Development and verification of a predictive nomogram to evaluate the risk of complicating ventricular tachyarrhythmia after acute myocardial infarction during hospitalization: A retrospective analysis.

Abstract Purpose The purpose of this study was to establish a nomogram to predict the risk of complicating ventricular tachyarrhythmia (VTA) in patients with acute myocardial infarction (AMI) during hospitalization and to verify the accuracy of the model. Clinical information and method The authors enrolled the information of 503 patients who were diagnosed as AMI from January 2017 to December 2019. The cohort was randomly divided into a training set and a testing set at a ratio of 70%:30%. A total of 13 clinical indicators were screened by the least absolute shrinkage and selection operator (LASSO) regression and Boruta arithmetic independently in order to figure out the optimal feature variables. Multivariable logistic regression analysis was applied to establish the prediction model represented by a nomogram incorporating the selected feature variables. The performance of the nomogram was assessed by discrimination, calibration and clinical usefulness. C-Statistics with the area under the receiver operating characteristic curve (AUC), calibration curve and decision curve analysis were used to evaluate the identification ability, calibration and clinical practicability respectively. The prediction model was verified on the testing set to ensure its accuracy. Results Five feature variables as percutaneous coronary intervention (PCI) timing after hospitalization, ejection fraction (EF), high-sensitive troponin T (hsTnT) score, infection and estimated glomerular filtration rate (eGFR) were selected by both LASSO regression and Boruta arithmetic. C-statistics with AUC was 0.764 (95% confidence interval: 0.690–0.838) in the training set while a slight increasing to 0.804 (95% confidence interval: 0.673–0.935) in the testing set. Calibration curve illustrated that the predicted and actually diagnosis of VTA probabilities were satisfactory on both training set and testing validation. Decision curve analysis indicated that the nomogram can be used in clinical settings as it has a threshold of between 4% to 90% along with a net benefit. Conclusion The nomogram with five variables is practical to clinicians in estimating the risk of complicating VTA after AMI during hospitalization.