| BackgroundThe metabolic syndrome is a kind of clinical syndrome,which is composed of a series of metabolic disorders.Metabolic syndrome mainly includes abdominal obesity,raised fasting plasma glucose,raised blood pressure,raised triglycerides and reduced high-density lipoprotein cholesterol.Metabolic syndrome is widespread in the world,and has a trend towards younger people.With the vigorous development of our economy,the process of urbanization accelerated and the aging of the population intensified,our diet and lifestyle have undergone tremendous changes.At the same time,the incidence of metabolic syndrome is also increasing yearly.Metabolic syndrome is extremely harmful to human health.Metabolic syndrome is not only characterized by the simultaneous aggregation of multiple diseases and complex clinical manifestations,which can bring great challenges to treatment,but also can increase the risk of cardio-cerebrovascular disease.Metabolic syndrome has become an increasingly prevalant global public health concern.Because of the serious harm to human health,the difficulties of radical treatment,as well as the heavy medical burden to individuals,families and society caused by the metabolic syndrome,it is very important to find the risk factors of MS as soon as possible and carry out health management to prevent the occurrence of MS according to the risk factors.At the same time,it is also of great significance to diagnose and treat early in high-risk groups.The relevant studies previously were based on the entire population,lacking of pertinence.This study was based on the high-risk groups of MS—people with phlegm dampness constitution.This study screened out the risk factors and established three diagnostic models of people with phlegm dampness constitution who were susceptible to metabolic syndrome,taking into account the advantages of various models.The results of this study provided a scientific basis and practical tools for achieving the tertiary prevention of MS,which includes early prevention,diagnosis and treatment,and led a new thinking and method for the realization of a certain sense of precise medical treatment.ObjectiveTo provide a scientific basis of health management programs in persons at high risk and practical tools of disease screening for achieving the tertiary prevention of MS,which includes early prevention,diagnosis and treatment,by screening the risk factors and establishing the diagnostic models of people with phlegm dampness constitution who were susceptible to metabolic syndrome.Method1.A case-controlled study was conducted.758 people with phlegm-dampness constitution were selected as the objects of the study.There were 240 patients with metabolic syndrome in case group and 518 cases without metabolic syndrome in control group.All of them were investigated by questionnaire,scale measurement,physical measurement and laboratory biochemical tests.Using EpiData3.1 for database establishment and SAS9.3 for statistical analysis.Univariate and multivariate Logistic regression analysis was used to select the risk factors of people with phlegm dampness constitution who were susceptible to metabolic syndrome.Variables which were statistically significant(P<0.05)were risk factors for people with phlegm dampness constitution who were susceptible to metabolic syndrome.2.Based on the above data and statistical analysis results,SAS9.3 and Ri386 3.3.2 were used to build the diagnostic models.The diagnostic models for people with phlegm dampness constitution who were susceptible to metabolic syndrome were developed,based on multivariate Logistic stepwise regression analysis,discriminant analysis,support vector machine(SVM)and a set of variables not requiring laboratory tests.According to the characteristics of the models,the performance of the diagnostic models were evaluated by different methods.Result1.Basic situation of the research objectsThe study included 758 subjects of phlegm dampness constitution,in which 240 patients with metabolic syndrome(31.66%)and 518 cases without metabolic syndrome(68.34%).There were 312 males(41.16%)and 446 females(58.84%)with an average age of 42.14 ± 10.50 years old.2.Screening risk factors of people with phlegm dampness constitution who were susceptible to metabolic syndromeUnivariate Logistic regression analysis showed that 33 variables were related to the occurrence of metabolic syndrome,and the results were as followed:28 variables such as region,gender,age,body weight,body mass index(BMI),waist circumference,hip circumference,hypertension,percentages of body fat,back subcutaneous fat thickness,abdominal subcutaneous fat thickness,fasting plasma glucose(FPG),fasting insulin(FINS),insulin resistance index(HOMA-IR),glycosylated hemoglobin(HbA1c),total cholesterol(TC),low-density lipoprotein cholesterol(LDL-C),triglyceride(TG),blood uric acid(BUC),family history of obesity,family history of diabetes,family history of hypertension,bedtime snacks,meat preference,smoking habits,drinking habits,sleeping habits,phlegm dampness constitution score were the risk factors.5 variables such as high-density lipoprotein cholesterol(HDL-C),daily feeding frequency,regular diet,exercise habits and exercise intensity were the protective factors.The differences were statistically significant(P<0.05).Multivariate Logistic regression analysis showed that 9 variables were related to the occurrence of metabolic syndrome,and the results were as followed in decreasing order:7 variables such as hypertension(OR=27.96,95%CI:14.22～54.97),triglyceride(OR=3.47,95%CI:2.56～4.69),fasting plasma glucose(OR= 1.95,95%CI:1.57～2.43),family history of hypertension(OR=1.88,95%CI:1.08～3.26),waist circumference(OR=1.07,95%CI:1.03～1.10),percentages of body fat(OR=1.05,95%CI:1.02～1.09),phlegm dampness constitution score(OR=1.04,95%CI:1.00-1.07)were the risk factors.Daily meal frequency(OR=0.30,95%CI:0.15-0.63)and high-density lipoprotein cholesterol(OR=0.01,95%CI:0.00～0.04)were the protective factors.The differences were statistically significant(P<0.05).3.Establishment of diagnostic models of people with phlegm dampness constitution who were susceptible to metabolic syndromeVariables included in the diagnostic models were as followed:Hypertension(X9),waist circumference(X10),family history of hypertension(X27),daily meal frequency(X34),phlegm dampness constitution score(X47).(1)Logistic regression modelAll data model:Model evaluation:Area under ROC curve of the Logistic regression model was 0.8655.Accuracy of self-validating of Logistic regression model was 81.53%,the sensitivity was 64.17%,the specificity was 89.58%。80%data model:Model evaluation:Area under ROC curve of the Logistic regression model was 0.8528.Accuracy of test set of Logistic regression model,which included 20%data was 84.11%,the sensitivity was 65.38%,the specificity was 93.94%.(2)Bayes Discriminant Analysis ModelAll data model:Model evaluation:Accuracy of self-validating of discriminant analysis model was 80.87%,the sensitivity was 65.42%,the specificity was 88.03%.Accuracy of jackknife method of discriminant analysis model was 79.95%,the sensitivity was 63.75%,the specificity was 87.45%.80%data model:Model evaluation:Accuracy of self-validating of discriminant analysis model was 81.05%,the sensitivity was 65.96%,the specificity was 87.83%.Accuracy of jackknife method of discriminant analysis model was 79.41%,the sensitivity was 63.30%,the specificity was 86.63%.Accuracy of test set of discriminant analysis model,which included 20%data was 82.12%,the sensitivity was 65.38%,the specificity was 90.91%.(3)Support vector machine modelSupport vector machine model does not output function.Model evaluation:Accuracy of self-validating of all data model was 87.47%,the sensitivity was 79.17%,the specificity was 91.31%.Accuracy of self-validating of 80%data model was 87.97%,the sensitivity was 79.79%,the specificity was 91.65%.Accuracy of test set of 80%data model,which included 20%data was 80.13%,the sensitivity was 63.46%,the specificity was 88.89%.Conclusion1.Hypertension,triglyceride,fasting plasma glucose,family history of hypertension,waist circumference,percentages of body fat and phlegm dampness constitution score were the risk factors of people with phlegm dampness constitution who were susceptible to metabolic syndrome.Daily meal frequency and high-density lipoprotein cholesterol were the protective factors of people with phlegm dampness constitution who were susceptible to metabolic syndrome.2.Based on Logistic regression analysis,Bayes discriminant analysis and support vector machine,the diagnosis of people with phlegm dampness constitution who were susceptible to metabolic syndrome could be realized.The models were simple,inexpensive,reliable,effective and noninvasive. |
PDF Full Text Request