Heart Failure Management in Dialysis Patients: Many Treatment Options with No Clear Evidence

Abstract

Heart failure with reduced ejection fraction (HFrEF) impacts approximately 20% of dialysis patients and is associated with high mortality rates. Key issues discussed in this review of HFrEF management in dialysis include dialysis modality choice, vascular access, dialysate composition, pharmacological therapies, and strategies to reduce sudden cardiac death, including the use of cardiac devices. Peritoneal dialysis and more frequent or longer duration of hemodialysis may be better tolerated due to slower ultrafiltration rates, leading to less intradialytic hypotension and better volume control; dialysate cooling and higher dialysate calcium may also have benefits. While high quality evidence exists for many drug classes in the non-dialysis population, dialysis patients were excluded from major trials, and only limited data exist for many medications in kidney failure patients. Despite limited evidence, beta-blocker and angiotensin converting enzyme inhibitor or angiotensin receptor blocker use is common in dialysis. Similarly, devices such as implantable cardiac defibrillators (ICDs) and cardiac resynchronization therapy that have proven benefits in non-dialysis HFrEF patients have not consistently been beneficial in the limited dialysis studies. Use of leadless pacemakers and subcutaneous ICDs can mitigate future hemodialysis access limitations. Additional research is critical to address knowledge gaps in treating maintenance dialysis patients with HFrEF.

Introduction

Heart failure is highly prevalent, currently affecting roughly 6.5 million US residents¹. Chronic kidney disease is also common; approximately 16 to 18 million United States residents have chronic kidney disease stage 3-5, including nearly 800,000 individuals with kidney failure treated by dialysis or transplant and approximately 125,000 new cases of kidney failure each year². Heart failure and kidney failure often occur in the same individual. Approximately 20% of hemodialysis patients and 10 to 20% of peritoneal dialysis patients have heart failure with reduced ejection fraction (HFrEF) based on administrative data and limited echocardiographic studies,^3–6 Chronic kidney disease (CKD) and heart failure impact each other - heart failure increases the risk of developing kidney failure,^7,8 and CKD may potentiate volume overload. Dialysis patients with heart failure have a lower 2 year survival after initiation of dialysis (65%) than those without heart failure (83%)³.

Strong evidence exists for several targeted medical and device therapies in the HFrEF population without advanced chronic kidney disease. These include beta-blockers, renin-angiotensin-aldosterone system (RAAS) blockade, and implantable cardiac resynchronization devices and/or defibrillators. However, their safety and efficacy in dialysis patients is uncertain. Here, we review the available evidence regarding treatment of chronic heart failure in dialysis, specifically focusing on HFrEF, as there are limited data for heart failure with preserved ejection fraction (HFpEF) in both the general and dialysis population.

Clinical Manifestations

The classic presenting symptoms of heart failure include dyspnea and fatigue, often accompanied by edema and other signs of volume overload.⁹ These non-specific signs and symptoms can also accompany kidney failure related volume overload, anemia and uremia; hence, distinguishing which disease is responsible for these symptoms can be challenging. Other clinical evidence, such as tachycardia and/or hypotension with minimal volume challenge should prompt consideration of HFrEF.¹⁰

Dialysis Modality and Dialysis-Related Treatment Strategies in HFrEF

Dialysis Modality Selection

The decision surrounding which dialysis modality is best for a patient is complex, taking into account many factors including patient preferences, living situation (including storage space), and social support among others. However, we support strong consideration of PD or extended hemodialysis in patients with HFrEF because these therapies provide more consistent volume control with lower ultrafiltration rates and greater hemodynamic stability, and there may be less myocardial stunning with these therapies.^{11, 12} Critically, volume overload, more so than the ultrafiltration rate or other factors, appears to be the greatest risk state for mortality among dialysis patients.¹³ Intradialytic hypotension, a major reason for failure to achieve a target dry weight, is not a feature of PD.

Myocardial stunning is common with standard thrice weekly hemodialysis, and, in one study of hemodialysis patients with normal cardiac function, repeated episodes of hemodialysis-induced myocardial stunning were associated with a subsequent reduction in left ventricular ejection fraction^12,14. This is likely due to decreased myocardial perfusion during hemodialysis, and occurred even in the absence of intradialytic hypotension and/or angiographically significant coronary artery disease.¹⁵ In contrast, myocardial stunning was not observed in one small study of 10 peritoneal dialysis. Only limited data exist on frequent hemodialysis, with the Frequent Hemodialysis Network (FHN) Daily Trial showing that hemodialysis five to six times per week reduced ventricular end systolic and diastolic volumes as well as left ventricular mass.¹⁶

There are no randomized clinical trials comparing hemodialysis and PD in the general dialysis population, let alone in the HFrEF population. Observational studies suggest that peritoneal dialysis in HFrEF patients is well tolerated and improves symptoms and functional status. Two prospective studies from Germany, one with 118 and the other with 159 patients with NYHA class 3 or 4 symptoms and diuretic resistance, demonstrated successful initiation of PD, improvement in symptoms within 6 months of starting PD, a low incidence of peritonitis or catheter dysfunction, and a decrease in hospitalizations for heart failure from 3 to 1.8 per year.^{17, 18}. There was no comparator hemodialysis group in either study. Similarly, in a prospective Austrian study of 40 patients initiating PD in the setting of refractory right-sided heart failure with or without left-sided heart failure, the one year survival rate was 55% and patients reported improved quality of life and/or fewer hospitalization days¹⁹. Additionally, 4 patients who were initially deemed ineligible for a left ventricular assist device (LVAD) due to severe right heart failure improved sufficiently following PD initiation to qualify for LVAD implantation.¹⁹ A propensity-matched retrospective cohort study of dialysis dependent HFrEF patients in Taiwan found a lower incidence of hospitalization for heart failure with PD as compared to hemodialysis (19.7 vs. 27.6 events per 1,000 patients-years, respectively),²⁰ while a recent systematic review summarizing results in 673 patients from 21 observation studies reinforced that peritoneal dialysis is a reasonable option in diuretic-resistant HFrEF patients with kidney failure, with annual mortality of approximately 20% across studies²¹.

Performing Peritoneal Dialysis in the Patient with HFrEF

The peritoneal dialysis prescription in those with HFrEF is similar to the prescription in patients with normal cardiac function, although there are additional considerations. First, initiation of peritoneal dialysis may be complicated by surgeon hesitancy to place a peritoneal dialysis catheter, particularly if this is done laparoscopically with general anesthesia. In these cases, non-guided catheter placement may be preferable depending on operator experience. Second, many patients with decompensated heart failure may have ascites. This is not a contraindication to peritoneal dialysis catheter placement, and, to the contrary, would be a setting in which peritoneal dialysis will be advantageous. Our local practice is to drain any significant ascites, supporting with colloid if this is a large volume, in order to reduce the risk of a catheter leak. Subsequently, we drain approximately 1L of fluid up to twice daily depending on the rate of accumulation, until PD is initiated. If PD needs to be initiated urgently, which is uncommon as there is some solute clearance and volume control simply from repeated draining of ascites, we use low volume exchanges and maintain the patient in a supine position as much as possible for the first 1-2 weeks following catheter placement. Our protocol for initiating peritoneal dialysis in patients with ascites is shown in Box 1.

Box 1. Protocol for PD Initiation with Ascites Present.

1. Catheter placement

   • Significant ascites may be drained intraoperatively

     • Lessens likelihood of a pericatheter leak

   • Consider colloid administration, similar to with large volume paracentesis, depending on volume status and hemodynamics

2. Early Considerations

   • Maintain supine when significant intraperitoneal volume pending tunnel tissue ingrowth to reduce the likelihood of a pericatheter leak

   • Consider spontaneous bacterial peritonitis prophylaxis in those with cirrhosis

3. Peritoneal dialysis initiation

   • If there is likely significant remaining ascites or rapidly reaccumulating ascites

     • Drain 1 L of fluid from the abdomen via the PD catheter twice a day

       • When drainage <500 cc, instill 750 – 1000 cc of 1.5% dialysate

         • Dwell for ~4 hours

         • Drain the instilled volume plus a maximum of 1L

     • If more than 1L effluent drained beyond instilled volume, do not instill additional dialysate

       • Continue to drain up to 1L of fluid via PD catheter twice a day until drainage below 500 cc, at which time begin to instill dialysate as above

     • If net negative balance remains less than 1L per exchange, initiate and titrate PD per usual protocols

       • If rapid start, continue lower volume dwells to reduce the likelihood of a pericatheter leak

   • If minimal reaccumulating ascites

     • Initiate PD per usual protocols

Maintenance therapy focuses on volume control. Sodium removal in peritoneal dialysis differs from hemodialysis, not only varying between patients but also varying significantly within each dwell. With high osmolar solutions, commonly seen with dextrose-containing dialysate, sodium-free water transport via aquaporins occurs early in the dwell, resulting in an immediate decrease in the concentration of sodium in the dialysate. After this initial period, sodium transport will supersede water transport, through both diffusion and convection via small pores in the peritoneal membrane. Enhancing this later phase may result in better sodium removal. Depending on transport characteristics, this can be accomplished with either longer dwells or use of iso-osmotic dialysate, such as icodextrin.²² With successful decongestion, cardiorenal physiology may improve, and it is possible that some patients may do well with fairly minimal peritoneal dialysis regimens.

Performing Hemodialysis in the Patient with HFrEF

Despite PD being often preferred, some patients cannot perform PD because of medical, social, psychiatric or other contraindications. Additionally, some patients prefer the outpatient hemodialysis setting. If hemodialysis is the selected dialysis modality for a patient with HFrEF, there are several potential strategies that may reduce myocardial stunning, reduce morbidity, and enhance the tolerability of hemodialysis in HFrEF. These include reducing the ultrafiltration rate, dialysate cooling, and utilizing higher dialysate calcium concentrations (Table 1).

Table 1.

Dialysis Strategies to Improve Symptoms and Dialysis Tolerability in Patients with Heart Failure

Strategy	Benefits	Limitations
Longer or more frequent dialysis	Lower interdialytic weight gain allows for slower ultrafiltration with smaller amounts of volume removal at each session	Patient preference, cost
Dialysate cooling	May lead to less myocardial stunning, less intradialytic hypotension, and less cardiac remodeling	Optimal temperature has not been defined. Occasional patient discomfort.
Maintain dialysate calcium ≥2.5 mEq/L	May lead to less myocardial stunning, less intradialytic hypotension, and fewer HF hospitalizations	Possible promotion of vascular calcification.
Higher dialysate sodium	Less intradialytic hypotension, resulting in better single session tolerance	Increased interdialytic weight gain, thereby worsening heart failure
Dietary sodium restriction	May decrease fluid retention and congestive symptoms	May be associated with higher hospitalization rates and increased neurohormonal activation
Midodrine	Less intradialytic hypotension. May decrease dyspnea presumably through better volume removal during dialysis	May be associated with increased mortality

Longer or more frequent dialysis refers to increasing dialysis time; this can be accomplished through longer hemodialysis sessions (including nocturnal), more hemodialysis sessions, and peritoneal dialysis.

Ultrafiltration Rate

By definition, the ultrafiltration rate can be lowered by increasing the frequency of treatments and/or the duration of each treatment. Longer treatments facilitate more gradual volume removal, reducing the likelihood of intradialytic hypotension and, in turn, increasing the ability to achieve a lower target weight. With more frequent treatments, the interdialytic weight gain is less, lowering the ultrafiltration rate.

In a cross-sectional study of 46 patients with mean LVEF 35-40% treated with thrice weekly in-center hemodialysis, frequent home or in-center hemodialysis (5 or more days/week), or home nocturnal hemodialysis (5 or more days per week), myocardial stunning was found in 100% of standard thrice weekly in-center patients, 92% of frequent in-center patients, 75% of frequent home hemodialysis patients, and 50% of home nocturnal dialysis patients. While these results could reflect that sicker patients are more likely treated with thrice weekly in-center hemodialysis, this study might also suggest that prolonged treatments rather than more frequent treatments may be associated with more consistent myocardial perfusion during dialysis sessions.²³ As noted above, only limited data exist on frequent dialysis, suggesting improvements of ventricular volumes and LV mass with daily in-center hemodialysis in a population without HFrEF¹⁶.

Dialysate Cooling

Cooling of dialysate is another strategy that may reduce intradialytic hypotension and myocardial stunning. A recent Cochrane review including 25 studies of dialysate cooling among 712 participants suggested a potential reduction in intradialytic hypotension with dialysate cooling while noting that there was substantial uncertainty based on the available trials to date²⁴.

Cooling of dialysate may have beneficial effects on cardiac function and remodeling. In a randomized, crossover study of 10 patients without HFrEF but who were prone to intradialytic hypotension, blood pressure was higher, intradialytic hypotension was less common, and left ventricular wall motion abnormalities were less frequent when dialyzing at 35°C than at 37°C.¹⁴ In another randomized controlled trial evaluating 73 patients without known heart failure, left ventricular mass was lower and there was less deterioration in diastolic function among those randomized to cool dialysate (defined as 0.5°C below their body temperature) as compared to those randomized to a standard dialysate temperature of 37°C.²⁵ No studies have examined effects of cooling in the HFrEF population specifically. A large pragmatic trial of the effects of dialysate cooling on cardiovascular outcomes is ongoing in Ontario, Canada (clinicaltrials.gov NCT02628366).

Dialysate Calcium

Use of higher dialysate calcium concentrations may also reduce the incidence of intradialytic hypotension and subsequent cardiac damage. Current guidelines suggest use of a dialysate calcium concentration of 2.5 to 3 mEq/L (1.25 to 1.5 mmol/L), but some facilities use lower concentrations of 2.0 to 2.25 mEq/L out of concern that the higher calcium concentration promotes vascular calcification. A retrospective matched cohort study comparing facilities that converted to a dialysate calcium concentration <2.5 mEq/L with those that used dialysate calcium concentration ≥2.5 mEq/L (predominantly 2.5 mEq/L) found no effect of dialysate calcium concentration on all-cause mortality or hospitalizations; however, rates of hospitalization for heart failure were higher in in the facilities using lower dialysate calcium.²⁶ Whether reductions in heart failure morbidity from better preserved hemodynamics in the high risk HFrEF population outweigh the theoretical risks of vascular calcification requires study.

Vascular Access Considerations

The many cardiac procedures performed that involve central veins, such as right heart catheterization and cardiac device implantation, decrease the likelihood of successful arteriovenous fistula (AVF) creation. Subcutaneous and epicardial implantable cardioverter defibrillators (ICDs) are alternative options to transvenously implanted devices, though they have not been well-studied in dialysis patients²⁷. Leadless pacemakers are increasingly being used and should be considered when available to minimize venous damage²⁸. Studies have shown that patients who have had prior peripherally inserted central catheters (PICCs) are less likely to achieve a mature AVF ²⁹. For HFrEF patients with CKD or on dialysis who require longer-term inotropes or other intravenous therapies, a small-bore tunneled central venous catheter rather than a PICC is always preferred. Vein sparing strategies should be used whenever possible.

High output heart failure is an underappreciated potential complication of upper arm arteriovenous fistulas (AVF) and grafts and may pose harm especially to the HFrEF patient. High-output heart failure occurs because a large proportion of arterial blood is shunted from the left-sided circulation to the right-sided circulation via the AVF. The increase in preload can lead to increased cardiac output, LV dilatation and/or hypertrophy and over time, reduced LVEF³⁰. Case reports describe worsening of heart failure symptoms in patients with or without pre-existing heart failure and pulmonary hypertension after AVF creation, with resolution following ligation of the AVF^31,32. In a retrospective cohort study of 137 dialysis patients who had echocardiogram before and after AVF creation, (88% upper arm and 12% lower arm), maladaptive changes, including left atrial dilatation, right ventricular dilatation, worsening right ventricular function, and a small reduction in LV systolic function, were observed following AVF creation. In a small subset of patients from this same study who subsequently underwent fistula ligation, right ventricle function remained stable or improved with no change in left ventricle function post-ligation. There was no comparator group of patients who did not have a fistula in this study and so it is unclear if these changes were related to the AVF per say though it is theoretically possible³³. How the risks of infection and thrombosis from a permanent central dialysis catheter weigh against the risk of worsening heart failure in the HFrEF patient is unclear. These concerns further elevate PD as the better maintenance dialysis modality in patients with HFrEF.

Other Considerations

It is unclear if modification of sodium, whether in the dialysate or in the diet, is beneficial in HFrEF. Higher dialysate sodium can reduce the incidence of intradialytic hypotension but also leads to higher interdialytic weight gains, which necessitates a higher UFR, increasing the risk of intradialytic hypotension and associated morbidities. ³⁴. The most recent guidelines for the management of HFrEF state that dietary sodium restriction is reasonable, as it may reduce fluid retention and congestive symptoms but may be associated with higher hospitalization rates thought to be mediated through increased neurohormonal activation^9,35,36.Potential harms of sodium restriction may be less applicable in dialysis patients, given low residual kidney function.

Midodrine is sometimes used in HFrEF patients with frequent intradialytic hypotension or chronic hypotension. There are a handful of observational studies, all of modest sample size, that have yielded conflicting results, with some showing symptom improvement and others suggesting possible increased mortality^37–40. Clinicians should be cautious in prescribing midodrine to hemodialysis patients with HFrEF until more data become available.

In sum, longer duration and more frequent dialysis, whether done with hemodialysis or PD, seems to improve heart failure symptoms and may help prevent further progression of HFrEF. Several hemodialysis strategies such as longer or more frequent treatments, dialysate cooling, higher dialysate calcium, higher dialysate sodium, dietary salt restriction, and midodrine may improve the tolerability of single hemodialysis sessions in HFrEF patients, but it is unknown if any of these improve clinical outcomes, and higher dialysate sodium in particular may trade a short term benefit for a longer term harm.

Pharmacologic Management of HFrEF

Standard therapies for non-dialysis patients with HFrEF include beta blockers, angiotensin converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARB), mineralocorticoid receptor antagonists (MRA), and loop diuretics (Table 2). While robust evidence for efficacy exists among patients with HFrEF without kidney failure, evidence among dialysis patients with HFrEF is lacking. Studies in the dialysis population are limited by their observational design, small sample size, or reliance on surrogate endpoints that may or may not be relevant to improving HFrEF clinical outcomes. Most of these agents lower blood pressure, which can limit ultrafiltration leading to chronic volume overload. This may result in disagreement between the dialysis physician and the heart failure physician, with the dialysis physician preferring to reduce dry weight to optimize volume control and the heart failure physician preferring neurohormonal agents despite their effects in potentially limiting the ultrafiltration volume. The optimal approach remains uncertain.

Table 2.

Pharmacologic Treatments for Heart Failure

Medication	Benefits in patients with HFrEF not on dialysis	Benefits in patients with HFrEF on dialysis
Beta blockers	Carvedilol, metoprolol, and bisoprolol increase survival, improve symptoms and decrease HF hospitalization	Uncertain. Possible improvement in symptoms
ACEi/ARB	Increases survival, improves symptoms, and decreases HF hospitalization	Uncertain
Mineralocorticoid receptor antagonists	Increases survival, improves symptoms, and decreases HF hospitalization	Uncertain
Loop diuretics	Reduces symptom burden	Minimizes weight gains between treatments
Digoxin	May improve symptoms, quality of life, and exercise tolerance in mild to moderate heart failure. No mortality benefit.	Risks likely outweigh any potential benefits.
SGLT2 Inhibitors	Increases survival and decreases HF hospitalization	Not studied; effects would likely need to be independent of kidney function
Angiotensin receptor-neprilysin antagonists	Increases survival and decreases HF hospitalization	Not studied

HFrEF, heart failure with reduced ejection fraction; SGLT2, sodium glucose cotransporter-2; ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker

The existing data for each drug class are presented below, with the exception of the newer heart failure therapies including angiotensin receptor blocker-neprilysin inhibitors (ARNI) and sodium-glucose transporter 2 (SGLT2) inhibitors. ARNIs are not contraindicated in dialysis patients but have not been studied. A single case report exists of the safe and successful use of sacubitril/valsartan in a hemodialysis patient who experienced improved heart failure symptoms and decreased filling pressures after ARNI intitiation.⁴¹ SGLT2-inhibitors would not be likely to be of benefit given their mechanisms of action.^42–44

Beta Blockers

Carvedilol, metoprolol succinate, and bisoprolol reduce mortality and hospitalizations in the general HFrEF population, likely including those with CKD stage 3.^45–47 While limited data exist regarding efficacy in dialysis, most beta blockers are relatively safe in dialysis patients, although dialysis clearance may differ among drugs in this class. Metoprolol, for example, is readily cleared by dialysis, whereas bisoprolol and carvedilol are less dialyzable.⁴⁸ While metoprolol and carvedilol are the most commonly used beta blockers in patients with HFrEF, one retrospective study found higher mortality with carvedilol compared to metoprolol in hemodialysis patients, perhaps related to a higher frequency of intradialytic hypotension in the carvedilol group.⁴⁹ A propensity-matched retrospective cohort study of 3,400 Taiwainese hemodialysis patients with heart failure showed lower all-cause mortality among patients receiving beta blockers, and even lower mortality among patients on both beta blockers and an ACE inhibitor or ARB, although the ability to utilize both of these agents concurrently suggests less hypotension and better cardiac function.⁵⁰

The only randomized trial, a 1 year study comparing carvedilol to metoprolol in 114 dialysis patients with HFrEF, showed no difference in mortality but better NYHA functional classification at study completion among those randomized to carvedilol.⁵¹ Sixteen percent of patients dropped out during the run-in phase or shortly after randomization due to hypotension, bradycardia, or worsening heart failure, reinforcing that some dialysis patients with HFrEF may be unable to tolerate beta-blockers.

There are several possible reasons why beta blocker use may be less beneficial in dialysis patients. Cardiac rhythm monitoring has shown that the most common clinically significant arrhythmias in hemodialysis patients are bradyarrhythmias, which occur in up to 20% of hemodialysis patients, and are most common in the first dialysis session of the week.⁵² Beta blockers increase the incidence of bradyarrhythmias. Moreover, as negative inotropes and chronotropes, beta blockers may precipitate or prolong IDH episodes. To this end, studies of beta blockers in hemodialysis patients have mixed results, with some showing lower risk of sudden cardiac death (SCD) while others so no differential effect or suggest higher risk of SCD.^53–55 A randomized controlled trial of beta blocker versus placebo was initiated in Australia but was unfortunately closed down very soon thereafter because of futility in recruiting patients, with many patients or their physicians unwilling to stop treatment with beta- blockers.^56,57 In sum, it remains uncertain whether beta-blockers are beneficial in dialysis patients with HFrEF, although, if tolerated, most physicians are likely to prescribe beta blockers in this clinical setting.

ACE inhibitors and ARBs

While RAAS inhibition is one of the mainstays of heart failure therapy, patients with advanced kidney disease often discontinue ACE inhibitor/ARB therapy due to hyperkalemia or worsening of GFR,⁵⁸ despite ACE inhibitors and ARBs potentially having beneficial anti-fibrotic effects in the heart and possibly the kidney.^59,60 Though dialysis patients can safely start or resume ACE inhibitors or ARBs, albeit with an increased risk of hyperkalemia, it is unclear if these medications provide the same benefits as seen in patients with heart failure not on dialysis.

A meta-analysis of 11 randomized clinical trials that included 1856 hemodialysis and PD patients found no reduction in the incidence of adverse cardiovascular events, defined as a composite of non-fatal or fatal MI, non-fatal or fatal stroke, or heart failure events, among those treated with ACE inhibitor/ARB therapy as compared with placebo or other blood pressure medications.⁶¹ In secondary analyses, both ACE inhibitors and ARBs were associated with fewer heart failure events. Results were not examined in the HFrEF subgroup separately. A retrospective analysis of the Dialysis Outcomes and Practice Patterns Study (DOPPS) showed a small, but statistically significant survival benefit associated with ACE inhibitor/ARB use as compared to no ACE inhibitor/ARB use regardless of HFrEF status.⁶² Conversely, a propensity-matched secondary analysis of the HEMO study found no difference in mortality, heart failure hospitalizations, or other cardiovascular outcomes associated with ACE inhibitor use⁶³. An observational study of 4,771 hemodialysis patients with chronic heart failure in Taiwan showed an association between ACE inhibitor/ARB use and lower all-cause and cardiovascular mortality.⁶⁴ Outcomes specifically in patients who already were known to have heart failure were not analyzed separately in any of these studies.

In sum, whether there is a mortality benefit associated with ACE inhibitor/ARB use in maintenance dialysis patients with heart failure remains unproven, although these agents are generally safe in dialysis patients, and, if tolerated, may have benefits.

Mineralocorticoid Receptor Antagonists

Data on the use of mineralocorticoid receptor antagonists in dialysis patients with heart failure are limited. While mineralocorticoid receptor antagonists have demonstrated a mortality benefit in patients with HFrEF, patients with kidney disease were excluded from these studies^65–67. Several randomized clinical trials have examined the safety and efficacy of spironolactone in maintenance hemodialysis patients. These studies are limited by few HFrEF patients, small sample sizes, under-recruitment, low event rates, and use of surrogate outcomes such as left ventricle mass and left ventricle ejection fraction. These trials have several notable findings: 1) a slightly higher frequency of moderate hyperkalemia and rare occurrence of life-threatening hyperkalemia;^68–70 and 2) no increase in hypotension requiring hospitalization.⁷¹ The Mineralocorticoid Receptor Antagonist in End Stage Renal Disease (MiREnDa) study found no difference in LV mass or LVEF between those on spironolactone versus placebo, though only 4% of patients had heart failure at study entry.⁷² Even less is known about mineralocorticoid receptor antagonists in PD patients, especially in those with heart failure, although they are an effective treatment for hypokalemia in PD patients.⁷³

In the general population, there are limited data suggesting a possible benefit of spironolactone in HFpEF, an entity that remains difficult to define and is essentially unstudied in the dialysis population. The Treatment of Preserved Cardiac Function Heart Failure With an Aldosterone Antagonist (TOPCAT) trial randomized 3,445 patients with symptomatic HFpEF to spironolactone versus placebo and found no difference in the primary composite outcome of death from cardiovascular causes, aborted cardiac arrest, or heart failure hospitalization.^74,75 A subgroup analysis suggested that the lack of a benefit may have been driven by worse results in Russian than in North American study participants. Consequently, these results should be viewed as exploratory.

Viewed in sum, these data suggest spironolactone generally is safe in dialysis patients with careful monitoring. However, whether they offer any benefit is unproven. Two large multicenter clinical trials, the ACHIEVE (Aldosterone bloCkade for Health Improvement EValuation in End-stage Renal Disease) Trial; NCT03020303) and ALCHEMIST (ALdosterone antagonist Chronic HEModialysis Interventional Survival Trial; NCT01848639), are ongoing, although these do not specifically focus on HFrEF.

Loop Diuretics

Many hemodialysis and PD patients still make urine, and loop diuretics may further enhance urine output and, in so doing, reduce the ultrafiltration volume required with dialysis. An observational study of 11,297 US hemodialysis patients comparing those receiving to those not receiving loop diuretics suggested that loop diuretic use was associated with a lower incidence of hospitalization, intradialytic hypotension, and interdialytic weight gain, but was not associated with a reduction in mortality at one year.⁷⁶ A trial of 61 PD patients randomized to either oral furosemide 250 mg daily or placebo showed a higher urine volume in the furosemide group (1070 mL/24 hours versus 733 mL/24 hours) but no difference in residual kidney function at one year.⁷⁷ As with patients without kidney failure, loop diuretic dosing should be titrated to achieve the desired degree of volume removal. In sum, there appears to be little downside to using oral loop diuretics in dialysis patients with urine output, and, given potential benefits, we use these agents routinely in our clinical dialysis practice.

Digoxin

Studies in patients with mild to moderate HFrEF not on dialysis show that digoxin improves symptoms, quality of life, and exercise tolerance. Digoxin also reduces the incidence of heart failure hospitalization, but does not affect mortality.^9,78 We recommend great caution with the use of digoxin in patients receiving dialysis as it is predominantly excreted by the kidneys, and it is not cleared with dialysis.⁷⁹ Observational studies, with limitations including potential confounding by indication, suggest a 28% higher risk of death in dialysis patients receiving digoxin.^80,81 It remains uncertain whether digoxin is of harm or of benefit in the dialysis patient with HFrEF.

Conclusions

Based on the above data, we recommend that traditional pharmacologic treatment of HFrEF can be considered in dialysis patients, as these medications are generally safe in dialysis. We avoid use of digoxin. However, the potential risks associated with traditional pharmacologic therapies, including hypotension, hyperkalemia, and decreased ability to remove volume during hemodialysis, must be weighed against what seem to be attenuated benefits relative to those seen in the broader non-dialysis population of HFrEF patients.

Cardiac Implantable Electronic Devices

ICD and CRT-D

HFrEF patients have a high risk of sudden cardiac death, with an annual incidence of about 15%.^82,83 In the appropriate patients with HFrEF, an implantable cardioverter defibrillator (ICD) or cardiac resynchronization therapy (CRT), with or without a defibrillator(CRT-D), can improve survival (Table 3).^9,84 ICD therapy is indicated for primary prevention of SCD in patients with symptomatic heart failure and LVEF ≤ 35%, while CRT is indicated when there is significant electrical dyssynchrony between the right and left ventricles. ICD therapy reduces the risk of death in HFrEF patients by over 20% ⁸⁵. Similarly, cardiac resynchronization not only improves mortality but also improves LVEF, symptoms, and quality of life⁸⁶.

Table 3.

Indications for primary prevention ICD and CRT.

ICD	CRT
LVEF ≤35%	LVEF ≤35%
≥40 days post MI if no revascularization, 3 months if revascularized post-MI	Sinus rhythm with left bundle branch block and QRS ≥130 ms
Expected survival >1 year	Sinus rhythm with non-LBBB and QRS ≥150 ms
≥NYHA class 2 symptoms on appropriate medical therapy	≥NYHA class 2 symptoms on appropriate medical therapy

LVEF, left ventricular ejection fraction; MI, myocardial infarction; LBBB, left bundle branch block; NYHA, New York Heart Association; ICD, implantable cardiac defibrillator; CRT, cardiac resynchronization therapy

Sudden cardiac death accounts for approximately 30-40% of deaths in kidney failure patients and may be even higher in kidney failure patients with HFrEF, though this has not been well-characterized.³ This risk is highest on dialysis days, particularly after a long interdialytic interval, presumably as a result of more marked electrolyte derangements, worse volume overload, and, if occurring after the dialysis session, more marked changes in each of these parameters as a result of the dialysis treatment.⁸⁷ Accordingly, dialysis patients seem an attractive population for ICD therapy for primary prevention of SCD. Unfortunately, there are no RCTs examining this question in dialysis patients with HFrEF.

A meta-analysis of RCTs of ICDs for primary prevention in the general population showed that the benefit of ICD therapy declined as GFR declined, though very few patients with CKD stage 4 or 5 and no dialysis patients were included.⁸⁸ Observational data suggest little to no benefit of ICDs in dialysis patients, and even a potential for harm. For example, a propensity-matched retrospective study of 303 hemodialysis patients with HFrEF, showed no mortality benefit associated with receiving an ICD for primary prevention, with one-year mortality rates of 43% in the ICD group as compared with 40% in the non-ICD group.⁸⁹ The only RCT to date was stopped prematurely when the effect size calculated in an interim analysis suggested it would be futile to continue.⁹⁰ Furthermore, there are additional risks with ICD therapy in dialysis patients that are important to consider. Most ICDs are placed transvenously, increasing the risk of ipsilateral central venous stenosis, which in turn may reduce the longevity of an existing vascular access or the potential to create a new vascular access, and the hardware and wires may predispose to infection.^91,92 Subcutaneous ICD implantation, in which a lead is tunneled under the skin in the left chest and remains extravascular, could mitigate some of the risks associated with transvenous ICDs in dialysis patients.⁹³

To our knowledge, there have been no RCTs of CRT-D in dialysis patients. A retrospective analysis suggested that CRT-D as compared to ICD alone improved mortality in 10,946 patients with moderate to severe CKD, including CKD stage 5.⁹⁴ The authors found no differential effect by CKD stage, though the sample size for the CKD stage 5 subgroup was small, and results have not yet been replicated in a larger study. Currently, it is unclear if dialysis patients should receive an ICD or CRT-D for primary prevention in heart failure.

LVADs

Kidney failure after left ventricular assist device (LVAD) is an uncommon occurrence.⁹⁵ This may be because LVAD therapy is contraindicated in patients at high risk for kidney failure, as it is known that LVAD patients who require dialysis have extremely high mortality (>60%).⁹⁶ In LVAD recipients who develop kidney failure, both PD and hemodialysis can be considered as treatment modalities, although specific management of LVAD patients on dialysis is complex and beyond the scope of this review. Several key principles of management of the LVAD patient with kidney failure include: 1) PD can be successfully done in LVAD patients; 2) LVAD patients undergoing hemodialysis can have successful AV fistula maturation; 3) Erythropoiesis stimulating agents likely can be used judiciously, applying the current KDIGO hemoglobin targets; and 4) LVAD patients with kidney failure can go on to successfully receive simultaneous heart and kidney transplants^97–100.

Conclusions

Despite a wealth of data and well-supported clinical practice guidelines for the treatment of heart failure in the general population, similar quality data are lacking for patients with concurrent heart failure and kidney failure. If a patient with HFrEF develops kidney failure, PD as an initial modality should be strongly encouraged. Those who opt for conventional hemodialysis may benefit from longer or more frequent treatments. While pharmacological treatments for heart failure generally appear safe in dialysis patients, there is no clear evidence that they improve symptom control, heart failure hospitalizations or survival. Similarly, the role for implantable electronic devices like ICD and CRT in patients with HF on dialysis is unclear.

In sum, dialysis patients with HFrEF represent a group at high risk for death for whom there are few proven effective treatments. The current paucity of data in this population underscores the need for adequately powered clinical trials examining the benefits and harms of conventional HF therapy in the context of both hemodialysis and PD. Only with additional research can we hope to improve the clinical outcomes of these patients.