Abstract
Introduction and Objective: The TrialNet Abatacept Prevention study of 12 months use of the CTLA-4 immunoglobulin to delay disease progression in Stage 1 type 1 diabetes (T1D) did not meet its primary endpoint. We adopted the oral minimal model (OMM) to quantify insulin secretion and assess if high vs low baseline secretion determined treatment response. Methods: Beta cell responsivity to glucose (Phitotal ) was computed from oral glucose tolerance test at baseline and every 6 months up to 48 months in placebo and abatacept arms. Baseline secretor groups were defined as high-secretors (Phitotal >33rd centile) and low-secretors (Phitotal ≤33rd centile). Risk regression analyses compared time to Stage 2 or 3 T1D between arms, by baseline secretor group. Age-adjusted ANCOVA compared Phitotal at pre-specified discretized visits. Results: In high-secretors, abatacept delayed progression by ~12 months (48 (IQR 26, 48) vs 35 (IQR 17, 48), p=0.009) with 51% lower hazard for progression among abatacept high secretors (HR 0.51, (95% CI 0.29, 0.89), p=0.018). High-secretors on abatacept maintained Phitotal after treatment suspension for up to 48 months, unlike placebo high-secretors (0.3% (95% CI -3.8, 4.5) vs -9% (95% CI -13.2, -4.8), p=0.002). Conclusion: The OMM reveals heterogeneity within Stage 1 T1D and identifies abatacept responders with high baseline beta cell function. This is consistent with the observation in murine models that abatacept is most effective in early disease.