Development and validation of a risk score heatmap for in-hospital adverse events in older adults with acute myocardial infarction

Background Older adults with acute myocardial infarction (AMI) represent a particularly high-risk population for early in-hospital deterioration. Life-threatening complications—including sustained ventricular tachycardia or fibrillation (sVT/VF), cardiogenic shock (CS), and acute heart failure (AHF)—are major contributors to in-hospital adverse events (IAE), as these events may herald an increased risk of poor post-discharge prognosis. However, a simple and visually intuitive tool for early risk stratification in this population remains lacking. Methods We conducted a retrospective cohort study of consecutive patients aged 60 years with AMI admitted between January 2019 and December 2024. The primary endpoint was a composite of in-hospital death, sVT/VF, CS, or AHF. After multiple imputation for missing data, a multi-stage feature selection strategy incorporating least absolute shrinkage and selection operator (LASSO) logistic regression and random forest analysis was applied. Results Among 4,897 included patients, 434 (8.9%) experienced the composite endpoint. The final model incorporated 10 readily available clinical variables: age, atrial fibrillation (AF), pulse rate, systolic blood pressure (SBP), neutrophil-to-lymphocyte ratio (NLR), albumin, N-terminal pro-B-type natriuretic peptide (NT-proBNP), estimated glomerular filtration rate (eGFR), glucose, and alanine aminotransferase (ALT). The model demonstrated good discrimination area under the receiver operating characteristic curve (AUC) 0.808, 95% confidence interval (CI) 0.786–0.830; p 0.001 and satisfactory calibration. The risk score heatmap allows intuitive estimation of individual risk based on cumulative point scores. Conclusion We developed an internally validated 10-variable prediction model, visualized as a risk score heatmap, to estimate IAE in older adults with AMI. This intuitive tool may support early bedside risk stratification and personalized management in clinical practice.