Urea is quantitatively the most abundant solute excreted by the kidney, and its levels rise higher than those of any other solute when the kidney fails. The chemical formula for urea is CH ₄ N ₂ O, thus incorporating two nitrogen atoms per molecule of urea, so it is relevant to note that clinical laboratory assays traditionally report urea concentrations in terms of the nitrogen content of urea in the sample (blood urea nitrogen or BUN, most often reported in units of mg/dL or mmol/L). Early studies suggested that urea causes only a minor part of acute uremic illness. In the most often cited of these studies, Johnson and colleagues dialyzed patients with kidney failure against bath solutions containing urea. They found that initiation of hemodialysis with or without urea added to the dialysate (at 90 mg/dL) improved uremic symptoms including weakness, fetor, and gastrointestinal upset, even when urea was added. Furthermore, in patients already on dialysis, increasing the BUN level to 140 mg/dL (50 mmol/L) did not cause recurrence of uremic symptoms. Increasing the BUN level >140 mg/dL caused nausea and headaches, and increasing the BUN level >180 mg/dL (64 mmol/L) caused weakness and lethargy; however, symptoms in dialyzed patients whose BUN values were increased to these levels were less severe than symptoms in undialyzed patients with high BUN values. Studies in patients without kidney failure suggest that urea by itself does not cause uremia. Lastly, uremic symptoms have not been observed in patients in whom BUN levels are maintained at approximately 60 mg/dL (21 mmol/L) by high protein intake or increased tubular urea absorption.

It is important to point out that these previous studies focused on the acute symptoms associated with elevated blood urea levels and not their chronic effects. The finding that uremia is not replicated by an isolated elevation of the plasma urea concentration does not mean that urea has no toxic effects. Indeed, urea’s dissociation product cyanate ( Fig. 51.1 ) can trigger a nonenzymatic, posttranslational protein modification known as “carbamylation.” The net result of the reaction is the addition of a “carbamoyl” moiety (-CONH ₂ ) to a nucleophilic functional group (e.g., the primary amino group on the side chain of lysine), which can subsequently change the charge, structure, and function of a broad array of proteins (not surprising when considering the ubiquitous nature of urea in the body). [It must be noted that several authorities state that the reaction in question is more correctly deemed “carbamoylation,” and, in fact, “carbamylation” denotes a different chemical reaction. Yet the biomedical literature has consistently described the addition of the carbamoyl moiety as “carbamylation,” so we will follow this convention.]

Examples of pathophysiologic pathways accelerated by carbamylation include atherosclerosis, vascular calcification, and fibrosis of kidney parenchyma. Carbamylation of low-density lipoprotein (LDL) can increase its atherogenic properties by decreasing its binding to the LDL receptor (preventing its clearance from circulation), a phenomenon seen in animal models ^, and in human CKD studies. Carbamylated LDL has also been shown to trigger inflammatory signaling, ^{, ,} induce endothelial cell apoptosis, ^, and stimulate vascular smooth muscle proliferation. Other studies demonstrate that carbamylated sortilin, elastin, uromodulin, and mitochondrial proteins all result in increased vascular pathology. In animal models, carbamylated proteins at physiologic concentrations can activate glomerular mesangial cells to a profibrogenic phenotype and result in renal tubular cell damage via stimulation of TGF-β, EGF, NF-κB, and endothelin-1. Together, these data highlight how a single uremic retention solute (urea) can mechanistically result in multiple disease processes with significant clinical impact, with particular implications for vascular and myocardial disease. Indeed, numerous epidemiologic studies from different groups, employing different cohorts and even using different biomarkers of global carbamylation burden, have shown that carbamylation load is an independent risk factor for mortality, cardiovascular disease, heart failure, sudden cardiac death, and CKD progression ( Table 51.1 ).

Carbamylated compounds, principally formed from protein exposure to urea, are considered among the top five uremic toxins with the highest toxicity evidence score. Of note, to measure carbamylation, most studies measure a specific carbamylated protein (e.g., carbamylated albumin), protein-bound homocitrulline, or circulating free homocitrulline. L-homocitrulline is the amino acid that is produced when lysine is carbamylated on the terminal amino group of its side chain. Hydrolysis or proteolytic degradation of protein will release protein-bound homocitrulline, and measurement of the protein homocitrulline content or, alternatively, measurement of the homocitrulline-to-lysine ratio is a quantitative indicator of the relative amount of protein carbamylation of biological samples. Likewise, measurement of free homocitrulline in circulation also represents an indicator of total body carbamylation, as it presumably reflects the amount of protein that is carbamylated by urea and degraded during protein proteolytic turnover. Free and protein-bound homocitrulline thus both represent measures of protein carbamylation. ^, If urea is the driver of harmful carbamylation, one could speculate that interventions reducing urea such as dietary modification or intensification of hemodialysis treatments should yield benefits. However, clinical studies that indeed stood to modify urea levels have typically remained null. A likely reason is the complexity in targeting the putative pathway of harm. Individual measurements of blood urea levels, which represent protein intake, renal excretion, and overall nitrogen balance of patients at a single point in time, in clinical practice and research studies, do not appear to accurately reflect carbamylation toxicity. In contrast, carbamylation of proteins represents a time-averaged measure of urea levels over the entire lifespan of that protein. Possibly because of these differences in kinetics, studies in hemodialysis patients have observed that the mortality risks associated with carbamylation were significantly greater than, and independent of, the risks associated with blood urea levels. Furthermore, a case (mortality within 1 year of starting dialysis) versus control (demographically matched survivors of >1 year on dialysis) study showed that while urea, dialysis clearance of urea measured by Kt/V, and protein catabolic rate were similar, the distribution of carbamylated albumin levels was significantly higher in subjects who died compared with survivors ( Fig. 51.2 ) In nondialyzed CKD patients, carbamylated protein levels strongly correlate with blood urea concentrations and both carbamylated albumin and blood urea levels are associated with increased mortality; despite their correlation, however, the mortality risk associated with carbamylated albumin is still significant even after adjusting for patients’ blood urea levels, suggesting that carbamylation is a function of more than just urea levels. In fact, it has been shown that carbamylated albumin levels in hemodialysis patients also correlated negatively with circulating concentrations of free amino acids, suggesting that protein-energy wasting and amino acid deficiencies associated with CKD may also contribute to protein carbamylation and its associated survival risk. ^, Lastly, the association between carbamylation and the sequelae of uremia is further suggested by the fact that when end-stage kidney failure (ESKF) patients were transitioned from standard intermittent hemodialysis to extended duration nocturnal hemodialysis (double the treatment time), it caused dramatic reduction in protein carbamylation in many patients and the reduction in carbamylation correlated with reductions in left ventricular hypertrophy and other favorable outcomes.

Patient and control distributions of carbamylated albumin, Kt/V, blood urea nitrogen (BUN), and normalized protein catabolic rate.

(A) Distribution of carbamylated albumin level measures in the study population by patients (blue; n = 122 who died within 1 year of starting dialysis) and matched controls (green; n = 244 who survived the first year of dialysis); P = 0.01). (B) Distribution of the average Kt/V measure in the study population by patients (blue) and controls (green; P = 0.55). (C) Distribution of the average BUN measure in the study population by patients (blue) and controls (green; P = 0.51). (D) Distribution of normalized protein catabolic rate measured in the study population by patients (blue) and controls (green; P = 0.15). P values represent the frequency distribution comparisons.

Reproduced from Kalim S, Trottier CA, Wenger JB, et al. Longitudinal changes in protein carbamylation and mortality risk after initiation of hemodialysis. Clin J Am Soc Nephrol. 2016;11:1809–1816.

Such discrepancy between the downstream toxic effect of urea (carbamylation) and simple blood urea measurements may explain why prior studies in dialysis and CKD employing interventions that lowered urea did not show the effects that would be expected had carbamylation specifically been targeted. Although the clinical risks associated with protein carbamylation combined with the basic experiments showing pathologic effects of carbamylation on cellular functions provide intriguing evidence, further investigation is needed to prove its causal role in the pathophysiology of uremia.

In comparison with urea, we know much less about most other potential uremic toxins. The d –amino acids exemplify this problem. Aggregate plasma concentrations of d –amino acids increase as kidney function declines. ^, However, the source, clearance, and toxicity of the d –amino acids found in the plasma are not well defined. d –Amino acids can be synthesized by mammalian cells, as well as derived from food, or produced by colonic bacteria. Circulating d –amino acids are filtered by the glomerulus and then, in varying proportion, reabsorbed intact, degraded by d –amino acid oxidase (DAO) or d –aspartic acid oxidase in the proximal straight tubule, or excreted unaltered in the urine. ^, The liver can also clear d –amino acids, but the relative importance of renal and hepatic clearance is unknown. Aggregate d –amino acid concentrations have been found to increase almost in proportion to the serum creatinine level in kidney failure, suggesting that renal clearance predominates. ^{, ,} However, measured concentrations of individual d –amino acids, such as d -serine, increase less than creatinine. ^, This discrepancy remains unexplained. Of interest, d -serine is an endogenous coagonist of the synaptic N -methyl- d -aspartate receptor. It is tempting to speculate that d –amino acids are cleared rapidly from the extracellular fluid (ECF) because they have toxic effects. In addition, it has long been presumed that high levels of d –amino acids could impair protein synthesis or function. d –amino acid accumulation could also interfere with the effects of endogenous d -serine and d -alanine on neuronal function, but no major ill effects of d –amino acid accumulation have been observed in DAO-deficient mice, which have higher d –amino acid levels than humans with impaired kidney function. ^, Exogenous d –amino acids have so far been shown to be toxic only when administered in large quantities. ^,

The kidney clears circulating dipeptides and tripeptides, which may comprise a significant portion of the extracellular amino acid pool. Filtered dipeptides and tripeptides can be broken down by brush-border peptidases and reabsorbed as amino acids or reabsorbed by a brush-border peptide transporter and then hydrolyzed within proximal tubule cells. Peritubular uptake, again followed by hydrolysis to amino acids, makes the renal clearance of many peptides higher than the glomerular filtration rate (GFR). ^, Small peptides are also taken up by other organs and generally do not accumulate in kidney failure. Peptides containing altered amino acids, which are normally cleared by the kidney, may be an exception to this rule.

The kidney plays a proportionally larger role in the clearance of larger peptides. Proteins with a molecular weight of 10 to 20 kDa, such as β ₂ -microglobulin and cystatin C, are normally filtered by the glomerulus and then endocytosed and hydrolyzed in the lysosomes of proximal tubular cells. ^, Their plasma concentrations therefore rise in parallel with the plasma creatinine level as kidney function declines. Indeed, the plasma concentration of cystatin C, which is released at a near-constant rate by nucleated cells, may yield a more reliable estimate of GFR than the concentration of creatinine. The role of the kidney in the removal of peptides with molecular weight between 500 Da and 10 kDa is less well defined. Peptides in this range are also filtered by the glomerulus and then hydrolyzed by brush-border peptidases or endocytosed, depending on their size and structure. Biologically active peptides such as insulin may also be cleared by peritubular uptake. Studies in patients with inherited dysfunction of proximal tubular endocytosis suggest that the normal kidney clears approximately 350 mg/day of peptides with molecular weights of 5 to 10 kDa from the circulation. The relative importance of renal to extrarenal clearance has not been defined for most substances in this size range. The extent to which circulating levels of such peptides are increased in kidney failure is therefore unpredictable. Even less is known about the kidneys’ contribution to the clearance of peptides in the range of 500 Da to 5 kDa.

Although the aggregate peptide levels in kidney failure remain ill defined, we have some knowledge of individual peptides that are retained in uremia. The best known are small proteins or fragments of large proteins for which immunoassays have been developed. Box 51.3 provides a partial list of these substances. Retinol-binding protein, α ₁ -microglobulin, and β-trace protein (also known as prostaglandin D ₂ synthase) are members of the lipocalin superfamily, and future studies may identify elevated levels of other proteins of this group. Lipocalins are a family of proteins that transport small hydrophobic molecules and have an eight-stranded, antiparallel, symmetrical β-barrel fold. This reflects a β sheet that has been rolled into a cylindrical shape. A ligand-binding site is present in this cylinder. Studies using proteomic techniques have yielded a more complete picture of individual peptides that are retained in uremia. These studies have shown that, as expected, uremic plasma contains a vast array of protein fragments that are normally cleared by the kidney. Many of these are derived from fibrinogen and the complement cascade. ^, One study has identified >1000 peptides with molecular weights from 800 Da to 10 kDa in the plasma of patients undergoing dialysis. The toxicity of these peptides and low-molecular-weight proteins, like that of smaller molecular uremic solutes, is largely unknown. It has been widely speculated that retained peptides can cause inappropriate activation of various hormone or cytokine receptors. For example, retained complement protein D (molecular weight, 24 kDa) could contribute to systemic inflammation and accelerated vascular disease observed in patients receiving dialysis. Complement factor D is a serine protease crucial for activating the alternative pathway by cleaving complement factor B (FB) to generate the C3 convertases C3(H ₂ O)Bb and C3bBb. Complement factor D is primarily produced by adipose tissue and circulates in its active form. The overall hypothesis that retained peptides contribute to uremia remains to be proven except for the case of β ₂ -Microglobulin. β ₂ -Microglobulin amyloidosis is the only disease attributable with certainty to a retained peptide, and its prevalence among the dialysis population appears to have declined over the past decades. Plasma levels of peptides such as procalcitonin, troponin I, N-terminal brain natriuretic peptide, and chromogranin A are, however, increasingly used as diagnostic markers. And “false-positive” results may be obtained if criteria are not adjusted for the elevation of their plasma levels, which normally accompanies loss of kidney function.

Among the compounds most frequently considered uremic toxins are guanidines, which, like urea, are derived from arginine (see Fig. 51.1 ). One group of guanidines that accumulate in uremia includes creatinine and its breakdown products. Creatinine is produced by nonenzymatic conversion of creatine, which in turn is made from guanidinoacetic acid. Creatinine itself appears nontoxic, and levels have been increased transiently to >100 mg/dL (8800 μmol/L) in subjects undergoing clearance studies. Instead, interest has been focused on the potential toxicity of various creatinine metabolites, especially creatol and methylguanidine (MG). ^, Creatol is a product resulting from the reaction of creatinine with the hydroxyl radical and is identified as a precursor of MG. Creatol and MG levels normally approximate 1% of creatinine levels. The production of these substances increases as plasma creatinine concentrations rise and may be stimulated by increased levels of intracellular oxidants. ^{, ,} MG is also produced by colonic bacteria, and its production may be increased by augmented dietary intake of protein or creatinine. Guanidinosuccinic acid (GSA) is a nitrogenous metabolite that is found in elevated levels in both serum and urine of patients with advanced CKD. GSA is a guanidine formed not from creatinine but from the urea cycle intermediate argininosuccinate. ^, Rising plasma urea concentrations impede the conversion of argininosuccinate to urea and increase the production of GSA. The production of GSA thus depends on dietary protein intake, as well as kidney function, and may also be stimulated by increased concentrations of intracellular oxidants. ^,

Creatol, MG, and GSA share the interesting property that their plasma concentrations rise out of proportion to urea and creatinine levels as the GFR declines. This is because they are cleared largely by the kidney, and their production rates increase when plasma creatinine and urea concentrations are elevated. ^{, ,} In addition, relative to creatinine, large volumes of distribution, combined with restricted intercompartmental diffusion, limit the removal of creatol, MG, and GSA by intermittent hemodialysis. In patients undergoing conventional hemodialysis, these compounds therefore exhibit high concentrations relative to normal. The finding that they are present in relatively high concentrations does not prove that they are toxic. However, the evidence for toxicity of various guanidines, although incomplete, is stronger than that for most other solutes. Administration of MG aggravates uremic symptoms in dogs, whereas GSA contributes to uremic platelet dysfunction, and a number of guanidines impair neutrophil function. In addition, various guanidines have been shown to accumulate in the brain and cerebrospinal fluid in uremia and may contribute to central nervous system (CNS) dysfunction.

The methylated arginines, asymmetric dimethyl arginine (ADMA) and symmetric dimethyl arginine (SDMA), also accumulate in kidney failure (see Fig. 51.1 ). The metabolism of methylated arginines is quite different from that of other uremic guanidines. ADMA and SDMA are formed by the methylation of arginine residues in nuclear proteins, which are released when these proteins are degraded. Interest has focused largely on ADMA because it inhibits nitric oxide (a potent endothelial-derived vasodilator) synthesis by the endothelial nitric oxide synthase (eNOS). The enzyme eNOS converts l -arginine to nitric oxide, and l -citrulline and is an endothelial-enriched messenger RNA and protein. ADMA inhibits eNOS enzymatic activity. ^, SDMA has been less extensively considered as a toxin, but recent evidence has associated its plasma levels with cardiovascular risk. The urinary clearance of ADMA is similar to that of creatinine, but most plasma ADMA is taken up and degraded in various tissues including the kidney. ^, The presence of nonrenal clearance limits the rise in plasma ADMA concentrations observed as kidney function declines, so ADMA concentrations rise to approximately two times normal early in the course of CKD and then do not increase much further as patients advance to ESKF. ^, Increases in plasma ADMA concentration, although modest in proportion to other uremic solutes, have been associated with an increased risk for cardiovascular disease, especially through impaired bioactive NO function, and death in patients with CKD. It should be noted that differences in assay methods and reported reference ranges for ADMA greatly complicate the interpretation of these studies.

Phenols are compounds that have one or more hydroxyl groups attached to a benzene ring. In discussions of uremia, phenols are usually considered together with other aromatic compounds, such as hippurates, and the term “phenols” is sometimes used loosely to include these other substances. The aromatic compounds normally found in the ECF compartment are, for the most part, derived from the amino acids tyrosine and phenylalanine or from aromatic compounds contained in dietary vegetables. Medications provide an additional source for patients. The compounds in the ECF are mostly metabolites; these are derived from their parent compounds by a combination of methylation, dehydroxylation, oxidation, reduction, and/or conjugation. Many of these reactions take place in colonic bacteria. The final step, which is usually conjugation with sulfate, glucuronic acid, or an amino acid, may take place in the liver, intestinal wall, or, to a lesser extent, kidney. ^, In general, conjugation tends to make the aromatic compounds both less toxic and more polar, thus facilitating their excretion by various organic ion transport systems.

These metabolic processes produce a bewildering array of aromatic compounds that are normally excreted in the urine or feces. The aggregate urinary excretion of aromatics is about 1000 mg/day and varies widely with the diet. The compounds normally excreted by the kidney accumulate in uremia and contribute to elevation of the anion gap, because most aromatic conjugates are negatively charged. The concentration of individual aromatic compounds in uremic patients ranges from barely detectable up to 500 μM. ^, The relatively few compounds that have been studied extensively, including the examples described as follows, are among those found in the highest concentration. Interest in the contribution of phenols and other aromatic compounds to uremic toxicity has been encouraged by reports that uremic symptoms are better correlated with plasma concentration of these compounds than with those of other solutes. ^, Evidence obtained so far on the toxicity of individual aromatic compounds is incomplete.

The most extensively studied aromatic uremic solute is hippurate (see Fig. 51.1 ). Hippuric acid (Gr. hippos, horse, ouron, and urine) is the aromatic waste compound normally excreted in the largest quantity. The concentration of free hippurate rises higher than those of other aromatic solutes in the plasma of patients with kidney failure. Hippurate is the glycine conjugate of benzoate, derived largely from vegetable foods, with only a small amount formed endogenously from the amino acid phenylalanine. ^, Levels of hippuric acid rise with the consumption of phenolic compounds (such as in fruit juice, tea, and wine). Diet therefore determines hippurate production, and hippurate excretion in aboriginal people eating vegetable diets may exceed hippurate excretion in people from industrialized nations by many fold. In persons with normal kidneys, active tubular secretion maintains a plasma hippurate concentration much lower than it would be if hippurate were cleared solely by glomerular filtration. There is a clinical correlate of this property of hippurate. Aminohippuric acid or para -aminohippuric acid (PAH), a derivative of hippuric acid, is a diagnostic agent used in the measurement of renal plasma flow. PAH is completely removed from the blood that passes through the kidneys given that PAH undergoes both glomerular filtration and tubular secretion. Therefore the rate at which the kidneys can clear PAH from the blood reflects total renal plasma flow. By itself, hippurate is not toxic. The plasma hippurate concentration in a normal person can be increased to equal that of a uremic patient, without apparent ill effect. Moreover, increasing the hippurate concentration by benzoate feeding in a patient with kidney failure does not aggravate uremic symptoms. A further clinical correlate addresses the potential toxicity of aromatics, in this case, an exogenous source. Toluene is a widely abused inhaled drug (e.g., glue sniffing) due, in part, to its acute neurologic effects of euphoria and hallucinations. Toluene is metabolized by cytochrome P-450 into benzoic acid and hippuric acid. The hallmarks of acute toluene intoxication are hypokalemic paralysis and metabolic acidosis due to distal renal tubular acidosis. Patients can also have hippurate crystals appear in the urine. They can appear either yellow-brown or clear and frequently take the form of needle-like prisms or plates.

Another extensively studied aromatic compound is p -cresol. In contrast to hippurate, which is derived from aromatic compounds in plants, p -cresol is formed by the action of colonic bacteria on tyrosine and phenylalanine. The portion of amino acids that escapes absorption in the small intestine may be increased in uremic patients, leading to increased production of p -cresol and other bacterial metabolites. ^, Studies have shown that p -cresol circulates largely as p -cresol sulfate with a much lesser portion in the form of p -cresol glucuronide. Reports of unconjugated p -cresol in the plasma of uremic patients now appear to have been the result of inadvertent hydrolysis of p -cresol sulfate during sample processing. ^, p -Cresol sulfate binds avidly to serum albumin, and the effect of different kidney replacement therapies on albumin-bound solutes has often been tested by measuring plasma p -cresol sulfate concentrations. High concentrations of p -cresol sulfate (often measured as p -cresol) have been associated with cardiovascular death in patients undergoing hemodialysis, and p -cresol sulfate has been related to indices of endothelial injury, including endothelial microparticle production. While the results of clinical studies are not uniform, these findings, along with in vitro evidence of p -cresol sulfate toxicity, have increased the focus on p -cresol sulfate as a potential uremic toxin.

Other aromatic uremic solutes have been identified in great numbers but studied less extensively. ^{, ,} Metabolites of tyrosine and phenylalanine, which also accumulate in uremia, include phenlylacetylglutamine, para -hydroxyphenylacetic acid, and 3,4-dihydroxybenzoic acid. Phenylacetylglutamine, like p- cresol sulfate, has been associated with cardiovascular disease and mortality in dialysis patients. The structural relationship of the aromatic amino acid metabolites to neurotransmitters has stimulated interest in their potential role as uremic neurotoxins. So far, 3,4-dihydroxybenzoate has been shown to cause CNS dysfunction in rats, but only at levels higher than those encountered in patients with kidney failure. Increased levels of 4-hydroxyphenylacetate were associated with impaired cognitive function in patients maintained on hemodialysis. The work of identifying toxic aromatic uremic solutes, however, remains daunting. Relatively little progress has been made, and caution is appropriate, particularly when interpreting studies relating levels of specific compounds to poor outcomes. Associations have necessarily been identified only for the few compounds extensively studied. Even if a positive association has been correctly identified, correlations of poor outcomes with abnormal blood levels could be confounded by associations with other unmeasured aromatic substance(s) evidencing related production and/or clearance rates.

Indoles are compounds containing a benzene ring fused to a five-membered nitrogen-containing pyrrole ring (see Fig. 51.1 ). Compounds with an indole ring structure are hydrophobic and, hence, are highly protein bound in the circulation. This makes them difficult to filter at the kidney glomerulus or remove with therapeutic dialysis via diffusion. Thus some indoles are waste products that accumulate when kidney function is impaired and have been considered to be, in part, uremic toxins. Many similarities are encountered when considering the indoles and phenols in uremia. As with phenols, some indoles are derived from plant foods and others are produced endogenously. However, the endogenous indoles are derived mostly from tryptophan, whereas the phenols are derived from phenylalanine and tyrosine. Diverse chemical modifications of l -tryptophan yield a remarkable variety of tryptophan-derived indole structures; more than 600 indoles are known. As with the phenols, minor chemical modifications in various combinations yield a remarkable variety of structures. Those with known physiologic function include the neurotransmitter 5-hydroxytryptamine (serotonin) and melatonin. Other indoles are considered to be waste products and are often conjugated before urinary excretion. These uremic indoles accumulate when kidney function is impaired.

The most extensively studied of the uremic indoles is indoxyl sulfate; this is produced from tryptophan in a manner reminiscent of the production of p -cresol sulfate from tyrosine and phenylalanine. Gut bacteria convert tryptophan to indole, which is then oxidized to indoxyl and conjugated with sulfate in the liver. There is evidence that indoxyl sulfate is toxic in vitro and in animal studies, but early studies of indoxyl sulfate infusion failed to replicate uremic symptoms. ^, Like p -cresol sulfate, indoxyl sulfate is extensively bound to plasma albumin. It has also been suggested that indoxyl sulfate is toxic to renal tubular cells, and the rising indoxyl sulfate levels accelerate the progression of renal disease. Controlled trials to lower indoxyl sulfate levels, however, have not been found to slow progression.

Other indoles that accumulate in uremia include indoleacetic acid, indoleacrylic acid, and 5-hydroxyindoleacetic acid. ^{, ,} As with the phenols, indoles are structurally related to potent neuroactive substances, including serotonin and (famously) d -lysergic acid diethylamide. This structural similarity has stimulated interest in the potential role of indoles as neurotoxins, but few uremic indoles have been administered to normal animals and none have convincingly been shown to alter CNS function at the levels encountered in patients with kidney failure. There is, however, increasing evidence that indoxyl sulfate contributes to vascular disease and particularly to thrombosis. ^, Some effects of indoxyl sulfate have been shown to be mediated by activation of the aryl hydrocarbon receptor, a transcription factor originally identified as upregulating xenobiotic disposal pathways. ^, This work puts our knowledge of the mechanisms by which indoxyl sulfate acts ahead of that of other putative uremic toxins.

Only a minor portion of dietary tryptophan is excreted as indoles. Most is metabolized by the kynurenine pathway, which allows tryptophan to be converted to glutarate and oxidized or, when necessary, used in the synthesis of nicotinamide. Kidney failure causes members of the kynurenine pathway, including l -kynurenine and quinolinic acid, to accumulate in the plasma. ^, Knowledge that these substances play a physiologic role in the modulation of CNS function has stimulated interest in their possible contribution to uremic toxicity. As usual, however, evidence that they are toxic at the levels encountered in patients has not been obtained.

The methylamines monomethylamine (MMA), dimethylamine (DMA), trimethylamine (TMA), and trimethylamine oxide (TMAO) are among the simplest compounds that have been considered to be uremic toxins. Reported serum levels are twofold to threefold higher in patients with ESKF compared with persons with normal or near-normal kidney function.147,148 However, available data and predictions based on their chemistry suggest that the methylamines are poorly removed by dialysis, and early data suggest that they may even be produced in excess in those with uremia.147-149

A large volume of distribution may contribute to poor removal of the methylamines by dialysis. These compounds are bases, with a pK _a ranging from 9 to 11. Thus they exist as positively charged species at a physiologic pH. The lower intracellular pH compared with extracellular pH should lead to their preferential intracellular sequestration, with volumes of distribution exceeding total body water. Indeed, measurements in experimental animals and humans have confirmed these predictions for both DMA and TMA. ^{, ,}

Because they circulate as small organic compounds that are not protein bound, these three amines are likely freely filtered. However, because they exist as organic cations, they also have the potential to be secreted by one or another of the family of organic cation transporters and may also travel through Rh channels. ^, Hence they may achieve clearances that are higher than the GFR. The chemically similar exogenous compound, tetraethyl ammonium, has long been a prototype test solute for organic cation secretion and is cleared at rates up to (and in one study higher than) the renal plasma flow. ^, Although formal renal clearances of DMA and TMA are not available, the total metabolic clearance of DMA and TMA by plasma disappearance of labeled compounds in rats approaches that of renal plasma flow. By contrast, the urinary clearance of MMA in normal subjects is about one-third that of creatinine, indicating no net secretion for this amine.

The biochemical pathways leading to MMA, DMA, and TMA are not well delineated. Both the host’s mammalian tissues and resident gut flora have been thought to contribute to the net appearance of these amines. However, plasma MMA and DMA concentrations were not different among patients with ESKF with and without colons. The dietary precursors for MMA, DMA, and TMA include choline, carnitine, betaine, and dimethylglycine. Production of these compounds may actually be increased with kidney failure, perhaps caused by overgrowth of intestinal bacteria. ^{, ,} Thus production of aliphatic amines may be increased in the face of impaired renal removal.

Incomplete data also implicate the amines as toxic. Reported effects of MMA include a variety of neural toxicities, hemolysis, and inhibition of lysosomal function. MMA has also been shown to be a potent anorectic agent when administered into the cerebrospinal fluid in mice. Despite toxicities in cells and animals, however, MMA and DMA were not associated with all-cause mortality in a cohort of patients with ESKF. Although of utmost importance, mortality is obviously a blunt metric of uremic toxicity. Other signs of toxicity of MMA and DMA should be explored.

The uremic fetor or fishy breath noted in uremic patients is attributable to TMA. Although the malodor may be of no major consequence in itself, the potentially important and well-described diminutions in taste (dysgeusia) and smell (dysosmia) among patients with kidney failure (which may contribute to poor nutritional status) may also be related to the amines. Plasma MMA concentrations were not related to olfactory defects in ESKF.

TMA can be both a precursor to TMAO and, as noted, a product of dietary TMAO. TMAO is cleared by secretion in the normal kidney; its levels rise as the GFR falls and, like those of many other secreted solutes, are increased out of proportion to urea levels in dialysis patients. TMAO was initially identified as a risk factor for cardiovascular disease in persons with normal renal function. ^, TMAO is a gut microbial metabolite that has been shown to be directly atherogenic. TMAO arises from gut microbiota metabolism following ingestion of diets rich in phosphatidylcholine (lecithin), the major dietary source of choline, and carnitine, an abundant nutrient in red meat. Studies employing microbial transplantation into recipients have confirmed a direct causal role for gut microbes in transmitting atherosclerosis susceptibility and overall TMAO production. Subsequent studies showed that accumulation of TMAO is also associated with atherosclerosis burden and cardiovascular events in patients with renal insufficiency and in patients maintained on dialysis. ^,

A wide variety of other compounds accumulate in kidney failure. One group is the polyols, of which the most extensively studied is myoinositol (see Fig. 51.1 ). ^, Myoinositol is different from most other uremic solutes in that it is normally oxidized by the kidney. Its accumulation in uremia therefore reflects impaired degradation and not impaired excretion. Evidence that myoinositol causes nerve damage, although stronger than most of the evidence for the toxicity of uremic solutes, is far from conclusive.

The purine metabolite uric acid is the only known organic substance with a plasma level actively regulated by variation of its renal excretion. With advanced CKD, the capacity of the kidney to increase the fractional excretion of uric acid is exceeded and uric acid levels increase, along with those of its precursor molecules, xanthine, and hypoxanthine. Observational studies have shown that high uric acid levels are associated with increased survival in dialysis patients, perhaps because the plasma uric acid is a marker of nutritional status. Other nucleic acid metabolites excreted by the kidney are produced in much lesser quantities. Many are derived from the modified nucleosides contained in transfer RNAs. They appear to be cleared largely by filtration and to accumulate in the plasma as the GFR falls. Pseudouridine is an isomer of the nucleoside uridine in which the uracil is attached via carbon–carbon instead of a nitrogen–carbon glycosidic bond. It has been suggested that pseudouridine, which is the most abundant of these variant nucleic acid metabolites, contributes to insulin resistance and altered CNS development but, as usual, the demonstration of its toxicity is not conclusive. ^,

Oxalate is also excreted by the kidney and accumulates in kidney failure. Oxalate is derived from catabolism of endogenous substances including vitamin C, as well as from plant foods. ^, The potential for oxalate deposition in tissues may limit our ability to maintain normal plasma vitamin C concentrations in patients receiving dialysis. ^, Additional substances excreted by the kidney that accumulate in kidney failure include various pteridines, dicarboxylic acids, isoflavins, and furancarboxylic acids including 3-carboxy-4-methyl-5-prophy-2-furanpropanoic acid. ^{, ,}

Indeed, studies continue to add new solutes to the list and the number reported increases rapidly with upgrades in metabolite screening methods. This list includes known and now even more unknown compounds that are at present identified only by their molecular mass using mass spectrometry and do not yet appear in standard databases of human metabolites. The possibility of toxicity is invariably considered when new solutes are identified, but experiments to test the solute toxicity are rarely performed. These challenges are common to metabolomic and proteomic studies in general but may be exacerbated in nephrology research. Within the realm of metabolomic kidney disease research, a well-recognized challenge is that a substantial portion of all detectable metabolites track with estimated glomerular filtration rate (eGFR), creating problems of confounding when making associations involving eGFR. While adjusting for eGFR’s influence on metabolites is considered necessary, analytes that strongly correlate with eGFR may still be mechanistically important and worthy of additional study. The sheer number of metabolites that can be assessed using modern technologies increases the risk of false associations due to chance alone from multiple testing. New biostatistical approaches and analytical techniques aim to reduce these risks, but replication remains an important burden for putative uremic toxins. ^, External validation of findings through replication studies with large and sufficiently powered sample sizes thus remains a critical component of metabolomic research, but meeting such methodologic demands in the setting of specific populations or phenotypes of interest where data can be limited is difficult. Translating findings for candidate markers from high throughput screening studies is even more difficult. These dilemmas highlight why progress in this field remains incremental.

Resting energy expenditure has been reported as increased, decreased, and normal in patients with kidney failure. The choice of control populations and other methodologic issues, such as corrections for altered body composition, have probably contributed to this uncertainty. The effect of dialysis treatment, which may transiently increase energy expenditure, must also be considered. Uremia, apart from dialysis, likely reduces resting energy expenditure. ^, Lower energy expenditure accords with observations of lower body temperatures in those with uremia, although additional factors may be at play in thermoregulation. However, dialysis treatment may itself speed metabolism and increase energy expenditure. Effects of inflammation in patients with ESKF add further complexity to the assessment of energy requirements. Using anthropometric techniques in patients with ESKF on chronic hemodialysis who were maintained for 3 months on carefully constructed diets, the energy requirements were found not too different from those reported for similarly sedentary normal subjects. The energy requirements in ESKF were, however, more variable than in normal subjects.

Knowledge of the physiologic control mechanisms for appetite, fat metabolism, and energy expenditure is potentially relevant to uremic energy metabolism. Studies have focused on the signaling molecules ghrelin, produced by the stomach, and leptin, produced by adipose tissue along with other adipose tissue–derived hormones (adipokines). Levels of these small proteins tend to rise in kidney failure because of reduced clearance by the kidney and possibly because of increased production. ^,

Insulin resistance is the most conspicuous derangement in uremic carbohydrate metabolism. The defect is clearly present in ESKF, but in cross-sectional studies, impairment can be detected when the GFR falls below 50 mL/min/1.73 m, with a graded relation to GFR. ^, There are probably several causes of this phenomenon. Surprisingly, some obvious possibilities do not seem to contribute. Insulin binds normally to its receptor in uremia, and receptor density is unchanged. ^, Moreover, excess levels of glucagon or fatty acids do not account for the disorder. As is the case with overall energy metabolism, interest has recently focused on the contribution of adipokines including leptin, resistin, and adiponectin to insulin resistance. Adipose tissue has also been identified as a source of inflammatory cytokines that impair insulin action in various experimental systems and circulate at increased levels in many patients with advanced renal insufficiency. However, correlations among levels of individual substances with measures of insulin resistance are poor, and the extent to which adipose tissue products contribute to uremic insulin resistance remains uncertain. ^{, ,}

Because dialysis, transplantation, and low protein diets tend to restore insulin responsiveness, it has also been suggested that unidentified nitrogenous product(s) mediate insulin resistance. Acidosis has been shown to provoke insulin resistance and accumulation of acid, as well as nitrogenous wastes, may contribute to insulin resistance in uremia. 11-β-Hydroxysteroid dehydrogenase type 1 provokes insulin resistance by regenerating glucocorticoids. In experimental uremia, liver and fat tissues express increased activity of this enzyme, as well as insulin resistance. The insulin resistance associated with 11-β-hydroxysteroid dehydrogenase can be mitigated with an inhibitor of the enzyme, all suggesting a role for this steroid pathway in insulin resistance. The cause for increased enzymatic activity is unknown. Resistin is a protein capable of inducing insulin resistance, and its levels are high when kidney function is impaired. However, the plasma concentrations of resistin are not associated with insulin resistance if the GFR is taken into account. Retinol-binding protein is also associated with insulin resistance (and also rises in ESKF) but is not related to markers of glucose control. Finally, physical inactivity and deconditioning may contribute to insulin resistance. Exercise programs have been shown to mitigate insulin resistance but must be relatively protracted and intensive to be effective. ^, Ongoing studies have shown exercise programs in CKD and ESKF to yield various benefits including improved physical strength and improved cardiovascular and kidney health measures. However, the mechanisms for these benefits independent of those found in the general population remain speculative and the impact specifically on uremic toxins as mediators is still unanswered. ^, Improvements in insulin sensitivity and reductions in systemic inflammation and oxidative stress are leading mechanistic contenders.

Insulin resistance may have several adverse effects. Most importantly, it has been recognized as a risk factor for cardiovascular disease. The connections between insulin resistance and vascular disease are not clear. A tendency to hyperglycemia is one presumably toxic effect. Some investigators have suggested that the sodium-retentive effect of insulin on the kidney remains intact, whereas other tissues become insulin resistant in uremia. Increased plasma insulin concentrations could thus contribute to arterial hypertension in patients with impaired kidney function. ^, Outside of vascular disease, the loss of insulin’s anabolic action may contribute to uremic muscle wasting. ^,

Even though insulin resistance is the rule in uremia, hypoglycemia can be a significant effect of renal insufficiency. Hypoglycemia is likely to occur, despite insulin resistance, for two main reasons. First, the kidney is a major site of insulin catabolism. Patients with diabetes mellitus treated with insulin, or insulin secretagogues (e.g., sulfonylureas), frequently become hypoglycemic if doses are not adjusted downward as the GFR declines. Second, the kidney is a major site of gluconeogenesis. The liver produces the bulk of glucose in postabsorptive and starvation states, but even in these situations, the kidney produces some glucose. With prolonged fasting, the kidney is responsible for approximately half of the total glucose production. ^, Thus advanced CKD may predispose to hypoglycemia by prolonging insulin action and reducing gluconeogenesis. These effects may become particularly apparent when other hypoglycemic factors, such as ethanol ingestion or liver disease, are also evident.

Nephrotic syndrome and even lower-grade proteinuria are regularly associated with hyperlipidemia. However, lipid abnormalities are modest when kidney function is impaired without significant proteinuria. Indeed, total cholesterol falls on average as the GFR drops below about 30 mL/min/1.73 m ² . Metabolite profiling in plasma of patients with ESKF using liquid chromatography and tandem mass spectrometry has revealed that lipid products deviate from normal levels far less than polar compounds. However, lower-molecular-weight triacylglycerols were generally decreased, and an increase in intermediate-weight triacylglycerols was observed. The causes and consequences of these changes in lipids are uncertain. The decline in total cholesterol is taken to reflect, at least in part, progressive reduction in food intake. Numerous hypotheses have been advanced to account for the finding that atherosclerosis is accelerated, but lipid levels are not markedly elevated in renal failure. Even though total lipid levels are not significantly elevated, there may be an increase in oxidized forms due to oxidant stress and reduced lipoprotein clearance rates. The high prevalence of cardiovascular disease in patients with ESKF has prompted several randomized clinical trials of statin agents. The largest trials performed to date have identified no clear benefit of statins in patients on dialysis, in sharp contrast to persons with normal or near-normal kidney function and, in one study, to patients with advanced, non–dialysis-requiring CKD. Furthermore, statins seem to provide no benefit in slowing the progression of kidney disease. Nevertheless, given the high cardiovascular risks in patients with nondialysis CKD, most recommend statin therapy as secondary prevention for nondialysis CKD patients with atherosclerotic cardiovascular disease (similar to those without CKD) and even for primary prevention in those older than the age of 50 or at high risk. For ESKF patients on dialysis, the evidence suggests not routinely initiating statin therapy but maintaining it if already receiving it at the time of dialysis initiation.

The normal kidney participates in the metabolism of several amino acids. ^, For example, the kidney converts citrulline to arginine, one of the necessary steps in the urea cycle. Loss of this function likely contributes to the increasing ratio of citrulline to arginine as the GFR declines below 50 mL/min/1.73 m ² . ^, Similarly, reduced renal production of serine from glycine probably underlies the rise in the plasma glycine-to-serine ratio. Increased concentrations of the sulfur-containing amino acids—cysteine, taurine, and homocysteine—are especially intriguing. Cysteine and homocysteine accumulate in the oxidized form, consistent with the concept that uremia is a state of oxidant stress, and homocysteine levels have been associated with the progression of cardiovascular disease. ^{, ,} Administration of folate to lower homocysteine levels neither restores the plasma redox state toward normal nor reduces the frequency of cardiovascular events. ^, The mechanism responsible for the accumulation of oxidized cystine and homocysteine remains unclear, but this change can be detected as the GFR drops below roughly half of normal and becomes more extreme as ESKF approaches.

Tissue protein loss reflected by muscle wasting is a major concern in patients with kidney failure. Factors that predispose to protein wasting include reduced appetite, along with insulin resistance and altered amino acid metabolism, as described earlier. Dialysis also results in some protein loss, with amino acids lost in the dialysate and plasma proteins and amino acids lost in the peritoneal dialysate. In the absence of other complications, the effect of uremia at the levels now seen clinically on protein metabolism is usually modest, at least over the short term. Patients with advanced CKD can maintain nitrogen balance on low protein diets as long as acidosis and inflammation are minimal. Several factors may combine with defects in insulin resistance and altered amino acid and adipose metabolism to produce muscle wasting. The best studied of these is acidosis, which has been shown to stimulate the ubiquitin–proteasome pathway of intracellular protein degradation. Activation of caspase-3 seems to be an important step in proteolysis, which is followed by disposal of protein cleavage fragments through the proteasome. In addition to these effects, acidosis contributes to insulin resistance and thereby attenuates protein anabolic actions of insulin. Base supplements can mitigate the catabolic effects of acidosis, but a long-term study establishing the value of normalizing bicarbonate levels in patients with impaired kidney function is lacking. ^, Newer guidelines no longer recommend a target serum bicarbonate level or a graded treatment recommendation. Instead, the guidelines simply suggest treating to keep serum bicarbonate levels >18 mEq/L.

Inflammation may be an even more important contributor to protein wasting than acidosis in patients with kidney failure. Muscle loss in patients with ESKF has been linked to an inflammatory state, characterized by increased serum concentrations of C-reactive protein and various cytokines. How these inflammatory mediators trigger net protein degradation in muscle and other tissues remains to be better elucidated, although their presence is regularly accompanied by muscle loss and theoretical pathways are proposed. High levels of inflammatory mediators are also associated with lower serum albumin concentrations, which have been attributed largely to reduced hepatic production of albumin. Both muscle loss and reduced albumin concentration predict early death. The exact cause(s) of inflammation remains elusive. In some cases, inflammation can be ascribed to known episodes of infection or other intercurrent illness, and in other cases, periodontal disease and gut dysbiosis can be invoked. In many cases, however, no cause can be identified. Occult inflammatory stimuli in these cases may include subclinical infection at hemodialysis catheter or arteriovenous graft sites, exposure to dialysate and various synthetic materials, and accelerated vascular injury that is common in ESKF. Oxidant stress has also often been invoked. Attempts to reduce inflammation with free radical scavengers and other agents, however, have been largely unsuccessful. ^{, ,} An interesting possibility is that organic solutes retained in kidney failure, although they do not regularly trigger inflammation over the short term, cause the late appearance of inflammation in a subset of patients with ESKF. Some evidence has suggested that accumulation of proteins modified by glycation and oxidation can trigger a self-perpetuating inflammatory loop in these cases. ^,

Another factor contributing to muscle wasting is inactivity. Johansen and associates have shown that self-reported physical activity in patients starting dialysis is at, or below, the first percentile for population reference range, and lower levels of physical activity are strongly associated with mortality. In these patients, inactivity may be caused by fatigue and loss of energy, which are invariable although difficult-to-measure features of uremia, and by depression and other comorbid illnesses, which are common in the ESKF population. Exercise and other physical modality interventions in CKD have shown mixed but promising results through a variety of proposed mechanisms.