Development and Validation of a Novel Nomogram to Predict Improved Left Ventricular Ejection Fraction in Patients With Heart Failure After Successful Percutaneous Coronary Intervention for Chronic Total Occlusion

BACKGROUND: Heart failure with improved left ventricular ejection fraction (HFiEF) is linked to a good clinical outcome. The purpose of this study was to create an easy-to-use model to predict the occurrence of HFiEF in patients with heart failure (HF), 1 year after successful percutaneous coronary intervention (PCI) for chronic total occlusion (CTO) (CTO PCI). METHODS: Patients diagnosed with HF who successfully underwent CTO PCI between January 2016 and August 2019 were included. To mitigate the effect of residual stenosis on left ventricular (LV) function, we excluded patients with severe residual stenosis, as quantitatively measured by a residual synergy between PCI with Taxus and Cardiac Surgery score (rSS) of >8. We gathered demographic data, medical history, angiographic and procedural characteristics, echocardiographic parameters, laboratory results, and medication information. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression models were used to identify predictors of HFiEF 1 year after CTO revascularization. A nomogram was established and validated according to the area under the receiver operating characteristic curve (AUC) and calibration curves. Internal validation was performed using bootstrap resampling. RESULTS: A total of 465 patients were finally included in this study, and 165 (35.5%) patients experienced HFiEF 1 year after successful CTO PCI. According to the LASSO regression and multivariate logistic regression analyses, four variables were selected for the final prediction model: age [odds ratio (OR): 0.969; 95% confidence interval (CI): 0.952–0.988; p = 0.001], previous myocardial infarction (OR: 0.533; 95% CI: 0.357–0.796; p = 0.002), left ventricular end-diastolic dimension (OR: 0.940; 95% CI: 0.910–0.972; p < 0.001), and sodium glucose cotransporter two inhibitors (OR: 5.634; 95% CI: 1.756–18.080; p = 0.004). A nomogram was constructed to present the results. The C-index of the model was 0.666 (95% CI, 0.613–0.719) and 0.656 after validation. The calibration curve demonstrated that the nomogram agreed with the actual observations. CONCLUSIONS: We developed an simple and effective nomogram for predicting the occurrence of HFiEF in patients with HF, 1 year after successful CTO PCI without severe residual stenosis.