Author + information
- Received May 18, 2012
- Revision received June 30, 2012
- Accepted July 24, 2012
- Published online November 13, 2012.
- Orly Vardeny, PharmD, MS⁎,⁎ (, )
- Dong Hong Wu, PhD†,
- Akshay Desai, MD, MPH†,
- Patrick Rossignol, MD‡,
- Faiez Zannad, MD‡,
- Bertram Pitt, MD§,
- Scott D. Solomon, MD†,
- RALES Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Orly Vardeny, University of Wisconsin School of Pharmacy, 777 Highland Avenue, Madison, Wisconsin 53705-2222
Objectives This study investigated the influence of baseline and worsening renal function (WRF) on the efficacy of spironolactone in patients with severe heart failure (HF).
Background Renal dysfunction or decline in renal function is a known predictor of adverse outcome in patients with HF, and treatment decisions are often on the basis of measures of renal function.
Methods We used data from the RALES (Randomized Aldactone Evaluation Study) in 1,658 patients with New York Heart Association functional class III or IV HF and an ejection fraction <35%. Participants were randomized to spironolactone 25 mg, which could be titrated to 50 mg, or placebo daily. Renal function (estimated glomerular filtration rate [eGFR]) was estimated by the Modification of Diet in Renal Disease equation. Worsening renal function was defined as a 30% reduction in eGFR from baseline to 12 weeks post-randomization.
Results Individuals with reduced baseline eGFR exhibited similar relative risk reductions in all-cause death and the combined endpoint of death or hospital stays for HF as those with a baseline eGFR >60 ml/min/1.73 m2 and greater absolute risk reduction compared with those with a higher baseline eGFR (10.3% vs. 6.4%). Moreover, WRF (17% vs. 7% for spironolactone and placebo groups, p < 0.001) was associated with an increased adjusted risk of death in the placebo group (hazard ratio: 1.9, 95% confidence interval: 1.3 to 2.6) but not in those randomized to spironolactone (hazard ratio: 1.1, 95% confidence interval: 0.79 to 1.5, p interaction = 0.009). The risk of hyperkalemia and renal failure was higher in those with worse baseline renal function and those with WRF, particularly in the spironolactone arm, but the substantial net benefit of spironolactone therapy remained.
Conclusions The absolute benefit of spironolactone was greatest in patients with reduced eGFR. Worsening renal function was associated with a negative prognosis, yet the mortality benefit of spironolactone was maintained.
Elevated aldosterone concentrations have known detrimental effects on the myocardium and renal vasculature (1). Two pivotal clinical trials in heart failure (HF) have demonstrated benefit with aldosterone receptor antagonists (2,3), and current HF treatment guidelines recommend the use of a mineralocorticoid receptor antagonist (MRA) in patients with moderately severe to severe symptoms of HF and reduced left ventricular ejection fraction and in patients after acute myocardial infarction (MI) complicated by left ventricular systolic dysfunction and HF (4,5). More recently, the EMPHASIS-HF study (Eplerenone in Mild Patients Hospitalization and Survival Study in Heart Failure) further supported the use of these agents in HF patients with milder HF symptoms (6).
Mineralocorticoid receptor antagonists are not recommended when the serum creatinine is raised above 2.5 mg/dl (or when creatinine clearance is <30 ml/min) or in those with serum potassium levels above 5.0 mmol/l. Renal dysfunction, even transient, is common in HF patients and is a known predictor of cardiovascular outcomes and mortality in patients with cardiac disease. Worsening renal function (WRF), otherwise known as cardiorenal syndrome type 2 (7), is defined as increases in creatinine or reductions in estimated glomerular filtration rate (eGFR). Worsening renal function has been associated with adverse outcomes in patients with HF, underuse of proven agents, and even discontinuation of beneficial medication (8,9).
Nevertheless, rises in creatinine or declines in eGFR are common in patients receiving inhibitors of the renin-angiotensin-aldosterone system, possibly due to alterations in renal hemodynamic status (10). In both the Survival And Ventricular Enlargement (11) and Studies of Left Ventricular Dysfunction (12) trials, worsening creatinine was not associated with adverse outcome in patients receiving the angiotensin-converting enzyme inhibitor (ACEI) captopril, whereas it was associated with increased risk in those receiving placebo. Recent data in post-MI patients from the EPHESUS (Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study) showed that early WRF with eplerenone was associated with an increased risk of adverse cardiovascular outcomes, yet the benefit of eplerenone was maintained (13). The prognostic significance of WRF in patients with moderate to severe HF receiving spironolactone, however, remains unknown.
We used data from the RALES (Randomized Aldactone Evaluation Study) to determine the influence of baseline renal function on efficacy of the MRA spironolactone in HF patients and the prognostic importance of WRF in HF patients receiving an aldosterone antagonist. We hypothesized that WRF would be associated with a more benign prognosis in patients receiving an MRA compared with patients receiving placebo.
The RALES study was a double-blind, randomized, placebo controlled trial that was designed to assess the efficacy of spironolactone on prevention of all-cause mortality and cardiac-related hospital stays in patients with New York Heart Association functional class III or IV HF. Participants were enrolled if they had a left ventricular ejection fraction <35% while taking background ACEIs and diuretics. Exclusion criteria were primary valvular disease, congenital heart disease, unstable angina, liver failure, listing for cardiac transplant, active cancer, or any other life-threatening disease. Patients with serum creatinine >2.5 mg/dl or potassium >5 mmol/l were also excluded. Participants were randomized to receive spironolactone 25 mg or placebo daily. After 8 weeks, the dose could be increased to 50 mg daily for patients with signs and symptoms of progression of HF without evidence of hyperkalemia. Serum potassium and creatinine were measured at 4, 8, and 12 weeks during the titration phase and every 3 months thereafter during the study and were available in 1,658 of the 1,663 patients enrolled in the study. Concomitant treatment with digoxin and vasodilators was allowed, and the use of potassium-sparing diuretics was not permitted. Oral potassium supplement use was discouraged unless hypokalemia (defined as a serum potassium concentration of <3.5 mmol/l) developed.
We defined reduced baseline eGFR as <60 ml/min/1.73 m2 and WRF as a 30% reduction in eGFR (14) from baseline (calculated by the Modification of Diet in Renal Disease equation) at any time during the titration phase (through week 12) after randomization. Baseline demographic data between participants with eGFR >60 ml/min/1.73 m2 and those with eGFR <60 ml/min/1.73 m2 and between those with WRF and those without WRF were compared with identify potential differences. Between-group assessments were performed with t tests for continuous variables and chi-square or Fisher exact tests, as appropriate, for categorical variables. Hyperkalemia was defined as a potassium level ≥5.5 mmol/l at any visit or a serious adverse event related to hyperkalemia at any time during study follow-up. For WRF, we performed a sensitivity analysis with another definition of >0.3 mg/dl increase in serum creatinine with qualitatively similar results (not shown).
Cox proportional hazards regression models were used to examine associations between baseline renal function and all-cause mortality and the combined endpoint of death or HF hospital stay as well as the effectiveness of treatment with spironolactone. Paired t-tests were used to assess differences in eGFR by treatment arm during the titration phase. A WRF during the titration phase (through week 12) was related to subsequent long-term outcomes in a landmark analysis. We further performed these analyses, adjusting for the following covariates: age, sex, race, HF etiology, history of diabetes, MI, angina, hypertension, baseline blood pressure, ejection fraction, baseline potassium and creatinine, treatment, baseline medications (aspirin, ACEI/angiotensin receptor blocker [ARB], beta-blocker, loop diuretic, digoxin), and eGFR or WRF × treatment interaction term. All analyses were conducted with Stata (version 11, StataCorp, College Station, Texas).
Of 1,658 patients included in these analyses, 792 (48%) had a baseline eGFR <60 ml/min/1.73 m2, and 866 (52%) had a baseline eGFR ≥60 ml/min/1.73 m2. Baseline characteristics by eGFR are shown in Table 1. Participants with a baseline eGFR <60 ml/min/1.73 m2 were older, more likely to be female, diabetic, Caucasian, and were more likely to have an ischemic etiology for HF and less likely to be taking aspirin, diuretics, and digoxin at study entry.
WRF occurred in 199 (12%) patients during titration (Table 2). The percentage of patients with WRF was significantly higher in the spironolactone group than in the placebo group (17% vs. 7%, p < 0.001). Although eGFR was similar in the spironolactone and placebo groups at baseline (65.3 ± 23.1 vs. 64.5 ± 22.8, p = 0.46), eGFR declined at 4 weeks in the spironolactone arm (62.6 ± 23.8 vs. 65.5 ± 26.1, p = 0.02) and remained reduced at the end of titration (62.4 ± 25.0 vs. 65.4 ± 23.6, p = 0.02). Renal function continued to decline in both groups; however, differences in eGFR between spironolactone and placebo groups were no longer significant by 6 months after randomization. The proportion of patients with WRF was similar in those who started with baseline eGFR below and above 60 ml/min/1.73 m2 (11% vs. 15%, p = 0.22). Patients with worsening eGFR were older and more likely to be female. There were no significant differences in baseline characteristics between the placebo and spironolactone groups in patients with or without WRF. Of note, blood pressure at the end of titration (week 12) was not different between participants with and without WRF. In addition, patients with WRF were more likely to have received ACEIs or ARBs (34% vs. 26%, p = 0.01), digoxin (22% vs. 14%, p = 0.004), or loop diuretics (41% vs. 33%, p = 0.02) during titration, although this did not differ by treatment group.
Outcomes by baseline eGFR and WRF
Baseline eGFR lower than 60 ml/min/1.73 m2 was associated with similarly increased risk for mortality in both treatment groups (Fig. 1A,Table 3), even adjusting for baseline covariates. The risk for the combined endpoint of death or HF hospital stay was similarly increased in crude estimates in both groups (Fig. 1B, Table 3) but not significantly different when adjusted for baseline covariates. There were a total of 670 deaths (386 placebo, 284 spironolactone), similarly distributed between eGFR < or ≥60 ml/min/1.73 m2 (402 vs. 436 deaths). For the combined endpoint of all-cause death or HF hospital stays, there were 909 events (522 placebo, 387 spironolactone); 483 for eGFR <60, and 422 for those with eGFR ≥60 ml/min/1.73 m2. At the end of titration (12 weeks), 63 deaths occurred in the placebo group, and 54 deaths occurred in the spironolactone group. For HF hospital stays, 105 occurred in the placebo group, and 51 occurred in the spironolactone group by the end of titration. When stratifying by baseline eGFR, 80 deaths occurred in eGFR <60, and 37 occurred in those with eGFR ≥60. For HF hospital stays, 74 occurred in eGFR <60, whereas 81 occurred in eGFR ≥60. Baseline renal function, modeled as a continuous or categorical variable, did not modify the benefits of spironolactone, with an approximately 30% relative risk reduction for mortality regardless of baseline eGFR, and a similar risk reduction for the combined endpoint of death or HF hospital stay (Fig. 1A, Table 4). The absolute risk reduction for mortality was substantially higher in patients with worse baseline eGFR (10.3% vs. 6.4%), and the absolute risk reduction for death or HF hospital stay at 2 years was 13.7% in the lower eGFR group, compared with 12.7% in the higher eGFR group.
Worsening renal function, defined as a 30% reduction in eGFR during the titration period, was associated with an increased subsequent long-term risk of death in the placebo group, even when adjusted for baseline covariates (hazard ratio: 1.9, 95% confidence interval [CI]: 1.3 to 2.6) (Figs. 2A and 2B, Table 3). In contrast, in the spironolactone arm, WRF was not associated with an increased risk for death, with a highly significant interaction between treatment and WRF with respect to outcome (hazard ratio: 1.1, 95% CI: 0.79 to 1.5; p interaction = 0.009). For the combined outcome of death or hospital stay for HF (Figs. 2C and 2D, Table 3), WRF was similarly associated with a significantly increased risk in the placebo group, and this risk was also substantially attenuated in the treatment arm (p-interaction 0.04). Patients randomized to spironolactone derived benefit whether or not renal function worsened during titration, with no attenuation of the approximately 30% reduction in mortality, and mild attenuation in the combined endpoint, when making even the most conservative comparison between the treatment group with WRF and the placebo group without WRF (Fig. 2, Table 4).
Hyperkalemia or adverse event rates by baseline eGFR and WRF
Hyperkalemia defined as potassium >5.5 mmol/l at a study visit or a hyperkalemia adverse event at any time during follow-up occurred more frequently in participants with reduced baseline eGFR and particularly more frequently in those with reduced eGFR who received spironolactone compared with those with reduced eGFR receiving placebo (25.6% vs. 8.5%, p < 0.001) (Table 5). The increased risk of hyperkalemia due to spironolactone was higher in those with lower eGFR compared with those with higher eGFR (odds ratio: 1.53, 95% CI: 1.16 to 2.02). Those with baseline eGFR <60 who were taking spironolactone were most likely to undergo a dose reduction or discontinuation during titration.
Patients with WRF had a higher overall risk of hyperkalemia, and this risk was even greater in those receiving spironolactone compared with those who received placebo (Table 5). The increased risk of hyperkalemia in patients receiving spironolactone was most evident if renal function worsened (odds ratio: 3.6, 95% CI: 1.5 to 8.6, compared with participants receiving spironolactone without WRF). Finally, patients with WRF were more likely to have their study drug dose reduced or discontinued during the titration period, although this also appeared true whether they were receiving placebo or spironolactone.
In this analysis of patients with moderately severe to severe HF randomized to an MRA or placebo, individuals with reduced baseline eGFR exhibited similar relative risk reductions in all-cause death and the combined endpoint of death or hospital stays for HF as those with a baseline eGFR >60 ml/min/1.73 m2 and greater absolute risk reduction compared with those with a higher baseline eGFR. Moreover, WRF—defined in this study as a 30% reduction in baseline eGFR—was associated with an increased risk of death in the placebo group but not in those randomized to spironolactone, and the risk of the combined endpoint of death or HF hospital stay in those with WRF was increased in the placebo group but markedly attenuated in those receiving spironolactone. The risk of hyperkalemia was higher in those with worse baseline renal function and those with WRF, particularly in the spironolactone arm, but there remained substantial net benefit.
Spironolactone was associated with an approximate 30% reduction in all-cause mortality in the RALES study, and the more selective MRA eplerenone has shown similar benefit in several other studies. In this analysis, we noted a similar relative benefit and greater absolute benefit in patients with baseline eGFR <60 ml/min/1.73 m2 treated with spironolactone, a finding that is similar to that observed in post-MI patients with lower eGFR treated with an ACEI (15) and in a small study suggesting improvement in myocardial systolic and diastolic left ventricular function in patients with chronic kidney disease when an MRA was used in addition to ACEIs or ARBs (16,17). Although we observed an increased risk of adverse events, particularly hyperkalemia, in these patients, there was still a substantial net benefit to the use of MRAs in HF patients with reduced eGFR.
Although WRF during the titration phase occurred more frequently in patients randomized to spironolactone, the risk associated with this worsening was greatest in patients in the placebo group and was markedly attenuated in those taking spironolactone. Indeed, WRF was not associated with an increased risk for mortality in patients randomized to spironolactone, and the risk for death or HF hospital stay was markedly attenuated in those receiving active treatment. These data suggest that elevation of creatinine (and worsening of GFR) in the setting of spironolactone therapy has far less prognostic importance than worsening of renal function without inhibiting the renin-angiotensin-aldosterone system, a finding that is similar to those seen with ACE inhibition post-MI and in HF (11,12). These findings suggest that the reduction in eGFR associated with renin-angiotensin-aldosterone inhibition do not necessarily reflect kidney injury but might result from MRA-induced reduction in blood pressure affecting renal blood flow, a notion that is further supported by the lack of worsening of renal function after an initial relatively early decline seen in both the EPHESUS (13) and RALES studies.
A number of potential mechanisms might explain the benefit of MRAs in the context of renal dysfunction, independent of the potential hemodynamic effects. Up-regulation of mineralocorticoid receptor density and activity in the heart and kidney might play an important role in aldosterone-mediated organ damage (18,19). Aldosterone induces proteinuria, glomerular mesangial injury, and tubulointerstitial fibrosis in rat models, and treatment with MRAs prevented renal injury secondary to aldosterone administration (20,21). Additionally, the incidence and magnitude of proteinuria or albuminuria are higher among patients with primary aldosteronism compared with patients with essential hypertension (22,23). The use of an MRA alone or in combination with ACE inhibition or ARB therapy was shown to produce additive effects to the reduction of proteinuria in small randomized studies, which could suggest that aldosterone might play an important role in causing renal injury (24–26). Whether these mechanisms account for the preservation of the benefit of MRAs in HF patients remains unclear, and it is likely that improved outcomes in patients with renal dysfunction are more attributable to a reduction in cardiac fibrosis and remodeling similar to patients with normal renal function.
Hyperkalemia is a well-recognized adverse outcome in patients treated with inhibitors of the renin-angiotensin-aldosterone system. We noted a significantly increased risk for hyperkalemia in patients with reduced baseline renal function and in those with WRF after spironolactone. Higher rates of hyperkalemia and related adverse events have been noted in clinical practice compared with clinical trials (27), a finding that might be secondary to less close monitoring, more frequent dietary indiscretions, or use of concomitant medications. These findings underscore the importance of close monitoring of electrolytes with MRAs, particularly in patients with renal dysfunction, and the careful assessment of risk versus benefit of MRAs in the setting of rising potassium levels.
Because this analysis was not pre-specified, the results should be interpreted with caution. In particular, a few of the subgroups in this analysis had small sample sizes, therefore point estimates noted are not definitive. However, these results are consistent with those from the EPHESUS study, which tested eplerenone in a post-MI HF population, and with those from the SAVE (Survival And Ventricular Enlargement) and SOLVD (Studies of Left Ventricular Dysfunction) studies, in which an ACEI was tested in a post-MI population and HF populations. Although we defined WRF as a 30% reduction in eGFR, other definitions have been used in prior analyses. Nonetheless, we performed a sensitivity analysis exploring an alternate definition of WRF, defined as a >0.3-mg/dl increase in creatinine, and noted qualitatively similar results (data not shown). Measurements of serum creatinine were not blinded during this study. As such, imbalances in the use ACEIs could have occurred between placebo and spironolactone groups such that more subjects in the placebo group received lower or no doses of ACEIs due to dose adjustments in response to rising serum creatinine. This might have accounted, in part, for the attenuation in risk noted in the spironolactone group. The RALES study excluded individuals with baseline serum creatinine >2.5 mg/dl (eGFR <30 ml/min) and those with serum potassium levels >5.0 mmol/l. We cannot extrapolate these results to patients with more severe renal dysfunction or eGFR <30 ml/min. We could not fully adjust for all potential confounding factors to worsening renal function in our statistical models, such as hospital stays, exposure to contrast or other nephrotoxic agents, or occurrence of acute renal injury. Because more hospital stays occurred in the placebo group, these other factors could have had a greater role in contributing to the outcomes. Beta-blocker usage in the RALES study was low, and beta blockers are known to enhance the risk for hyperkalemia and might affect the relative benefit of MRAs. As such, it is unclear whether the same degree of benefit among those with renal dysfunction and WRF would have been observed if more participants were also receiving beta blockers. Finally, in this analysis, we assessed WRF only during the titration period of the study, although our data suggest that declines in renal function associated with spironolactone occurred early in the course of treatment.
We found that in patients with advanced HF, those with reduced baseline eGFR receiving spironolactone exhibited similar risk reduction in all-cause mortality and the combined outcome of all-cause mortality and hospital stay for HF, compared with patients with higher baseline eGFR. Moreover, individuals randomized to spironolactone derived benefit, regardless of whether renal function worsened during the titration period. Nevertheless, these benefits occurred at the expense of an increased risk of hyperkalemia, which was more common in patients with reduced baseline eGFR and those with WRF, particularly when randomized to spironolactone. These findings suggest that patients with HF and renal dysfunction still benefit from an MRA yet argue that close monitoring of electrolytes is warranted in this setting.
The RALES trial was funded by Searle (Skokie, Illinois). Dr. Desai is a consultant for Novartis, Boston Scientific, Reata, and Intel; and received a research grant from AtCor Medical, Inc. Dr. Rossignol has received a travel grant from Pfizer. Dr. Zannad received consultancy fees from Bayer, BiomÃ©rieux, Biotronik, BostonScientific, CVCT, Novartis, and Pfizer, Resmed, Servier, and Takeda; and received grants to institution BG-Medicine, Roche Diagnostics. Dr. Pitt is a consultant for Pfizer, Merck, Bayer, Novartis, Takeda, Lilly, Bristol-Myers Squibb, Relypsa, BG-Medicine, and AstraZeneca. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme inhibitor
- angiotensin receptor blocker
- confidence interval
- estimated glomerular filtration rate
- heart failure
- myocardial infarction
- mineralocorticoid receptor antagonist
- worsening renal function
- Received May 18, 2012.
- Revision received June 30, 2012.
- Accepted July 24, 2012.
- American College of Cardiology Foundation
- Pitt B.,
- Stier C.T. Jr..,
- Rajagopalan S.
- Hunt S.A.,
- Abraham W.T.,
- Chin M.H.,
- et al.
- Jose P.,
- Skali H.,
- Anavekar N.,
- et al.
- Testani J.M.,
- Kimmel S.E.,
- Dries D.L.,
- Coca S.G.
- Rossignol P.,
- Cleland J.G.,
- Bhandari S.,
- et al.
- Bellomo R.,
- Ronco C.,
- Kellum J.A.,
- Mehta R.L.,
- Palevsky P.
- Tokmakova M.P.,
- Skali H.,
- Kenchaiah S.,
- et al.
- Edwards N.C.,
- Steeds R.P.,
- Stewart P.M.,
- Ferro C.J.,
- Townend J.N.
- Navaneethan S.D.,
- Nigwekar S.U.,
- Sehgal A.R.,
- Strippoli G.F.