Analysis of risk factors for heart failure in patients with type 2 diabetes mellitus and acute ST-segment elevation myocardial infarction after percutaneous coronary intervention

Objective This study aimed to investigate the independent risk factors for in-hospital heart failure (HF) following percutaneous coronary intervention (PCI) in patients with type 2 diabetes mellitus (T2DM) and acute ST-segment elevation myocardial infarction (STEMI), and to develop and validate a personalized risk prediction model. Methods A retrospective cohort study was conducted, enrolling 362 consecutive patients with T2DM and STEMI who underwent primary PCI between January 2022 and June 2025. Patients were categorized into an HF group ( n = 74) and a non-HF group ( n = 288) based on the occurrence of in-hospital HF. Baseline clinical data were collected. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors. A nomogram prediction model was constructed based on the regression results. Its discriminative ability and calibration were assessed using the area under the receiver operating characteristic curve (AUC) and a calibration plot, respectively. Missing data were handled using multiple imputation. Sensitivity analysis, including different variable selection methods and variance inflation factor (VIF) calculation, was conducted to evaluate the model’s robustness. Results Multivariate logistic regression identified seven independent risk factors for in-hospital HF (all p 0.05): prior myocardial infarction (OR = 4.187, 95% CI: 2.374–7.389), smoking history (OR = 2.683, 95% CI: 1.630–4.415), decreased left ventricular ejection fraction (per 1% decrease, OR = 0.944), elevated white blood cell count (per 1 × 10 9 /L increase, OR = 1.107), decreased hemoglobin level (per 1 g/L decrease, OR = 0.976), elevated platelet count (per 50 × 10 9 /L increase, OR = 1.200), and atrial fibrillation (OR = 2.121, 95% CI: 1.085–4.146). The nomogram incorporating these seven factors demonstrated good predictive performance, with an AUC of 0.845 (95% CI: 0.792–0.898) and good calibration according to the Hosmer-Lemeshow goodness-of-fit test ( p = 0.712). Sensitivity analysis, using different variable selection methods (Forward and Backward stepwise regression) and VIF calculation, confirmed the stability of the model and the absence of significant multicollinearity (all VIF 1.5). Conclusion The constructed nomogram model shows good discriminative ability and calibration in this single-center internal validation. However, external validation is required before it can be considered a clinical-ready tool. It provides a preliminary quantitative tool for the early identification of high-risk patients and requires external validation in multi-center settings.