Mineralocorticoid Receptor Antagonists in High‐Risk Heart Failure Patients With Diabetes Mellitus and/or Chronic Kidney Disease

Introduction

The steroidal mineralocorticoid receptor antagonists (MRAs) spironolactone and eplerenone have been shown in large‐scale prospective double‐blind trials to reduce cardiovascular mortality and hospitalizations for heart failure in patients with a reduced left ventricular ejection fraction (HFrEF).1 Their use in patients with midrange and preserved left ventricular function (HFpEF) remains controversial. Despite designation as a class 1 indication in current US2 and European3 guidelines for patients with HFrEF, their use remains suboptimal in comparison to β‐adrenergic blocking agents, angiotensin‐converting enzyme inhibitors, or angiotensin receptor blocking agents.4 The relative underuse of MRAs in guideline‐eligible patients with HFrEF5 has in large part been attributed to the fear of inducing hyperkalemia and/or renal insufficiency (RI), especially in patients with diabetes mellitus (DM) and/or chronic kidney disease.6 In part these fears have been amplified by questions raised as to their effectiveness and safety in “real world” settings based upon retrospective analysis of data in various registries.7, 8 While large‐scale prospective randomized trials are recognized as the criterion standard for evaluating the safety and efficacy of new therapies for approval, the application and relevance to clinical practice obtained from these studies to the “real world” has been questioned because of various restrictions on inclusion into the large‐scale randomized trials and the intensity of patient monitoring and follow‐up in comparison to clinical practice in “real world” settings. “Real world” data regarding the safety and efficacy of a given intervention based upon retrospective analysis of registry data have, however, been criticized as suffering from bias by intent to treat. In an observational study, observations are collected without assignment, and no adjustment, however elaborate it may be, can protect against the “unavoidable risk of selection bias and of systematic differences in outcomes that are not due to the treatment itself.”9 Nevertheless, registry‐based analysis of safety and efficacy of a given therapeutic intervention is valuable in pointing out areas for further investigation.

In this issue of JAHA, Cooper et al,10 using the Get With The Guidelines registry linked to Medicare outcome data, have evaluated the first‐time use of MRAs in high‐risk patients with either a history of DM or RI at hospital discharge after an episode of HF. In contrast to some previous registry studies, they included both patients with a left ventricular ejection fraction <35% (HFrEF) as well as those with borderline (midrange: heart failure with midrange left ventricular ejection fraction‐HFmrEF) or preserved ejection fractions (HFpEF). They evaluated the factors favoring or not favoring the use of a MRA at hospital discharge as well as the 30‐day, 1‐year, and 3‐year incidence of mortality, hospitalizations for heart failure, and hospitalization for hyperkalemia and RI. Of 16 848 eligible patients, 2067 (12.3%) were prescribed a MRA for the first time at hospital discharge, of whom 60.5% had HFrEF. In the overall population after inverse probability weighting, the use of a MRA in these relatively high‐risk patients with a history of either DM and/or RI was not associated with a significant reduction in 30‐day, 1‐, or 3‐year mortality. MRA use was, however, associated with a 1‐ and 3‐year lower risk of hospital readmissions. This benefit of MRA use was associated with a greater risk of readmissions for hyperkalemia and RI at 1 and 3 years. The increased rate of readmissions for RI was, however, restricted to those with borderline and preserved ejection fractions.10 In the TOPCAT (Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist) study of patients with HFpEF randomized to spironolactone, there was a significant increase in the incidence of doubling of serum creatinine but no increase in the need for renal dialysis.11 It should also be pointed out that the absolute increase in the incidence of hospitalizations for hyperkalemia over 3 years in the analysis by Cooper et al10 was relatively low and that previous studies have suggested that the use of a MRA is associated with a reduction in mortality in comparison to placebo in patients with hyperkalemia, at least up to a serum K+ value of 5.5 mmol/L.12

What Are the Implications of the Findings by Cooper et al From This “Real World” Analysis?

An analysis of “real world” data such as that by Cooper et al10 cannot provide a definitive answer as to the efficacy or safety of MRAs in high‐risk patients with HF, as pointed out above. It can, however, suggest the need for further prospective study in these high‐risk HF patients with DM and/or RI. The evidence available from the prospective randomized trials of MRAs in patients with HFrEF would support the use of a MRA in these high‐risk guideline‐eligible patients13 (inclusion if serum K+ <5.0 mmol/L and estimated glomerular filtration rate >30 mL/min per 1.73 m²). Cooper et al10 also point out that “the benefits of therapy may outweigh the risks in this high–risk population” since there was an overall decrease in the risk of readmissions in patients treated with a MRA. One should, however, closely monitor these high‐risk patients. Current evidence suggests that many patients prescribed a MRA for HF do not undergo recommended serial monitoring of serum potassium (K+) and renal function.14 Monitoring of serum K+ and renal function would allow appropriate dose adjustment and/or discontinuation of a MRA as well as discontinuation of other medications that might affect renal function and thus serum K+. Those who do undergo appropriate serial monitoring often discontinue their MRA because of an increase in serum K+ and/or creatinine. While clearly one would like to avoid the increased risk of ventricular arrhythmias and sudden cardiac death associated with the occurrence of hyperkalemia, discontinuing a MRA could potentially expose the patient to an even greater risk since the steroidal MRA eplerenone has been shown to reduce all‐cause mortality and all‐cause hospitalizations in patients with HFrEF1 including those patients with DM and/or chronic kidney disease.13 The situation in regard to the use of a MRA in those patients with borderline and/or preserved ejection fractions (HFmrEF or HFpEF), especially those with DM and/or RI, is less certain. Although spironolactone appeared to reduce cardiovascular mortality and HFHs (Heart Failure Hospitalizations) in the Americas region in TOPCAT, the overall results including patients from Russia and the Republic of Georgia, who have been found not to have had the mortality rate we expect in patients with HFpEF and in whom there is evidence that many patients randomized to spironolactone and claiming to have taken it did not take it,15 were not significant. Until further strategies are identified that clearly reduce cardiovascular mortality and hospitalizations for heart failure in patients with HFmrEF and HFpEF, one should consider the use of spironolactone in these patients as suggested by the class II indication in current US guidelines.2 Further information in regard to the efficacy and safety of spironolactone in patients with midrange left ventricular function and HFpEF should be forthcoming from the prospectively randomized open label SPIRRIT‐HFpEF (Spironolactone Initiation Registry Randomized Interventional Trial in Heart Failure with Preserved Ejection Fraction) (n=3200) in which the primary outcome is cardiovascular mortality.16 In those patients with DM and/or RI, one might consider it prudent to await further data from the SPIRRIT trial16 before considering a MRA in view of the increased risk of RI noted by Cooper et al in these patients.10 However, even in these high‐risk patients there was a decrease in the incidence of readmissions without any increase in mortality, which would favor the use of a MRA. While the decision as to whether or not to administer a MRA to these high‐risk patients with borderline and/or preserved ventricular function remains uncertain, the data by Cooper et al10 are of value in focusing our attention on the need to consider new strategies to reduce the risk of hyperkalemia and RI in these high‐risk patients who are at great risk of dying of HF. The availability of new safe and well‐tolerated potassium‐lowering agents such as patiromer should avoid the risks of hyperkalemia associated with MRA use and potentially could allow the long‐term use of a MRA in these high‐risk patients despite the occurrence of hyperkalemia. The long‐term risks and benefits of such a strategy will, however, require further prospective evaluation.17 The use of new nonsteroidal MRAs such as Finerenone, which in early studies has been shown to be as effective as spironolactone in reducing brain natriuretic peptide and N‐terminal pro‐brain natriuretic peptide in HF patients but associated with a lower incidence of hyperkalemia,18 also holds promise for the high‐risk patients with HF complicated by DM and/or chronic kidney disease identified by Cooper et al.10 Whether Finerenone improves cardiovascular and renal outcomes in patients with diabetic nephropathy (Figaro n=6400, clinicaltrials.gov NCT02540993), (Fidelio n=4800, NCT02545049) is currently being evaluated.

Finally, the trend toward the design of more pragmatic prospective randomized trials with fewer exclusions should hopefully in the future reduce the differences between “real world” data obtained retrospectively from registries such as that by Cooper et al10 and the prospective randomized trials used for drug approval. Until then, both the large‐scale prospective randomized trials and the retrospective analysis of registries to provide “real world” data such as that by Cooper et al10 have a role in informing clinical practice—but in a different manner.

Disclosures

Dr Pitt reports personal fees (consulting) from Bayer, KBP Pharmaceuticals, AstraZeneca, Boehringer Ingelheim, Merck, Takeda, Relypsa, Sanofi, sc Pharmaceuticals, Sarfez Pharmaceuticals, Stealth Peptides, Vifor, and Tricida; stock options from KBP Pharmaceuticals, scPharmaceuticals, Sarfez Pharmaceuticals, Relypsa, and Tricida; patent pending for site‐specific delivery of eplerenone to the myocardium. Dr Rossignol reports personal fees (consulting) from Novartis, Relypsa, AstraZeneca, Grünenthal, Stealth Peptides, Fresenius, Vifor Fresenius Medical Care Renal Pharma, Vifor, and CTMA; lecture fees from Bayer and CVRx, and is a cofounder of CardioRenal.

J Am Heart Assoc. 2017;6:e008054 DOI: 10.1161/JAHA.117.008054.29275377