See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/322690500

Dialysis Membranes Today

Article  in  The International journal of artificial organs · May 2002

DOI: 10.1177/039139880202500516

CITATIONS READS
23 1,690

1 author:

Sudhir K Bowry
Dialysis Advisory
70 PUBLICATIONS   1,011 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dialysis Therapy - Delivery, Outcomes & Payability View project

All content following this page was uploaded by Sudhir K Bowry on 10 February 2018.

The user has requested enhancement of the downloaded file.

 The International Journal of Artificial Organs / Vol. 25 / no. 5, 2002 / pp. 447-460

Biomaterials and their Application

Dialysis membranes today

S. K. BOWRY

Fresenius Medical Care, Bad Homburg - Germany

ABSTRACT: In recent years, hemodialytic therapies have evolved from the simple, diffusion-
dependent removal of small molecular weight substances from blood to advanced therapy
modalities involving the convective removal of larger uremic sloutes. The clinical benefits of
removal of substances such as ß2-microglobulin (ß2-m) have been reported by several authors:
elimination of large-molecular weight “uremic toxins” is now widely accepted as being beneficial
to the overall quality of life of patients.
This trend would not have been possible without parallel technical developments, especially that
of new membranes having more open pore structures resulting in higher sieving coefficients and
increased hydraulic permeability. Not all polymer types are suitable for the manufacture of high-
flux membranes required for convective therapies in which large fluid volumes are exchanged.
Amongst the more important criteria are: the selected polymer must be able to undergo steam
sterilisation, have high endotoxin retention capabilities, be versatile for the fabrication of a range
of hydraulic permeabilities and, of course, have high blood compatibility.
The aim of this paper is, firstly, to review the major membrane development phases over the last
quarter of a century. Secondly, the suitability of current membrane materials to meet the
aforementioned requirements will be examined. Thirdly, in view of the recent, rapid proliferation of
polysulfone-based membranes, dialysis membranes of the polysulfone 'family' are placed under
scrutiny; membranes of this class represent a significant portion of the product portfolio of dialyser
manufacturers today, yet, few end-users are able to distinguish between the salient features of the
respective products because of a combination of confusing membrane nomenclature,
classification, tradenames and product claims. (Int J Artif Organs 2002; 25: 447-60)

KEY WORDS: Membrane nomenclature, Polysulfone, ß2-microglobulin, High-flux dialysis

1. Early trends in dialysis membrane development number of therapeutic options that have resulted in
necessitated by biocompatibility issues improved patient outcomes and survival rates (1-3).

The development and improvement of dialysis Cellulosic membranes and the biocompatibility
membranes over the last quarter of a century has issue
basically kept apace with the demands of new scientific
and medical findings. While the basic principles of size- In the early eighties, the major impetus towards the
dependent retention or elimination of substances improvement of membranes was necessitated by
across a semi-permeable barrier are still retained, reports in the scientific literature that membranes made
dialysis membrane technology has advanced from regenerated cellulose activated the complement
significantly, offering the clinician – and the patient – a pathway of blood and also caused transient leucopenia

©
Wichtig Editore, 2002 0391-3988/447-14 $07.00/0
Dialysis membranes today

Fig. 1 - Transient leucopenia and complement activation by Fig. 2 - Modified-cellulose membranes are derived by the
regenerated cellulose dialysis membrane (Cuprophan®) that substitution of a small, defined percentage of hydroxyl groups
essentially triggered the biocompatibility debate in within the cellobiose structure with chemical groups such as
hemodialysis. From Hoenich et al (11). diethyaminoethyl (DEAE), acetate (Ac) or benzyl groups (for
'synthetically' modified cellulose, SMC). Although a reduction of
complement activation results, a sufficient number of -OH groups
still remain and activate complement.

(Fig. 1). The initial findings of Craddock et al (4, 5) on such as those of the cellulose acetate family,
the phenomenon were subsequently reproduced and Hemophan® (made by substituting -OH moieties with
validated by several researchers world-wide (6, 7). anion-exchange groups) and later "synthetically"
Cellulose, and more precisely cuprammonium, from modified cellulose (SMC; -OH groups substituted by
which the majority of all early dialysis membranes benzyl groups) became available (Fig. 2).
were made, thus acquired the highly undesirable label
of possessing poor biocompatibility. Introduction of the low-flux polysulfone
Extensive research activities began to be membrane: synthetic membranes
undertaken to improve the poor biocompatibility at a
time when the potential of hemodialysis was However, almost parallel with the first developments
beginning to be appreciated. It soon became apparent of modified-cellulose membranes, another membrane,
that the activation of complement by Cuprophan ® the Fresenius Polysulfone®, began to make its mark in
could be attributed to the polyhydroxyl functionality 1983. It would transform the future development of
present in the basic cellulose structure. During blood- dialysis membranes and associated therapies (12).
material contact, the labile hydroxyl groups of the Membranes manufactured from the synthetic polymer
cellulose molecule interact chemically with the polysulfone (PSu) and polyvinylpyrrolidone (PVP)
sulfhydroxyl group on the C3a- complement molecule blends, result in extremely low complement activation
and, through a cascade pathway, leads to sufficient (8), lower than even the modified cellulosic
systemic complement activation to produce transient membranes (9-11) that were offered as alternatives
neutropenia (5, 6, 8). Cuprophan® (Fig. 3). The PSu - PVP membrane blend
An obvious remedy for the bioincompatibility of soon came to be regarded as the 'Golden Standard' in
regenerated cellulose membranes was, therefore, to terms of its highly favourable all-round
reduce the number of the -OH groups by their biocompatibility profile and not only established a new
substitution with other chemical groups (9-11) This dialysis membrane segment ('synthetic membranes'),
development strategy resulted in a clear reduction of but also set the standards for future competitor
the degree of complement activation as membranes development strategies to emulate (1, 2, 11).

448
Bowry

exceedingly thin wall thickness (15µm down to 8 µm).

Secondly, the abundant polar hydroxyl groups allow a
strong interaction with water, making it a very hydrophilic
polymer resulting in a hydrogel-like material that leads to
high rates of small molecule diffusion (13). However, low-
flux membranes were not suitable for the removal of
larger uremic toxic substances.

Solute removal targets: identification of specific

uremic toxins for removal

In recent years, considerable clinical research

effort has been directed towards the identification
Fig. 3 - Modified cellulose membrane lead to some of substances that cause the uremic syndrome (14),
improvements in biocompatibility but as yet do not match that of
polysulfone membranes. From Hoenich et al (11). characterised by a deterioration of biological and
physiological functions, in parallel with the progression
of renal failure (15). A better understanding of the role
2. The current trend in convective hemodialysis of the various uremic toxins could, eventually, enable
modalities targeted strategies to be developed, such as the
development of membranes for the adsorptive
In the early era of hemodialysis only small-molecular elimination of specific uremic toxins (16).
weight substances were eliminated from the patients'
blood, initially by cellulose-based membranes before Uremic toxins
synthetic membranes such as polysulfone were made
available. The solute removal mechanisms for these As the uremic syndrome is the result of the
low-flux membranes depended entirely upon diffusive cumulative retention of a broad spectrum of
mass transfer. Two facts contribute to the suitability of compounds having innumerable biological effects, it is
cellulose membranes for diffusion-dependent solute now widely accepted that a more effective treatment by
removal: firstly, the unique fibril ultrastructure of cellulose renal replacement therapies is achieved by removal of
offers the material considerable mechanical stability, the small water-soluble and protein-bound compounds
allowing it to be cast into hollow-fibre membranes with (MW < 300) such as urea, creatinine, potassium and

TABLE I - MAIN CLASSES OF SOME UREMIC SOLUTES RETAINED IN RENAL FAILURE. 'UREMIC TOXINS'
ARE NORMALLY ARBITRARILY DIVIDED ACCORDING TO THEIR MOLECULAR WEIGHT (MW).
CLASSIFICATION IN THE LITERATURE MAY THEREFORE VARY

Low MW molecules (< 300 D)

Water-soluble, non-protein-bound e.g. urea (60 D), creatinine (113 D), guanidines, oxalate,
hypoxanthine, uric acid

Protein-bound e.g. p-cresol (108 D), indoxyl sulfate (251 D), phenol, indoles,
hippuric acid, tryptophan, homocysteine

Middle MW range (300 - 12 000 D) e.g. Parathyroid hormone (9223 D), peptide-linked AGEs,
ß2-microglobulin (11 800 D)

High MW solutes (> 12 000 D) e.g. leptin (16 000 D), complement factor D (23 750 D)

Table compiled from reference (15) and personal communication with Prof. R .Vanholder, Gent

449
Dialysis membranes today

phosphate as well as the large “middle” molecules (MW

300 - 5000) and small proteins such as ß2-microglobulin
(ß2-m), parathyroid hormone (PTH) and various
granulocyte inhibitory proteins (16). Table I summarises
the various categories of substances currently regarded
as uremic toxins.
Because ß2-m is only removed by membranes with
larger mean pore sizes, it is considered to be
representative of kinetic behaviour in other large
molecules, although conclusive data are not available
to demonstrate a direct correlation between removal of
ß2-m and the removal of other compounds in the same
molecular weight range. However, one issue related to
Fig. 4 - Efficiency of various diffusive and convective treatment
membranes with more open pores and removal of
dialytic modalities on the clearance of solutes of different
molecules of the ß2-m size-range, is the potential to molecular weights. If it is considered equally important to clear
create albumin losses which has to be precisely solutes over the entire molecular weight spectrum, both transport
principles need to be utilised, as in HDF (17).
controlled.

Treatment modalities for the removal of different In clinical practice, this goal is achieved by
uremic retention solutes hemodiafiltration (HDF) treatment modalities. Metabolite
removal by convective processes alone is performed by
As previously mentioned, classic low-flux hemofiltration (HF). Thus, HDF is a form of 'hybrid'
hemodialysis (HD) has been demonstrated to be therapy, lying in-between the purely diffusive (HD) and
effective for the removal of small water-soluble and the purely convective (HF) treatment modalities (17).
protein-bound molecules by diffusive mechanisms only. The relative efficacy of each of the three modalities is
For the additional removal of larger compounds that shown in Figure 4. The flux of small solutes depends
contribute to uremic toxicity, the diffusive solute largely upon the dialysate flow rate and is maximised
clearances have to be increased by a convective when hemodialysis (HD) is used. The flux of large
component related to plasma water ultrafiltration (2). solutes depends primarily on convective transport

TABLE II - EFFECTIVENESS OF DIFFERENT THERAPY MODES AND THE ROLE OF TREATMENT

PARAMETERS ON SOLUTE REMOVAL. THE IMPORTANCE OF TREATMENT TIME (T) AND BLOOD
FLOW RATE (Q B ) IS VALID FOR ALL THERAPY MODES. THE SELECTION OF THE
ULTRAFILTRATION RATE (QF), TOGETHER WITH FLUID SUBSTITUTION RATE (QS) IS DECISIVE
TO ENHANCE EFFICIENT REMOVAL OF LARGE MOLECULAR WEIGHT SUBSTANCES. DIALYSATE
FLOW RATE (QD) IS ANOTHER VARIABLE IN ALL MODES EXCEPT IN HEMOFILTRATION (HF)

Therapy Mode Parameters Effectiveness

Low-Flux Hemodialysis QB, QD, (QF), T Effective for small molecules

High-Flux Hemodialysis QB, QD, (QF), T Very effective for small, and
effective for large molecules

High-Flux Hemofiltration QB, QF, QS, T Very effective for large molecules

High-Flux Hemodiafiltration QB, QD, QF, QS, T Very effective for small and large molecules

QB = Blood flow rate; QD = Dialysate flow rate; QF = ultrafiltration rate; QS = substitution rate; T = treatment time

450
Bowry

Fig. 5 - Principle of online-HDF which is a safe, economical and Fig. 6 - Relative clinical efficiency of high-flux hemodialysis (HD),
practical way to supply the large quantities of substitution fluid hemofiltration (HF) and hemodiafiltration (HDF) in the clearance
needed to achieve the full benefits of HDF and HF. The usage of of low- (urea), middle- (ß2-microglobulin) and large- (amylase)
a dual polysulfone-based ultrafilter set-up ensures a high degree molecules. Data from Reference (19).
of microbiological safety of large volumes of substitution fluid
and the dialysate.

compensated by high flow rates of substitution fluid and shows that HDF is not only more effective than HD in
is maximised when using either hemofiltration or the removal of low-molecular weight solutes, but shows
hemodiafiltration (Tab. II). Maximal fluxes for both small higher removal rates of larger proteins (ß2-m) than
and large solutes can effectively be achieved using pure convection-dependent HF (19). The removal of
post-dilution HDF in favour of the pre-dilution procedure larger solutes is clearly beneficial in terms of a reduced
(18). On-line HDF, in which the substitution solution is incidence of adverse reactions experienced by the
prepared during the HDF treatment (Fig. 5), allows patient during the treatment (20, Fig. 7). Altieri et al (21)
usage of high filtration rates that directly influence the further showed online-HDF to have even more
efficiency of removal of uremic toxins. pronounced advantages over high-flux HD in terms of
A number of clinical studies have demonstrated the reduced intra-dialytic symptoms (Tab. III).
improved efficacy of HDF over HD and HF in the
removal of small and large solutes (19 - 23). Figure 6 Effect of membrane characteristics on convective
mass transfer
TABLE III - A SIGNIFICANT REDUCTION OF
INTRADIALYTIC SYMPTOMS OBSERVED Because the diffusivity of a solute decreases with
DURING ONLINE-HEMODIAFILTRATION increasing molecular size, the contribution of
(IN PRE-DILUTION MODE) COMPARED
TO HIGH-FLUX HEMODIALYSIS (HD) convection has to be maximised to gain high total
USING ULTRAPURE DIALYSATE solute removal. Thus, for HDF and HF, high-flux
membranes with high fluid and solute permeabilities
High-Flux HD Online-HDF p
are required (17). The convective mass transfer is a
Hypotension 61 39 0.003 function of the ultrafiltration rate (QF) and the solute
Hypertension 30 26 0.04 sieving coefficient (24). The filtrate flux (QF) of pure
Arrhythmia 11 2 0.04
Dyspnea 4 4 n.s. water through a membrane increases linearly with the
Muscular cramps 33 17 0.006 applied transmembrane pressure, with the slope of this
Headache 41 35 0.02
Pruritus 9 11 n.s. curve representing the hydraulic permeability of the
Nausea 17 4 0.02 membrane; the hydraulic permeability is strictly a
Vomiting 9 9 n.s.
function of the membrane structure, depending upon
From reference (21) the pore size, pore distribution, and the membrane

451
Dialysis membranes today

Fig. 7 - Clinical benefits of

convective therapies (hemo-
filtration): reduced adverse
reactions are observed
compared to standard he-
modialysis (HD). Data from
1000 clinical observations,
from Reference (20).

thickness. It decreases significantly after exposure to

blood because of the adsorbed proteins on the
membrane surface.
A measure of the permeability of dialysis membranes
to a particular solute is expressed as the sieving
coefficient that essentially determines which size
molecule may or may not traverse the membrane. The
SC is predominantly a function of the pore size, i.e. that
fraction of membrane pores that are large enough to
allow solutes to pass through unhindered at the
innermost, separating layer of a membrane. Assuming
no adsorption of the solute in question to the
membrane occurs, the SC profile is an indicator of the
Fig. 8 - Comparison of the pore size distribution of low- and
membrane's capability to remove solutes of a specific high-flux polysulfone dialysis membranes with that of a cellulose
size range. Membrane fabrication procedures and membrane. Unlike cellulose, polysulfone is a highly suitable
technology are decisive towards defining these critical material for the manufacture of a range of membranes having a
narrow, selected pore size distribution according to the desired
membrane-related parameters that determine the range of molecules to be removed (Data from Fresenius Medical
convective transport characteristics of a particular Care).
membrane.
Figure 8 compares the pore size distributions of
cellulosic membranes with those of low- and high-flux The advantages of polysulfone as a polymer
polysulfone membranes. Unlike this specific material for high-flux membranes used in
polysulfone, the structure of which permits the convective therapies
controlled modulation of pore size dimensions
according to the size range of solutes to be removed, The introduction of polysulfone as a synthetic
the restrictions imposed by the cellulose structure polymer for dialysis membranes (12) coincided with the
enable a broad “pore” size distribution only, making the biocompatibility debate triggered by the undesirable
material conceptually unsuitable for the efficient interactions of early cellulosic membranes with blood.
removal of particularly large substances (i.e. having The obvious advantages of non-cellulosic polymers
high ß2-m sieving coefficients). were soon appreciated and other synthetic polymers

452
Bowry

(man-made non-cellulosic polymers) began to be used material for convective therapies involving high
for the commercial production of dialysis membranes exchange volumes is severely restricted due to the fact
(13); these include polyacrylonitrile (PAN), poly- that high ultrafiltration coefficients (KUF) matching
methylmethacrylate (PMMA), ethylvinylalcohol (EVAL), those of the synthetic membranes cannot be attained.
polyamide (PA) and polycarbonate (PC). For example, the highest KUF quoted for cellulose-
It would be pertinent, at this point, to examine the triacetate dialysers (1.9 m2) is only 36 ml/hr/mmHg,
reasons as to why none of these synthetic polymers has whereas it not uncommon today for a synthetic
made an impact comparable to that of polysulfone: for membrane dialyser having a surface area of only 1.2
example, data from the National Surveillance of m2 to have a KUF of well over 50 ml/hr/mmHg. The
Dialysis-Associated Diseases in the United States, ultrastructure of certain modified cellulose membranes
1999, show that 50% of all patients were treated with allows for high porosity that aids small molecule
high-flux polysulfone dialysers (e.g. F60, HF80) and diffusion, whilst the small pore size limits ultrafiltration
20.9% with low-flux polysulfone dialysers (e.g. F5, F8) to levels just above the traditional range and constrains
while various cellulose-based dialysers constituted the transport of permeants in the middle molecular
altogether 27.3% of the dialyser types used (25). Such range. Thus, cellulose acetate membranes fail to meet
overwhelming dominance of polysulfone membranes the first two criteria of modern membranes listed above,
over other materials is attributed to a number of distinct namely high biocompatibility and appropriateness for
features of polysulfone as a dialysis membrane material. high-volume convective therapies. In addition, cellulose
Foremost is the versatility of polysulfone as a acetate membranes undergo an ageing process which
material for the fabrication of a comprehensive range of has been linked to severe adverse reactions, e.g.
membranes required for the diverse diffusive and decreased vision and hearing when dialysis treatment
convective solute removal therapies. Although is carried out with aged dialysers (26).
polysulfone is not the only polymer that is being used The inability of a number of synthetic polymers to be
for low- and high-flux hemodialysis, hemodiafiltration or sterilised by steam further restricts the number of
hemofiltration, it is one of the few polymers that meets synthetic polymers that meet the essential criteria listed
the crucial demands associated with these therapeutic above. As ethylene oxide is associated with acute
modalities in use today. hypersensitivity reactions (27) and irradiation can lead
Ideally, polymers for modern dialysis membranes to structural and visual changes of polymers or liberate
must meet the following fundamental criteria: polymer by-products, steam sterilisation is clearly the
1. High biocompatibility to minimise acute and long- most favoured mode of sterilisation of medical devices.
term complications; Dialysis membrane polymers such as polyacrylonitrile
2. Suitability for multiple therapeutic modalities (low- (PAN) and polyamide (PA) do not, therefore meet this
and high-flux HD, HDF, HF); important requirement. Polymers of the polysulfone
3. Steam sterilisation; family are the only synthetic polymers that can be
4. High endotoxin retention capabilities. sterilised by all three sterilisation modes (steam,
The issue of biocompatibility has been discussed in ethylene oxide and gamma radiation).
earlier sections, and this criterion essentially relegates
cellulose-based membranes to second choice Considerations associated with the use of high-
compared to synthetic materials. Even substituted flux membranes
cellulosic membranes maintain a sufficient number of
hydroxyl entities to activate complement (Figs. 2, 3), The usage of high-flux membranes for convective
thereby never reaching the minimal complement- therapies is not without additional problems that need
activating properties of synthetic membranes (9, 11); in to be addressed. The phenomenon of back-transport
addition, the substituted groups may often compromise (i.e. back-filtration + back-diffusion) is unavoidable for
the thrombogenic properties of the membrane. membranes with high permeability leading to the
Furthermore, the usage of cellulose as a membrane potential risk of endotoxin transfer (lipopolysaccharide

453
Dialysis membranes today

material from the outer wall of gram-negative bacteria) not readily met by any other polymer material.
from contaminated dialysate fluid (28). The water The obvious question that arises is whether the end-
quality of most dialysis units – even in developed user is able to distinguish between the quality and
countries – often does not meet the established water features of the various polysulfone-based membranes
purity standards (in terms of either bacterial colony currently available. The confusion is increased by the fact
forming units, CFU, or as endotoxin units), thereby that the new polysulfone membranes have, on the one
posing a threat to the safety of the patient (29). Even hand, to compete with the proven image and long-term
with the most advanced water treatment systems, experience of the classic Fresenius Polysulfone®, while on
variations in the quality of the incoming water render it the other hand, there is the obvious necessity to
difficult to assure adequate quality at all times. differentiate the membrane with apparently novel features.
All modern-day dialysis membranes are therefore Table V lists the various dialysis membranes of the
expected to meet the explicit requirement of being polysulfone “family” from different manufacturers,
impermeable to pyrogenic / endotoxic material. together with the key features of each product. The first
However, not all polymer membranes are able to point to note is that not all manufacturers offer the low-
prevent endotoxin passage especially when several and high-flux varieties of polysulfone membranes
litres of substitution fluid are infused into the blood covering the HD, HDF and HF therapy range: only two
stream of the patient. Table IV shows that polysulfone dialyser manufacturers (Fresenius Medical Care and
membranes have unparalleled endotoxin retention Gambro) are able to cover the complete spectrum of
capabilities compared to other polymers (30). membrane variants. Secondly, only one membrane
(Fresenius Polysulfone®) from the polysulfone family is
Polysulfone-based dialysis membranes (Tab. V) available with the more efficient in-line steam-
sterilisation procedure. Regarding the dimensions of the
In recent years, there has been a proliferation of membranes, the wall thickness varies from 30 µm to 50
polysulfone-based dialysis membrane introduced onto µm while the fibre inner diameter ranges from 200 to
the market. Virtually every dialyser manufacturer and 215 µm. Generally, a thinner wall thickness is associated
supplier has been compelled by the state-of-the-art with increased diffusive mass transfer and, for reasons
convective therapies to include a polysulfone-based considered above, is more of an advantage for low-flux
dialyser as a major part of their product portfolio. The membranes rather than for high-flux membranes
reasons for this trend are apparent: the aforementioned, intended for HDF and HF which rely predominantly on
overriding advantages associated with polysulfone are convective solute transport mechanisms.

TABLE IV - ADSORPTION CAPACITY OF DIFFERENT DIALYSIS MEMBRANES FOR PSEUDOMONAS

AERUGINOSA ENDOTOXINS (CHEMICALLY LIPO PLYSACCHARIDE, (LPS), THE LIPID DOMAIN
OF WHICH IS LIPID A). THE ADSORPTION WAS MEASURED ON THE FIBRE LUMEN SIDE AFTER
A 2 hr INCUBATION PERIOD

Membrane Type (Dialyser type) Lipid A adsorbed onto membrane (µg/m2) LPS adsorbed onto membrane (ng/m2)

Cuprophan (E3) 4.5 ± 3.4 91 ± 29

Low-flux polysulfone (F4) 36.8 ± 8.1 360 ± 42
Cellulose acetate (Acephal 1300) 10.7 ± 2.3 162 ± 38
Cellulose acetate (Altraflux) 11.0 ± 1.3 90 ± 42
± ±
Polyamide (Polyflux 110) 61.1 ± 10.8 352 ± 123
AN 69 (Filtral 12) 12.8 ± 4.7 109 ± 84
High-flux polysulfone (F40) 50.3 ± 7.7 502 ± 86
High-flux polysulfone (F60) 55.5 ± 7.1 567 ± 80

Data from Weber et al (Ref. 30)

454
Bowry

Nomenclature of the polyarylsulfone-based common features and are differentiated chemically only
membranes by the fact that polyethersulfone lacks the -CH3 groups
which will result in an alteration of the chemical and
The manufacturer-specified trade name of a dialysis physical properties of the polymer. Further dissimilarities
membrane does not always provide an obvious between the polysulfone or the polyethersulfone
indication of the base polymer used for the manufacture membranes arise due to the selection of the membrane
of the membrane, and data sheets rarely specify the fabrication procedures (i.e. spinning conditions) and
exact chemical nature of the membrane. A more also due to the inclusion of additional of co-polymers in
detailed analysis reveals that all the polysulfone the spinning solution. The result of these manufacturer-
membranes listed in Table V are made from two base specific procedures is that the membranes have vastly
polymers only, either polysulfone or polyethersulfone. variable final characteristics e.g. membrane structure
Both these polymer types are classified under the and dimensions which significantly affect the different
polyarylsulfone group, in which the sulfone and arylether solute transport characteristics of membranes even
chemical groups are common to both polysulfone and having similar polymer compositions.
polyethersulfone (Fig. 9). Another anomaly is that few Figure 10 shows the membrane structures of different
recognise the fact that polyethersulfone and membranes of the polyarylsulfone family; it is apparent
polyarylethersulfone are exactly synonymous and do not that the fluxes, water permeability, the solute removal
represent two different polymer types. Thus, as shown and endotoxin retention characteristics cannot be
in Table V all polysulfone-based dialysis membranes expected to be identical for such vastly different
(being either polysulfone or polyethersulfone) possess membrane wall structures. Yet, striking similarities

TABLE V - SUMMARY OF THE MAIN CHARACTERISTICS OF THE VARIOUS POLYSULFONE DIALYSIS

MEMBRANES CURRENTLY ON THE MARKET. ALL MEMBRANES IN THIS CLASS BELONG TO
THE POLYARYLSULFONE FAMILY AND ARE CHEMICALLY EITHER POLYSULFONE OR
POLYETHERSULFONE – DESPITE TRADE NAMES OF CERTAIN MEMBRANES IMPLYING
OTHERWISE (E.G. POLYAMIDE STM)

Membrane Name Membrane Polymer Wall thickness (µm) Fibre inner diameter (µm) Low-/ High-Flux Sterilisation mode
(Manufacturer)

FMC Polysulfone® Polysulfone 40 200 LF / HF In-line steam

(Fresenius Medical Care)
Helixone® Polysulfone 35 185 HF In-line steam
(Fresenius Medical Care)
Asahi Polysulfone® Polysulfone 45 200 HF Gamma-ray
(Asahi )
Toraysulfone® Polysulfone 40 200 HF Gamma-ray
(Toray Industries)
Clirans PS Polysulfone 40 200 HF Autoclave
(Terumo)
Althane Polysulfone Polysulfone 23 ? HF Gamma-ray
(Althin/Baxter)
DIAPES Polyethersulfone 30 200 LF / HF Gamma-ray
(Membrana)
Synphan Polyethersulfone 35 200 HF Gamma-ray
(Membrana)
PEPA Polyethersulfone + Polyarylate 30 210 HF Gamma-ray
(Nikkiso)
Polyamide STM Polyarylethersulfone# 50 215 LF / HF Steam
(Gambro)
Arylane Polyarylethersulfone# 50 215 HF Gamma-ray
(Hospal Cobe)

# Polyarylethersulfone is simply another name for polyethersulfone

455
Dialysis membranes today

Fig. 9 - Chemical structures of the generic polysulfones: both Fig. 10 - Various membrane structures of generic polysulfone
polysulfone and polyethersulfone belong to the poly(aryl)sulfone dialysis membranes. Vastly different membrane thicknesses,
group, possessing the sulfone and arylether chemical groups. pore structures and their distribution leads to each membrane
The presence of the -CH3 groups in the polysulfone structure having its characteristic properties even if they belong to the
probably contributes to unique properties of polysulfone same polymer group. A: Fresenius Polysulfone® (40 µm);
membranes (e.g. high retention of endotoxins). B: Gambro Polyamide STM (50 µm); C: Hospal-Cobe Arylane
(50 µm); D: Membrana (Baxter/Bellco) DIAPES (30 µm).

between the key membrane properties are commonly Nomenclature and terminology of dialysis
reported in the literature. Membranes from the same membranes
polymer family in Table V, despite having similarities
between the base polymers, therefore possess distinct The major historical dialysis membrane developments
characteristics with respect to biocompatibility, endotoxin discussed in this paper would be incomplete without
retention ability, sterilisation and solute removal mention of the background to the common naming of
characteristics (e.g. elimination of ß2-m). To illustrate the membranes and the misconceptions that frequently
point, the permeability of polysulfone (Fresenius arise due to the selection and usage of product trade
Polysulfone ®, F60 dialysers) and polyethersulfone names. The correct classification of the three main
(DIAPES, Bellco BLS 814G dialysers) to endotoxin was membrane groups are:
compared in an in vitro model. The endotoxin transfer 1. Cellulose membranes (e.g. Cuprophan®, regenerated
from the dialysate side to the blood side was measured cellulose);
by re-circulating water contaminated with endotoxin (2 x 2. Modified-cellulose membranes (e.g. cellulose
104 IE/ml) on the dialysate side at a flow rate of 500 acetates, Hemophan®, SMC);
ml/min. In the blood side, distilled water was re-circulated 3. Synthetic membranes (e.g. polysulfones, PAN,
at a rate of 300 ml/min. Figure 11 shows the endotoxin EVAL, PMMA).
measurements on samples taken at various times from The polymeric constitution of the membrane is not
the blood-side compartment. Unlike the polysulfone obvious from the commonly-used (trade)name of each
membrane (F60), the DIAPES membrane demonstrates membrane: membrane names like Cuprophan ® ,
significant passage of endotoxin even after 30 minutes, Bioflux, Hemophan® do not provide a direct indication
reaching extremely high levels after a four-hour period of of the polymer used as do membrane names such as
re-circulation. Thus DIAPES, a high-flux membrane on a cellulose acetate, polysulfone or polyacrylonitrile
polysulfone basis, exhibits permeability to LAL-positive (AN69). Similarly, different trade names are used for
material even under purely diffusive conditions and does membranes made from the same polymer: e.g.
not meet one of the more important criteria (listed above) polysulfone-based membranes have trade names such
required of modern membranes. as Fresenius Polysulfone ® , Toraysulfone ® , Asahi

456
Bowry

Fig. 11 - Comparison of the endotoxin retention capabilities of Fig. 12 - Analysis of the chemical composition of Polyamide
two polysulfone membranes (F60 and FX60 dialyser series) with S TM. by GPC-coupled FTIR spectroscopy. Although the
DIAPES membrane (Bellco BLS 816G dialysers). Differences in membrane name suggests that polyamide is a major constituent,
the physicochemical properties of the membrane (due to the results show Polyamide S TM contains no traces of this
differences in the chemical structures of polymers and the polymer. Instead, the membrane is comprised solely of
membrane fabrication conditions) leads to a leakage of polyarylethersulfone (i.e. polyethersulfone).
endotoxin by the DIAPES membrane only. Data from Fresenius
Medical Care, St Wendel (in vitro recirculation model, using LAL
turbidometric assay (sensitivity 0.001 IE/ml).

Polysulfone APS ® and Terumo Clirans PS; the a strong resemblance to PSu – the recognised and
obviousness of the type of polymer used for these chemically correct abbreviation of polysulfone.
membrane types is immediately apparent. Another case of membrane terminology resulting in
The “modified-cellulose” class of membranes are confusion to the end-user is the Polyamide S TM
often, referred to as “synthetically-modified” cellulosic membrane manufactured by Gambro (Hechingen -
membranes. The background to this misnomer is the Germany). The obvious implication is that the
biocompatibility issue whereby the cellulosic membrane polymer is polyamide, characterised by the
membranes, because of their poor biocompatibility had presence of the amide group (- CO - NH - ), and that
to be modified in attempts to simulate the highly the major component of the membrane would be either
favourable biocompatibility profiles of synthetic an aliphatic or an aromatic polyamide. While all product
membranes: association with the term “synthetic” or brochures describe the membrane of the dialysers as
even with “polysulfone” (PSu) image rather than with Polyamide STM, a footnote in the product inserts for
“cellulose” had therefore obvious positive connotations. dialyser specifies the Polyamide S membrane as
The term “synthetic” in the context of cellulosic “polyarylethersulfone”. As the chemical and physical
membranes is misleading in that few recognise the fact properties of polyamides and polyarylethersulfones are
that the synthetically-modified cellulose membranes entirely different, it is unclear whether the functional
maintain the cellulose backbone and that well over 95% characteristics of Polyamide STM are to be attributed to
of the membranes are still comprised of cellulose – polyamide, or, to polyarylethersulfone. The choice of
which explains why cellulose modifications do not the base polymer used for the manufacture of dialysis
match the biocompatibility of membranes like membranes influences not only the membrane spinning
polysulfone (11). Infact, the term “synthetically modified conditions but also the attainable membrane
cellulose” refers to both, the modified cellulose family dimensions and structure (wall thickness, porosity, pore
as well as a specific product, namely, SMC produced size and distribution) that determine the mass transfer
by Membrana. Confusingly, Baxter has named the characteristics of the membrane. Furthermore, as noted
same SMC membrane as Polysynthane (PSN) – the above, polyamides cannot be sterilised with steam.
abbreviation for this cellulose-based membrane bearing Chemical analysis undertaken to ascertain the

457
Dialysis membranes today

Fig 13 - Comparison of the pore size and distribution of Fig. 14 - Sieving coefficient profiles of high-flux Fresenius
Fresenius Polysulfone® and Helixone®, the new, advanced high- Polysulfone ® and Helixone ® which has an increased ß2-m
flux membrane designed specifically for the removal of larger sieving coefficient (0.8 compared to 0.65 for the Fresenius
solutes (e.g. ß2-microglobulin). Helixone ® has a more Polysulfone ® ) while maintaining the low albumin loss
homogeneous distribution of pores with a larger average pore characteristics.
size, in a narrow size range distribution.

chemical nature of Polyamide S TM using several The correct and unambiguous designation of dialysis
analytical techniques (gel permeation chromatography- membranes is not just a matter of terminology and
coupled Fourier transformation infrared spectroscopy, nomenclature: the influence of the chemical nature of
nuclear magnetic resonance spectroscopy, differential the dialysis membrane on patient outcome and survival
scanning calorimetry and quantitative elementary is the subject of analyses of several national and
analysis) failed to show any traces of polyamide in international registries (33). The outcome of these
Polyamide S TM (Fig. 12). No aliphatic or aromatic studies may be seriously undermined if confusing
polyamide chemical entities were detected in any of the membrane specifications, cited in the literature or by
samples tested (relative to polyamide reference manufacturers, are available to the investigators of
materials). The polyamide membrane was thus found to such bodies.
contain the polymer polyethersulfone (i.e.
polyarylethersulfone) only (excluding PVP which is a New membrane developments for convective
copolymer of most synthetic membranes) – a fact that therapies
was obviously not recognised by clinical investigators
and reported in the literature. In the study of Hoenich et As described in the aforementioned sections, the
al (31) the membrane of the Polyflux 14S dialyser type current preference and dominance of polysulfone-
is mentioned as “the Polyflux membrane” and further based materials for dialysis membranes is attributed to
described as being “manufactured from a blend of this polymer group complying with the well-defined,
polyamide, polyarylethersulfone and polyvinyl- modern therapy- and technology-related requirements
pyrrolidone (PVP)”, while Ward et al (32) clearly refer to in the dialysis field. Numerous attempts to develop
the “steam-sterilized polyamide membrane (Polyflux 14 alternative materials have met with limited success to
S, Gambro)” in their prospective clinical study. Thus, the extent that product differentiation is even sought at
the descriptions of both authors for a single product are the expense of aspects such as membrane
totally divergent, and neither designation complies with nomenclature rather than on measurable and verifiable
either of the manufacturers’ specifications, i.e. that product features.
the Polyamide S TM membrane is made from The basis of further membrane-related refinements of
polyarylethersulfone polymer. dialysis treatments/therapies lies therefore on two

458
Bowry

fundamental approaches. Firstly, existing materials the innermost separating layer of Helixone® (35). The
having active functional groups or biological entities results of these refinements are that the membrane
attached to their surface may provide advantages inner surface has an increased average pore size and
related to a specific physiological function, or, aid the pore size uniformity as well as an even distribution of
targeted removal of individual toxins. An important point the geometrically well-defined pores at the innermost
to bear in mind with this approach is that the surface (Fig. 13) for the selective removal of molecules
modifications often result in a deterioration of the in the size range of ß2-m (35). The result is that the
material’s original surface properties or may result in sieving coefficient curve is steeper, providing the high
side-effects. Secondly, based upon a deeper ß2-m sieving coefficient of 0.8 (Fig. 14). Furthermore,
understanding of the mechanisms and factors the increased membrane surface porosity within the
influencing solute transfer, the approach is to fine-tune initial nano-scale region of the membrane offers
the structure of the effective part of the membrane reduced resistance for solute removal – without
surface to enable an increase in the selectivity and compromising the albumin retention characteristics of
efficiency of removal of solutes in a narrow size-range. the membrane.
A product of the latter approach is Helixone ® In conclusion, although current membranes for
(Fresenius Medical Care, Bad Homburg, Germany), an hemodialysis therapies are far from being an adequate
advanced polysulfone-based dialysis membrane surrogate to the natural membrane of the glomerulus,
specifically dedicated to high-flux dialysis and they enable routine renal replacement therapies to be
associated convective therapies and developed by successfully carried out based on their size exclusion
using nanofabrication techniques. Nanotechnology has principles. Synthetic membranes offer the versatility of
today emerged as a discipline in its own right that is at providing the sieving functions for the removal of a
the fore-front in some novel and progressive range of molecules that accumulate during renal failure.
developments of new materials, devices, processes The choice of polymers for the fabrication of dialysis
and techniques in every scientific field - medical, membranes used in modern-day therapies is
technological and industrial. Recently, the journal nevertheless limited as few polymers meet the crucial
Science even devoted a special issue (Vol 290, Nov requirements defined by the medical and scientific
2000) to advances and current trends in this rapidly communities.
expanding area.
As the thin, innermost “skin” layer of a membrane
determines the sieving characteristics of a dialysis
Reprint requests to:
membrane (34), membrane spinning conditions Sudhir K. Bowry, PhD
involving a combination of nanotechnological principles Fresenius Medical Care
Else-Kröner-Strasse 1
were used to define more precisely the pore structure D-61352 Bad Homburg, Germany
(size and geometry) as well as the pore distribution for e-mail: sudhir.bowry@fmc-ag.com

REFERENCES
1. Cheung AK, Leypoldt JK. The hemodialysis membranes: Contrib Nephrol 2001; 133: 81-104.
a historical perspective, current state and future prospect. 3. Bloembergen WE, Hakim RM, Stannard DC, et al.
Semin Nephrol 1997; 17: 196-213. Relationship of dialysis membrane and cause-specific
2. Hoenich NA, Katopodis KP. Hemodialysis membranes: a mortality. Am J Kidney Dis 1999; 33: 1-10.
matter of fact or taste? In: Ronco C, Winchester JF eds. 4. Craddock P, Fehr J, Dalmasso A, et al. Hemodialysis
Dialysis, Dialyszers and Sorbents. Where are we going? leucopenia: pulmonary vascular leucostasis resulting from

459
Dialysis membranes today

complement activation of dialyser cellophane 21. Altieri P, Sorba GB, Bolasco PG, et al. On-line predilution
membranes: J Clin Invest 1977; 59: 879-88. hemofiltration versus ultrapure high-flux hemodialysis: a
5. Craddock PR, Fehr J, Brighman KL et al. Complement multicentre prospective study in 23 patients. Blood Purif
and leucocyte-mediated pulmonary dysfunction in 1997; 15: 169-181.
hemodialysis. N Engl J Med 1977; 296: 769-74. 22. Locatelli F, Marcelli D, Conte F, et al. Comparison of
6. Falkenhagen D, Bosch T, Brown G, et al. A clinical study mortality in ESRD patients on convective and diffusive
on different cellulosic dialysis membranes. Nephrol Dial extracorporeal treatments. The Registro Lombardo Dialisi
Transplant 1987; 2: 537-45. e Trapianto. Kidney Int 1999; 55: 286-95.
7. von Sengbusch G, Bowry S, Vienken J. Focussing on 23. Lornoy W, Becaus I, Billiouw JM, Sierens L, van Malderen
membranes. Artif Organs 1993; 17 (4): 244-53. P. Remarkable removal of beta-2-microglobulin by on-line
8. Vienken J, Diamantoglou M, Hahn C, et al. hemodiafiltration. Am J Nephrol 1998; 18 (2):105-8.
Considerations on developmental aspects of bio- 24. Berland Y, Brunet P, Dussol B. Ultrafiltration and
compatible dialysis membranes. Artif Organs 1995; 19: depurative parameters. In: Bonomini V, Berland Y eds.
398-406. Dialysis Membranes: Structure and Predictions. Contrib
9. Hoenich NA, Woffindin C, Cox PJ, et al. Clinical Nephrol 1995; 113: 45-59.
characterisation of Dicea: a new cellulose membrane for 25. Tokars JI, Miller ET, Alter MJ, Arduino MJ. National
hemodialysis. Clin Nephrol 1997; 48: 253-9. Surveillance of Dialysis-Associated Diseases in the
10. Lucchi L, Bonucchi D, Acerbi MA, et al. Improved United States, 1999. National Center for Infectious
biocompatibility of modified cellulosic membranes: the Diseases, Centers for Disease Control and Prevention,
case of Hemophan. Artif Organs 1989; 13: 417-21. Atlanta, Georgia.
11. Hoenich NA, Woffindin C, Stamp S, et al. Synthetically 26. Hutter JC, Kuehnert MJ, Wallis RR, et al. Acute onset of
modified cellulose: an alternative to synthetic decreased vision and hearing loss traced to hemodialysis
membranes for use in hemodialysis? Biomaterials 1997; treatment with aged dialysers. JAMA 2000; 283: 2128-34.
18: 1299-303. 27. Tielemans C, Gastaldello K, Goldman M, et al. Acute
12. Streicher E, Schneider H. Polysulfone membrane hemodialysis membrane-associated reactions. Nephrol
mimicking human glomerular basement membrane. Dial Transplant 1996; 11 (suppl 2): S112-5.
Lancet 1983; ii: 1136. 28. Baurmeister U, Travers M, Vienken, et al. Dialysate
13. Lysaght MJ. Evolution of hemodialysis membranes. In contamination and back filtration may limit the use of high-
Bonomini V, Berland Y eds. Dialysis Membranes: Structure flux dialysis membranes. Trans ASAIO 1989; 35: 519-22.
and Predictions. Contrib Nephrol 1995; 113: 1-10. 29. Lonneman G, Krautzig S, Koch KM. Quality of water and
14. Hoerl WH. Uremic toxins: new aspects. J Nephrol 2000; dialysate in hemodialysis. Nephrol Dial Transplant 1996;
13 (suppl 3): S83-8. 11: 946-9.
15. Vanholder R, De Smet R, Lameire NH. Redesigning the 30. Weber C, Linsberger I, Rafiee-Teharani M, et al.
map of uremic toxins. In: Ronco C, Winchester JF eds. Permeability and adsorption capacity of dialysis
Dialysis, Dialyszers and Sorbents. Where are we going? membranes to lipid A. Int J Artif Organs 1997; 20.144-52.
Contrib Nephrol 2001; 133: 42-70. 31. Hoenich NA, Stamp S, Roberts SJ. A micro-domain-
16. Vanholder R, De Smet R, Lameire N. Protein-bound structured synthetic high-flux hollow fibre membrane for
uremic solutes: the forgotten toxins. Kidney Int 2001; 59 renal replacement therapy. ASAIO J 2000; 46: 70-5.
(suppl 78): S266-70. 32. Ward RA, Schmidt B, Hullin J, et al. A comparison of on-
17. Ledebo I. Principles and practice of hemofiltration and line hemofiltration and high-flux hemodialysis: a
hemodiafiltration. Artif Organs 1998; 22 (1): 20-5. prospective clinical study. J Am Soc Nephrol 2000; 11:
18. Ahrenholz P, Winkler RE, Ramelow W, et al. On-line 2344-50.
hemnodiafiltration with pre-and postdilution: a comparison 33. Bloembergen WE, Hakim RM, Stannard DC, et al.
of efficacy. Int J Artif Organs 1997; 20: 81-90. Relationship of dialysis membrane and cause-specific
19. Wizeman V. Hemofiltration - an avenue to shorter mortality. Am J Kidney Dis 1999; 33: 1-10.
dialysis? Contrib Nephrol 1985; 44: 49-56. 34. Mulder M. Characterisation of membranes. In: Basic
20. Quellhorst E, Schünemann B. Haemofiltrationzur principles of membrane technology, Kluwer Academic,
Behandlung der chronischen Niereninsuffizienz. In: Franz Dordrecht 1998; pp. 157- 209.
HE, Hörl WH, eds. Blutreinigungsverfahren, Technik und 35. Ronco C, Nissenson AR. Does nanotechnology apply to
Klinik. Stuttgart: Georg Thieme 1997; 484-95. dialysis? Blood Purif 2001; 19: 347-52.

460
View publication stats