Changes in A1C Versus GMI Across Glycemic Categories in Clinical Trials of Type 1 Diabetes

Context The glucose management indicator (GMI) is an estimated glycated hemoglobin (A1C) derived from sensor glucose. Though it is being used to approximate A1C in clinical trials, there are no data on direction and magnitude of change in GMI vs A1C after an intervention. Objective To evaluate the magnitude and direction of changes in A1C compared to GMI across different baseline glycemic categories in type 1 diabetes (T1D) clinical trials. Methods Baseline and 3-month central laboratory-measured A1C and estimated GMI from sensor glucose were collected from T1D clinical trials (DCLP3, DCLP5, and WISDM), encompassing children, adolescents, adults, and older adults. Magnitude and direction of changes (baseline-3 months) in A1C vs GMI were compared overall across the studies and by stratified baseline A1C (<7%, 7%-9%, >9%). Results A modest correlation was found between changes in A1C and GMI (r = 0.34). Participants with baseline A1C > 9% had larger reductions in A1C compared to GMI (−1.2 [−2.1 to −0.6] vs −0.6 [−0.94 to 0], P < .01). Those with baseline A1C between 7% and 9% showed a greater decline in A1C than GMI (−0.4 [−0.9 to −0.1] vs −0.12 [−0.49 to 0.21], P < .01). No significant difference was observed for baseline A1C < 7%. Conclusion Change in GMI is influenced by the baseline A1C of the participants and it underestimates the true change in A1C. Use of GMI as an endpoint in clinical trials may not reliably capture efficacy of an intervention in T1D trials or real-world studies.