Should We Let Dialysis Patients Eat Their Fruits and Veggies?

Hyperkalemia is a frequent abnormality in patients treated with thrice weekly maintenance hemodialysis (MHD) (1). Its clinical relevance is underscored by potassium’s physiologic effects on the membrane potential and the acute arrhythmogenicity of its excursions (both hypo- and hyperkalemia) (2), and the widely described association of abnormal serum potassium concentrations with adverse clinical outcomes in diverse patient populations, including patients with kidney failure treated with MHD (3). As a result, clinical guidelines recommend both acute interventions to lower elevated serum potassium and preventative measures to avert its (re)occurrence in patients who are prone to its development (4). Among the latter, a core intervention is restricting the amount of dietary potassium intake, based on the fact that hyperkalemia results from either higher intake, lower excretion, redistribution from the intracellular space to the extracellular space, or a combination of these (4). The seemingly rational concept of dietary potassium restriction is especially compelling in patients with limited kidney function, such as those with CKD or kidney failure treated with MHD, in whom tilting the potassium balance toward normal with enhanced excretion is less feasible. In particular, patients with kidney failure treated with MHD who are anuric have to rely on a fixed amount of dialytic potassium excretion and on enhanced colonic potassium excretion to maintain long-term potassium mass balance, neither of which may be able to serve as a rapid physiologic counter-balance to high dietary potassium loads. Dietary potassium restriction in patients with kidney failure treated with MHD is thus widely used in clinical practice; while the exact amount of what the restricted intake should be is ill-defined, a typical “low potassium” kidney failure diet would contain about 50–77 mEq (or approximately 2–3 g) of potassium per day (5). The logic of this intervention seems undeniable, and hence few dare to question it, let alone to recommend against implementing it. This is despite the lack of compelling evidence supporting the effectiveness of dietary restriction in preventing hyperkalemia, let alone its ability to improve clinical outcomes linked to hyperkalemia.

In this issue of CJASN, Bernier-Jean et al. set out to fill this void by examining the association of dietary potassium intake with hyperkalemia and with clinical outcomes (all-cause, cardiovascular, and noncardiovascular mortality) in 8043 adult patients receiving thrice weekly hemodialysis, recruited from several European countries and from Argentina (6). The investigators also examined whether hyperkalemia may mediate the putative association between (higher) potassium intake and (worse) clinical outcomes, and they also attempted to tease out the complex interrelationships of dietary food groups, potassium intake, and clinical outcomes, since the food groups that are highest in potassium intake (e.g., whole fruits and vegetables) have numerous other components that could confound the association of potassium intake with clinical outcomes. Potassium intake was quantified from the Global Allergy and Asthma European Network food frequency questionnaire, which determines an individual’s dietary composition averaged over the preceding 1 year, the reproducibility of which has been validated, albeit not in populations with kidney failure treated with MHD. In a series of complex, but well-planned and executed, statistical analyses, the authors concluded that higher potassium intake showed favorable associations (lower risk of noncardiovascular deaths) when adjusting for baseline demographic and various clinical characteristics, but the association was attenuated to nil once adjusting for food groups, implying that the favorable associations were due to confounding by the ancillary benefits afforded by high-potassium containing food groups (e.g., higher amount of antioxidants, vitamins, fiber, alkali content, etc.) (5). The authors also found no association of higher dietary potassium intake with higher serum potassium concentrations, or with hyperkalemia events, and detected no mediation effects of serum potassium in the association of dietary potassium intake and clinical outcomes.

The findings of Bernier-Jean et al. are thus questioning the validity of the above-mentioned paradigm that higher dietary potassium leads to higher serum potassium, which then leads to worse clinical outcomes. Their study is by far the largest observational study examining these questions, and offers important new knowledge to complement and extend the findings of previous smaller studies (7,8). While the lack of an association between dietary potassium intake and hyperkalemia seems counterintuitive, it is possible (and indeed likely) that foods that are high in potassium have several ancillary effects that can mitigate or negate the impact of their higher potassium content on serum potassium concentration. These include effects that lower the absorption of the potassium from the gut by decreasing constipation and shortening intestinal transit time, and effects that facilitate the distribution of potassium toward the intracellular space through alkalinization or by decreasing insulin resistance (9). The net impact of these could indeed be a neutral effect on serum potassium of diets that otherwise contain higher amounts of potassium.

One should then rightfully ask whether we should abandon the age-old practice of potassium restriction, and whether we should in fact promote heart-healthy diets despite their higher potassium content. Notwithstanding the findings of Bernier-Jean et al., we may not have reached the point where we can recommend such changes. First and foremost, we need to examine the reliability of the results provided by Bernier-Jean et al. Their analytical approach was sophisticated, and the robustness of their findings is supported by a series of carefully planned subgroup and sensitivity analyses. The prospective design and the use of validated tools further strengthen fidelity in their findings. What could give us pause is that the authors describe an association between serum potassium concentration and clinical outcomes that fails to replicate findings from scores of previous observational studies, namely, a U-shaped association, with both hypo- and hyperkalemia being associated with higher all-cause and cardiovascular mortality (1,3). Hypothetically, the fact that only 29% of the cohort had available serum potassium measurements, that these measurements were obtained from mid-week specimens (as opposed to specimens obtained after the long interdialytic interval), that they were extracted from the dialysis provider’s clinical database (as opposed to protocol-driven research collections), and that patients’ serum potassium levels were characterized using a single measurement could all have had an impact on the presented results. Furthermore, using a dietary assessment tool that examines average intake over the preceding year and that is unable to account for hidden sources of potassium, such as food or other additives, may be unable to account for acute fluctuations in potassium intake, which may be important in causing transient hyperkalemic events. Finally, there may be selection bias in enrolling patients in a prospective cohort of diet assessment. While Bernier-Jean et al. used advanced analytical approaches to mitigate many shortcomings of the data, some of the limitations are generic and cannot be negated analytically.

What could then be done to move us toward a more informed clinical application of dietary potassium manipulation? One method would be confirmation of the presented findings from other studies, but it is unlikely that we will have larger or better conducted cohorts of dietary assessment in populations with kidney failure treated with MHD. What we truly need are clinical trials designed to test the viability of the paradigm that potassium restriction (versus no restriction) can (or cannot) impact serum potassium levels, which could then provide sufficient knowledge to engage in longer-term trials of potassium manipulations aimed at clinical outcomes in patients with kidney failure treated with MHD. Engrained clinical practices and long-held beliefs about the need for dietary potassium restriction in patients with kidney failure treated with MHD can make the implementation of such trials difficult, but not impossible. The advent of novel potassium binders now allows us to pharmacologically lower the amount of potassium absorption even in patients with kidney failure treated with MHD (10), which can offer a security blanket for trials assigning individuals to high-potassium containing foods. Indeed, such trials are now in progress (ClinicalTrials.gov identifier: NCT04997161) and offer hope that we will soon see a conclusion to the vexing question of dietary potassium manipulation in dialysis patients. We may yet come to recommend our dialysis patients to eat their fruits and veggies.

Disclosures

C.P. Kovesdy has received consulting fees from Abbott, Akebia, AstraZeneca, Bayer, Cara Therapeutics, CSL Behring, Rockwell, and Vifor; royalties from Springer and UpToDate; research funding from AstraZeneca, Bayer, Gilead, and GSK; and honoraria from Abbott, Akebia, AstraZeneca, Bayer, Cara, CSL Behring, Rockwell, and Vifor. C.P. Kovesdy serves as an Associate Editor of CJASN and Nephron and serves on the editorial boards of American Journal of Kidney Disease, International Urology Nephrology, Kidney International Reports, Kidney Medicine, and Nephrology Dialysis Transplantation.

Funding

None.