Relation of Serum Potassium to Cardiovascular Events in Patients With Heart Failure and Preserved Ejection Fraction: “Mind the Gap”

There has been considerable attention focused on the relationship between serum or plasma potassium (sK, pK) and cardiovascular (CV) events over the past several years.1 In individuals without known CV disease, especially those <65 years of age, it appears that a relatively wide range of sK values may be tolerated without a significant increase in CV risk.2 However, in patients with CV disease such as chronic kidney disease, heart failure (HF), diabetes mellitus or even hypertension alone the optimal range of sK appears to be narrower. In some studies 3.5–5.0 meq/l or 4.0–5.0 meq/l while in others the range may be even narrower and or include an upper limit of 5.5 or even 6.0 meq/l meq/l. Although there have been several studies evaluating the relationship between sK or pK in patients with HF and a reduced left ventricular ejection fraction (HFrEF), there is less information regarding this relationship in patients with HF and preserved left ventricular ejection fraction (HFpEF). Nishihara et al.3 in this issue of the Journal report on the relationship between sK and CV events in 506 consecutive patients in a single Japanese center discharged from hospital after an episode of HFpEF. The authors divided their patients into 4 quartiles of sK: Q1 sK < 4.1 meq/l, Q2 4.1–4.4 meq/l, Q3 4.4–4.7 meq/l, and Q4 sK ≥ 4.7 meq/l. Kaplan–Meier analysis revealed a significantly increased probability of CV death- and HF-related events in Q1, Q3, and Q4 compared with Q2. Although the authors point out that the use of diuretics might predispose to hypokalemia, they did not find a significant difference in their use among the four quartiles, nor the use of RAAS-Is, with the exception of an increased use of mineralocorticoid receptor antagonists (MRAs) in Q4, possibly associated with the increased risk of hyperkalemia in this group. The relationship between sK and CV events in this study in patients with HFpEF is similar to that in the Americas portion of TOPCAT,4 with the exception that in TOPCAT, there was an increased risk of CV events at a sK <3.5 rather than <4.0 meq/l, suggesting that sK should be maintained within a relatively narrow range in patients with HFpEF. The number of patients and CV events in both of these studies3,4 is, however, relatively small and further studies in a far larger number of patients with HFpEF will be required to provide confidence as to the exact level of sK below and above the normal range which should be of concern and trigger intervention. It is, however, likely that in a given individual the level of sk associated with an increase in CV risk will vary depending upon the rate of change in sK, the intracellular level of potassium and magnesium, the level of serum calcium, pH, and other factors. Thus, the data and thresholds for risk derived from the current and other retrospective studies evaluating the relationship between sK and CV risk should at best serve as a guide and decisions regarding initiation or a change in therapy should depend upon individual circumstances and clinical judgment. Nishihara et al.3 also attempted to determine if there was any significant difference in the relationship between sK and CV events in patients with HFrEF as compared with those with HFpEF. In an analysis of 200 patients from their center with HFrEF, there was an increased risk of CV events only in the Q4 group. Once again, this difference could be related to the relatively small number of patients and CV events in both those with HfrEF and HfpEF in this study. At the moment we believe the evidence suggests that there is little if any difference between the relationship of sK and CV events in patients with HfrEF as compared with those with HFpEF. Should further studies in a larger number of patients find a significant difference in this relationship between those patients with HFrEF, HF with mid-range ejection fraction (HFmrEF), and HFpEF we believe it will likely reflect significant differences in the use of CV drugs in those with HFrEF compared with those with HFmrEF or HFpEF, such as a possible increased use of diuretics or their higher doses in patients with HFpEF who are often older, have a higher incidence of uncontrolled hypertension and chronic kidney disease than those with HFpEF or HFmrEF or conversely a possible higher use of MRAs in patients with HFrEF, due to current HF guideline recommendations.5,6 Importantly, it is likely that the underlying degree of renal dysfunction and or presence and duration of DM determines the relationship between sk and CV risk rather than the presence of HF alone. These differences in sK thresholds for CV risk while potentially important may be less important than a further understanding of the gaps in our current understanding regarding sK such as whether or not sK is a risk marker or risk factor for an increase in CV events and if it is a risk factor how should we best reduce this risk. It is likely that below the optimum range, sK is a risk factor. Nishihara et al. point out that low SK levels even within the conventional normal range may impair myocardial function.3 The situation in patients with a sK above the normal range is, however, less certain. At the extremes, for example, a sK >6.0 or 6.5 meq/l sK may be a true risk factor whereas in those with a sk >5.0–<6.0 meq/l in some cases it may be a risk marker indicative of the underlying renal dysfunction while in other cases depending upon some of the factors enumerated above such as the rate of change and pH it could be a risk factor. In patients with HF or other CV disease with a sK below the optimal range such as <4.0 meq/l should we first add a potassium supplement such as KCL or should we add or switch to a K sparing diuretic such as amiloride or spironolactone? In those patients with CV disease and a sK or pK greater than the normal range such as >5.0 meq/l in whom it is thought that sK is a risk factor should we increase the dose of diuretic to reduce sK, decrease the dose or possibly discontinue RAAS-Is, especially MRAs or should we consider adding one of the newer potassium-lowering agents, such as ZS9 and Patiromer, both of which have been approved for use by the FDA and EMA should be available in most parts of Europe and the United States in the near future and have been shown to be relatively well tolerated and effective-over at least 1 year for Patiromer.7 In situations where one considers decreasing the dose or discontinuing a RAAS-I should the decision rest on the sK level alone or should it take into account a change in serum Cr? An increase in serum Cr may, however, be due to transient hemodynamic changes associated with the initiation or an increase in the dose of an ACE-I, ARB, and or MRA. The reason for an increase in serum creatinine after initiation of an RAAS-I, especially an MRA, is well understood and is associated with a decrease in vascular tone. There is no evidence that this decrease in renal vascular tone is detrimental and associated with adverse renal structural changes because RAAS-I withdrawal is associated with a return in serum creatinine toward normal.8 However, the discontinuation of a RAAS-I in a patient with HF, especially those with HFrEF, could place the patient at increased risk of death or further hospitalizations for HF. Therefore, it would be important to determine whether or not the reason for an increase in serum Cr and or sK was due to a transient hemodynamic effect that can be easily reversed, therefore enabling the continued use of the RAAS-I likely associated with long-term renal protection or whether it signifies increased renal damage and therefore the need to down-titrate or discontinue the RAAS-I and consider other strategies. In conclusion, we are indebted to Nishihara et al. and the many other investigators over the last few years who have focused our attention on the relationship between sk or pk and CV events in patients with HF as well as chronic kidney disease, DM, and hypertension. However not only do we need further data in a larger number of patients with HFpEF and other CV diseases to determine with greater certainty the level of sK or pK associated with an increase in CV risk in a given individual but as or more importantly we need further understanding as to when and how to intervene both below and above the optimal range of sk or pK. Despite the many recent studies evaluating the relationship between sk or pK and CV events in patients with CV disease9–11 large gaps in our knowledge remain providing an opportunity to further reduce CV mortality and morbidity. B.P. received personal fees (consulting) from Bayer, KBP Pharmaceuticals, AstraZeneca, Boehringer Ingelheim, Merck, Takeda, Relypsa, Sanofi, sc Pharmaceuticals, Sarfez pharmaceuticals, Stealth Peptides, Vifor and Tricida; has stock options from KBP Pharmaceuticals, sc Pharmaceuticals, Sarfez pharmaceuticals, Relypsa, and Tricida; has patent pending for site-specific delivery of eplerenone to the myocardium. P.R. received personal fees (consulting) from Novartis, Relypsa, AstraZeneca, Grünenthal, Stealth Peptides, Fresenius, Vifor Fresenius Medical Care Renal Pharma, Vifor and CTMA; received lecture fees from Bayer and CVRx; is a cofounder of CardioRenal.