Brain metabolism of amino acids and ammonia in patients with chronic renal insufficiency

Fatigue in Patients Receiving Maintenance Hemodialysis: A Review

Fatigue surrounding hemodialysis treatments is a common and often debilitating symptom that impacts patients' quality of life. Intradialytic fatigue develops or worsens immediately before hemodialysis and persists through the dialysis treatment. Little is known about associated risk factors or pathophysiology, although it may relate to a classic conditioning response. Postdialysis fatigue (PDF) develops or worsens after hemodialysis and may persist for hours. There is no consensus on how to measure PDF. Estimates for the prevalence of PDF range from 20%-86%, likely due to variation in methods of ascertainment and participant characteristics. Several hypotheses seek to explain the pathophysiology of PDF, including inflammation, hypothalamic-pituitary-adrenal axis dysregulation, and osmotic and fluid shifts, but none is currently supported by compelling or consistent data. PDF is associated with several clinical factors, including cardiovascular and hemodynamic effects of the dialysis procedure, laboratory abnormalities, depression, and physical inactivity. Clinical trials have reported hypothesis-generating data about the utility of cold dialysate, frequent dialysis, clearance of large middle molecules, treatment of depression, and exercise as potential treatments. Existing studies are often limited by sample size, lack of a control group, observational design, or short intervention duration. Robust studies are needed to establish the pathophysiology and management of this important symptom.

Branched-Chain Amino Acids Depletion during Hemodialysis Is Associated with Fatigue

BACKGROUND

Fatigue is one of the most debilitating symptoms reported by maintenance hemodialysis (MHD) patients. Hemodialysis causes marked depletion in plasma essential amino acids. We studied the cross-sectional relationship of pre- and post-hemodialysis branched-chain amino acids (BCAAs) concentrations with fatigue in MHD patients.

METHODS

MHD patients self-reported fatigue during a dialysis session using the Brief Fatigue Inventory. Pre- and post-dialysis plasma levels of BCAAs (valine, leucine, and isoleucine) were measured using HPLC-mass spectrometry.

RESULTS

The mean age of study participants (n = 114) was 54.8 ± 12.8 years. Plasma levels of BCAAs decreased significantly post-dialysis compared to pre-dialysis (303.8 ± 9.4 vs. 392.1 ± 9.4 μM/L, p < 0.0001). Fatigue score increased as a function of age (p = 0.015). There was no association between pre-dialysis plasma levels of BCAAs and fatigue. A significant negative correlation was found between post-dialysis plasma levels of BCAAs and fatigue (p < 0.05).

CONCLUSIONS

These preliminary findings suggest that disruption in BCAAs homeostasis may play a role in precipitating fatigue.

The role of Neurotransmitters in the Genesis of Uremic Encephalopathy

To classify the influence of neurotransmitters in the genesis of uremic encephalopathy we studied cerebrospinal fluid (CSF) and plasma (P) amino acid (AA) concentration, in patients undergoing various dialytic treatments (hemodialysis = HD, intermittent and continuous peritoneal dialysis = IPD and CAPD).

HD causes a significant decrease in CSF/P ratios of branched chain AA (BCAA) and a significant increase in CSF Glycine/Valine ratio, suggesting an augmented brain uptake of Glycine at detriment of Valine. In IPD the general trend of Aromatic AA/BCAA ratio suggests a preferentilal transport of Aromatic AA through the blood brain barrier. The differences between IPD and HD are confirmed by data concerning metabolites of Serotonin and Dopamine: CSF concentrations of 5-Hydroxyndoleacetic acid and Homovanillic acid are low in HD but high in IPD.

So, a reduced (in HD) and an increased (in IPD) activity of monoamine systems could be at the basis of some neurological disturbances appearing in uremia.

Hyperprolactinemia as a Marker of Neurotransmitter Imbalance in Uremic Population

Serum prolactin (PRL) levels are elevated in patients with chronic renal failure (CRF) but the mechanisms responsible for these abnormalities are not fully understood. PRL secretion is undoubtedly influenced by many substances, which can be variously altered in uremia: monoamines, endogenous opiates and PTH. Our data suggest that in early renal failure PRL levels are already significantly high and the 24-h pattern of PRL secretion is significantly different from that in controls.

PRL derangements could be due in mild renal failure, to unknown factors (GABA?); in severe CRF, to a major change in dopaminergic activity; in hemodialysis (HD), to a low turnover of monoamines, and in peritoneal dialysis (PD) to increased activity of sero-toninergic and dopaminergic systems.

Insulin resistance and the metabolism of branched-chain amino acids in humans

Peripheral resistance to insulin action is the major mechanism causing the metabolic syndrome and eventually type 2 diabetes mellitus. The metabolic derangement associated with insulin resistance is extensive and not restricted to carbohydrates. The branched-chain amino acids (BCAAs) are particularly responsive to the inhibitory insulin action on amino acid release by skeletal muscle and their metabolism is profoundly altered in conditions featuring insulin resistance, insulin deficiency, or both. Obesity, the metabolic syndrome and diabetes mellitus display a gradual increase in the plasma concentration of BCAAs, from the obesity-related low-grade insulin-resistant state to the severe deficiency of insulin action in diabetes ketoacidosis. Obesity-associated hyperinsulinemia succeeds in maintaining near-normal or slightly elevated plasma concentration of BCAAs, despite the insulin-resistant state. The low circulating levels of insulin and/or the deeper insulin resistance occurring in diabetes mellitus are associated with more marked elevation in the plasma concentration of BCAAs. In diabetes ketoacidosis, the increase in plasma BCAAs is striking, returning to normal when adequate metabolic control is achieved. The metabolism of BCAAs is also disturbed in other situations typically featuring insulin resistance, including kidney and liver dysfunction. However, notwithstanding the insulin-resistant state, the plasma level of BCAAs in these conditions is lower than in healthy subjects, suggesting that these organs are involved in maintaining BCAAs blood concentration. The pathogenesis of the decreased BCAAs plasma level in kidney and liver dysfunction is unclear, but a decreased afflux of these amino acids into the blood stream has been observed.

Acute supplementation with keto analogues and amino acids in rats during resistance exercise

During exercise, ammonia levels are related to the appearance of both central and peripheral fatigue. Therefore, controlling the increase in ammonia levels is an important strategy in ameliorating the metabolic response to exercise and in improving athletic performance. Free amino acids can be used as substrates for ATP synthesis that produces ammonia as a side product. Keto analogues act in an opposite way, being used to synthesise amino acids whilst decreasing free ammonia in the blood. Adult male rats were divided into four groups based on receiving either keto analogues associated with amino acids (KAAA) or a placebo and resistance exercise or no exercise. There was an approximately 40 % increase in ammonaemia due to KAAA supplementation in resting animals. Exercise increased ammonia levels twofold with respect to the control, with a smaller increase (about 20 %) in ammonia levels due to exercise. Exercise itself causes a significant increase in blood urea levels (17 %). However, KAAA reduced blood urea levels to 75 % of the pre-exercise values. Blood urate levels increased 28 % in the KAAA group, independent of exercise. Supplementation increased glucose levels by 10 % compared with control animals. Exercise did not change glucose levels in either the control or supplemented groups. Exercise promoted a 57 % increase in lactate levels in the control group. Supplementation promoted a twofold exercise-induced increase in blood lactate levels. The present results suggest that an acute supplementation of KAAA can decrease hyperammonaemia induced by exercise.

Application of branched-chain amino acids in human pathological states: renal failure

During renal failure, abnormalities of BCAA and branched-chain keto acid (BCKA) metabolism are due to both the lack of renal contribution to amino acid metabolism and the impact of renal failure and acidosis on whole-body nitrogen metabolism. Abnormal BCAA and BCKA metabolism result in BCAA depletion as reflected by low plasma BCAAs and cellular valine. BCAA metabolic disturbances can alter tissue activities, particularly brain function, and nutritional status. In dialysis patients, BCAA oral supplementation can induce an improvement of appetite and nutritional status. During chronic renal failure, the aims of nutritional interventions are to minimize uremic toxicity, avoid malnutrition and delay progression of kidney disease. BCAA and BCKA supplements have been proposed to decrease further protein intake while maintaining satisfactory nutritional status. In this setting, BCAAs or BCKAs have not been administrated solely but in association with other essential AA or keto analogs. Therefore, the proper effects of BCAAs and/or BCKAs have not been studied separately. Protein restriction together with keto acids and/or essential AAs has been reported to improve insulin sensitivity and hyperparathyroidism and to be compatible with a preservation of nutritional status. Nonetheless, a careful monitoring of protein-calorie intake and nutritional status is needed. A recent meta-analysis concluded that reducing protein intake in patients with chronic renal failure reduces the occurrence of renal death by approximately 40% as compared with larger or unrestricted protein intake. The additional effect of essential amino acids and keto acids on retardation of progression of renal failure has not been demonstrated.

Ammonia functions as a regulatory molecule to mediate adjustments in glomerular filtration rate in response to changes in metabolic rate

In mammals, varying in size from mouse to elephant there is a close linkage between metabolic rate (MR) and glomerular filtration rate (GFR). This linkage has evolved because of the necessity to excrete nitrogenous wastes generated predominately by amino acid catabolism. A similar linkage is observed in birds, reptiles and fish, although in the latter vertebrate group gill excretion rate and GFR need to be considered as an interdependent co-ordinated system. A change in the rate of amino acid catabolism will result in a change in the rate of production of nitrogenous wastes requiring excretion and hence GFR (or GFR plus gill excretory rate in fish) must be appropriately re-set. It is hypothesized that ammonia functions as a regulatory molecule to mediate adjustments in excretory rate (GFR or GFR plus gill excretory rate) in response to such changes in amino acid catabolism.

Central nervous dysfunction in uremia

The mechanisms of central nervous system dysfunction in uremia are multifactorial and only partially characterized. Studies using sealed presynaptic nerve terminals (synaptosomes) for in vitro ion transport and metabolism of neurotransmitter in chronic renal failure (CRF) neuronal cell culture and in vivo brain structure microdialysis generated significant new information. An increase in total calcium content of the cerebral cortex accompanied by increased levels of cytosolic calcium ([Ca(2+)]i) in synaptosomes are common findings in rats with CRF. Mechanisms leading to the increase in [Ca(2+)]i include increased calcium uptake mediated by parathyroid hormone and decreased activity of Na(+),K(+)-adenosine triphosphatase (ATPase) and Ca(2+)-ATPase of synaptosomes in CRF rats. Moreover, these synaptosomes respond inappropriately to depolarization, which can impair neurotransmitter metabolism. Brain gamma-aminobutyric acid content, norepinephrine, and acetylcholine release uptake and degradation are affected by uremia. These may lead to certain somatic, behavioral, and motor dysfunctions in uremia. Many derangements of the central nervous system in uremia appear to be mediated by secondary hyperparathyroidism of CRF because parathyroidectomy of animals with CRF prevented the increase in basal levels of [Ca(2+)]i and derangements in neurotransmitter metabolism. The role of other neurotoxins, such as guanidinosuccinic acid, are also reviewed.

Inter-organ substrate exchanges in the critically ill

Metabolic inter-organ exchange is a major field of research for improving the treatment of the critically ill. Adapting regional blood flows is the first regulatory step, although the relationships between hypoperfusion and metabolic disorders are matter of controversy. Metabolic steady state results from a vast inter-organ interplay and several nutrients or metabolites are signalling molecules in the regulation of gene transcription. Inter- or intra-organ substrate recycling shares or delays the mandatory need for aerobic ATP synthesis in some conditions. Nitrogen metabolism is highly compartmentalised in an inter-organ co-operation and liver, muscle, kidney and gut are the most important organs. By remodelling the amino acid mixture delivered to peripheral cells after intestinal absorption, the liver plays a determinant role in whole body protein synthesis. Albumin turnover increases after brain injury. Since the location of synthesis is different to that of breakdown this turnover can be viewed as an inter-organ exchange. The metabolic side of pH homeostasis is also an inter-organ exchange mainly shared by liver, kidney and muscle.

Mechanism of increases in L-kynurenine and quinolinic acid in renal insufficiency

Marked increases in metabolites of the L-tryptophan-kynurenine pathway, L-kynurenine and quinolinic acid (Quin), were observed in serum and cerebrospinal fluid (CSF) of both the rat and human with renal insufficiency. The mechanisms responsible for their accumulation after renal insufficiency were investigated. In patients with chronic renal insufficiency, elevated levels of serum L-kynurenine and Quin were reduced by hemodialysis. In renal-insufficient rats, Quin and L-kynurenine levels in serum, brain, and CSF were also increased parallel to the severity of renal insufficiency. Urinary excretion of Quin (3.5-fold) and L-kynurenine (2.8-fold) was also increased. Liver L-tryptophan 2,3-dioxygenase activity (TDO), a rate-limiting enzyme of the kynurenine pathway, was increased in proportion to blood urea nitrogen and creatinine levels. Kynurenine 3-hydroxylase and quinolinic acid phosphoribosyltransferase were unchanged, but the activities of kynureninase, 3-hydroxyanthranilate dioxygenase, and aminocarboxymuconate-semialdehyde decarboxylase (ACMSDase) were significantly decreased. Systemic administrations of pyrazinamide (ACMSDase inhibitor) increased serum Quin concentrations in control rats, demonstrating that changes in body ACMSDase activities in response to renal insufficiency are important factors for the determination of serum Quin concentrations. We hypothesize the following ideas: that increased serum L-kynurenine concentrations are mainly due to the increased TDO and decreased kynureninase activities in the liver and increased serum Quin concentrations are due to the decreased ACMSDase activities in the body after renal insufficiency. The accumulation of CSF L-kynurenine is caused by the entry of increased serum L-kynurenine, and the accumulation of CSF Quin is secondary to Quin from plasma and/or Quin precursor into the brain.

Arteriovenous difference of blood ammonia in uremic patients under hemodialysis

In this study the ammonia concentration was determined in arterial and venous blood samples pre- and posthemodialysis (HD) in 18 uremic patients and in 18 health subjects (controls). The mean values (+/- SD) of ammonia in the arterial blood of uremic patients pre-HD were 98.32 +/- 26.55; post-HD, 63.18 +/- 17.09; and in control group patients, 72.37 +/- 10.09 micrograms/dl. In venous blood they were pre-HD, 71.70 +/- 20.68; post-HD, 58.05 +/- 16.73; and in control patients, 74.46 +/- 12.0 micrograms/dl. According to our findings, the ammonia concentration in the arterial blood of uremic patients pre-HD exceeds the normal limits and is significantly higher (p < 0.001) than that post-HD and that of control patients. The ammonia contents of venous blood pre- and post-HD ranges were within normal values, but the post-HD range was significantly lower than the pre-HD range (p < 0.05) and the control range (p < 0.01). Comparison between ammonia levels from arterial and venous blood showed significant and positive arteriovenous differences pre-HD (p < 0.001), which disappeared post-HD and were not observed in the control patients. In conclusion, uremic patients under HD present pre-HD high levels of ammonia in arterial blood with a significantly positive arteriovenous difference. In contrast, the post-HD ammonia levels in arterial and venous blood are decreased, and the arteriovenous difference is not significant.

Skeletal muscle protein synthesis and degradation in patients with chronic renal failure

Muscle protein turnover and amino acid (AA) exchange across the forearm were studied in nine postabsorptive patients with chronic renal failure (CRF) under unrestricted calorie-protein diets and eight controls by using the arterio-venous difference technique associated with the 3H-phenylalanine kinetics. In patients with CRF: (1) the rate of appearance (Ra) of phenylalanine (Phe) from the forearm, reflecting proteolysis, was 27% increased in comparison with controls (P < 0.01). Also the rate of disposal (Rd) of Phe, reflecting protein synthesis, was increased in patients (P < 0.01). As a consequence of these counterbalanced alterations, net balance of Phe across the forearm, that is, net proteolysis, was not changed. (2) The release of total AA from the forearm was not different from controls. Valine and ketoisocaproate release was reduced (P < 0.05). Serine uptake was not detectable. (3) Net proteolysis and the Rd/Ra ratio were inversely and directly, respectively, related to arterial [HCO3-] (P < 0.02 and P < 0.03, respectively). (4) Moreover, net proteolysis and Phe Rd/Ra ratio were directly and inversely, respectively, correlated with plasma cortisol (P < 0.01 and < 0.005, respectively). Plasma cortisol was in the normal range and inversely related to arterial [HCO3-] (P < 0.02). (5) While in controls phenylalanine appearance from the forearm was inversely related to insulin levels, no correlation was found in patients. In conclusion, in patient with CRF, forearm Phe kinetics indicate the existence of an increased muscle protein turnover. Changes in protein synthesis and degradation are well balanced and net proteolysis is not augmented.(ABSTRACT TRUNCATED AT 250 WORDS)

Erythropoietin treatment and plasma levels of corticotropin-releasing hormone, delta sleep-inducing peptide and opioid peptides in hemodialysis patients

An improvement of quality of life and objective brain function has been reported in patients receiving regular hemodialysis treatment (RDT) during treatment with recombinant human erythropoietin (r-huEPO). The mechanisms explaining this improvement are unknown. In this study the plasma levels of peptides known to be involved in CNS functions, namely corticotropin-releasing hormone, delta sleep-inducing peptide, beta-endorphin, methionine-enkephalin, beta-lipotropin and alpha-melanocyte-stimulating hormone, were measured by radioimmunoassay in seven stable RDT patients before the start of r-huEPO therapy and during 28 weeks' treatment. All patients responded with significantly increased hemoglobin concentrations. An improvement of well-being, state of mood and physical fitness was reported by the patients. There were no significant changes during the study in the plasma concentrations of any of the peptides measured. However, as the plasma levels of neuropeptides will not necessarily reflect the local concentrations in the vicinity of the nerve terminals, changes in the intracerebral concentrations of these peptides might occur in response to r-huEPO.

Renal ammoniagenesis in humans with chronic potassium depletion

Renal ammonia production and distribution and ammonia precursor utilization were evaluated in eight patients with chronic potassium depletion (CPD) and aldosterone-producing adenoma and in 20 controls. In CPD, urinary ammonia excretion and ammonia added to renal venous blood were about twofold higher than in controls; thus, total ammonia production was significantly augmented (88.0 +/- 10.3 mumol/min.1.73 m2 vs. 45.0 +/- 2.6 in controls). Total ammonia production was inversely correlated with serum potassium and directly correlated with urine flow. Stepwise multiple regression analysis showed that both factors, mainly serum potassium, significantly influence ammonia production and account for 61.4% of variations in ammonia production. Renal extraction of glutamine was significantly increased (56.6 +/- 5.9 mumol/min.1.73 m2 vs. 34.6 +/- 3.1 in controls), and this could account for ammonia production. The ratio of urinary ammonia excretion to total ammonia production, an index of the intrarenal ammonia distribution, was similar in patients and controls, and was significantly correlated with urine pH and true renal blood flow (RBF). Stepwise multiple regression analysis showed that RBF, urine pH and urine flow also significantly affected ammonia distribution. However, these factors accounted for only 41.7% of variations in intrarenal ammonia partition, urine pH having a minor role. We conclude that in patients with CPD other factors besides urine pH, urine flow and RBF intervene in the ammonia partition between urine and blood.

Improvement of brain function in hemodialysis patients treated with erythropoietin

To evaluate the effects of recombinant human erythropoietin (rHuEPO) on brain function, 15 chronic hemodialysis patients were studied by event-related P300, stimulus-related evoked potentials, and trailmaking before (hematocrit 22.7%) and after rHuEPO (hematocrit 30.6%). P300 peak latency elicited by a tone discrimination paradigm improved (391 before vs. 366 ms after; Cz = vertex; P less than 0.01) confirming beneficial effects on cerebral cognitive processing. P300 amplitude (13.6 vs. 15.8 microV; P = 0.06) and trailmaking tended to improve (55 vs. 43 s). P300 measures were influenced by low hemoglobin levels before rHuEPO (P less than 0.01), suggesting that severe anemia may contribute to uremic brain dysfunction. Furthermore, decrease of stimulus-related auditory brainstem I-V interpeak latency (4.28 before vs. 4.17 ms after; P less than 0.05) and increase of somatosensory N20/P25 amplitude (4.8 vs. 7.0 microV; P less than 0.05) pointed to improvement of sensory pathways by mechanisms unrelated to cognition. Brain dysfunction in chronic hemodialysis patients may, beside other factors, in part be caused by severe anemia and can be improved by rHuEPO treatment.

CAPD with an amino acid dialysis solution: a long-term, cross-over study

This prospective cross-over study was undertaken to evaluate the safety and efficacy of a 1% amino acid dialysis solution on the nutritional and metabolic changes, plasma amino acid profiles and peritoneal membrane function of patients on CAPD. Six CAPD patients had one exchange a day with two liters of this solution over a six month period. Every month there was a medical examination, anthropometric measurements and dietary inquiry were made, blood biochemistry tests were done. Every three months renal function, peritoneal function, aminograms of plasma and dialysate and nitrogen balance were determined. Data were compared with those obtained one month prior to and three months after withdrawal of amino acid administration. Nitrogen balance, which was negative (-1.3 g/day) became positive (+3.1 g/day). Patients who were already overweight increased in weight, both in fat and lean mass. Plasma cholesterol and triglycerides significantly decreased and the amino acid profile moved towards normal; plasma urea levels increased and pH and bicarbonate decreased slightly but significantly (P less than 0.05). Plasma protein concentrations did not change. All the above parameters turned towards basal values when amino acids were discontinued. We conclude that amino acids can be used as osmotic agents for CAPD since they do not cause toxic effects or impair peritoneal membrane function. Moreover, they can help the nutritional status, provided that an increase in weight is prevented and the slight worsening of systemic acidosis is corrected.

[15N]aspartate metabolism in cultured astrocytes. Studies with gas chromatography-mass spectrometry

The metabolism of 2.5 mM-[15N]aspartate in cultured astrocytes was studied with gas chromatography-mass spectrometry. Three primary metabolic pathways of aspartate nitrogen disposition were identified: transamination with 2-oxoglutarate to form [15N]glutamate, the nitrogen of which subsequently was transferred to glutamine, alanine, serine and ornithine; condensation with IMP in the first step of the purine nucleotide cycle, the aspartate nitrogen appearing as [6-amino-15N]adenine nucleotides; condensation with citrulline to form argininosuccinate, which is cleaved to yield [15N]arginine. Of these three pathways, the formation of arginine was quantitatively the most important, and net nitrogen flux to arginine was greater than flux to other amino acids, including glutamine. Notwithstanding the large amount of [15N]arginine produced, essentially no [15N]urea was measured. Addition of NaH13CO3 to the astrocyte culture medium was associated with the formation of [13C]citrulline, thus confirming that these cells are capable of citrulline synthesis de novo. When astrocytes were incubated with a lower (0.05 mM) concentration of [15N]aspartate, most 15N was recovered in alanine, glutamine and arginine. Formation of [6-amino-15N]adenine nucleotides was diminished markedly compared with results obtained in the presence of 2.5 mM-[15N]aspartate.

Uremic encephalopathies: clinical, biochemical, and experimental features

Patients with renal failure may manifest a variety of neurologic disorders. Patients with chronic renal failure who have not yet received dialytic therapy may develop a symptom complex progressing from mild sensorial clouding to delirium and coma, with tremor, asterixis, multifocal myoclonus, and seizures. After the institution of adequate maintenance dialysis therapy, patients may continue to be afflicted with more subtle nervous dysfunction, including impaired mentation, generalized weakness, and peripheral neuropathy. These central nervous system disorders are referred to as uremic encephalopathy. The dialytic treatment of end-stage renal disease has itself been associated with the emergence of two distinct, new disorders of the central nervous system; dialysis dysequilibrium and dialysis dementia. The dialysis disequilibrium syndrome consists of headache, nausea, muscle cramps, obtundation, and seizures, and is a consequence of the initiation of dialysis therapy in some patients. Dialysis dementia is a progressive, generally fatal encephalopathy which affects patients on chronic hemodialysis. There are at least three different forms of dialysis encephalopathy: sporadic, epidemic; and that associated with renal disease in children. In addition to the foregoing neurologic diseases which are specifically related to uremia and/or dialysis, a number of other neurologic disorders occur with increased frequency in patients with end-stage renal disease on chronic hemodialysis. These include subdural hematoma, electrolyte disorders, vitamin deficiencies, drug intoxication, hypertensive encephalopathy, and acute trace element intoxication. Renal transplantation is associated with a variety of central nervous system infections, reticulum cell sarcoma, and central pontine myelinosis. The present manuscript will review the clinical, structural, and biochemical components of those neurologic disorders which are peculiar to the uremic state and its treatment with dialysis.

Long-term effects of a new ketoacid-amino acid supplement in patients with chronic renal failure

Nine patients with severe chronic renal failure (mean glomerular filtration rate 4.8 ml/min; mean serum creatinine 11.3 mg/dl) who were previously on a protein-restricted diet were treated with a diet containing an average of 33 kcal/kg and 22.5 g/day of mixed quality protein, supplemented by a combination of amino acids and mixed salts formed between basic amino acids and keto-analogues of essential amino acids. The supplement was designed to minimize or reverse the amino acid abnormalities of chronic renal failure rather than to meet the normal requirements for the essential amino acids; it contained tyrosine, ornithine, and a high proportion of branched-chain ketoacids, but no phenylalanine or tryptophan and very little methionine. Within one month, serum urea nitrogen fell and serum albumin and transferrin rose significantly; serum creatinine fell slightly. Hyperphosphatemia (present in three patients) was corrected. Nitrogen balance, measured in seven of the nine patients, on the average was neutral, as it was in a preceding control period on a 40 to 50 g/day protein diet. Plasma tyrosine and threonine, which were subnormal before therapy, rose to normal or high normal levels. Branched-chain amino acids did not change. During a total of 63 patient-months of therapy, no side effects or toxicity were observed, and serum albumin and transferrin did not change further. It is concluded that this specially designed supplement added to a 20 to 25 g/d protein diet is an acceptable regimen which can improve or maintain protein nutrition in patients with severe chronic renal failure who would otherwise require dialysis.