| Part One: Reference Range of Glycated Albumin for Adults in Shaanxi ProvinceBackground: At present,many laboratories in different regions at home and abroad have established their own normal reference ranges for glycated albumin(GA)based on the characteristics of the local population,but the GA reference range has not been established in Shaanxi.Objective: To establish the reference range of serum GA for adults in Shaanxi.Methods: The subjects of the study were selected from the China National Diabetes and Metabolic Disorders Survey(CNDMDS)during 2016～2017.From 1101 subjects,we excluded 332 people with missing GA value,182 people with prediabetes,250 people with diabetes,2 people with abnormal liver function,1 people with impaired renal function,9people with outlier GA value.Totally 325 subjects were included.The median age of the subjects is 48.0±10.4 years.There are 105 males and 220 females.Enzymatic method was used to detect serum GA levels From June 2019 to December 2019,and the results were double-entered using Epi Data 3.1 software.Results: Two-sided 95% percentile method was used to establish the GA reference interval because GA does not conform to the normal distribution in the population of our study.The normal reference interval for GA in Shaanxi province was 10.8%～17.3%.There was no significant difference in GA levels between men and women.There was no significant difference in GA levels between people from middle area and northern area of Shaanxi.GA leves were positively correlated with age(r=0.353,P<0.001)and increased along with aging.The normal reference ranges of GA for young,middle-aged,and old people were established as follows: 10.6%～15.1%,11.1%～17.4%,11.2%～17.9%.There was no significant difference in GA levels among different body mass index(BMI)groups.Conclusion: The reference interval of serum GA for adults in Shaanxi is 10.8%～17.3%.GA level has a positive association with age and the normal reference of GA for young,middle-aged,and old people were: 10.6%～15.1%,11.1%～17.4%,11.2%～17.9%respectively.Part Two: Comparison of Glycated Albumin with Glycated Hemoglobin for Diagnoses of Type 2 Diabetes and Prediabetes: Cross-Sectional StudyBackground: GA and Hb A1 c are both effective indicators for evaluating blood sugar.However,the differences of GA and Hb A1 c in diagnosing type 2 diabetes mellitus(T2DM)and prediabetes,and whether GA is involved in the process of prediabetes progressing to T2 DM have not been studied.Objective: To compare the differences of GA with Hb A1 c in the diagnosis of T2 DM and prediabetes,and the relationships with the progression of prediabetes to T2 DM.Methods: The subjects of the study were selected form 2016～2017 CNDMDS and2014～2016 Xi’an CBMDE project.From 4102 subjects,114 people with missing main variables were excluded,and a total of 3988 subjects were finally included for the study.T2 DM and prediabetes were diagnosed according to the 1999 WHO criterion.Receiver operating characteristic(ROC)and area under the curve(AUC)were used to evaluate the diagnostic abilities of biomarkers.GA and Hb A1 c were converted to binary categorical variables with the median as the cut point and logistic regression was used to analyze the relationships between GA and the progression of prediabetes to T2 DM.Results: For patients with type 2 diabetes,after multivariate adjustment,BMI was positively correlated with Hb A1c(β=0.110,P<0.001),but was negatively correlated with GA(β=-0.113,P<0.001).There was no significant difference in ROC-AUCs between GA and Hb A1 c in the diagnosis of prediabetes(0.661 VS 0.655,P=0.7279),but the ROC-AUC of GA in the diagnosis of T2 DM was significantly greater than that of Hb A1c(0.898 VS0.846,P<0.001).For patients with hyperuricemia,the ROC-AUC of Hb A1 c in the diagnosis of prediabetes was significantly greater than that of GA(0.732 VS 0.593,P=0.015);for people with normal blood uric acid,the ROC-AUCs of Hb A1 c and GA in the diagnosis of prediabetes have no significant differences(0.658 VS 0.674,P=0.3559).For T2 DM with only 2 hour postprandial blood glucose(2h-PBG)reaching the diagnostic criteria,the ROCAUC of GA in the diagnosis of T2 DM was significantly higher than that of Hb A1c(0.860 VS 0.791);for T2 DM with only fasting blood glucose(FBG)meeting the diagnostic criteria,the ROC-AUC of GA and Hb A1 c in the diagnosis of T2 DM have no significant difference(0.756 VS 0.756,P=0.9944).After multivariate adjustment,GA and Hb A1 c were negatively correlated with HOMA-β(P<0.001),and positively correlated with HOMA-IR(P<0.001);GA > 14.4%(OR=7.852,P<0.001)and Hb A1 c > 5.7%(OR=2.842,P<0.001)were independently positively correlated with the development of prediabetes to T2 DM.Conclusion: GA is more capable of diagnosing type 2 diabetes than Hb A1 c.Hb A1 c is a better indicator for diagnosing prediabetes in people with hyperuricemia,and GA is a better indicator for diagnosing T2 DM with elevated postprandial blood glucose.Both GA and Hb A1 c are independent risk factors for the development of prediabetes to type 2diabetes.Part Three: Correlations between Glycated Albumin and Type 2 Diabetes and Prediabetes Combined with Cardiovascular and Cerebrovascular Diseases: Cross-sectional studyBackground: At present the relationship between GA and T2 DM combined with cardiovascular and cerebrovascular diseases(CCVD)is still controversial.In addition,many studies found that even if blood sugar does not meet the diagnostic criteria for prediabetes,the risk of CCVD in prediabetic people is significantly increased.However,there is insufficient evidence to prove that GA is independently associated with prediabetes combined with CCVD.Objective: To analyze the correlation between GA and T2 DM and prediabetes combined with CCVD.Methods: This study was based on a cross-sectional database.The subjects of the study were from the 2016～2017 survey of CNDMDS in Shaanxi area(1101 people)and all participants(3001 people)of 2014～2016 Xi’an CBMDE.114 people with missing main variables,2138 people with normal glucose metabolism were excluded from them,and totally 1850 people were finally included,including 883 people with prediabetes and 967 people with T2 DM.T2DM and prediabetes were diagnosed according to the 1999 WHO standard.In the present study,CCVD includes previous myocardial infarction,unstable angina pectoris,cerebral infarction or transient cerebral hemorrhage which were diagnosed through self-reporting histories.According to whether they had CCVD,the prediabetic population was divided into pre-DM group(834 people)and pre-DM combined CCVD group(49 people).T2 DM patients were divided into T2 DM group(859 people)and T2 DM combined CCVD group(108 people).GA and Hb A1 c were converted to binary categorical variables with the median as the cut point and logistic regression was used to analyze the relationships between GA and prediabetes/T2 DM with CCVD.Results: The GA level(15.27±2.06% VS 14.56±2.17%,P=0.020)of the pre-DM combined CCVD group was significantly higher than that of the pre-DM group,but the Hb A1 c level of the two groups(5.8±0.6% VS 5.7±0.5%,P >0.05)had no significant difference.The logistic regression of prediabetes combined with CCVD showed that only age(OR=1.191,P<0.001),gender/male(OR=1.987,P<0.001)and GA>16%(OR=1.357,P=0.001)entered the regression equation.There was no significant difference in the levels of GA(18.67±5.63% VS 18.63±4.97%,P=0.593)and Hb A1c(6.6±1.6% VS 6.7±1.9%,P=0.597)between T2 DM group and T2 DM combined with CCVD group.The logistic regression of T2 DM combined with CCVD showed that neither GA nor Hb A1 c entered the regression equation.Conclusion: GA but not Hb A1 c is an independent risk factor for prediabetes complicated with CCVD.Both GA and Hb A1 c have no correlation with T2 DM complicated with CCVD.Part Four: The Predictive Utility of Glycated Albumin for Type 2 Diabetes and Prediabetes: A Retrospective Cohort StudyBackground: Due to the lack of evidence of GA-related longitudinal studies at home and abroad,especially in China,where the development of cohort research is relatively lagging,it is still unclear whether GA can effectively predict T2 DM and prdiabetes.Objective: To study the predictive value of GA for T2 DM and prediabetes through a retrospective cohort design and to establish prediction models.Methods: Among 885 people in the 2007～2008 CNDMDS who attended the 2016～2017 CNDMDS follow-up survey,87 subjects with diabetes,161 subjects with prediabetes,8 subjects with liver function impairment,1 subjects with abnormal renal function,107 subjects with missing variables,5 persons with GA outlier were excluded,and 516 subjects were finally included.During follow-up from 2016 to 2017,51 persons developed T2 DM and 92 persons developed prediabetes,373 people remained normal blood glucose.T2 DM and prediabetes were diagnosed according to the 1999 WHO criterion.Multivariate adjusted COX proportional hazard models and relative risk(RR)were used to analyze the relationship between GA groups and the risk of T2 DM and prediabetes.R program is used to establish restricted cubic splines and T2 DM prediction models.ROC-AUCs of different prediction models were compared using the "Delong" method.Results: Compared with the normal control group,the baseline GA levels of the prediabetes group(P<0.05)and T2 DM group(P<0.001)were significantly higher.After multivariable correction,the RR(95% CI)of T2 DM according to GA quartiles from low to high were 1,2.43(0.60-6.29),4.68(1.25-13.04),5.39(1.86-16.74)respectively,and the trend was significant(P<0.001).The RR value of T2 DM with the maximum GA value is approximately 10 times higher than T2 DM with the minimum GA value.Although the RR value of prediabetes gradually increased with GA quartiles,the trend was not significant after multivariate adjustment(P=0.609).ROC-AUC of GA for T2 DM was not significantly different from the ROC-AUCs of FBG(0.698 VS 0.655,P=0.367)and 2h-PBG(0.698 VS0.725,P=0.552).Using GA to replace FBG(0.856 VS 0.877,P=0.215)or 2h-PBG(0.856 VS 0.874,P=0.417)or both FBG and 2h-PBG in the T2 DM prediction models(0.856 VS0.867,P=0.635)had no significant influences on the ROC-AUCs of the prediction models.Conclusion: GA is an useful indicator for predicting T2 DM,but its ability to predict prediabetes is limited.The prediction models established in this study have excellent performances and can effectively predict T2 DM. |
PDF Full Text Request