Reduced Risk of Hyperkalemia During Treatment of Heart Failure With Mineralocorticoid Receptor Antagonists by Use of Sacubitril/Valsartan Compared With Enalapril: A Secondary Analysis of the PARADIGM-HF Trial

Key Points

Question  For patients with heart failure and reduced ejection fraction (HFrEF), is the risk of hyperkalemia associated with the use of mineralocorticoid receptor antagonists (MRAs) lower among those treated with sacubitril/valsartan or those treated with enalapril?

Findings  In this secondary analysis of a randomized clinical trial, we examined the incidence of hyperkalemia (potassium level >5.5 mEq/L) and severe hyperkalemia (potassium level >6.0 mEq/L) according to use of MRA at baseline among patients with HFrEF randomly assigned to treatment with enalapril or sacubitril/valsartan as part of the PARADIGM-HF (Prospective Comparison of ARNI With an ACE-Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial. Severe hyperkalemia was more common in MRA-treated patients assigned to enalapril than in those assigned to sacubitril/valsartan, even after accounting for incident MRA use during the course of the trial.

Meaning  Neprilysin inhibition with sacubitril/valsartan may attenuate the risk of hyperkalemia when MRAs are combined with other inhibitors of the renin-angiotensin-aldosterone system for patients with HF.

Importance  Consensus guidelines recommend the use of mineralocorticoid receptor antagonists (MRAs) for selected patients with symptomatic heart failure and reduced ejection fraction (HFrEF) to reduce morbidity and mortality; however, the use of MRAs in combination with other inhibitors of the renin-angiotensin-aldosterone system increases the risk of hyperkalemia.

Objective  To determine whether the risk of hyperkalemia associated with use of MRAs for patients with HFrEF is reduced by sacubitril/valsartan in comparison with enalapril.

Design, Setting, and Participants  The PARADIGM-HF (Prospective Comparison of ARNI With an ACE-Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial randomly assigned 8399 patients with chronic HF, New York Heart Association class II to IV symptoms, and a left ventricular EF of 40% or less to treatment with enalapril 10 mg twice daily or sacubitril/valsartan 97/103 mg twice daily (previously known as LCZ696 [200 mg twice daily]) in addition to guideline-directed medical therapy. Use of MRAs was encouraged but left to the discretion of study investigators. Serum potassium level was measured at every study visit. The incidence of hyperkalemia (potassium level >5.5 mEq/L) and severe hyperkalemia (potassium level >6.0 mEq/L) among patients treated or not treated with an MRA at baseline and the risk of subsequent hyperkalemia for those newly treated with an MRA during study follow-up were defined in time-updated Cox proportional hazards models. Analyses were conducted between August 1 and October 15, 2016.

Main Outcomes and Measures  Incident hyperkalemia and severe hyperkalemia.

Results  In comparison with the 3728 patients (44.4% of enrolled participants [21.6% female]) not taking an MRA at baseline, the 4671 patients (55.6% [22.0% female]) taking an MRA tended to be younger, with a lower EF, lower systolic blood pressure, and more advanced HF symptoms. Among those taking an MRA at baseline, the overall rates of hyperkalemia were similar between treatment groups, but severe hyperkalemia was more common in patients randomly assigned to enalapril than to sacubitril/valsartan (3.1 vs 2.2 per 100 patient-years; HR, 1.37 [95% CI, 1.06-1.76]; P = .02). In analyses including patients who newly started taking MRAs during the PARADIGM-HF trial, severe hyperkalemia remained more common in those randomly assigned to enalapril than to those randomly assigned to sacubitril/valsartan (3.3 vs 2.3 per 100 patient-years; HR, 1.43 [95% CI, 1.13-1.81]; P = .003).

Conclusions and Relevance  Among MRA-treated patients with symptomatic HFrEF, severe hyperkalemia is more likely during treatment with enalapril than with sacubitril/valsartan. These data suggest that neprilysin inhibition attenuates the risk of hyperkalemia when MRAs are combined with other inhibitors of the renin-angiotensin-aldosterone system in patients with HF.

Trial Registration  clinicaltrials.gov Identifier: NCT01035255

The use of mineralocorticoid receptor antagonists (MRAs) is associated with a reduction in the risk of death or hospitalization for heart failure (HF) among patients with symptomatic HF and reduced ejection fraction (HFrEF).^1,2 Accordingly, current treatment guidelines recommend the addition of an MRA for most patients who remain symptomatic despite treatment with β-blockers and inhibitors of the renin-angiotensin-aldosterone system.^3,4 Despite the acknowledged benefits, enthusiasm for the use of MRAs is often limited by concern for the heightened risk of hyperkalemia, particularly in patients with chronic kidney disease.^5,6 Accordingly, many eligible patients are not treated with MRAs in clinical practice.⁷

In the PARADIGM-HF (Prospective Comparison of ARNI With an ACE-Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial, patients with symptomatic HFrEF randomly assigned to receive sacubitril/valsartan (previously known as LCZ696) had lower rates of death and hospitalization for HF than those randomly assigned to receive enalapril.⁸ The incremental benefits of sacubitril/valsartan were consistent among subgroups of patients treated or not treated with an MRA at baseline, and they were associated with a concomitant reduction in overall rates of clinically important hyperkalemia. To examine whether treatment with sacubitril/valsartan might reduce the risk of hyperkalemia associated with the use of MRAs for heart failure, we examined the postrandomization risk of hyperkalemia among PARADIGM-HF patients treated or not treated with an MRA at baseline.

The detailed design, methods, and principal results of the PARADIGM-HF trial have been previously reported.⁹ In brief, the trial was a randomized, double-blind, prospective comparison of the effect of sacubitril/valsartan (97/103 mg twice daily [LCZ696, 200 mg twice daily]) vs the effect of enalapril (10 mg twice daily) among 8399 patients with chronic heart failure (New York Heart Association [NYHA] classes II-IV) and a left ventricular EF of 40% or less. Prior to randomization, all participants underwent single-blind, sequential run-in periods to ensure tolerability of both study drugs at target doses. Eligible patients were those treated with an appropriate regimen of background HF medications at stable doses for at least 4 weeks, including an angiotensin-converting enzyme (ACE) inhibitor or an angiotensin receptor blocker (dose equivalent to enalapril 10 mg daily or greater) and β-adrenergic receptor blocker (unless not tolerated). The use of MRAs was left to the discretion of the investigators but encouraged if tolerated. Key exclusion criteria included symptomatic hypotension; a systolic blood pressure lower than 100 mm Hg at screening or 95 mm Hg at randomization; an estimated glomerular filtration rate of less than 30 mL/min/1.73 m² at screening/randomization or a decrease in the estimated glomerular filtration rate higher than 25% (amended to 35%) between screening and randomization; and a serum potassium level higher than 5.2 mEq/L (to convert to millimoles per liter, multiply by 1.0) at screening or higher than 5.4 mEq/L at randomization. The serum potassium level was measured at every study visit, with additional postrandomization checks of potassium advised for any patient with a serum potassium level higher than 5.3 mEq/L to confirm stability without an increase to the range of concern (≥5.5 and <6.0 mEq/L) or potential danger (≥6.0 mEq/L). The protocol recommended withdrawal of study drug for any patient with a potassium level of 5.5 mEq/L or higher confirmed by repeated testing. The primary composite end point for the trial was death from cardiovascular causes or first hospitalization for HF. The trial was approved by the institutional review board or ethics committee at each participating site, and all participants provided written informed consent prior to participation.

We examined all postbaseline measurements of serum potassium to assess the incidence of hyperkalemia among patients treated or not treated with an MRA at baseline during the median 27-month follow-up in the PARADIGM-HF trial. Incident hyperkalemia was defined as having a potassium level higher than 5.5 mEq/L, and severe hyperkalemia was defined as having a potassium level higher than 6.0 mEq/L. Patients meeting both definitions were counted in both analyses. The times to incident hyperkalemia and severe hyperkalemia were evaluated using the Kaplan-Meier method, with hazard ratios (HRs) derived from Cox proportional hazards models. Multivariable models were adjusted for all characteristics presented in Table 1, including known predictors of hyperkalemia such as age, diabetes, renal function, and concomitant medications, as well as potassium levels at screening and randomization. Separate analyses using competing risk methods were conducted to account for potential bias introduced by differential rates of death among patients assigned to enalapril vs those assigned to sacubitril/valsartan. The risk of subsequent hyperkalemia for those newly initiated to treatment with MRA during follow-up was defined in time-updated Cox proportional hazards models, in which patients were not considered at risk for hyperkalemia until the day of MRA initiation. Separate models were constructed to define the predictors of hyperkalemia (potassium level >5.5 mEq/L) among participants treated with MRA at baseline for both randomized groups, forcing in known predictors including age, diabetes, estimated glomerular filtration rate, sex, and baseline potassium level. All statistical analyses were conducted in STATA version 14.2 (StataCorp). P < .05 was the threshold for statistical significance.

Of the 8399 validly randomized participants in the PARADIGM-HF trial, 4671 (55.6%) were taking an MRA (4124 of 4671 [88.3%] were taking spironolactone) at baseline, including 2271 of 4187 patients randomly assigned to sacubitril/valsartan (54.2%) and 2400 of 4212 patients randomly assigned to enalapril (57.0%) (Figure 1). Based on current American College of Cardiology/American Heart Association guidelines,³ 7005 participants (83.4%) were eligible for MRA treatment, of whom 56.5% were taking an MRA at baseline; of 1394 participants not meeting guideline-recommended criteria for an MRA, 717 (51.4%) were using an MRA. The clinical characteristics of those taking or not taking an MRA at baseline are summarized in Table 1. Compared with the participants not taking an MRA at baseline, the participants taking an MRA at baseline were younger, with a lower EF, lower systolic blood pressure, less severe ischemic heart disease, greater likelihood of prior hospitalization for HF, and more advanced HF symptoms. Whether or not participants were taking an MRA at baseline, the patient characteristics were similar for those randomly assigned to enalapril and those randomly assigned to sacubitril/valsartan (Table 1). Treatment with sacubitril/valsartan reduced the risk of cardiovascular death regardless of MRA use at baseline (MRA nonusers: HR, 0.75 [95% CI, 0.63-0.89]; MRA users: HR, 0.84 [95% CI, 0.73-0.98]; P = .30 for interaction).

The serum potassium and creatinine levels by study visit among MRA-treated patients in each randomized group are summarized in Figure 2. As shown, changes in potassium levels did not parallel changes in renal function over time. For the population as a whole, the incidences of hyperkalemia (10.0 vs 7.3 per 100 patient-years; HR, 1.33 [95% CI, 1.19-1.48]; P < .001) and severe hyperkalemia (2.7 vs 2.0 per 100 patient-years; HR, 1.35 [95% CI, 1.11-1.64]; P = .003) were higher among MRA-treated patients than untreated patients during study follow-up. For patients not taking an MRA at baseline, the rates of hyperkalemia and severe hyperkalemia were similar among patients randomly assigned to sacubitril/valsartan and patients randomly assigned to enalapril (Table 2). For patients taking an MRA at baseline, the rates of hyperkalemia (10.6 vs 9.4 per 100 patient-years; HR, 1.12 [95% CI, 0.98-1.28]; P = .11; risk difference, 1.2 per 100 patient-years [95% CI, −0.1 to 2.6 per 100 patient-years]) were similar between treatment groups; however, incident severe hyperkalemia (3.1 vs 2.2 per 100 patient-years; HR, 1.37 [95% CI, 1.06-1.76]; P = .02; risk difference, 1.0 per 100 patient-years [95% CI, 0.3-1.6 per 100 patient-years]) was more common in patients randomly assigned to enalapril than patients randomly assigned to sacubitril/valsartan (Table 2). Reanalysis of these data using competing risk methods did not substantively alter the estimated HRs for either hyperkalemia or severe hyperkalemia. Although the incremental risk of hyperkalemia with enalapril was slightly greater among MRA-treated patients than untreated patients, no formal statistical interaction between treatment assignment and MRA use at baseline with regard to the risk of hyperkalemia was apparent. Kaplan-Meier curves depicting the time to development of hyperkalemia and severe hyperkalemia according to baseline MRA use and treatment assignment are displayed in Figure 3. Overall, 520 participants experienced 2 or more episodes of hyperkalemia, of whom 303 (161 randomly assigned to enalapril and 142 randomly assigned to sacubitril/valsartan) were treated with an MRA at baseline.

For participants randomly assigned to enalapril, diabetes, screening potassium level, baseline potassium level, baseline estimated glomerular filtration rate, region, and lower body mass index remained statistically significant predictors of hyperkalemia in multivariable models. For participants randomly assigned to sacubitril/valsartan, diabetes, screening potassium level, baseline potassium level, higher natriuretic peptide levels, and higher (worse) NYHA class were predictors. The only statistically important interaction between treatment assignment and predictors of hyperkalemia was noted with NYHA class, which appeared to be a predictor for participants randomly assigned to sacubitril/valsartan but not for participants randomly assigned to enalapril.

Of the 3728 participants who were not taking an MRA at baseline, 791 (21.2%) started receiving an MRA during the course of study follow-up. Postrandomization initiation of MRA occurred more often among patients randomly assigned to enalapril (423 of 1812 [23.3%]) than among patients randomly assigned to sacubitril/valsartan (368 of 1916 [19.2%]) (P = .002). There was no apparent difference between groups in the change in blood pressure or rates of hospitalization for HF prior to MRA initiation. Including both the patients who were receiving MRA at baseline and those newly initiated during the trial, the risk of subsequent severe hyperkalemia was also more common among those randomly assigned to enalapril than those randomly assigned to sacubitril/valsartan (3.3 vs 2.3 per 100 patient-years; HR, 1.43 [95% CI, 1.13-1.81]; P = .003).

As anticipated, incident hyperkalemia was more frequent among MRA-treated patients in the PARADIGM-HF trial; however, the risk of serious hyperkalemia (potassium level >6.0 mEq/L) was lower when MRAs were used in combination with sacubitril/valsartan rather than an ACE inhibitor. This lower risk was consistent among participants who were not treated with an MRA at baseline but initiated MRA therapy during the course of the study. Because the reduction in the rates of death and hospitalization for HF with sacubitril/valsartan seen in the PARADIGM-HF trial was consistent among MRA-treated patients and untreated patients, these data further support the rationale for the substitution of sacubitril/valsartan for an ACE inhibitor or angiotensin receptor blocker among eligible patients, as recommended by current treatment guidelines.¹⁰

The overall rate of hyperkalemia (18% over 27 months) among MRA-treated patients observed in the PARADIGM-HF trial was similar to that observed among spironolactone-treated patients in the Randomized Aldactone Evaluation Study of patients with HFrEF and symptoms of NYHA classes II to IV (19% over 24 months)¹¹ but was significantly higher than that observed among eplerenone-treated patients in the EMPHASIS-HF (Eplerenone in Mild Patients Hospitalization and Survival Study in Heart Failure) trial of patients with HFrEF and symptoms of NYHA classes I to II (11% over 21 months).¹² These rates of hyperkalemia were observed despite the use of a run-in period prior to randomization in which patients who developed hyperkalemia with titration of enalapril or sacubitril/valsartan to target doses were selectively excluded from participation. Rates of hyperkalemia (18.7%) and serious hyperkalemia (6.1%) were highest in the group treated with the combination of MRA and enalapril, and were nearly one-third lower in the group treated with sacubitril/valsartan; the rates of severe hyperkalemia among patients treated with the combination of MRA and sacubitril/valsartan were comparable to those seen among patients treated with enalapril alone. Little difference between treatment groups was seen in the rates of hyperkalemia among patients not treated with an MRA at baseline, suggesting that sacubitril/valsartan may selectively attenuate MRA-associated increases in serum potassium levels. Because the rates of hyperkalemia observed during treatment with ACE inhibitors and angiotensin receptor blockers are consistent in previous studies, these data point to a clinically important effect of sacubitril on potassium homeostasis.

Despite compelling evidence for the clinical benefit of MRAs and the guideline-based directives for MRA use among patients with symptomatic HFrEF, they remain underused among eligible patients in clinical practice.¹³ A key concern is the risk of hyperkalemia, which is enhanced when MRAs are added to a regimen including antagonists of the renin-angiotensin-aldosterone system. These data from the PARADIGM-HF trial suggest that the substitution of sacubitril/valsartan for an ACE inhibitor or angiotensin receptor blocker may enhance the safety and tolerability of MRAs in suitable patients with HFrEF by moderating the risk of serious hyperkalemia. In this fashion, they challenge the suggestion in the recent European Society of Cardiology guidelines that initiation of sacubitril/valsartan should be considered only after patients have been successfully treated with an MRA.⁴

Our analysis must be viewed in the context of its limitations. First, randomization in the PARADIGM-HF trial was not stratified according to MRA use, and, accordingly, we cannot exclude residual confounding by differences in the baseline risk for hyperkalemia among MRA-treated patients assigned to enalapril vs sacubitril/valsartan. Second, patients with advanced chronic kidney disease and those who developed hyperkalemia during the run-in period were systematically excluded from randomization, so the rates of hyperkalemia reported here may underestimate those in clinical practice. Finally, there are no reliable data available regarding MRA dosing or adherence over the course of the study.

In summary, these data from the PARADIGM-HF trial suggest that neprilysin inhibition attenuates the risk of hyperkalemia when MRAs are combined with other inhibitors of the renin-angiotensin-aldosterone system in patients with HF. Accordingly, the use of sacubitril/valsartan in preference to enalapril among eligible patients with HFrEF may enhance the ability to use MRAs safely, allowing patients to reap the incremental benefits at less incremental risk. Given the marked reductions in morbidity and mortality associated with the use of MRAs for patients with HFrEF, these observations provide additional rationale to consider replacing ACE inhibitors with sacubitril/valsartan for suitable patients in clinical practice.

Accepted for Publication: September 26, 2016.

Corresponding Author: Akshay S. Desai, MD, MPH, Cardiovascular Division, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115 (adesai@partners.org).

Published Online: November 14, 2016. doi:10.1001/jamacardio.2016.4733

Author Contributions: Drs Desai and Claggett had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the analyses.

Concept and design: Desai, Packer, Swedberg, Zile, Lefkowitz, Shi, Solomon.

Acquisition, analysis, or interpretation of data: Desai, Vardeny, Claggett, McMurray, Packer, Rouleau, Zile, Lefkowitz, Shi, Solomon.

Drafting of the manuscript: Desai, Claggett.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Desai, Claggett, Solomon.

Obtained funding: Packer, Lefkowitz, Solomon.

Administrative, technical, or material support: Desai.

Supervision: McMurray, Packer, Swedberg, Lefkowitz, Solomon.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Drs Lefkowitz and Shi are employees of Novartis Pharmaceuticals Corporation. All the other authors have consulted for or received research support from Novartis, sponsor of the PARADIGM-HF trial. In addition, Dr Desai consulted for Novartis, Relypsa, Janssen, Sanofi, Merck, and St Jude Medical. Dr Packer has consulted for Amgen, Boehringer Ingelheim, BioControl, Cardiorentis, Celyad, Daiichi Sankyo, GlaxoSmithKline, Novartis, Takeda, Relypsa, and ZS Pharma. Dr Swedberg has received honoraria and research support from Amgen, AstraZeneca, Novartis, Servier, ZS Pharma, and Vifor Pharma. Dr Rouleau has consulted for Novartis and Bayer. Dr McMurray’s employer, the University of Glasgow, was paid by Novartis for his time spent as cochairman of the PARADIGM-HF trial.

Funding/Support: The PARADIGM-HF trial was funded by Novartis Pharmaceuticals.

Role of the Funder/Sponsor: Novartis participated in the design and conduct of the PARADIGM-HF trial; collection, management, analysis, and interpretation of the data; and reviewed/approved the manuscript prior to submission for publication. The sponsor otherwise had no role in the preparation of this manuscript or the decision to submit the manuscript for publication.