Right ventricular to pulmonary artery coupling: an independent marker in a prediction model for heart failure with improved ejection fraction

Introduction Heart failure (HF) with improved ejection fraction (HFimpEF) was recently defined as a distinct clinical phenotype within the HF spectrum. Despite increasing recognition, its prognostic determinants remain incompletely characterized. Right ventricular (RV) to pulmonary artery (PA) coupling is acknowledged as an important hemodynamic parameter quantifying the interaction between RV contractile performance and RV afterload. A practical, non-invasive indirect marker of RV-PA coupling is the echocardiographic ratio of the tricuspid annular plane systolic excursion (TAPSE) to pulmonary artery systolic pressure (PASP). Purpose Our study aimed to assess the prognostic role of TAPSE/PASP alongside other biomarkers for left ventricular EF (LVEF) improvement after hospitalization for HF with reduced EF (HFrEF). Methods This is a retrospective cohort study including HFrEF patients previously hospitalized for HF and subsequently reevaluated between January 2024 - December 2025, with either improved or persistently reduced EF. HFimpEF was defined as baseline LVEF ≤40%, improved to >40%, with a ≥10% increase. Results We included 202 HFrEF patients (mean age 68.76±10.57 years; 68.81% male); 32.67% developed HFimpEF. Rhythm control via atrial fibrillation cardioversion was a direct predictor of LVEF recovery (OR 9.28, p<0.001). History of acute coronary syndrome (OR 0.25, p<0.001), chronic obstructive pulmonary disease (COPD) (OR 0.14, p<0.001), obstructive sleep apnoea syndrome (OSAS) (OR 0.20, p=0.005), peripheral arterial disease (OR 0.42, p=0.03) were inversely correlated with HFimpEF. Lower NYHA class (AUC 0.75, p<0.001), PASP (AUC 0.79, cut-off ≤36mmHg, p<0.001), NT-proBNP (AUC 0.66, p<0.001) and higher LVEF at baseline (AUC 0.60, p=0.016), TAPSE (AUC 0.79, cut-off>18mm, p<0.001), and TAPSE/PASP (AUC 0.84, cut-off >0.45, p<0.001) were correlated with HFimpEF. The independent parameters associated with HFimpEF in multivariable analysis were used to build the baseline prediction model: successful cardioversion (HR 10.02, p<0.001), NYHA class (HR 0.31, p<0.001), COPD (HR 0.25, p=0.017), and OSAS (HR 0.11, p=0.024). When added to the model, TAPSE/PASP >0.45mm/mmHg (HR 5.97, p<0.001) outperformed and replaced TAPSE >18mm (HR 2.88, p=0.018) and PASP ≤36mmHg (HR 3.46, p=0.06) as individual markers (Figure 1). The prediction model containing TAPSE/PASP >0.45mm/mmHg, successful cardioversion, NYHA class, COPD and OSAS had the best association with HFimpEF in ROC analysis, outperforming all the significant individual variables (Figure 2). Conclusions RV-PA coupling, quantified in our cohort by a TAPSE/PASP >0.45mm/mmHg, was an independent predictor of LVEF improvement in HFrEF patients. It outperformed individual assessments of PASP and TAPSE and improved the risk stratification model containing successful cardioversion, NYHA class, COPD, and OSAS as independent predictors of HFimpEF.