet autoantibodies are useful to predict the natural history of children with new-onset diabetes. METHODS: We prospectively studied 72 children with new-onset diabetes (median follow-up: 8 months) by applying the Aβ classification system ("A+": islet autoantibody positive, "β+": random serum C-peptide ≥1.3 ng/mL at diagnosis). Beta-cell function was assessed longitudinally with 2 hours postprandial/stimulated urinary C-peptide-to-creatinine ratio (UCPCR) 3-12 weeks (V1) and 6 to 12 months after diagnosis (V2). We obtained a type 1 diabetes genetic risk score for each participant, and compared characteristics at baseline, and clinical outcomes at V2. RESULTS: The cohort was 50% male. Racial distribution was 76.4% White, 20.8% Black, and 2.8% Asian or other races. A total of 46.5% participants were Hispanic. Median age (Q1-Q3) was 12.4 (8.3-14.5) years. The Aβ subgroup frequencies were 46 A+β-(63.9%), 1 A-β-(1.4%), 4 A+β+(5.6%), and 21 A-β+(29.2%). Baseline serum C-peptide correlated with UCPCR at both V1 (r = 0.36, P = .002) and V2 (r = 0.47, P < .001). There were significant subgroup differences in age, race, frequency of diabetic ketoacidosis, and type 1 diabetes genetic risk score (P < .01). At V2, the 2 β-subgroups had lower UCPCR and higher hemoglobin A1c compared with the 2 β+ subgroups (P < .001 and P = .02, respectively). Thirty-eight percent of A-β+ but none of the other subgroups were insulin-independent at V2 (P < .001). CONCLUSION: C-peptide and islet autoimmunity at diagnosis define distinct phenotypes and predict beta-cell function and insulin dependence 6 to 12 months later in racially/ethnically diverse children with new-onset diabetes.