This review found that eplerenone did not appear to be more effective in improving clinical outcomes for patients with left ventricular dysfunction (heart failure) compared with older aldosterone antagonists. The conclusion appears generally appropriate but was weakened by being based on indirect comparisons and may not be applicable to all patient groups.

To compare the relative effectiveness of eplerenone versus other aldosterone antagonists for improving clinical outcomes in patients with left ventricular dysfunction.

The authors searched PubMed, EMBASE, CINAHL and Cochrane Central Register of Controlled Trials to July 2011. Search terms were reported. There were no language restrictions. Reference lists of included studies and relevant reviews were searched.

Randomised controlled trials (RCTs) that compared an aldosterone antagonist (eplerenone, spironolactone or canrenone) with a specified comparator (placebo, angiotensin-converting enzyme inhibitor, angiotensin-receptor blocker or beta-blocker) in adult patients with left ventricular dysfunction were eligible for inclusion. RCTs that compared two aldosterone antagonists were excluded. Trials had to be at least eight weeks duration. The primary outcome was all-cause mortality at the end of follow-up. A secondary outcome was cardiovascular mortality and safety outcomes were hyperkalaemia and gynaecomastia.

Interventions in included trials were spironolactone (12.5mg to 100mg), canrenone (25mg or 50mg) or eplerenone (25mg to 100mg). Comparators were placebo (in most trials), metoprolol, usual care, captopril or ramipril. Four studies involved patients with left ventricular dysfunction after myocardial infarction; participants in other trials had heart failure. Mean participant age ranged from 55 to 68.7 years and 54% to 87% were men.

Two authors selected studies for inclusion; disagreements were resolved by consensus.

Quality was assessed using the Cochrane Collaboration risk of bias tool. It appeared that two reviewers independently assessed quality and disagreements were resolved by consensus.

Data were extracted to calculate risk ratios and associated 95% confidence intervals.

Two authors extracted data; disagreements were resolved by consensus.

The authors reported that pooled risk ratios were calculated using a random-effects model (Mantel-Haenszel). Results from both random-effects and fixed-effect analyses were presented. Heterogeneity was assessed using Χ² tests and the Ι² statistic. Publication bias was investigated using funnel plots and Egger's test. An adjusted indirect comparison of eplerenone versus other aldosterone antagonists was performed for all-cause mortality only.

Sixteen RCTs were included: three (9,489 participants) for eplerenone and 13 (3,016 participants) for other aldosterone antagonists. Risk of bias was rated high in one trial, intermediate in seven and low in eight. Follow-up ranged from two to 24 months.

Eplerenone (RR 0.85, 95% CI 0.77 to 0.93; two RCTs) and other aldosterone antagonists (RR 0.74, 95% CI 0.66 to 0.83; six RCTs) significantly reduced all-cause mortality compared with control. Heterogeneity was low. The adjusted indirect comparison favoured other aldosterone antagonists over eplerenone (p=0.009).

Eplerenone and other antagonists significantly reduced cardiovascular mortality compared with control. The magnitude of reduction was greater in the other antagonists group (RR 0.75, 95% CI 0.67 to 0.84; four RCTs) than for eplerenone (0.83, 95% CI 0.75 to 0.92; two RCTs).

Eplerenone was associated with an increased risk of hyperkalaemia (RR 1.72, 95% CI 1.19 to 2.47; Ι²=57%; three RCTs) but risk with other antagonists did not differ significantly from control (RR 1.80, 95% CI 0.83 to 3.91; Ι²=50%; 10 RCTs).

For gynaecomastia, eplerenone was not associated with increased risk (RR 0.74, 95% CI 0.43 to 1.27; two RCTs) but other antagonists were (RR 6.26, 95% CI 3.38 to 11.57; six RCTs).

Eplerenone did not appear to be more effective in improving clinical outcomes compared with older aldosterone antagonists.

The review question and inclusion criteria were generally clear. Exclusion of direct comparisons of different aldosterone antagonists probably reflected the authors' prior knowledge that no such trials had been done; otherwise this would have been inappropriate. The search covered a range of relevant resources without language restrictions. Risk of publication bias was assessed but only limited results were reported. Study quality was assessed appropriately. It appeared that review processes were conducted by at least two reviewers to reduce potential for errors and bias. Some relevant study details were reported but there were few details of participants so it was difficult to assess generalisability of the findings. Statistical methods used in the synthesis seemed appropriate but there was a discrepancy between the type of analysis reported in the methods and that presented in the results for some analyses.

The authors' conclusions reflected the evidence presented but were weakened by being based on an indirect comparison (rather than a direct comparison or a mixed treatment meta-analysis). Limited reporting of patient characteristics made it difficult to rule out differences between participants in the two groups of trials. However, the authors noted that control group death rates were higher in trials of other antagonists than in eplerenone trials, which suggested that eplerenone had generally been tested in healthier patients. Overall, the authors' conclusion that eplerenone did not appear to be superior to other aldosterone antagonists seems appropriate.

Practice: The authors stated that eplerenone should be prescribed cautiously pending further evaluation of its effectiveness and cost-effectiveness.

Research: The authors stated that there is a need for a direct comparison between eplerenone and other aldosterone antagonists, including a detailed pharmacoeconomic evaluation.

None.