Risk prediction in heart failure using invasive hemodynamics.

Abstract

Aims: Risk stratification in patients with heart failure patients is crucial. The prognostic value of invasive hemodynamic parameters measured by right heart catheterization compared to established risk scores remains unknown.

Methods and results: This retrospective analysis included 883 patients. The combined endpoint was all-cause mortality, heart transplantation or left ventricular assist device implantation. A Cox proportional hazards model assessed the impact of invasive parameters, cardiac biomarkers, and patient characteristics, comparing them with the Seattle Heart Failure Model (SHFM) and the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) Score. A new score was created including mean pulmonary arterial (PA) pressure, mean right atrial pressure, mean pulmonary artery wedge pressure (PAWP), age, N-terminal pro-brain natriuretic peptide (NT-proBNP), high-sensitivity troponin T (hsTnT), mixed venous oxygen saturation (SVO2), creatinine, and presence of ischemic cardiomyopathy. Mean, diastolic, and systolic PA pressure, mean right atrial pressure, mean PAWP, SVO2 and cardiac index were significant predictors for the primary endpoint reached by 467/883 (53%) patients, in a multiple Cox proportional hazards model (p < 0.001). The predictive value was diminished in a subgroup of patients with ischemic cardiomyopathy. We used invasive parameters, age, NT-proBNP, hsTnT_, creatinine presence of ischemic cardiomyopathy and sex to develop a new model for risk stratification. This new score showed better performance compared to the SHFM and MAGGIC score in predicting the primary endpoint at 6, 12 and 24 months (area under the curve 0.76, 0.78 and 0.77 vs 0.71/0.69, 0.70/0.68 and 0.70/0.70).

Conclusion: Invasive hemodynamics provides valuable measurements for predicting outcome in heart failure with reduced ejection fraction and show better performance than established risk models when combined with cardiac biomarkers and other clinical variables in this particular cohort.