Introduction and Objective: Detecting dysglycemia at early stages of T1D can prevent DKA and improve long-term outcomes. OGTT, currently recommended for monitoring progression, can be challenging in clinical care. HbA1c is not sufficiently sensitive, particularly in young children. We aimed to identify CGM metrics predicting progression to stage 3 T1D. Methods: 171 participants prospectively followed in the ASK and DAISY studies with CGM, OGTT and HbA1c were included. Bayesian joint models for longitudinal and survival outcomes tested associations between longitudinal trajectories of CGM metrics (mean glucose, SD glucose, CV glucose, TAR140) and HbA1c with time to progression to stage 3. Results: Participants were 2-45 yrs old, 43% male, 49% progressed to stage 3. All CGM metrics and HbA1c were strongly associated with time to progression (p0.001). HbA1c had best model fit (HR 3.58 95% CI 2.39-5.65 per SD) while CV glucose had greatest effect size (HR 4.75 95% CI 2.72-10.70 per SD). In a combined model, HbA1c (p=0.005) and CV glucose (p0.001) were associated with time to progression, while the HR for CV glucose remained higher than HbA1c (3.31 1.72-8.11 vs 2.04 1.24-3.41). AUC was 0.67 for HbA1c alone, and 0.84 for combined HbA1c and CV glucose. The model allows individualized risk prediction over customized time intervals (Figure). Conclusion: CGM variability metrics independently predict progression to stage 3 T1D and should be integrated into standard monitoring protocols. Disclosure T. Vigers: None. L. Pyle: None. F. Dong: None. T. Fleury: None. M. Rewers: Consultant; Current; Sanofi. Advisory Panel; Current; Vertex Pharmaceuticals Incorporated. B. Frohnert: None. A. Steck: Consultant; Current; Sanofi. Funding Breakthrough T1D (SRA-2024-1603-SB, SRA-2024-1590-MB, 3-SRA-2024-1590-M-B), NIH (U01 DK106993, 5R01DK032493)
Vigers et al. (Sat,) studied this question.