Objective: To investigate the associations between the patterns of change of self-monitored blood glucose (SMBG) and the parameters of intraday blood glucose variability [mean absolute glucose excursions (MAGE), mean postprandial glucose excursion (MPPGE) and standard deviation of blood glucose (sBG)] measured by the continuous glucose monitoring system.
Methods: A 72-hour continuous glucose monitoring was performed in a sample 105 people with impaired glucose tolerance (IGR, n=51) and newly-diagnosed type 2 diabetes mellitus (T2DM, n=54) to calculate MAGE, MPPGE and sBG. Meanwhile, fingertip blood glucose self-monitoring was performed to determine fasting blood glucose (FBG), blood glucose after breakfast (BG(AB)), blood glucose before lunch (BG(BL)), blood glucose after lunch (BG(AL)), blood glucose before supper (BG(BS)), blood glucose after supper (BG(AS)), and blood glucose before sleeping (BG(BR)) at the same period of time. Multiple stepwise regression analysis was performed to generate equations for predicting MAGE, MPPGE and sBG with age and the self-monitoring blood glucose parameters in 80% of the subjects (41 IGR and 44 T2DM, randomly selected from the overall sample). These equations were then cross-validated in the remaining 20% subjects (10 IGR and 10 T2DM).
Results: BG(AA), BG(AB), BGAL and FBG entered into the regression equations predicting MAGE, sBG and MPPGE for the IGR subject, while age only entered into the regression equations predicting MPPGE and sBG. For the subjects with T2DM, BG(AS), BG(AL) and age entered into the equation predicting MAGE; BG(AS), BG(AL), BG(BL) and BG(BS) entered into the equation predicting MPPGE; BG(AS), BG(AL) and FBG entered into the equation predicting sBG. The cross-validation study showed that the differences between predicted and observed values of MAGE in the subjects with IGR and T2DM were 4.1% and 8.2%, respectively; the differences between predicted and observed values of MPPGE in the subjects with IGR and T2DM were 23.1% and 1.3%, respectively; and the differences between predicted and observed values of sBG in the subjects with IGR and T2DM were 1.2% and 6.8%, respectively. Except for MPPGE in the subjects with IGR, the goodness of fit between predicted and observed values were good.
Conclusion: The MAGE and sBG in people with IGR and the MAGE, MPPGE and sBG in patients with T2DM can be well predicted with age and self-monitored blood glucose.