Establishment And Evaluation Of Male Coronary Heart Disease Risk Prediction Model On Laboratory Results

Objectives:To explore the correlation between laboratory results and male coronary heart diseases diagnosis,screen potential independent risk factors for male coronary heart diseases,and construct a risk prediction model.Methods:1.Selecting 359 male patients with coronary heart diseases(CHD)diagnosed in the Department of Cardiology,First Affiliated Hospital of Kunming Medical University,Yunnan Province from June 2018 to December 2018,age distribution 26-90 years old(average 60.35);359 healthy males who received a physical examination at the same time,aged 22-72 years old(average 40.42).The results of blood cell analysis was included in the study,including:white blood cell(WBC),Neutrophil%(NEUT%),Lymphocyte%(LYM%),absolute neutrophil count(NEUT#),absolute lymphocyte(LYM#),red blood cell count(RBC),hemoglobin(HGB),hematocrit(HCT),mean corpuscular volume(MCV),mean corpuscular hemoglobin(MCH),mean corpuscular hemoglobin concentration(MCHC),red blood cell volume distribution width(RDW),red blood cell volume distribution width-CV(RDW-CV),platelet count(PLT),platelet hematocrrit(PCT),mean platelet volume(MPV),platelet volume distribution width(PDW),platelet large cell ratio(PLR);And the results of biochemical indicators,including:total protein(TP),albumin(ALB),globulin(GLB),alanine aminotransferase(ALT),aspartate aminotransferase(AST),urea(Urea),creatinine(CRE),uric acid(UA),urea(Urea),glucose(Glucose,GLU);total cholesterol(TC),triglyceride(TG),High Density Lipoprotein cholesterol(HDL-C),Low Density Lipoprotein cholesterol(LDL-C)results.Observe all these variables distribution and characteristics;2.Analyze the internal correlation of each variable and the difference between the CHD group and the control group;3.Binary logistic multivariate regression analysis was used to screen the independent risk factors related to male coronary heart disease,and a predictive model of male coronary heart disease risk was constructed.Result1.Find ALB,GLB,HCT,HDL-C,HGB,LDL-C,LYM#,LYM%,MCH,MCHC,MCV,MPV,NEUT%,PCT,PLT,RBC,RDW-CV,TC,TP,UA,UREA,WBC basically conform to the normal distribution by histogram,K-S test,and trend PP map,ALT,AST,CRE,GLU,NEUT#,PDW,TG obviously do not conform to the normal distribution,so their subsequent analysis needs to perform logarithmic transformation;2.Pre-analytical stage found that InALT,MCHC,MPV,TC,UA,WBC,BNP and CHD were positive correlation,ALB,HCT,LDL-C,LYM%,NEUT%,lnPDW and CHD were negatively correlated.However,the relationship between these factors and CHD remains to be confirmed by a large number of epidemiological data;3.Through strict statistical analysis to exclude some variables,as far as possible to avoid confounding factors,select age,InAST,NEUT%,InPDW,TC as the male CHD independent correlation factors,while age,InAST,NEUT%was positively correlated with CHD,and lnPDW and TC were negatively correlated.Two kinds of prediction models were constructed,and the ROC curve evaluation model was used to verify the performance.The AUC of model 1 was 0.996,and the model 2 was 0.997.4.Establish a verification population and finally confirm that age,InAST,NEUT%,InPDW,and TC are independent correlation factors.The final prediction equation is:Probability(CHD-final)=1/{1+exp-(-6.511+0.229age+2.7441nAST+0.131NEUT%-6.714nPDW-0.82TC)}Conclusion:1.In blood cell analysis and clinical biochemical indicators,there are certain differences between the blood cell analysis and biochemical indicators in the CHD group and the control group;2.age?lnAST?NEUT%?lnPDW?TC may be the independent correlation factor of male coronary heart diseases.3.Through the statistical analysis of strict control,the common blood cell analysis and blood biochemical indicators can be properly screened to obtain the ideal CHD risk prediction model.