| Background and Objective:Due to the development of modern society,diabetes has become one of the biggest issues of public health worldwide and the prevalence is still rising.Prediabetes,usually referred to people whose glucose level is higher than normal but lower than diabetic level,is one of the most important risk factors for diabetes.Recently,the prevalence of prediabetes is rising as well.Now more than one third of Chinese population were prediabetic.According to previous researches,there are 5%-10%prediabetic population who proceed to diabetes each year.If no action is taken,it will pose a great challenge against the prevention and treatment of diabetes in China.There are numbers of researches,both domestic and international,shown that early intervention on prediabetic people could delay or reverse their progress to diabetes.Thus,recognizing and intervening people at risk with early identification and risk prediction would be essential.There is still less research on screening strategies and risk prediction for prediabetes in China,and applying foreign screening or prediction methods directly in China is usually less effective because of the differences in the population,more researches are needed to build better screening strategies and prediction models for the prediabetic population in China.In the same time,how to conduct large-scale intervention on prediabetic population in Chinese communities is also something researchers focus on.Based on National Key R&D Program,this study focused on prediabetic population,aiming to understand the epidemic and related risk factors of prediabetes and evaluate the effectiveness of lifestyle interventions in Jiangsu Province,China.Therefore,this study could help to explore appropriate screening strategies and intervention methods for prediabetic population in communities of Jiangsu Province,China.Methods:A stratified multi-stage sampling method was used among local residents aged 18-65 with no previous diagnosis of diabetes in Yandu District,Yancheng City,and Jurong City,Jiangsu Province.In the sampling process,three townships/streets were randomly selected in Yancheng and Jurong,and then within the jurisdiction of each township health clinics/community health service center eight to nine neighborhoods/villages were randomly selected.Finally,100 eligible villagers from different households were picked by simple random sampling within the selected neighborhoods/villages.General physical examination,questionnaires and laboratory tests were performed after informed consent was obtained.Those who met ADA criteria for a pre-diabetes diagnosis in the baseline survey were included in the lifestyle intervention cohort,which was divided into intervention group and control group by regions.The intervention group received lifestyle intervention which included exercise,diets and peer educations,and the control group received general health education.Baseline survey was conducted in 2017,then subjects at baseline were followed up in 2018.This study contained cross-sectional study and cohort study.In the cross-sectional study,Pearson correlation analysis and Bland-Altman plots were used to analyze the correlation of capillary glucose with venous plasma glucose.The discrimination ability of capillary glucose was evaluated using ROC curve and its AUC.In the cohort study,the effect of lifestyle interventions and prediction models for1-year risk of diabetes were evaluated using logistic regression and the calculation of OR.All statistical analyses were performed using SAS 9.4 software.Results:(1)A total of 3736 people were included in the screening study,of whom 1407 were men and 2320 were women.The mean age of the study population was 51.25±8.64 and the mean BMI was 25.48±3.54 kg/m~2.According to the WHO diagnostic criteria,there were 349 people with IFG and 965 people with IGT with a combined prevalence of prediabetes of 35.1%.Age and BMI of people with abnormal glucose metabolism was significantly higher than those with normal blood glucose levels.There was also significant difference in the prevalence of abnormal glucose metabolism between men and women.Capillary glucose was strongly positively correlated with venous plasma glucose.The Pearson coefficient between fasting capillary glucose and fasting venous plasma glucose was 0.75.The Pearson coefficient between 2hOGTT capillary glucose and 2hOGTT venous plasma glucose was 0.92.The AUC of fasting capillary glucose for screening IFG was 0.722,the optimal cut-point was 5.8,and the sensitivity and specificity were 0.691 and 0.635,respectively.The AUC of 2hOGTT capillary glucose for screening IGT was 0.916,the optimal cut-point was 8.2,and the sensitivity and specificity were 0.856 and 0.841,respectively.The discrimination ability of fasting capillary glucose and 2hOGTT capillary glucose are both better than model with only non-invasive variable.(2)There are a total of 2865 people included in the cohort study at baseline.1944 people were assigned to intervention group with a mean age of 52.15±8.95 and a mean BMI of 25.97±3.62 kg/m~2.921 people were assigned to intervention group with a mean age of 53.70±7.61 and a mean BMI of25.19±3.20 kg/m~2.A total of 2086 people were followed up in 2018.The follow-up rate was 72.8%.After excluding the subjects with missing values,a total of 2005 people were included in the analysis,of which 1405 people are in the intervention group,with 516(36.7%)having normal blood glucose at follow-up,and 600 people are in the control group,207(34.5%)having normal blood glucose at follow-up,no statistical difference was seen in conversion rate between the two groups.the decline in BMI,waist circumference and fasting plasma glucose levels was significantly higher in intervention group.After adjusting for other variables,the OR for intervention was 1.37(P=0.008),meaning that the intervention had led to some effect.In addition,females were more likely to reverse to normoglycemia than males,and those who were younger and had lower BMI and plasma glucose levels at baseline were more likely to reverse to normoglycemia.(3)A total of 2005 people were included in the risk prediction study,with a mean age of53.56±7.85 and a mean BMI of 25.78±3.46 kg/m~2.There are 165 new cases of diabetes with an incidence of 8.23%.70%of the study subjects were selected as the training set for building the risk prediction model,and another 30%were used to validate the effectiveness of the model.After adjusting for sex,age,BMI and intervention,the OR for fasting plasma glucose,2hOGTT plasma glucose and glycated hemoglobin were 4.349(95%CI 2.706-6.989),1.539(95%CI 1.340-1.744)and10.212(95%CI 5.638-18.496),respectively.The AUC for predicting 1-year risk of progress to diabetes for fasting plasma glucose,2hOGTT plasma glucose and glycated hemoglobin were 0.717(95%CI 0.668-0.767),0.712(95%CI 0.663-0.760)and 0.755(95%CI 0.712-0.798),respectively.The predictive effect of those indicators in combination was almost always better than that of using them alone.The optimal cut-point for fasting plasma glucose to predict 1-year diabetes risk was 6.0mmol/L with sensitivity and specificity of 0.617 and 0.668,respectively;the optimal cut-point for 2hOGTT plasma glucose was 7.0 mmol/L with sensitivity and specificity of 0.800 and 0.464,respectively;the optimal cut-point for glycated hemoglobin to predict diabetes risk was 5.5%with sensitivity and specificity of 0.826 and 0.528,respectively.Conclusion:(1)As a low-cost screening tool for prediabetes,capillary glucose is with good discrimination ability and is ideal for large-scale screening.(2)Comparing to general health education,lifestyle intervention could help prediabetic people reverse to normal glucose level.Age and weight loss is associated with the reversion in prediabetic people.(3)Fasting plasma glucose,2hOGTT plasma glucose and glycosylated hemoglobin could be used as predictors of 1-year diabetes risk with a good predictive effect,no matter they are used independently or together. |
PDF Full Text Request