| Metabolic syndrome (MetS) refers to pathological state of carbohydrates, protein, fat and other substances, and it is a complex metabolic disorder syndrome. Metabolic syndrome is one of the major diseases affecting the health of the population. Because of lacking standards to risk assessment and scientific personalized intervention prescription in MetS health management, it is difficult to realize truly personalized management.Thus, the health management of MetS can only be determined by a physician based on medical laboratory test, proposing health intervention guidance subjectively and generally. To achieve the target of truly personalized health management, this paper based on the research group "Shandong Province Medical Association Health Management and Health Insurance Professional Committee" building "a large multi-center longitudinal monitoring health management cohort", and then applied Generalized estimating equations and Mixed effects models to explore important health management biomarkers. We used exploratory factor analysis (EFA) to extract latent factor, and then use these potential factors as variables to construct a Cox regression model and to evaluate its effectiveness. Finally, We convert the model to two simple tools by using Framingham scoring method and risk assessment matrix.The results are as follows:1. In 2007～2014, The MetS prevalence rates of both male and female is have been fluctuating year by year, but basically a downward trend.2. In each the non-MetS and MetS crowd, there was significant difference between physical indicators.3. Indicators selected to build metabolic syndrome predictive model are different between male and female (male 17, female 19). Among them, the same indicators for male and female 12 (body mass index, systolic blood pressure, diastolic blood pressure, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, fasting glucose, mean hemoglobin content, mean platelet volume, monocytes and the absolute value of serum uric acid); In addition, male have five special indicators platelet count, percentage of cells basophils, total bilirubin, direct bilirubin and indirect bilirubin, while female have seven special indicators:white blood cell count, hemoglobin, absolute neutrophil count, Eosinophils Percentage, lymphocyte count, alanine aminotransferase and blood urea nitrogen.4. Extracted potential predictors by EFA are not all the same between male and female,9 in men and 10 in female. Male and female have the same 6 predictors (cholesterol factor, blood pressure factor, platelet-activating factor and lipid metabolism factor, glucose metabolism and blood viscosity factor factor); in addition, there are other three potential predictors in male:bilirubin male factor, leptin resistance factor and weight factor; there are other four potential predictors in female: inflammatory factor, renal function factor, liver function body quality factor and granulocyte factor.5. Application of constructed five-year Cox regression model, area under the ROC curve:male 0.762 (95% CI:0.742-0.782), female 0.763 (95% CI:0.734-0.792); external validation display:male AUC 0.753 (95% CI:0.686-0.812), female AUC 0.892 (95% CI:0.867-0.914).6. The application of Framingham scoring system and risk assessment matrix to convert the Cox regression model to two simple tools, scoring system and MetS risk assessment matrix to predict MetS.7. Applying MetS prediction model,79,151 individuals were evaluated:before 45-year-old, the percentage of high-risk individuals grow as line, and higher in male than female; after 45-year-old, the percentage of high-risk individuals is stable, and gradually lower than female.Conclusion:1. In the non-MetS and MetS crowd, there was significant difference in most physical indicators.2. The potential factor are different between male and female. Cox regression model for different gender have good prediction effect.3. Risk scoring system and risk assessment matrix can be easily and intuitively used in clinical practice. |
PDF Full Text Request