Using HbA1c to reflect AG has limitations. The AGR has been proposed as a correction factor to compute a personalized estimated average glucose (peAG). Three CGM datasets with contemporary HbA1c values (mostly White individuals with T1D [REPLACE BG - RBG], and mainly Hispanic/Latinos with T2D and prediabetes [FARMING FOR LIFE-FFL and DIGITAL HEALTH - DH]) were used to estimate AGR. The RBG dataset was also used to compare peAG vs eAG. No significant differences in AGR were found across age, BMI, gender, or HbA1c in the FFL (n=105, 91% Hispanic/Latino, 32% T2D, age 54 ± 12 yr) and DH (n=26, 100% Hispanic/Latino, 39% T2D, age 50 ± 12 yr) data. However, the median AGRs (72.3 [69, 76.2] ml/g in FFL, and 75.3 [72.8, 83.8] ml/g in DH) were higher than the reference of 65.1 ml/g (p<0.005) from predominantly White individuals. Using 3-month CGM data from RBG (n=57, 90% White, 100% T1D, age 50 ± 13 yr) to compute AGR, a Bland-Altman plot showed that confidence intervals narrowed when using peAG (Figure). Also, AG had a higher correlation with peAG compared to eAG. Bias in peAG/eAG-AG lost significance when HbA1c was measured 3 months after the CGM data for AGR.Hispanic/Latinos with or at risk of T2D may have higher AGR levels than the previously suggested value, implying an overestimation of AG values from HbA1c. The use of AGR and peAG in stable glycemia can provide a more accurate representation of AG than eAG.

Disclosure

S. E. Sato Imuro: None. D. Kerr: Consultant; Sanofi-Aventis U. S. W. C. Bevier: None. A. Sabharwal: None.

Funding

U.S. Department of Agriculture (2018-33800-28404); National Science Foundation (1648451); Elsevier Foundation; Sun Life Financial

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 http://www.diabetesjournals.org/content/license.