The added and interpretative value of cgm-derived parameters in type 1 diabetes depends on the level of glycemic control

Objective: In type 1 diabetes mellitus (T1DM) management, continuous glucose monitoring (CGM)-derived parameters can provide additional insights, with time in range (TIR) and other parameters reflecting glycemic control and variability being put forward. This study aimed to examine the added and interpretative value of the CGM-derived indices TIR and coefficient of variation (CV%) in T1DM patients stratified according to their level of glycemic control by means of HbA1C. Methods: T1DM patients with a minimum disease duration of 10 years and without known macrovascular disease were enrolled. Patients were equipped with a blinded CGM device for 7 days. TIR and time spent in hypoglycemia and hyperglycemia were determined, and CV% was used as a parameter for glycemic variability. Pearson (r) and Spearman correlations (r(s)) and a regression analysis were used to examine associations. Results: Ninety-five patients (age: 45 +/- 10 years; HbA1C level: 7.7% +/- 0.8% [61 +/- 7 mmol/mol]) were included (mean blood glucose [MBG]: 159 +/- 31 mg/dL; TIR: 55.8% +/- 14.9%; CV%: 43.5% +/- 7.8%) and labeled as having good (HbA1C level 8% [>64 mmol/mol]; n = 31) glycemic control. HbA1C was significantly associated with MBG (r(s) = 0.48, P < .001) and time spent in hyperglycemia (total: r(s) = 0.52; level 2: r = 0.46; P < .001) but not with time spent in hypoglycemia and CV%, even after an analysis of the HbA1C subgroups. Similarly, TIR was negatively associated with HbA1C (r = -0.53; P < .001), MBG (r(s) = -0.81; P < .001), and time spent in hyperglycemia (total: r(s) = -0.90; level 2: r(s) = -0.84; P < .001) but not with time in hypoglycemia. The subgroup analyses, however, showed that TIR was associated with shorter time spent in level-2 hypoglycemia in patients with good (r(s) = -0.60; P = .007) and moderate (r(s) = -0.25; P = .047) glycemic control. In contrast, CV% was strongly positively associated with time in hypoglycemia (total: r(s) = 0.78; level 2: r(s) = 0.76; P < .001) but not with TIR or time in hyperglycemia in the entire cohort, although the subgroup analyses showed that TIR was negatively associated with CV% in patients with good glycemic control (r = -0.81, P < .001) and positively associated in patients with poor glycemic control (r = +0.47; P < .01). Conclusion: The CGM-derived metrics TIR and CV% are related to clinically important situations, TIR being strongly dependent on hyperglycemia and CV% being reflective of hypoglycemic risk. However, the interpretation and applicability of TIR and CV% and their relationship depends on the level of glycemic control of the individual patient, with CV% generally adding less clinically relevant information in those with poor control. This illustrates the need for further research and evaluation of composite measures of glycemic control in T1DM. (C) 2020 AACE. Published by Elsevier Inc. All rights reserved