Prediction Model For Acute Myocardial Infarction Complicated With Venous Thromboembolism: Development And Validation

Objective:To investigate the risk factors of acute coronary syndrome(ACS)patients complicated with Venous Thromboembolism(VTE)by analysing the demographic characteristics and laboratory inspections,and to establish and evaluate a predictive model for ACS complicated with VTE.Methods:A retrospective analysis of 232 patients with ACS complicated VTE admitted to The First Hospital of Jilin University from January 2013 to June 2018 was set as an observation group.At the same time,406 patients(406/15763,3%)with ACS but without VTE were randomly selected as the control group.We collected the clinical features(sex,Age,Basic diseases,etc),and test results(blood routine,BNP,D-dimer,etc)of the two groups,then screened variables that are significant for predicting VTE and established a risk prediction model by logistic regression.Finally,we evaluated the prediction ability by comparing this risk prediction model with the Wells score.Results:(1)About 1.9% of the ACS patients were complicated with VTE(295/15763).(2)Univariate analysis showed that elder(≥60 years),female,tumor history,cerebral infarction history,long-term bed rest,malignant tumor,definite coronary artery disease(CAD)history,VTE history,D-dimer,BNP were positively correlated with VTE(P<0.05).Smoking history and alcohol were negatively correlated with VTE(P<0.05).(3)Multivariate logistic regression analysis showed that the elder(≥60 years old),female,smoking history,cerebral infarction history,definite CAD history,VTE history,D-dimer,BNP were independent predictors for VTE.(4)The risk prediction odel contains six variables:elder(95%CI:2.153-6.489,P<0.001),female(95%CI :1.099-2.628,P=0.017),smoking(95%CI : 0.289-0.712,P=0.001),definitive CAD history(95%CI:2.539-7.424,P<0.001),history of VTE(95%CI:2.391-41.252,P=0.002),D-dimer(95%CI:5.447-13.022,P<0.001).(5)After weights,the predicted model has a score range of-1 ～ 14,and the ACS patients can be divided into low-risk group(score <3),middle-risk group(score 3-5)and high-risk group(score >5).The sensitivity of the risk score model was 84.9,the specificity was 72.2.The positive predictive value(probability of VTE in high risk)is 8.4,the negative predictive value(probability of without VTE in low risk)is 97.4,and the area under the ROC curve is 0.863,greater than Wells score.Conclusion: The incidence of VTE in ACS patients is about 1.9%.We identified independent predictors of VTE in ACS patients by regression analysis,including elder(≥60 years old),female,smoking history,cerebral infarction history,definite CAD history,VTE history,D-dimer,BNP.The prediction model established with the six variables: elderly(≥60 years old),female,smoking history,definite CAD history,VTE history,D-dimer,could be used to screen patients complicated with VTE,and may provide an effective means for early warning of adverse events.The risk prediction model established in this study is superior to the Wells score in predicting VTE in ACS patients.