The effect of spironolactone on morbidity and mortality in patients with severe heart failure 

Authors Pitt B, Zannad F, Remme W, Cody R, Castaigne A, Perez A, Palensky J, Wittes J, for the Randomized Aldactone Evaluation Study (RALES) investigators.
Source New England Journal of Medicine 341:709-17. September 2, 1999. 
Institutions University of Michigan (USA); Clinique de Nancy (France); Hopital Henri Mondor (France); STICARES Cardiovascular Research Foundation (Netherlands); Searle Pharmaceuticals (USA); Statistics Collaborative (USA).
Support Searle Pharmaceuticals.



Aldosterone has multiple negative effects in patients with congestive heart failure, including the retention of sodium, depletion of potassium and magnesium, sympathetic activation, and causing myocardial and vascular fibrosis and baroreceptor dysfunction.

Aldosterone receptor antagonists have not been widely used in severe CHF, however, because of the belief that ACE inhibitors already inhibit the production of aldosterone and the fear that combining aldosterone blockers with ACE inhibitors would lead to hyperkalemia.

There is evidence, however, that ACE inhibitors do not completely block the production of aldosterone and that low doses of aldosterone blockers in conjunction with ACE inhibitors do not often lead lead to serious hyperkalemia.

The Randomized Aldosterone Evaluation Study (RALES) was designed to investigate the effect of adding a relatively low dose of spironolactone to loop diuretics and ACE inhibitors, in patients with severe congestive heart failure.



Eligibility criteria

  • NYHA class IV heart failure and EF <= 35% within the preceeding 6 months and class III or IV at time of enrollment.
  • Treatment with a loop diuretic as well as with an ACE inhibitor (if tolerated).

Exclusion criteria

  • Cardiac: primary valvular or congenital heart disease; unstable angina; cardiac transplantation.
  • Potassium > 5.0 mmol/l  or  creatinine > 2.5 mg/dl.
  • Primary hepatic failure; active cancer.

Concommitent therapy

  • Treatment with digitalis and vasodilators was allowed.
  • Treatment with potassium sparing diuretics was not permitted.
  • Potassium supplementation was not recommended unless serum potassium was less than 3.5 mmol/l.


Patients were randomized to either 25 mg of spironolactone daily or placebo.  This dose could be increased to 50 mg daily after 8 weeks if there was evidence for progression of heart failure in the absence of hyperkalemia.

The dose could be decreased to every other day if hyperkalemia developed, but it was recommended that the dose of ACE inhibitors be decreased first.  Study medication could be withheld for medical reasons, including hyperkalemia or creatinine > 4.0 mg/dl. 



Clinical and laboratory evaluation were carried out every four weeks for the first twelve weeks, every 3 months for the rest of the first year and every 6 months thereafter.  An additional serum potassium measurement was performed at 9 weeks in those patients in whom the dose was increased at 8 weeks.



The primary endpoint was all-cause mortality, analyzed by intention to treat.

Secondary endpoints included cardiac mortality, hospitalization for cardiac causes, death or hospitalization for cardiac causes and change in NYHA class.

Predefined subgroups for analysis of efficacy were ejection fraction, cause of heart failure, serum creatinine, age, use of ACE inhibitor and digoxin.




A total of 1663 patients from 195 centers were randomized (822 to spironolactone, 841 to placebo), between March 24, 1995 and December 31, 1996. Analysis was terminated on August 24, 1998 (16 months early) after interim analysis demonstrated efficacy of the study drug exceeding preset limits.

The groups were well matched at baseline. Pertinent characteristics included (approximate average of both groups, unless otherwise indicated):

  • Age 65; white 86%; male 73%
  • BP 122/75; heart rate 81
  • NYHA
    • Class III: placebo - 69%;  spironolactone - 72% (not significantly different)
    • Class IV: placebo - 31%;  spironolactone - 27% (nsd)
  • LV ejection fraction: 25.5%;  cause of CHF ischemic in 55%
  • Medications
    • All were on loop diuretics
    • 95% were on an ACEI. Mean dose of captopril was 62 mg, of enalapril 15 mg, of lisinopril 14 mg.
    • Digitalis 74%
    • Aspirin 37%
    • Beta blockers 10%
    • Potassium supplements 28%

Discontinuation of treatment occurred in 200 patients in the placebo group and in 214 in the spironolactone group (for lack of response, adverse effects or administrative reasons).

Mean follow-up was 24 months.



Cause of death



RR and P-value

All-cause 284 (35%) 386 (46%) 0.70;  p<0.001
Cardiac 226 314 0.69;  p<0.001
Progressive CHF 127 189 0.64;  p<0.001
Sudden death 82 110 0.71;  p=0.002
MI 17 15  
Non-cardiac or unknown 58 72  

The relative risk reduction was similar across all of the following subgroups: median age (< or > 67), LV ejection fraction (< or > 26%), cause of CHF (ischemic or not), median creatinine (< or > 1.2 mg/dl), use of digitalis, use of ACEI, sex, median serum potassium (< or > 4.2 mmol/l), NYHA class (III or IV), beta-blocker use, and use of potassium supplements.



A total of 336 patients in the placebo group (40%) were hospitalized for cardiac causes, vs. 260 in the spironolactone group (32%), for a RR of 0.70 (p<0.001).  Most of these patients were hospitalized for worsening CHF.


NYHA class

Change in NYHA class Placebo Spironolactone
  Improved   33%   41%
  Same   18%   21%
  Worsened or died   48%   38%


Adverse effects

During the course of the study, the heart rate, blood pressure, weight and serum sodium concentrations did not differ between the two groups.

In the spironolactone group, the median potassium concentration rose by 0.3 mmol/l and the serum creatinine rose by approximately 0.1 mg/dl.  There was no change in the placebo group.

Overall, adverse events were reported in 79% of patients taking placebo and in 82% of those on spironolactone.  Discontinuation because of adverse events occurred in only 5% of placebo patients and 8% of spironolactone patients.  Most of the adverse events occurred with equal frequency in both groups, including serious hyperkalemia (10 patients in the placebo group, 14 in the spironolactone group).

The only adverse effect which happened significantly more often in patients in the spironolactone group was gynecomastia or breast pain, which occurred in 10% of men in this group, vs 1% in the placebo group. 


Author's discussion

The authors note that the beneficial effects of spironolactone on morbidity and mortality demonstrated in this study were observed after 2 to 3 months of treatment and persisted throughout the study.

Recent research has shown that aldosterone can cause myocardial and vascular fibrosis, direct vascular damage, baroreceptor dysfunction and can prevent the uptake of norepinephrine by the myocardium.  The authors speculate that the beneficial effects of aldosterone receptor blockade on progressive heart failure and sudden death are in part due to blocking these effects of aldosterone.  The beneficial effects were seen in patients who were almost all receiving ACE inhibitors; this implies that the reduction of aldosterone production by blockade of the angiotensin system alone is not sufficient to prevent the ill effects of aldosterone in heart failure.

Although some of the effects of spironolactone could be due to natriuresis and retention of potassium, the authors point out that there was no evidence of a significant hemodynamic effect of spironolactone and they do not consider the rise in serum potassium in spironolactone treated patients to be clinically significant.

Significant hyperkalemia was rare, which the authors attribute to the relatively low dose of spironolactone used (mean dose 26 mg) and to the exclusion of patients with elevated serum creatinine or potassium at baseline. Although gynecomastia and breast pain were significant problems in men treated with spironolactone, they only led to discontinuation of the drug in 2% of men and may well be a much lesser problem with newer aldosterone receptor blockers that are under investigation.




This study of spironolactone in patients with severe congestive heart failure revealed a significant reduction (by about 30% relative risk) in death and hospitalizations among treated patients.  These patients were receiving reasonable therapy with diuretics, ACE inhibitors and digitalis, and, although only 10% were on beta-blockers, the benefit was, if anything, greater in this subgroup.  The results of the study should be applicable to a fairly large segment of patients with severe CHF, since the exclusion criteria should not rule out the majority of patients in clinical practice.

The authors do not state whether or not baseline aldosterone levels were obtained in these patients.  If they were, it would be interesting to see whether the benefit of spironolactone was concentrated in patients with higher aldosterone levels.  If so, the drug could be targetted to those who would benefit the most.

I was initially surprised to see a major pharmaceutical company like Searle sponsoring a trial of such an old drug, but Searle is also developing a newer aldosterone receptor blocker, eplerenone, which may cause less gynecomastia. Whether or not this newer drug will be as effective in CHF remains to be seen.

September 7, 1999


References related to this article from the NLM's PubMed database. 

Reader Comments

September 12, 1999
From: Hans Liedholm []

Dear Michael,

We read with interest the publication of the results of the RALES study as this adds to the knowledge on the pharmacological treatment of severe heartfailure (1). Earlier this year two trials on treatment with selective beta-blockers were published (2,3). Bisoprolol (CIBIS- II study) and metoprolol (MERIT-HF trial) respectively showed a significant effect on mortality compared to placebo when added to conventional treatment. The impact of the results of these trials is considerable and will basically change our treatment of (systolic) heart failure.

We want to make two points. Firstly, none of these trials has been properly reported according to the CONSORT guidelines for reports of randomised controlled trials (4). How many patients were initially involved and from where did they come? The omission of information on eligible patients is hard to understand, as the external validity depends on the representativity of included patients. Secondly, it would be clarifying if all major trials also reported the NNTs (with 95 % confidence intervals).

Yours Sincerely,

Hans Liedholm, MD PhD, Agneta Björck Linné, B Pharm
The Drugs and Therapeutics Committee, Malmö University Hospital, S-205 02 Malmö, Sweden.


1. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky J, Wittes J, for the Randomized Aldactone Evaluation Study Investigators. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341: 709-17.

2. CIBIS-II Investigators. The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised trial. Lancet 1999;353:9-13.

3. MERIT-HF Study Group. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL randomised intervention trial in congestive heart failure (MERIT-HF). Lancet 1999;353:2001-7.

4. Altman DG. Better reporting of randomised controlled trials: the CONSORT statement. BMJ 1996;313:570-1.

Date: September 22, 1999
From: Sergio Emanuel Kaiser []

In the latest issue of Circulation (Circ 1999;100:1311) the SOLVD investigators published a restrospective analysis addressing the relationship between use of diuretics and arrhythmic deaths. The relative risks of deaths according to the class of diuretic used is shown below. The influence of non-potassium sparing diuretics was independent of group assignment, i.e., enalapril or placebo.

Relative risk of arrhythmic death according to diuretic use in the SOLVD population

No diuretic 1.00
Any diuretic 1.37 (1.08–1.73) p=0.009
Non–potassium-sparing diuretic 1.33 (1.05–1.69) p=0.02 
Potassium-sparing diuretic 0.90 (0.61–1.31) 0.6

Therefore, it is possible that part of the mortality reduction observed in the RALES study could be attributed to the potassium-sparing action of spironolactone.

I also have some doubts whether we might extrapolate the RALES results to a larger population under beta-blocker use, since only 10% were effectively taking such drugs. Nowadays, the use of beta-blockers in heart failure is consistently increasing and it remains to be proved whether spironolactone adds incremental benefit beyond that of a combination of ACEI and beta-blocker to standard digitalis and diuretic therapy.

Best regards

Sergio Kaiser, MD


These are important points. The SOLVD study looked at the benefit of enalapril in patients with CHF.  The article referred to by Dr. Kaiser is a retrospective study of patients enrolled in SOLVD and found that use of non-potassium sparing diuretics, but not potassium-sparing diuretics, was associated with an increased risk of arrhythmic death.  Unfortunately, the exact type of potassium-sparing diuretics was not specified, so it is impossible to know whether the majority of patients taking potassium-sparing diuretics were taking spironolactone or triamterene (not an aldosterone antagonist).

If most of these patients were taking spironolactone, then much of the benefit seen in this subgroup in SOLVD might actually be due to aldosterone antagonism (as hypothesized here, in RALES), rather than to the prevention of potassium depletion.  If, on the other hand, a significant benefit was also present in patients taking triamterene, then one would have to postulate a beneficial effect simply due to the prevention of potassium loss.  Prevention of potassium loss alone, however, probably does not account for all of the benefit seen in RALES, since an effect was seen on death from progressive CHF also, not only from arrhythmia.

As for beta-blockers and aldosterone antagonism, a beneficial effect of spironolactone was also seen in the subgroup of patients taking beta-blockers, but as you point out, this represents only 10% of the study population.  --mj

Date: November 11, 1999
From: Agostino Colli []

In RALES the number of patients in atrial fibrillation is not stated. Even
if one could suppose it from the number of patients on digitalis ( 70%)
which is just the same in both groups, balancing with respect to this
feature is not shown.

It is true that the number of patients in atrial fibrillation was not specified.  Presumably digitalis was given for heart failure, not for fibrillation, in the majority of patients studied here.  However, I'm not sure that spironolactone would be likely to have more or less of an effect in patients in fibrillation, so even if the groups were unbalanced in this regard, I don't think it would have affected the outcome.  -- mj

January 24, 2000

From: Javier Ena (
Received: November 30, 1999

I read with great interest your article and I was surprised by the large proportion of patients that dropped out from study (24% in the placebo group and 26% in the spironolactone group). Could you detail in numbers the reasons for the drop-outs. In addition, there is some less degree of cardiac failure in patients taking spironolactone (27% in class IV in the spironolactone group; 31% in class IV in the placebo group).  Were the differences in mortality and hospitalization maintained after controlling for NYHA class?. Thanks for your answer.

J. Ena MD (Alicante. Spain)

As you note, about one quarter of patients dropped out during the course of the study.  The authors do not detail the reasons for dropping out, other than to say it was "because of a lack of response, because of adverse events, or for administrative reasons".  These patients were included in the analysis of the results.

Although there were slightly more patients in severe CHF in the placebo group, the reduction in mortality was almost identical when subgroups according to NYHA class were looked at.  -- mj

January 24, 2000

Letters to the Editor about this article, from the January 13 New England Journal of Medicine.  Topics covered include:

  • Was conventional therapy maximized?
  • Danger of hyperkalemia in diabetics
  • Danger of hyperkalemia in the geriatric population
  • Role of the beta-adrenergic system


Submit a comment

Journal Club homepage

Site Meter