ABSTRACT
Background@#Among the various glycemic indices in current use, glycemic variability has the greatest contribution in the development of microvascular and macrovascular complications in Type 2 Diabetes mellitus (T2DM). Most metrics that are currently used to measure glycemic variability are derived from continuous glucose monitoring (CGM) data. However, CGM is burdensome to the patient due to its relatively high cost as well as the need for multiple visits with the health care provider. With the use of serum 1,5-anhydroglucitol (1,5-AG) as a biomarker of glucose fluctuations, physicians and patients alike could have an easier surrogate measure of glycemic variability thus aiding in achieving target glucose control. This study aims to determine the diagnostic accuracy of 1,5-AG as compared to the glycemic variability metrics derived from CGM as a surrogate measure of glycemic variability among adult Filipinos with T2DM.@*Methods@#Retrospective analysis of data of adult patients aged 20 years old and above diagnosed with T2DM referred for CGM at the Diabetes, Endocrine, Metabolic, and Nutrition Center of Cardinal Santos Medical Center from January 2017 to October 2021 who underwent serum 1,5-AG level determination within 2 weeks of CGM were collected. Diagnostic accuracy was obtained by computing the sensitivity, specificity, positive (PPV) and negative predictive values (NPV), and Youden index. Pearson correlation coefficient was used to determine the correlation of 1,5-AG and the different metrics. Analysis of variance (ANOVA) was used to check for statistical significance with 99% confidence interval and a p < 0.05 considered as statistically significant.@*Results@#This study involving 37 subjects showed a good diagnostic accuracy of serum 1,5-AG levels with the different measures of glycemic variability derived from CGM namely mean amplitude of glycemic excursion (MAGE), continuous overlapping net glycemic action at 1-hour intervals (CONGA-1), and mean of daily differences (MODD) with significant correlation among patients with HbA1c ≤ 7%. Subjects were on CGM for approximately 6 ± 1 day with statistically significant difference between the good and poor glucose control group (p<0.05). Determination of diagnostic accuracy between 1,5- AG and MAGE showed good accuracy (Sensitivity = 95.3%, Specificity = 100%, PPV = 100%, NPV = 75.43%, Diagnostic accuracy 96%, and a Youden Index of 92.3) with a statistically significant correlation among subjects with HbA1c level ≤ 7% (p=0.021). There is likewise good diagnostic accuracy between CONGA-1 and 1,5-AG level (Sensitivity = 99%, Specificity = 75.29%, PPV = 89.1%, NPV = 97%, Accuracy = 89.50% and Youden index of 58.41) with a statistically significant correlation among subjects with HbA1c ≤ 7% (p=0.038). Comparison with interday glycemic variability showed fair diagnostic accuracy between MODD and 1,5-AG (Sensitivity = 79.17%, Specificity = 78%, PPV = 97%, NPV = 32%, Accuracy = 76.89%, and Youden index of 49.07) and a statistically significant correlation among subjects with HbA1c ≤ 7% (p=0.009).@*Conclusion@#There is good diagnostic accuracy of serum 1,5-AG levels with the different measures of glycemic variability derived from CGM namely MAGE, CONGA-1, and MODD with significant correlation among patients with HbA1c ≤ 7%. Among diabetics with HbA1c ≤7%, 1,5-AG could be used as a surrogate measure of glycemic variability and excursions.