One echocardiographic or hemodynamic parameter improves mortality prediction by the ESC/ERS four-strata model: a flexible approach to follow-up risk stratification in pulmonary arterial hypertension

Introduction: The ESC/ERS-recommended four-strata mortality risk stratification of prevalent patients with pulmonary arterial hypertension (PAH) does not incorporate echocardiographic or right heart catheterization (RHC) measures. Methods: Within the multicentre, observational study FOCUS-PAH, we selected those patients who underwent both echocardiography and RHC (primary analysis) or at least one of these exams (secondary analysis) in the first 12 months after PAH diagnosis and with known vital status over the following 6 years. The association of echocardiographic and RHC variables with all-cause mortality was evaluated by Cox regression. Risk prediction model performance was assessed using ROC curves and net reclassification improvement. Results: The primary analysis included 197 patients (58% female; median age 55 years). In multivariable analyses, right atrial area (RAA, HR 1.09 per 1-cm2 increase) and the percentage change in pulmonary vascular resistance (ΔPVR, HR 1.01 per 1% change) at follow-up were independent predictors of all-cause mortality when added to ESC/ERS categories. The four-strata model AUC was 0.66 and improved to 0.73 (P=0.02), 0.74 (P=0.002), and 0.77 (P=0.0005), respectively, with RAA, ΔPVR, or both. Including RAA and ΔPVR reclassified 25% of patients into more appropriate risk groups. RA pressure, instead of RAA, also predicted mortality (HR 1.08 per 1-mmHg increase) and similarly enhanced ESC/ERS four-strata discrimination (AUC 0.73 vs 0.66, P=0.009). Moreover, RAA (AUC 0.73, P=0.007) and/or ΔPVR (AUC 0.68, P=0.004) improved ESC/ERS risk prediction in the secondary analysis including 394 patients. Conclusions: Adding one echocardiographic or RHC parameter interchangeably improves mortality risk prediction by the ESC/ERS four-strata model in PAH.