Membranes and Filters for Use in Acute Renal Failure

Overview on Devices and Membranes

The contemporary design of hollow-fiber dialyzers consists of a single fiber bundle contained in a housing made of biocompatible materials (e.g., polycarbonate, polyethylene). The bundle is encapsulated (“potted”) at both ends with a silicone material before being sterilized and put in the housing. Two caps cover the end surfaces of the bundle at both sides. The housing contains inlet and outlet ports for blood (directly on the housing or on the end caps) and one or more additional ports for the effluent compartment, depending on the mode for which the filter is conceived. Based on a specific manufacturer's approach, blood flow enters the filter in parallel with respect to the fiber bundle. The design, size, and geometric characteristics of the fiber bundle are the primary determinants of the performance characteristics for the entire filter.

Because of the size of the global market, most filter development activities have occurred for chronic hemodialysis therapy. Different bundle configurations have been developed in the past to maximize treatment efficiency, including rectangular block arrangements, cross-flow configurations, multiple bundles, spiral fibers surrounding a central core, and warp-knitted hollow-fiber mats. Today, the most common configuration is the Moiré structure, in which an undulation of the fiber bundle with a specific periodicity is applied.

The major geometric characteristics of hollow fibers are length, mean inner radius, and wall thickness. The porosity of the whole bundle, an important determinant of diffusive solute removal in conventional hemodialysis, is determined by the pore density (number of pores/unit surface area) in each fiber multiplied by the total number of fibers in the bundle.

This and other membrane properties influence several filter characteristics, which are important considerations when prescribing a certain therapy, including surface area, filter priming volume (volume of blood compartment), and total priming volume (sum of volumes of blood and effluent compartments).

Membrane Materials

The most important parameter determining the chemical and physical behavior of a membrane is the material of which it is composed. A wide spectrum of filters together with a multitude of different membrane materials are currently available on the market.

In Fig. 140.4 the main materials that make up membranes used for dialysis are summarized. Dialysis membranes can be classified according to their chemical structure. Natural and synthetic polymers are used currently worldwide for this application because of their characteristics of chemical resistance, sterilizability, industrial processing, and biocompatibility.

Natural polymeric membranes can be further subclassified as unmodified cellulose based and modified cellulose based. In the latter category, hydroxyl groups (–OH) have been reduced by substitution (e.g., cellulose acetate) or by bonding of synthetic materials to the cellulose. In chronic hemodialysis, the use of unmodified cellulosic membranes now is exceedingly rare and the use of synthetic membranes continues to increase.

Membranes applied for CRRT are almost exclusively synthetic. Synthetic polymers for membranes are classified as thermoplastics. This class of materials results in less complement activation with respect to natural polymers such as cellulose because of their hydrophobic nature. Almost all synthetic polymers currently used for CRRT are hydrophobic: the exception is AN69. Hydrophobic membranes in general are relatively biocompatible but typically require a hydrophilic pore-enlarging agent (e.g., polyvinylpyrrolidone, PVP) to achieve the desired permeability.

In general, synthetic polymeric membranes have an asymmetric structure ( Fig. 140.5 ); again the exception here is AN69. A typical asymmetric membrane wall is at least 20 µm in thickness and consists of a thin (~1 µm) inner “skin” layer in contact with the blood compartment. This layer is the primary determinant of the solute removal properties for the membrane. The remainder of the membrane wall is characterized by a much thicker spongy region, with interstices that cover a wide size range, as determined by the manufacturing process and the polymer composition.

Contemporary synthetic membranes can be manufactured with a relatively sharp curve of pore size distribution, allowing closer simulation of the filtration provided by the native kidney and greater removal of middle-high molecular weight molecules. Modern high-flux membranes, in the “virgin” state, allow filtration of molecules reaching almost 50 to 60 kDa while restricting the passage of albumin to a great extent.

Device Performance

Solute transport across membranes occurs by two different mechanisms: diffusion and convection. In addition, there also can be an interaction between the solutes and the membrane surface, leading to their binding (adsorption).

The relative contributions to total solute transport from diffusion and convection are determined by the treatment modality. For example, in hemodialysis the dominant mode of solute transport is diffusion, whereas convection is the prevailing mechanism in hemofiltration. Adsorption occurs with all treatment modalities and can be considered either as a negative attribute (membrane fouling leading to a reduction in transmembrane transport) or a positive attribute (removal of low-molecular proteins or peptides).

Mathematically, diffusive solute transport across the membrane is governed by Fick's law, depending largely on the concentration gradient of the solute between the blood compartment and the dialysate side. On the other hand, convective transport is transmembrane solute movement resulting from the bulk movement of solvent in the presence of a pressure gradient (“solvent drag”). Adsorption reflects the interaction at a molecular level between the material surface and the molecule; it is determined by the hydrostatic forces and the nature of the compound. It can be quantified by either the Freundlich equation or the more commonly used Langmuir equation for monolayer adsorption on a surface. These equations are based on a number of assumptions, including (1) adsorption can only occur at a fixed number of definite localized sites; (2) each site can hold only one molecule; (3) all sites are equivalent with no interaction between adsorbed molecules.