Predictors of early dyspnoea relief in acute heart failure and the association with 30-day outcomes: findings from ASCEND-HF.

AIMS: To examine the characteristics associated with early dyspnoea relief during acute heart failure (HF) hospitalization, and its association with 30-day outcomes. METHODS AND RESULTS: ASCEND-HF was a randomized trial of nesiritide vs. placebo in 7141 patients hospitalized with acute HF in which dyspnoea relief at 6 h was measured on a 7-point Likert scale. Patients were classified as having early dyspnoea relief if they experienced moderate or marked dyspnoea improvement at 6 h. We analysed the clinical characteristics, geographical variation, and outcomes (mortality, mortality/HF hospitalization, and mortality/hospitalization at 30 days) associated with early dyspnoea relief. Early dyspnoea relief occurred in 2984 patients (43%). In multivariable analyses, predictors of dyspnoea relief included older age and oedema on chest radiograph; higher systolic blood pressure, respiratory rate, and natriuretic peptide level; and lower serum blood urea nitrogen (BUN), sodium, and haemoglobin (model mean C index = 0.590). Dyspnoea relief varied markedly across countries, with patients enrolled from Central Europe having the lowest risk-adjusted likelihood of improvement. Early dyspnoea relief was associated with lower risk-adjusted 30-day mortality/HF hospitalization [hazard ratio (HR) 0.81; 95% confidence interval (CI) 0.68-0.96] and mortality/hospitalization (HR 0.85; 95% CI 0.74-0.99), but similar mortality. CONCLUSION: Clinical characteristics such as respiratory rate, pulmonary oedema, renal function, and natriuretic peptide levels are associated with early dyspnoea relief, and moderate or marked improvement in dyspnoea was associated with a lower risk for 30-day outcomes.

Duke Scholars

Author Gary Michael Felker Medicine, Cardiology

Author Robert John Mentz Medicine, Cardiology

Author Adrian Felipe Hernandez Medicine, Cardiology

Author Victor Hasselblad Biostatistics & Bioinformatics, Division of Biostatistics

Author Christopher Michael O'Connor Medicine, Clinical Pharmacology