Preoperative Risk Factors Analysis And Prediction Model Construction For In-hospital Death After Surgery For Stanford Type A Aaortic Dissection

ObjectiveTo explore the risk factors of in-hospital death after surgical repair by analyzing the preoperative clinical data of patients with Stanford type A aortic dissection(TAAD),construct a prediction model through logistic regression analysis and develop an individualized nomogram for validation,as well as predict the probability of in-hospital death of TAAD patients after surgical repair using the clinical data at admission and the constructed model,so as to provide a reliable basis for clinicians to develop perioperative diagnosis and treatment strategies and arrange perioperative management reasonably.MethodsTAAD patients undergoing surgery in the First Affiliated Hospital of Zhengzhou University from January 2013 to September 2019 were included in this study as subjects.Based on the inclusion and exclusion criteria,393 patients were finally selected.According to development group:validation group=6:4,250 patients were randomly selected to the development group(accounting for about 60%of the total data),and 143 to the validation group(accounting for about 40%of the total).In this study,the differences in observation indexes were analyzed using the t-test,nonparametric test(Mann-Whitney U test)or?2 test.The risk factors of in-hospital death in TAAD patients after surgery were screened using single-and multi-factor logistic regression analysis and a prediction model was constructed.The nomogram was drawn by R software.Through drawing receiver operating characteristic(ROC)curve,Hosmer-Lemeshow goodness of fit test and decision curve analysis,the constructed prediction model was validated and evaluated from three aspects of discrimination,calibration and clinical effectiveness.Result1.Differences in observation indexes between patients with in-hospital death and rehabilitation and discharge after TAAD surgery in development group:The t test,nonparametric test or?~2test showed statistically significant differences in age,history of hypertension,smoking history,leukocyte count,platelet count,neutrophil count,lymphocyte count,neutrophil-to-lymphocyte ratio(NLR),aspartate aminotransferase(AST),alanine aminotransferase(ALT),AST/ALT,albumin,creatinine,urea and plasma D-dimer(D-D)level between the two groups(P<0.05).No statistically significant differences were found in the constituent ratio of gender,history of diabetes or globulin between the two groups(P>0.05).2.Risk factors selected by single-and multi-factor logistic regression analysis:Single-factor logistic regression analysis presented that age,history of hypertension,smoking history,leukocyte count,platelet count,neutrophil count,lymphocyte count,AST,ALT,albumin,creatinine,urea,NLR,AST/ALT and plasma D-D level were statistically correlated with postoperative in-hospital death in TAAD patients.Multi-factor logistic regression analysis demonstrated that age,smoking history,platelet count,ALT and plasma D-D level were the risk factors of postoperative in-hospital death in TAAD patients.3.The logistic regression model was developed in the development group based on the following equation:P=1/(1+exp[-(-2.512+1.514×age+1.166×smoking history-1.297×platelet count+0.870×ALT+1.880×plasma D-D level)]).4.Validation and evaluation results of prediction model:The area under the ROC curve(AUC)for the risk of in-hospital death after TAAD surgery in the development group and the validation group was 0.868(95%CI:0.822?0.918)and0.899(95%CI:0.839?0.943),respectively.Hosmer-Lemeshow?~2was 6.689 and4.158,respectively,in the development group and the validation group(P=0.571 and0.843,respectively).The decision curve analysis of the prediction model showed that when the probability threshold was in the range of 0.4?0.8,the prediction results of the patients had high net income.Conclusions1.The individualized prediction model of in-hospital death after TAAD surgery improves the ability of early identification of patients with high risk of postoperative in-hospital death.2.In clinical work,close attention should be paid to the influences of age,smoking history,platelet count,ALT and plasma D-D level on in-hospital outcome of TAAD patients after surgery.3.The individualized prediction model of the risk of in-hospital death after TAAD surgery constructed in this study has high discrimination ability and consistency,as well as certain clinical practicability.