VIEWS, VISIONS AND VISTAS IN DIALYSIS

New Methods and Technologies for Measuring Uremic Toxins and Quantifying Dialysis Adequacy Raymond Vanholder, Jente Boelaert, Griet Glorieux, and Sunny Eloot Nephrology Section, Department of Internal Medicine, Ghent University Hospital, Gent, Belgium

ABSTRACT This publication reviews the currently available methods to identify uremic retention solutes, to determine their biological relevance and to quantify their removal. The analytical methods for the detection of uremic solutes have improved continuously, allowing the identification of several previously unknown solutes. Progress has been accelerated by the development of comprehensive strategies such as genomics, proteomics and the latest “omics” area, metabolomics. Those methodologies will be further refined in future. Once the concentration of solutes of interest is known based on targeted analysis, their biological relevance can be studied by means of in vitro, ex vivo, or animal models, provided those are representative for the key complications of the uremic syndrome. For this to come to pass, rigid protocols should be applied, e.g., aiming at free solute concentrations conform those found

in uremia. Subsequently, the decrease in concentration of relevant solutes should be pursued by nondialysis (e.g., by influencing nutritional intake or intestinal generation, using sorbents, modifying metabolism, or preserving renal function) and dialysis methods. Optimal dialysis strategies can be sought by studying solute kinetics during dialysis. Clinical studies are necessary to assess the correct impact of those optimized strategies on outcomes. Although longitudinal studies of solute concentration and surrogate outcome studies are first steps in suggesting the usefulness of a given approach, ultimately hard outcome randomized controlled trials are needed to endorse evidence-based therapeutic choices. The nonspecificity of dialysis removal is however a handicap limiting the chances to provide proof of concept that a given solute or group of solutes has definite biological impact.

The uremic syndrome is characterized by the retention of various solutes, which are normally excreted by the kidneys. If these retention solutes exert adverse biochemical/biological effects, they are called uremic toxins. They are preferentially classified according to their physicochemical characteristics, which have an impact on their removal during dialysis. This classification focuses on three types of solutes, i.e., the small water-soluble compounds (MW < 500 Da), the middle molecules (MW > 500 Da), and the protein-bound compounds (1). Research on uremia has resulted in the identification of many retention solutes, including uremic toxins. In 2003, the European Uremic Toxin (EUTox) Work Group classified 90 retention solutes which had been quantified at that time (1). This list has been extended with an additional 56 solutes in 2012 (2). Nevertheless, many uremic retention solutes have not been identified yet.

The current challenge is to identify uremic retention solutes, to determine their biological relevance, and to quantify their removal. Detection and Identification of Uremic Retention Solutes Recently, -omic technologies have been explored in the search for new uremic retention solutes (3–9). Genomics has indispensably initiated the omics cascade and paved the way for transcriptomics, proteomics, and metabolomics. In the context of uremia, proteomics, and metabolomics have been the main omics applications (3–5,7–11). Proteomics is suited for the study of peptides and proteins (middle molecules), while metabolomics focuses on small molecules. Although “-omic” strategies are complementary, analysis of the metabolome is particularly useful as an approach for identifying pathways that are perturbed in a given pathology (12,13). Since metabolites are downstream of both transcription and translation, they are typically more closely associated with disease processes than proteins, mRNA or genes, and for that reason also particularly suited for the study of uremic toxicity, next to proteomics which is more suited for the larger peptides (middle molecules). Next to the metabolome’s close biological proximity to the

Address correspondence to: Raymond Vanholder, Ghent University Hospital, Nephrology Section, De Pintelaan 185, 9000 Gent, Belgium, Tel.: +32-9-332-45-25, or e-mail: [email protected]. Seminars in Dialysis—2014 DOI: 10.1111/sdi.12331 © 2014 Wiley Periodicals, Inc. 1

2

Vanholder et al.

phenotype of the system, metabolomics offers several other advantages such as relatively low price on a per-sample basis, relatively high throughput and possibilities for automation (13). Metabolomics can be nontargeted or targeted. Nontargeted metabolomics aims to capture as much information as possible. Since the goal is nonbiased detection of unknowns, these assays are only semiquantitative at best and not validated. As there is currently no single technology available that allows the comprehensive analysis of the metabolome, nontargeted metabolomics is characterized by the use of multiple analytical techniques (13). In contrast, targeted methods measure one or several well-defined compounds, and are quantitative and validated. Quantification can be done using calibration curves and stable isotopically labeled internal standards. In general, nontargeted methods are only hypothesisgenerating of which the results require follow-up with targeted approaches. The workflow in nontargeted metabolomics consists of experimental design, sampling, storage and transfer, sample treatment, analytical determination, data processing, statistical analysis, metabolite identification, and biological interpretation (Fig. 1). The analytical techniques in metabolomic studies involve nuclear magnetic resonance spectroscopy (NMR) and mass spectrometry (MS), usually preceded by a chromatographic separation step. These hyphenated techniques include gas chromatography (GC), liquid chromatography (LC), and capillary electrophoresis (CE) coupled with mass spectrometry (MS). GC-MS has long been used for metabolic profiling due to its high sensitivity, reliability, and the

Fig. 1. Workflow in nontargeted metabolomics.

availability of extensive databases facilitating metabolite identification (12). GC-MS is not suitable for nonvolatile, thermolabile, and/or highly polar compounds; therefore, derivatization of metabolites is often needed to yield volatile and thermostable analytes. GC-MS provides separation of low molecular weight metabolites which includes a range of relatively polar metabolite classes, such as amino acids, organic acids, amines, amides, and sugars. LC-MS can separate and identify highly polar, thermally labile, and/or high molecular weight mixture compounds. Because LC-MS does not require derivatization, sample preparation is simple compared with GC-MS (14). Moreover, LC is a flexible technique; many modes of separation are available. Although LC is inferior to GC in terms of separation efficiency, technological advancements in LC have resulted in improved performance. The introduction of ultra (high) performance liquid chromatography (U(H)PLC) substantially increased the available chromatographic resolution and the number of metabolites detected compared with traditional LC (15). Coupling UPLC to mass spectrometers of high mass resolution and high mass accuracy provides a tool for putative metabolite identification. Capillary electrophoresis is particularly suited for the separation of polar and charged compounds. The orthogonality in separation principle underscores the relevance of CE as a complementary tool to the more established chromatographic techniques; in many cases, samples that cannot be easily resolved by GC or LC can be separated by CE. The use of CE-MS has increased considerably over the

NOVEL METHODS IN ASSESSING UREMIA

past years, but is still rather limited compared to the other analytical techniques (16,17). Poor concentration sensitivity and reduced migration time reproducibility most probably still hamper the wider use of CE-MS for metabolomic studies. Nevertheless, fast, highly efficient separation, without requiring rigorous sample pretreatment, can be obtained. NMR provides a rapid, nondestructive, highthroughput method and can offer direct identification and quantification of a range of abundant analytes (12). The technique is usually more robust and reproducible than LC-MS and can allow the utilization of libraries and interlaboratory validation studies. Compared with MS, NMR is less sensitive and requires more expensive instrumentation. Technological improvements in instrument sensitivity and accuracy allow identification of metabolites by far higher confidence and at levels that were undetectable only a few years ago. Nevertheless, the process of metabolite identification in MS-based nontargeted metabolome studies is still a bottleneck in deriving biological knowledge from data. In the near future, developments such as the extension of databases for metabolite identification, synthesis of metabolite standards, and improvements in dataprocessing software are expected to facilitate metabolite identification. After identification of solutes, the understanding and interpretation of metabolite data in terms of the underlying biochemical mechanisms is essential to fathom their patho-physiologic importance. An obvious step is to map and visualize the metabolic pathways and other general biological networks affected by the newly detected metabolites, i.e., graphically representing the relationships among metabolites and/or enzymes (18). This will make it possible to assess whether the identified metabolites are involved in the same biological pathway or if they are close to each other in the metabolic network. For this purpose several software applications are available like KEGG (Kyoto Enclycopedia of Genes and Genomes) and MetaCyc (database of metabolic pathways and enzymes). This approach can help to select metabolites of greater interest for evaluation of pathophysiological effects for which targeted quantitative analytical methods should be developed since nontargeted metabolome studies are semi-quantitative at best, providing relative differences instead of absolute concentrations. If concentrations applied in further toxicity tests exceed those encountered in uremia, conclusions might have relatively little clinical relevance (1). Therefore, quantification of the confidently identified metabolites of interest should be performed by targeted methods before toxicity testing becomes possible, and then only provided that authentic standards are available. Assessment of the pathophysiologic role of these newly detected metabolites will enable identification of novel key culprits for the uremic syndrome as a first step to pursue their specific removal.

3

In summary, both proteomics and metabolomics have emerged as potentially useful strategies in the search for uremic retention solutes. Improvements in technology and lessons from genomics, transcriptomics, and proteomics have fueled the pace of discovery in metabolomics, the youngest of the -omics fields. In the future, we expect further developments such as the extension of databases for metabolite identification, synthesis of metabolite standards, and improvements in data processing software, facilitating the identification of to date unknown uremic retention solutes. Identification data complemented with absolute concentrations obtained by targeted methods will offer valuable input for further uremic toxicity research. Biological Relevance of Uremic Retention Solutes Once uremic concentrations of the newly detected and identified metabolites are available, metabolites can be evaluated for their biological activity in in vitro, ex vivo, and/or in vivo experimental settings. Selected and pathophysiologically relevant parameters involved in the highlighted pathways should be evaluated to search for mechanisms that impact survival, hospitalization and/or quality of life of patients. In the context of CKD these are mainly mechanisms involved in (a) progression of CKD; (b) inflammation; (c) nutritional status, and (d) cardiovascular disease. Effects on functional activities of different cell types/systems/organs involved in the above comorbidities, reflected by specific parameters, can be evaluated in in vitro/ex vivo, or in vivo experimental settings. Many exemplary types of test methods evaluating the effect of the uremic milieu or specific retention solutes in the context of CKD progression and the increased risk for cardiovascular disease can be found in the literature as summarized in Tables 1 and 2 respectively (19–44). As shown in the table, many of these effects are closely linked to oxidative stress and inflammation. The European Uremic Toxin (EUTox) work group proposed in 2007 a standardized approach for testing the biological effect of uremic retention solutes using relevant concentrations and appropriate control conditions, taking into account, especially for protein-bound solutes, the albumin content of the test medium, and excluding confounding factors like contamination by bacterial derivatives (45). Unfortunately, still as of today, in many publications these elementary rules are not satisfied. In such a situation, the application of systematic search methods as applied in evidence based medicine and guideline development may be useful (46). The ultimate salient examples are indoxyl sulfate and p-cresylsulfate, both protein-bound uremic toxins for which a host of toxic effects have been demonstrated. Recently, a systematic review procedure

4

Vanholder et al. TABLE 1. Functional evaluation of renal tubular cells and kidneys for effects on progression of CKD

Cell type in animal experiments Renal proximal tubular cells

Functional evaluations Inflammation Inflammation/fibrosis

Oxidative stress

Cell adhesion/oxidative stress Aging Epithelial-to-mesenchymal transition

Organ in animal experiments Kidneys

Fibrosis Fibrosis/Epithelial-tomesenchymal transition Fibrosis/Renin-AngiotensinAldosterone system Aging Oxidative stress

Cell adhesion/oxidative stress Tubular secretion

Parameters Expression of cytokines (Tgfb1, Fasl, Il6/15, Il15, Csf1/3 and Cxcl10); intracellular signaling (Stats, Smads, Nfkb2, Ikbkb, Bcl2 and Bax connected to Tgfb 1) (PCR array) (20) mRNA expression of transforming growth factor (TGF)-beta1, tissue inhibitor of metalloproteases-1, and Pro-a1 (I) collagen and active TGF-beta1 protein secretion (21) qPCR and western blotting for studying p53-TGF-beta1 -Smad3 signaling (22) Expressions of nuclear factor (erythroid-derived 2)-like 2 (Nrf2) through activation of Nuclear Factor-jB (23) MCP-1 expression through production of Reactive oxygen species (ROS) and activation of NF-jB, p53, ERK, and JNK (24) mRNA and protein expression of p22phox and Nox4 (21) qPCR for renal expression of ICAM-1 through production of ROS such as superoxide and activation of NF-jB and p53 (24) DNA methyltransferase expression - klotho expression (25) RT-PCR and immunoblotting of the epithelial markers E-cadherin and zonula occludens-1 (ZO-1) and the mesenchymal marker a-smooth muscle actin (a-SMA) (26,27) Renin, angiotensinogen, angiotensin 1 and 2 (AT1 en 2) receptor expression (27) Masson’s trichrome staining (21,25,28) Immunostaining of TGF-beta1 (28) and Smad3 positive areas (22) Idem above (26,27) Expression of the EMT-associated transcription factor Snail (27) Renin, angiotensinogen, angiotensin 1 and 2 (AT1and 2) receptor expression; TGF-beta pathway (27) Idem above (25) Immunostaining for expression of Nrf2 and its downstream target genes, heme oxygenase-1 (HO-1) and NAD(P)H:quinone oxidoreductase 1 (NQO1) and expression of 8-hydroxydeoxyguanosine (8-OHdG), a marker of ROS (21,23) Immunohistochemistry and mRNA expression of MCP-1 (24) Immunoblotting for renal expression of ICAM-1 (24) Effect on expression of organic anion transporters (29); Effect on function of tubular efflux transporters (30)

qPCR, quantitative polymerase chain reaction; ICAM-1, intercellular adhesion molecule; MCP-1, monocyte endothelial chemotactic protein-1; ERK, Extracellular-regulated kinase; JNK, C-Jun N-terminal kinase

was used for exploring the methodologically sound studies unbiased by erroneous conditions related to solute or albumin concentrations. This approach brought down an original number of 397 retrieved candidate studies to 27, whereby the finally yielded selection allowed to confirm definitely the toxicity of especially indoxyl sulfate but also of p-cresylsulfate, supporting their role in vascular and renal disease progression (47). In addition, the selected papers were also scored for their quality, based on the following criteria: six or more experiments, confirmation by more than one experimental approach, neutralization of the biologic effect by counteractive reagents or antibodies, use of a real-life model, and use of dose–response analyses in vitro and/or animal studies (47). In summary, next to the relevance of the tested concentrations, qualitative criteria related to the studies considered should also be taken into account when evaluating biological effects of uremic metabolites. In addition, the presumed effects should be cross-checked in clinical studies, either observational, or, if possible, controlled (see below). This approach will identify the most important uremic

retention solutes or groups of solutes for which intervention strategies or specific removal strategies might be considered, preferably to be used in early CKD. Of note, at this moment most removal strategies are not selective, and as long as such a situation prevails, it will remain impossible to prove a causal relationship between specific uremic toxins and clinical outcomes. Nondialysis Removal of Uremic Retention Solutes It may reasonably well be assumed that the strength of the biological (toxic) effect of uremic retention solutes is related to their concentration, which is affected not only by dialysis, but also by nondialysis factors that should necessarily be considered to fully understand solute mechanistics in end stage renal disease (ESRD). Endogenous metabolism generates many uremic solutes (48), especially the small water soluble compounds, such as the guanidino compounds or the purines, and the middle molecules, a group mostly

5

NOVEL METHODS IN ASSESSING UREMIA

TABLE 2. : Functional evaluation of cell types, systems, and organs involved in uremia related cardiovascular complications Cell types Endothelial cells

Functional evaluations Cell proliferation and repair Cell dysfunction Oxidative stress Cell senescence Cytotoxicity Thrombotic state

Smooth muscle cells

Cell proliferation Injury-related thrombosis Cytotoxicty

Parameters 3

5-bromo-2-deoxy-uridine (31,32); ( H)-thymidine incorporation (33) Microparticle release (34) Reactive oxygen species (ROS); NADPH-oxidase; Nitric Oxide-synthase; xanthine oxidase (35) Senescence associated-b-galactosidase (33) Lactate dehydrogenase (LDH) release (32) Tissue factor and aryl hydrocarbon receptor pathway (36) 4-(3-(4-lodophenyl)-2-4(4-nitrophenyl)-2H-5-tetrazolio-1,3-benzene disulfonate) assay; (3H)-thymidine incorporation (37) Tissue factor (38) LDH release; ATP-assay (32)

Leukocytes

Oxidative stress Cell mobility Inflammation

ROS (39) Chemotaxis (40); surface molecule expression (32) Tumor Necrosis Factor-alpha, Interleukin (IL)-6; IL-1beta mRNA, Mitogen-activated Protein Kinase (MAPK) and Nuclear Factor (NF)-jB pathway (41)

Cardiac fibroblasts/ myocytes

Hypertrophy Fibrosis

3

Adipocytes/myotubes

Insulin resistance

Insulin-stimulated glucose uptake (incorporation of tritiated 2-deoxy-glucose uptake), nsulin signaling pathways, cellular characteristics (size, count, DNA content) Incorporation of (14C)-acetate Glycerol released in presence or absence of isoproterenol (19)

Lipogenesis Lipolysis Cell systems in animal experiments Vessels Aortic calcification and stiffness Cell senescence

Leukocytes-endothelial cell cross-talk

Cell activation/mobility Glycocalyx integrity

Organs in animal experiments Heart Cardiac fibrosis

White adipose tissue and various organs

H-leucine incorporation, MAPK and NF-jB pathway (41) Collagen synthesis, 3H-proline incorporation, MAPK and NF-jB pathway (41)

Von Kossa staining; Expression of osteoblast specific proteins (28) Immunohistochemistry of senescence-associated b-galactosidase (SA-b-gal), and senescence-related proteins such as p16INK4a, p21WAF1/CIP1, p53 and retinoblastoma protein (Rb) (42) Rolling/Adhesion/Extravasation (43) Heparan sulfate (43)

Oxidative stress

Masson’s trichrome (MT); Transforming growth factor (TGF)-b1, a-smooth muscle actin (SMA), type 1 collagen (44) NADPH oxidase Nox 4, malondialdehyde (MDA), and 8-hydroxydeoxyguanosine (8-OHdG) and decreased staining for nuclear factor (erythroid-derived 2)-like 2 (Nrf2) and heme oxygenase-1 (HO-1) (44)

Insuline resistance Ectopic lipid redistribution

Metabolic profiling, Insulin signaling pathways (Phosphorylation of Protein Kinase B/Akt and Extracellular Regulated Kinase 1/2) Lipid content (19)

constituted of peptides (1). Some, like parathyroid hormone, are actively secreted and therefore not exclusively depending on retention and removal (49). The predominant role of the intestine in generation of some uremic solutes has been highlighted only recently (48,50). Some molecules, like the advanced glycation end products (AGEs), are present in food and are absorbed unmodified (48). Many other molecules, especially protein-bound solutes (51) but also volatile compounds (52), are metabolites of which the primary building stones are generated by the intestinal microbiota from natural breakdown products of digestion (48,50). They are then further transformed by the intestinal wall or liver via conjugation, a primary detoxification process that nevertheless gives rise to bioactive compounds (47): tyrosine is modified by the intestinal

microbiota into p-cresol, to be further metabolized in the body to p-cresylsulfate and p-cresylglucuronide (53–55). The key role of the intestine in clearing uremic toxins has recently been stressed by metabolomic studies. When dialysis patients with and without intact colons were compared, the concentration of several solutes was markedly lower in the latter (56). In rats given the intestinal sorbent AST-120 (KremezinR, Kureha Corporation, Tokyo, Japan) or not, again marked differences were found for a large array of compounds (57). In addition, the composition of intestinal microbiota is modified by uremia (58), resulting in alterations of toxin generation (assimilation disturbances) (59). Finally, removal is not a question of dialysis alone. One alternative pathway is via bile secretion,

6

Vanholder et al.

whereby molecules may undergo enterohepatic recycling (60), but the most obvious mechanism is through residual renal excretion, which even in patients on dialysis plays a key role (61) and relates to survival (62). For several solutes, residual renal function together with nutritional intake has a more significant impact on solute concentration than hemodialysis adequacy estimated by Kt/Vurea (urea clearance multiplied by dialysis time and normalized for distribution volume) (63). Although GFR or its estimated equivalent (eGFR) are currently considered as the most stringent parameters of kidney function, eGFR correlates deceivingly with concentration of a host of uremic toxins (64–66). For several of these, renal tubules are more important determinants of removal than the glomeruli (67), by involvement of tubular pump systems (68,69). All these data together indicate that removal of uremic solutes by dialysis is only one way to decrease concentration of uremic retention solutes, and that it may be relevant to influence concentration not only by extracorporeal removal strategies, but also by trying to affect intestinal uptake, metabolism or renal clearance. Dialysis Removal and Kinetics of Uremic Retention Solutes Uremic toxins are generated by metabolic processes throughout the body and their removal during renal replacement therapy, as well as by the kidneys, depends on equilibration and redistribution between different body compartments during blood purification. Hence, for purposes of describing solute mass transfer within the body, the body can be considered to consist of different compartments separated by semi-permeable membranes, with transport between the compartments occurring by both passive and active transport mechanisms, including diffusion, carrier pumps, and channels/transporters. In parallel kinetic models (Fig. 2A (70)), organs are allocated to either a high flow system (comprising kidneys, heart, brain, portal system, lungs, and

A

blood volume, containing only 21% of total body water but perfused by 80% of the cardiac output) or a low-flow system (comprising muscles, bones, skin, and fat, containing 79% of total body water, but perfused by only 20% of cardiac output) (71– 73). Each system can be further subdivided into an intra- and extracellular volume with for each solute a specific transport rate between compartments. In this serial chain of diffusive and/or convective transport processes, overall solute removal is determined by the slowest process. Serial kinetic models (Fig. 2B) consist of one or more interconnected compartments, characterized by a volume and a homogenous solute concentration, both of which change in time as a result of different transport processes, including exchange between adjoining compartments, and solute generation and metabolic elimination within the compartment. For each compartment, mass balance can be expressed in terms of an ordinary differential equation. The solution to the differential equation(s) describes the compartmental concentration as function of time and the model parameters. Since kinetic models are purely mathematical, model parameters can be nonphysiological. Independent of model structure, use of a model requires determination of the model parameters from the literature or adjustment of the model parameters so that the output fits experimental data (Fig. 3). If a satisfactory fit cannot be obtained, the model may need to be changed by, for example, adding an additional compartment or mass transfer process. Once a calibrated kinetic model is available for a given solute, it can be used to predict intraand interdialytic solute concentrations for different dialysis strategies (Fig. 3). Studies of adequacy of dialysis using kinetic methods should include the frequency and duration of treatment. In addition, to achieve the therapeutic goal on an individual basis, a number of dialyzer characteristics and the blood and dialysis fluid flow rates through that dialyzer should be entered into the model. For convective therapies, the convection volume must also be selected. Delivery of the cho-

B G2 E 2

G1 E 1

K21 12

V2,C2

V1,C1

KD KND

V Fig. 2. (A) The regional blood flow model (70). He: extracellular high flow system; Hi: intracellular high flow system; Le: extracellular low flow system; Li: intracellular low flow system; QL: low flow; QH: high flow; Qc: cardiac output; Qf: fistula flow; Qx: extracorporeal blood flow. (B) Classic serial two-compartment model. V1: plasmatic volume; V2: nonplasmatic volume; V: total distribution volume; C1: plasmatic concentration; C2: nonplasmatic concentration; G1 and G2: generation rate in V1 and V2; E1 and E2: metabolic elimination in V1 and V2; KD: dialysis clearance; KND: nondialysis clearance; K12: intercompartmental clearance.

7

NOVEL METHODS IN ASSESSING UREMIA

concentration

In vivo blood and dialysate sampling

4

0,08

3

0,06

2

0,04

1

0,02

0

0

0

60

120

180

240

time (min)

Calibration of the kinetic model by fitting on the in vivo data

4

2

1

0

0

60

120

180

240

0,08

Cpredialysis

3

concentration

concentration

4

In vivo validation of the kinetic model: Measuring TAC, TSR, C predialysis in patients on different dialysis strategies

3

0,06

TSR=AUC 2

TAC

1

0,02

0

time (min)

0,04

0

0

60

120

180

240

time (min)

Simulations with calibrated kinetic model Calculation of TAC, TSR, C predialysis

20 18

Cpre dialysis

16

TAC

14 12

C1

10

C2

8 6 4 0

2000

4000

6000

8000

10000

time (min)

3x4h/week dialysis

Short daily (6x2h/week)

Fig. 3. Flow chart of the definition, calibration and validation of a kinetic model to perform simulations of different dialysis strategies. Cpredialysis, predialysis blood concentration; TAC, time-averaged concentration; TSR, total solute removal; AUC, area under the curve.

sen prescription then determines the dose of dialysis, which can be thought of as the amount of uremic solutes removed during a dialysis session. Hemodialysis dose is traditionally expressed in terms of single-pool Kt/Vurea, which was introduced in 1985. With growing interest in increasing the frequency of dialysis Gotch introduced in 1998 the standard Kt/V, which expresses the weekly dose of therapy (74). Increasing emphasis on removing solutes other than urea has led to more general expressions of dialysis dose, such as the equivalent renal clearance (EKR), defined as generation rate divided by the time-averaged concentration (TAC), which provides a measure of clearance on a continuous

basis. To provide a measure of total body clearance of a solute, the reduction ratio (i.e., change in concentration of a solute from pre- to postdialysis) is a commonly used metric. For larger solutes and protein-bound solutes with a sieving coefficient