A multicenter study of the correlation between advanced glycation end products in the lens and type 2 diabetes
周薇,唐文彬,吴静,岳星,李岚,郑璠,赵颖,王彦平,刘绍辉
DOI: https://doi.org/10.3760/cma.j.cn115624-20211112-00684
2022-01-01
Abstract:Objective:To investigate the relationship between advanced glycation end products (AGEs) in the lens and type 2 diabetes mellitus.Methods:226 subjects were recruited between August 14 to September 14, 2018 from the Endocrinology Department of Central South University Xiangya Hospital, the Third Hospital of Changsha City, and the Fourth Hospital of Changsha City. The OGTT test, combined with clinical indicators, were used as the gold standard. Subjects were screened for type 2 diabetes using both the lens AGE fluorescence assay and the gold standard. Drawing the receiver operating characteristic (ROC) curve, we calculated the area under the curve (AUC) and its 95%
CI and calculated the AGE for the diagnosis of type 2 diabetes. Sensitivity, specificity, Youden index, Kappa value, and its 95%
CI, and the optimal cut-off value were determined according to the Youden index. Taking diabetes as the outcome indicator and AGE as the binary indicator, three logistic regression models were constructed. Stratified by age and sub-center, the differences between fasting blood glucose and 2 h postprandial blood glucose were compared between the AGE-negative and AGE-positive groups to determine the relationship between AGE and diabetes.
Results:The area under the ROC curve was 0.86(95%
CI: 0.81-0.91). According to the Youden index, the optimal cut-off point for AGE was 0.24. At this time, the sensitivity was 82.86(95%
CI: 77.81-87.91), the specificity was 77.06(95%
CI: 71.43-82.7), the Youden index was 59.92(95%
CI: 53.36-66.49), the Kappa value was 79.62(95%
CI: 74.22-85.02). Except for the 20-39-year-old group, the fasting blood glucose and 2 h postprandial blood glucose of the AGE-positive group in different age groups, different sub-centers, and the general population were higher than those of the AGE-negative group (all
P<0.05). After adjusting for the confounding effects of age, gender, and sub-center (model 3), the relative risk of diabetes in the AGE-positive group was 11.75 times higher than the AGE-negative group (95%
CI: 5.61-24.60), all with
P<0.001.
Conclusion:There was a high correlation between AGE in the lens and the risk of type 2 diabetes. When the cut-off point of AGE is 0.24, it had high sensitivity and specificity and could be used as a practical tool for early screening of type 2 diabetes.