A propitious time for initiating clinical trials in patients with heart failure with reduced ejection fraction and an estimated glomerular filtration rate <30 mL/min with an mineralocorticoid receptor antagonist and a K⁺binder: ‘the forbidden fruit’

Cardiovascular disease (CVD) remains the most common cause of premature death in the chronic kidney disease (CKD) population.1,2 Patients with CKD are at 10–20 times greater risk of cardiac death than controls without CKD, despite stratification for age, race, sex, and diabetes. Enhanced CVD mortality begins with mild kidney disease and increases progressively with more advanced kidney disease.2–4 There is a substantive overlap between CKD and CVD.5–8 The annual data report of the United States Renal Data System (USRDS) in 2009 reported that the prevalence of CVD reached 63% in CKD patients compared with 5.8% of those without CKD, stratified for the association with both CKD severity and age.9 A recent analysis of the European Society of Cardiology Heart Failure Pilot Survey reported a strong association between comorbidities such as CKD and adverse clinical outcomes.8 Chronic kidney disease, anaemia, and diabetes remained significantly and independently related to all-cause mortality and heart failure hospitalizations.8 Longitudinal estimated glomerular filtration rate among 1 120 295 adults within a large integrated system between 1996 and 2000. It is readily apparent that as chronic kidney disease progresses, the prevalence of heart failure increases as baseline estimated glomerular filtration rate declines progressively. Data derived from Go et al.1 Despite the increased CVD mortality, there has been a dearth of interventional cardiovascular RCTs in the CKD population, especially those with Stages 4–5 CKD.11–13 Furthermore, many patients with kidney disease have been excluded from the majority of mainstream cardiovascular interventional trials. While recently published RCTs on traditional and non-traditional risk factors including dyslipidaemia have added to the available pool of clinical data level 1 clinical evidence remains significantly lacking. A decade ago, two systematic reviews of randomized clinical trials of cardiovascular interventions published from 1985 through 200512,13 highlighted the marked underrepresentation of patients with kidney disease in randomized clinical trials of cardiovascular interventions.12,13 Coca et al. reported that 86 of 153 trials excluded patients with CKD (56.2%) and Charytan and Kuntz reported that 69 of 86 trials (80.2%) excluded patients with CKD.12,13 What was most startling, merely 6 of 86 (7.0%) to 15 of 153 (9.8%) studies even bothered to report patients' kidney function at the start of the clinical trial, a truly worrisome omission. Recently, Konstantinidis et al.14 published findings from an updated systematic review on the representation of patients with CKD in randomized clinical trials of cardiovascular interventions that were published from 2006 through 2013. Of the 371 trials included in this study, 212 (57.1%) excluded patients with kidney disease (defined based on varying thresholds of serum creatinine level, estimated glomerular filtration rate, and/or creatinine clearance; the need for renal replacement therapy). Renal function at baseline was reported in only 156 (42.0%) of trials. While one can argue that these percentages constitute progress since the previous systematic reviews in 200612,13 clearly we still are not where we should be. In light of the persistent underrepresentation of CKD patients in such trials despite strong recommendations for greater inclusion of patients with CKD into cardiovascular trials, we must ask why should that still be the case? Are there valid reasons why patients with CKD should be excluded from cardiovascular trials? Several possibilities come to mind. We and others have emphasized that the pathophysiologic features of CVD are altered in patients with CKD and that consequently they may be resistant to conventional treatments.2,15,16 Consistent with this formulation, it is conceivable that some investigators may fear that the inclusion of CKD patients in cardiovascular trials may compromise the results, theoretically by shifting the results toward the null.13,16 Alternatively, we believe a more likely reason for the reluctance to enroll patients with CKD in trials of cardiovascular interventions is the concern and apprehension for increased risks of adverse effects—especially hyperkalemia and worsening renal function (WRF). The striking disconnect between persistent underrepresentation of CKD patients and recommendations for greater inclusion of this cohort into cardiovascular trials is worrisome because the numbers that comprise this patient group are not trivial. Chronic kidney disease patients are not a small cohort; rather the very opposite exists. Impaired renal function is highly prevalent among patients with CVD, with estimates ranging from 30 to 60%.1,2 Consequently, it is very likely that in a real world setting, a wide spectrum of clinicians—not only cardiologists and nephrologists, but internists, diabetologists, and hospitalists will face the challenge to manage patients with both CVD and CKD. As Ishida and Johansen appropriately and elegantly summarized the problem: ‘Excluding patients with CKD from cardiovascular trials may be viewed as the path of least resistance—or even greatest prudence—when designing cardiovascular trials because of concerns about the possibility of less benefit or greater risk. However, given that patients with KD are at higher risk of death and morbidity from CVD, they stand to garner the greatest absolute benefit from treatment if the relative benefit is similar among those with and without CKD’.16 This cohort has long been avoided in all the major cardiovascular trials. Yet, there is increasing evidence that the greatest benefit might accrue in this previously forbidden group. With this background we wish to turn our attention to the feasibility of conducting clinical trials in patients with chronic HFREF (heart failure with reduced ejection fraction) with an estimated glomerular filtration rate (eGFR) <30 mL/min, what we label the ‘forbidden fruit’ cohort. The primary data to support the contention that MRAs might be effective in attenuating cardiovascular events in patients with CKD derive from several pre-specified analyses of subgroups of the major MRA clinical trials in patients with chronic HFREF: RALES,17 the EMPHASIS–HF study,18 and the Swedish Heart Failure Registry.19 In a sub-study of RALES, Vardeny and colleagues17 investigated the effect of baseline and WRF on the efficacy of spironolactone in patients with severe heart failure. Patients with a reduced baseline eGFR had similar relative risk reductions in all-cause death and the combined endpoint of death or hospitalization for heart failure compared with those with a baseline eGFR of >60 mL/min/1.73 m2, and greater absolute risk reduction than did those with a higher baseline eGFR (10.3% vs. 6.4%). They concluded that the absolute benefit of spironolactone was greatest in patients with reduced eGFR. Although WRF was associated with a negative prognosis, the mortality benefit of spironolactone was maintained.17 Eschalier and colleagues18 recently conducted a pre-specified analysis of subgroups in the EMPHASIS-HF study, cardiovascular mortality, and hospital admission for heart failure. The findings from these subgroup analyses provide compelling evidence for the use of eplerenone in all patients with heart failure and reduced ejection fraction that meet the inclusion and exclusion criteria of the EMPHASIS-HF study. Although these studies excluded patients with an eGFR of less than about 30 mL/min/1.73 m2, they clearly showed that eplerenone was safe and associated with a cardiovascular benefit in patients with less severe CKD (as low as 35 mL/min/1.73 m2). The most compelling evidence to support the premise that MRA intervention may confer benefit in patients with HFREF and an eGFR <30 mL/min, derives from the recent report from the Swedish Heart Failure Registry. Edner and colleagues19 have assessed data from the Swedish Heart Failure Registry collected between 2000 and 2013. About 80 variables were recorded at discharge or outpatient visits upon entry into the registry. Of 24 283 patients, 2410 had a creatinine concentration exceeding 221 μmol/L or creatinine clearance of <30 mL/min (Cockcroft–Gault method). Approximately two-thirds of these patients were treated with renin–angiotensin inhibitors, despite the currently prevailing caveats in the guidelines. About 25% of patients were given an MRA. The investigators created a matched cohort based on propensity scores for RAS antagonist use. All patients had an ejection fraction of 39% or lower and about half had an ejection fraction of <30%. Importantly, mean creatinine clearance was ∼23 mL/min in patients with severe renal insufficiency and the hazard ratios (HRs) for overall survival for RAS antagonist use were almost identical in those with and without severe renal insufficiency. This carefully designed study provides strong support for the postulate that the benefits of both renin–angiotensin inhibition and MRA treatment still occur in patients with concomitant severe renal dysfunction, even those with Stages 4–5 CKD. No interaction with age was noted, suggesting that renin–angiotensin antagonists might be associated with benefit not only in patients with Stages 4–5 CKD but also in elderly patients. The relative risk reduction in mortality throughout the 5-year follow-up was 24%, and the number needed to treat for 1 year to save one life was 10. Schematic diagram summarizing the current status of MRA therapy in patients with heart failure with reduced ejection fraction and diminishing renal function. As can be seen, current guidelines exclude patients with an estimated glomerular filtration rate <30 mL/min/1.73 m2 from receiving mineralocorticoid receptor antagonist therapy. We propose that the previous formidable constraint (i.e. hyperkalaemia) no longer constitutes a barrier because it can now be countered by the availability of new ‘enablers’ (the new non-absorbable K+ binders, both patiromer and ZS-9). We postulate that treatment with a K+ binder, by acting as an ‘enabler’ would facilitate sustained MRA therapy (avoiding down titration or discontinuation) with consequent improved outcomes. As an example, recent studies have clearly demonstrated that treatments with patiromer were able to sustain normokalaemia for 52 weeks in hyperkalaemic patients with diabetic CKD20 and resistant hypertension,21 thereby avoiding down-titration or discontinuation and consequently enabling continuous RAASi treatment. The risk of inducing WRF in patients whose ambient eGFR is already reduced to 30 mL/min merits comment. We should ask—what is the possible ‘downside’? In the absence of a randomized trial, the optimal timing of initiation of chronic dialysis remains elusive.22 There is a trend in the nephrology community toward earlier initiation of dialysis.23 Whereas some prior retrospective studies have suggested improved patient survival associated with earlier initiation of chronic dialysis,24 a more recent retrospective analysis showed early dialysis initiation based only on eGFR without clinical indications cannot be recommended.22 Given this lack of consensus when to initiate dialysis, it is reasonable to contend that an occasional worsening of eGFR when a patient is almost at the brink of dialytic therapy, is a risk worth accepting in order to provide the potential benefits of RAASi therapy in patients who are at the highest risk of death. In brief, we face a tradeoff of potentially conferring robust cardiovascular benefit vs. the downside of hastening initiation of dialytic therapy in patients already on a clinical trajectory for renal replacement therapy. Supportive data for our proposal may accrue by analogy from recent studies that assessed the effects of ACE inhibitors and ARBs in patients with advanced CKD. At earlier stages of CKD, blockade of the renin–angiotensin system is a mainstay of therapy; ACEIs and ARBs clearly slow the decline in renal function, especially in patients with proteinuria. Until recently, the use of ACEIs or ARBs in advanced CKD remained uncertain. This question was recently investigated by Hsu et al.25 In analogy with MRAs, the reason for hesitating to prescribe ACEIs or ARBs is the potential risk of producing hyperkalaemia, as well as short-term reductions in GFR. Hsu et al.25 conducted a prospective cohort study to assess the effectiveness and safety of ACEI/ARB use for advanced predialysis CKD. They utilized a comprehensive Taiwanese database that captures health care utilization data of >95% of the hospitals in Taiwan and 99% of the country's population. The investigators studied 28 497 patients with stable, hypertensive, advanced CKD who had inferred serum creatinine levels >6 mg/dL. Of these, 14 117 used ACEIs/ARBs, and 14 380 were nonusers. Use of ACEIs/ARBs was associated with a lower HR for long-term dialysis (0.94 [95% CI, 0.91–0.97]) and the composite outcome of long-term dialysis or death (0.94 [0.92–0.97]). More hyperkalaemia-associated hospitalizations were found among ACEI/ARB users than among nonusers (9.2% vs. 6.7%; P < .001]). However, the development of hyperkalaemia did not significantly increase the risk of predialysis death (HR, 1.03 [95%CI, 0.92–1.16]; P = .30). These observational data are consistent with randomized clinical trial data among patients with Stages 4–5 CKD.26,27 We believe that the Hsu study provides reassuring information that RAAS blockade can be safely implemented even in patients with advanced CKD. Clinical guidelines recommend close monitoring of renal function and electrolyte levels throughout the course of MRA therapy in patients with heart failure therapy.28 Regretfully, these guidelines are often not implemented. As detailed in a recent analysis of a study population of 10 443 Medicare beneficiaries with prevalent heart failure who newly initiated MRA therapy, Cooper et al.29 reported that rates of appropriate laboratory monitoring after MRA initiation were low. We suggest that the laboratory surveillance protocol suggested by Rossignol et al.30 may constitute a reasonable starting point for developing an adequate monitoring programme. As detailed in the eplerenone titration algorithm used in EMPHASIS-HF (table 1 of Ref. 30) monitoring was obtained after 1 week, 1 month, and every 4 months thereafter. Clearly, the frequency of monitoring should be increased in patients with an eGFR <45. In conclusion, the formulation we espouse is that a treatment strategy of add-on of an MRA in conjunction with a newly available K+ binder will reduce CV mortality and morbidity compared with current therapy, wherein current guidelines exclude patients with an eGFR <30 mL/min/1.73 m2 from receiving an MRA. It is appropriate to reevaluate this topic in the new era of capabilities for obviating hyperkalaemia. In brief, we propose that the previously ‘forbidden’ HFREF may now be transformed to the ‘low hanging fruit’. These hypotheses will, however, require careful prospective evaluation in adequately powered randomized clinical trials. M.E. performed statistical analysis, and supervision, acquired the data; Both M.E. and B.P. conceived and designed the research, drafted the manuscript, and made critical revision of the manuscript for key intellectual content. We thank Kim Jaimes for the excellent editorial support in the preparation of the manuscript, and Larry Gilstad for expert assistance with the media and artwork. Conflict of interest: M.E. is a consultant for Bayer Healthcare, OPKO Healthcare, Relypsa Inc., and Rockwell Medical. B.C.P. is a consultant for KBP Biosciences, Tricida, and Aurasense.