Glycemic variability in gestational diabetes mellitus and its association with β cell function

Abstract

Maternal hyperglycemia in gestational diabetes mellitus (GDM), especially hyperglycemic excursions, is associated with increased risks of adverse pregnancy outcomes. Continuous glucose monitoring (CGM) system (CGMS) is better than intermittent self-measurements in detecting detailed glucose profiles on the magnitude and duration of glucose fluctuations. Hyperglycemia resulted from impaired β cell function. This study analyzed the characteristics of glycemic variability in GDM with 24-28 gestational weeks and its association with β cell function. Thirty GDM with 24-28 gestational weeks (GDM group) were included in this study, and 20 normal gestational women (NGW group) and 20 normal glucose regulation non-pregnant women (NGRW group) were set as controls. The three groups were monitored using the CGMS for consecutive 72h. The parameters of glycemic variability included the standard deviation of blood glucose (SDBG), mean of continuous 24-h blood glucose (MBG), mean amplitude of glycemic excursions (MAGEs), and mean of daily differences (MODDs). Homeostasis model assessments were applied to access the insulin resistance (HOMA-IR). The early insulinogenic index (ΔI30/ΔG30) and the area under the curve of insulin (AUCI180) derived from 75-g oral glucose tolerance test were applied to evaluate the early-phase insulin secretion and second-phase insulin secretion, respectively. After CGM, MAGE and MBG value increased progressively from NGRW, NGW to GDM group (p<0.05); MODD and SDBG values of GDM group were all higher than those of NGRW and NGW groups (p<0.05), but there are no differences in MODD and SDBG between NGRW and NGW groups (p>0.05). After comparison of β cell function, ΔI30/ΔG30 decreased progressively from NGRW, NGW to GDM group (p<0.05); HOMA-IR and AUCI180 increased progressively from NGRW, NGW to GDM group (p<0.05). MAGE value was correlated with ΔI30/ΔG30 and HOMA-IR in GDM group (r=-0.78 and 0.65, respectively, p<0.05). Multiple linear stepwise regression analysis showed that ΔI30/ΔG30 and HOMA-IR were the independent factors of MAGE (β=-0.61, 0.34, respectively, p<0.05). Glycemic variability in GDM was higher than in normal pregnant women, and glycemic variability evaluated by MAGE correlates well with impaired early-phase insulin secretion in GDM. Further large-scale studies are needed to formulate treatment strategies to make up for the impaired early-phase insulin secretion and flat glycemic variability, and analyze the association between pregnancy outcomes improvement and glycemic variability remission in GDM.