WO2005099789A1

WO2005099789A1 - Treatment of sepsis

Info

Publication number: WO2005099789A1
Application number: PCT/GB2005/001501
Authority: WO
Inventors: Stephen Robert Tennison; Gary James Phillips; Sergey Victorovich Mikhalovsky; Andrew William Lloyd; James Macniven Courtney
Original assignee: Mast Carbon International Limited
Priority date: 2004-04-16
Filing date: 2005-04-18
Publication date: 2005-10-27
Also published as: GB0408521D0

Abstract

Apparatus and a method for removing endo and exo toxins from blood comprises passsing blood from a patient through a porous monolithic carbon structure to form a filtrate stream and a retentate stream and combining the streams for re-injection back into the patient.

Description

Treatment of Sepsis

The present invention relates to a novel device for the treatment of inflammatory conditions particularly sepsis and its sequalae and to materials for use in that device.

Sepsis is a major cause of morbidity and mortality in intensive care units, not only within the UK but also throughout the world. As many as 30,000 cases of sepsis occur each year in the UK alone with mortality rates being particularly high for those patients developing multi-organ dysfunction syndrome (MODS). Human sepsis can be initiated by the presence of endotoxin or exotoxin from microorganisms, parasites, exogenous substances including chemical agents, drug overdose and snake toxins or can be a consequence of trauma. Endotoxin is a constituent of the outer cell membrane of gram negative bacteria and via a series of events the endotoxin stimulates the release of a number of inflammatory mediators including cytokines and activation products¹. Exotoxins are released from the gram positive bacterial cell upon cleavage of a single peptide. They can act as potent pyrogens and superantigens, stimulating the expansion of host T cell populations and related inflammatory mediators independent of antigenic specificity. When released, these inflammatory species evoke a systemic inflammatory response, which may result in septic shock, organ failure, multiple organ dysfunction syndrome (MODS) and death. Gram negative sepsis accounts for between 50-60% of cases of sepsis while gram positive sepsis is responsible for 35-40% of reported cases. The remaining cases are the result of fungal, viral or protozoan insult². Compensatoiy anti-inflammatory reaction (CARS) may also be important in the pathogenesis of sepsis with the release of inappropriate amounts of anti-inflammatory mediators such as IL-4, IL-10, IL-11, IL- 13, IL-1 receptor agonist and soluble TNF-alpha receptors³. To counteract the process of sepsis it is necessary to reduce the systemic level of both endotoxin and exotoxin and the circulating pro and anti-inflammatory mediators. Multiple organ dysfunction syndrome represents the main cause of death in surgical intensive care units with approximately 90% mortality occurring when three or more organs fail. MODS can be divided into primary and secondary MODS depending on the time to manifestation and the associated pathophysiological events. Primary MODS occurs in the hours immediately following trauma and results from systemic whole body inflammation, either as a direct consequence of tissue damage or severe tissue hypoxia and vascular endothelial injury and is, therefore, not dependent upon the presence of bacteria or endo/exotoxins.

Secondary MODS develops as a consequence of the host response to various stimuli rather than the original insult itself. As a consequence, secondary MODS generally occurs between day 4 and 14 following the initial trauma event and is based on the systemic inflammatory response (SIRS). This inflammatory response is a result of the uncontrolled synthesis and release of pro-inflammatory cytokines such as TNF-α and, IL-lβ, IL-6 and IL-8 by the activation and continuous stimulation of tissue macrophages and monocytes. This process is tenned cytokinemia and is a result of bacteremia and/or exo/endotoxemia following bacterial invasion of the host either early after injury, through compromised skin barriers or via bacterial translocation (e.g. from the gut), or in the late post-traumatic period during infectious episodes.

Under normal circumstances, inflammation is a localised event that acts to contain and eradicate infecting organisms or to cleanse damaged tissue of cell debris. Uncontrolled inflammation, however, can become detrimental with the destruction of local tissues with eventual major tissue damage if the process continues unchecked. In addition, the pro-inflammatory response can become more widespread with the activation of multiple effector cells and humoral protein cascades such as that in the coagulation and complement systems. The intravascular response may lead to more widespread damage of the vascular endothelial system, which, in turn, potentiates distant organ failure. Ultimately, systemic inflammation becomes self perpetuating due to the continued release of locally or systemically produced pro-inflammatory mediators into the circulation.

At present the best approach to combat MODS is with the use of antibiotic therapy. However, the antibiotic-mediated lysis of bacteria may release large amounts of exo- or endotoxin into the systemic circulation leading to an exacerbation of the systemic inflammatory response. It is apparent, therefore, that the elimination of both toxins and the inflammatory mediators from the systemic circulation of patients with sepsis would be beneficial. The most suitable approach is to remove the mediators from the circulation with the greatest level of specificity possible. A number of approaches have recently been used to reduce the systemic levels of both endotoxins and cytokines. One such approach has been the use of anti-inflammatory therapy to block the actions of single pro-inflammatory cytokines. It has previously been demonstrated, using experimental models, that TNF-α, acting synergistically with other cytokines such as IL-lβ and IFN-γ, is a key mediator in the induction of the process of septic shock⁴⁽¹⁾. As a result, a number of clinical trials have been undertaken targeting the production of TNF-α or IL-lβ (e.g. anti-TNF antibodies, IL- lβ receptor agonist and IL-10). However, these trials have failed to demonstrate any significant therapeutic benefit ^,5' '⁷'⁸' ' ' ^(1"8'⁾. In addition, the use of monoclonal antibodies against endotoxin has proved unconvincing as they are only believed to bind to a specific endotoxin and are, therefore, ineffective against endotoxin derived from other bacterial strains.

It is now generally assumed that the simultaneous removal of multiple pro- inflammatory cascades is necessary if patients are to have the best chance of survival ^{5,6,7,8,9,10,11 (2}-⁸⁾ _^ ^ different approach, therefore, has attempted to increase the clearance of both endotoxins and cytokines using extracorporeal systems. A variety of continuous renal replacement therapies (CRRT) including haemofϊltration and haemodialysis have been investigated for their potential to reduce cytokine and endotoxin levels in MODS. Some studies have shown limited removal of IL-lβ, TNF-α, 11-6 and IL-8¹²'¹³'^{14 (9"π)}. However, clearance in all cases has not been sufficient to reduce plasma levels by clinically significant amounts and there is an increasing belief among practitioners that CRRT does not result in increased survival rates in septic patients with MOF¹²'¹³'^{14 (9"π)}. The failure of renal replacement therapies to significantly reduce plasma levels of cytokines and endotoxins appears to be due to the relatively low molecular weight cut-off of even high flux haemofiltration and haemodialysis membranes (~20kD)¹⁴'^{15 (11, 12)}. Other methods of removal are required, therefore, to counteract the inflammatory response in patients suffering from severe sepsis and MODS.

A more promising extracorporeal strategy for the treatment of sepsis is the use of plasma exchange to remove high molecular weight cytokines and endotoxins. A number of recent case studies have suggested that plasmapheresis may be useful to control the progress of sepsis¹⁶'^{17,18 {}-^{l3 5} This process involves filtering blood through membranes with a high molecular weight cut-off (~100kD) to separate a patient's cells from their plasma which may then be exchanged for healthy donor plasma. Preliminary data suggests that the removal of plasma from septic patients in this way may reduce the deleterious effect of the inflammatory response by the removal of both endotoxins and cytokines¹⁸'^{19 (15}' ¹⁶⁾. The removal of inflammatory mediators via this system of blood purification is poorly characterised. However, Gardlund demonstrated that acute plasmapheresis resulted in a significant clearance of TNF-α in patients with primary septic shock^{20 (17)} and Skerrett and co-workers reported the reduction of TNF-α and IL-6 by 66% and 55% respectively in patients with primary dysfunction syndrome of hepatic transplants^{21 (18)}. Plasmapheresis does, therefore, offer a potentially beneficial approach to the treatment of sepsis. However, this technique is costly (partly due to the need for replacement plasma) and its lack of selectivity results in the removal of products important for the host defence and coagulation processes, which are not replaced by donor plasma. Risks of infection and severe immune response of the patients' immune system to donor plasma proteins/antigens are also higher with this technique. An alternative method for removing endogenous levels of both endotoxin and cytokines is the use of adsorbents for haemoperfusion. In 1996, an elegant study demonstrated the extracorporeal use of microspheres with covalently-linked polyclonal antibodies for the specific adsorption of pro-inflammatory cytokines²² ^¹⁹ This microsphere-based detoxification system allowed the efficient removal of both TNF-α and IL-lβ from spiked human plasma. Such a system, however, is likely to be expensive for clinical use and would have to be used in conjunction with other adsorbents capable of binding endotoxins if it were to interfere sufficiently with the progress of sepsis. A more practical approach would be to use a less expensive, readily available adsorbent that can be manipulated chemically to enhance specific adsorption.

For a number of years activated carbons from natural sources and amberlites, have been known to have adsorbent properties and have been studied in relation to the development of extracorporeal bioartificial liver devices. Such adsorbents have been demonstrated to adsorb IL-lβ, IL-6 and TNF-α although with limited efficiency²³'²⁴'²⁵'²⁶'^{27 (20"24} Indeed, Hughes and co-workers have completed a clinical study²⁸'^{29 (}-²⁵' ⁶⁾ of existing medical haemoperfusion adsorbents and have highlighted a number of important findings. Firstly, all existing medical adsorbents have low affinities for the major pro-inflammatory cytokines (apparently due to the spatial constraints of their microporous structure which is unable to retain substances the size of pro-inflammatory cytokines (10-60kD) and the uptake kinetics for the fraction adsorbed are slow. Secondly, amberlite-type adsorbents have a very broad pore size distribution, rendering them non-selective on the basis of molecular size. Many circulating components can compete for the available binding sites, therefore making the selective binding of an individual cytokine molecule unlikely. Finally, all the adsorbents are poorly biocompatible resulting in thrombogenesis, the activation of endogenous complement or direct toxic leaching into the circulation³⁰ ^²⁷⁾. Coating the carbons to avoid direct blood contact effectively inhibits the adsorption of molecules with a molecular weight greater than 2kD and results in slow adsoφtion kinetics ¹⁹'^{28 (22}' ²⁵ .

In 1999 Steczko and co-workers³ ^³⁷⁾ reported the use of a push-pull sorbent-based pheresis system (the BioLogic-DTPF system) for the removal of cytokines and endotoxin from bovine serum albumin (BSA) solution and bovine plasma. This system combined the BioLogic-DT hemodiabsorption system in series with the BioLogic PF push-pull pheresis system. The PF membranes were used to separate plasma for direct contact between plasma proteins and the sorbents. Equilibrium binding studies in BSA revealed relatively high capacity of cytokine binding (ILl-β and TNF-ot by powdered charcoal (70-90ng/g). Kinetic binding studies in plasma also demonstrated that adsorption of the two cytokines by charcoal was relatively rapid. However, both binding capacity and rate of adsorption for the two cytokines was shown to be even greater to powdered silica, although the efficiency depended on particle size. This study also demonstrated that cholesτyramine was more efficient at removing endotoxin than either charcoal or silica, however none of the adsorbents employed in this study were specifically engineered for maximising adsorption efficiency. More recently Tetta and co-workers^ ' demonstrated an extracoφoreal coupled plasma filtration-adsoφtion system that was reported to improve survival rates in a rabbit model of endotoxic shock. The extracoφoreal circulation consisted of plasma filtration followed by the passage of the plasma filtrate through a hydrophobic sorbent and reinfusion into the venous line. This system, however, employed a non-selective hydrophobic resin as the stationary adsorbent phase. The non-selectivity of the adsorbent used will clearly limit the efficiency of such a system.

There is a requirement therefore for an efficient and cost effective extracoφoreal device for the treatment of sepsis which includes a separation system to separate the blood cells from plasma, combined with an effective adsorbent system for the removal of the inflammatory mediators including cytokines and endo/exotoxins and which would then allow the two streams to be recombined and passed directly back to the body eliminating the requirement for replacement plasma. For such a device to be effective it is also essential that the materials used do not cause either a further immune response, or excessive platelet activation (blood coagulation) on the feed side of the filter.

We have now devised a device based on the use of a new 3 dimensional carbon system which achieves this requirement.

According to the invention there is provided a method for the treatment of blood which comprises (1) passing the blood through a monolithic porous carbon structure, (2) allowing the plasma components to pass through the walls of the monolith to form two streams, a plasma permeate stream passing through the walls of the monolith and a retentate stream containing the majority of the blood cells, (3) adsorbing the contrary substances from the plasma permeate stream in the walls of the monolith and (4) combining the filtrate stream and the retentate stream.

After recombination the combined stream can be passed back into a body from where the blood was taken.

The invention also provides apparatus for treating blood which comprises (i) a porous monolithic carbon having at least one channel formed in it (ii) an inlet for the blood stream to enter the channel whereby blood passes down the channel and passes through the walls of the channel to form two streams, a filtrate stream and retentate stream (iii) means to combine the retentate and filtrate stream and (iv) an outlet for the combined stream.

Preferably there are two channels, an inlet channel which is the retentate channel and a filtrate channel into which the filtrate enters after passing through the walls of the inlet channel; the filtrate channel is connected to the outlet. The contrary substances which can be removed from the blood include endotoxins and exotoxins and the circulating pro and anti-inflammatory mediators e.g.IL-lβ, IL- 6, TNFα, IL-4, IL-10, IL-11, IL-13, IL-1 receptor agonist, soluble TNF-alpha receptors and LPS.

By "monolithic" is meant that the porous carbon is in a single piece i.e. not granular or not composed of granular carbons bound together by a binder etc. The monolithic carbon additionally contains large transport channels through which the feed /retentate steam flows and by which means the pressure drop in the device can be controlled.

Preferably the monolithic carbon has a continuous channel structure.

For a symmetrical monolith (figure 5) the continuous channel structure is defined by the channel dimension, W, and the wall thickness, t, or for an asymmetric monolith (figure 4) by R and F as well as t where R and F are dimensions of the feed (F) and retentate (R) channels. These fix the ratio of open to closed area and therefore the flow velocity along the tube.

Preferably the walls of the monolithic carbon have a macroporous structure.

By "macroporous" is meant that the carbon has continuous voids or pores through which liquid, specifically blood plasma, can pass from the feed to the retentate with a mean pore size of greater than 500mn. The macropore structure in the wall of the monolith is controlled by the particles used to form the monolith. When the tube is made from macro-particles with a mean particle size of Dp the macropore size is typically 20% of the size of the precursor resin particle. This can be varied over a wide range from a maximum particle size of approximately 10% of the wall thickness, t, to a minimum practical particle size of around 10 microns. This gives a macropore size within the wall for a 1mm wall thickness between approximately 2 and 20 microns. The pore size fixes the filtration characteristics of the wall which is required to filter blood cells from the whole blood and to allow plasma containing middle molecular weight proteins to pass through. There is an optimum particle size that will maximise the permeability of the wall whilst minimising the passage of blood cells. As the retentate and filtrate streams are ultimately recombined it is not essential that 100% of the blood cells are removed from the filtrate but a high separation is beneficial. The maximum liquid flow through the wall for a given pressure differential or flow linear velocity is defined by the Blake-Kozeny equation:-

Preferably the carbon particles making up the walls of the monolith contain micro (<2nm) meso (2-50nm) and small macropores (50-200nm) suitable for the adsoφtion of small, middle and high molecular weight molecules with a mean pore size between 0.5 and 200nm.

Preferably the monoliths are either tubular with a central channel of between 1 and 10mm and a wall thickness of between 0.2 and 2mm or are rectangular channel monoliths with a cell structure (cells per square cm -cpcm) where the channel size is between 100 and 2000 microns and the wall thickness is between 100 and 2000 microns and with an open area of between 30 and 60% to give a good carbon packing density per unit volume and acceptable mass transfer characteristics. More preferably the cell structure should be asymmetric to provide control over the flow and pressure drop in the feed and filtrate channels. It is also required that the filtrate channels be closed off at the monolith inlet.

The monoliths can be produced in lengths from around 1mm to 200cm but for use in the present invention this will depend on the use. The monolithic porous carbon can be made by partially curing a phenolic resin to a solid, comminuting the partially cured resin, extruding the comminuted resin, sintering the extruded resin so as to produce a form-stable sintered product and carbonising and activating the form-stable sintered product.

EP 0 254 551 gives details of methods of forming the porous carbons suitable for forming the porous carbon used in the present invention and its contents are included herein by reference. The process comprises (a) partially curing a phenolic resin to a solid, (b) grinding the solid to form particles, (c) forming the resulting ground product into a dough and extruding at a pressure in the range 0 to 20 MPa, (d) sintering the shaped solid so as to produce a form-stable sintered product. The sintered product can then be carbonised and activated. PCT/GB02/01142 gives details of producing monolithic structures using the sintered resin structures of EP 0 254 551 and this is included herein by reference.

Phenolic resins are well known materials. They are made by the reaction of a phenol and an aldehyde, e.g. formaldehyde. The condensation is initially carried out to produce a partially condensed product. The condensation may be carried out so as to produce a resin which is fully curable on further heating. Alternatively the condensation may be carried out so as to produce a novolak resin which is only curable when an additional cross-linking agent is mixed with it e.g. hexamethylene tetramine (known as "hexamine" or "hex"). It is preferred to use hexamine-cured novolak resins in the process of the present invention.

The resin cure should be controlled so that it is sufficient to prevent the resin melting during subsequent carbonisation but low enough so that the particles produced during the milling step can sinter during subsequent processing. Preferably the temperature and duration of the partial curing step is selected as to give a degree of cure sufficient to give a sinterable product, and being such that a sample of the partially cured solid when ground to produce particles in the size range 106-250 microns and tabletted in a tabletting machine gives a pellet with a crush strength which is not less than 1 N/mm. Preferably the pellet after carbonisation has a crush strength of not less than 8 N/mm.

By "sintering" we mean a step which causes the individual particles of phenolic resin to adhere together without the need for a separately introduced binder, while retaining their individual identity to a substantial extent on heating to carbonisation temperatures. Thus the particles must not melt after forming so as to produce a molten mass of resin, as this would eliminate the internal open porosity of the article. The open porosity (as opposed to the closed cells found in certain types of polymer foams) is believed to be important in enabling formed articles to retain their shape on carbonisation.

In one embodiment the comminuted resin particles have a particle size of 1 to 250 microns, more preferably from 10 to 100 microns. Preferably the resin powder size is between around 20 microns and 75 microns which provides for a macropore size of between 4 and 15 microns with a macropore volume of around 40%. The size of the resin is selected so as to give an optimum balance between the permeation rate of the plasma component through the filter and the separation of the blood cells from the plasma.

As disclosed in PCT/GB02/01142 the milled powder can then be extruded to produce polymeric monolithic structures with a wide range of cell structures, limited only by the ability to produce the required extrusion die and suitable dies are available commercially. At this stage the monolith has a bimodal structure - the visible channel structure with either the central channel in a simple tube or the open cells in a square channel monolith of around 100 to 2000 microns cell dimension and cell walls with thickness between around 100 and 2000 microns - and the macropore structure within the walls generated by the sintered resin particles. The carbonisation steps take place preferably by heating above 600°C for the requisite time e.g. 1 to 48 hours and takes place under an inert atmosphere or vacuum to prevent oxidation of the carbon.

On carbonisation the material loses around 50% weight and shrinks by around 50% volume but, provided the resin cure stage was correctly carried out, this shrinkage is accommodated with no distortion of the monolith leading to a physical structure identical to that of the resin precursor but with dimensions reduced by approx 30%. The macropore size is also reduced by ~30% although the macropore volume (ml/ml) remains unaltered.

During carbonisation the microstructure of the porous carbon develops. After carbonisation the monolith behaves as a molecular sieve due to partial blocking of the microstructure by the decomposition products from the carbonisation process. These blockages must be removed to provide rapid access to the internal structure of the carbon that is essential for the operation of the monoliths as adsoφtion devices.

By "microporous" is meant pores within the carbon particles suitable for the adsoφtion of low molecular weight molecules with a mean pore size of less than 2nm.

By "mesoporous" is meant pores within the carbon matrix suitable for the adsorption of middle molecular weight molecules with a mean pore size of between 2 and 50nm.

By "macroporous" is meant pores within the carbon particles suitable for the adsoφtion of larger protein bound molecules of greater than 50nm.

The micro and mesoporosity of the walls of the monolithic structures fix the adsoφtion characteristics of the carbon for the various materials that are to be removed from the plasma filtrate. The micropores are responsible for the removal of small molecules such as billirubin. Efficient removal of these molecules requires pores in the range of 0.5 to 2nm. These are generated from the resin matrix material and also partly from the secondary additive carbon.

After carbonisation the monolithic porous carbon can be activated to provide the necessary micropore volume and surface area. Activation can take place in either steam or carbon dioxide at temperatures above approximately 700°C and 800°C respectively or in combinations of these gases. The activation process is carried out for a time that varies with the temperature and the activation gas composition, such that a carbon weight loss of between 20 and 40% is achieved. Preferably the activation is carried in CO₂ at 850 to 1000°C.

The only drawback to this production route is that the phenolic resin derived carbons can only be produced with pores of approximately 0.6-1. Onm which is too small for the removal of the higher molecular weight cytokines and endotoxins and protein bound molecules that are the primary goal of this application.

Patent application WO 02/12380 A2 discloses a means for producing bead form phenolic resin derived carbons with a bimodal pore structure comprising micropores and meso/macropores using a solvent pore forming route and this patent is included herein by reference. However these materials cannot be formed directly into monolithic structures as the high degree of cure required to generate the meso/macro pores prevents the resin particles from sintering.

Patent application PCT/GB01/03560 discloses a method for the production of a meso/macroporous monolithic carbon by using a meso/macroporous phenolic resin, produced according to patent WO 02/12380 A2, in conjunction with novolak phenolic resin as a binder and this application is incoφorated herein by reference. If the meso/macroporous resin powder is mixed with a powder comprising thermoplastic novolak and hexamine curing agent, and tins is then extruded using methocell as an extrusion agent, the material can be dried, carbonised and activated to give a bimodal micro plus meso/macroporous carbon monolithic structure. The drawbacks to this route are that the novolak resin binder tends to partially infill the meso/macro pore structure of the cured resin component leading to a reduction in pore volume and the mesopore size of the cured resin in the l-5nm pore size range.

Patent publication WO 2004/087612 discloses a method for the production of monolithic structures comprising mixtures of the sintered phenolic resin, produced according to patent EP 0 254 551, with other materials such as powdered activated carbons and this is included herein by reference. According to this method it is possible to produce a monolithic structure using the partially cured resin powders of EP 0 254 551, and the monolithic structures of PCT/GB02/01142 producing channel structures where the walls comprise a mixture of micropores, derived from the resin component, and controlled structure meso/macro pores derived from the secondary activated carbon or other porous material component.

The binary route of WO 2004/087612 can also be used with mixtures of the sinterable phenolic resin powder according to EP 0254551 and the meso/macro porous resin beads of WO 02/12380 A2. In this way the carbon formed comprises the micropores from the sinterable resin component and the bead form resin combined with the meso/macro pores derived from the bead form resin. This method has the advantage that it can be used with much higher loadings of the mesoporous beads than is possible with the mesoporous powdered activated carbons.

The carbon monoliths have a surface area of at least 600m2/g, preferably in excess of 1000m2/g with a controlled distribution of micro, meso and macro pores.

The adsoφtion of cytokines, endotoxins, exotoxins and other smaller toxic molecules is also influenced by the surface chemistry of the adsorbent matrix. The phenolic resin derived carbons of this invention provide a unique degree of control over this where the surface oxygen functionality can be controlled over wide ranges through controlled activation processes using gas phase or liquid phase oxidation processes, including approaches such as electrochemical oxidation.

The invention combines the filtration properties of a novel 3 -dimensional monolithic carbon matrix with the adsoφtive potential of novel pyrolysed carbons produced with defined porosity. Phenolic resin derived mesoporous carbon beads³¹ have been demonstrated to have extremely high biocompatibility due to their high mechanical strength, which reduces the formation of 'fines,' and absence of toxic leachables. This biocompatibility allows them to be used uncoated, and so display much faster adsoφtion kinetics when compared to their nearest coated equivalents. Other polymer-based pyrolysed carbons have previously been used in the former USSR for haemoperfusing patients with acute poisoning³². We have demonstrated that these polymer based carbons efficiently adsorb the inflammatory mediators TNF-α, IL-lβ and IL-6 and we believe that such adsorbents can be used to reduce systemic levels of the cytokines and endotoxins.

This has the major advantage that there is no loss of fluid through the process and therefore there is no requirement for sterile make up fluids. This represents an entirely unique approach to the problem of sepsis and its associated pathological consequences.

It is a feature of the present invention that it can provide apparatus for control of sepsis whereby:- a) The plasma filtration characteristics of the device are controlled by the macropore structure of the monolith walls, primarily generated through control of the primary resin particle size and the thickness of the walls. b) The adsoφtion properties of the device for the removal of the inflammatory mediators including pro- and anti-inflammatory cytokines, endotoxins and exotoxins are generated by the secondary mesoporous component that can either be a mesoporous powdered activated carbon or mesoporous resin. c) The adsoφtion properties of the device for toxic small molecules such as creatinine are generated by the micropore structure of the carbon derived from the primary resin. d) The surface chemistry of the activated carbon is controlled by gas or liquid phase activation. e) The pressure drop for blood flow is controlled by the size of the transport channels in the monolithic structure.

And the apparatus is used for the control of sepsis by:- a) first filtering the blood stream to partially separate the blood cells from the plasma, b) removing the inflammatory mediators including cytokines, endotoxins, exotoxins and other toxic components from the plasma component within the walls of the filter devices, c) re-combining the cleaned plasma permeate and the partially separated retentate stream containing the blood cells, d) re-injecting the combined stream back into the patient.

These then also provide a route to the manufacture of a device for control of the sequalae of sepsis including single organ failure (including renal and/or hepatic failure) and whereby the monolithic system can be combined with existing methods of blood purification. These include plasmapheresis and/or renal replacement therapies such as continuous or intermittent methods of haemofiltration or haemodialysis, or any other method of extracoφoreal blood purification system benefiting from the addition of a separate or combined adsoφtion unit to allow:

i.) the removal of both endogenous small molecular weight substances including uraemic toxins, middle molecular weight toxins (including those bound to proteins) and exogenous toxins including drugs (both free and protein bound) taken as overdose,

ii.) the removal of pro-inflammatory and anti-inflammatory mediators including cytokines and other humoral components,

iii) the removal of bacterial toxins.

These then also provide a route to the manufacture of a unique device for the removal of circulating antibodies and antigen-antibody complexes which contribute to the pathology of various auto-immune diseases such as Myasthenia gravis, systemic lupus erythematosus, Waldenstrom's macroglobulinemia and Guillain-Barre syndrome.

The invention is illustrated in the accompanying drawings in which :-

Fig. 1 shows the principle of the invention schematically

Fig. 2 shows one embodiment using a plurality of tubes

Fig. 3 shows an embodiment using a carbon monolith Fig. 4 shows an embodiment with rectangular channels

Fig. 5 shows the dimensions of the structure of fig. 4

Fig. 6 shows the macroporous carbon wells for small scale in-vitro testing

Fig 7 shows the pore structure of the mesoporous activated carbons, mesoporous resin bead and microporous granular resin carbons Figs. 8 to 13 show the results of the examples.

Referring to fig. 1 the device comprises a tube formed of a porous monolithic carbon formed by the method of example 1 of Patent application number WO 2004/087612 with a feed channel (2) passing through walls (4). The whole blood (1) enters the feed channel (2) at one end of the device. The filtrate (7), comprising essentially plasma, passes through the channel wall (4) into a parallel channel and is cleaned of endotoxins and cytokines by adsoφtion within the wall. The retentate (5), comprising a blood cell concentrate leaves the channel exit. At the exit of the device the filtrate (3) and retentate (5) streams are recombined (6) and returned to the venous blood and systemic circulation.

Referring to fig. 2 this shows a multi-tubular device which comprises monolithic carbon tubes (11) (made as in fig. 1) with walls (20) and inlet channels (19) mounted in a reactor (14) with feed (21) and retentate (13) collection chambers. The filtrate passes through the walls of the tubes and is collected in the body of the reactor leaving as stream (8) via the retentate pressure controller (9). The control of the pressure drop can be achieved by the external pressure control devices (18) and (9) in the retentate (16) and filtrate (8) streams respectively. The feed and permeate linear velocities are controlled by the free volume on the feed and permeate sides of the tube, the feed velocity and the permeation rate. As in the simple device in figure 1 the retentate and filtrate streams are recombined in stream (10) and returned to the venous system.

In operation it is essential that the blocking of the filter surface by blood cells and other high molecular weight species is minimised. This can be achieved by careful control of the linear velocity of the feed stream and the differential pressure across the channel wall.

Preferably the channels are formed in a porous synthetic carbon monolith as shown in figures 3 and 4 where the filtration surface area as a function of the device volume can be maximised. This can either comprise multiple round channels as shown in figure 3 or, preferably, square or rectangular channels as shown in figure 4.

Referring to fig. 4 this structure is shown in side view in fig. 4a and end view in fig. 4b. In this structure the smaller feed channels (22) pass directly through the monolith. These channels are made sufficiently small that the high linear velocity both reduces deposit formation on the walls and maintains sufficient pressure drop along the channel to force the filtrate through the monolith walls (23) into the filtrate collection channels (26). The filtrate channels are blocked at the blood inlet end of the device (25) to ensure that all the incoming blood passes through the feed channels. The filtrate channels are sized so that there is negligible pressure drop in the filtrate flow. It can be seen that the relative retentate: feed flow velocity is fixed by the dimensions of the feed (F) and retentate (R) channels such that the ratio of the cham el cross sections for a unit cell is given by R²/(2RF) = R/2F.

It is not possible to differentiate the feed and permeate channels as in fig. 2. Control of the flow pathways and the pressure drop (dP) is only possible by closing of one end of the permeate channels (22) to restrict the whole blood feed just to the feed channels and then controlling the dP by adjusting the size of the feed and permeate channels.

The effectiveness of the device is thought to derive from the unique combination of controlled channel, macro, micro and meso pore structures. These are shown in fig. 5 for the square channel monolith in figure 4.

For a symmetrical monolith of fig. 4, fig. 5 shows the macroporous channel structure which is defined by the channel dimension, W, and the wall thickness, t, or for an asymmetric monolith as shown in figure 3 by R and F as well as t. As described above these fix the ratio of open to closed area and therefore the flow velocities along the tube. This pore size then fixes the permeability of the wall and the maximum liquid flow through the wall form a given pressure differential or flow linear velocity as defined by the Blake-Kozeny equation:-

Moulded carbon forms were made using the apparatus of fig. 6 in which there is a Perspex (RTM) holder (31) in which is placed a carbon well (32) made from the' sintered phenolic resin used to produce the monolith carbons. The macro, micro and meso porous structure of the carbon well could be controlled to be identical to that of the channel monoliths described in figures 1 to 5.

To test the carbons a small volume of blood (34) is placed in the top well and the permeate (33) that passes through the base of the carbon cell is then collected in the Perspex holder. Both the retentate and permeate are then analysed.

Figure 7 shows the pore structures of the carbons used in the production of the binary materials. These comprise the commercially available mesoporous carbon powders supplied by CECA (CXN) and PICA (EPII and SC10). The EPII was also available in two particle sizes, 8-15microns and 15-30 microns, although this had no impact on the pore structure. The resin derived carbons comprise the carbon from the mesoporous resin beads, prepared according to WO 025/12380 A2, and the sinterable, non porous resin prepared according to EP0254551 that forms the underlying structure. It can be seen that the pore size increases in the order:- Microporous resin < EPII < SC10 < CXN < mesoporous resin beads

The results are illustrated in the Examples.

Example 1 Pore Structure of Binary Carbon Wells

The following example demonstrates the structures that can be produced through the use of the binary combinations of the non porous resin with either the commercial carbons or the mesoporous resin beads. The amount of commercial activated carbon powder that can be incoφorated into the wells is limited by the shrinkage of the resin matrix. By contrast the wells can be produced using up to 100%) of the mesoporous beads. The formulations used are summarised in the table 1 :-

Table 1 Component weight loading

For the mesoporous material the formulation is parts microporous resin/parts mesoporous resin. The pore structures of these binary wells are shown in figure 7. Whilst the binary resin structures can be produced with ratios varying between 100% of each component the amount of carbon that can be incoφorated is limited to approximately 30%) volume.

Contrary Substance Adsorption

The following examples demonstrate the removal of the four contrary substances IL- 8, IL-6, TNF, IL-lβ and LPS from plasma. LPS molecular mass varies. Subunits are 10-20kDa, but aggregate into complexes with plasma proteins. Complex molecular mass ranges from 100-lOOOkDa. TNF (17kDa) forms dimers and trimers and has a molecular size of approximately 51 kDa. In plasma IL-6 (21-28kDa) can complex with the carrier protein Alpha2M to give 42-45kDa, IL-lβ has a size of 17kDa, and IL-8 is the smallest at around 8kDa. Example 2. Use of Carbon powders and Carbon wells to remove FITC-labelled LPS from plasma

Fluorescein-isothiocyanate labelled-lipopolysaccharide (FITC-LPS, Sigma) was dissolved in phosphate-buffered saline (PBS) with 10% human plasma (NBS) to give a concentration of lOμg/ml. Carbon powders (n=5, Table 2) were weighed (0.2 g) and equilibrated in PBS overnight prior to addition of 1 ml of spiked plasma (lOμg/ml). Controls consisted of spiked plasma, or plasma only with no adsorbent present and Adsorba 300C (Norit RBX cellulose coated carbon, Gambro). Materials were incubated at 37 °C while shaking (90 rp ). After 2 hours, samples were centrifuged (125 g) and the supernatant collected, the concentration of the FITC-LPS remaining was determined by fluorescence spectrophotometery (excitation 495nm, emission 520nm). Spiked plasma (1.8 ml) was added to the surface of the carbon wells, tissue culture plastic, and Adsorba 300C (Norit RBX cellulose coated carbon, Gambro) controls and incubated at 37°C. At timed intervals (1.5, 5, 24 hrs) plasma filtrate was collected and the concentration of the FITC-LPS remaining was determined by fluorescence spectrophotometery (excitation 495nm, emission 520nm).

The FITC-LPS adsoφtion profile for the carbonized powders 1-14 is displayed in Figure 8. Carbon powder 1 & 5 displayed efficient removal of FITC-LPS from plasma, along with the carbon powder activation range 11-14. The larger size of LPS in comparison to inflammatory cytokines, restricts their uptake by the cellulose coated Adsorba 300C (material 8). The ability of the binarycarbon wells to remove FITC-LPS from human plasma is presented in Figure 9. The pattern of LPS removal was similar to that found for the equivalent powder adsoφtion of FITC-LPS (Fig 8). Saturation effects may have increased measurable LPS present in solution at the two later time points. Carbon wells C, D and E (mesoporous) displayed superior adsoφtion capacity when compared to microporous well A, and to the Adsorba 300C control, particularly at the first time point. Table 2

The tubes were then carbonised in flowing nitrogen at 800C. Example 3. Use of Carbon powders to remove TNF, IL-6, IL-8 and IL-lβ from plasma

Fresh frozen human plasma (NBS) was defrosted and spiked with the recombinant human cytokines; TNF, IL-6, IL-8 and IL-lβ (BD Biosciences) at a concentration of 1000, 2000, 500 and 1000 pg/ml respectively. The adsorbents studied (Table 2) were weighed (0.2 g, n=5) and equilibrated in PBS overnight prior to addition of 500 μl of spiked plasma. Controls consisted of spiked plasma, or plasma only with no adsorbent present, and Adsorba 300C (Norit RBX cellulose coated carbon, Gambro). Materials were incubated at 37 °C while shaking (90 rpm). After 2 hours, samples were centrifuged (125 g) and the supernatant collected and stored at -20 °C prior to ELIS A analysis for the presence of cytokines. Samples were diluted 1 :4 (TNF, IL-8), 1:10 (IL-6) and 1 :2 (IL- 1 β) in assay diluent prior to analysis.

The four cytokine adsoφtion profiles for the carbon powders 1-14 are displayed in Figures lOa-d. The most mesoporous Carbon powder 5 displayed efficient removal of all four cytokines, whilst the mesoporous carbon powder 3 also exhibited significant removal of IL-6, IL-8 and IL-lβ. The more microporous carbon powders 1 & 2 removed significantly less of the two larger cytokines TNF and IL-6. It can clearly be seen from Figure 10a, that as the degree of activation increases in the adsorbent range 9-14, removal of TNF the largest of the cytokines examined increased accordingly from 70-97%). Activation increases the number of micropores within the structure of the carbon at the larger end of the micropore range (<0.2nm) and clearly results in greater uptake. This micropore size range may be critical for the cytokine size range (21-51 kDa) of interest in this study, and goes some way to reinforcing the relationship between pore size and adsoφtion capacity when the adsorbant is at a critical molecular weight. It is important to note, however, that activation in air or carbon dioxide alters the functional groups present at the surface of the carbon and thereby may influence the degree of adsoφtion. It can be seen from Figures lOb-d, that the smaller IL-6 (21.5-28 kDa), IL-8 (8 kDa) and IL-lβ (17kDa) were efficiently adsorbed by the adsorbent range 5-14, with 100% removal in most cases. It is noticeable that all the carbon adsorbents tested were superior at removing cytokines from plasma than Adsorba^® 300C (material 8). This is consistent with the fact that these carbons are coated with cellulose to improve their biocompatibility, which reduces the diffusion into the carbon, of molecules of high molecular weight such as the four cytokines tested here.

Example 4 Use of Carbon wells to remove TNF, IL-6, IL-8 and IL-lβ from plasma Fresh frozen human plasma (NBS) was defrosted and spiked with the recombinant human cytokines; TNF, IL-6 & IL-8 (BD Biosciences) at a concentration of 1000, 2000 and 500 pg/ml respectively. Spiked plasma (1.8 ml) was added to the surface of the binary carbon wells and tissue culture plastic controls and incubated at 37 °C (21 hours). Plasma filtrate that had passed through the carbon matrix was collected and tested by ELISA for the presence of cytokines. Samples were diluted 1:8 (TNF, IL-8) and 1 :20 (IL-6) in assay diluent prior to testing.

The ability of the carbon wells to remove the three cytokines (TNF, IL-6 & IL-8) from human plasma is presented in Figure 11. All three cytokines were removed from human plasma to some extent as a result of filtration/adsoφtion by carbon wells A-D. Carbon well A comprising a largely microporous structure, was effective at removing the smallest cytokine IL-8 from spiked plasma but failed to adsorb TNF and IL-6 as efficiently, perhaps as a result of size exclusion, and the concentration of IL-8 was lower than for the other two cytokines. The lower concentration and smaller size of IL-8 also explains why this cytokine was removed efficiently by carbon wells B-D, which contained an increasing number of mesopores. This increasing mesoporosity also explains the improved ability of carbon wells B-D to remove the larger molecular weight TNF and IL-6. The results show a significant difference between removal of TNF and IL-6 by carbon wells C & D and the control TC plastic (Student T test, p<0.05 for TNF, and pθ.001 for IL6, n=3). These results correlate with the previous carbon powder adsoφtion assay. Carbon well B is composed of 5%> adsorbent 5 and demonstrated good adsoφtion ability both in the well form and carbonised powder form, with significant removal of all three cytokines compared to the control. Carbon wells C and D are composed of adsorbent 9 (80 and 100%) respectively) which also demonstrated good adsoφtion capacity in the carbonised powder form. Their superior adsoφtion capacity compared to well B may be explained by their higher percentage mesoporous component.

Example 5 Use of Sintered Carbon Structures as Filtration Devices

Figures 12, 12A and 12B demonstrate the control of the macropore structure of the sintered resin derived carbons. Figure 12 shows the relationship between the particle size of the resin used to produce the carbon and the macro pore size determined by mercury porosimetry of the finished carbon. It can be seen that there is a direct linear relationship with the mean pore size approximately 20% of the resin particle size. This is the relationship that would be expected from approximately close packed particles as shown in the diagram in figure 12 A. Figure 12B then shows the liquid phase filtration efficiency of these materials using a British Standard water based particulate challenge. It can be seen that the particulate removal efficiency is related to the particle size of the material used to produce the carbon. The cut of efficiency is actually considerably better than might be expected from the pore size of the material reflected in the data in figure 12 A.

Example 6 Production of Asymmetric Carbon Membrane Systems using a Mesoporous Resin Precursor

Resin tubes were produced from cured 30 micron resin powders produced according to EP 0 254 551 , and the extrusion method described PCT/GB02/01142. These were then cured at 150C overnight. After drying the tubes were placed in a spray device that rotated the tubes at a fixed rate whilst moving the spray head along the tube. The coating solution comprised a solution of partially cured novolak resin in ethylene glycol, prepared according to WO 02/12380 A2, and further diluted with isopropyl alcohol to give a viscosity suitable for spraying. The tubes were then sprayed in two stages with an intermediate drying and curing step at 120C between the coating steps. Depending on the thickness of the membrane layer required a varying number of coats can be used in each spraying step with just a period of ambient temperature air drying between spray coats to allow the majority of the IP A to evaporate. In these tubes either 3 or 2 first coats followed by 2 second coats were used. The tubes produced are summarised below:-

Example 7 Blood Filtration using Extruded Carbon Tubes

In the filter system of this invention the operation is based on the initial separation of the blood into a plasma permeate stream and a cell enriched retentate stream. The contrary substances are then adsorbed from the plasma stream as it permeates through the membrane structure (see figure 1 and 2). It is not necessary in this system that total removal of the blood cells occurs, as both streams are recombined before being passed back into the body. The requirement is to remove sufficient cells to allow effective adsoφtion of the contrary molecules from the plasma during permeation. However it is also then essential that degradation of these blood cells does not occur during permeation. The blood filtration tests were carried out using a tubular membrane system as shown in figure 1. This was sealed into a Perspex holder with feed, retentate and permeate connections. The feed flow and the pressure in the feed and permeate side of the membrane could be controlled separately whilst the retentate and permeate streams could be collected and analysed.

The red blood cell concentration was determined in the feed and permeate streams respectively. A reduction in the permeate cell concentration does not confirm conclusively that the cells have been removed as they could be destroyed during passage through the filter. However if this happens haemoglobin is released that can then be detected either directly by assay or by centrifugation of the permeate to remove the blood cells. If there was haemoglobin present due to cell destruction the filtrate after centrifugation would be pink. The tubes evaluated are summarised below and comprise plain microporous resin based tubes (E789 to E840) based on 40 micron resin, binary tubes produced from either 9 micron (E786) or 40micron resin (E788/2) with 20% added mesoporous carbon and asymmetric spray coated tubes based on 30 micron resin.

The haemocrit concentration in the blood, retentate and filtrate is shown in figure 13 a. It can be seen that the levels of blood cells passing through the membrane is quite variable but for several of the tubes there are essentially no blood cells in the permeate. However comparison with figure 13 b, which shows the levels of haemoglobin in the feed and permeate streams shows that in most of these cases this was associated with extensive cell degradation. The only sample that showed a good combination of trans membrane flux, blood cell removal and no cell degradation was 120205 which was an asymmetric structure with a spray coated membrane layer in contact with the blood, whilst the 10 micron resin carbon powder binary tubes showed good fluxes and some cell degradation. In this case the enhanced permeability relative to the 40 micron powder based tubes can be attributed to defects in the structure introduced by the active carbon powder. The very low flows found with the 40 micron resin systems can probably be attributed to the use of unclassified powders where the smaller particles partially infill the voids between the larger particles resulting in reduced porosity and increased tortuosity. The extensive cell degradation may then be due to the higher turbulence encountered in this more tortuous pathway.

References

¹ Stegmayr BG. Apheresis as therapy for patients with severe sepsis and multiorgan dysfunction syndrome. Therapeutic Apheresis 2001 ;5(2):123-127.

² Paterson RL, Webster NR. Sepsis and the systemic inflammatory response syndrome. J.R.Coll.Surg.Edinb 2000;45:178-182.

³ Steczko J, Ash SR, Blake DE, Carr DJ, Bosley RH. Cytokines and endotoxin removal by sorbents and its application in push-pull sorbent-based pheresis: the biologic-DTPF system. Artificial Organs 1999;23(4):310-318.

⁴ Volk HD, Reinke P, Krausch D, Zuckermann H, Asadullah K, Muller JM, Docke WD, Kox WJ, (1996) Monocyte deactivation - rationale for a new therapeutic strategy in sepsis. Intensive Care Medicine 22 (4): S474-S481 Dinarello CA. (1995) Interleukin-1 and interleukin-1 receptor antagonist. Nutrition. 11 (5): 492-494

⁶ Guirao X, Lowry SF. (1996) Biologic control of injury and inflammation-much more than too little or too late. World J. Surg. 20 (4): 437-446

⁷ Quezado ZMN, Banks SM, Natanson C. (1995) New strategies for combating sepsis - the magic bullets missed the mark....but the search continues. Trends biotechn. 13 (2): 56-63

⁸ Henderson B, Poole S, Wilson M. (1996) Microbial/host interactions in health and disease - who controls the cytokine network. Immunopharm. 35 (1): 1-21

⁹ Natanson C, Hoffman WD, Suffredini AF, Eichacker PQ, Danner RL. (1994) Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis. Ann Intern Med. 120 (9): 771-783 Verhoef J, Hustinx WMN, Frasa H, Hoepelman AIM. (1996) Issues in the adjunct therapy of severe sepsis. J. Antimicr. Chemother. 38(2): 167-182)

¹¹ Cohen J, Heumann D, Glauser MP. (1995) Do monoclonal-antibodies and anticytokines still have a future in infectious diseases? Am. J. Med. 99 (1): S6A, S45-S52

¹² Bellomo R, Tipping P, Boyce N, (1993) Continuous veno-venous hemofiltration with dialysis removes cytokines from the circulation of septic patients. Critical Care Medicine 21: 522-526

¹³ Hoffman JN, Hartl WH, Deppisch R, Faist E, Jochum M & Inthorn D, (1995) Hemofiltration in human sepsis: Evidence for elimination of immuno-modulatory substances. Kidney International. 48: 1563-1570

¹⁴ G.J. Phillips, S.L. Moore, J.G. Davies, C.J. Olliff, M.K. Street (1997) Cytokine clearance from septic patients during haemofiltration with hollow fibre polyacrylonitrile membranes. J. Pharm. Pharmacol. 49 (4): 51

¹⁵ Yamamoto C, Kim ST, (1996) Endotoxin rejection by ultrafiltration through high flux hollow-fibre membranes. J. Biomedical Materials Research. 32 (3): 467-471

¹⁶ Haupt W, Fritzsche H, Hohenberger W, Zirngibl H. (1996) Selective cytokine release induced by serum and separated plasma from septic patients. European J. of Surgery. 162 (10): 769-776

¹⁷ Scholmerich J. (1996) Interleukins in acute pancreatitis. Scandinavian J. of Gastroenterology. 31 (S219): 37-42

¹⁸ Koepp A, Lampert R, (1996) Plasmapheresis in a 62-year-old male patient with severe sepsis.

Infusiontherapie und transfusionmedizin. 23 (2): 92-96

19 Stegmayer BG, (1996) Plasma exchange in patients with septic shock including acute renal failure.

Blood Purification. 14 (1): 102-108

²⁰ Weber C, Falkenhagen D, (1996) Extracorporeal removal of proinflammatory cytokines by specific adsoφtion onto microspheres. ASAIO J. 42 (5): M908-911

²¹ Skerrett D, Mor E, Curtiss S, Mohandras K. (1996) Plasmapheresis in primary dysfunction of hepatic transplants. J. Clinical Apheresis 11 (1): 10-13

²² Weber C, Falkenhagen D, (1996) Extracorporeal removal of proinflammatory cytokines by specific adsorption onto microspheres. ASAIO J. 42 (5): M908-911

²³ Hughes RD, Williams R, (1993) use of sorbent columns and haemofiltration in fulminant hepatic failure. Blood Purification. 11 (3): 163-169 ²⁴ Ash SR. (1991) Treatment of acute hepatic failure with encephalopathy: a review. 14: 191-195

²⁵ Chang TMS (1972) Haemoperfusion over microencapsulated adsorbent in a patient with hepatic coma. Lancet 2: 1371-1372

²⁶ Malchesky PS. (1994) Non-biological liver support: historic overview. Artif. Organs. 18: 342-347

²⁷ Soeda K, Odaka M, Tabata Y, Kobayashi S, Itoh S, Isono K. (1991) Efficacy and limitations of plasma exchange in patients with hepatic failure; comparing with haemoadsoφtion and with impaired regeneration syndrome. Bio at. Art. Cells Immob. Biotech. 19: 203-211

²⁸ Nagaki M, Hughs RD, Keane HM, Lau JYN, Williams R. (1992) In vitro plasma perfusion through adsorbants and plasma ultrafiltration to remove endotoxin and cytokines. Circ. Shock. 38 (3): 182-188

²⁹ Nagaki M, Hughs RD, Lau JYN, Williams R (1991) Removal of endotoxin and cytokines by adsorbents and the effect of plasma-protein binding. Int. J. Artif. Organs. 14 (1): 43-50

³⁰ Courtney JM, Lamba NMK, Gaylor JDS, Ryan CJ, Lowe GDO. (1995) Blood clotting biomaterials: bioengineering viewpoints. Artif. Organs. 19 (8): 852-856

³¹ S Sandeman, CA Howell, GJ Phillips, JG Davies, AW Lloyd, SV Mikhalovsky, S.R Tennison, O Kozynchemko, J.D.S Gaylor, J.M Courtney. 2003. A Novel Carbon Device for the Treatment of Sepsis. Therapeutic Apheresis and Dialysis. 7(5):44 (A 13).

³² Nikolaev VG. (1984) Hemcarboperfusion in the USSR. Int. J. Artif. Organs. 7 (5): 289-296

³³ G.J. Phillips, S.L. Moore, J.G. Davies, S. Mikhalovsky, CJ. Olliff, A.W. Lloyd, M.K. Street, C Kingswood (1997) Cytokine adsoφtion by uncoated carbon absorbants. Int. J. Artificial Organs. 20 (9): 514

Claims

1. A method for the treatment of blood which comprises (1) passing the blood through a monolithic porous carbon structure, (2) allowing the plasma components to pass through the walls of the monolith to form two streams, a plasma permeate stream passing through the walls of the monolith and a retentate stream containing the majority of the blood cells, (3) adsorbing the contrary substances from the plasma permeate stream in the walls of the monolith and (4) combining the filtrate stream and the retentate stream.

2. A method as claimed in claim 1 in which the monolithic porous carbon structure has (i) continuous voids or pores through which blood plasma can pass with a mean pore size of greater than 500nm and (ii) pores within the carbon matrix suitable for the adsorption of middle and high molecular weight molecules with a mean pore size between 2 and 500nm.

3. A method as claimed in claim 1 or 2 in which the monolithic porous carbon structure is tubular with a central channel of between 1 and 10mm and a wall thickness of between 0.2 and 2mm.

4. A method as claimed in claim 1 or 2 in which the monolithic porous carbon structure comprises rectangular channel monoliths with a channel size between 100 and 2000 microns and the wall thickness is between 100 and 2000 microns and with an open area of between 30 and 60%.

5. A method as claimed in claim 4 in which the cell structure is asymmetric.

6. A method as claimed in any one of claims 1 to 5 in which the monolithic porous carbon is made by partially curing a phenolic resin to a solid, comminuting the partially cured resin, extruding the comminuted resin, sintering the extruded resin so as to produce a form-stable sintered product and carbonising and activating the form- stable sintered product.

7. A method as claimed in any one of the preceding claims in which the monolith has a surface area of at least 600m2/g.

8. A method of treating blood from a patient as claimed in any one of the preceding claims to control sepsis comprising (i) filtering the blood stream to partially separate the blood cells from the plasma (ii) removing the inflammatory mediators and (iii) re- combining the cleaned plasma permeate and the partially separated retentate stream containing the blood cells for re-injection of the combined stream back into the patient.

9. Apparatus for treating blood which comprises (i) a porous monolithic carbon having at least one channel formed in it (ii) an inlet for an inlet blood stream blood to enter the channel whereby blood passes down the channel and a filtrate passes through the walls of the channel to form two streams a filtrate stream and retentate stream (iii) a means to combine the retentate and filtrate streams and (iv) an outlet for the combined stream.

10. Apparatus as claimed in claim 9 in which there are two channels, an inlet channel which is the retentate channel and a filtrate channel into which the filtrate enters after passing through the walls of the inlet channel and in which the filtrate channel is connected to the outlet.

11. Apparatus as claimed in claim 10 which comprises a monolithic porous carbon structure having at least one inlet retentate channel and at least one outlet filtrate channel formed in it.

12. Apparatus as claimed in claim 11 having a plurality of inlet and outlet channels.

13. Apparatus as claimed in any one of claims 9 to 12 in which the monolithic carbon has a macro porous structure.

14. Apparatus as claimed in any one of claims 9 to 13 in which the monolithic carbon has (i) continuous voids or pores in the walls through which liquid can pass with a mean pore size of greater than 500nm and (ii) pores within the carbon matrix suitable for the adsorption of middle and high molecular weight molecules with a mean pore size between 5 and 500nm.

15. A method as claimed in any one of claims 9 to 14 in which the monolithic porous carbon structure is tubular with a central channel of between 1 and 10mm and a wall thickness of between 0.2 and 2mm.

16. An apparatus as claimed in any one of claims 9 to 14 in which the monolithic porous carbon structure comprises rectangular channel monoliths with a channel size between 100 and 2000 microns and a wall thickness of between 100 and 2000 microns and with an open area of between 30 and 60%.

17. An apparatus as claimed in any one of claims 9 to 16 in which the cell structure is asymmetric.

18. An apparatus as claimed in any one of claims 9 to 16 in which the filtrate channels are closed off at the monolith inlet.

19. An apparatus as claimed in any one of claims 9 to 18 in which the monolithic porous carbon is made by partially curing a phenolic resin to a solid, comminuting the partially cured resin, extruding the comminuted resin, sintering the extruded resin so as to produce a form-stable sintered product and carbonising and activating the form- stable sintered product.

20. An apparatus as claimed in any one of claims 9 to 19 in which the monolithic porous carbon is made by mixing the partially cured, milled phenolic resin powder with a mesoporous carbon powder, extruding the mixed resin and carbon powder, sintering the extruded mixture so as to produce a form-stable sintered product and carbonising and activating the form-stable sintered product.

21. An apparatus as claimed in any one of claims 9 to 19 in which the monolithic porous carbon is made by mixing the partially cured, milled phenolic resin powder with a mesoporous resin powder, extruding the mixed resin and mesoporous resin powder, sintering the extruded mixture so as to produce a form-stable sintered product and carbonising and activating the form-stable sintered product.

22 An asymmetric carbon membrane structure in which the surface of the channel in the extruded resin monolith in contact with the feed stream is coated with a resin layer by either spray coating, dip coating or flowing a solution through the channel to be coated. The asymmetric resin structure is then cured, carbonised and activated. The resin can be a phenolic resin or any other carbonisable resin system that can be dissolved in a coating solution.

23. Apparatus for treating blood as hereinbefore described and illustrated in the accompanying drawings.