| Objective:1.Based on the cohort study data of Uygur population in rural areas of Xinjiang,the factor model of metabolic syndrome(MS)was determined,the internal relationship circumference between multiple metabolic abnormalities of MS was analyzed,and the potential influencing factors and aggregation characteristics of Uygur MS in Xinjiang were explored;2.Using MS factor as the main factor,the risk prediction model of cardiovascular diseases(CVD)was constructed to estimate the risk of CVD in Xinjiang Uygur population,so as to provide reference basis for the prevention and treatment of CVD in this population;3.Using MS factors related to CVD onset to draw a 5-year onset risk nomogram to assess individual onset risk;4.Compare the CVD risk prediction model constructed by MS factor and original variable,and evaluate the feasibility of constructing CVD risk prediction model with MS factor.Methods:1.Using the information from 2010 to 2012 in Jiashi County,Xinjiang and the 2016 information of the51 st Regiment of the Third Division as the baseline data,we conducted 2 follow-up visits to the research site in Jiashi County in 2013 and 2017,and in 2019,2020,and 2021.Taking the information of Jiashi county from 2010 to 2012 and the 51 st regiment of the third division in 2016 as the baseline data,the research site of Jiashi county was followed up twice in 2013 and 2017,and the research site of 51 st regiment was followed up three times in 2019,2020 and 2021.The follow-up content was the same as the baseline,mainly focusing on the collection of outcome events.At the same time,in order to ensure accuracy,social security data and hospitalization information of Jiashi county from 2010 to2017 and 51 regiment from 2016 to 2021 were continuously collected as outcome data;2.Taking the population(10160 people)followed up for 5 years in the cohort study as the research object,selected training samples(6708 people)and validation samples(3452 people)by random method.Based on the method of factor analysis,2522 people with MS at the baseline of training samples were analyzed to determine the aggregation mode of MS in Xinjiang Uygur;3.Cox regression was used to construct CVD risk prediction model based on training sample MS factor.The samples were verified,and ROC curve and calibration curve were used to evaluate the area and calibration of the model;4.Using the predictive factors related to CVD risk extracted from the training sample Cox regression,the 5-year CVD risk nomogram can be used to quickly estimate the individual risk;5.Based on the training samples,the MS factor and the original variable were used to build a CVD risk prediction model,and the ROC curve was drawn to compare the area under the curve,and to evaluate the feasibility of using the MS factor to build a CVD risk prediction model.Results:1.This study included 10,160 Uyghurs who were followed up for 5 years.The incidence of CVD was8.78%(6.84% in males and 10.79% in females),which was higher in females than in males(P<0.001).Males were higher than females in age,waist circumference,hip circumference,SBP,DBP,TC,TG,FPG,Scr,UA,AST,ALT,LDH,α-HBDH,TBIL,IBIL,smoking and drinking,and the difference was statistically significant.The incidence of CVD in the training sample was 8.80%(6.58% in men and11.05% in women),which was higher in women than in men(P<0.001).The incidence of CVD in the validation sample was 8.75%(7.32% in men,10.28% in women)compared with men(P<0.001).In training samples and validation samples,males were higher than females in age,waist circumference,hip circumference,SBP,DBP,TG,FPG,Scr,UA,AST,ALT,α-HBDH,TBIL,IBIL,smoking,and drinking.2.Both male and female MS are explained by seven factors: Obesity factor(BMI,waist circumference,hip circumference),liver function factor(TBIL,ibil),myocardial enzyme factor(LDH,α-HBDH),liver enzyme factors(AST,ALT),blood pressure factors(SBP,DBP),blood lipid and blood glucose factors(TC,TG,FPG)and renal metabolic factors(SCR,UA).The contribution rates of male factors were 18.858%,12.279%,11.234%,10.017%,8.515%,7.468% and 6.288% respectively.The contribution rates of female factors were 18.677%,14.506%,10.708%,8.464%,9.721%,7.109% and7.836% respectively.3.Cox regression was used to construct a CVD risk prediction model based on the extracted 7 MS factors.Combined with follow-up CVD outcome data,a ROC curve was drawn to evaluate model discrimination.The area under the ROC curve(AUC)for male training samples was 0.775(95%CI:0.744-0.805),and the AUC for female training samples was 0.762(95%CI: 0.736-0.788).A calibration curve was drawn to evaluate the calibration degree of the CVD risk prediction model.The male and female consistency indices were 0.778(0.748-0.808)and 0.761(0.736-0.787),respectively.The results of different genders showed that the predicted values of the two calibration curves were consistent with the actual values.And the two lines are close.4.The male validation sample AUC was 0.766(95%CI: 0.711-0.822)and the female validation sample AUC was 0.774(95%CI: 0.742-0.805).The calibration degree of the CVD risk prediction model was evaluated by the calibration curve,and the consistency indices for men and women were 0.762(0.706-0.818)and 0.751(0.718-0.781),respectively.The results showed that the predicted values and actual values of the two calibration curves of different gender validation samples were trending Consistent and the two lines are close.5.In this study,the predictors related to the risk of CVD were extracted by Cox regression of training samples.Men included age,alcohol consumption,obesity factors,myocardial enzyme factors,and renal metabolic factors into the 5-year CVD risk nomogram,and women included age,obesity factors,liver enzyme factors,and renal metabolic factors were included in the 5-year CVD risk nomogram to assess individual risk.6.The Cox stepwise regression method was used to analyze the original variables of the training samples and the main factor data of MS,and the prediction models of CVD incidence risk were constructed respectively.The results show that the male training sample MS main factor AUC is 0.770(95%CI:0.739-0.801)higher than the original variable model AUC is 0.705(95%CI: 0.670-0.739)(Z=2.708,P=0.007);female training sample The main factor AUC of MS was 0.756(95%CI: 0.731-0.784),which was higher than the original variable model AUC of 0.716(95%CI: 0.689-0.743)(Z=2.211,P=0.039).Conclusions :1.CVD incidence rate and MS prevalence rate of Uygur in Xinjiang were higher than that of the national level,and women were higher than men.2.The main factors of MS population of different genders were extracted by factor analysis.Both male and female were obesity factors(BMI,waist circumference,hip circumference),liver function factors(TBIL,IBIL),myocardial enzyme factors(LDH,α-HBDH),liver enzymes Factors(AST,ALT),blood pressure factors(SBP,DBP),blood lipid and blood sugar factors(TC,TG,FPG),renal metabolic factors(Scr,UA),the cumulative variance contribution rates of males and females were 74.659% and77.021%,respectively.3.The Uyghur CVD risk prediction model was constructed based on the MS factor of the training samples.The ROC curve and the calibration curve were used to evaluate the discrimination and calibration of the model,and the validation samples were used to verify that the constructed model was suitable for the prediction of CVD in this population.4.The nomogram is drawn with the predictors related to CVD risk extracted by Cox regression,and individual CVD risk can be obtained. |
PDF Full Text Request