The risk of hyperkalaemia in patients with heart failure has increased in the past few years together with the evolution of pharmacological treatment for these patients. This significant change has been associated with the introduction of angiotensin converting enzyme (ACE) inhibitors, angiotensin II receptor blockers (ARBs), and aldosterone antagonists. High potassium concentrations in heart failure could lead to life threatening events, and therefore should be taken seriously. In this review we summarise the information about potassium homeostasis in heart failure and the current risk of developing potentially serious hyperkalaemia, particularly in association with the use of aldosterone antagonists.
- Heart failure
Statistics from Altmetric.com
Heart failure treatment has evolved considerably over the past decades. New treatments have resulted in significant improvements in outcomes. However, we now see some side effects that were rarely or not seen before, such us hyperkalaemia. Many years ago, heart failure treatment was based on high doses of diuretics and digoxin, and consequently hypokalaemia was a frequent and feared side effect.1 2 The neurohumoral hypothesis induced a significant change in heart failure treatment, and several medications which interfered in the renin–angiotensin–aldosterone (RAA) system were then found to be of therapeutic value.3 These led to a reduction in diuretic doses, resulting in less hypokalaemia.4 5 In the past few years the risk of hyperkalaemia has significantly risen in these patients as a consequence of the interference of the RAA system. The purpose of this review is to analyse the relationship between current heart failure treatment and the risk of developing hyperkalaemia, particularly with the use of aldosterone antagonists.
Pathophysiology of hyperkalaemia in heart failure
Under normal conditions, potassium is filtered through the kidney glomerulus and approximately 90% is reabsorbed in the proximal convoluted tubule and loop of Henle. In the distal convoluted and collecting tubules, intraluminal potassium concentration is directly related to the amount of proximally non-reabsorbed potassium plus the actively secreted potassium. Potassium elimination is partly regulated by the RAA system,6 7 through the activation of the aldosterone receptor, that leads to increased potassium and magnesium elimination.8 9 Aldosterone plasma concentration increases by the stimulus of low intrarenal vascular pressure, β -adrenergic stimulus, prostaglandin I2, and low plasma sodium concentration. However, several circumstances may reduce aldosterone secretion or its renal effect, resulting in an increase of potassium in plasma. Low plasma aldosterone concentrations can result from intrinsic renal parenchyma diseases, due to renin synthesis reduction, and primary or secondary hypoaldosteronism. The latter may result from angiotensin converting enzyme (ACE) inhibitors, angiotensin II receptor blockers (ARBs), β-blockers, and other drugs. It may also occur as a result of a functional reduction in the aldosterone receptor due to pharmacologic antagonism (with espironolactone or eplerenone) or renal diseases.
Another mechanism that reduces potassium elimination is a significant drop in the sodium concentration in the distal convoluted tubule.8 The abnormally high adrenergic system and RAA system activation in heart failure, in combination with other hormones such as endothelins and vasopressin, induce water and sodium reabsorption in distal convoluted and collecting tubules. In mild heart failure potassium elimination is usually not altered; however, in advanced heart failure, RAA system activation is remarkably high leading to an exaggerated water and sodium reabsorption resulting in a significant drop in sodium intraluminal concentration. The sodium available for exchange is very low and potassium elimination is not adequate, even in the presence of high aldosterone concentrations. Therefore, heart failure per se could lead to hyperkalaemia even without other evidence of renal dysfunction (ie, increased creatinine concentrations). In summary, there are many alterations that may lead to high potassium concentrations in heart failure: first, renal parenchyma may be altered in many concomitant diseases, including diabetes, hypertension or chronic ischaemia; second, the use of ACE inhibitors, ARBs, β-blockers and spironolactone; and third, the high adrenergic stimulus increases the risk of hyperkalaemia interfering with the mechanism of potassium elimination (figure 1).8–10
Potassium abnormalities in heart failure
The treatment of heart failure based on diuretics and digoxin frequently induced hypokalaemia.2 11 In 1976 Davidson et al reported an incidence of hypokalaemia of 15.6% in a heart failure population, even though many patients were on potassium supplements.2 Hypokalaemia frequently induced serious episodes of arrhythmia, particularly in patients treated with digoxin. However, in the past few years episodes of hypokalaemia have become less frequent and hyperkalaemia has been increasingly reported in heart failure. This fact can be associated with several changes in heart failure management. First, we already mentioned that loop diuretic dose was reduced as the clinical condition of many heart failure patients improved, and the indication for potassium sparing diuretics is now unusual.12 13 As a result of this, less urinary potassium elimination is expected. Second, the introduction of ACE inhibitors and ARBs has raised the potassium concentration in heart failure patients. However, hyperkalaemia is unusual with the use of these medications.5 14–16 Finally, spironolactone was found to increase survival in heart failure, and its use is now widespread in these patients. This medication, combined with ACE inhibitors or ARBs, definitively changed the pattern of potassium concentrations in heart failure patients, resulting in an increase incidence of hyperkalaemia.
Aldosterone antagonism: the evidence favouring its use
Aldosterone values are high in patients with heart failure, and this has been shown to be detrimental for the myocardium.17 This led to the hypothesis that its antagonism could improve outcomes in this disease. The RALES investigators performed a safety study regarding the use of spironolactone on a small chronic congestive heart failure (CHF) cohort. Of note, the control group in this study presented hypokalaemia in 10% of the cases.18 Contrarily, those patients assigned to 50 and 75 mg of spironolactone had no episodes of hypokalaemia, and there was just one episode in those patients receiving 25 mg/day. A bigger study then showed that spironolactone significantly reduced mortality in heart failure patients with New York Heart Association (NYHA) functional class III/IV and low ejection fraction (<35%) (box 1).19This study showed a significant reduction in mortality of approximately 30%. The impact on mortality was consistent in both progressive heart failure death and sudden death. Subgroup analysis showed similar improvement in mortality regardless of creatinine or potassium concentration or the use of potassium supplements. The indication for spironolactone use in heart failure substantially increased after the trial was published. A recent heart failure population analysis regarding the use of spironolactone stressed the importance of this medication in improving prognosis.20 Interestingly, it was expected that spironolactone would increase the effects of ACE inhibitors and loop diuretic dose reduction on potassium values, as the safety study suggested. However, the RALES study did not report an increment in the incidence of severe hyperkalaemia (>6 mEq/l). The spironolactone group showed a non-significant difference compared to controls (2% (n=14) vs 1% (n=10)). No data were provided regarding milder hyperkalaemic episodes (ie, between 5.5–6.0 mEq/l). Some matters should be considered in support of these findings. Electrolyte check-up was rigorous after starting with the randomised treatment. During the initial 4 weeks of randomisation potassium was checked weekly, and then every 3 months for 1 year and every 6 month thereafter; therefore severe hyperkalaemia episodes may have been prevented by detection of milder ones and the treatment modifications secondary to them. It should also be mentioned that the use of spironolactone in other diseases such as hypertension did not induce higher rates of hyperkalaemia, either used alone or in combination with other diuretics or ACE inhibitors.21–23
Box 1 The RALES study
Population: 1663 left ventricular ejection fraction ≤35% and New York Heart Association (NYHA) functional class III or IV.
Design: randomised, double blind (25 mg of spironolactone vs placebo).
Primary end point: death from any cause.
Published: September 1999.
Follow-up: mean of 24 months.
Primary end point results:
25% relative risk reduction (95% CI 15% to 33%).
11% absolute risk reduction (95% CI 7% to 16%).
Number needed to treat (NNT)=9 (95% CI 6 to 15).
Hyperkalaemia >6 mEq/l: 1% vs 2% NS (placebo vs treatment group).
Hyperkalaemia <6 mEq/l: were not reported.
Eplerenone, a new aldosterone antagonist, has been studied in the past few years. The randomised trial EPHESUS showed a significant reduction in death from any cause in patients with low ejection fraction and heart failure after a myocardial infarction.24 Aldosterone antagonism induced an increase of serious (>6 mEq/l) hyperkalaemia episodes (5.5 vs 3.9% vs placebo, p<0.001) in the first year of follow-up. However, this was balanced by a significant reduction in hypokalaemia (<3.5 mEq/l) episodes (8.4% vs 13.1% in the eplerenone and placebo group, respectively). This effect could theoretically lead to a significant reduction in the risk of arrhythmia in heart failure, which can be even greater since aldosterone antagonism reduces renal magnesium excretion.25 In fact, EPEHSUS showed a 21% reduction in the relative risk of sudden death, and therefore it is likely that the favourable hypokalaemia effect was much more important than the potentially dangerous hyperkalaemia risk. Regarding this point, it is interesting to consider mortality according subgroups. When death was analysed according to potassium concentration (with a cut-off value of 4 mEq/l) there was no difference in all cause mortality (p=0.29), but this was significant when considering the cardiovascular death or hospitalisation for cardiovascular events (p=0.02), favouring those patients with potassium concentration >4 mEq/l. When the analysis was done according to creatinine value (< or >1.1 mg/dl) there was a significant effect (p=0.02) in all cause of death (with improvement only in those patients with low creatinine), but this was not the case when analysing cardiovascular death or hospitalisation for cardiovascular events. Extreme caution should be taken when considering the results of these subgroup analyses due to the lack of power of the trial, but they flag the complex interaction between electrolytes, renal function and aldosterone antagonism. Finally, these findings also stress the importance of checking electrolytes and renal function when using aldosterone antagonists in order to avoid hypo- or hyperkalaemia. Although the overall effect of aldosterone antagonism in randomised trials suggests strong benefits with their use, it cannot be ruled out that this benefit is lost in those patients presenting with severe hyperkalaemia due to the presence of serious arrhythmic events.
These two randomised studies definitely provide a role for aldosterone antagonism in the treatment of heart failure, but it should be pointed out that they addressed different populations. The RALES study included patients with advanced and chronic heart failure and the EPHESUS included post-myocardial infarction patients with low ejection fraction complicated by heart failure. Despite this distinction, the use of spironolactone is more widespread than eplerenone in the treatment of heart failure, and therefore the epidemiologic impact on the potassium concentration is mainly from the former medication.
Aldosterone antagonism: the evidence relating to potassium
In contrast to the findings of the randomised trials regarding potassium, numerous population based papers have now reported a higher rate of hyperkalaemia in heart failure patients treated with spironolactone and ACE inhibitors.12 18 19 26–36 Table 1 shows the published papers that looked at the association between hyperkalaemia and spironolactone in heart failure. Rates of hyperkalaemia range from 5–36% after the publication of the RALES study. For example, Cruz et al published their observation showing an incidence of 14% of serious hyperkalaemia (>6 mEq/l).27 Cases occurred only in those patients treated with the combination of spironolactone and ACE inhibitors. The odds of presenting with hyperkalaemia were 24-fold higher for spironolactone treated patients. The findings of this and many of the studies shown in table 1 are different from the RALES study because they are observational and retrospective in nature, but also because they include a different population. In many cases differences are evident in age, diabetes or renal function. There were also differences in medical treatment, where in some studies the spironolactone dose was higher, or patients were taking β-blockers or potassium supplements more frequently than the patients in the RALES study.
Considering the results of RALES and these previous studies, it is interesting to remark the experienced published by Tamirisa et al.31 These authors performed a case–controlled analysis of hyperkalaemia in patients with heart failure on spironolactone. Remarkably, the indication for spironolactone and the laboratory follow-up was in accordance with the RALES protocol. They found that plasma potassium values increased significantly in the spironolactone population compared to basal values, but severe hyperkalaemia occurred in only 1.6% of the cases, a value similar to the one found in the RALES study. The authors concluded that under appropriate medical supervision spironolactone can be safely prescribed in heart failure, and it is certainly possible that adopting a follow-up strategy in patients at risk could cut down the incidence of severe hyperkalaemia.
One problem that can be noted in table 1 is that there is a wide discrepancy in the information available regarding the true incidence of hyperkalaemia in heart failure. However, an insight into the real problem can be seen from some of the recently published population studies. In these papers, the rate of prescription before and after the publication of the RALES study was analysed in large provincial databases, but not specifically in heart failure cohorts. Juurlink et al did a retrospective analysis of patients with heart failure >65 years medicated with ACE inhibitors in the province of Ontario, Canada.13 They focused on the period between 1993 and 2001 and analysed the prescription rate of spironolactone and admissions and hospital death secondary to hyperkalaemia. The result showed that after publication of the RALES study in August 1999 the prescription rate for spironolactone rose significantly from 4% to 14.9%. Concomitantly, admissions related to hyperkalaemia increased from 0.24% to 1.1%, and death associated with hyperkalaemia increased from 0.03% to 0.2%. Due to the observational nature of the study it is not possible to infer causality but it strongly suggests that the use of spironolactone is at least partly related to the higher incidence of complications related to hyperkalaemia.
Two subsequent reports showed similar results. Another Canadian study showed a threefold increase in spironolactone prescription rate after RALES study.37 A US study found a sevenfold increase in spironolactone prescription.38 Therefore, a significant increase in the episodes of hyperkalaemia in heart failure patients probably has happened since the publication of the RALES study (figure 2). These papers also suggest that the prescription of spironolactone expanded to different populations than those described in RALES, and to patients without any evidence proven benefit. Based on the inclusion criteria for RALES, these studies showed a high incidence of inappropriate prescription of spironolactone (30.9%,37 58%12 and 75%38). Reasons to consider prescriptions as inadequate were patients without assessment of ventricular function, ejection fraction >35%, or previous high creatinine or potassium values. Additionally, mean age and spironolactone dose were higher than in RALES.38 Juurlink et al also reported an increased prescription of spironolactone in non-heart failure patients, which may be associated with its use in hypertension.13
The analysis of the available information shows some variables that could be potential risk factors for hyperkalaemia in patients with heart failure treated with spironolactone and ACE inhibitors. Advanced age, previously high creatinine or potassium values, low ejection fraction and diabetes have been found to be associated with episodes of hyperkalaemia after the introduction of spironolactone. Regarding concomitant medications, the combination of ACE inhibitors and spironolactone increased the risk, in particular with the use of high doses of the latter.18 33 According to the findings of Shah et al, patients with both medications and creatinine values between 1.5–2.0 mg/dl had a 35% risk of hyperkalaemia compared to those with normal renal function, scaling up to 46% if creatinine ranges from 2.0–2.5 mg/dl.34 Other drugs may also slightly predispose to hyperkalaemia in CHF patients, including non-steroidal anti-inflammatory drugs, β-blockers and heparin.18 28 29 31–34
The hypothesis regarding the benefits of combining ACE inhibitors and ARBs in the treatment of heart failure has been tested. The potential risk of inducing severe hyperkalaemia if this combination is used should be taken into consideration, particularly if a triple combination (ACE inhibitor + ARB + spironolactone) is indicated. In the CHARM-Added study, the use of candesartan with ACE inhibitors significantly increased the risk of discontinuation of the study drug due to hyperkalaemia (3.4% vs 0.7%, p<0.0001).39 Although it was suggested that the clinical impact was small due to the low frequency of severe hyperkalaemia, the number of patients on spironolactone was also small (17%). Recently, a detailed analysis of hyperkalaemia in the CHARM programme was published.40 Overall, there was a significant increase in the rate of hyperkalaemia with the combination of candesartan plus ACE inhibitors. The addition of spironolactone to these medications further increased the risk. The rate of hyperkalaemia per 1000 patients/year was 6.6 in the placebo group, 10.9 in the candesartan plus ACE inhibitor, and 21.8 in those who were also taking spironolactone. Therefore, although no definite conclusion can be drawn from the available information with the use of triple combination (particularly regarding the effect on mortality), a close look at potassium values should be taken, and a rigorous follow-up should be implemented.
Is this problem serious?
All this information strongly suggests a role for spironolactone in the trend towards a higher incidence of severe hyperkalaemia in heart failure. But what is the risk of clinical events secondary to hyperkalaemia in heart failure? The current recommendation for serum potassium concentration in heart failure is between 4.5–5.5 mEq/l.41 Beyond these values, there may be an increasing incidence of serious adverse events. Several studies reported complications related to hyperkalaemia in heart failure.28 42–45 Wrenger et al reported 44 patients with heart failure admitted with severe hyperkalaemia (mean 7.7 mEq/l) and renal failure and found that five patients required cardiopulmonary resuscitation, two patients died, and six patients required chronic dialysis.43 Bozkurt et al reported that three patients (3%) required temporary pacemakers because of bradyarrhythmias related to hyperkalaemia.28 Finally, Berry et al reported 84 heart failure patients with hyperkalaemia, and two patients died suddenly secondary to severe hyperkalaemia.45 Larger, prospective series of hyperkalaemia in heart failure analysing life threatening complications are lacking, but these reports suggest that extreme caution should be taken whenever a patient presents with high potassium concentrations.
Another important question regarding hyperkalaemia related to spironolactone treatment is the time of presentation after starting the medication. There are no precise analyses regarding this issue, but it is likely that the risk is greater at the beginning of treatment. Cruz et al reported that the mean time to the first episode of hyperkalaemia after starting with spironolactone was 15 days.27 However, Saito et al reported that the potassium concentration increases after 3 months of spironolactone treatment, and that many episodes of hyperkalaemia occurred even after 12 months.33 Therefore, it is likely that the closer to the start of spironolactone treatment the higher the risk of hyperkalaemia, but surveillance should probably continue while the patient is on this medication, particularly if renal failure develops.
Prevention of hyperkalaemia in heart failure
In our daily clinical practice we presume that severe hyperkalaemia may evolve into serious complications, in particular if left untreated. We also consider that mild to moderate degrees may predict the development of serious hyperkalaemia. However, evidence for the evolution of potassium values in patients with heart failure is lacking, in particular after the introduction of spironolactone. If we find a patient with a potassium concentration of 5.5 mEq/l, what are the chances that this patient will develop serious hyperkalaemia (>6.0 mEq/l) in the next hours or days? More importantly, what are the chances of developing life threatening events? Circadian variation in plasma potassium concentration could be as high as 10% in healthy subjects.46 It has been reported that hyperkalaemia develops after exercise in heart failure patients, even to a higher extent than in healthy controls.47 However, we do not know what happens in patients treated with spironolactone. Since we do not have information regarding potassium variation in heart failure patients, the usual practice is to recheck potassium values after a short time (ie, 24–48 h) and to take some preventive or treatment measures depending on the degree of hyperkalaemia (ie, dietary changes, spironolactone dose reduction, spironolactone suspension, etc). Measures to avoid the potential risk of developing hyperkalaemia are now an important part of the management of heart failure (box 2).
Box 2 Clinical recommendations for preventing serious hyperkalaemia in heart failure
Consider RALES enrolment criteria for indicating spironolactone in heart failure: (ejection fraction <35%, NYHA functional class III–IV, creatinine <2.5 mg/dl, K<5.0 mEq/l).
Careful monitoring when spironolactone is given in combination with other medications that may increase potassium concentrations
Spironolactone dose reduction whenever serum potassium is between 5.5–6.0 mEq/l and no electrocardiographic changes are noted.
Do not increase spironolactone above suggested dose (12.5–50 mg/day)
Renal function assessment by glomerular filtration rate instead of creatinine values in borderline cases.
Frequent controls in diabetic and renal patients.
Careful assessment on potassium intake (foods, supplement),
Use of loop diuretics for increasing urine potassium elimination.
Aldosterone antagonism showed significant benefits in patients with heart failure without inducing serious side effects. Long term improvement in survival was observed with both spironolactone and eplerenone. Spironolactone was rapidly adopted for the treatment of heart failure and numerous patients are now on this medication, even without having strict indications for its use. It is also likely that monitoring potassium in patients on spironolactone has not been followed as suggested. Therefore, the widespread prescription of spironolactone and the lack of proper follow-up of potassium values may be serious for a minority of patients at risk, as suggested by many cohort studies. Furthermore, the indication for aldosterone antagonism is expanding to the treatment of hypertension, and potentially the problems with high potassium concentrations may increase. The results of the randomised trials show that the relatively low number of hyperkalaemia episodes do not surpass the significant benefits that aldosterone antagonism in heart failure offers, and therefore the risk of hyperkalaemia should not preclude the use of these medications in heart failure. However, in order to gain the maximum benefit from aldosterone antagonists we need to individualise their use; we must also monitor electrolytes carefully whenever a patient with heart failure is being treated with one of these medications, particularly in patients with known risk factors for hyperkalaemia.
▶ Pitt B, Zannad F, Remme WJ, et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N Engl J Med 1999;341:709–17.
▶ Pitt B, Remme W, Zannad F, et al. Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med 2003;348:1309–21.
▶ Juurlink DN, Mamdani MM, Lee DS, et al. Rates of hyperkalemia after publication of the Randomized Aldactone Evaluation Study. N Engl J Med 2004;351:543–51.
▶ Palmer BF. Managing hyperkalemia caused by inhibitors of the renin-angiotensin-aldosterone system. N Engl J Med 2004;351:585–92.
▶ Desai AS, Swedberg K, McMurray JJV, et al. Incidence and predictors of hyperkalemia in patients with heart failure: an analyis of the CHARM program. J Am Coll Cardiol 2007;50:1959–66.
Multiple choice questions (true (T)/false (F); answers after the references)
1. What was the main finding of the RALES study?
A significant mortality reduction
An improvement of two points in the functional class.
High potassium concentrations in heart failure patients medicated with spironolactone
Low incidence of sudden death with spironolactone use
2. Which of the following does not relate to hyperkalaemia in heart failure?
Advanced heart failure
3. Which of the following statements is not suggested in order to avoid hyperkalaemia in heart failure patients medicated with an aldosterone antagonist?
Frequent laboratory check-ups
Assessment of changes in food intake
Spironolactone dose reduction.
4. Digoxin increases the risk of sudden death in heart failure with:
5. Population studies suggest that spironolactone use after publication of the RALES study increase the incidence of:
Atrioventricular (AV) node block
Heart failure progression
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.