The history of dietary protein restriction for renal failure shows manifold influences—personalities, philosophic orientation, economics, professional identity, cultural issues—on the growth of medical knowledge and the evolution of medical practice. I briefly outline this history, with emphasis on fluctuations in research and practice. This article is not based on an exhaustive literature review, nor do I assess the clinical merits or scientific basis of this dietary treatment.

Early History of Dietary Protein Restriction

Dietary protein restriction for renal failure was suggested as early as 1869. In a textbook on diseases of the urinary tract, Beale wrote:

A large proportion of excess of meat taken passes off from the body in the form of urea and other urinary constituents, which it is the special work of the kidney to remove from the blood. It is obviously of the utmost importance to relieve the kidneys of at least this unnecessary and useless work in cases in which they are diseased, when their working power is seriously impaired [1].

At that time an important school of medical opinion emphasized the "healing power of nature". The physician's role was to provide—by diet, hygiene in general, and medication—an environment in which such healing could occur [15]. Dietary restriction was also consistent with "depletive" treatment programs, popular in the 19th century, to decrease the "morbidly animated patient to a healthy, natural state [16]".

Early in this century, some patients with renal failure receiving extreme dietary protein restriction fared poorly, presumably because of protein malnutrition [17]. High protein diets were recommended partly in reaction to these cases; partly in accord with a reversal in therapeutic practice from "depletive" to stimulating or "repletive" regimens [18]; partly because high-protein diet was known to induce renal hypertrophy [5], and because renal disease increased urinary protein losses [19]. Confusion about appropriate dietary protein allowance for patients with kidney disease also stemmed from controversy about the minimum protein allowance (0.5 g/kg daily) recommended for normal persons [20, 21]. Many observers, dating back to Voit in 1882 [22], believed that additional protein conferred strength and well-being. The greater protein intake of the developed countries was correlated with their greater "success". One authority declared that "a large protein allowance is the right of civilized man [23]". The opinion that protein builds strength may even now underlie physicians' reluctance to limit dietary protein [13], regardless of the established minimum daily requirement.

In the first four decades of this century, interest in nutrition burgeoned, driven by research in the vitamin deficiencies [14]. Nutrition was taught as an independent course, or a prominent component of biochemistry and physiology, in medical schools. Hospitals offered as many as fifty-five specialized diets [24]. Dietary protein research for renal disease intensified in the 1920s, when Newburgh showed that high protein (but not urea) feeding could produce the pathologic lesions of "Bright's disease" in rabbits [2] and in rats [3]. Addis and colleagues confirmed that renal hypertrophy could be stimulated by unilateral nephrectomy [4] and by protein feeding [5]. In 1927, Moise and Smith [6], noting that Newburgh's observation had been extensively followed up with controversial results [25, 26], showed that high-protein diet in uni-nephrectomized rats produced hypertrophy, proteinuria, and extensive pathologic changes, including capsular adhesions and segmental glomerulosclerosis. The authors [6] cautioned against clinical application of their work, however, noting that the dietary protein had been high (85%) and that results obtained with uni-nephrectomy would not apply to most clinical cases. They did not make the now-familiar analogy of the remnant kidney to chronic renal disease of other causes. In 1932, Chanutin and Ferris [7] produced an influential model of chronic renal failure in the rat. They tied off the upper and lower poles of both kidneys in sequential operations, achieving a 50% to 75% nephrectomy. The rats developed proteinuria, progressive azotemia, renal hypertrophy, hypertension, and cardiac hypertrophy. The rate of development of these changes was proportional to dietary protein, which was varied from 10% to 80% [9, 10]. Inconsistent results, however, led Chanutin [9, 10] and Newburgh and Johnston [27] to doubt the specific role of protein. In 1939, Farr and Smadel [8, 11] showed that dietary protein restriction was protective in experimental (Masugi) nephritis. They fed rats diets of 5%, 18%, and 40% protein and noted the development of chronic nephritis in the high-protein groups as well as recovery from nephritis (and, unfortunately, malnutrition) in the 5% protein group.

These studies constituted a formidable body of work in animals on the prevention or exacerbation of renal disease by manipulations of diet. But interest in this area abated until the 1980s. In trying to account for this hiatus, it is useful to look at textbooks written by influential authors in the nascent specialty (not yet named) of nephrology. These authors expressed and shaped the predominant attitudes of their time.

Addis's Glomerular Nephritis

The large body of work in renal hypertrophy in response to renal ablation and to protein feeding culminated in Addis' rather idiosyncratic volume of 1948, Glomerular Nephritis, Diagnosis and Treatment [17]. Addis unified previous observations of protein-induced hypertrophy and of renal ablation under the rubric of "renal work". He noted,

... the practical identity of the curves of increase in weight, whether the increase in work is produced by an increase in protein consumption ... or by a reduction in the amount of working renal tissue leaving the protein consumption unchanged ... Such an identity indicates that, as far as the kidney is concerned, both of these situations are alike, that both induce the same sort and rate of growth, and that the effect of the increase in work common to both is not appreciably altered by the additional factors that in one of these cases accompany an increase in protein consumption [28].

Addis used Beale's term: In both types of hypertrophy the remaining nephrons were subject to increased "work". His insight is strikingly modern, entirely analogous to the concept of nephron hyperfiltration with normal protein intake in renal insufficiency and with high-protein intake in healthy persons. Osmotic work [29] was the thermodynamic work needed to achieve final urine solute concentrations, given that initial concentrations presented to the kidney were those of plasma. Urea, the most plentiful solute in urine, accounted for the bulk of this osmotic "work" and for the renal burden of high protein diet. Addis cited the benefit of dietary protein restriction in Masugi nephritis [8, 11] to support the view that nephron reduction by ablation and by disease were analogous. He provided data on the effects of dietary protein restriction on renal work and on a restricted set of end points: kidney weight, blood urea nitrogen, proteinuria, hematuria, and urinary casts. He did not do controlled studies in humans to show the advantage of protein restriction; convinced of its efficacy, he thought that such studies would be unethical [17].

There are several reasons why Addis's ideas were not pursued. First, his suggestions foundered on a specific scientific issue, osmotic work. Most of the work of the kidney was expended in processes, such as the filtration and reabsorption of electrolytes and glucose, that did not affect final urine concentrations. Renal oxygen consumption did not increase during osmotic diuresis [30, 31]. Homer Smith [32] put a definitive end to the concept of osmotic work in 1951: "Physiologically, the work represented by the composition of final urine is an almost negligible fraction of the work it is known that the kidney must do in order to make that urine, and the calculation of the so-called minimal thermodynamic work has no more significance". Smith concluded [32] that low-protein dietary therapy could not be defended on the ground that it reduced osmotic work. His tone indicated that he considered protein-induced renal injury a muddled affair: "The belief has been expressed from time to time that a high protein diet is injurious to the kidney in renal disease". Smith did not take a position in this debate. Only a single sentence in his massive text, The Kidney, concerns the deleterious effects of high dietary protein on renal structure [33]. Perhaps Smith thought that if the concept of osmotic work was discredited, the concept of protein-induced renal injury was discredited as well.

A second reason for failure to pursue Addis's ideas was his own confusion of the purposes of dietary protein restriction. At times, as in a paper on the effect of protein infusion in acute renal failure, he seemed to advocate dietary protein restriction to alleviate the accumulation of uremic toxins: residual nephrons failed to keep pace with work demanded and a "relative" renal failure ensued [34]. At other times, as in dietary protein restriction for acute glomerulonephritis, he was concerned with preservation of nephrons; the goal of protein restriction was to promote "complete healing of the lesion [35]". The use of low-protein dietary therapy to minimize the generation of uremic toxins in acute renal failure became an active issue in the 1940s [34, 36-38] and, paradoxically, eclipsed the earlier work on glomerulosclerosis. Some subsequent authors attributed the benefits of low-protein dietary therapy found in previous studies to the effect of minimizing the generation of uremic toxins and assumed that the effect on kidney function had been misinterpreted or overstated. This is how Merrill assessed Addis's work in his textbook of 1955 (revised in 1965). Merrill [39] argued that hypertrophy might be a good thing because it provided more working renal tissue; protein restriction was needed only when obvious nitrogen retention (by which he meant uremia) was present. Merrill followed Homer Smith in dismissing the concept of osmotic work, and he interpreted Addis's work as if it referred only to minimizing uremic toxins rather than to preventing renal injury.

Finally, Addis failed to persuade others of the utility of dietary protein restriction because he did not provide convincing pathologic correlations. Although he emphasized kidney weight and preservation of renal mass, he did not apply the pathologic findings of the 1920s and 1930s to his clinical work: he did not show glomerulosclerosis in patients. Renal hypertrophy was an ambiguous end point, a possibly appropriate response. Addis would have been more convincing with a histopathologic approach.

Homer Smith and the Tradition of Claude Bernard

Homer Smith was the founder of modern renal physiology. It is striking that, in his textbook of 1951, he devoted so little attention to protein-induced renal injury [33] and so much to the effects of protein feeding to increase glomerular filtration rate and to stimulate renal hypertrophy [40]. One reason was that he doubted the concept of osmotic work. Another was that he thought in terms of homeostatic responses. The increase in glomerular filtration rate or renal mass after protein feeding was appropriate. Renal injury was not. He did not attach much importance to protein-induced glomerulosclerosis.

Smith's outlook was admirably summed up in his preface to The Kidney [41]. He began with a long quotation from Claude Bernard, culminating in Bernard's declaration that:

(t)he perpetual changes of external conditions cannot reach it (the organism); it is not subject to them but is free and independent ... All the vital mechanisms, however varied they may be, have only one object, that of preserving constant the conditions of life in the internal environment.

In this line of thinking, dietary variation was one of those "changes of external conditions" of which the internal milieu was independent. Bernard's [42] seminal experiments showed that during fasting the liver produced glucose and maintained a constant plasma concentration; thus, the blood sugar was independent of the dietary sugar. Bernard declared explicitly in his Introduction to Experimental Medicine, one of the founding documents of scientific medicine [42], that the external environment (and, by implication, the healing power of nature) could profitably be ignored.

Smith agreed, extending the concept of the constant internal milieu to renal physiology.

In the last analysis, composition of the plasma is determined not by what the body ingests but by what the kidneys retain and excrete.

This outlook was fundamental to modern understanding of fluid and electrolyte balance (consider the current approach to hyponatremia and hyperkalemia) and to establishing nephrology as a discipline. But it was not helpful to investigating the deleterious effects of protein feeding. The kidneys reacted appropriately to increased dietary protein by increasing the glomerular filtration rate, eliminating nitrogenous wastes, thus maintaining constancy of the internal milieu. The logic of homeostasis did not accommodate the observation that protein feeding adversely affected renal function. The doctrine of the autonomous internal milieu, appropriated to nephrology by Homer Smith and so important to the rise of scientific medicine, eclipsed the role of dietary treatment.

Peters and van Slyke: Quantitative Clinical Chemistry

A third viewpoint was provided by the 1946 classic text of John P. Peters and Donald van Slyke [43]. In contrast to Addis (whose view of the work of others was personal and idiosyncratic) and to Homer Smith (who ignored the tradition of protein-induced renal injury), these authors thoroughly reviewed the relevant literature. Like Chanutin [9, 10] and Newburgh [27], they questioned whether the apparent nephrotoxicity of high-protein diet could be due to other factors (they suggested choline deficiency, excessive cholesterol, saturated fat, or biotin). They worried about the adverse effect of low-protein diet therapy on general nutrition. Chanutin's and Smadel's rats had improvement of renal failure at the cost of malnutrition. Peters and van Slyke stated for the patient with chronic glomerular nephritis,

Complete recovery is now no longer to be expected and therefore the general condition and well being of the patient assume relatively more importance than the state of his kidneys. To condemn him to invalidism and malnutrition on the doubtful pretext of preserving or prolonging the anatomical and function integrity of the renal tissue is hardly a defensible policy ... . Dietary treatment should be governed by the general condition, the state of nutrition, and the symptoms of the patient and should be aimed to maintain nitrogen equilibrium. To meet these requirements the diet of an adult should contain one gram of protein per kilogram of the body weight ... [44].

Overestimating the daily protein requirement, they believed that malnutrition due to dietary protein restriction was too high a price to pay for preservation of the kidneys. In addition, they perceived a "distinct danger in unbalanced diets" [45]. Dietary restrictions risked an increase of "nephrotoxic factors" or a deficiency of "protective factors". Their views were characteristic of an era in which nutrition research centered on the vitamin deficiencies. Although they came at the problem from a different angle than Homer Smith, whose implicit assumption was that the kidneys adapted appropriately to dietary variation, the practical effect of their views was similar. Smith and Peters trained a remarkable proportion of future leaders of American nephrology; Addis trained few if any. Perhaps Smith and Peters did not actively discourage research in this area; but their writings also did little to stimulate it.

Emergence of the Modern Paradigm

The importance of nutrition in research and practice diminished after the second World War. The classic deficiency diseases "ceased to be a public health problem," and nutrition "was relegated to a low place in the curriculum and no longer was taught as an independent course [46]". Nutrition was taught in the disciplines of biochemistry and physiology, but they shifted toward cell biology and molecular biology. Increasingly, nutrition was taken up by medical popularizers and by the lay press. Attention to diet was probably inconsistent with the professional identity of scientific medicine during these years. Diet was passive therapy, harking back to the healing force of nature and to restoration of natural equilibrium; but active modification of physiologic processes and correction of deviations from normality had become the professional standard. Hence, the oft-repeated complaint, to which the medical profession was indifferent, that doctors "knew nothing" about diet. Knowledge of nutrition was not, the profession seemed to believe, what made one a doctor.

Research on the effects of protein feeding on the kidneys stagnated after the 1940s, even though considerable interest existed in low-protein dietary therapy to treat acute and chronic uremic symptoms. Renal ablation in the rat was recognized as a good model of the systemic and renal responses to human chronic renal failure [47, 48] but not as a model for loss of nephrons from other causes. Some protein feeding studies were done in various models of renal injury [49, 50]; but there was no sense that protein feeding was a model for progressive renal failure in general. Protein feeding was regarded as a special form of nephropathy, a "nutritional glomerulosclerosis".

Several conceptual changes helped to renew interest in dietary protein restriction for prevention of progressive renal failure. First, the minimum daily protein requirement to prevent negative nitrogen balance was clarified. Giordano [51] and Giovanetti [52] showed that uremic patients on a virtually nitrogen-free diet could maintain nitrogen balance; the diet could minimize or reverse uremic symptoms when patients were given essential amino acids or protein of high biological value. Thus, protein restriction was compatible with neutral nitrogen balance not only in normal persons [21, 53] but also in patients with renal failure [51, 52, 54]. Modifications of the Giordano-Giovanetti diet to improve palatability and compliance were made in various other centers [38].

A second important shift was the slowly emerging concept that chronic renal failure was a unitary condition with unique determinants and pathologic and pathophysiologic features. In 1950, Oliver [55] suggested that it would be helpful to regard the failing kidneys as collections of diverse and "fantastic" structures, the altered nephrons. This insight implied that chronic renal failure was a pathophysiologic entity in itself, with specific nephron characteristics. Bricker and colleagues [56, 57] developed this idea, showing some of the common physiologic features of "intact nephrons" in chronic renal failure. One of these features was glomerular hyperfiltration, demonstrated by Bricker and colleagues [56, 57] and by Morrison and Howard in 1966 [58]. In 1969, Nagle and colleagues [59] published an analogous observation in structural terms: Glomerular obsolescence had a similar appearance by light as well as by electron microscopic examination in various renal diseases. Thus, a unitary view of chronic renal failure emerged. It was as if "Bright's disease," a term that fell out of use about 1950 [60], had been resurrected after decades of discrimination between the various causes of renal disease. Perhaps the emergence of a class of patients united by their eligibility for dialysis or renal transplantation encouraged this unitary view. In any case, chronic renal failure came to be viewed as a discrete entity with its own determinants and course rather than the outcome of various diseases. It was then possible to see the analogy (which Addis had seen years before) between renal ablation and loss of nephrons by disease: In either case, the remaining nephrons were prone to common conditions, such as hyperfiltration.

A third factor was the emerging notion in diverse clinical areas that degenerative diseases (congestive heart failure, cirrhosis, and chronic obstructive pulmonary disease as well as renal failure) tended to progress after resolution of initial injury. Initially appropriate adaptations to injury could have adverse late consequences. In cardiology, hypertrophy (teleologically an apparently appropriate response) was recognized as a precursor of heart failure; reducing the work of the heart by "unloading" systemic blood pressure was beneficial. These words and concepts were analogous to those used by Beale and Addis: Low dietary protein intake reduced the work and hypertrophy of the kidneys. The advent of unloading therapy for heart failure was nearly contemporaneous with Bricker's "trade-off hypothesis" [61], which was based on an analogous argument: A useful adaptation (increased parathyroid hormone to facilitate phosphate excretion) could have unwanted consequences (hyperparathyroid bone disease). Phosphate restriction could prevent or ameliorate secondary hyperparathyroidism in renal failure. Although the trade-off hypothesis did not bear directly on dietary protein restriction for chronic renal failure, it did imply that appropriate defenses of the internal milieu had potentially dangerous consequences.

Altogether, these shifts in perception created a different climate in the 1970s. Walser described patients in whom dietary protein restriction and supplements of essential amino acids or ketoanalogues slowed progression of renal failure [62-64]. Similar observations (of renal protection by diet) were familiar to several workers in the field but had been ascribed to other causes [38]. The crucial assumption, for which Oliver's and Bricker's work had prepared the ground, was that renal failure progressed in itself. Walser suggested that calcium-phosphate complexes were deposited in the kidney as a consequence of the hyperphosphatemia of renal failure; the benefit of low-protein dietary therapy was that such diets were also phosphate poor. This hypothesis predicted a self-generating mechanism of progressive renal failure: Loss of renal function increased serum phosphate, enhancing calcium-phosphate deposition, and further decreasing renal function. The hypothesis subsequently garnered experimental support in an animal model [65].

In 1975, on the basis of pathologic studies in the remnant kidney model, Shimamura and Morrison [66] speculated that hyperfiltration in remaining glomeruli could mediate glomerulosclerosis and could be a final common pathway of renal failure. Several ideas had converged in Morrison's work. Single nephron hyperfiltration was a more persuasive example of renal work than Addis's osmotic work had been. Developments in cardiology and in chronic renal failure were precedents for the idea that useful adaptations could have adverse consequences. Most important was the consensus that chronic renal failure was a common pathway with common structural and functional features, a process or mechanism rather than an outcome. All of these developments enabled Shimamura and Morrison [66] to see the analogy between the remnant kidney model and clinical chronic renal failure.

Reception of the Hyperfiltration Hypothesis

When taken up by Brenner in 1981 [12], the hyperfiltration hypothesis [66] generated great excitement among nephrologists. Brenner and colleagues showed [12] that renal ablation in the rat caused hyperfiltration in remaining nephrons; that glomerulosclerosis occurred in the remnant kidney; and that hyperfiltration and glomerulosclerosis were prevented or ameliorated by protein restriction. The observations on glomerulosclerosis and the benefit of low-protein diet therapy had been made many years before [2-11], and the causal role of hyperfiltration was suggested, but not proved, in Brenner's work [67, 68].

Nevertheless, the effect of this work was profound. Why? The hypothesis was presented with supporting data rather than as speculation. Sophisticated scientific technique was employed and dietary modification was not merely providing a hygienic environment but was an active physiologic manipulation. Dietary therapy was thereby integrated into scientific professional identity. An appealing mechanism was offered for what had previously been a puzzling observation. In the absence of such a mechanism, the use of low-protein dietary therapy seemed merely empirical, "hygienic," and out of date.

Why was the hyperfiltration hypothesis so attractive? Fundamental to the hypothesis was the irony that an apparently appropriate short-term adaptation (hyperfiltration to facilitate excretion of nitrogenous wastes by a diminished nephron population) caused long-term injury (glomerulosclerosis). The mechanism was, so to speak, "antihomeostatic," a homeostatic response ultimately mediating injury. This mechanism coincided with a change of outlook since the time of Homer Smith: a new suspicion or doubt of the unalloyed usefulness of homeostatic responses; a new attentiveness to the external environment. The "adaptations" of modern life engendered bigger problems: Industrialization compromised the global environment; dialysis spawned a new set of diseases, such as aluminum intoxication and dialysis amyloidosis, as well as a new set of economic and social problems. In this climate, the idea that renal adaptation mediated renal disease was appealing.

Role of Dialysis and Transplantation

Despite the success of dietary protein restriction in clinical and laboratory studies, its use was swamped by the advent of dialysis and transplantation. As Bergstrom has written:

In spite of the success of our group and many others with low protein diets and amino acids or keto analogues, this form of treatment was not universally accepted as an alternative treatment. Reasons for this may have been more general availability of chronic dialysis and, in USA and some other countries, reimbursement systems which made alternatives to dialysis less urgent and less profitable for the patient as well as for the doctor. It is more difficult to understand why dietary treatment is not used more frequently in countries where resources for chronic dialysis for economical reasons have remained inadequate [38].

Dietary protein restriction for uremic symptoms was recommended in standard textbooks [69, 70] but often passed over in clinical practice in favor of dialysis and transplantation [13, 38]. Posing the question "Which is preferable—long-term dietary therapy and then dialysis, or earlier dialysis?" Berlyne [71] answered that "Modern dialytic therapy is preferable to prolonged dietary therapy, and if dialysis is available it should be started sooner rather than later". The burden of dietary restrictions and the difficulty of compliance were the reasons for this preference. Giovanetti [13] identified four reasons for reluctance to use dietary treatment in preference to dialysis: "the opinion that a high protein intake is necessary for good health"; doubt that patients could comply with the diet; unreimbursed time and effort required of the physician; and economic incentives to dialysis and transplantation. In lectures that I attended, Brenner suggested (to the consternation of clinical nephrologists) that economic and professional incentives to dialysis and transplantation had wrongly shifted attention away from dietary therapy. However, high remuneration for dialysis and transplantation could have been a consequence of the evaluation that patients, physicians, and policymakers made of these technology-intensive treatments, an evaluation more favorable than the one they made of dietary treatment.

As the advent of dialysis and transplantation had eclipsed dietary protein restriction for renal failure, increasing dissatisfaction with the end-stage renal disease program renewed interest in dietary treatment. The program was more costly than had been anticipated, quality of life was restricted, and rehabilitation (especially keeping a job) was disappointing. Some critics thought that dialysis profits and utilization were excessive, the worrisome consequences of private enterprise [72, 73]. Although dietary modification could not suffice when the kidneys failed completely, it might slow progression to renal failure or ameliorate uremic symptoms when some kidney function remained; if so, it could reduce costs of the end-stage renal disease program. In 1978 and 1980, "the rapid growth of this program and the substantial costs associated with long-term dialysis and transplantation resulted in a congressional mandate (expressed in Public Laws 95-292 and 96-499) to the secretary of health and human services to... develop and carry out a demonstration project to determine 1) the extent to which the commencement of nutritional therapy in early renal failure, utilizing [but not limited to] controlled protein substances, can retard or arrest the progression of the disease with a resultant substantive deferment of dialysis, and 2) the administrative, financial and other aspects of making such nutritional therapy generally available as part of the benefits under title XVIII of the Social Security Act" [74].

Government payors, together with "expert advisory groups" [74] of academic nephrologists, initiated a multicenter trial of dietary protein restriction (Modification of Diet in Renal Disease) now in progress. The study was justified on the ground that previous clinical trials of protein restriction to prevent renal failure were inconclusive and "deficient with respect to experimental design" [74]. Although these deficiencies did exist, lack of definitive scientific information was not enough to explain lack of enthusiasm in clinical practice. Compare the enthusiastic use of coronary artery bypass grafting long before its appropriate use was defined by controlled studies. On the contrary, it was lack of enthusiasm for dietary protein restriction that, in part, motivated the Modification of Diet in Renal Disease trial. Desultory use of dietary treatment increased payors' (government) incentive to obtain more scientific information. Had use of dietary treatment for renal failure been greater, a Congressional mandate to commit substantial resources to this issue would not have existed. From the payors' point of view, the prestige of a randomized, controlled study could induce physicians to use this therapy and justify third-party reimbursement for it.

The recent resurgence of interest in dietary protein restriction for renal failure is part of renewed interest, especially by government agencies and patient groups [75], in nutrition to prevent chronic degenerative diseases. But medical practitioners and educators have not fully shared this interest. Despite many government-sponsored conferences, increased funding for nutrition research and training, and several federal agency reports on the inadequacy of medical education in nutrition, surveys have shown no change in the nutrition curriculum and have shown a perception among medical students of inadequate nutrition education [76]. Preventive measures in general occupy a small portion of American medical expenditure, and physicians could rightly point out that their efforts in prevention have not been rewarded by the reimbursement system. Thus, the Modification of Diet in Renal Disease trial is an important milestone: a government-sponsored study to get more information about, and thereby increase the use and reimbursement of what is essentially a preventive measure.

Nutrition and Scientific Medicine

Perhaps nephrologists have not been prepared to forego the high technology that has fostered recognition, by themselves and the public, of the power of medicine, and that is now at the core of their professional identity. Dialysis and renal transplantation are active interventions that normalize blood chemistry and volume, the internal milieu; they exemplify the scientific medical model that Claude Bernard and Homer Smith adumbrated. Dietary treatment has continued to be seen mainly as a means of maintaining a healing environment, of avoiding nutritional deficiencies. Many nephrologists have believed that their patients were not capable of, or could not be happy with, dietary protein restriction, a prophesy that an unpersuaded physician could, wittingly or not, make self-fulfilling. Underlying these attitudes were deeper motives: professional commitment to and abundant resources to support the technology of dialysis and renal transplantation; cultural attitudes about the virtue of high-protein intake; mistrust of "passive" dietary treatment; and, perhaps, unwillingness to restrict food consumption (or consumption of high technology) in a society committed to unlimited growth.

Most important has been physicians' ambivalence to nutritional treatment, which for many remains old-fashioned and merely hygienic. Such treatment has not accorded with their modern professional identity; scientific medicine has discounted the "healing power of nature". But new paradigms of the relation of diet to scientific medicine have been developed: diet to prevent chronic degenerative disease; diet to make physiologic manipulations; diet to avoid adaptations that would ultimately be deleterious; diet to save costs. Perhaps these paradigms, including economic constraints, will speed integration of dietary treatments into the armamentarium of scientific medicine. In any case, cultural, social, and economic factors assure that dietary treatment of renal failure (and dietary treatment in general) will undergo many vicissitudes in years to come, as they have in the past.