The Cohort-based Study Of The Roles Of Glucose Metabolic Indicators In The Diagnosis And Risk Assessment Of Type 2 Diabetes Mellitus

BackgroundThe prevalence of type 2 diabetes mellites(T2DM)is soaring and becomes a social problem threatening human being' s health and survival globally.According to the latest data issued by IDF in 2019,there are 463 million adult diabetic patients worldwide,with 116 million distributed in China,which owns the largest diabetic population.In addition,much larger prediabetic population also exists in China,which is an amazing reserve of diabetic population.Enough evidence has shown that effective intervention towards prediabetes individuals can reduce the possibility for their progression to diabetes.Thus,the key to prevent diabetes is to identify the prediabetic people timely and perform effective management.Up to now,regarding the diagnostic criteria,no consensus has been achieved.World Health Organization(WHO)1999 criteria and American diabetes association(ADA)2010 criteria are the most influential ones.Recently updated China Guideline for Type 2 Diabetes(2017 edition)adopted WHO 1999 criteria,while whether ADA2010 criteria is suitable for Chinese still remains to be explored.Therefore,it is necessary to compare the different clinical applications of WHO 1999 and ADA2010 criteria in Chinese population,compare the metabolic features of patients stratified with different glucose levels and find the optimal HbAlc cut-off point for Chinese population.In 2019,the article Dubious diagnosis published in Science questioned the concept of prediabetes,stating that the prediabetes concept,a man-made disease,brought in millions of new patients,although the perniciousness was so minor that only less than 2%of prediabetes would progress to diabetes.Lately,a research showed that the prevalence of prediabetes in Brazil was so high as 59%,according to AD A2010 criteria.After a 3.7-year follow up,IFG population and impaired glucose tolerance(IGT)population defined by WHO 1999 criteria,instead of ADA2010 criteria,exhibited higher risk of diabetes.Since identical diagnostic standard was used in both criteria,we drew the conclusion that the prediabetic population defined by WHO 1999 was more likely to progress to diabetes.Similarly,the 4C study,a 5-year program based on ADA2010 criteria,suggested that only IGT population showed higher risk in diabetes and cardiovascular disease.From the two researches above,we can see that the reason why prediabetic population defined by ADA2010 criteria shows lower risk may be that plenty of healthy individuals,classified by WHO 1999 criteria,may fall to the definition of prediabetes when they are re-evaluated via ADA2010 criteria.So,it is urgently to clarify how to include individuals with higher risks of diabetes in a much smaller new prediabetic population.Part 1 Comparison of the clinical applications of WH01999 and ADA2010 diabetes criteria in Chinese adultsObjectiveType 2 diabetes mellitus has become a public issue threatening the health and survival of human beings.However,an international consensus of diagnosis criteria about diabetes have not been achieved.Dramatic difference in diabetes and pre-diabetes prevalence was shown when different diagnostic criteria were chosen.China now is adopting the WHO 1999 criteria,since whether ADA2010 criteria is suitable for Chinese population remains vague.Here,we compared the clinical application of WHO1999 and ADA2010 criteria in Chinese population and analyzed the glucose metabolic characteristics,pancreatic ? cell function and insulin resistance status when subjects were stratified with diverse lab results,in order to find the optimal criteria to filtrate high-risk population for diabetes.MethodsThe present work was one part of the baseline survey from REACTION study.8,239 subjects underwent HbA1c measurement and OGTT after an overnight starvation.We divided those subjects into different groups of glucose metabolic status,via WHO1999 or ADA2010 criteria.Subsequently,we compared the baseline data of different groups,the compositions of diabetic and prediabetic patients,and structures and metabolic features when we stratified those subjects with diverse lab results.ResultsCompared with WHO1999 criteria,ADA2010 criteria led to 67.03%drop in the proportion of normal glucose tolerance(NGT)group,214.58%augment in prediabetic group and 60.02%in diabetic group.For baseline data,except genders,increasing tendency was observed in age,BMI,waist,systolic blood pressure,diastolic blood pressure,heart rate,FPG,2hPG,HbAlc,circulatory insulin,HOMA-IR index,Cr,TG,TC,LDL-c,ALT,AST and GGT with the worsening of glucose metabolic states,while decreasing tendency was shown in HOMA-? index and HDL-c.Regarding compositions of diabetic and pre-diabetic patients,the criteria "HbA1c?6.5%"attributed to 60%increasement for diagnosis of diabetes.For the diagnosis of pre-diabetes,"only meeting the AD A-IFG but not WHO-IFG standard" resulted in 29.22%augment,"only meeting the HbAlc between 5.7%-6.4%" contributed to 122.06%increasement,"meeting both the ADA-IFG but not WHO-IFG standard and HbAlc between 5.7%-6.4%" led to 63.30%rise.When it came to structures and metabolic features when we stratified those subjects with diverse lab results,if we stratified subjects with FPG,with the growing of FPG,2hPG,HbAlc,HOMA-IR index were also increasing while HOMA-? index decreasing with statistical significance.If we stratified subject with 2hPG,with the increasement of 2hPG,FPG,HbA1c,HOMA-IR index were growing,though HOMA-? declining with statistical significance.If we stratified subjects with HbAlc,with the augment of HbAlc,the FPG,2hPG and HOMA-IR index were increasing with statistical significance.As for HOMA-p index,no statistically significant difference was found between group "HbAlc<5.70%" and group "HbA1c between 5.70%-6.4%".But statistically significant difference was shown between group "HbAlc?6.5%" and group "HbA1c<5.7%" or group "HbAlc between 5.7%-6.4%".ConclusionDramatic increase in prevalence of diabetes and pre-diabetes was exhibited when we chose ADA2010 criteria,compared to WHO 1999 criteria,strongly indicating the necessity to find the optimal criteria to diagnose diabetes and pre-diabetes in Chinese population.Part 2 Diagnostic efficiency of HbAlc for newly diagnosed diabetes and pre-diabetes in Chinese adultsObjectiveIn 2010,clinical practice recommendations from ADA,EASD,and IDF have incorporated HbA1c for diagnosing diabetes mellitus.Although the initiative was supported by the results of many large-scale cross-sectional epidemiological surveys,whether HbAlc can be converted from a diabetic monitoring indicator to a diagnostic indicator has been the focus of debate.Europeans and Americans are gradually accepting the HbAlc threshold of 6.5%for diagnosing diabetes considered by the ADA,but the cut-off of HbA1c for the Chinese population remains unclear.We evaluated the diagnostic efficiency of HbAlc for diagnosing newly diagnosed diabetes and pre-diabetes in community-based Chinese and attempted to find the optimal diagnostic cut-off point.MethodsThe present work was one part of the baseline survey from REACTION study.8,239 subjects(5,496 women)aged 40 to 90 years old underwent HbAlc and OGTT measurement after an overnight fast.Diabetes and pre-diabetes were defined by the WHO criteria.The correlation between HbAlc level and diabetes or pre-diabetes was illustrated by fitting smooth curve.The area under the receiver operating characteristic(ROC)curve(AUC)was used to evaluate the diagnostic efficiency of HbAlc,and the optimal cut-off was defined as the point with the largest Youden index on the ROC curveResultsAdjusting for gender and age,HbAlc and diabetes showed a good positive correlation,especially in the range of 6-8%of HbAlc,which was almost a straight upward trend.HbA1c was not as significantly related to pre-diabetes as diabetes,but it is significantly more significant in HbAlc 5.5-7%range,it also showed a straight upward trend.For diagnosing diabetes,the AUC was 0.857(95%confidence interval,0.841-0.873),and when the largest Youden index 0.581 was meeting,the optimal cut-off for HbA1c was 6.3%,with sensitivity 72.2%,specificity 85.9%,positive predictive value 0.378,and negative predictive value 0.963.For diagnosing pre-diabetes,the AUC was 0.681(95%CI,0.666-0.697),and when the largest Youden index 0.280 was meeting,the optimal cut-off for HbAlc was 5.9%,with sensitivity 64.5%,specificity 63.5%,positive predictive value 0.317,and negative predictive value 0.867.Using the new cut-off point of HbA1c,the number of patients with diabetes increased from 880(10.68%)to 1,919(23.29%),the number of patients with pre-diabetes increased from 1,564(18.98%)to 2,696(32.72%),and the normal glucose metabolism population increased from 5,795(70.34%)declined to 3,624(43.99%).The average ?-cell function index of the HbAlc>5.9%and<6.3%group was slightly higher than that of the HbA1c<5.9%group,while the average ?-cell function index of the HbA1c?6.3%group was significantly lower than the other two groups.The average insulin resistance index of the three groups showed a gradual increase.ConclusionsAn HbAlc threshold of 6.3%was highly valuable for diagnosing newly diagnosed diabetes,and 5.9%was weakly valuable for diagnosing pre-diabetes in community-based Chinese adults aged 40 years or older.Part 3 Predictive value of different pre-diabetic diagnostic indicators for the risk of progressing to diabetes among Chinese adultsObjectiveAn HbA1c value of>6.5%has been adopted for the diagnosis of diabetes in the majority countries,according to the clinical practice recommendations from ADA,EASD,and IDF,partly based on the association of HbAlc with retinopathy.However,the use of HbA1c values between 5.7-6.4%for diagnosing pre-diabetes is not widely accepted because the adequacy of HbA1c as a reliable diagnostic tool for pre-diabetes remains under debate,and the cut-off points vary by race.Our previous cross-sectional study proposed an HbAlc cut-off point of 6.3%to diagnose diabetes and 5.9%to diagnose pre-diabetes in Chinese adults.However,that study provided insufficient evidence regarding the utility of HbAlc to predict future risks of diabetes.In this article,we performed a retrospective cohort study to estimate the annual incidence rate of diabetes and to assess the ability of HbAl c to predict progression to diabetes in Chinese adults aged 40 years or older.MethodsThe present study was one component of the baseline and 3-year follow-up surveys conducted for the Risk Evaluation of cAncers in Chinese diabeTic Individuals:a lONgitudinal(REACTION)study.In all,2,778 non-diabetic subjects(including 1,901 women)underwent HbAlc testing and OGTT at baseline and after 3 years.Diabetes and pre-diabetes were defined using the WHO 1999 criteria.Different HbAlc cut-off points were evaluated to predict the future risks of diabetes.The correlation between FPG,2hPG and HbAlc levels and the risk of progressing to diabetes after 3 years was analyzed using a fitted smooth curve.Relative risk(RR)was calculated via chi-square test.The area under the receiver operating characteristic(ROC)curve(AUC)was utilized to evaluate the predictive efficiency of FPG,2hPG and HbAlc for progression to diabetes.A superior cut-off point was defined as the point on the ROC curve with a larger Youden index.ResultsOverall,7.53%(210/2,778)of subjects progressed to diabetes,yielding an annual 2.51%diabetes incidence rate.Additionally,4.49%(100/2,227)of subjects with normal glucose tolerance(NGT)and 19.61%(110/561)of subjects with pre-diabetes progressed to diabetes,and the relative risk of progression to diabetes was 5.188 times higher in subjects with pre-diabetes than in subjects with NGT(p<0.001),after ages and genders being adjusted.Compared to subjects with HbAlc values ?5.6%,the RRs of progression to diabetes in subjects whose HbA1c ranged from 5.7 to 5.8%,5.9 to 6.2%,6.3 to 6.4%,and>6.5%were 1.057?2.281?4.976 and 14.495,respectively.However,the RRs for subjects with HbAlc ranging from 5.7 to 5.8%and those with HbAlc<5.6%showed no significant difference(p=0.855).FPG,2h-PG,HbAlc were all positively correlated with the risk of developing diabetes after 3 years by fitting smooth curve analysis.The AUCs for predicting diabetes after 3 years by FPG,2hPG and HbA1c were 0.752(95%CI,0.718-0.787),0.710(95%CI,0.671-0.748)and 0.756(95%CI,0.720-0.793),respectively.Both FPG value at 6.1 mmol/L and 2hPG value at 7.8 mmol/L exhibited lower sensitivity but higher specificity.The HbA1c cut-off point of 5.9%(sensitivity of 0.771 and specificity of 0.580)may better identify individuals at high risk of progression to diabetes than the 5.7%value(sensitivity of 0.862 and specificity of 0.371)due to the former's larger Youden index of 0.351,which exceeded the indices for FPG and 2hPG.ConclusionsThe use of HbAlc values>5.9%may provide greater accuracy in evaluating the risk of progression to diabetes and identify individuals with pre-diabetes with greater reliability among Chinese adults.Part 4 Nomogram for risk estimation in progression to diabetes after 3 years in Chinese adultsObjectiveAt present,the international prediabetic diagnosis standards are not uniform,leading to the greatly different prediabetic population based on different standards.How to include more people at high risk for future diabetes in a smaller pre-diabetes population is an issue we need to solve urgently.In this study,a cohort of Qilu Hospital Branch of Shandong University,an epidemiological study of the tumorigenic risk of type 2 diabetes patients in China(REACTION study),was used as the research object.By analyzing the risk factors for the onset of diabetes,a nomogram for predicting the risk of progression to diabetes was established,providing a quantitative predictive tool for early clinical screening of high-risk population of diabetes.MethodsThis study was one component of the baseline and 3-year follow-up surveys conducted for the REACTION study.In all,2,778 non-diabetic subjects(including 1,901 women)underwent HbAlc testing and OGTT at baseline and after 3 years.diabetes and pre-diabetes were defined via the WHO 1999 criteria.All participants were randomly divided into a modeling cohort(75%)and a validation cohort(25%).The modeling cohort was used to build a predictive model for assessing the risk of progress to diabetes,while the validation cohort was used to verify the prediction efficiency of the model we established in previous cohort.According to the Akaike information criterion,a logistic regression model was used to construct a nomogram that predicted the risk of diabetes.The ROC curve was utilized to evaluate the predictive value of the nomogram.Decision curve analysis was employed to asses the clinical utility of the predictive model.Statistical analysis was performed by two-sided Student-t test,with P<0.05 regarded as difference statistically significant.All statistical analyses were performed using the statistical software package R and Empower Stats.ResultsAfter 3 years of follow-up,210 of the 2,788 participants progressed to diabetes.This subset was composed of 150 of the 2,094 in modeling cohort and 60 of the 494 in validation cohort.In the modeling cohort,univariate logistic regression analysis showed that age,BMI,SBP,FPG,2hPG,HbAlc,Cr,LDL-c,TG,ALT,AST,and GGT were risk factors for diabetes,while female,HDL-c were protective.Multivariate Logistic regression analysis exhibited that only age,FPG,2hPG,HbA1c were independent risk factors for diabetes.We used the four factors above as independent variables to construct a nomogram that predicted the risk of progression to diabetes.The area under the ROC curve of the nomogram constructed in the modeling cohort was 0.825(95%CI,0.787-0.864),with the sensitivity of 68.00%,specificity of 86.32%,positive predictive value of 27.72%,negative predictive value of 97.22%,positive likelihood ratio of 4.970,and negative likelihood ratio of 0.371;the area under the ROC curve of the nomogram established above in the validation cohort was 0.781(95%Cl,0.716-0.846),with the sensitivity of 80.00%,specificity of 66.72%,positive predictive value of 18.53%,negative predictive value of 97.24%,positive likelihood ratio of 2.404,and negative likelihood ratio of 0.300.Decision curve analysis suggested that the nomogram constructed in this study had high clinical application value.ConclusionThe nomogram constructed in this study can effectively predict the risk of future progression to diabetes and be adopted to identify patients with high-risk of diabetes so that individualized interventions can target them.