The different methods of assessing glycemic variability, quality of glycemic control and glycemic risk cannot be interpreted as equivalent in clinical practice

Objective: Several methods are available to calculate glycemic variability (GV), quality of glycemic control (QGC) and glycemic risk (GR). However, clinicians do not easily interpret these data. This study evaluates whether the results of the different methods can be interpreted as equivalent. Methods: A prospective study was performed including outpatients with DMT2 evaluated at the San Ignacio Hospital and the Colombian Diabetes Association in Bogotá, Colombia. From six-day continuous glucose monitoring data, GV (SD, CV, IQR, MODD, MAGE), QGC (M-value, J-index) and GR (LBGI, HBGI) were calculated. Reference values ​​were generated, classifying the patients according to GV control quartiles (excellent, good, fair or poor). The concordance between the different indices was evaluated. Results: In total, 140 patients (68.9 ± 11.2 years) were included. The agreement levels (Kappa) between GV indices were moderate, 0.40 (CI 95%:0.29–0.51), 0.42 (CI 95%:0.31–0.53) and 0.39 (CI 95%:0.28–0.50), for CV versus SD, IQR and CONGA respectively. The levels of agreement between GV and QGC indices were minimal (Kappa CV vs. M-value, 0.15CI 95%:0.046–0.26) and weak between the GV and GR indices (Kappa CVvs.LBGI 0.37CI95%:0.26–0.48). The estimators did not improve significantly when the analysis was performed with linearly weighted or quadratic weighted Kappa. Conclusions: The present study demonstrates that the concordance between the clinical interpretation of the different GV, QGC and GR indices is poor, suggesting that they cannot be assumed as equivalent, so different indices evaluating different concepts, must be evaluated simultaneously to analyze adequately each patient. New studies are needed to evaluate which of the methods better predicts hypoglycemia and microvascular or macrovascular complications.