Diagnosis And Warning Model Of Acute Aortic Dissection

Obiectives : Acute aortic dissection(AD)is a severe cardiovascular disease demonstrating the characteristics of acute onset,rapid development and high mortality.Early diagnosis and reasonable treatment is the key to reduce the mortality of patients with acute aortic dissection.Since be reported for the first time by Morgagni in 1761,The diagnosis of acute aortic dissection makes unceasing progress,but the clinical misdiagnosis rate remains as high as 39%.At present,the diagnosis of acute aortic dissection is mainly based on the single pattern of imaging examination,and lacks of laboratory tests to assist diagnosis.As the most commonly used clinical test projects,liver and kidney function,blood routine,blood coagulation function and other laboratory items are fast and cheap,and has been widely spread in primary care centers.What's more,blood lipids,white blood cell count,neutrophil count,platelet count,D-dimer,fibrinogen and so many indicators have been confirmed to change significantly in blood of patients with acute aortic dissection by lots of studies.The purpose of this research was to determine whether laboratory items and part meidcal history information allow to be used in early diagnosis and evaluation of AAD sufferer.Methods: This study is divided into two parts: model building and model evaluating.Model building: According to the inclusion and exclusion criteria,this part involved 392 patients,including 230 AAD patients(including 175 survivor in the acute stage and 55 death in the acute stage;113 Stanford A patents and 113 Stanford B patents)and 162 controls(including 55 patients with AMI,39 patients with APE,68 patients with AAA).Firstly,laboratory items and part meidcal history information were univariate analyzed between AAA group and different control group,survivor group and death group,Stanford A group and Stanford B group.Then,variables with significant differences were included in the multivariate study,and determine the regression coefficients,standard errors,Wald chi square values,P values,their corresponding OR values and their 95% confidence intervals for each variable.Put the results of multivariate analysis into the nomogram function and make the nomograms of the control groups and subgroups by R software,which are diagnostic model between AAD and AMI(DMDI),diagnostic model between AAD and APE(DMDE),diagnostic model between AAD and AAA(DMDA),diagnostic model between Stanford A AAD and Stanford B AAD(DMAB),modle of death risk evaluation for AAD sufferer(MDRE).Model evaluating : According to the inclusion and exclusion criteria,this part involved 159 patients,including 60 AAD patients(including 34 survivor in the acute stage and 26 death in the acute stage;32 Stanford A patents and 28 Stanford B patents)and 99 controls(including 36 patients with AMI,20 patients with APE,43 patients with AAA).Put the variables of different group and subgroup into the five modles.Then calculate the agreement ratio(Ar),sensitivity(Se),specificity(Sp)of the modles according to the diagnostic test.Results:(1)The study obtained the DMDI,which includes seven variables: Age,HDL,PLT,D-Dimer,FIB,Neutrophil to Lymphocyte ratio(NLR),Triglyceride/HDL(TG/HDL).The Ar=80.49%,Se=74.29%,Sp=85.11%.(2)The study obtained the DMDE,which includes six variables: Age,Smoke,WBC,Mono,Gran,PT.The Ar=66.18%,Se=61.90%,Sp=68.09%.(3)The study obtained the DMDA,which includes five variables: Age,Cholesterol,WBC,Triglyceride,LDL,Gran.The Ar=78.31%,Se=72.22%,Sp=82.98%.(4)The study obtained the DMAB,which includes five variables: Age,Gran,D-Dimer,FIB,APTT.The Ar=72.34%,Se=76.19%,Sp=69.23%.(5)The study obtained the MDRE,which includes fivevariables:WBC,Neut_Ratio,PLT,D-Dimer,Age.The Ar=76.59%,Se=78.13%,Sp=73.33%.Conclusion: This five modles,DMDI,DMDE,DMDA,DMAB,MDRE can be five convenient and efficient tools,which can assist doctor to diagnose AAD and evaluate the death risk AAD sufferer.