Drug and device effects on peak oxygen consumption, 6-minute walk distance, and natriuretic peptides as predictors of therapeutic effects on mortality in patients with heart failure and reduced ejection fraction.

BACKGROUND Although peak oxygen consumption (peak VO(2)), 6-minute walk distance (6MW), and natriuretic peptides (BNP and NT-proBNP) are predictors of mortality in heart failure (HF) patients, it is not known whether therapy-induced changes in these measures can predict therapeutic effect on mortality. The objective of this analysis is to quantitatively assess the relationship between therapeutic effects on commonly proposed short-term markers in HF trials and therapeutic effects on long-term outcome in patients with HF and left ventricular dysfunction. METHODS AND RESULTS We identified drug or device therapies for which there exists at least 1 randomized, controlled trial (RCT) assessing mortality over at least 6 months in at least 500 patients. For each of these therapies, we identified RCTs assessing the short-term changes in VO(2), 6MW, BNP, and NT-proBNP (few of the mortality RCTs assessed the short-term changes in markers). For each intervention, we calculated the odds ratio for mortality (using random effect meta-analysis when necessary), as well as the trial level average drug- or device-induced change in the markers. We assessed the correlation between the odds ratio for death with the placebo-corrected change in the functional parameter or biomarker across the interventions. We identified mortality RCTs of 27 distinct therapies (n=73 267 patients) with a median follow-up of 19 months, that directed the search for RCTs of the effect of those interventions on the functional markers and biomarkers. There were 54 peak VO(2) trials (n=4646 patients), 34 6MW trials (n=6995 patients), 15 BNP trials (n=7233), and 6 NT-proBNP trials (n=1946) included in this analysis. There was no significant correlation between the average therapy-induced placebo-corrected change in peak VO(2) and the odds ratio for mortality (r=0.158, P=0.26). Increased drug or device-induced average change in 6MW was correlated with increased odds ratio for mortality (r=0.373, P=0.036). There was no significant correlation between the average therapy-induced, placebo-corrected change in the natriuretic peptides and the odds ratio for mortality (BNP: r=-0.065, P=0.82, NT-proBNP: r=-0.667, P=0.15). There was no apparent relation between change in the functional parameter or biomarker and categorical effect on mortality. CONCLUSIONS This analysis, limited to trial level data from different therapeutic eras, suggests that drug- or device-induced effects on peak VO(2), 6MW, and natriuretic peptides found in short-term trials do not predict the corresponding average long-term therapeutic effects on mortality for patients with HF and left ventricular dysfunction.