Impacts of glycemic variability on the relationship between time in range and estimated glycated hemoglobin in patients with type 1 diabetes mellitus

Abstract

Objective: This study is to investigate the relationship between time in range (TIR) and glucose management indicator (GMI), and the impact of glycemic variability (GV) on their relationship in patients with type 1 diabetes mellitus (T1DM). Methods: The CGM data were collected from a multicenter randomized clinical trial of adults (≥18 years old) with T1DM, including 83 T1DM patients, respectively from the Third Affiliated Hospital of Sun Yat-sen University (72 cases), Drum Tower Hospital Affiliated to Nanjing University School of Medicine (2 cases), and the First Affiliated Hospital of University of Science and Technology of China (9 cases). All subjects wore the iProTM2 system for 14 days at baseline (0-2 weeks), 3 months (12-14 weeks), and 6 months (24-26 weeks). Data derived from iProTM2 sensor was used to calculate CGM parameters. Correlation between TIR and GMI was explored according to different stratification of glycemic variability assessed by glucose coefficient of variation (CV). Predicted TIR in the fixed GMI value was calculated via the linear regression equations performed in the respective interquartile group of CV. Results: From November 2017 to June 2021, a total of 233 CGM data were collected with 83 collected from baseline, 80 from the 3-month follow-up, 70 from the 6-month follow-up. Patients including 27 males had a median (Q1, Q3) age of 30.69 (25.22, 38.43) years, with a diabetes duration of 10.05(4.46, 13.92) years. The median (Q1, Q3) and effective wearing time of available CGM data was 13.92 (13.02, 14.00) days and 91.61% (84.96%, 95.94%), and the value of TIR, GMI and CV was 60.34%±13.03%, 7.14%±0.61% and 41.01%±7.64%, respectively. There was a strong negative correlation between TIR and GMI (r=-0.822, P<0.001). Multiple linear regression analysis showed that the predictive value of TIR calculated from a given GMI was 8.352% higher when CV was up to standard (36%) than that when CV was down to standard. Based on the multiple linear regression equations generated from quartiles of CV, the predicted TIR value was decreased across the ascending quartiles with 69.98 % in the lowest quartile of CV (≤35.91%), 64.57 % in 25th-50th quartile of CV (35.91%<CV≤40.08%), 60.96% in 50th-75th quartile of CV (40.08%<CV≤45.86%) and 56.44% in the highest quartile of CV (>75th quartile, CV>45.86%) when GMI was set as 7%. Conclusions: There is a strong correlation between TIR and GMI in adult patients with T1DM in patients with type 1 diabetes mellitus. CV influenced the relationship between TIR and GMI.