Abstract 16091: Left Ventricular Diastolic Function and Peak Exercise Oxygen Consumption Contribute Independently to Predicting Clinical Outcomes in Patients With Systolic Heart Failure in Heart Failure: A Controlled Trial Investigating Outcomes of Exercise TraiNing (HF-ACTION)
The HF-ACTION trial is the largest randomized trial to evaluate the effect of exercise training on clinical events in HF. We examined the ability of resting echocardiographic (echo) measurements at study entry to independently predict clinical events and response to exercise training in the subset of the 2,331 participants who had evaluable echo measurements of left ventricular (LV) size, mass, systolic (ejection fraction) and diastolic (mitral valve [MV] E/A and E/E’ ratios) function, left atrial (LA) size, and mitral regurgitation (MR) severity. The primary clinical endpoint was all-cause death and hospitalization; a secondary endpoint was cardiovascular disease (CVD) death and HF hospitalization. All 7 echo variables were significant univariate predictors of the primary endpoint; however, in multivariable analyses, singly adding each to a basic demographic/clinical model only increased the C-index from 0.60 to 0.61, compared with an increase to 0.63 after adding peak exercise oxygen consumption (VO2) to the model. For the secondary endpoint, addition of either all 7 echo variables or peak VO2 to a basic model added significantly more information (C-index increase from 0.70 to 0.73 for peak VO2; 0.70 to 0.73 for the 7 echo variables; p < 0.0001 for each) for predicting HF hospitalization and CVD mortality (see Table). Adding both 7 echo variables plus peak VO2 increased the C-index to 0.75. Among echo variables, measures of LV diastolic function_e.g., MV E/A and E/E’ ratios_were more powerful incremental predictors than was LV ejection fraction for both primary and secondary endpoints. No echo variable was significantly related to 0-to-3 month change in exercise peak VO2. We conclude that echo variables reflecting LV diastolic function and peak VO2 add similarly and incrementally to a basic demographic/clinical model for predicting clinical outcomes (especially, CVD) in patients with systolic HF, but echo variables do not predict 3-month change in peak VO2.
- © 2011 by American Heart Association, Inc.