OBJECTIVE: Determine the relationship between mean glucose (MG), as assessed by continuous glucose monitoring (CGM), and hemoglobin A1c (A1C) in insulin-requiring adults in a clinical practice setting and examine the persistence of this relationship over time. METHODS: In this retrospective record review in a diabetes practice, a linear regression model was developed using data sets from all patients with ≥1 available download of a Dexcom SevenPlus CGM device in which there was >50% utilization in the 60 days prior to a laboratory A1C. Persistence of the MG to A1C relationship was examined in patients with ≥2 data sets available. RESULTS: A total of 139 patients had ≥1 data set available for evaluation, and 101 patients had ≥2 data sets (range, 2 to 6; total, 279). The slope of the MG versus A1C curve was 19.5 mg/dL for each 1% change in A1C, with an intercept of 17.7 mg/dL. Although 88% of the measured MG values were within 15% of the A1C-predicted MG, there was substantial variation in individuals, with differences as large as ±26%. The MG to A1C (MG:A1C) ratio, which is a measure of glycation, was normally distributed, with a median of 21.6. Spearman correlation coefficients for the MG:A1C ratio on repeated measures ranged from 0.52 to 0.73, demonstrating persistence. CONCLUSION: The relationship between MG and A1C is linear in a population but can vary between individuals. The MG:A1C ratio was normally distributed, tended to persist in individuals over time, and thus could be useful to identify apparent high and low glycators. Glycemic goals may need to be modified in such patients.
OBJECTIVE: Determine the relationship between mean glucose (MG), as assessed by continuous glucose monitoring (CGM), and hemoglobin A1c (A1C) in insulin-requiring adults in a clinical practice setting and examine the persistence of this relationship over time. METHODS: In this retrospective record review in a diabetes practice, a linear regression model was developed using data sets from all patients with ≥1 available download of a Dexcom SevenPlus CGM device in which there was >50% utilization in the 60 days prior to a laboratory A1C. Persistence of the MG to A1C relationship was examined in patients with ≥2 data sets available. RESULTS: A total of 139 patients had ≥1 data set available for evaluation, and 101 patients had ≥2 data sets (range, 2 to 6; total, 279). The slope of the MG versus A1C curve was 19.5 mg/dL for each 1% change in A1C, with an intercept of 17.7 mg/dL. Although 88% of the measured MG values were within 15% of the A1C-predicted MG, there was substantial variation in individuals, with differences as large as ±26%. The MG to A1C (MG:A1C) ratio, which is a measure of glycation, was normally distributed, with a median of 21.6. Spearman correlation coefficients for the MG:A1C ratio on repeated measures ranged from 0.52 to 0.73, demonstrating persistence. CONCLUSION: The relationship between MG and A1C is linear in a population but can vary between individuals. The MG:A1C ratio was normally distributed, tended to persist in individuals over time, and thus could be useful to identify apparent high and low glycators. Glycemic goals may need to be modified in such patients.
Authors: Roy W Beck; Crystal G Connor; Deborah M Mullen; David M Wesley; Richard M Bergenstal Journal: Diabetes Care Date: 2017-08 Impact factor: 19.112