Construction and validation of a nomogram to predict in-hospital mortality in the patients with acute myocardial infarction complicated by cardiogenic shock

Song Lin; He Quan

doi:10.3969/j.issn.1002-1949.2025.08.005

Abstract：Objective To construct and validate risk prediction model of in-hospital mortality in the patients with acute myocardial infarction(AMI)complicated by cardiogenic shock(CS).Methods The 513 patients with AMI complicated by CS in the MIMIC-Ⅳ database were randomly divided into training set and validation set according to 6∶4 ratio,and 99 patients from the Cardiology Department in the First Affiliated Hospital of Chongqing Medical University from 2022 to 2023 was used for external validation set.The vital signs and laboratory test results within the first 24 h of admission were collected,retaining their maximum and minimum values.Variables for the model were selected by using best subset regression(BSR),and a predictive model and nomogram were subsequently developed.The predictive performance of the model was comprehensively assessed by the receiver operating characteristic(ROC)curves,calibration curves,and decision curve analysis(DCA).Results The prediction model included four variables:age,minimum systolic blood pressure,maximum blood urea nitrogen,and maximum lactate.The area under ROC curve(AUC)was 0.813 in the training set and 0.800 in internal validation set,showing a good discriminative ability.In the external validation set,the AUC was 0.816,which was higher than that of the existing the intra-aortic balloon pump in cardiogenic shock Ⅱ(IABP-SHOCK Ⅱ)score and CardShock score,but the difference was not statistically significant.In addition,calibration curve indicated that the predictive values of the model were highly consistent with the actual results,DCA further confirmed the potential value of the model in clinical applications.Conclusions Our model,requiring only four easily accessible variables,provides a simple and effective tool for predicting in-hospital mortality in the patients with AMI complicated by CS.