Low protein diets in patients with chronic kidney disease: a bridge between mainstream and complementary-alternative medicines?

Dialysis and transplantation are life-long, life-sustaining treatments requiring multiple drug therapies, and a high technological level of “mainstream medicine” [1‐3]. However, like all life-long, chronic diseases they require not only integration of complex therapies into everyday life (multiple drugs in the earlier phases of the diseases, and then drugs and dialysis or drugs and a kidney transplant, not necessarily in this order), but also a change in lifestyle, including diet.

In this regard, CKD treatments merge technology (dialysis, transplantation, multiple drug therapy), which belongs to/“mainstream medicine”, and lifestyle interventions (exercise, weight loss, low-protein diets) which are conversely often considered a part of the holistic approach attributed to “complementary and alternative medicine” (CAM).

While several comprehensive reviews have addressed the issue of the efficacy of various therapeutic approaches, including diet, to date, no reviews have been dedicated to the relationship between a mainstream and a CAM approach to low protein diets in CKD patients.

Indeed, nutritional treatments share these two facets: several recent papers aimed at a vast readership of family practitioners which assessed diverse diseases cite the addition of different foods or the use of specific diets as part of CAM approaches [4‐7].

For at least a century chronic uraemia has been considered an example of “protein intoxication”. Thus, in the history of “renal medicine”, diets have been attempted as a way to correct the metabolic alterations of kidney failure as much as possible [8, 9]. Indeed, the first systematic studies on low-protein diets in uraemia patients started with the observation that a protein-restricted diet was effective in reducing the symptoms of “uraemic toxicity” and that it was even able to prolong life [10‐12]. This role of life-prolonging therapies was maintained in western countries until dialysis became widely available; availability of dialysis, no more a problem in Europe, is however non universal on a global scale [12‐21]. Reconsidering the goals and modalities of low-protein diets in progressive renal diseases and keeping in mind the lessons of CAMs may offer an interesting outlook in the context of “high-tech” CKD care [20, 22‐24]. At a time in which growing interest (and business) in CAMs may divert attention from high-tech medicine, “playing in the two fields” may also aid the physicians in helping patients find effective coping strategies.

At the beginning of the new millennium, “renal insufficiency” was redefined and CKD became an umbrella term encompassing several affections having variable evolution rates towards end-stage kidney disease (ESRD) [25‐27]. The staging of CKD underlines the similarities of all advanced kidney diseases, despite different progression rates which depend upon race, age and underlying nephropathy [25‐35].

Therapies aimed at slowing CKD progression have to deal with high baseline heterogeneity that hinders the demonstration of their long-term advantages and drawbacks [30‐36]. Even the method that is used to measure progression rate is a matter of controversy. Furthermore, differences in the hard outcomes, namely start of dialysis or death, require trials with long observation times which are expensive and are often affected by many confounding factors [28‐36].

Several studies have focused on different therapeutic goals which may partially overlap: these goals include removing the cause of CKD, treating the metabolic derangements, correcting lifestyle habits that are not necessarily consequent to CKD but that may potentially worsen the diseases, and reducing the work-load on the remnant nephrons [20, 37, 38].

Dietary interventions in general, and low protein diets in particular may play a role at all these levels by, for example, treating obesity, correcting hypophosphataemia, avoiding processed foods, and most importantly, reducing the work-load on the remnant nephrons.

Furthermore, the outcome measure is non-univocal: slowing the progression of CKD is not synonymous of delaying dialysis [20].

Long-term follow-up of either homogeneous populations with similar baseline progression rates or of very large cohorts is ideally needed to demonstrate an effect of dietary interventions (and, indeed, one of the observations of the famous MDRD study was that it would have probably given different results over a longer follow-up [39‐41]). The initial GFR decrease, achieved by “renal rest”, or correction of the relative hyperfiltration may interfere with short-term analyses [42, 43].

Conversely, delaying dialysis requires attaining stable metabolic balance that is, at least partly, independent of GFR. However, the analysis of this short term goal is only apparently easier since it is deeply entangled with the pharmacological “mainstream approach” and reflects various policies of dialysis start which are not merely based upon (relatively) simple measures, such as GFR. The new Canadian guidelines on dialysis start clearly underline that whenever an “intent to defer” policy is chosen, wise clinical judgement cannot be replaced by formulae [44].

Heterogeneity of diseases, heterogeneity of progression, and heterogeneity of goals and measures add to the charm, but also to the complexity of the analysis of dietary approaches in CKD patients. Without a change in the approach, the issue will probably remain open forever.

Many of the questions posed in the 1986 paper published in the Lancet entitled: “Dietary treatment for chronic renal failure, ten unanswered questions” are currently still unanswered. Among them, some highlight the complexity and heterogeneity of chronic kidney disease (Does chronic renal failure always progress?), while others underline the difficulties in standardization of low protein diets in clinical practice (When should a low-protein diet start? Which low-protein diet? How should we assess progression, nutritional status or compliance?) [45]. However, in the literature there are several evidence-based demonstrations of the possible role low protein diets may play in delaying ESRD, including several RCTs and a Cochrane review [46‐58]. Furthermore, we are not aware of any large studies demonstrating a higher progression rate in non-malnourished, on-diet patients; in addition, data on lower costs are concordant [48, 49, 51], and there is no evidence showing that being on the diet results in any survival disadvantage after the start of dialysis [51, 59, 60].

The risk of malnutrition is frequently cited, but the data are at best as inconclusive as those cited by the diet “advocates” [61‐68].

Low protein diets are however still underutilized [20].

Table 1 summarizes some of the reasons why even “best quality” evidence is often considered non conclusive or does not affect the clinical management of CKD patients. Many of these reasons underline how the prescriptive modality, which is typical of high-tech medicine, may be less suitable than the educational modality, which is typical of CAMs, for the management of diets in CKD patients.

The first reason why the treatment with the best cost-benefit ratio is underutilized in the expensive CKD scenario may be the difference between the amount of time it takes to prescribe a drug or define the parameters of a dialysis treatment (mainly requiring technical skills), and to recommend lifestyle changes (especially in the case of diet) which are deeply rooted in personal preferences and social habits. In this regard, the barriers for the implementation of LPDs may be the result of the difficulty in conciliating the prescription and follow-up of a diet with a classic “mainstream medicine” approach. Some comments regarding the MDRD study, the largest randomised controlled study on LPDs in CKD, are in line with this interpretation: some experts claimed that MDRD actually demonstrated that prescribing a diet is not enough, since patients tend to lean to the dietary habits of their choice [39‐43]. Hence, the results of the “intention to treat” analysis (diet as a “mainstream prescription”) are negative (no advantage of an LPD on CKD progression), while the results of the “per protocol” analysis (diet as an educational intervention) suggest a positive role for LPDs in slowing CKD progression (Table 1).

The outstanding results obtained in some centres with a great deal of experience in selecting and following-up patients on low protein diets have rarely been replicated in other settings [51‐59]. The selection issue is fundamental: randomization is felt to be unethical on dialysis: this is why peritoneal dialysis and hemodialysis have never been randomised and their results are acknowledged as a matter of prescription and therapeutic adherence. Strangely enough, the same caveats are not applied to diets. Diet, as well as dialysis, depends upon the patients’ choices, and the few RCTs that were performed on dialysis frequency (3 versus 6 dialysis sessions per week) and those carried out on very low protein diets share an extremely low enrolment rate (in the 10–30 % range) [52‐54, 59, 69‐71]. Accordingly, one of the main criticisms of the unique Brunori study on diet versus dialysis is that it enrols about 30 % of the population screened for recruitment (and the percentage of cases randomised in the Garneata trial on very low protein diets is about 10 %) (Table 1). Such a low prevalence demonstrates the presence of a strong patient preference and suggests that adapting the prescription to the individual may be more successful than randomizing therapies. These issues merge with those related to quality of life, as will be further discussed [72‐77].

Wikipedia, the world’s largest source of on-line information, states “diet is the sum of food consumed by a person or other organism”; however, in clinical practice, the word “diet” is often interpreted with a restrictive (or even punitive) connotation, meaning the reduction of at least some foods [77, 78]. Limiting, reducing, and abolishing, are probably not the best slogans for a “marketing strategy” to improve compliance among CKD patients. Indeed, some authors presently prefer the terms “nutritional treatments” or “nutritional interventions” to define strategies aimed at changing dietary habits and not at restricting the use of some (many-most) foods [77‐80].

The concept of low-protein diets is presently undergoing a substantial change, mainly as a consequence of the changes in the definition of “adequate” protein intake which has shifted in recent years from 1–1.2 g/Kg/day to 0.8 g/Kg/day, according to the most recent FDA Reference Daily Intake or Recommended Daily Intake [81, 82]. The issue of dietary protein content in the overall population is complex, and is entangled with that of healthy lifestyle; the rediscovery of traditional and Mediterranean diets, and the attention to reducing additives and processed foods are two emerging aspects of this issue [83‐90].

A low protein diet is not necessarily a healthy diet; however, the cultural trend towards reducing proteins and canned, preserved and processed foods plays in favour of our patients, for whom the older “moderate” protein restriction (0.8 g/Kg/day) is now synonymous of a “biologically normal/adequate” protein intake, and who may choose a vegan-vegetarian or traditional diet (for example “Mediterranean”) with a lower risk of being “discriminated” in a carnivorous world [77].

Table 2 shows the main “menus” of low protein diets as defined according to; the daily protein intake relative to body weight, to the main foods, to the room for personalization and to the type of approach (“mainstream prescription” versus “CAM education”).

On Google, the term “mainstream medicine” retrieves over 3 million results and there are several different definitions among the first hits. Some of them are tautological and some are centred on “who” practices the medicine: “Medicine as practised by holders of M.D. or D.O. degrees and by their allied health professionals” [146]. Other definitions focus on the type of treatment “A system in which medical doctors and other healthcare professionals (…) treat symptoms and diseases using drugs, radiation, or surgery” [147]. Others are more complex and include an explanation of the pathogenesis of the diseases: a general term for conventional healthcare based on the “western model” of evidence-based practice for diagnosing and treating disease. Mainstream medicine assumes that all physiologic and pathological phenomena can be explained in concrete terms, and “best practice” is the final result of a stream of objective analyses. They begin with non-human model systems, evolve through blinded studies and a statistical analysis of the results, finally leading to guidelines to which doctors should adhere in order to achieve optimal patient outcomes [148]. In addition, they also include social elements: “The approach to healthcare as practised in developed nations is based on scientific data for diagnosing and treating disease; mainstream medicine assumes that all physiologic and pathological phenomena can be explained in concrete terms; the tools of mainstream medicine include non-human model systems, blinded studies, and statistical analysis to ensure reproducible results” [149].

Diet or nutritional approaches are never cited within these definitions of mainstream medicine. Conversely, on Medline plus, along with a general definition of Complementary and alternative medicines, CAM is the term for medical products and practices that are not part of standard care. Standard care is what medical doctors, doctors of osteopathy, and allied health professionals, such as nurses and physical therapists, practice. Dietary supplements are cited, while natural products are in first place in the classification of CAMs in the National Institute of health website [150, 151].

The concept of diet may however be different in mainstream and complementary medicine: in the first case, it is usually based upon specific, measurable prescriptions, while in the second case it is more commonly part of a lifestyle, often with emphasis on disease prevention [7]. The limits are elusive, and no wonder, therefore, if a broad search strategy on Medline, combining the terms “Chronic kidney disease” and “alternative” or “complementary” or “allied” medicine retrieves, along with papers on CAMs, several studies on diets which are often defined as “alternative” to early dialysis, or “complementary” to pharmacological treatments. Thus, even if diet in CKD is not strictly considered a CAM, the terms currently employed for defining CAMs (complementary or alternative) identify many relevant recent papers on the dietary treatment of CKD [152‐157].

The use of these terms is probably not only a semantic exercise: diet is often cited as an alternative to early dialysis or as a treatment that is complementary to ACE-inhibition or to other drug therapy (for instance, diuretics) [20, 158‐161]. In this regard, a critical observation from the CAM point of view may offer some insights for a better understanding of the limits of applying “conventional” evidence-based medicine (EBM) analysis to low-protein diets in CKD, as well as into the possible role for a “whole system” analysis and for a narrative, personalized approach to maximize compliance (or, perhaps better stated as “concordance”) to low-protein diets in CKD patients [162‐165].

In the era of personalised medicine, we might wonder whether a randomised controlled trial is the best way to demonstrate the effects of changes in dietary habits and, more than this, if it is ethically sound to randomise such an intrusive lifestyle change [166‐169].

In this regard, mainstream medicine should reflect on the lesson learned from CAMs and the focus might shift from proving the superiority of a single treatment (diet vs no diet; moderately- versus severely-restricted diets etc.) to an observational multidimensional “whole system” approach [7, 170‐177].

A very interesting health technology assessment on the role of the preferences of professionals and patients in randomised controlled trials states that “All RCTs in which participants and/or professionals cannot be masked to treatment arms should attempt to estimate the participants’ preferences….” [178]). Current approaches may not be fit to highlight differences, and other study designs, in particular the new “partially randomized patient preference (PRPP) trial model”, which was developed for the analysis of CAMs, have already been applied to several fields of mainstream medicine [179‐186].

No study is without pitfalls, and there are sound criticisms to virtually every study design [187]. However, according to an educational report in the BMJ: “when the effectiveness of the intervention depends on the subject’s active participation (…) a randomised trial may be inappropriate because the very act of random allocation may reduce the effectiveness of the intervention” [173].

If the focus on diets has to shift from issues regarding effectiveness to those of feasibility, then the implementation strategy should probably be personalized and should take into account the features of each patient [142]. This “narrative” approach may be seen as the other side of the moon as compared to a randomised trial approach. However, evidence-based medicine is fully compatible with narrative, individualized approaches provided that flexible, comprehensive and innovative study designs are chosen [188‐190]. Paraphrasing a JAMA paper “Scientific reports (or RCTs) are genuinely dispassionate, characterless, and ahistorical. But their translation and dissemination (implementation in clinical practice) should not be” [191]. Dealing with personalized medicine, we should probably learn from patients’ preferences, and be happy with the fact that a treatment may have positive results, at least in experienced centres, in the absence of sound evidence on severe drawbacks and side effects. In this regard, the lessons learned from dialysis, which is chosen on the basis of the patients’ preference in the setting of an overall equivalence of side effects, may be extended to the diet, which is probably wrongly considered less intrusive in daily life than the choice of the different types of renal replacement therapy [69‐71].

Talking about diet, broadly defined on Wikipedia as “everything we eat”, is extremely complex and no review of a single issue, such as protein intake, can be exhaustive without touching upon other aspects, such as caloric intake, vitamins and oligo-elements, additives and contaminants. In particular, this is crucial in CKD where the presence of kidney disease impairs the activation of vitamin D, leading to a deficient absorption of calcium and to hyperphosphataemia; low-protein diets are usually deficient in vitamin B12, vitamin D, and iron. Conversely, hyper-vitaminosis may be a threat, in particular for vitamin D [192‐194]. The toxicity of additive and trace elements, such as pesticides, has only recently been studied; negative effects are likely to be enhanced by reduced renal depuration [194‐196]. We know less about individual elements, such as selenium, which is important in the oxidative pathway; furthermore, the network of saturated-unsaturated fatty acids is extremely complex and only partly known, and bicarbonate supplementation is still an open issue, while the discussion on all these items is beyond the scope of this review [197‐199]. Again, a flexible holistic approach, which is more common in CAMs than in mainstream medicine, may be of interest. Allowing at least one (and in our experience up to three) free meals per week may help avoid deficits, generically speaking by enriching the diet, but the idea that “body wisdom” can help identify which foods the body needs is appealing [24, 200]. A holistic approach not only to the patient, whose diet is an important part of daily life, but also to the diet as a whole and not only as a “reduction of” may reduce the risk of focusing on single elements, and thus losing sight of the whole problem.

There is probably at present no best low-protein diet, but, in the absence of contraindications, the best approach is probably to offer the patients a choice of diets with different options and protein content, bearing in mind that the most important issue is not the optimal prescription, but the optimal compliance and target achievement (Table 2).

In this regard, the prescriptive approach that deals with low protein diets as it does with drug therapies may be combined with the educational approach that is typical of CAMs by integrating dietary habits into daily life [201‐203].

Patients should be aware that there is a high degree of uncertainty with regard to the choice of their dietary treatment and that its aim is to reduce the signs, symptoms and complications of renal failure and then to delay the need for dialysis. Furthermore, even if this is beyond the scope of this review, they should be warned that a low-protein diet is not necessarily a healthy diet, and that several other factors, such as type of renal disease, type of proteins, salt and phosphate intake, use of canned and processed food, and effective energy intake can modulate the risks and the clinical success [194‐196].

Long-term interventions that require a profound change in everyday life are difficult to assess, at least with the current study strategies. Clinical research needs long-term observations and long term follow-up, whenever possible even after the start of dialysis, to check for carry-over effects and for attrition biases. Since the report and publication biases suggest that only experienced centres with sound, positive data should publish their series and studies, efforts should be made to build large multi-centre series involving centres with different types of experience on low-protein diets. Multi-centre studies should take into account the important cultural differences in eating habits and cuisine which are often present even within the same country. Both “per protocol” and “intention to treat” analyses may be relevant in chronic interventions and both should probably be analysed [203, 204].

Since it is somewhere between a mainstream prescription and a complementary therapy, the study of diet in chronic kidney disease should merge the suggestions of both approaches. Integration among various study designs that take into account patients’ preferences (observational, randomised, partially randomised, etc.) is more promising than focussing on a single type of analysis, be it randomised or observational.