Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Solute and Water Transport in Hemodialysis : Dialyzers, Flow Distribution, and Cross-Filtration
Objectives
This chapter will:
- 1.
Discuss the determinants of the two main dialysis mass transfer mechanisms, diffusion and convection, along with the factors influencing ultrafiltration.
- 2.
Explain the concept of solute clearance and the various ways in which it is expressed.
- 3.
Describe the basic characteristics of hollow fiber dialyzers and highlight the major features influencing ultrafiltration properties and small and larger solute removal capabilities.
Conventional hemodialysis remains an important renal replacement modality for critically ill patients with acute kidney injury. Because prescription of hemodialysis requires establishing goals for the rate and extent of solute and fluid removal, an understanding of the mechanisms by which solutes and fluid are removed during hemodialysis is necessary. This chapter provides an overview of basic mechanisms of solute and fluid transfer during conventional hemodialysis. The major characteristics of hollow fiber membranes influencing solute and water removal are discussed. Within this section, the chemical composition and physical characteristics of commonly used dialysis membranes and the features determining their solute and water permeability properties are reviewed. Flow distribution inside the dialyzer, internal filtration, and backfiltration phenomena are discussed in the last part of the chapter.
Diffusion and Convection
The basic physical mechanisms leading to removal of solute and water through a semipermeable membrane have been discussed already in other sections of this book. Diffusion is the dominant mass transfer mechanism mediating small solute removal in conventional hemodialysis.
Diffusion is a process in which molecules randomly move in all directions. Statistically this phenomenon results in net movement of solutes from a more concentrated area to a less concentrated one. In addition to the concentration gradient (dc), the solute diffusive flux per unit of area (J d ) through a semipermeable membrane depends directly on the diffusivity (D) of the solute (which is a function of temperature, viscosity of the fluid, and an approximate solute radius) and is inversely proportional to the membrane thickness (dx), as shown by the following equation:
On the contrary, convection is related to ultrafiltration of plasma water and involves solute transfer through fluid movement in response to a transmembrane pressure gradient based on a process termed “solvent drag.” Therefore the solute convective flux (removal rate per unit area, J c ) depends on the ultrafiltration flux (ultrafiltration rate per unit area, Q UF ), the solute concentration in plasma water (C Pi ), and the solute sieving coefficient (SC), as shown in the following equation:
These definitions of diffusion and convection (together with ultrafiltration) imply the two phenomena are separate. In fact, since the dawn of dialysis, they have been combined in an attempt to replace renal function. The knowledge of diffusion came from industrial chemistry, and dialyzers were designed to be ideal countercurrent exchangers, whereas the potential clinical advantages of convection were recognized later. In current clinical practice, the combined effect of diffusion and convection is exploited commonly. Although it is impossible to define precisely the contribution of these individual processes in the removal of solutes because of their continuous interactions, this principle applies not only to hemodiafiltration but also to standard high-flux hemodialysis.
Blood flow greatly affects the clearance of small solutes such as urea, whereas the influence of ultrafiltration rate is relatively greater for the removal of larger solutes. An increase in dialysate flow rate becomes important only with large surface area dialyzers and mostly affects the clearance of small solutes. In addition to the above aspects related to modality and flow rates, the type of membrane used and the hydraulic conditions within the hemodialyzer also must be considered.
Membranes and Filters for Intermittent Hemodialysis
Membranes used in dialysis are of natural or synthetic origin. Table 150.1 shows a simple but comprehensive comparison between membrane properties in these two classes. Different membranes have been generated from numerous basic materials and have been used subsequently in extracorporeal therapy over several decades. Table 150.2 presents an overview of existing membrane materials with the defining characteristics for each.
TABLE 150.1
Basic Comparison Between Classes of Membranes for Renal Replacement Therapy
| PARAMETER | NATURAL | SYNTHETIC |
|---|---|---|
| Structure | Homogeneous | Mainly asymmetric |
| Porosity | Hydrogel | Microporous |
| Interaction with water | Hydrophilic | Hydrophobic |
| Thickness | Small | Large |
| Biocompatibility | Low | High |
| Electrical charges | Mixed | Negative |
| Hydraulic permeability | Low-flux | High-flux |
TABLE 150.2
Membrane Materials for Renal Replacement Therapy
Modified from Saito A KH, Yamashita AC, Mineshima M. High-Performance Membrane Dialyzers. Basel: Karger; 2011.
| MATERIAL | MEMBRANE CHARACTER |
|---|---|
| Cellulosic | |
| Cellulose | |
| Cellulose diacetate | Hydroxyl groups replaced with acetate |
| Cellulose triacetate (CTA) | Hydroxyl groups replaced with acetate |
| Hemophane | Hydroxyl groups replaced with diethylaminoethyl radicals |
| Synthetic Membranes | |
| Ethylvinyl alcohol (EVAL) | Hydrophilic |
| Polysulfone (PS) | Hydrophobic |
| Polymethylmethacrylate (PMMA) | Hydrophobic |
| Polyacrylonitrile (PAN) Polyamide (PA) |
Hydrophobic Hydrophobic |
| Polyethersulfone (PES) Polyarylethersulfone (PAES) Polyesther polymer alloy (PEPA) |
Hydrophobic Hydrophobic Hydrophobic |
An obvious difference between synthetic and cellulosic membranes is chemical composition. Unlike naturally occurring cellulose membranes, synthetic membranes are manufactured polymers that are classified as thermoplastics. As reported previously in Table 150.1 , another feature differentiating cellulosic and synthetic membranes is wall thickness ( Fig. 150.1 ). Synthetic membranes have wall thickness values of at least 20 µm and may be structurally symmetric (e.g., AN69, PMMA) or asymmetric (e.g., polysulfone, polyamide, polyethersulfone, polyarylethersulfone/polyamide). For asymmetric structures, a very thin “skin” (approximately 1 µm) contacting the blood compartment lumen acts primarily as the membrane's separating element with regard to solute removal. The structure of the remaining wall thickness (“stroma”), which determines the thermal, chemical, and mechanical properties, varies considerably among the different synthetic membranes.
Cellulosic Membranes
The relatively long duration of popularity of cellulosic membranes can be explained largely by their particular suitability for a diffusion-based procedure such as hemodialysis. The underlying hydrogel structure of these membranes and their tensile strength allow the combination of thinner walls (from 5 to 15 µm Fig. 150.2 ) and high porosity to be achieved in the fiber spinning process. These characteristics allow the attainment of high rates of diffusive membrane transport and efficient removal of small, water-soluble uremic solutes, such as urea and creatinine. Another characteristic feature of these membranes is symmetry with respect to composition, implying an essentially uniform resistance to mass transfer over the entire wall thickness.
The most commonly used cellulosic dialyzers contain cellulose acetate (rigorously, cellulose diacetate) membranes, in which approximately 75% of the hydroxyl groups on the cellulosic backbone are replaced with an acetate group. As compared with a hydroxyl group, an acetate group does not bind avidly to a C3 molecule to initiate activation of the complement cascade. Consequently, in dialysis using cellulose acetate membranes, complement activation is attenuated, as is the leukopenic response, in comparison with dialysis using unmodified cellulosic membranes. Because production of cellulose triacetate membranes involves complete hydroxyl group substitution with acetate groups, further attenuation of complement activation and leukopenia is achieved.
Synthetic Membranes
Synthetic membranes were developed essentially in response to concerns about the narrow scope of solute removal and the pronounced complement activation associated with unmodified cellulosic dialyzers. The AN69 membrane, a copolymer of acrylonitrile and an anionic sulfonate group, was employed first in flat sheet form in a closed-loop dialysate system in the early 1970s. Since that time, a number of other synthetic membranes have been developed, including polysulfone, polyamide, polymethylmethacrylate (PMMA), polyethersulfone, and polyarylethersulfone/polyamide.
Largely in relation to the interest in hemofiltration as a therapy for end-stage renal disease, along with the inability to use low-flux unmodified cellulosic dialyzers for this therapy, synthetic membranes initially were formulated with high water permeability. The large mean pore size and thick wall structure of these membranes allowed the high ultrafiltration rates necessary in hemofiltration to be achieved at relatively low transmembrane pressures.
However, dialyzers with these highly permeable membranes were used subsequently in the diffusive mode as high-flux dialyzers. This latter mode continues to be the most common application of these membranes, although they increasingly are being employed for long-term hemodiafiltration.
Properties of Hemodialyzer Membranes That Influence Dialyzer Performance
Although not strictly correct, hollow-fiber dialyzer membrane function can be approximated with a model having straight cylindric pores, all of the same radius (r) and all with a directionality perpendicular to the flow of blood and dialysate.
The major determinants of plasma ultrafiltrate flow rate through the pores are the number of pores (i.e., number per unit area of membrane surface area), transmembrane pressure, and pore size. With regard to pore size, the rate of ultrafiltrate flow depends on the fourth power of the pore radius (r 4 ), consistent with application of the Hagen-Poiseuille equation to an individual pore. Mean pore size also directly influences water permeability.
Membrane wall thickness is one important determinant of diffusive transport. The relatively thin-walled structure of cellulosic membranes (usually 5–15 µm) is largely responsible for their particular suitability in the setting of diffusive hemodialysis. The other major determinant of diffusive transport is porosity, also known as pore density. Membrane porosity is directly proportional to the number of pores and the square of the pore radius (r 2 ). Therefore the smaller dependence of membrane porosity on pore size, relative to the case of water permeability, implies a relatively greater importance of pore number in determining diffusive permeability.
In fact, flux and diffusive permeability can be independent of each other for a particular hemodialysis membrane, because of their differing major determinants ( r 4 for the former and number of pores, r 2 and wall thickness for the latter). Such is the case for cellulosic high-efficiency dialyzers, which typically have very high diffusive permeability values for small solutes but low water permeability.
A membrane represented by the cylindric pore model previously described deviates from an actual membrane used for clinical hemodialysis, in that the latter actually has a distribution of pore sizes. Ronco et al. have discussed the manner in which pore size distribution may differ among hemodialysis membranes and the resultant influence on a membrane's sieving properties. In Fig. 150.3 , which has been reproduced from their study, the membrane represented by curve A on the left diagram has a large number of relatively small pores, whereas the membrane represented by curve B has a large number of relatively large pores. On the basis of the relatively narrow pore size distributions, the solute sieving coefficient versus molecular weight profiles for both membranes (right diagram) have the desirable sharp cutoff, similar to that of the native kidney. However, the molecular weight cutoff for membrane A (approximately 10 kDa) is consistent with a high-efficiency membrane (high diffusive permeability but low hydraulic permeability), whereas that of membrane B (approximately 60 kDa) is consistent with a high-flux membrane (membrane ultrafiltration coefficient K UF >25 mL/hr/mm Hg/m 2 ). In addition, primarily because of the large number of pores, both membranes would be expected to demonstrate favorable diffusive transport properties. On the other hand, the membrane represented by curve C exhibits a pore size distribution that is unfavorable from a diffusive transport and sieving perspective. The relatively small number of pores accounts for the poor diffusive properties. In addition, the broad distribution of pores explains not only the “early” drop-off in sieving coefficient at relatively low molecular weight but also the “tail” effect at high molecular weight (right diagram). This latter phenomenon is highly undesirable, because it may lead to unacceptably high albumin losses across the membrane. In practice, all highly permeable membranes have measurable albumin sieving coefficient values so that the design of this type of membrane involves striking a balance between optimized removal of high-molecular-weight toxins and minimal loss of albumin.
As suggested earlier, the most common classification scheme for membranes used in hemodialysis traditionally has included low-flux, high-efficiency, and high-flux groups. High cutoff membranes are the most recent addition to this scheme. Although these membranes are used commonly in CRRT, they also have been employed in hemodialysis, most commonly for patients with myeloma-associated AKI (“cast nephropathy”). In the virgin state, these membranes may allow passage of molecules as large as approximately 300 kDa, thus providing significant clearance of free light chains. Although the effective molecular weight cutoff is much lower after blood exposure, relatively substantial albumin loss (as much as 30 g per treatment) still occurs with use of these membranes. Thus a risk/benefit determination is important when these membranes are employed for myeloma-associated AKI or other disorders.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here