The aim of this study was to assess the relationships among indices of glycemic variability obtained by continuous
glucose monitoring (CGM). CGM was performed in 88 patients with diabetes (20 type 1 and 68
type 2 diabetes, age 59 ± 15 years) admitted to our hospital (Keio University Hospital, Tokyo, Japan) between 2010 and 2012. Mean
glucose,
glucose standard deviation (
SDglu) and other glycemic indices such as index of
glycemic control (ICG), J-index, mean of daily differences (MODD), continuous overlapping net glycemic action 1 (CONGA1), mean amplitude of glycemic excursions (MAGE) and M value were calculated from CGM data, and the correlations among these indices were assessed. There were strong correlations between
SDglu and the indices MAGE, CONGA1, MODD and M value (all r > 0.8, P < 0.05). On the other hand, mean
glucose was strongly correlated with J index and M value (both r > 0.8, P < 0.05).
SDglu and other glycemic variability indices were more strongly correlated with
hypoglycemia than was mean
glucose, and the combination of mean
glucose and
SDglu was useful for predicting
hypoglycemia in patients with diabetes. In this study, we demonstrated the characteristics of various glycemic variability indices in relation to mean
glucose and
SDglu. This information will help physicians to understand the characteristics of various glycemic variability indices and to select an appropriate index for their purpose. Our results also underpin the importance of glycemic variability in relation to risk of
hypoglycemia in patients with diabetes.