The Tianjin Cohort Study Of Non-communicable Diseases And Establishing Prediction Models

ObjectiveWith the development of economy and the improvement of living standard,non-communicable diseases（chronic diseases）,caused by poor lifestyle,pose a serious threat to people’s health and social and economic development.Our research object in Tianjin,mostly are senior citizens,are in the stage of high rates of chronic disease,help us to study the causes of chronic disease,risk factors,which will provide a scientific basis for prevention and control of chronic diseases.1.The purpose of our research is to analyze risk factors of some chronic diseases including diabetes mellitus（DM）,hyperuricemia（HUA）,and metabolic syndrome（MS）.2.Investigate the cumulative incidence of MS,DM,and HUA in Tianjin through a prospective cohort study;Decipher independent risk factors of MS,DM,and HUA.3.Based on the cohort study,we built 5-year prediction models for type 2 diabetes （T2DM）and HUA.Methods1.Subjects in this study were selected from”the The Tianjin Cohort study of noncommunicable diseases”,which was established in 2006.By 2017,the total number of sunjects in the cohort was 18790（age 57.54±15.81 years）.In 2017, the number of participants was 12383（age 55.54±16.81 years）,prevalences of MS,DM,and HUA were calculated.Using multi-factor unconditioned Logistic regression analysis,we explored the independent risk factors for those diseases.2.In this study,all subjects were selected from the“The Tianjin Cohort study of noncommunicable diseases”.Follow-up checking were performed in 2007,2011, and 2015,respectively.A total of 1752 people,including 1535 people without the metabolic syndrome,1706 people without diabetes,and 1532 people without hyperuricemia,were included in this study with age of 60.05±12.63 years. Cumulative incidence of chronic diseases were calculated in the follow-up study.3.Three hundred and five（305）subjected were chosen for risk prediction from the “The Tianjin Cohort study of noncommunicable diseases”,none of them had T2DM or HUA in 2010.In the follow-up study in 2015,131 people still did not have T2DM,HUA,while 75 people had T2DM and 99 had HUA.Genotyping of the selected candidate genes were performed.This study is based on the cohort using unconditioned Logistic regression model,the OR value（β）natural logarithm weighted method,T2DM and HUA risk prediction equations were established for non-genetic factors,genetic factors,and combined factors, respectively.Result1.In 2017,prevalences of MS,DM,pre-diabetes,and HUA were 21.5%,13.1%,27.2%,and 19.4%,respectively in the cohort.2.We carried out multifactor non-conditional Logistic regression analyses for MS, DM,and HUA in the Tianjin cohort.The factors influencing metabolic syndrome were gender,age,proteinuria,serum uric acid,total protein,and alanine aminotransferase.The risk factors of diabetes are gender,age,BMI（Body Mass Index）,blood pressure,high cholesterol,high blood triglycerides,low high-density lipoprotein levels,blood urine nitrogen,estimated glomerular filtration rate（eGFR）,proteinuria,and alanine aminotransferase（ALT）.The risk factors of HUA were:blood urine nitrogen,total protein,ALT,hypertension, hypercholesterolemia,hypertriglyceridemia,and BMI;and protective factors wereage and eGFR.3.The present study found that the cumulative incidences of MS in the four-year period from 2007 to 2011 and from 2011 to 2015 were 10.6%and 21.6% respectively.For DM,were 6.8%and 6.0%;For HUA were 10.2%and4.7%, respectively.4.Cox regression was of MS performed on MS with male as the exposure factor （P=0.015,HR=1.287（1.048¹.580））.Cox regression was performed with the elderly（age≥60）as the exposure factor（P=0.001,HR=1.354（1.131¹.621））.Cox regression was performed with HUA as the exposure factor（P=0.001,HR=1.377（1.084¹.749））.5.Cox regression of DM was performed with the elderly as exposure factor,and the results were:P=0.000,and HR value was 1.877（1.415².491）;Cox regression of DM was performed with the obesity（BMI≥28kg/m²）as exposure factor,and the results were:P=0.000,and HR value was 1.840（1.326².553）.Cox regression was performed on DM in the presence of hypertriglyceridemia.Results:P=0.001, and the HR value was 1.540（1.184².004）.6.Cox regression of HUA was performed on the male as the exposure factor,and the results were:P=0.000 and the HR value was 3.122（2.120-4.598）;The results were:P=0.000,and the HR value was 1.791（1.299².468）.The results were:P=0.014,and the HR value was 1.391（1.068¹.810）.In order to determine whether hypertriglyceridemia was the exposure factor,the Cox regression was performed on HUA,and the results were:P=0.002,and the HR value was 1.480 （1.153¹.899）.7.T2DM risk prediction equation was established in the non-elderly population: logistP=-8.214+2.529×（0.077×rs5215Gi+0.068×rs972283Gi+0.322×rs10830963 Gi+0.266×rs163182Gi+0.199×rs13266634Gi+0.122×rs4402960Gi+0.322×rs1083 0963Gi+0.095X₁+0.630X₂+0.693X₃+0.683X₄+0.182X₅+1.030X₆）; AUC is 0.880（0.808⁰.951）,sensitivity is 96.2%,sepcificity is 73.7%. T2DM risk prediction equation was established in the elderly population:logistP=-5.845+1.39×（0.077×rs5215Gi+0.068×rs972283Gi+0.322×rs10830963Gi+0.266×rs163182Gi+0.095X₁+0.630X₂+0.693X₃+0.683X₅+1.030X₆）;AUC is 0.756（0.671⁰.842）,sensitivity is 81.6%,sepcificity is 62.0%.remark:X₁=gender,X₂=age,X₃=obesity,X₄=hypertension,X₅=hypertriglyceridemia,X₆=high-normal glucose8.HUA risk prediction equation was established in the non-elderly population:logistP=-1.016+3.131×（0.336X₁+0.119X₂+0.336X₃+0.345X₄+0.157X₅+1.435X₆）;AUC is 0.921（0.866⁰.975）,sensitivity is 97.4%,sepcificity is 77.8%.Establish the HUA risk prediction equation in the elderly population:logistP=-6.630+2.013×（0.747×rs505802Gi+0.18×rs606458Gi+0.04×rs2941484Gi+0.05×rs1967017Gi+0.336X₁+0.119X₂+0.336X₃+0.345X₄+1.435X₆）;AUC is 0.921（0.866⁰.975）,sensitivity is 97.4%,sepcificity is 77.8%.remark:X₁=gender,X₂=obesity,X₃=hypertension,X₄=hypertriglyceridemia,X₅=hypercholesteremia,X₆=high-normal glucose uric acid Conclusion1.The cumulative incidence of MS in the physical examination population of tianjin was higher from 2011 to 2015 than from 2007 to 2011,showing an upward trend.The cumulative incidence of DM and HUA was lower from 2011 to 2015 than from 2007 to 2011,showing a decreasing trend.2.The results of cohort study showed that gender,age and HUA were independent risk factors for MS.The age,obesity,and hypertriglyceridemia were independent risk factors for DM.Men,obesity,high blood pressure and high triglycerides were independent risk factors for HUA.The results showed that the above-mentioned chronic diseases have some common risk factors,chronic diseases are closely related,suggesting that there may have common pathogenic bases.3.We succeed in establishing nongenetic factors,genetic factors,and comprehensive prediction models for T2DM.Prediction value for the non-genetic model is better than the genetic model,and the comprehensive model yielded a better prediction value than either genetic and non-genetic models.The prediction values of the three models is non-elderly are higher than that of the elderly.4.We have successfully established a model of HUA risk prediction model for non-genetic factors among non-elderly people,and we successfully established three types of HUA prediction models of genetic factors,non-genetic factors,and comprehensive factors among the elderly.In the three models of the elderly,the non-genetic prediction effect is stronger than that of genetic factors,and the prediction effect of the comprehensive factor model is higher than that of the pure genetic and non-genetic prediction.Compared with other domestic and foreign models,our prediction models showed higher or at least equivalent sensitivity, specificity,and areas under ROC cruves.