CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985 5

CLIN. CHEM. 31/1, 5-13 (1985)

Uremic Toxins, and Their Effect on IntermediaryMetabolismMichael R. Wills

In the late stages of chronic renal damage the functionalmass of the kidney is reduced and there is progression to.renal insufficiency, usually called uremia, in which all aspectsof renal function are affected. The complexity of the biochem-ical aspects of the syndrome of uremia is a manifestation ofthe wide variety and nature of the individual disorders thatcontribute to the pathogenesis of the final clinical syndrome.One major feature is the retention of metabolic end productsand their effects, as toxins, on intermediary metabolism. Theretained end products, working singly or in combination,probably affect metabolic pathways by some modification ofenzymic reactions. They act at the cell membrane level.Although “middle molecules” have been incriminated asuremic toxins, recent attention has also focused on traceelements-especially aluminum, which has been implicatedin the pathogenesis of two major disorders, osteomalacicdialysis osteodystrophy and dialysis encephalopathy.

AddItIonal Keyphrases: renal function “middle molecules”

aluminum dialysis patients osteodystrophy . enceph-

alopathy trace elements . kidney disease hor-

mones chronic renal failure uremia lipoproteins

Besides being the dominant organs in the homeostaticcontrol of the constitution of the internal environment, thekidneys are also major endocrine organs, secreting varioushormones that act on the respective target tissues. Thekidneys also indirectly influence other endocrine systems,by affecting the clearance and degradation of other hor-mones. In reviewing the biochemical aspects of chronic renalfailure-with specific reference to toxins-one must have aclear concept of the role that the kidneys play as homeostat-ic organs. However, the complexity of the syndrome ofuremia is due to the role played by the kidneys, not only ashomeostatic but also as endocrine organs.

As homeostatic organs the kidneys are responsible for thevolume and the ionic and molecular composition of theinternal environment. The concept of an internal fluidenvironment in which the cells of the body live and functionwas introduced by Claude Bernard in 1878(1). The vital roleof the kidneys in maintaining the constitution of thatenvironment was subsequently recognized by Peters in 1935(2), who said: “the kidneys appear to serve as the ultimateguardians of the constitution of the internal environment.”

Excretion of metabolic end products, a major homeostaticfunction of the kidneys, involves both glomerular ultrafil-tration and renal tubular secretion. During the development

Departments of Pathology and Internal Medicine, University ofVirginia, Charlottesville, VA 22908.

Received August 8, 1984; accepted October 1, 1984.

of the nephron unit the tubular secretory systems evolved toexcrete, into an aqueous environment, waste products ofhigh relative molecular mass that could not escape from thebody by the process of simple diffusion (3). Maintainingoptimal conditions in the fluid environment of the cells isvitally important; the failure of the kidneys to perform thistask is manifested in the clinical syndrome called uremia.

The term “uremia” was introduced in 1840 by Piorry andl’H#{233}ritier(4). In their treatise on alterations in the bloodthey proposed use of the general suffix “-emia” to denote theblood compartment and qualifying it with a specific prefix todenote the presence of an abnormality in that compartment.Thus in their terminology “uremia” literally means “urinein the blood,” which reflects their view that the toxicmanifestations of renal failure were a form of poisoning ofthe blood, a consequence of reabsorption of urine. Today,uremia is used clinically to describe a complex syndromethat has many interrelated features.

In the specific context of toxins, the word “uremia” is usedto describe the state associated with the retention of nitroge-nous metabolic end products and is characterized by anincreased concentration of urea in the blood. Despite aconsiderable amount of time and effort devoted to a searchfor the uremic toxin, no one individual compound has, so far,been incriminated. It may well be that no one uremic toxinwill ever be identified as the toxin. The clinical syndrome ofuremia should be recognized as a composite problem, involv-ing all of the body’s systems and reflecting biochemicalalterations in all aspects of the constitution of the internalenvironment. In this view the alterations that produceuremia would reflect not only accumulation of metabolicend products but also associated changes in water, electro-lyte, and acid-base homeostasis; disturbances in endocrineand nutritional status; and associated abnormalities inmetabolism of fat, carbohydrate, and protein.

High concentrations in serum of the end products ofprotein catabolism are generally considered to be a majorfeature of the uremic state; some of the intermediate break-down products are also believed to accumulate and play arole in the development of toxicity. Organic substancesknown or reported to accumulate in uremic blood include:urea, creatinine, guanidines and related compounds, uricacid, creatine, certain amino acids, polypeptides, poly-amines, cyanate, indican (indoles), hippuric acid, phenolsand conjugates of phenol, phenolic and indolic acids andtheir conjugates, organic acids of the tricarboxylic acid cycle,aliphatic amines, guanidine bases, pseudouridine, acetoinand 2,3-butylene glycol, j3-hydroxybutyrate, glucuronic acid,carnitine, myoinositol, “middle molecules,” sulfates, andphosphates (5). Given the improvement in clinical symp-toms in uremic patients on institution of adequate therapywith dialysis, and the associated decrease in the concentra-

Creatinine

6 CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985

tions of these retained organic metabolic end products, thereis no doubt that the retained metabolites do play an impor-tant role in the pathogenesis of the clinical syndrome ofuremia. The retained organic metabolic products probablyplay a major role as toxins in the pathogenesis of uremia,either working singly or in combination.

In this review I will deal with those compounds that havebeen of major interest in the past as toxins or are currentlyof interest.

Urea

Among the specific compounds incriminated and investi-gated as uremic toxins, urea has received considerableattention. The increase in the urea concentration in blood isperhaps the most striking abnormality of the body fluids inrenal failure, although not the most important functionally.

Urea is formed only in the liver and may be regarded asthe end product of protein catabolism, whether the proteinoriginates from the diet or the tissues. Patients with acuterenal failure show a relatively good correlation between theseverity of the illness and the concentration of blood ureanitrogen. In chronic renal failure, however, the concentra-tion of creatinine in serum seems to be a better index of theseverity of the degree of failure, particularly when thepatient is on a low protein diet.

The role of urea in the pathogenesis of the clinicalsyndrome of uremia has been controversial. In 1827 an earlyclinical chemist, Dr. John Bostock, who did the chemicalanalyses for Dr. Richard Bright, reported that in patientswith chronic renal failure the serum “was found to consist inpart of an animal matter possessing peculiar propertieswhich seemed to approach those of urea” (6). AlthoughBright recognized the markedly increased concentrations ofurea in patients with chronic renal failure, it is interestingthat he considered that it “may be but in part a cause ofgeneral derangement of the system” (7). As has been subse-quently demonstrated, the administration of urea to normalsubjects in amounts sufficient to raise their blood concentra-tions to values similar to those in patients with chronicrenal failure is associated only with thirst and polyuria, andnone of the other clinical manifestations of uremia. Inpatients with chronic renal failure, therefore, most of theavailable evidence supports the proposal that an increasedurea concentration in blood does not itself have a major toxiceffect. In support of this proposal, Merrill et al. (8) hemodia-lyzed chronic uremic patients against solutions with highurea concentrations and noted an excellent clinical responseeven though the blood urea concentration was unaltered.

At one time, a high urea concentration in the blood wasimplicated in the pathogenesis of the “dialysis disequilibri-um syndrome.” This syndrome is characterized by thedevelopment of headache, confusion, muscle twitching, pro-gressing occasionally to convulsions, coma, and (rarely)death, either during or towards the end of an otherwisebiochemically successful dialysis. In the mechanism pro-posed by Kennedy et al. (9), an osmotic gradient wascreated, during dialysis, between the brain tissues, thecerebrospinal fluid, and extracellular fluid compartments,such that water would shift into the first of these; this eventwould be associated with an increase in intra-cranial pres-sure, which would account for the features of the syndrome.Since then, the dialysis disequilibrium syndrome has beenfound to relate to changes in sodium concentration duringdialysis (10) and can be prevented by the use of a highsodium concentration in the dialyzing fluid (11).

Although urea may not itself be a major toxin in patientswith chronic renal failure, it is one of the retained metabo-lites known to act as enzyme inhibitors (see below).

An increase in creatimne concentration in the plasma isanother diagnostic feature of chronic renal failure; and it isroughly correlated with the degree of failure. A formula hasbeen proposed that allows for age and weight and predictsthe endogenous creatinine clearance ratio from the serumcreatimne concentration (12). Like urea, creatinine maynot, per so, play a role as a uremic toxin. Its retention is ofimportance in the genesis of other toxic metabolites, aconsequence of an alteration in its normal metabolic path-way. Creatinine metabolism and excretion is altered inpatients with chronic renal failure. In normal individuals,the amount of creatimne excreted is chiefly influenced bylean body mass and diet. Patients with chronic renal failureexcrete less creatiine in their urine than would normallybe expected. The serum creatiine concentration increasescorrespondingly in chronic renal failure, although an in-creasing proportion reportedly is cleared by an extrarenalmechanism, and this accounts for the reduction in urineexcretion (13). The proposed extrarenal clearance mecha-nism involves two potential pathways: recycling of creati-nine to creatine, irreversible degradation to products otherthan creatine, or both. Any alteration in creatiine metabo-lism during chronic renal failure, with an extrarenal clear-ance route, is of particular interest in that it represents anadaptative change in a metabolic pathway in response to thelong-term decrease in functional renal mass.

Uric and Oxalic Acids

Urate retention is one of the recognized biochemicalfeatures of chronic renal failure. The uric acid in serumincreases in uremia, but correlates poorly with the incre-ment in creatinine concentration (14). As overall renalfunction deteriorates, the excretion and clearance of uricacid by the functional renal remnant markedly increases(14). An increase in the tubular secretion of uric acid andincomplete reabsorption of the ifitered fraction-which isnormally almost completely reabsorbed-would account forthe lack of correlation with the increase in creatimneconcentration. Again: the changes in the tubular secretionand reabsorption of uric acid represent adaptations to thedecrease in functional renal mass. These adaptative changesin the remaining nephrons of the chronically diseasedkidney with respect to uric acid transport may be promptedby a uricosuric factor in uremic serum (15, 16).

In addition to an adaptative change in the renal tubularhandling of uric acid there is also evidence that uremicpatients develop an alternative route for either the metabo-lism or the clearance of uric acid (17). The alternative routeinvolves the extrarenal elimination by the process of uricol-ysis (18), which takes place entirely in the intestinal tract, iscatalyzed by bacterial enzymes, and appears to becomeincreasingly important as the plasma uric acid concentra-tion increases. Here again, this induction of alternativeroutes for the handling of uric acid, either by variations inrenal tubular handling or the process of uricolysis, is ofinterest because of what it tells us of biochemical adapta-tions in response to decreased functional renal mass.

Retention of oxalic acid is a recognized feature in patientswith chronic renal failure (19), and it reportedly is associat-ed with the deposition of oxalate crystals in the myocardiumand renal tissues (20-22). Crystal deposition in these tissuescould be of importance in the genesis of some of the clinicalfeatures of the syndrome of uremia. It is also of interest inthe overall pathogenesis of the clinical syndrome of uremiathat oxalic acid is among the retained metabolites that are

CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985 7

known to act as enzyme inhibitors-in this particular

instance, lactate dehydrogenase (23).

Myoinositol

Myoinositol has a specific effect on nerve conduction.Some patients with chronic renal failure have a markedincrease in serum myoinositol concentration and also fail toclear it at the normal rate after an oral load. Changes inmyoinositol concentration and clearance rate could resultfrom an impairment of the renal catabolism (24). Myoinosi-tol is a precursor and constituent of a class of phospholipids,the phosphoinositides, whose metabolism has been linkedwith the functional activity of nerve. Rats with hypermyoin-ositolemia from oral loading reportedly have a significantdecrease in the velocity of sciatic motor-nerve conduction(24). On the basis of these findings it was suggested that theabnormally high myoinositol concentrations found in ure-mics may be a factor in the development and progression ofa polyneuropathy. Most of the evidence, however, wouldnow support the view that, although hypermyoinositolemiamay depress nerve conduction velocity (25) and its concen-trations in serum are markedly increased in patients withchronic renal failure, there is no indication that myoinositolis the neurotoxin of uremia (26). This view was based on thefact that there was no correlation between nerve-conductionvelocities and either serum or cerebrospinal fluid myoinosi-tel concentrations: there was also no correlation betweenelectroencephalographic changes and myoinositol concen-trations in either of these two compartments (26).

Guanidines and Related Compounds

Guanidines and related compounds have long been impli-cated as uremic toxins. Experimental animals, chronicallyintoxicated with methylguanidine, reportedly lose weight ata rate that suggests that it exerts a catabolic action (27). Inthese animals the late stages of intoxication with methyl-guanidine were associated with disturbances in the gastro-intestinal tract, the cardiovascular system, the lungs, andthe central and peripheral nervous systems. The potentialimportance of methylguanidine as a toxin in uremia is thatit is a constituent of creatine (methylguamdoacetic acid) andits derivative, creatinine.

Although the concentrations of guanidine, methylguani-dine, and 1,1-dimethylguanidine in serum may not bemarkedly increased in uremic patients, the urinary excre-tion of methylguamdine is significantly increased (28, 29).There is also evidence that the rate at which methylguani-dine is produced metabolically is increased in chronic renalfailure and that this is associated with increased concentra-tions in the tissues (30). Thus it was proposed that methyl-guanidine retention plays a role in the pathogenesis of theuremic syndrome (31). If methylguanidine is a uremic toxin,it presumably exerts its toxic actions in association with apreferential distribution in the intracellular compartment.

Although some experimental studies suggest that methyl-guanidine is a uremic toxin, other evidence supports theproposal that its toxic role remains to be defined. Thisproposal is based on observations that high concentrationsof methylguanidine caused no significant inhibition of oxy-gen uptake in vitro in tissue-respiration studies of slices ofrat cerebral cortex, kidney, and liver (32). The concentra-tions of methylguanidine used in those studies were up tofour times those in serum of patients with chronic renalfailure. In studies of glucose uptake and utilization bypreparations of rat diaphragm, in vitro methylguanidinehad a small enhancing effect (33) rather than a toxicinhibitory action. Guamdine and guanidinoacetic acid alsohave been studied individually and shown to have small

enhancing effects on glucose uptake and utilization (33).Guanidinoacetic acid is a precursor of creatinine, and itsconcentration increases during chronic renal failure as adirect consequence of the decreased clearance that can beeffected by whatever functional renal tissue is left. Anincrease in guanidinoacetic acid potentially leads to thetransfer of the amidine group of arginine to aspartateinstead of glycine, with the formation of guanidinosuccinicacid and ornithine.

The urinary excretion of guanidinosuccinic acid increasesin chronic renal failure (34), and the increase has beenattributed to the development of an alternative pathway forthe detoxication of ammonia and urea synthesis, whichinvolves guanidinosuccinic acid. The mechanism for thisalternative pathway would involve repression of normalenzyme activity and either the activation of a dormantenzyme or the appearance of a new enzyme. The alternativemetabolic pathway for urea synthesis in chronic renalfailure (35) is consistent with an increase in methylguani-dine concentration. In subsequent studies Stein et al. (36)reported increased concentrations of guanidinosuccinic acidin serum and cerebrospinal fluid and confirmed its highurinary excretion in patients with chronic renal failure.These observations have been confirmed by other workers,who reported that not only does guanidinosuccimc acidaccumulate in chronic renal failure, there was also evidenceof increased production in these patients as compared withnormal individuals (37). However, the role of guanidinosuc-ciic acid as a toxin in uremia remains to be clarified;animal experiments gave (38) no evidence of guanidiniosuc-cinic acid being a uremic toxin.

Dimethylamine

In any review of the metabolites retained in the serum ofpatients with chronic renal failure it should be recognizedthat not all these compounds are derived either from auremia-induced derangement of the usual metabolic path-way or from endogenous tissue sources. An example isdiethylamine, which increases in serum in uremia. Itsconcentration in duodenal contents is also increased ascompared either with normal healthy subjects or with othertypes of patients (39). This increase is attributable to theuremic state per so, a consequence of uremia-induced alter-ations in the bacterial flora of the gastrointestinal tract. Inuremic patients, part of the choline is transformed bybacteria in the gut to trimethylamine, which is reabsorbedand then either oxidized by trimethylanune dehydrogenase(EC 1.5.99.7) or demethylated to dimethylamine (DMA) inthe liver (40). Dimethylamine enters the circulation and isexcreted in bile and urine. The increased concentrations ofdimethylamine and trimethylamine in the breath of ure-mics are correlated with their classic “fishy odor” (41).Possibly, alterations in the gastrointestinal flora as a directconsequence of the uremic state are of importance in theformation of various other metabolites, which then areabsorbed into the extracellular fluid compartment, resultingin an increased circulating concentration of them, withpotential toxic and nutritional sequelae.

Parathyrin (Parathyroid Hormone) as a Uremic Toxin

Normally functioning kidneys play a major role in theenzymic degradation and clearance of hormones from thecirculation, and because of this the kidneys may be regardedas indirect endocrine organs. Hormone status is disturbed inpatients with chronic renal failure, and the associatedconsequences may be regarded as features of the clinicalsyndrome of uremia. An established biochemical feature ofuremia is a markedly increased concentration of carboxy-

8 CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985

terminal immunoreactive parathyroid hormone (i-PTH),which not only reflects a disturbance in the rate of secretionof parathyrin but also a decrease, or even a failure, in therenal excretion of the peptide fractions resulting fromenzylnic degradation. This accumulation of i-PTH degrada-tion products, together with the known actions of thehormone, has led to the proposal that it may be a majoruremic toxin.

In 1977, Massry (42) proposed that many of the manifes-tations of the uremic state could be accounted for by anexcess of circulating parathyrin and that this polypeptidecould be an important uremic toxin. The manifestations ofuremia that were attributed to parathyrin included disor-ders of the central nervous system, soft-tissue calcification,soft-tissue necrosis, bone disease, pruritis, hyperlipidemia,anemia, and sexual dysfunction (43). The major potentialrole of parathyrin as a uremic toxin is as a neurotoxin, and aconsiderable amount of evidence supports this view. Avramet al. (44) reported a significantly decreased motor-nerveconduction velocity in uremic patients with above-normali-P’FH concentrations in their serum as compared with age-matched uremic patients with normal or only slightlyabove-normal i-PTH. There were no significant differences,in these two groups of patients, between the mean values forcalcium and creatinine in serum. In subsequent studies thesame workers added further support for the role of para-thyrin as a neurotoxin with their observations that, afterparathyroidectomy, motor nerve conduction velocity in-creased in a group of uremic patients on maintenancehemodialysis (45).

In experimental animals, an excess of parathyrin report-edly increases peripheral nerve calcium with a simulta-neous decrease in motor-nerve conduction velocity (46), andthese changes were reversed when the administration ofparathyrin was stopped. Marked abnormalities in electroen-cephalographic patterns have been found in patients withacute renal failure who have concurrent increases in i-PTH;it was proposed that the abnormalities might be ascribed toa direct effect of parathyrin on the brain, causing anincrease in calcium content (47). The electroencephalo-graphic abnormalities show a direct relationship with thei-PTH values (48). In these studies, however, the signifi-cant correlation was with the values for the amino-terminali-PTH fragment, and not with the carboxy-terminal frag-ment, in serum. Moreover, after six months of treatmentwith 1,25-dihydroxycholecalciferol, there was a markeddecline in the concentrations of the amino-terminal i-PTHfragment in serum, which was associated with a significantchange in the electroencephalogram toward normal in someof the patients. Thus the currently available evidence wouldsupport the concept that parathyrin or its immunoreactivedegradation products, or both, act as a neurotoxin in uremia.

An anemia of the normochromic normocytic type is al-most always a feature of chronic renal failure. The kidneysplay a major role in erythropoiesis and in the incorporationof iron into erythrocytes, by their production of erythropoie-tin. Nevertheless, the anemia of chronic renal failure has amultifactorial etiology; it is not simply the end result of adecrease in the renal production of erythropoietin. Thepathogenesis of the anemia of chronic renal failure involvesmultiple factors, parathyrin among them. It is, however,important to recognize that among these the uremic stateper se is of major importance; uremic serum has beenrecently reported to be toxic and inhibitory to erythropoiesis(52). The suppressive activity against erythroid colonygrowth in uremic serum appeared to be contained in com-pounds with molecular masses ranging from 47000 to>150 000 Da. The importance of these observations is that

inhibitors in this molecular-mass range would not be re-moved by conventional dialysis techniques and would not beclassified as “middle molecules” (see below). An anemia ofthe normochromic normocytic type is a recognized clinicalfeature of patients with primary hyperparathyroidism. Al-though the anemia of chronic renal failure is of a differenttype, secondary hyperparathyroidism, with an excess ofi-PTH, has been implicated as a factor in the pathogenesis ofthe anemia found in these patients (42). In uremic patients,subtotal parathyroidectomy, with a resulting decrease inserum i-PTH, is followed by significantly improved hems-tological indices(49-Si). There is no evidence that an excessof parathyrin has a direct toxic effect on erythropoiesis.Rather, the improvement in the anemia after parathyroid-ectomy in uremic patientsseems to be related to a decreasein the amount offibroustissuein their bone marrow-an

increase in which is known to be induced by an excess of

parathyrin.

Middle Molecules

Most investigationsinto the pathophysiology of uremiahave been directed towards the search for and isolation ofeithera singletoxicsubstance or a group ofsubstances,theretention of which would account for all of the clinicalfeatures of the syndrome. The most recent development inthis search has been termed the “middle molecule” hypothe-sis, a concept that has developed as the direct consequence ofthe failureto incriminate such small molecules as urea andcreatinine.

The middle molecule hypothesis is that a group of un-known compounds with molecular masses in the range of300 to 1500 Da are of importance as toxic metabolites inuremia at relatively low concentrations (53, 54). The hy-pothesis is based on (a) a comparison of the clinical resultsobtained on using either hemodialysis or peritoneal dialysisand, more specifically, (b) the relationship between thenumber of hours of dialysis and the development of thesymptoms of neuropathy. The assumption underlying thedevelopment of the middle molecule hypothesis is based onthe difference between the molecular permeability of thecellulosemembranes used in hemodialysis equipment ascompared with that ofthe peritonealmembrane. The latter,which isknown toleak protein, would presumably allow theremoval of retained toxic metabolites with a higher molecu-larweight than those removed by dialysis across a cellulosemembrane. The removal of toxic compounds of highermolecular mass would account for the better clinical results,as assessed by the alleviation of the symptoms of uremia,obtained with the peritoneal dialysis technique. A majorfeature of the middle molecule hypothesis is the concept thatthese unknown molecules exert their toxic effects at rela-tively low concentrations. Reportedly, middle moleculesonly accumulate in quantities that are sufficient to bemeasured in patients with advanced chronic renal failure,as assessed by an endogenous creatinine clearance rate ofless than 11 mL/min (55).

Several recent studies have been aimed at identifring themiddle molecules. One procedure revealed up to 10 identifi-able subpeaks in the middle-molecule range in uremic sera,urine, and erythrocyte hemolysates (56). The techniqueused was a modification of a two-stage chromatographicprocedure, involving use of a molecular sieve followed byion-exchange chromatography (57). Differences were seen inthe distribution of middle molecules among the three differ-ent materials studied (56). Although the concentrations ofthe middle molecules were increased in the serum of pa-tients with uremia, the values for erythrocytes of patientswith the syndrome were similar to those for normal subjects.

CLINICAL CHEMISTRY, Vol.31,No. 1,1985 9

Thus Chapman et al. (56) concluded that “caution is neces-sary in interpreting results of middle molecule analyses.”One biochemical aspect of their findings to which they drewattention is that if any of the species they identified issubsequently shown to be a uremic toxin, the differences intheir distribution among the compartments of the body areof potential importance in the choice of an appropriatemathematical model for optimizing dialysis therapy.

Compounds in the middle-molecule range are said to be inhigher concentrations in the serum of”sick” uremic patientsthan in an asymptomatic group of uremics (58,59), althoughthere is a considerable overlap between the data on individ-ual middle-molecule fractions for the two groups (60). Theseauthors could not correlate the accumulation of any specificmiddle-molecule fraction with the occurrence of any specificuremic symptom in the “sick” group of patients. In contrast,using a two-stage chromatographic technique-gel perme-ation followed by anion-exchange chromatography-onegroup of workers (61) has correlated the amplitude of onespecific middle-molecule peak with the occurrence of uremicneuropathy. After successful renal transplantation, uremicmiddle molecules rapidly disappear from serum, even morerapidly than the creatiine concentration declines (62).Collectively these observations add support to the conceptthat compounds in the middle-molecule range may indeedplay a role as uremic toxins.

A major problem in the development and understandingof the potential toxic nature of middle molecules has beenthe failure of various attempts to define both their chemicaland biological nature. Menyh#{225}rtand Gr#{243}f(63) proposed thata large group of unknown peptides with limited diffusibilitythrough hemodialysis membranes constituted most of theuremic middle molecules. They investigated the molecularcomposition of uremic middle molecules in the 500-5000range of relative molecular mass. Using cation-exchangecolumn chromatography, they separated three fractions ofserum containing these molecules. Subsequent one-dimen-sional paper chromatography of these fractions demonstrat-ed that each of them could be further resolved into at leastseven ninhydrin-positive subfractions. The qualitative ami-no acid composition of the subfractions differed from eachother and from those of known peptides with which theywere compared. Some of these peptides appeared to becompletely removed from uremic serum by hemodialysis;others were only partly removed or were completely unal-tered by this treatment.

The kidneys play a major role in the enzymatic degrada-tion and clearance of hormones from the circulation. Adisturbance in the normal clearance patterns of hormonespotentially could play a role in the pathogenesis of theuremic syndrome. According to this concept, retention ofeither the hormones themselves or of their subsequentlyrenal-cleared peptide degradation products could be regard-ed as potential uremic toxins, the unknown peptides report-ed by Menyhart and Gr#{243}f(63).

Although the clinical evidence I have mentioned supportsthe middle-molecule hypothesis, in the pathogenesis of theclinical syndrome of uremia their precise role still remainsto be clarified. The middle-molecule hypothesis currently isused to refer to serum solutes in the relative molecular massrange of 500 to 2000 Da. In a recent study, uremic ultrafil-trates were fractionated by means of gel ifitration and thenfurther investigated by combined analytical techniques:“high-performance” liquid chromatography, liquid chroma-tography, gas chromatography, mass spectrometry, andisotachophoresis (64). These studies led the authors topropose that gel filtration was inappropriate as an analyti-cal technique for determination of middle molecules, be-

cause they found that ultrafiltrate fractions in the middle-molecular-mass range also contain a considerable amount ofsubstances of low molecular mass, such as carbohydrates,organic acids, amino acids, and ultraviolet-absorbing sol-utes. The earlier work on middle molecules was largelybased on gel ifitration studies, so they recommended recon-sideration of the importance of low-molecular-mass com-pounds and their role as uremic toxins. Other workers (65)have also concluded that “despite the extensive studies inthis field, results are still confusing concerning the role ofso-called ‘middle molecules’ in uremic toxicity.” They (65)drew specific attention to the fact that a major problem inthis field of investigation has been the differences betweenthe analytical techniques used by different groups of work-ers, which has made it impossible to inter-compare thereported middle-molecule fractions. They proposed that “itappears necessary to develop a chromatographic referencesystem which would be the basis for the evaluation of the invitro and in vivo effects of the various molecules ranging inthe still poorly defined M.W. fraction” (65). It can only beconcluded at this time that, until the chemical identity ofthe retained uremic middle molecules is known and theirbiological toxicity has been demonstrated by experimentalin vivo and in vitro studies, their role as uremic toxins isonly hypothetical. Chemical identification and characteriza-tion of middle molecules will probably require mass spec-

trometry but this will have to await further technological

advances in this field (66).

Retained Uremic Metabolites as Enzyme Inhibitors

A potential role of the retained serum metabolites astoxins in the pathogenesis of the uremic syndrome is asenzyme inhibitors. Urea, in concentrations comparable tothose found in uremia, significantly inhibits monoamineoxidase (67). This enzyme has a major role in the destructionof both serotonin and catecholamines and in the regulationof amine metabolism in nervous tissue, and its inhibition byurea could be of significance in the pathogenesis of certainclinical features of the uremic syndrome. Urea also inhibitsoxygen uptake by brain slices (68), but only at very highconcentrations as compared with those in blood and after a3-h lag period. Urea and creatinine reportedly inhibit glu-cose uptake and utilization in rat diaphragm in vitro; therewas enhancement of the inhibitory effect when these twocompounds were used together in the system (69). Theselatter findings are consistent with the proposal that theretained metabolites have a cumulative enyzme inhibitoryeffect in vivo.

Phenolic acids affect cerebral metabolism, as measured bythe rate of respiration and anaerobic glycolysis of guinea pigbrain slices, and they also inhibit the activity of someselected enzymes (70). The enzymes studied were the decar-boxylases of 3,4-dihydroxyphenylalanine, 5-hydroxytrypto-phan, and glutamic acid; aspartate aminotransferase; 5’-nucleotidase; amine oxidase; and lactate dehydrogenase.Many aromatic acids, especially those with an unsaturatedside chain, depressed enzyme reaction rates. The concentra-tions of phenolic acids used by Hicks et al. (70) were higherthan those found in uremic serum, but they proposed thatthe lower concentrations of phenolic acids in uremic serummight possibly exert an effect by virtue of having beenpresent for a longer time than the relatively high concentra-tions used in their enzyme inhibition studies. Alternatively,it could be that these compounds exert a cumulative inhibi-tory effect in vivo.

Aromatic compounds retained in uremia may exert anenzyme inhibitory action by a summation effect in vivo; botharomatic and aliphatic amines can cause enzyme inhibition

10 CLINICALCHEMISTRY, Vol. 31,No. 1,1985

(71). However, compared with the phenolic acids, theamines were less effective inhibitors of glutamic acid anddihydroxyphenylalamne decarboxylases. Amines cross theblood-brain barrier more readily than do acids (72) andtherefore may be more effective in vivo than in vitro. Theincreased concentrations of aromatic amines in serum inuremia approximately correspond with the increase in bloodurea nitrogen concentrations.

Although the precise biochemical role and the potentialmechanisms of the retained metabolites as enzyme inhibi-tors in the etiology of the clinical syndrome of uremiaremain to be clarified, these compounds clearly do accumu-late in the extracellular-fluid compartment, and many ofthem can inhibit enzymes. Nevertheless, collectively theymay exert their actions at a final common site. In one reviewof uremia, a unifying hypothesis for the toxicity of uremiawas proposed, with derangements in membrane transport asthe final common pathway (5). This hypothesis was based ona perspective of the wide range of known and potential toxicmetabolites retained in uremic serum. Skeletal resistance tothe action of exogenous parathyrin is a recognized feature ofthe disturbance in calcium homeostasis in patients withuremia. Wills and Jenkins (74), who studied parathyrin-induced bone resorption by using an in vivo organ culturesystem, reported that some known retained uremic metabo-lites, including phosphate, when tested individually, inhib-ited the calcium-mobilizing action of the hormone on bone,but the inhibitory effect of many of the individual metabo-lites was only significant at concentrations higher thanthose found in the serum of uremic patients. Serum collectedfrom uremic patients prior to hemodialysis totally inhibitedthe calcium-mobilizing action of parathyrin; that collectedfrom the same patients just after dialysis was not inhibitory.Thus they (74) proposed that if the uremic metabolites havean inhibitory effect on the action of parathyrin, it must be acumulative phenomenon, which potentially could involveeither the blockage of membrane receptors to parathyrin orhormone-induced effects on cell membrane transport.

Aluminum

The potential role of aluminum as a toxic agent inpatients with normal renal function is controversial. Still, ithas been established that in chronic renal failure there is anincreased total body burden of aluminum and that this isassociated with toxic sequelae involving bone, brain, anderythropoietic tissues (75). Aluminum salts are extensivelyused in the therapeutic management of the hyperphosphate-mia that is a relatively constant feature of patients withchronic renal failure. Aluminum is absorbed from thegastrointestinal tract. In normal subjects, after an oral loadits concentration in serum increases, followed by its in-creased excretion in the urine (76). Increased aluminumconcentrations in the serum of some patients with chronicrenal failure was first reported in 1970 (77). It is nowestablished that hyperaluminemia may occur in patientswith end-stage chronic renal failure and is associated withthe accumulation of aluminum in various tissues. Hyperalu-minemia and the associated toxic sequelae may occur inpatients on treatment with either hemodialysis or peritone-al dialysis and in some patients who have not been dialyzedbut are being treated with orally administered aluminumsalts (75, 78). The high values for aluminum in serum andtissue result from the intestinal absorption of aluminumsalts taken by mouth and from passage of aluminum acrossthe dialysis membrane. The aluminum content of the dialyz-ing fluid obviously also depends on the Al content of thewater used as the solvent. Some domestic tap-water con-tains aluminum in high concentration, either naturally or

because aluminum has been added as a flocculant in thepurification process. Acid rain markedly increases the “nat-ural” aluminum content of water.

Aluminum is now recognized as a major, if not the major,toxic factor in the pathogenesis of a progressive fatal neuro-logical syndrome in patients with chronic renal failure thatwas first reported in 1972 (79). The syndrome was latertermed “dialysis encephalopathy” or “dialysis dementia,”and these patients showed increased amounts of aluminumin the brain, muscle, and bone tissue (80). Aluminumpotentially exerts its neurotoxic action by inhibiting dihy-dropteridine reductase (EC 1.6.99.7) (81). Such inhibitionwould result in a decrease in the tetrahydrobiopterin, tyro-sine, and neurotransmitters in brain. The neurotoxicity ofaluminum may also involve alterations in the major post-synaptic enzymes of cholinergic neurotransmission (82).Although the precise mechanism possibly remains to bedefined, much evidence indicates that aluminum is neuro-toxic for patients whose functional renal mass is decreased.The latter leads to a reduction in the normal renal clearanceof aluminum with a consequent increase in its concentrationin serum and body tissue.

Bone pain, as a consequence of metabolic bone disease, isa common symptom in patients with chronic renal failurewho are receiving long-term treatment with intermittenthemodialysis. The progressive metabolic bone disease inthese patients is usually termed “dialysis osteodystrophy,” aterm that distinguishes it, and some aspects of its pathogen-esis, from renal osteodystrophy in the undialyzed patient.The osteomalacic component of dialysis osteodystrophy is amajor clinical problem because it is associated with a highincidence of fractures. An increase in the Al content of bonein some patients with end-stage chronic renal failure wasfirst reported in 1970 (83). Only years later was an associa-tion reported between the occurrence of dialysis encephalop-athy and osteomalacia: exposure to high amounts of alumi-num was a factor common to both of those complications inuremic patients (84, 85). Clearly, aluminum plays a majoretiological toxic role in one particular type of osteomalacicdialysis osteodystrophy (75), a type that usually occurs inpatients on dialysis treatment but may also occur in non-dialyzed patients (86). The mechanism for the disorderedbone formation induced by an excess of aluminum remainsto be clarified. It may involve a disturbance either in theformation of calcium apatite or in the bone-mineralizationprocess (75,87). There is also evidence from in vitro studiesthat suggests that aluminum may affect the activities of thebone enzymes, acid and alkaline phosphatase, and modiItheir response to parathyrin and 1,25-dihydroxycholecalci-ferol (88).

In patients on dialysis treatment who develop aluminumtoxicity the major source of the aluminum is the tap-waterused to prepare the dialysate. In addition there is also someintestinal absorption from the aluminum-containing phos-phate-binding gels; in some patients this route appears tohave been dominant. The driving force for aluminum trans-fer during dialysis appears to be the effective concentrationgradient between the dialysate aluminum and the freediffusible aluminum fraction in serum (89). Transfer ofaluminum from the dialyzing fluid across the dialyzingmembrane appears to occur even when concentrations of themetal in the fluid are low (90, 91). The major portion, if notall, of aluminum in blood is bound to serum proteins and anas-yet-unidentified low-Mr species (91). The identification ofthe latter may be of importance in understanding themechanisms of tissue accumulation and consequent toxicity.

Although aluminum has attracted much attention inrecent years in this regard, it is important to recognize that

CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985 11

other trace metals in the serum and total body may beaffected and may also play a role in the pathogenesis of someof the clinical features of uremia. Uremic patients show anincreased nickel concentration in serum, but this does notappear to be associated with a corresponding increase intissues, in contrast to the hyperaluminemia of chronic renalfailure (92). As for other trace metals and their potentialeffects in chronic renal failure, it is important to recognizethat deficiency states may also play a contributory role inthe pathogenesis of the clinical syndrome of uremia. Zincdeficiency is a feature of uremia (93) and has been implicat-ed as a major cause of the impairment in cellular immunitywhich is seen in these patients (94). The rates of sodium andpotassium transport in vitro are normally influenced by theextracellular concentration of zinc; the cellular membranetransport of sodium reportedly is defective in leukocytes anderythrocytes from uremic patients. A low serum zinc concen-tration, however, does not account for the defect in mem-brane sodium transport in uremia (95), nor is it associatedwith an intracellular deficiency (95).

Conclusion

The biochemical disturbances-and in particular the roleof the retained metabolites as toxins-in the pathogenesis ofthe syndrome of uremia are complex. The complexity is alsomanifested by the variety and nature of the individualdisorders that make their individual contributions to thefeatures seen in the final clinical syndrome. The complexityis a reflection of the importance of the homeostatic andendocrine roles of the kidneys in maintaining the constitu-tion of the internal environment. In any review of uremia itis apparent that none of the body systems is spared. Theconstitution of the internal environment is disturbed notonly by the retention of metabolic end products but also bythe effects of the retained metabolites on intermediarymetabolism. In addition to the many metabolites that areretained in the body fluids as a consequence of chronic renalfailure there are major disturbances in total body electrolytecomposition and the distribution of body water (96, 97) andhormone-control mechanisms (98). Brennan et al. (96) re-ported that in patients with end-stage chronic renal failurethere were significant increases in total body sodium, chlo-ride, water, and extracellular fluid volume. These changesin total body water and sodium are linked both withalterations in their respective hormonal control mecha-nisms and with the clinical problem of hypertension inuremia. The role of the changes in extracellular fluidvolume and electrolytes in the causation of some of the otherbiochemical and metabolic disturbances in the syndrome ofuremia is not clearly defined.

The complexity, interactions, and consequences of theuremic state and the potential role of toxins are exemplifiedby a consideration of the disorder in carbohydrate andinsulin metabolism in chronic renal failure. Patients withchronic renal failure have a diminished ability to handle aglucose load, the degree of glucose intolerance correlatingroughly with the severity of the uremia (3). The disturbancein glucose metabolism is the end result of the summation ofseveral interrelated factors, including toxins. The dominantfeature in the disturbance would appear to be an impair-ment of the insulin-dependent transfer of glucose into extra-hepatic tissues and its subsequent utilization, with a defectin insulin synthesis or release (or both) playing a relativelyminor role. The primary defect in the sequence appears to bean insensitivity of peripheral tissue to insulin, probably adefect either in intracellular metabolism or in the glucosetransport system (99). The importance of the disturbance incarbohydrate and glucose metabolism is not in any way the

direct result of variations in blood glucose concentration-hyperglycemia is rare-rather, it is ascribable to the impor-tant etiological role it potentially plays in the disturbance inlipoprotein metabolism in uremia. The precise cause of thedyslipoproteinemia of chronic renal failure is not clear,although the disorder in carbohydrate metabolism is proba-bly a major factor. The severity of the dyslipoproteinemiahas shown no consistent correlation with the nature of theunderlying renal disease, the degree or duration of renalfailure, or the diet. The patterns of the disturbance inlipoprotein metabolism in patients with chronic renal fail-ure take several forms: an increase in the concentration andcholesterol content of very-low-density lipoprotein; an in-crease in the size and triglyceride content of low-densitylipoprotein particles; a decreased concentration of cholester-ol in high-density lipoprotein; and an increased prevalenceof an electrophoretic subclass (late pre-/3-lipoprotein) ofvery-low-density lipoprotein (100, 101). The importance ofthe dyslipoproteinemia of uremia is that it is related to theaccelerated atherogenesis and premature coronary arterydisease that is a clinical feature of patients with chronicrenal failure, particularly those being treated by long-termintermittent hemodialysis.

Uremia is a multi-system clinical and biochemical prob-lem that still requires extensive clarification. It would seemunlikely, in the final analysis, if any one individual toxinwill ever be identified as “the” toxin; rather, the syndrome ofuremia probably will be attributed to a cumulative effect invivo, to which all of the retained metabolites contribute. Inthe process of evolution the kidneys have developed, duringthe past 600 million years, into complex organs, sharply incontrast to the simple open-ended tubes that drained wasteproducts from the coelomic cavity of our early remoteancestors (3). As complex, sophisticated homeostatic andendocrine organs the kidneys play the dominant role in themaintenance of the internal environment. The failure toremove metabolic waste end products from the urine, theirconsequent retention in the blood and tissue compartments,and their potential cumulative action as toxins are majorfeatures in the pathogenesis of the clinical syndrome ofuremia.

References1. Bernard C. Lecons sur les Phenomnes di Is Vie Commune auxAnimaux et aux Vegetaux, 1, Balli#{233}re,Paris, France, 1878.2. Peters JP. Body Water, Balli#{233}re,Tindall and Cox, London, U.K.,1935.3. WiI1 MR. Metabolic Consequences of Chronic Renal Failure, HM+ M Publishers, Ltd., Aylesbury, Buckinghamshire, U.K., 1978.

4. Piorry PA, L’H#{233}ritierD. Trait#{233}des Alterations du Sang, Bail-here, Paris, France, 1840.5. Schreiner GE, Winchester JF. Uremia-1978 perspective. ClinNephrol 11, 52-55 (1979).6. Bostock J. Observations on the chemical properties of urine,Quoted by R Bright in Reports of Medical Cases, Selected with aView of Illustrating the Symptoms and Cure of Diseases by aReference to Morbid Anatomy, 1, Longman, Rees, Orme, Brown andGreen, London, U.K., 1827, pp 75-85.7. Bright it Reports of Medical Cases, Selected with a View ofIllustrating the Symptoms and Care of Diseases by Reference toMorbid Anatomy, 2, Longman, Rees, Orme, Brown and Green,London, U.K., 1831.

8. Merrill JP, Legrain M, Hoigne R. Observations on the role ofurea in uremia. Am J Med 14, 519-520 (1953).

9. Kennedy AC, Linton AL, Eaton JC. Urea levels in cerebrospinalfluid after hemodialysis. Lancet i, 410-411 (1962).

10. Wakim KG, Predominance of hyponatremia over hypo-osmolal-ity in simulation of the dialysis disequilibrium syndrome. MayoClinic Proc 44, 433-460 (1969).11. Port FK, Johnson WJ, Klass DW. Prevention of dialysis dis-

12 CLINICAL CHEMISTRY, Vol.31,No. 1,1985

equilibrium syndrome by use of high sodium concentration in thedialysate. Kidney mt 3, 327-333 (1973).12. Cockcroft DW, Gault MR. Prediction of creatimne clearancefrom serum creatinine. Nephron 16, 31-41 (1976).

13. Mitch WE, Collier VU, Walser M. Creatinine metabolism inchronic renal failure. Clin Sci 58, 327-335 (1980).14. Gresham GE, Keller MD. Hyperuricemia and chronic renaldisease. J Chronic Die 23, 755-762 (1971).15. Danovitch GM, Weinberger J, Berlyne GM. Uric acid inadvanced renal failure. Clin Sci 43, 331-341 (1972).16. Danovitch GM. Uric acid transport in renal failure. A review.Nephron 9, 291-299 (1972).17. Richet G. Someaspectsof uric acid metabolism in chronic renalfailure. In Nutrition in Renal Disease, GM Berlyne, Ed., Living-stone, Edinburgh, U.K., 1968, pp 133-139.18. S#{248}rensenLB. The elimination of uric acid in man. Scand J ClinLab Invest 12, 1-2 14 (1960).19. Zarembski PM, Hodgkinson A, Parsons FM. Elevation of theconcentration of plasma oxalic acid in renal failure. Nature (Lon-don) 212, 511-512 (1966).20. Bennett B, Rosenblum C. Identification of calcium oxalatecrystals in the myocardium in patients with uremia. Lab Invest 10,947-955 (1961).21. Bennington JL, Haber SL, Smith JV, et al. Crystals of calciumoxalate in the human kidney. Am J Clin Pathol 41, 8-14 (1964).

22. Fanger H, Esparza A. Crystals of calcium oxalate in kidneys inuremia. Am J Clin Pathol 41, 597-603 (1964).

23. Emerson PM, Wilkinson JH. Urea and oxalate inhibition of theserum lactate dehydrogenase. J Clin Pathol 18, 803-807 (1965).

24. Clements Jr RS, DeJesusJr PV, Winegrad Al. Raised plasma-myoinositol levels in uraemia and experimental neuropathy. Lanceti, 1137-1141 (1973).25. Reznek RH, Salway JG, Thomas PK. Plasma-myoinositol con-centrations in uraemic neuropathy. Lancet 1, 675-676 (1977).26. Blwnberg A, Esslen E, B#{252}rgiW. Myoinositol-a uremic neuro-toxin? Nephron 21, 186-191 (1978).27. Giovanetti S, Biagini M, Balestri PL, et al. Uraemia-likesyndrome in dogs chronically intoxicated with methylguanidineand creatinine. Clin Sci 36, 445-452 (1969).28. Stein IM, Perez G, Johnson R, Cummings NB. Serum levelsand urinary excretion of methylguanidine in chronic renal failure.JLab Clin Med 77, 1020-1024 (1971).29. Stein IM, Michlus MJ. Concentrations in serum and urinaryexcretion of guanidine, 1-methylguanidine, and 1,1-dimethylguani-dine in chronic renal failure. Clin Chem 19, 583-585 (1973).30. Giovanetti 5, Barsotti G. Methylguanidine in uremia. ArchIntern Med 131, 709-713 (1973).31. Giovanetti 5, Barsotti G. Uremic intoxication. Nephron 14,123-133 (1975).32. Felgate RA, Taylor WH. The effects of methylguanidine ontissue respiration in vitro. Clin Sci 41, 85-87 (1971).33. Balestri PL, Rindi P, Biagini M, et al. Effects of uraemic serum,urea, creatinine and methylguanidine on glucosemetabolism. ClinSci 42, 395-404 (1972).

34. Natelson 5, Stein I, BonasJE. Improvements in the method ofseparation of guarndino organic acids by column chromatography.Isolation and identification of guanidinosuccinic acid from humanurine. Microchem J 8, 371-382 (1964).35. Cohen BD, Stein IM, Bonas JE. Guarudinosuccinic aciduria inuremia. A possible alternative pathway for urea synthesis. Am JMed 45, 63-68 (1968).36. Stein IM, Cohen BD, Kornhauser RS. Guanidinosuccinic acidin renal failure, experimental azotemia and inborn errors of theurea cycle. N EngI J Med 280, 926-930 (1969).37. Dobbelstein H, Edel HH, Schmidt M, et al. Guanidinberstein-sflure and Uramie. I. Mitteilung: Klimsche Untersuchungen. KlinWochenschr 49, 348-357 (1971).38. Dobbelstein H, Grunst J, Schubert G, et al. Guanidinberstein-s#{228}ureand Ur#{228}mie.II. Tierexperimentelle Befunde. Kim Wo-chenschr 49, 1077-1083 (1971).39. Simenhoff ML, Saukkonen JJ, Burke JF, et al. Amine metabo-lism and the small bowel in uremia. Lancet ii, 818-821 (1976).40. Simenhoff ML, Saukkonen JJ, Burke JF, et al. Importance of

aliphatic amines in uremia. Kidney Int 13, S16-S19 (1978).41. Simenhoff ML, Burke JF, Saukkonen JJ, et al. Biochemicalprofile of uremic breath. N Engl J Med 297, 132-135 (1977).42. Massry SO. Is parathyroid hormone a uremic toxin? Nephron19, 125-130 (1977).43. Massry SO, Goldstein DA. The search for uremic toxin(s) “X”-“X” = PTH. ClinNephrol 11, 181-189 (1979).44. Avram MM, Feinfeld DA, Huatuco AR. Search for the uremictoxin. Decreased motor-nerve conduction velocity and elevatedparathyroid hormone in uremia. N EngI J Med 298, 1000-1003(1978).45. Avram MM, Morrow IP, Feinfeld D, et ah. Uremic syndrome inman: New evidence for parathormone as a multi-system nuero-toxin. Clin Nephrol 11, 59-62 (1979).

46. Goldstein DA, Chui LA, Massry SG. Effect of parathyroidhormone and uremia on peripheral nerve calcium and motor-nerveconduction velocity. J Clin Invest 62, 88-93 (1978).47. CooperJD, Lazarowitz VC, Arieff A!. Neurodiagnostic abnor-malitiesinpatients with acute renal failure. Evidence for neurotox-icity of parathyroid hormone. J Clin Invest 61, 1448-1455 (1978).48. Goldstein DA, Feinstein El, Chui LA, et al. The relationshipbetween the abnormalities in electroencephalogram and bloodlevels of parathyroid hormone in dialysis patients. J Clin Endo-crinol Metab 51, 130-134 (1980).49. Zingraff J, Drueke T, Marie P, et al. Anemia and secondaryhyperparathyroidism. Arch Intern Med 138, 1650-1652 (1978).50. Barbour GL. Effect of parathyroidectomy on anemia in chronicrenal failure. Arch Intern Med 139, 889-891 (1979).51. Podjarny E, Rathaus M, Kormts Z, et al. Is anemia of chronicrenal failure related to secondary hyperparathyroidism? Arch In-tern Med 141, 453-455 (1981).52. Freedman MH, Cattran DC, Saunders EF. Anemia of chronicrenal failure: Inhibition of erythropoiesis by uremic serum. Neph-ron 35, 15-19 (1983).53. Babb AL, Popovich RP, Christopher TO, et al. The genesisofthe square metre-hour hypothesis. Trans Am Soc Artif Int Organs17, 81-91 (1971).

54. Babb AL, Farrell PC, Uvelli DA, et al. Hemodialyzer evalua-tion by examination of solute molecular spectra. Trans Am SocArtiflnt Organs 18, 98-105 (1972).

55. Asaba H. Accumulation and excretion of middle molecules.Clin Nephrol 19, 116-123 (1983).56. Chapman GV, Ward RA, Farrell PC. Separation and quantifi-cation of the “middle molecules” in uremia. Kidney mt 17, 82-88(1980).57. FUrst P, Zimmerman L, Bergstrom J. Determination of endoge-nous middle molecules in normal and uremic body fluids. ClinNephrol 5, 178-188 (1976).

58. Bergstrom J, Furst P, Zimmerman L. Uremic middle moleculesexist and are biologically active. Clin Nephrol 11, 229-238 (1979).59. Asaba H, Alvestrand A, Bergstrom J, et al. Uremic middlemolecules in non-dialyzed azotemic patients: Relation to symptomsand clinical biochemistries. Clin Nephrol 17,90-95 (1982).

60. Asaba H, Alvestrand A, Furst P, et al. Clinical implications ofuremic middle molecules in regular hemodialysis patients. ClinNephroi 19, 179-187 (1983).61. Faguer P, Man N-K, Cueille G, et al. Improved separation andquantification of the “middle molecule” b4-2 in uremia. Clin Chem29, 703-707 (1983).62. Asaba H, Bergstrom J, Ftirst P, et al. The effect of renaltransplantation on middle molecules in plasma and urine. ClinNephrol 8, 329-334 (1977).

63. Menyhart J, Grof J. Many hitherto unknown peptides areprincipal constituents of “uremic” middle molecules. Clin Chem 27,1712-1716 (1981).

64. Schoots AC, Mikkers FEP, ClaessensHA, et al. Characteriza-tion of uremic “middle molecular” fractions by gaschromatography,mass spectrometry, isotachophoresis, and liquid chromatography.Clin Chem 28, 45-49 (1982).65. Navarro J, Contreras P, Touraine JL, et al. Are ‘middlemolecules’ responsible for toxic phenomena in chronic renal failure?Nephron 32, 301-307 (1982).66. Fenselau C. Mass spectrometry of middle molecules. AnalChem 54, 105A-116A (1982).

CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985 13

67. Giordano C, Bloom J, Merrill JP. Effect of urea on physiologicsystems. I. Studies on monoamine oxidase activity. JLab Clin Med59, 396-400 (1962).68. Lacefles PT, Taylor WH. The effect upon tissue respiration invitro of metabolites which accumulate in uremic coma. Clin Sci 31,403-413 (1966).69. Balestri PL, Rindi P, Biagini M, et al. Effects of uremic serum,urea, creatiine and methylguanidine on glucose metabolism. ClinSci 42, 395-404 (1972).70. Hicks JM, Young DS, Wootton IDP. The effect of uremic bloodconstituents on certain cerebral enzymes.Clin Chim Acta 9, 228-235 (1964).71. Young DS, Wootton IDP. The retention of ammes asa factor inuremic toxaemia. Clin Chim Acta 9, 503-505 (1964).72. Schwerin P, Bessman SP, Waelsch H. The uptake of glutamicacid and glutamine by brain and other tissues of the rat and mouse.JBiol Chem 184, 37-44 (1950).73. Morgan RE, Morgan JM. Plasma levels of aromatic amines inrenal failure. Metabolism 15, 479-481 (1966).74. Wills MR, Jenkins MV. The effect of uraemic metabolites onparathyroid extract-induced bone resorption in vitro. Clin ChimActa 73, 121-125 (1976).75. Wills MB, Savory J. Aluminum poisoning Dialysis encephalop-athy, osteomalacia and anaemia. Lancet ii, 29-34 (1983).76. Mauras Y, Allain P, Riberi P. Etude de l’absorption digestive del’hydrocarbonate d’aluminum chez l’individu sain. Therapie 37,593-605 (1982).

77. Berlyne GM, Ben-An J, Pest D, et al. Hyperaluminaemia fromaluminum resins in chronic renal failure. Lancet ii, 494-496(1970).78. Alfrey AC. Aluminum intoxication. N Engi J Med 310, 1113-1115 (1984).79. Alfrey AC, Mishell JM, Burks J, et al. Syndrome of dyspraxiaand multifocal seizures associated with chronic hemodialysis. TransAm Soc Artif Intern Org 18, 257-261 (1972).80 Alfrey AC, L.eGendre OR, Kaehny WD. The dialysis encepha-lopathy syndrome. Possible aluminum intoxication. N Engl J Med294, 184-188, (1976).81. Leeming RJ, Blair JA. Dialysis dementia, aluminium andtetrahydrobiopterin metabolism. Lancet i, 556 (1979).82. Marquis J. Aluminum neurotoxicity: An experimental perspec-tive. Bull Environ Contam Toxicol 29, 43-49 (1982).83. Parsons V, Davies C, Goode C, et al. Aluminium inbone frompatients with renal failure. Br Med J iv, 273-275 (1971).84. Ward MK, Feest TO, Ellis HA, et al. Osteomalacic dialysisosteodystrophy:Evidence for a water-borne aetiological agent prob-ably aluminium. Lancet i, 841-845 (1978).

85. Parkinson IS, Ward MK, Feest TO, et al. Fracturing dialysisosteodystrophy and dialysis encephalopathy. An epidemiologicalsurvey. Lancet i, 406-409 (1979).

86. Kaye M. Oral aluminum toxicity in a non-dialyzed patient withrenal failure. Clin Nephroi 20, 208-211 (1983).87. Thomas WC, Meyer JL. Aluminum-induced osteomalacia: Anexplanation. Am J Nephrol 4, 201-203 (1984).88. Lieberherr M, Grosse B, Cournot-Witmer G, et al. In vitroeffects of aluminum on bone phosphatases: A possible interactionwith bPTH and vitamin D metabolites. Calcif Tissue mt 34,280-284 (1982).89. Graf H, Stummvoll HK, Meisinger V. Dialysate aluminiumconcentration and aluminium transfer during haemodialysis. Lan.oct i, 46-47 (1982).90. Kaehny WD, Alfrey AC, Holman RE, et al. Aluminum transferduring hemodialysis. Kidney mt 12, 361-365 (1977).91. Savory J, Wills MR. Dialysis fluids as a source of aluminumaccumulation. In Trace Elements in Rena! Insufficiency, GM Ger-lyne, S Giovanetti, Eds., Karger, Basle, Switzerland, 1984, pp 12-23.92. Savory J, Brown S, Bertholf R, et al. Serum and lymphocyte,nickel and aluminum concentrations in patients with extracorpore-al hemodialysis. Ann Clin Lab Sci 14, 413-414 (1984).

93. Lindemann RB, Baxter DJ, Yunice An, et al. Serum concentra-tion and urinary excretions of zinc in cirrhosis, nephrotic syndromeand renal insufficiency. Am J Med Sci 275, 17-31 (1978).

94. Antomou LD, Shaloub RJ, Schechter GP. The effect of zinc oncellular immunity in chronic uremia. Am J Clin Nutr 34, 1912-1917 (1981).95. Michael J, Hilton PJ, Jones NF. Zinc and the sodium pump inuremia. Am J Ciin Nutr 31, 1945-1947 (1978).96. Brennan BL, Yasumura 5, Letteri JM, et al. Total bodyelectrolyte composition and distribution of body water in uremia.Kidney mt 17, 364-371 (1980).97. Mitch WE, Wilcox CS. Disorders of body fluids, sodium andpotassium in chronic renal failure. Am J Med 72, 536-550 (1982).98. Bonomini V, 0mm 0, Stefoni S, et al. Hormonal changes inuremia. Clin Nephrol 11, 275-280 (1979).99. De Fronzo RA. Pathogenesis of glucose intolerance in uremia.Metabolism 27, 1866-1880 (1978).100. Norbeck HE. Serum lipoproteins in chronic renal failure. ActaMed Scand 1, Suppl 649, 1-49 (1981).101. Chan MK, Varghese Z, Moorhead JF. Lipid abnormalities inuremia, dialysis and transplantation. Kidney mt 19, 625-637(1981).