Influencing factors and construction of a prediction model for poor prognosis in patients with acute myocardial infarction complicated by heart failure

Objective To explore the influencing factors for poor prognosis in patients with acute myocardial infarction (AMI) complicated by heart failure (HF), construct a nomogram prediction model, and validate its performance.

Methods A total of 252 patients with AMI complicated by HF were selected as training set and divided into poor prognosis group (60 patients) and good prognosis group (192 patients) based on 1-year follow-up results. Additionally, 86 patients with AMI complicated by HF, with a ratio approximately 1∶3 to the training set, were selected as validation set. Cox regression models were used to analyze the influencing factors for poor prognosis. A nomogram model was constructed based on the screening results and underwent internal and external validationHosmer-Lemeshow test was used to assess goodness of fit, calibration curves were plotted to evaluate calibration, receiver operating characteristic (ROC) curves were drawn to analyze discriminative ability, and decision curve analysis (DCA) was conducted to assess clinical utility.

Results There were no statistically significant differences in clinical data between the training set and validation set (P>0.05). The poor prognosis group had higher levels of serum creatinine and cardiac troponin T (cTnT), higher proportions of patients aged ≥60 years, with time from onset to admission ≥4 hours, with heart function grades Ⅲ to Ⅳ, and a lower left ventricular ejection fraction (LVEF) compared with the good prognosis group (P < 0.05). Multivariate Cox regression analysis showed that time from onset to admission, heart function grade, serum creatinine, cTnT, and LVEF were independent influencing factors for poor prognosis in patients with AMI complicated by HF (P < 0.05). Based on these results, a nomogram model was constructed. Internal validation results showed that the model had good goodness of fit (χ²=13.966, P=0.083), excellent calibration, and good discriminative abilityarea under the curve (AUC) was 0.831. External validation results also showed that the model had good goodness of fit (χ²=6.465, P=0.136), excellent calibration, and good discriminative ability (AUC was 0.884). DCA results indicated that the nomogram model had good clinical net benefit within a high-risk threshold range of 0.02 to 0.98.

Conclusion Influencing factors for poor prognosis in patients with AMI complicated by HF include time from onset to admission, heart function grade, serum creatinine, cTnT, and LVEF.The constructed nomogram model has high predictive value for poor prognosis in these patients.