Response to “Plant-based diets and postprandial hyperkalemia”

Abstract

Nutrition Reviews recently published our article on novel, etiology-based dietary strategies for managing hyperkalemia risk in people with chronic kidney disease (CKD).¹ Overall, the approach was drastically different from the traditional low-potassium diet and largely consistent with the current trend toward plant-based diets for people with kidney disease. In fact, one of the proposed strategies was consuming a high-potassium, not a low-potassium, diet.¹

In this editorial, we respond to the additional perspective offered by Varshney et al.² Because translating sparse nutrition research into dietary guidance is challenging, alternate viewpoints are generally welcome and necessary to move the field forward. This is why it was disappointing to read the commentary by Varshney et al.² Rather than expanding understanding with new insights and interpretations, they present a refined version of the analyses from our article. Depending on what is removed, refinement can produce a higher quality product. Unfortunately, as with refined grains, in this case, many of the most important components were lost in the refining process. Therefore, this editorial shall enrich their perspective, and in doing so, demonstrate that while their product may be more palatable, it may not best for the field.

Varshney et al² argue that our article¹ overstates the risk of postprandial hyperkalemia from high-potassium foods in people with CKD, specifically plant foods. Their commentary implied that the concerns were based solely on potassium kinetic studies and that this failed to account for the fact that most of these studies used large doses of potassium salts, which do not apply to plant foods that are rich in carbohydrates, alkali, and dietary fiber. Having proposed these concepts in the first place,³ we certainly hope they are correct. However, the limitations and gaps in the underlining research remain an issue.

Despite the known shortcomings of nutrient kinetic studies, they are often used in the field of nutrition to provide insight into the mechanisms and determinants of nutrient handling that have important dietary implications. Take the example of the oral glucose tolerance test. Postprandial glycemic response to meals depends on dietary factors other than glucose and exhibits considerable between-individual variability that is not fully explained by oral glucose tolerance test results.⁴ And yet, the oral glucose tolerance test is still used in clinical practice to identify people with impaired glucose tolerance who may be at risk of postprandial hyperglycemia and to guide treatment accordingly. In an analogous manner, we argued that “kinetic studies using potassium salts indicate that people with CKD have impaired potassium tolerance that may make them susceptible to transient increases in plasma potassium levels from dietary potassium,”¹ while also acknowledging that “other dietary factors may modify the effect of dietary potassium intake on postprandial kalemic response in people with CKD.”

The brief literature review provided by Varshney et al² contained substantive errors and omissions. The authors’ panache for brevity is appreciated, but not at the expense of clarity. Although mostly immaterial to this discussion, issues that may meaningfully affect interpretation of study findings are noted in Table 1^5–12 to correct the record and to highlight potential concerns in the rigor of the analysis of Varshney et al² of this research.

In our article,¹ we proposed that dietary factors present in plant foods may protect against postprandial hyperkalemia in people with CKD. It is surprising that Varshney et al² should so readily embrace this aspect of this theory given that the underlining research is no stronger than the kinetic studies they criticize as being irrelevant to food. Moreover, several practical considerations are pointed out that limit the ability of these factors to offset potassium load. Varshney et al² do not address these issues in their commentary. But, because they are central to our contrasting perspectives, they bear repeating here.

Broadly, the problems relate to scientific premise and clinical relevance. In terms of premise, both carbohydrates and alkalinity are supported by kinetic studies in people with CKD,^10,13 whereas the potential of dietary fiber to blunt kalemic response is largely based on shifts in potassium excretion seen in healthy individuals that may (or may not) be due to reduced potassium bioavailability.^14–15 The already tenuous case for dietary fiber is made more concerning by an important outlier in the research.¹⁶ In this feeding trial, potassium (∼782, 1564, and 2346 mg/d) from unfried potatoes, fried potatoes, and potassium gluconate was added to a background diet containing ∼2346 mg/day. Contrary to expectations, the vast majority of additional potassium from potatoes was recovered in urine (∼95%), similar to potassium gluconate, regardless of dose. Because potatoes are the largest contributor to vegetable intake in the US diet, with portion sizes often exceeding the standard serving size of 1 cup (∼140 g), these findings warrant consideration.¹⁷

In terms of clinical relevance, the doses of carbohydrates and alkali used in kinetic studies often exceeded the amounts normally present in high-potassium plant foods,^10,13 not to mention that some of these foods contain very little carbohydrate (eg, green leafy vegetables, avocados, nuts, seeds) and/or possess low base-forming to mild acid-forming potential (eg, green leafy vegetables, nuts, seeds, legumes). What is more, these relatively high doses of carbohydrates and alkali were still unable to prevent increases in plasma potassium concentrations.^10,13 As for dietary fiber, even if the aforementioned potato study is ignored¹⁶ and it is assumed that the changes in potassium excretion are entirely caused by lowering potassium bioavailability, the observed dose-response relationship (∼25 mg potassium/g fiber)¹⁴ is small compared with the ratio of potassium and fiber in most high-potassium plant foods (eg, 138 mg potassium/g fiber in bananas; US Department of Agriculture National Database no. 9040).¹⁸ Along with this, although potassium bioavailability may be reduced by dietary fiber, the majority of ingested potassium still appeared to be absorbed from plant-rich diets.^14,15

Importantly, the mechanistic research is limited, and the study of nutrients in isolation cannot fully capture the complex interactions that occur in food. In this, Varshney et al² rightfully highlight the recent feeding study by Byrne et al¹² as providing the best insights to date on postprandial hyperkalemia in people with CKD. Because the perspective of Varshney et al appears to be heavily influenced by this 1 study, it deserves careful examination.

The study by Byrne et al¹² was an unpowered, single-day, nonrandomized sequence, crossover, pilot feeding study (n = 8) that excluded individuals with diabetes mellitus and hypo- or hyperphosphatemia. The study was designed to evaluate a modified phosphorus diet that included (1) adding plant proteins (namely, pulses, nuts, and whole grains), (2) lowering the phosphorus to protein ratio, and (3) reducing phosphate-additive-containing foods. In fact, the title of the study makes no allusion to potassium or hyperkalemia. As such, although their findings are relevant, the study diets were not specifically designed to examine postprandial hyperkalemia.

In the Byrne et al study,¹² postprandial plasma potassium levels were measured in the afternoon, so only the breakfast and lunch meals are relevant. As shown in Table 2,¹² the modified diet does not reflect a simple substitution of plant for animal products. Additionally, the prescribed study diets only represent what was offered, not consumed, and at least 1 participant reported eating outside foods. Mean estimated potassium intake was noted to be slightly lower on the modified diet, albeit nonsignificantly (–89 mg/d; 95%CI –328 to 150 mg/d). However, this should be interpreted with caution because the potassium contents of the diets were calculated and many of the foods were prepared using wet cooking methods that are known to remove variable, but potentially substantial, amounts of potassium, including double-boiling potatoes.¹⁹ For example, the calculated phosphorus content of the modified diet in this study was found to overestimate phosphorus when compared with chemical analysis.¹²

The overview by Varshney et al² indicated that “although hyperkalemia occurred, it occurred in the standard diet with less plant foods, which illustrates the role of fiber, dietary alkali and carbohydrates in hyperkalemia—all of which are not found in potassium supplements.” Although postprandial hyperkalemia (>6.0 mEq/L) was more common on the standard diet (n = 5 of 8), it also occurred on the modified diet (n = 1 of 8).¹² As such, although the findings from Byrne et al¹² are certainly promising, they fall considerably short of demonstrating that dietary potassium from plant foods cannot contribute to postprandial hyperkalemia in people with CKD. Of interest, the conclusions of Byrne et al¹² appear to more closely align with ours (St-Jules and Fouque¹) than with Varshney et al.² In their discussion, Byrne et al state, “This result also indicates that caution may be required in permitting patients to save their dietary allowances for one meal, as with special occasions, and it may be more prudent to spread allowances over the day, from…a potassium load safety viewpoint.”¹²

A key aspect of translating nutrition research that is not informed by controlled feeding studies is the dietary habits and practices of free-living people. Although a balanced, portion-controlled, plant-rich diet such as the modified diet in the Byrne et al¹² article may afford the best protection against postprandial hyperkalemia in people with CKD (this is generally what was recommend in our article¹) it is unlikely that most patients will fully adhere to such a diet long term. This is why the approach was broken down into smaller, more manageable dietary strategies that can be tailored to the patient, and focused on general concepts like balance and portion control. The fact is, without studying them, one cannot accurately predict how different patient populations will respond to changes in dietary guidance and how this will affect their health outcomes.

The devices needed to examine dietary determinants of postprandial hyperkalemia in free-living people are finally being developed,²⁰ but the studies have yet to be conducted. In lieu of this standard, the observations of clinicians can be used to provide clues and insight. It is plausible, and certainly occurs to some extent, that people conditioned to believe that hyperkalemia is caused by dietary potassium will falsely attribute hyperkalemic episodes to recently consumed high-potassium foods. However, the observations of trained clinicians who are dealing with these problems regularly is more than mere anecdote and should not be disregarded.

When St-Jules et al³ first questioned the necessity of restricting high-potassium plant foods to manage hyperkalemia in people with kidney disease, the immediate response on the RenalRD (kidney dietitian) listserv was not elation.²¹ Instead, reactions ranged from general concern to describing the article as brainless and ridiculous. One comment (May 29, 2016) in particular stands out. They wrote, “Wow I can’t believe what I am reading. If empirical data isn’t out there showing that a high potassium diet causes life threatening hyperkalemia it can only be due to the fact that none of us have shared out experiences and published it.”

A systematic review of published case reports on diet-induced hyperkalemia included only a few cases attributed to high-potassium plant foods.²² But publication bias, particularly in case reports that tend to focus on novel, rare, and/or strange observations, is likely. In the years since the 2016 article by St-Jules et al³ was published, a consistent theme that has been reported by front-line dietitians, prominent leaders in the field of kidney nutrition, and others has been their experiences with hyperkalemia linked to high-potassium foods.^21,22 Although this does not constitute an unbiased analysis from a representative sample, it is in agreement with a survey of perspectives and practices of renal dietitians related to liberalizing restrictions on plant-based foods (n = 187).²³ In failing to address this issue, Varshney et al² imply that these clinicians’ observations do not count as evidence.

The final and perhaps most difficult aspect of translation is weighting the totality of evidence. In this, one must consider not only the confidence in dietary strategies but also the consequences of being wrong. Although we recommended that kidney dietitians continue to look for excessive intakes of high-potassium foods when conducting dietary assessment,¹ we specifically argued against the traditional practice of providing patients with lists of high-potassium foods to limit and avoid. Instead, we advised that clinicians inform patients considered to be at high risk of hyperkalemia that their condition may make them less resilient to abnormal dietary practices and, to avoid potential problems, they are encouraged to consume a balanced diet (ie, complete meals) and control their portions so they are not getting too much any of 1 food at a time.¹

Although this approach presents some restrictions of patients’ dietary practices, it is generally consistent with a healthy eating pattern and unlikely to cause much harm. If proven wrong, the field can continue on the current path toward liberalization. However, although we acknowledged that the “clinical relevance of postprandial hyperkalemia from dietary potassium is not well understood,”¹ and the risk from high-potassium plants foods is likely low at normal intake levels, the risk is certainly not zero, particularly because patients sometimes adopt atypical eating habits, often with their health in mind. If Varshney et al² are wrong, not only may promoting high-potassium plants foods as safe expose patients to risk of fatal hyperkalemia, it may necessitate backtracking (again) on dietary guidance. The field of kidney nutrition is already going through a protracted paradigm shift that is upending long-standing dietary guidance for people with CKD. By once again getting ahead of the evidence, it may compound the risk to the credibility of the field and, in the process, make the already difficult job of the kidney dietitian harder.

We thank Varshney et al² for sharing their perspectives. This response was lengthy, but we deemed it necessary to address concerns with their commentary and any corresponding misconceptions about the rationale for our approach.¹ Furthermore, the interpretation by Varshney et al is emblematic of a disturbing tendency for unbridled enthusiasm regarding plant-based dietary liberalization in the field. Although this work is theoretical, the implications for patients is all too real. This troubling trend appears to represent hopeful thinking that serves the field poorly. Because the approach we proposed is more cautious, the burden of proving high-potassium foods are safe is not on them. Instead, because it is not yet possible to say with enough confidence that high-potassium foods do not contribute to hyperkalemia in people with CKD, researchers and clinicians should resist the temptation of what they want to be true, and exercise patience with the research and caution with your patients. In this, the call from Varshney et al² for research on the postprandial kalemic response to whole, plant food is wholeheartedly supported and may be expanded slightly to consider the whole diet in free-living individuals. These questions could be answered in the near future by new techniques such as a 24-hour continuous subcutaneous monitoring of potassium, in the same way glucose is monitored by FreeStyle devices technology.²⁴

CONCLUSION

Sixty years ago, researchers and clinicians stopped short of conducting definitive dietary interventions, and the field is still dealing with the consequences of this decision. Understanding of dietary determinants of hyperkalemia in people with kidney disease has grown considerably since that time, particularly in the past few years, and the appearance of breakthrough studies like that of Byrne et al¹² provide some comfort that the necessary research is coming. In closing, the field should learn from the mistakes of the past by remembering that it is a Hippocratic oath, not a “Hubricratic” (“hubris”) oath, that is meant to guide patient care.

Acknowledgments

Author contributions. D.E.S. and D.F. wrote and approved this editorial.

Funding. The Nevada Agricultural Experiment Station in the College of Agriculture, Biotechnology & Natural Resources at the University of Nevada, Reno, supports the research of D.E.S.

Declaration of interest. The authors have no relevant interests to declare.

References