Contributing Factors Analysis And Risk Prediction Model Establishment Of In-Hospital Mortality In Patients With Acute Myocardial Infarction Aged ≥60 Years

ObjectiveThe eICU database was used to study the contributing factors of in-hospital mortality in patients with acute myocardial infarction(AMI)aged≥60 years and to build a model to predict the risk of in-hospital mortality.MethodsThe eICU database was used to screen out patients with AMI.The patients’ clinical data(including basic patient information,clinical characteristics,laboratory examination results,etc.)are extracted from the database.A total of 6922 patients were included according to inclusion and exclusion criteria.Based on the admission time to the study,patients were divided into a training cohort(3379 patients admitted in 2014)to construct a risk prediction model and a validation cohort(3543 patients admitted in 2015)to verify the validity of the model.The primary endpoint was in-hospital mortality,which was defined as all-cause mortality during hospitalization.SPSS26.0 software was used to analyze the clinical data of patients in the training cohort to study the risk factors of in-hospital mortality of AMI patients aged>60 years old.Multivariate logistic analysis was used to determine the independent risk factors and build a risk prediction model.The model can be visualized by drawing nomogram for clinical application.The predictive ability of the model was verified by internal and external validation cohorts(including MIMIC-III database and data from Qilu Hospital of Shandong University).ResultsIn-hospital mortality was 12.3%in the training cohort(n=3379)and 11.5%in the validation cohort(n=3543).In this study,10 variables were identified as independent factors for the risk of in-hospital death in AMI patients aged≥60 years,5 of which were associated with a higher risk of death in hospital:age[OR(95%CI),2.036(1.611-2.573)],heart rate[OR(95%CI),1.015(1.010-1.021)],respiratory rate[OR(95%CI),1.026(1.008-1.045)],white blood cell count[OR(95%CI),1.017(1.002-1.032)],serum creatinine[OR(95%CI),1.223(1.125-1.329)].However,body temperature[OR(95%CI),0.703(0.606-0.815)],systolic blood pressure[OR(95%CI),0.988(0.982-0.994)],oxygen saturation[OR(95%CI),0.981(0.9660.997)],serum albumin[OR(95%CI),0.494(0.400-0.610)]and bicarbonate[OR(95%CI),0.965(0.941-0.990)]were associated with lower risk of death during hospitalization.In the training cohort,the AUC of the risk prediction model constructed based on the above variables was 0.788(95%CI,0.765-0.811),and Hosmer-Lemeshow P>0.05,indicating that the model was well calibrated and had good forecasting ability.In the internal validation cohort,the AUC of the model was 0.773(95%CI,0.750-0.796).In the two external validation cohorts,the AUC of MIMIC-III database and date from Qilu Hospital of Shandong University were 0.697(95%CI,0.659-0.736)and 0.843(95%CI,0.731-0.954),respectively.ConclusionAge,body temperature,respiratory rate,heart rate,systolic blood pressure,blood oxygen saturation,white blood cell count,serum creatinine,albumin and bicarbonate were independent risk factors for in-hospital mortality in patients with AMI aged≥60 years.Age,heart rate,respiratory rate,white blood cell count and serum creatinine were risk factors,while body temperature,systolic blood pressure,blood oxygen saturation,albumin and bicarbonate were protective factors.The risk prediction model constructed based on the above variables is well calibrated and has good predictive ability(AUC=0.788),and also showed good discriminability in both internal cohort and external validation cohort.The nomogram can be used in clinic and provide reference for clinical decision making.