Abstract

OBJECTIVE—Recent studies show the importance of controlling blood glucose variability in relationship to both reducing hypoglycemia and attenuating the risk for cardiovascular and behavioral complications due to hyperglycemia. It is therefore important to design variability measures that are equally predictive of low and high blood glucose excursions.

RESEARCH DESIGN AND METHODS—We introduce the average daily risk range (ADRR), a variability measure computed from routine self-monitored blood glucose (SMBG) data. The ADRR was constructed using a development dataset for 39 and 31 adults with type 1 and type 2 diabetes, respectively. The formula was then fixed, and the ADRR was compared against other variability measures using an independent validation dataset containing ∼4 months of SMBG for 254 and 81 adults with type 1 and type 2 diabetes.

RESULTS—From the 1st month of validation SMBG data, we computed the ADRR, blood glucose SD and coefficient of variation, daily blood glucose range and interquartile range, mean amplitude of glycemic excursion, M-value, and lability index. Then all measures were tested as predictors of low blood glucose (<2.2 mmol/l; <3.9 mmol/l) and high (>10 mmol/l; >22.2 mmol/l) events in the subsequent 3 months. The ADRR was the best predictor of both hypoglycemia and hyperglycemia, with a 6-fold increase in the likelihood of hypoglycemia and 3.5-fold increase in the likelihood of hyperglycemia across its risk categories.

CONCLUSIONS—In a large SMBG database, the ADRR showed strong association with subsequent out-of-control glucose readings. Compared with other variability measures, the ADRR demonstrated a superior balance of sensitivity to predicting both hypoglycemia and hyperglycemia. This prediction was independent from type of diabetes.