Glucose management indicator (GMI) was developed using a linear regression formula to estimate A1C from mean glucose. The study population that was used to develop and validate GMI had a mean A1C of 7.3 ± 0.8%. Therefore, we hypothesize that GMI may overestimate A1C in those with A1C <6.5%. Non-pregnant adults (≥18 years) with T1D for ≥2 years and using Dexcom G6 for ≥6 months from Barbara Davis Center for Diabetes (n=93, 40.6 ± 12.6 years, female 58.4%, A1C 5.9 ± 0.4%) and subjects from “Glucose Sensor Profile in Healthy Non-diabetic Subjects” study (N=153, 31.2± 21.0 years, female 66.0%, A1C 5.1 ± 0.3%) were included. All participants used Dexcom G6 continuous glucose monitor (CGM) . Up to days (healthy participants) and 15 days (T1D patients) of CGM data was compared with laboratory A1C. Discordance in GMI and A1C is presented by A1C <5.7% (healthy participants) and A1C between 5.7-6.4% (T1D patients) . The degree of discordance between GMI and A1C was higher in those with A1C<6.5% compared to original GMI development cohort [Figure 1A]. The differences between A1C and GMI were largely negative in both A1C categories (A1C <5.7% and A1C of 5.7-6.4%) . On average, GMI was 0.59% higher in the A1C < 5.7% group and 0.49% higher in the A1C 5.7 - 6.4% group (both p < 0.0001) [Figure 1B and 1C]. The current GMI formula may overestimate A1C and may need to be re-assessed for those with A1C <6.5%.


V.Shah: Advisory Panel; Medscape, Sanofi, Consultant; Dexcom, Inc., Research Support; Dexcom, Inc., Eli Lilly and Company, Insulet Corporation, Novo Nordisk. T.B.Vigers: None. L.Pyle: None. S.Dubose: None. P.Calhoun: None. R.M.Bergenstal: Advisory Panel; Hygieia, Medtronic, Roche Diabetes Care, Zealand Pharma A/S, Consultant; Abbott Diabetes, Ascensia Diabetes Care, Bigfoot Biomedical, Inc., CeQur SA, Dexcom, Inc., Eli Lilly and Company, Novo Nordisk, Onduo LLC, Sanofi, United HealthCare Services, Inc., Research Support; Abbott Diabetes, Dexcom, Inc., Eli Lilly and Company, Insulet Corporation, Medtronic, Novo Nordisk, Sanofi.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at