The filmadsorber : some aspects of its development,
manufacturing and use for bloodpurification
Citation for published version (APA):
van Berlo, G. M. W. (1986). The filmadsorber : some aspects of its development, manufacturing and use for bloodpurification. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR245822
DOI:
10.6100/IR245822
Document status and date: Published: 01/01/1986 Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
providing details and we will investigate your claim.
THE FILMADSORBER
THE FILMADSORBER
Some aspects of Its development, manufacturing
and use for bloodpurification.
PROEFSCHRIFT
TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN, OP GEZAG VAN DE RECTOR MAGNIACUS, PROF. DR. F.N. HOOGE, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HET OPENBAAR TE VERDEDIGEN OP
VRUDAG 16 MEI 1986 TE 16.00 UUR DOOR
GERARDUS MARIA WILHELMUS VAN BERLO
GEBOREN TE SINT OEDENRODE
promotoren : Prof.Dr.Ir. F.M. Everaerts
Technische Hogeschool Ei:ndhoven
Prof.Dr.Sc.Med. H. Klint.ann
Wilhelm Pieck UniversitAt Rostock eo-promotor: Dr.Ir. A.H.M. Verkooyen
for the dream of yesterday is the hope of today and the reality of tomorrow.
Robert H. Goddard
Aan Corien Aan mijn ouders
Scholco Textielfabrieken bv te Almelo. De firma Norit nv te Amersfoort en de firma Gambro bv te Breda droegen bij in de drukkosten van dit proefschrift. Cover-design: Corien van Berlo-Deeben
I INTRODUCTION
1.1 Principles and techniques of bloodpurification
1.2 Application and adequacy of bloodpurification 3 1.2.1 Technique of extracorporeal bloodpurification 3
1.2.2 Adequacy of treatments 4
1.2.3 Unselective, selective and specific removal 5 1.3 Objectives of the filmadsorber project
1.4 Programme selection II HAEMOPERFUSION
2.1 Hj~torical background 2.2 Activated carbon/charcoal
2.2.1 Sources and preparation 2.2.2 Terminology
2.2.3 The properties of activated charcoal 2.2.4 Types of activated charcoal granules 2.2.5 Uncoated granular activated charcoal 2.2.6 Polymer-coated activated charcoal granules 2.2.7 Immobilised activated powder charcoal 2.3 Resins and other adsorbents
2.3.1 Ion exchange resins 2.3.2 Macroreticular resins 2.3.3 Pyrolised resins 2.3.4 Other adsorbents 7 9 10 10 10 10 11 11 12 13 14 17 24 24 25 26 26
2.4 Complex sorbent systems 27
2.4.1 Artificial cells 27
2.4.2 Affinity chromatography with immunoadsorbents 28 2.4.3 Other affinity chromatography systems 33
2.4.4 Enzyme reactors 35 2.5 Clinical applications 37 2.5.1 Acute poisoning 37 2.5.2 Uraemia 39 2.5.3 Hepatic failure 41 2.5.4 Immunological diseases 44
2.5.5 Other clinical utilisation 45
2.6 Perfusion of other fluids 46
3.2 Previous studies of mass transfer in adsorption devices 48
3.3 Methodology and assumptions 50
3.4 A structural model of adsorption 51
3.4.1 Adsorption theory 51
3.4.2 External diffusion 53
3.4.3 Intraparticle diffusion 54
3.4.4 Effects of particle size 56
3.4.5 Total set of equations 56
3.5 A formal model of adsorption 59
3.5.1 Transition from structural to formal model 59 3.5.2 Mathematical equations for two hypotheses 62 3.5.3 Determination of K7 and K7;K8 through curve-fitting 64 3.5.4 Estimation of mass tr~nsfer coefficients 66 3.5.5 Relation between formal and structural model 68 3.5.6 Influence of design and process parameters
3.6 Pharmacokinetic considerations 3.6.1 Definition and relevance 3.6.2 Pharmacokinetic models
3.7 A kinetic model of the liver's excretory function 3.7.1 Introduction
3.7.2 Mathematical equations 3.7.3 Some simulations
3.7.4 Calculation of required clearance capacity IV MANUFACTURE OF THE FILMADSORBER
4.1 Introduction 4.2 Historical review
4.2.1 The original concept 4.2.2 Further improvements
4.3 Design of a new film production system 4.3.1 Good Manufacturing Practice 4.3.2 Original outlines
4.3.3 The intermediate film production system 4.4 The semi-automatic manufacturing system
4.4.1 Requirements and specifications 4.4.2 Suspension storage
4.4.3 Carrier storage and supply 4.4.4 Coating section 4.4.5 Drying section 4.4.6 Extraction section 4.4.7 Windinq section 69 72 72 73 74 74 74 76 78 80 80 81 81 81 83 83 83 87 89 89 90 91 92 93 93 96
4.5 Processing in the winding section 4.5.1 Introduction
4.5.2 Measurements in the production machine 4.5.3 The winding section
4.6 The sterilisation system 4.6.1 Introduction
4.6.2 Current sterilisation methods
4.6.3 Gamma sterilisation of the filmadsorber 4.6.4 Steam sterilisation of the filmadsorber 4.6.5 Sterilisation through forced circulation 4.6.6 A comparative study of sterilisation time 4.6.7 Chemical stability during sterilisation 4.7 Device-characteristics
4.7.1 Geometry 4.7.2 Pressure drop
4.7.3 Residence time distribution 4.7.4 Particle release
4.7.5 Trace element analysis
V IN-VITRO EVALUATION OF COLUMN PERFORMANCES 5.1 Introduction
5.2 Adsorption isotherm experiments 5.2.1 Experimental equipment 5.2.2 Experimental methods 5.2.3 Results
5.3 Measurement of intraparticle diffusion rates 5. 3. 1 Theory
5.3.2 Materials and methods 5.3.3 Results
5.4 Recirculation experiments 5.4.1 Aim of the experiments 5.4.2 Experimental set-up 5.4.3 Results
5.5 Haemodialysis-haemoperfusion in series 5.5.1 Introduction
5.5.2 Materials and methods 5.5.3 Results
5.6 A short bloodcompatibility study 5.6.1 Introduction
5.6.2 Materials and methods 5.6.3 Results 100 100 100 102 104 104 105 106 108 112 116 118 121 121 123 124 126 128 130 130 130 130 132 133 138 138 141 142 145 145 145 146 146 146 148 149 149 149 150
6.1 Introduction 154 6.2 Normal and disturbed liver function 154 6.2.1 Functions of the normal liver 154 6.2.2 Potentially toxic metabolites in hepatic coma 157 6.3 Animal models of fulminant hepatic failure 157
6.3.1 Requirements 157
6.3.2 Previous studies with animal models 158 6.3.3 The animal model in this study 160 6.4 Effects of haemoperfusion with the filmadsorber 162
6.4.1 Materials and methods 162
6.4.2 Results 164
6.5 A pilot study with an extended animal model 170 6.5.1 Extension of the model in Rostock 170 6.5.2 Results w~th one short haemoperfusion 171
6.5.3 Histological investigation 173
6.6 A biocompatibility study in a healthy pig 173
6.6.1 Introduction 173
6.6.2 The control study 173
6.6.3 Haemoperfusion with the filmadsorber 175 6.6.4 Histological investigation after haemoperfusion 177 VII MARKET AND MARKETING, FINANCIAL PLAN AND PATENTABILITY 178
7.1 Introduction 178
7.2 Market study in the Netherlands 178
7.2.1 Design of the study 178
7.2.2 General results of the study 179 7.2.3 Technical results of the study 181 7.3 A marketing report in collaboration with Smith&Nephew Ltd 183 7.3.1 Market expectations for Haemocol 183
7.3.2 The proposed collaboration 186
1.4 Financial plan 7.4.1 Introduction 7.4.2 Adsorption costing
7.4.3 Feasibility of the company 7.5 Patentability
VIII GENERAL DISCUSSION AND CONCLUSIONS 8.1 Framework for discussion
188 188 188 189 190 191 191
8.3 Kinetic modeling of haemoperfusion 192
8.4 Manufacture of the filmadsorber 194
8.5 In-vitro evaluation of column performances 196 8.6 In-vivo evaluation of column performances 197
8.7 Market and marketing, financial plan and patentability 198
8.8 Overall conclusions 199
IX REFERENCES 201
X APPENDICES 231
10.1 Calculation of cost of the filmadsorber 231
10.2 Financial calculations 233
10.2.1 Analysis of product cost 233
10.2.2 Budgets 235
10.2.3 Statement of earnings 236
10.2.4 Balance sheet 237
10.2.5 Statements of changes in financial position 238 10.3 Future work with the filmadsorber 238
10.3.1 Introduction 238
10.3.2 Interesting areas for application 239 10.3.3 Further technological research and development 241
ACAC albumin collodion activated charcoal AChR "' acetyl choline receptor
ADP adenosinediphosphate
AGA acinetobacter glutaminase aspar~g~nase AIDS acquired immune deficiency syndrome ALAT alaninaminotransferase
AN acrylonitrile AP alkalic phosphatase ASAT aspartataminotransferase BE base excess
BSA bovine serum albumin CnBr CAPD DMAEMA DNA dsDNA EEG GABA GBM GOT GPT HBAb HDL HEMA IgA INDH LDL MCH MCHC MCV MMA PAN pC02 PCV PEG PEI PMMA PVA RA RBC rpm SAC SLE st HC0 3 TRCS UDP VLDL WBC cyanogen bromide
=
continuous ambulant peritoneal dialysis~ dimethylaminoethylmethacrylate - desoxyribonucleic acid
double stranded DNA
= electroencephalogram gamma amino butyric acid
=
qlomelur basement membrane=
glutamate oxalacetate transaminase=
glutamate pyruvate transaminase hepatitis B antibody= high density lipoprotein
=
hydroxyethylmethacrylate= immune globulin A (also E,G,M) ~ indolyl 3-alkane a hydroxylase
low density lipoprotein
=
mean concentration haemoglobin=
mean cell haemoglobin=
mean cell volume methylmethacrylate polyacrylonitrile partialco
2 tension=
packed cell volume (haematocrit) polyethyleneglycol= polyethyleneimine polymethylmethacrylate
=
poly vinyl alcohol=
rheumatoid arthritis red blood cells=
revolutions per minute=
staphylococcus aureus cowans= systemic lupus erythematosis
=
standard HC0 3=
thrombocytesuridine-5'-diphosphate very low density lipoprotein
1.1 Principles and techniques of blood purification
The idea to eliminate toxic or pathogenic substances from the body has been a fundamental goal of medicine throughout its existence. In its oldest concept as bloodletting it has often been prescribed to cure nearly every ailment (Klinkmann, 1983). The exploitation of bio-logical laws in combination with developments in technology enabled the multidisciplinarily collaborating clinicians, scientists and en-gineers to develop different detoxification and correction methods for clinical use during the last 50 years.
They are all based on the rationale that due to diseases of organs or organ systems, toxic and harmful substances accumulate in the body. Body fluids, both within the cells (intracellular) and outside the cells (extracellular), should therefore be purified or corrected to maintain life. Only some of the body fluids are easily accessible: blood through the circulation system, gastro-intestinal fluid through the stomach and blood-ascites-lymphatic fluid through the peritoneum. Blood purification by passing blood from the circulation through an extracorporeal device constitutes about 80\ of the routinely used procedures at the moment. Its relative safety and effectiveness in comparison with accessing other body fluids are the reasons for this. However, as a consequence of intravascular blood purification, other body fluids are purified partially and with a certain time differen-ce. This is due to limitations in diffusion into the intravascular
space of the substances to be removed.
Three basic principles of physics are applied in blood purification techniques: adsorption, diffusion and convection (fig. 1.1).
Haemodialysis is based on diffusion of uraemic toxins and convection of water and toxins through a semipermeable membrane. It is applied in about 75\ of all treatments for renal failure. Peritoneal dialy-sis, applied in approximately 25\ of renal patients, uses the perito-neal membrane for diffusion of toxins and osmosis of excessive water. With haemofiltration body fluids are filtrated through a membrane (e.g. 20 1), while simultaneously replacement fluid (e.g. 18 1) is
PERITONEAL DIALYSIS HAEMODIALYSIS
fig. 1.1 Principles and techniques of blood purification returned to the patient. Ultrafiltration is sometimes applied to re-move excessive water only. Therapeutic plasmapheresis employs such large pores in its membrane, that only blood cells are retained. All proteins, including the pathogenic, are removed. With haemoperfusion blood of the patient is simply perfused over an adsorbent, which can retain toxin(s) unselectively, selectively or specifically. Dialysa-te, haemofiltrate or plasma can also be perfused over adsorbent(s) for regeneration, if techniques are combined. In fig. 1.2 a survey of different combinations is given. Although advantages of the different techniques can be taken together, clinical use of integrated systems has been limited, due to high cost or insufficient function.
1.2 Application and adequacy of blood purification
Chronic renal patients undergoing twice or thrice weekly haemodialy-sis or baemofiltration need good blood access. Main types of vascular access in these patients are the Quinton-Scribner shunt, the Cimino-Brescia arterio-venous fistula and arterio-venous grafts. A typical haemodialysis blood circuit is illustrated in fig. 1.3.
fig. 1.3 Typical haemodialysis blood circuit
With the use of the a-v fistula, blood is pumped from the patient via his access site through the dialyser and back to the patient. Pressu-res in the blood are usually measured between the patient's access and the blood pump and at the venous air trap located downstream of the dialyser. Heparin, for anticoagulation, is typically infused downstream of the bloodpump. The dialysis fluid circuit will not be discussed here, since it is not used in haemoperfusion.
In principle the haemoperfusion blood circuit is the same, except that in acute circumstances (intoxications, acute liver failure) per-cutaneous catheterisation of the femoral or subclavian vein must be used for blood access.
Two main components are to be included in the concept of adequacy: efficiency and patient comfort (Funck-Brentano, 1985). Although inde-pendent of one another, they have mutual feedback.
a. ;!!!~!~!!~~
For decades the blood urea nitrogen level has been considered as the best means of evaluating the degree of efficiency of a dialysis treatment. Recently it has been demonstrated that with urea kinetic modeling the optimal "dialysis dosis" can be calculated (Gotch et al, 1985). In the past decade "middle molecules• were also attributed to be important for efficient removal, but evidence for this has not been given yet. Now, it is well known that high convective transfer
must be combined with diffusive transfer to obtain optimum clearing efficiency in uraemic patients.
Efficiency of haemoperfusion in acute poisoning is merely assessed by
a judgement. of clinical improvement based on subjective criteria in general. Only few groups have studied the adsorptive capacities and the extraction efficacies of haemoperfusion devices. In addition, the therapeutic efficiency of haemoperfusion is dependent on the distri-bution of toxins over the body compartments, which is rarely taken
into account.
Much more complicated is the efficiency of treatment of acute hepatic failure. There is no single endogenous toxin, which has an absolute correlation with the state of the disease. Other clinical measure-ments for determining, e.g., the residual functioning liver mass are most difficult to perform. Moreover, little is known about production rate of liver toxins in relation to liver de/re-generation. Suitable animal models for studying etiology and treatment of acute hepatic failure are difficult to establish.
The latter is also manifest in immunologically related diseases.
Besides, treatment by plasmapheresis/plasmaperfusion have begun re-cently and in mostly uncontrolled clinical studies.
b. ~~~!~!!~-~~~!~~~
Patient comfort includes good biocompatibility of the material and the maintenance of a physiological humoral equilibrium. In chronic
treatment it also includes selection of a procedure that is
conveni-ent for the paticonveni-ent's daily life.
~!~~~!~~!!~!!!~~
According to Klinkmann (1984) there is still no material available that meets all requirements for biocompatibility:
- absence of thrombogenic, toxic, allergic or inflammatory reactions; - no destruction of cellular elements;
- no immunological reaction; - no carcinogenic effect;
- no deterioration of adjacent tissue.
Therefore, chronic haemodialysis still causes side-effects in some
patients. Leukopenia and complement activation occur, particularly in cellulosic membranes thereby resulting in hypersensitivity reac-tions at the start of treatment. Spallation of particles, residuals of sterilants and desinfectants and aluminium are also related to problems. In haemoperfusion over charcoal and resins these problems are of minor importance since most haemoperfusions are performed a few times only. In plasmaperfusion over immunoadsorbents, particular-ly protein A, side-~ffects were very seve~e .. This is mostly attribu-ted to release of ligands and complement activation.
!~~~~~~-~!~!~2!~~!-~~~~g~!!!~~!~!
This problem mainly occurs in haemodialysis treatments due to too high water and sodium transfer.
In this work activated charcoal is studied as adsorbent in blood purification. In chapter II it is reported that this adsorbent is unselective when brought into contact with body fluids or dialysate which contain various endogenous substances. In this section the question is discussed whether unselective, selective or specific removal are desired or not. In a simple train of thoughts one might first ask the question: How many diseases do we understand complete-ly? The answer might be given in fig. 1.4a. This does not represent quantitative data. The next question to be asked would be: In how many diseases circulate accumulated toxins, harmful to the body?
-unknown known-known no toxins-unknown unknown known no toxins no purification toxins no puri-fication purification
fig. 1.4 Train of thoughts
unknown -taxi ns purification unknown known known taxi ns no toxins no purification known taxi ns purification known taxi ns no adequate purification known unknown unknown taxi ns unknown toxins no adequate purification unknown - - - t o x i n s adequate purification toxins adequate purification
The answer might be represented in fig. 1.4b. Then the consideration can be: In which diseases is blood purification applied as (partial) treatment? A possible answer is given in fig. 1.4c. A final question to be asked in this series might be: In which diseases is blood puri-fication applied as (partial) treatment adequately? As appears in fig. 1.4d there might be only few diseases, that we treat with blood purification adequately. There might be more diseases, that we do not treat adequately and also some diseases with circulating harmful sub-stances that we do not yet treat.
In uraemia it is not yet clear which substances must be removed. Hae-modialysis is not always adequate. Combined haemodialysis-haemoperfu-sion has given some improvements compared~to haemodialysis alone. In acute hepatic failure it is not clear at all which substances should be removed. Yet, some success is achieved, when this disease is treated with unselective techniques, such as plasmapheresis or haemoperfusion.
If acute poisoning occurs, sometimes haemoperfusion has to be started even· when there is no time available for determination of the toxin. The use of an unselective adsorbent is advantageous in these cases. In many auto-immune diseases the etiology is not known. Unselective plasmapheresis and selective immunoadsorption have demonstrated some benefits. Specific immunoadsorption, though in its infancy due to engineering problems, might be the treatment of choice in diseases in which specific pathological substances are known.
In conclusion: since not all diseases are completely understood and in aost cases the exact pathological substances are not known, blood purification with unselective and selective techniques will remain the best alternative for the present. However, techniques with speci-fic removal can also help to solve questions related to etiology. 1.3 Objectives of the filmadsorber proiect
The haemoperfusion devices containing activated charcoal or resins that are currently available are capable of removing unknown patholo-gical substances from patients. However, uptake rate and adsorption capacity are rather small, which is mainly due to the large size of the sorbent particles used.
, \
was first advocated by Davis et al (1974). Van Zutphen (1975) demon-strated that powder charcoal embedded in collodion film has excellent · in-vitro adsorption characteristics. However, his manufacturing
sys-tem for the so-called filmadsorber was not able to produce a mechani-cally strong and clinimechani-cally safe film. Moreover, he made no effort to optimise the filmadsorber by varying any of the design parameters. Verkooyen proposed some changes in the construction in order to im-prove mechanical strength in 1978 (unpublished results). In compari-son with commercially available haemoperfusion devices, the filmad-sorber demonstrated superior in-vivo adsorption of bile acids, when referred to the amount of charcoal used (Tangerman et al, 1980 a and b).
Although several groups demonstrated the theoretical advantages of extremely small particles, in the past decade no haemoperfusion sys-tems employing extremely small particles have been commercialised. As a consequence of these developments it was decided to undertake further research and development with the filmadsorber concept. The objectives of the total project were:
- optimisation of device geometry with respect to adsorption perfor-mance;
- development of a new manufacturing system, including filmproduc-tion, quality control and sterilisation according to international-ly accepted rules for Good Manufacturing Practice of sterile medi-cal disposables;
- in-vitro and in-vivo evaluation of device performances in different areas of application;
realisation of a marketing plan and feasibility study of commercia-lisation;
exploration of new areas of application using the filmadsorber con-cept;
1.4 Programme selection
The programme of work adopted for this thesis was designed to comply with the project objectives:
a. A review of the various adsorbents which have been employed in haemo- and plasmaperfusion in order to explore new areas of appli-cation using the filmadsorber concept (chapter II).
b. Kinetic modeling of haemoperfusion using both a structural and a formal model of the adsorption process; a kinetic model of the li-ver's excretory function (chapter III).
c. Design of a new manufacturing process (intermediate system) and subsequent design of a prototype of a semi-automatic manufacturing system, including a production machine, controlsystem and sterili-sation system; device characteristics (chapter IV).
d. In-vitro evaluation of column performances: adsorption characte-ristics and recirculation experiments in human plasma; haemodialy-sis-haemoperfusion in series; a single aspect of bloodcompatibili-ty (chapter V).
e. In-vivo evaluation of column performances: discussion on animal models of acute hepatic failure and test in two of them; a short biocompatibility test in a healthy pig (chapter VI).
f. Market and marketing study, financial plan (chapter VII). g. General discussion and conclusions (chapter VIII).
II. HAEMOPERFUSION
2.1 Historical background
The concept of purifying blood by extracorporeal passage over a sor-bent is not new. Apparently, the first to experimentally perfuse blood through adsorbents were Muirhead and Reid (1948). They tried to reduce the level of urea by passing the blood of dogs with experimen-tal acute renal failure through a cation exchange resin. They were followed by De Marchi and Bronniman, (1951} and Bronniman and Pini, (1955), who employed these sorbents clinically. Potassium was removed from dogs by Kessler et al (1953}, Schechter et al (1958a}, Me Laugh-lin et al (1959) and Cohn et al (1960) and cLaugh-linically by Sandler et al (1961). Directed towards the treatment of hepatic coma was the re-moval of ammonia by cation exchange resins in dogs (Schechter et al, 1958 b and Chaudhry et al, 1962) and clinically (Nealon and Ching, 1962). Among the studies dealing with the treatment of drug overdose were investigations on the removal of barbiturates in dogs by anion exchange resins (Cohn et al, 1954 and Palotta and Koppanyi, 1960). Despite the early emphasis on ion exchange resins, there is little doubt that the impetus which led to a sharp increase of haemoperfusi-on activity came from the work of Yatzidis (1964) and Chang (1966) with activated charcoal. Yatzidis directed attention towards activa-ted charcoal as a haemosorbent and Chang provided important guide-lines for the development of haemoperfusion systems.
The use of charcoal in medicine is very old, beinq mentioned in an Egyptian papyrus from 1550 BC (Hassler, 1951). At the time of Hippo-crates, wood chars were used to treat various ailments. However, the adsorptive properties of activated carbon were not appreciated until the twentieth century.
2.2 Activated carbon/charcoal
Activated carbon is a highly porous material prepared by carbonisati-on and activaticarbonisati-on of a source material. The source materials include:
bones, coal, coconut shells, fruit pits, graphite, lignite, peat, pe-troleum, pulp mill blackash, seaweed, sugar and wood. The variations in material and the activation process offer scope for varying the properties of the product (Asher, 1980). The first step in the prepa-ration of an activated carbon is usually the carbonisation of the source material, which is pyrolysis with the exclusion of air. This step is followed either by controlled oxidation with air, carbon di-oxide and steam (physical activation) or by addition of substances capable of restricting the formation of tar (chemical activation). The activation process creates a large number of pores so that the total surface area of the pore walls, i.e. the internal surface of the carbon, is very large and this is the principle reason for its effective adsorption capacity. In comparison, the area of the outer surface of the carbon, the geometrical surface, is almost negligible.
The terms •activated charcoal", •activated carbon• and "active car-bon" are generally used interchangeably. Since almost all charcoals in modern practical use have been purposely activated, the adjective •activated" is most appropriate and will be employed hereafter. Con-cerning the second half of the term, to many the designation "char-coal" implies vegetable or animal origin and it is argued that •car-bon• is a better term, as it includes all source materials. However, regardless of the source, no charcoal is purely carbon, but is rather a combination of carbon plus many impurities. Thus the term "carbon" is not strictly correct. For this reason and because of its overwhel-ming traditional usage in the medical literature, the term •activated charcoal" is used throughout this work.
The adsorptive properties of activated charcoal are determined not only by its porosity (specific surface area, pore shape and pore size distribution) but also by its particle size distribution and its che-mical composition. Disturbances in the microcrystalline structure of the activated charcoal caused by the activation process, can influen-ce the adsorptive properties, particularly for polar or polarisable
substances. Also non-carbon or hetero-atoms arlsinq from chemical binding or ash can have their influence. Chemically bonded substances
may be derived from the source material and remain in the structure
of activated charcoal as a result of imperfect carbonisation. Additi-onally, substances may become bonded to the surface during activation or as a result of oxidation of the charcoal surface. Numerous types of functional groups can be present on the surface of activated char-coal. These include carboxyl groups, ether, peroxide and ester groups. Thus, an activated charcoal can fall anywhere in the spectrum of nonpolar to polar sorbent, though being a nonpolar sorbent in principle. It will preferentially adsorb nonpolar solutes from a po-lar solvent. Of significance is the general rule of mutual affinity of substances. On the basis of the forces involved in binding the sorbate to the sorbent surface, adsorption can be separated into phy-sical adsorption and chemical binding. The explanation of the forces of attraction is not discussed here.
Activated charcoals, particularly those with oxygen on the surface, may also have ion exchange activity. Finally, it should be mentioned that a sorbent-sorbate reaction may lead to conversion of sorbate to other products.
Although not so relevant to this work, the different types of granu-les employed in most haemoperfusion devices are summarised briefly: • Irregular-shaped granules: These are produced by breaking and
sie-ving Chang Smith col)
activated charcoal into a narrow range of particle sizes. (1969) used coconut-based granules in his first experiments. & Nephew Ltd introduced the first commercial device (Haemo-on the market, which c(Haemo-ontained irregular-shaped coc(Haemo-onut-based granules. The macropore structure, which is characteristic for this type, is formed during carbonisation.
• Spherical granules: This type can be obtained by starting with a non-spherical source material and introducing a sphere-forming sta-ge to the granule penetration or by carbonisation and activation of a spherical source material. Haemoperfusion has primarily utilised spherical charcoal granules derived from a petroleum source materi-al (Amano et materi-al, 1978; Odaka et materi-al, 1980; Nakabayashi et materi-al, 1980).
Spherical charcoal granules derived from polymeric beads have been investigated by Rosenbaum et al, 1978 and 1980.
• Extruded· granules: This type of charcoal is produced by extruding previously carbonised powdered material with a binder, followed by carbonisation of the binder and activation. In this case, any large macropores present in the original carbonised material are destroy-ed by grinding the base powder. Therefore, the largest pores in ex-truded charcoal are the interstices between the charcoal particles. Extruded charcoals are more regular in shape and size and have a relatively smooth surface.
Yatzidis (1964) investigated haemoperfusion over uncoated charcoal granules in the clinical treatment of uraemia and acute poisoning. It was folowed by other workers in patients (Dunea and Kolff, 1965), in rabbits (Ragstam et al, 1966) and in dogs (De Myttenaere et al, 1967; Barakat and MacPbee, 1970). These early investigations demonstrated that haemoperfusion offered the following potential advantages: - simplicity of the method;
- removal of certain drugs, which were difficult to remove otherwise; - removal of solutes implicated in uraemia, thereby improving the
symptoms of renal failures.
However, these advantages were surpassed by two disadvantages resul-ting from the contact of blood with uncoated charcoal granules: - damage to the blood, in particular severe platelet depletion; - generation of charcoal microparticles leading to deposition in
va-rious organs.
In order to alleviate these problems the approach of coating the gra-nules with a polymer has been adopted.
An alternative to polymer coating is to produce an activated charcoal
with a hardness and a pore size distribution, which minimises blood damage and fine particle generation. Such sorbents are available in the form of bead-shaped charcoal (Denti and Walker, 1980).
Another alternative was proposed by Hill et al (1976 and 1977). They immobilised charcoal qranules on top of a polymer film in order to minimise fine particle generation. This approach has been adopted in
the production of the Becton-Dickinson Hemodetoxifier. In this devi-ce, granules of 0.3-0.8 mm are sandwiched between layers of adhesive and wound into a spiral coil. Althouqh described as an undoated fixed-bed charcoal device, Chang (1977) suggested that the granules might be coated by an ultrathin chlorosulphonated polyethylene ~dhe sive layer. The alternative employed in this work is the incorpo;rati-on of the sorbent within a polymer film.
In contrast to the early development period of haemodialysis, when the membranes were almost exclusively regenerated cellulose, there has been no general agreement on the type of polymer coating to be used in haemoperfusion over activated charcoaL Coatings have varied in chemical composition, hydrophilicity and thickness. The performan-ce of polymer-coated charcoal granules is influenperforman-ced by the coating, coating technique, nature of the qranules and preparation of the gra-nules prior to coating. The importance of the latter has been well recognised by washing the granules before and after coating (Van Wa-genen et al, 1975; Fennimore et al, 1975 and 1977; Chang, 1976; Wal-ker et al, 1976; Tijssen et al, 1980).
The pioneering work of Chang has strongly influenced haemoperfusion over polymer-coated granules (Chang, 1964, 1966, 1972a and 1977). He evaluated different polymers as membrane materials for the encapsula-tion of suspensions or soluencapsula-tions in the preparaencapsula-tion of artificial cells: nylon, collodion, cellulose acetate, radiation-grafted heparin-cellulose and silicone. He selected collodion (cellulos~ ni-trate) as the coating for the charcoal granules and also adopted the approach of albumin pretreatment for improvement of biocompatibility. The charcoal granules are coated with collodion by stirring the gra-nules in a solution of the polymer in a mixture of ether and ethanol. The granules are dried, washed and steam-sterilised. Coating thick-ness is quoted as 0.05 pm. After cooling, a supplementary coating of albumin is adsorbed onto the collodion. This albumin coating, Which is an essential feature of the ACAC (Albumin Collodion Activated Charcoal) system, does not require immobilisation with glutaraldehyde (Chang, 1977). Chang has used his procedure to coat irregular-shaped,
coconut-based and spherical petroleum-based granules. Odaka et al (1978) coated spherical granules with collodion using Chang's proce-dure. Amano et al (1978) applied collodion with a spray coating tech-nique to coat bead-type petroleum charcoal with 0.5 pm collodion by
spraying a 0.1\ pyroxylin ethanol solution. The charcoal granules are hard and have a smooth surface. Asahi's Hemosorba cartridge is based on this approach.
Andrade et al (1971 and 1972) selected polymers with low interfacial energy and prepared three different coatings: glutaraldehyde cross linked albumin, polyHEMA and Hydron Biomedical Polymer prepared by cross-linking polyHEMA. This group conducted extensive washing and washability studies of USA charcoals. In perfusion experiments with sheep they demonstrated a 20-50\ platelet drop with albumin coated charcoal and 20\ or less with polyHEMA and Hydron Biomedical Polymer coated charcoal. The latter one was later used in the Hydron Hemoper-fusion Cartridge, produced by Kuraray.
Fennimore et al (1975) adapted a concept similar to that of Andrade, which led to the first commercially available adsorber: the Haemocol of Smith & Nephew Ltd (UK). PolyHEMA is applied by a spray coating procedure (Watson, personal communication 1983) on irregular-shaped coconut-based granules. The coating thickness should be 5 pm with a coating weight of 2\.
PolyHEMA has been used as a coating for spherical charcoal granules by immersing the granules in a solution of the polymer in ethanol or by contacting the granules with a solution of the monomer, followed by polymerisation (Nakabayashi et al, 1980).
In another utilisation of HEMA (Stefoni et al, 1979} an attempt has been made to increase hydrophilic! ty, while preventing deterioration in mechanical strength. HEMA was copolymerised with an unsaturated heterocyclical compound that is highly hydrophilic in the presence of appropriate cross-links and an additional monomer providing hydropho-bic and elastomeric properties (Stefoni et al, 1982).
At the University of Stratchclyde much experience has been gained in membrane synthesis and evaluation of both haemodialysis membranes and haemoperfusion coatings. From the synthesis of haemodialysis membra-nes the advantages of a copolymer system had been demonstrated in
which one monomer contributes to sensitivity or reactivity and the other monomer to stability and strength (Courtney et al, 1978; Court-ney and Falkenhaqen, 1984). Polymerisation of DMAEMA confers water sensitivity and cationic properties. Therefore, attention was focus-sed on AN and MMA as copolymers. AN gives strength and water-insensi-tivity but is toxic; MMA is less soluble in water, less polar and less toxic. Norit RBX1 charcoal granules were coated with DMAEMA-AN and DMAEMA-MMA by dissolving in a mixture of ethanol and acetone and addinq the solution to a rotating glass vessel containing the granu-les. Preclinical evaluation suggested that the DMAEMA-MMA coating is most suitable with respect to solute removal and blood compatibility. The deacetylation of cellulose acetate to cellulose is an established procedure for the manufacture of regenerated cellulose haemodialysis membranes. In a similar manner, activated charcoal granules can be coated with a form of cellulose by applying cellulose acetate and then deacetylating, an approach first reported by Denti et al (1975). They placed charcoal granules in a rolling pan and spray-coated them intermittently with 3% cellulose triacetate in 9:1 V/V chloroform-ethanol solution. Subsequent deacetylation was achieved by hydrolysis with a 5% KOH aqueous solution treatment.
A modification of the approach of Denti et al has been used by Gam-bro, Sweden, in the production of the Adsorba haemoperfusion device (Thysell et al, 1976 and Martin et al, 1977). In this case, the cel-lulose acetate is applied by spray-coating with a solution of the
po-lymer in acetic acid and the deacetylating agent is sodium hydroxide. The coating weight has been designed to be 2% (membrane thickness 3-5 ~m) for peat-based extruded charcoal granules.
In Rostock activated charcoal was coated with a solution of cellulose acetate in acetone to give coating weights of 0.5 and 1.0\.
Subse-quently, a different coating technique has been adopted. The cellulo-se acetate is dissolved in a 9:1 mixture of chloroform and ethanol and the charcoal granules are spray-coated in a rotator to yield the coating weights of 0.5 and 1\ (Courtney et al, 1978).
An ultrathin cellulose acetate coating technique has been developed by Tijssen et al (1980). Peat-based, extruded granules are spray-coated with a solution of cellulose acetate in acetone to give a
ba-sis of the Hemopur 260 device produced by Organon Teknika. Other
ethyl 1981al
polymer coatings which have been used include those based on cellulose {Odaka et al, 1980), polysiloxane (Piskin et al, poly(ethylene) glycol (Piskin and Ozdural, 1981b) poly propy-lene oxide (De Koning, 1982) and silicone (Otsubo et al, 1983). Seve-ral types of coatings have also been developed in the USSR (Lopukhin and Molodenkov, 1979; Nicolaev and Ternovoy, 1983 and Nicolaev, 1984) and in China (In unpublished survey Van Berlo, 1984).
fig 2.1 Davis' device
The first report in literature about the use of activated powder charcoal in a haemoperfusion device is from Davis et al (1974). Powdered activated charcoal (Norit USA) was cast into a polymeric filament which is crosswound onto a cen-tral mandrel. Blood enters down the per-forated central core of the coil, passes radially through the coil and is collec-ted in grooves in the outer casing. In fig. 2.1 a cutaway drawing of this car-tridge is depicted. The priming volume was about 300 ml, when 100 g. of char-coal was used. A detailed analysis of capacity and kinetics of adsorption of these charcoal filaments was given by Holland et al (1977).
In 1975 Rietema and Van Zutphen publish-ed the development of the preliminary filmadsorber, describpublish-ed in de-tail in the thesis of Van Zutphen (1975). Merck activated charcoal
(particle size 40 ~m) was embedded in a 150 ~m thin collodion film. This was achieved by spreading out a charcoal collodion-ether-ethanol suspension over the cylindrical wall of a rotating drum. Simultane-ously glass beads (225 ~m) were spread over the still liquid film. When dry, after extracting ether and ethanol in a water bath, the
film was wound into a spiral with the beads acting as spacers. Malchesky et al described in 1976 the so-called coal sorbent fibers, which had been developed by Enka. Small charcoal particles were pla-ced inside hollow fibers made of Cuprophan. These fibers were similar to conventional hollow fibers employed in haemodialysers. Unlike a dialyser however, the blood passed over the outside of the fibers. The fibers were 300 to 330 ~m outside and 200 ~m inside diameter. The wall of the fibers had two distinct sections, a thin inner wall (ad-jacent to the adsorbent) which consisted of Cuprophan membrane and a thicker outer layer which was an open Cuprophan matrix.
Nose and Malchesky gave some data in 1977 on testing a typical canis-ter prototype (membrane surface area 1.1 m2, charcoal content 35 g. and priming volume 50 ml). More details about testing the coal sor-bent fibers in canisters with 114.5 g. of activated charcoal were pu-blished by Malchesky et al in 1977. They concluded that substances generally less easily dialysed but most easily adsorbed,, can be re-moved at higher rates.
Malchesky et al (1977) also reported about the additional availabili-ty of so-called sorbent and dialysing hollow fiber (Cuprophan Hollow Fiber SD: 300 ~m internal and 400 ~m outer diameter, 5 to 10 ~m inner wall of Cuprophan and 40 ~m sorbent in the outer wall) and the sor-bent dialysing tube (Cuprophan Tube SD: width of 15 cm and overall charcoal content of about 40\), see fig. 2.2.
Gurland et al (1978) were the first to test flat sheet sorbent mem-branes, produced by Enka. The membranes were built in the same confi-guration as a flat plate dialyser in the so-called Sorbiclear by
Me-dical Inc, Minnesota-USA. Surface area of the sorbent membrane unit (1.28 m2) was slightly less than that of the Cuprophan dialyser (1.40
m2) and the charcoal content was 23 g. In-vitro they found signifi-cantly higher clearances, particularly in the middle molecular weight range, with the sorbent systems. Gambro also manufactured such a de-vice, which showed similar performance characteristics (Gurland et al, 1979). In in-vitro experiments the effect of charcoal was consi-derable, although this was seriously diminished in the time course
(Ozdural and Piskin, 1980). However, in a limited number of clinical tests no particular improvements compared to conventional dialysers could be seen (Gurland et al, 1980). In two long-term crossover
stu-fig. 2.2 Scheaatic representation of the cuprophane membranes containing sorbents.
dies sorbent membrane dialysis was even less effective than conven-tional haemodialysis in controlling metabolite levels in uraemic pa-tients (Randerson et al, 1982). According to the authors, the results suggested that 23 g. of activated charcoal was insufficient to com-pensate for the reduced dialytic removal.
Klein et al (1978) incorporated activated powder charcoal, urease and macroporous ion exchange resins in much more permeable cellulose ace-tate hollow fibers. The ion exchange resin was a new zeolite exchang-er that binds NH4+ but has, unlike zirconium phosphate, no affinity f or ea 2+ . T e a1m was to ma e a system t at wou d a sorb creat1n1ne h . k h 1 d . . and convert urea to sodium bicarbonate. Fibers were prepared by eo-extruding a solution of cellulose acetate in acetone/formamide
sol-vent, and a dispersion of the activated charcoal (Norit SX and Norit USP), urease and zeolite. The fibers were tested in-vitro in saline and were shown to provide adequate capacity and to operate within a time scale comparable to haemodialysis. The zeolite had higher speci-ficity and capacity for ammonium than zirconium phosphate. oxystarch and oxystarch derivatives were tested for direct urea adsorption and were found unsuitable for this application.
The encapsulation of powder adsorbents in agarose beads has yielded another alternative for haemoperfusion with activated powder char-coal. Holloway et al (1979) describe a method to produce enlarged a-garose beads containing powder adsorbents: Powdered aqarose is prepa-red as 4-6\ w/v suspension in cold distilled water. This suspension is melted at 90°C and powder adsorbent is added. The mixture is drawn into a syringe and injected into an ice-cold mixture of toluene (1000 ml), chloroform (400 ml) and hexane (200 ml). The size of the beads can be regulated by syringe outlet or injection rate. The enlarged beads were initially produced to improve biocompatibility of agarose beads by binding albumin covalently (LOsgen et al, 1978). Harstick et al (1979) compared the adsorptive properties of these beads (5-6 mm), containing Merck activated powder charcoal in 2 hours incubation ex-periments with large granule adsorbents. They found the agarose-encapsulated activated powder charcoal to be much more efficient for drugs. The same was found for the adsorption of bile acids (Brunner et al, 1980). Holloway et al (1981a) considered the aqarose-encapsu-lated adsorbent system as a single step plasmapheresis and plasma treatment, since proteins could also diffuse through the matrix. This should be an advantage for adsorbinq protein-bound substances. Agaro-se gel is known to have an exclusion limit for components in excess of 106 daltons. However, when large adsoibents are encapsulated in aqarose beads they loose a lot of their efficiency. This should be due to a kinetic barrier at the gel/adsorbent surface (Holloway et al, 1981b). The reasonable blood compatibility was reproduced by a study of Dreier et al (1981), who saw slight haemolysis and modest. thrombocytopenia and leukopenia in in-vitro experiments ..
A different approach was proposed by Lauterburq et al in 1979. They attached activated powder charcoal onto glass beads for use in
plas-maperfusion. With this concept they wanted to avoid the relatively poor biocompatibility of the haemoperfusion systems developed so far. Moreover, they were looking for a better adsorbent for bile acids
that should be responsible for the disabling itching in patient with inoperable severe cholestasis. They compared different adsorbents: USP charcoal (mesh 100, Mallinckrodt Inc.) was added at a ratio of 3: 1 (w/w) to a slurry of a 10\ (w/v) solution of polyvinylpyrrolidone in water and glass beads to form the charcoal-coated beads; charcoal mesh 100-120 and different resins; covalently bound human albumin on-to CnBr activated Sepharose 68; covalently bound anti-cholyl glycine antibodies onto CnBr activated Sepharose 48. USP charcoal-coated beads demonstrated a high extraction of bile acids for a prolonged period of time. The extraction of bile acids by immobilised albumin rapidly dropped. The antibody column would have been comparable to the charcoal coated beads if 3 times more antibody could be bound per g. of gel without steric hindrance of the antigen binding. For the resins and the various charcoals the particle size appeared to be the main determinant of their bile acid adsorbing capacity. From studies with other bile acids the affinity for charcoal appeared to increase with the number of hydroxyl groups. The charcoal also adsorbed a sub-stantial fraction of unconjugated bilirubin, thereby demonstrating a higher capacity than the immobilised albumin. comparable albumin co-lumns prepared by Plotz et al (1974) appeared to be saturated after between 200 and 300 ~g of bilirubin had been adsorbed per g. of gel. The results of this study are in strong contrast with those of Hew et al (1978) who didn't see efficient removal of bile acids and biliru-bin with Becton-Dickinson's Hemodetoxifier.
Pineda et al (1984) recently published the results of the use of their charcoal coated glass beads column. At 34 patients with intrac-table pruritus a total of 250 plasma perfusions were performed with an average plasma volume perfused of 4075 ml. The averaqe retention by the column was 565 pmol of bile acids. An average of 373 ~mol of bile acids was calculated to be mobilised from the extravascular com-partment per procedure. Twenty-nine patients experienced a beneficial therapeutic effect with relief of pruritus,that varied from slight to complete and lasted from 24 h to 12 months (averaqe 9 weeks).
filmad-sorber was continued 4 years after the first paper on the preliminary filmadsorber. Tangerman et al (1980 a and b) demonstrated the superi-ority of small charcoal particles in the removal of bilirubin and bi-le acids from dogs, when compared to commercial adsorbers. Further studies (Van Berlo et al, 1982 a - b, 1983, 1984 a - b and 1985 a, b - c) are described in detail in this thesis.
Gundermann et al published the results of a study in which a new Enka Cuprophan coal sorbent (CS) fiber had been tested for the treatment of hepatic failure (1981, 1982 and 1983). The fiber consisted of a 335 ~m thick core and a 16 ~m thick cellulose wall. The charcoal con-tent was 32.7 g/kg. The CS fiber was wound in a cross coil onto a 6 cm diameter core. In in-vitro experiments blood of sacrificed dogs, which had acute hepatic failure by devascularisation of the liver, was perfused over different columns. The high-molecular weight enzy-mes GOT, GPT and AP remained unchanged. Uric acid and creatinine were adsorbed quickly by all columns (including Haemocol, Hemosorba and Adsorba). Proteins, bilirubin, free fatty acids and free phenols were poorly adsorbed. Most amino acids were removed within the first hours by the Hemosorba and the CS fibers. After 6 hrs all the charcoals showed a similar adsorptive amino acid profile. The fiber manifested the best in-vitro haemocompatibility.
Cooney and Kane (1982) also elaborated on the original idea of Riete-ma and Van Zuthphen (1975). Different activated powder charcoals were tested, from which Amoco PX-21 appeared to be the best. Their thin-filmadsorber was manufactured as follows: Ethyl alcohol was added to charcoal in a ratio of 10 ml to 3 g; 25 ml. of collodion (4 wt\) was added to this mixture. A film knife (thin metal block with an open interior) was employed to spread the film. There are 8 slots, with gaps ranging from 0.0125 to 0.125 cm in depth, on the top and bottom of each wall of the film knife. The films were spread on a long glass plate and the knife was pulled at a constant rate of roughly 1 cm/sec by means of a wire attached to the rotating shaft of a drive motor. Ether evaporates from the film, which is sprayed with distilled water to prevent excessive drying. Once the film had been made it could be cut and peeled off the qlass plate. Two different film and spacer ar-rangements were used: a mesh fabric (155-200 ~·thick) which was
wound along with the film to provide a uniform gap for flow and large charcoal particles (30-35 mesh) which were sprinkled on the film sur-face immediately as it was spread. The first one caused a periodical-ly occurring problem: channeling in the devices when they were wound too loosely and conversely, too high pressure drops when wound too tightly. The second one, based on the idea of Van Zutphen (1975), who used glass beads as spacer material, and the idea of the B-D Hemode-toxifier with granules stuck to a strip of polyester, gave better re-sults. However, problems with nonuniformity of laboratory production still caused channeling , resulting in only partly perfusing the co-lumns. Yet the thin-filaadsorbers proved to have superior or similar initial removal rates compared to commercial devices, in-vitro. In further tests (Cooney and Kane, 1983) pressure drops were low, but the priming volume only marginally acceptable. Significant flow non-uniformities also existed in tracer studies. Despite this, the over-all mass transfer resistance values were lower than those for commer-cial columns. It was measured that the slowest diffusion step in the charcoal-loaded film involves the charcoal particles themselves. Activated powder charcoal has been immobilised into polymer gel par-ticles by Ozdural et al (1982). Powder charcoal and sodiumalginate
(10-1 or 10-2.5 g.) were blended and mixed with water. Droplets were produced by syringe pump extrusion and put in a calcium chloride so-lution for 2 hrs. The exchange of sodium ions with calcium ions cause the gel formation. Then, these elastic charcoal spheres weretreated with poly-1-lysine and PEI to increase stability. Capacity and rate of adsorption for different test solutes were higher than those ob-tained for large charcoal granules. However, owing to the expensive preparation of these spheres, another procedure was developed by Ki-remitci and Piskin (1984). PEG is dissolved in HEMA at 40 °C. Activa-ted powder charcoal is blended with this mixture, which is dropped into cold petroleum ether by using a syringe pump. The supercooled glassy particles in the petroleum ether are irrediated with a co-60 source to polymerise and crosslink the spheres. In the last stage the particles are washed and swollen by distilled water. Adsorption isotherms and in-vitro bloodcompatibility were superior to large activated charcoal granules.
Another way of embedding activated powder charcoal in a sheet was proposed by Kawanishi et al (1982 and 1984): A suspension of powder charcoal in polyetherurethane/tetrahydrofuran solution (8\ wt/vol} is spread over a polypropylene mesh (100 pm thick). The layer on the mesh is solidified in water. After extraction of the solvent in water at 60
°c
a porous polyetherurethane sheet- 200 pm thick- is formed, containing 70% water, 20\ charcoal and 10\ polyetherurethane. Haemo-perfusions were performed in bile duct ligated dogs. Conjugated and unconjugated bilirubin decreased markedly, from 13.2 to 2.8 mg/dl and 4.3 to 1.0 mg/dl resp. Clearances were 25 ml/min to 10 ml/min. at 50 ml/min. flow rate. Platelets dropped 28\ initially to 19\ finally, whereas leucocytes decreased from 70 to 23\. Also in this device pri-ming volume was relatively large compared to charcoal content, while channeling problems occurred as well. Recently, Kawanishi et al (1985) performed a total of 32 haemoperfusions with this device in 19 hepatic failure patients; 37\ total bilirubin and 30\ bile acids were removed and 5 patients who were treated in coma grade II, survived. An approach which looks like an intermediate between those of cross-wound filaments and films/sheets is the charcoal cloth (Courtney and Falkenhagen, 1984}. The cloth consists of woven filaments with an a-verage diameter of 20 ~m, resulting in a specific surface area of 870 m2/g. In a haemoperfusion device a nylon mesh has to be wound to-gether with this composite to provide enough space between the win-dings. Only limited adsorption experiments were done. Blood compati-bility studies (in-vivo) indicated that platelet drop with charcoal cloth is likely to be similar to that caused by granular charcoal, but that blood loss due to entrapment represented a potential hazard. An alternative geometry is therefore investigated.2.3 Resins and other adsorbents
An ion exchange resin can be regarded as a matrix containing diffusi-ble ions. The properties of the ion exchange resin are influenced by the nature of the matrix.
groups balanced by diffusible cations. An anion exchange resin has a matrix with fixed insoluble cationic groups balanced by diffusible anions. An amphoteric exchange resin has fixed anionic and cationic groups in the matrix and can exchange both cations and anions. The ion exchange resins may have an organic or inorganic matrix and, de-pending on the nature of the functional groups, are termed strong, intermediate or weak exchangers. The principle operational difference from adsorbents is that materials are exchanged rather than simply adsorbed and that commonly ionic materials can be removed.
The use of ion exchange resins in the early years of baemoperfusion has been described in 2.1. Since that time, ion exchange resins have found more application in the removal of solutes from the qastro-intestinal tract (after oral intake) or in dialysate regeneration. Direct cation exchange haemoperfusion can be dangerous because of the non-specific removal of calcium, magnesium and potassium and the in-duction of hypotension. However, it has been demonstrated (Maini et al, 1976) that with appropriate selection of resin and pie-equilibra-tion, direct cation exchange haemoperfusion may be useful in the treatment of acute poisoning. Poor biocompatibility has been an ob-stacle to the wide use of unmodified ion exchangers for haemoperfusi-on. Poly-HEMA acrylic polymers arid albumin have been used as coating material of Dowex 1x2 by Sideman et al (1977 and 1979) for bilirubin removal. Much research to improve the biocompatibility of these re-sins was done in the Sowjet Union (Lopukhin and Molodenkov, 1979). However, it seems that plasmaperfusion is still to be preferred when applying ion exchange resins.
A macroreticular resin is a polymeric adsorbent usually produced in the form of nonionic beads. The most~commonly used types in haemoper-fusion are the Amberlite materials produced by Rohm and Haas. Amber-lite XAD-2 is an uncharqed, macroreticular styrene-divinylbenzene co-polymer with an adsorbent surface area of about 330 m2Jg. XAD-4 is chemically identical but differs in physical properties with a larger surface area of 750 m2/g. These sorbents have a highly aromatic structure and are effective for removing organic solutes having hy-drophobic and hydrophilic parts. Therefore, these resins have an at
traction for lipid-soluble molecules. Amberlite XAD-2 and XAD-4 have found application in the treatment of acute poisoning (Rosenbaum, 1980; Rosenbaum and Raja, 1983). Columns containing XAD-4 have been commercialised (Extracorporeal and Braun).
Amberlite XAD-7 is a polymeric adsorbent in the form of a metacrylic acidstyrene copolymer. This sorbent derives adsorptive properties from its porosity, surface area of 450 m2/g and the aliphatic nature of its structure. XAD-7 is more hydrophilic than XAD-2 and XAD-4. In adsorption of organic molecules, the hydrophilic and hydrophobic por-tions of the solute are both adsorbed on XAD-7.
As a means of improving blood compatibility, Amberlite adsorbents have been coated with albumin (Ton et al, 1979; Hughes et al, 1979), cellulose acetate or polyelectrolyte and a cellulose-albumin combina-tion (De Koning, 1982).
Sideman et al (1980a) have carried out much research to find a good bilirubin adsorbent. A particular macroreticular resin was found to be superior to others. It was coated by albumin via cross-linking with glutaraldehyde (Sideman et al, 1981a-b, 1982 and 1983).
Pyrolysis of resins may be used to produce a carbonaceous sorbent. An example of this is the pyrolysis of an uncharged polystyrene-based
r.esin in the production of Amberlite XE-336, a sorbent with adsorpti-ve pro[>erties similar to those of activated charcoal (Rosenbaum et al, 1978).
• Alumina
Active alumina is a polar adsorbent ft>rmed by the thermal dehydra-tion of hydrated aluminum oxide. surface areas of over 400 m2
tg
have been obtained. It is capable of both chemical binding and
physical adsorption of water. • Magnesia
Active magnesia is prepared by thermal decomposition of magnesium carbonate or magnesium carbonate tribydrate with evolution of resp. carbon dioxide or carbon dioxide and water. surface areas of
100-350 m2/g have been obtained. The surface of magnesium oxide is very polar, but wide spread medical application is questionable because of its lower stability in the presence of water.
• Silica gel
Silica gel is a polar adsorbent with surface areas of about 600 m2/g. It is thought to consist of a network of sio4 with the tetra-hedral configuration of silicon maintained at the surface by hydro-xyl groups. The Si-O-H configuration produces the polar surface. • Molecular sieves
The pore structure of the alumino-silicate zeolite molecular sieves is very uniform. They consist of crystals made up of unit cells
r
with interconnecting apertures in the faces of the unit cells. The total surface area of these crystals is usually in the range of 500-750 m2/g. There is no access to the internal surface of the crystals other than through the apertures, resulting in selective adsorption based on size and perhaps shape. The surface is very po-lar and they may be equally used for ion exchange as adsorption. • Chemical adsorbents
Chemical adsorbents are very selective as they rely on the formati-on of specific chemical bformati-onds. Polyaldehydes (called oxystarch and oxycellulose, remove little urea and ammonia under physiologic con-ditions and have been used as suspensions for oral intake.
2.4 Complex sorbent systems
The most important feature of the work of Chang (1964, 1966) was the potential development of systems for using macromolecules and suspen-sions in a convenient manner, without an adverse loss of biological activity and severe immunological reactions. Artificial cell membra-nes can be formed using emulsification followed by membrane formation around each microdroplet by means of a number of approaches (Chang, 1983a). Almost any material can be occluded within artificial cells: enzymes, cofactors, immune cells, vaccins, hormones and adsorbents for detoxification. The principle of artificial cells was used to en-capsulate activated charcoal for hemoperfusion, mainly with albumin collodion (see II.2.6), and collodion alone. The latter approach, a
