Abstract 17026: Serum Potassium and Outcomes in Heart Failure: Benefit of Aldosterone Antagonists Independent of Baseline Serum Potassium Levels
Introduction: The benefits of aldosterone antagonists (AAs) for patients with heart failure (HF) are well known and have been demonstrated in large outcome trials, the Randomized Aldactone Evaluation Study (RALES) and the Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study (EPHESUS). There are several mechanisms through which AAs are believed to confer their benefits, one of which may be their ability to normalize serum potassium (SK) levels. In a post-hoc analysis of RALES and EPHESUS by starting SK values, we examined the predictor and mortality association of SK.
Hypothesis: AAs provide survival benefit regardless of starting SK in HF patients.
Methods: Using Cox proportional regression in each of 8 SK strata and each of the whole study populations, hazard ratios were used to assess SK at baseline as a predictor of mortality. The validity of the Cox model assumptions was assessed: 1) statistically, by testing the time x study drug interaction and 2) visually, by plotting the log(-log(survival)) against log(time) curves of each randomization group.
Results: In both studies, the majority of patients died from cardiovascular (CV) causes (RALES 583/668, 87.3%, EPHESUS 890/1032, 86.2%), thus the results for all-cause and CV mortality showed similar trends. Despite the loss of power induced by 8 subgroups, the beneficial effect of spironolactone (SPI) was observed across all strata and reached significance for all-cause mortality (primary endpoint) for SK strata within the normal range, 4.11–4.50 mmol/L (46.9% placebo vs 34.1% SPI, p=0.007) and 4.51–5.00 mmol/L (45.1% placebo vs 29.4% SPI, p=0.003), which also encompassed the greatest number of patients. Similar effects were observed for eplerenone (EPL), although there was a non-significant negative effect in the extreme strata with fewer patients (Figure 1).
Conclusions: The benefit of AAs on all-cause and CV mortality was retained across a broad range of baseline SK levels.
- © 2010 by American Heart Association, Inc.