Establishing A Predictive Model Of Coronary Artery Disease Risk Based On Biochemical Indexes

BACKGROUND&OBJECTIVE:Coronary artery disease?CAD?is the leading cause of death and disability in global diseases.The incidence and mortality of CAD in China are also ranked first,much higher than tumor and respiratory diseases.Therefore,prevention and treatment of CAD is an urgent task.Through the study of CAD risk factors,the relationship between biochemical indicators and CAD was explored.And the differences in composition among different gender groups were discussed.The risk prediction model for CAD was established to establish an effective reference strategy for CAD risk assessment.Methods:A total of 864 CAD cases?mean age 68.62 years?were diagnosed in the cardiology department,and 454?mean age 58.81 years?in the non-CAD group.After gender stratification,570 male patients?mean age 67.08 years?;female case group295?mean age 71.60 years?;male control group 208?mean age 58.37 years?,female control group 245?mean age 59.18 years?.Recorded age,systolic blood pressure?Sp?,diastolic blood pressure?Dp?,blood glucose?GLU?,hemoglobin?HGB?,red blood cell?RBC?,white blood cell?WBC?,platelet?PLT?,hematocrit?HCT?,total cholesterol?TC?,triglyceride?TG?,high-density lipoprotein cholesterol?HDL-C?,low-density lipoprotein cholesterol?LDL-C?,apolipoprotein A?ApoA?,apolipoprotein B?ApoB?,apolipoprotein E?ApoE?,C-reactive protein?CRP?,blood homocysteine?HCY?.Using SPSS 22.0,?1?Observing the distribution trend of all variables,logarithmically convert the indicators that without conforming the normal distribution;?2?Analyzing the differences between the cases and the controls in the total population,male population and female population by monofactor analysis;?3?Using Pearson correlation to analyze the internal correlation between variables;?4?Using logistic regression to establish CAD prediction models for correlation factors;?5?Using ROC and AUC verify the fit of the model.Results:?1?In the total population,age,HGB,LnTG were positively correlated with CAD;Sp,TC,LnApoE were negatively correlated with CAD.The predictive model for CAD:Probability?CAD?=1/{1+exp-?-1.297+0.049age-0.007Sp+0.012HGB-0.184TC+0.997LnTG-0.897LnApoE?},AUC was 0.807,with well fitting effect.?2?In male population,age,LnTG,LnHCT were positively relevanced with CAD;HGB and LnApoE were negatively associated with CAD,and male CAD predictive model:Probability(CAD_male)=1/{1+exp-?-18.110+0.042age-0.048HGB+6.572LnHCT+1.032LnTG-0.975LnApoE?},AUC was 0.776,with good fitting effect.?3?In female population,age,LnTG,LnGLU were positively correlated with CAD;Sp,LDL-C,LnCRP were negatively correlated with CAD,and female CAD predictive model:Probability(CAD_female)=1/{1+exp-?-3.070+0.058age-0.008Sp-0.476LDL-C+LnGLU+0.877LnTG-0.197LnCRP?},AUC was 0.845,with well fitting effect.