A prediction model for the risk of developing HFpEF during hospitalization in patients with acute myocardial infarction.

Aims: This study aimed to establish a clinical prediction model for evaluating the risk of heart failure (HF) with preserved ejection fraction (HFpEF) occurrence in patients with acute myocardial infarction (AMI). Methods: We selected AMI patients with left ventricular ejection fraction ≥ 50% without a history of HF from the MIMIC-III database (training cohort) and those from The First Affiliated Hospital of Jinan University (validation cohort). Eligibility criteria for identifying patients who develop HFpEF were based on ICD-9 diagnosis codes in the training cohort and the 2021 European Society of Cardiology HF Guidelines in the external validation cohort. The LASSO-multivariate logistic regression analysis was performed to identify the predictive factors for establishing the nomogram prediction model. Then, this model was evaluated by discrimination, calibration, clinical effectiveness, and internal validation. Finally, using the validation cohort, the model was externally validated and updated. Results: 201 patients were included in the training cohort and 65 in the validation cohort. Older age, higher heart rate, higher neutrophil-to-lymphocyte ratio, and lower total bilirubin were identified as independent predictors. The area under the receiver operator characteristic curve (AUC) of the model and the bootstrap validation were 0.823 and 0.804, respectively. Both the prediction model's and the bootstrap validation's calibration curve were in good agreement with the standard reference line, and the Hosmer-Lemeshow test showed Conclusion: Our prediction model provides an additional perspective into the early identification of high-risk patients for HFpEF occurrence following AMI.