METHOD FOR THE PURIFICATION OF STARCH HYDROLYSATES FOR THE PRODUCTION OF GLUCOSE POLYMERS FOR PERITONEAL DIALYSIS

Description

[0001] The present invention relates to a process for decontaminating starch hydrolysates from which glucose polymers will be prepared for the manufacture of peritoneal dialysis solutions.

[0002] For the purposes of this invention, "decontamination process" means a process that removes contaminating microorganisms (live and / or spore-forming microorganisms such as yeasts, molds, and bacteria) and substances that can cause peritonitis (aseptic or non-aseptic), a major complication of peritoneal dialysis, from starch hydrolysates. These substances may include: Lipopolysaccharides (LPS) are toxic macromolecular complexes constitutively present in the outer membrane of all Gram-negative bacteria. Structurally, LPS consist of a lipid A and a polysaccharide portion extending beyond the outer membrane. Lipid A possesses toxic properties and corresponds to the endotoxin of Gram-negative bacteria, which is released in large quantities only after bacterial lysis. β-glucans are D-glucose polymers linked by β-glucosidic bonds. β-glucans are a diverse group of molecules with varying molecular weights, solubility, viscosity, and three-dimensional configurations. β-glucans are notably constituents of the cell walls of plant cells, yeasts, and certain molds and bacteria. Peptidoglycans (PGs) are polysaccharide components of the cell walls of Gram-positive bacteria.Also called mureins, or mucocomplexes, or mucopeptides, petidoglycans are formed of a polysaccharide part and a peptide part. The polysaccharide is a glycosaminopeptide polymer where N-acetylglucosamine and N-acetylmuramic acid are linked by β 1-4 glycosidic bonds.

[0003] The invention relates more particularly to the modification of the conduct of the conventional processes of production of starch hydrolysates, in the sense that serial filtration steps or special activated carbon treatment steps are added to said conventional processes, coupled or not with enzymatic treatment steps.

[0004] These additional steps are therefore intended to guarantee the safety of the starch hydrolysates thus prepared, i.e. to guarantee a content of contaminating substances well below the quantification threshold of classical methods of measuring said contaminating substances.

[0005] Peritoneal dialysis is a type of dialysis that aims to remove waste products such as urea, creatinine, excess potassium, or excess water that the kidneys are unable or no longer able to filter from the blood plasma. This medical treatment is indicated in cases of end-stage chronic kidney failure.

[0006] This is an intracorporeal purification method that uses the peritoneum as a dialysis membrane. Toxic waste products from the blood pass through the semi-permeable peritoneal membrane into a solution called dialysate. The dialysate is introduced into the peritoneal cavity via a permanent catheter. There are two types of peritoneal dialysis: CAPD (continuous ambulatory peritoneal dialysis), a treatment based on the passage of 4 bags of dialysate per day according to medical prescription. APD (automated peritoneal dialysis), a continuous nocturnal treatment which corresponds to approximately 15 liters of dialysate per 8 hours according to medical prescription.

[0007] The most commonly used dialysates are composed of a buffer solution (lactate or bicarbonate) at acidic pH (5.2 - 5.5) or physiological pH (7.4) to which electrolytes (sodium, calcium, magnesium, chloride) and an osmotic agent (glucose or a glucose polymer, such as "icodextrin" present in the EXTRANEAL ®< ambulatory peritoneal dialysis solution marketed by BAXTER) are added.

[0008] The electrolytes and the osmotic agent each play a role in the exchange mechanism, according to their respective physicochemical properties: The waste products of metabolism (such as urea or creatinine) or other electrolytes in excess that the kidney no longer eliminates or does not sufficiently eliminate via the urinary system and urine, will be extracted from the blood plasma by diffusion of the elements into the dialysate, the concentration levels of which are lower; the excess water, which the kidney normally eliminates for the regulation of plasma volume, will be attracted by osmolarity depending on the concentration of the dialysate in glucose or glucose polymer: the more concentrated the solution, the more water present in the body will be taken up by the dialysate.

[0009] Glucose polymer, such as the icodextrin mentioned above, is preferred to glucose as an osmotic agent because, although glucose has the advantage of being relatively safe and inexpensive, it has a number of disadvantages.

[0010] Due to its small size, glucose rapidly crossing the peritoneum leads to the loss of osmotic gradient within 2 to 4 hours of infusion.

[0011] The ultrafiltration characteristics of peritoneal dialysis solutions are therefore considered better by replacing glucose with high molecular weight substances, such as glucose polymers, as will be demonstrated below.

[0012] Standard glucose polymers are produced by acid or enzymatic hydrolysis of cereal or tuber starch.

[0013] The completely random acid hydrolysis of starch, or its somewhat more ordered enzymatic hydrolysis, yields mixtures of glucose (monomers) and glucose chains that include very short molecules (oligomers) with a low degree of polymerization (DP), as well as very long molecules (polymers) with a high DP. Glucose polymers also exhibit extremely varied molecular weights.

[0014] In the more specific field of using glucose polymers for continuous and ambulatory peritoneal dialysis, it quickly became apparent that these starch hydrolysates (a mixture of glucose, oligomers and glucose polymers) could not be used as such.

[0015] European patent application EP 207.676 teaches that glucose polymers are preferred, forming clear and colorless solutions at 10% in water, having a weight average molecular weight (Mw) of 5,000 to 100,000 daltons and a number average molecular weight (Mn) of less than 8,000 daltons.

[0016] Such glucose polymers also preferably comprise at least 80% glucose polymers with a molecular weight between 5,000 and 50,000 daltons, little or no glucose or glucose polymers with a DP less than or equal to 3 (molecular weight 504), and little or no glucose polymers with a molecular weight greater than 100,000 (DP close to 600).

[0017] In other words, preferred glucose polymers are glucose polymers with a low polymolecularity index (value obtained by calculating the Mw / Mn ratio).

[0018] It is indeed easy to understand for this application that low molecular weight monomers or polymers quickly cross the peritoneal wall and are therefore of no lasting interest for the creation of an osmotic pressure gradient, and that very high molecular weight polymers, devoid of osmotic power, are to be avoided and even prohibited since they are potentially dangerous if they happen to precipitate following their retrogradation.

[0019] The processes proposed in this patent application EP 207,676 for obtaining these low polymolecularity index glucose polymers from starch hydrolysates consist of: either to perform a fractional precipitation of a maltodextrin using a water-miscible solvent, or to perform a molecular filtration of this same maltodextrin through different membranes having an adequate cut-off or exclusion threshold.

[0020] In both cases, these processes aim to eliminate both very high molecular weight polymers and low molecular weight monomers or oligomers.

[0021] However, these processes are not satisfactory both in terms of their implementation and in terms of the yields and quality of the products they produce.

[0022] Concerned with developing a process for manufacturing a glucose polymer that is completely soluble in water and has a low polymolecularity index preferably less than 2.5, preferably having an Mn less than 8,000 daltons and possessing an Mw between 12,000 and 20,000 daltons, a process that is free from the disadvantages of the prior art, the Applicant company has sought to solve this problem in its patent EP 667,356, starting from a hydrolyzed starch, rather than from a maltodextrin.

[0023] This process consists of: acid hydrolyze a waxy starch milk to a DE between 8 and 15; optionally complete this acid hydrolysis with enzymatic hydrolysis using bacterial alpha-amylase to a DE between 11 and 18; chromatograph this double acid-enzyme hydrolysate on macroporous strong cationic resins in alkali or alkaline-earth form; collect the glucose polymer excluded during this chromatography step.

[0024] In this patent, to obtain a glucose polymer with a polymolecularity index of less than 2.5, the glucose polymer excluded during this chromatography step is collected in a weight yield of around 60% of the starch hydrolysate used upstream of the chromatography step.

[0025] This glucose polymer then preferably contains less than 3% of glucose and glucose polymers with a DP less than or equal to 3 and less than 0.5% of glucose polymers with a DP greater than 600.

[0026] Application WO2007 / 099212 also describes a process for preparing glucose polymers suitable for peritoneal dialysis. The process concludes with fractionation steps in which low molecular weight fractions are removed, particularly those with a molecular weight below 9000, while the remaining, higher molecular weight fractions are recovered. More specifically, the process comprises an activated carbon treatment step (NORIT SX+), fractionation with a cutoff threshold of 9000 daltons, retentate recovery, and then a demineralization and activated carbon treatment step (NORIT SX+).

[0027] It is now finally accepted by experts in the field of peritoneal dialysis that these glucose polymers, used for their osmotic power, give complete satisfaction.

[0028] However, there are regrettable risks of microbial contamination of these preparations intended for peritoneal dialysis.

[0029] For example, a catheter implanted in the peritoneal cavity is a potential entry point for germs. The numerous manipulations of the catheter during infusion and drainage increase the risk of local or systemic infection.

[0030] In addition, an additional risk factor for contamination may be directly related to impurities that can contaminate glucose polymers used as osmotic agents.

[0031] It is indeed known that glucose polymer production circuits can be contaminated by microorganisms, or by pro-inflammatory substances contained in said microorganisms.

[0032] For example, in starch manufacturing, the contamination of corn or wheat starches by microorganisms such as yeasts, molds, and bacteria, and more specifically by acid-thermophilic bacteria, is described. Alicyclobacillus acidocaldarius (extremophile bacteria that thrive in warm, acidic areas of the circuit).

[0033] The major risk for the patient who receives these contaminated products is then peritonitis.

[0034] Clinical suspicion of peritonitis is diagnosed when a disorder in the dialysate develops, associated with variable clinical manifestations including abdominal pain, nausea, vomiting, diarrhea, and fever.

[0035] These episodes of peritonitis are caused by intraperitoneal bacterial infections, and the diagnosis is usually easily established by positive dialysate cultures.

[0036] "Sterile peritonitis", also described as aseptic, chemical, or culture-negative peritonitis, is typically caused by a chemical irritant or a foreign body.

[0037] Since the introduction of icodextrin for the preparation of peritoneal dialysis solutions, isolated cases of aseptic peritonitis have been reported, which may be linked to various causes, including induction by potentially present pro-inflammatory substances.

[0038] However, these pro-inflammatory substances are not detected by the tests usually carried out to determine the safety of such preparations.

[0039] The tests described today in Pharmacopoeias for the detection of pyrogenic substances are indeed the following: The so-called LAL test, for the detection of bacterial endotoxins (LPS), major components of Gram negative bacteria, The rabbit pyrogen test, for the detection of bacterial endotoxins (LPS) as well as β glucans, components of the walls of fungal flora (yeasts and molds).

[0040] Although generally reliable, both of these tests have their limitations.

[0041] The rabbit pyrogen test is based on the indirect detection of pyrogenic substances by measuring a temperature rise in the rabbit to which the product containing these substances has been injected (febrile response).

[0042] This test may produce false negatives if the undesirable substance has too low a biological activity or concentration to induce a systemic pyrogenic response.

[0043] However, this substance may possess sufficient biological activity or concentration to produce a local inflammatory reaction.

[0044] Other biological impurities (DNA, etc.) are not detected. The same is true for peptidoglycans, the major components of the cell membranes of Gram-positive bacteria.

[0045] The manifestation of aseptic peritonitis observed with peritoneal dialysis solutions containing icodextrin therefore demonstrates, in some cases, how certain substances can escape the tests described in pharmacopoeias and can cause adverse clinical effects.

[0046] To remedy this situation, BAXTER proposed to focus efforts on detecting Gram-positive microbial contaminants in its peritoneal dialysis solutions, or directly on purified glucose polymers intended to be used in the composition of said peritoneal dialysis solutions.

[0047] In particular, in its patent EP 1,720,999, the company BAXTER proposes to develop a method based on the detection of peptidoglycans.

[0048] This detection is then recommended directly on preparations for peritoneal dialysis, or on purified glucose polymers.

[0049] This method involves performing the following on glucose polymers: a "bioburden" assay to detect the thermophilic acidophilic Gram-positive microorganism Alicyclobacillus acidocaldarius,then a sterilization of said glucose polymers, then a test consisting of adding a reagent capable of reacting with peptidoglycans to induce a serine-protease cascade reaction, a quantification of said peptidoglycans.

[0050] In other words, the first step detects the presence of living microorganisms that may have contaminated the glucose polymer purification circuit, the second, by sterilization, eliminates them, and the third detects peptidoglycans from cellular debris not eliminated by the previous sterilization step.

[0051] If it is determined that the quantity of these peptidoglycans sought in the glucose polymers is well below a certain threshold (10 ng / ml of a 7.5% glucose polymer solution, i.e. 133 ng / g of glucose polymers), these glucose polymers are then used to prepare the peritoneal dialysis solution itself.

[0052] In other words, to prevent the occurrence of these episodes of aseptic peritonitis, the company BAXTER, in its patent EP 1,720,999, proposes to ensure the control of the manufacture and use of peritoneal dialysis solutions, using a protocol for detecting peptidoglycans in the peritoneal dialysis solution.

[0053] BAXTER has also proposed no fewer than four other processes for detecting or avoiding the presence of peptidoglycans in components intended for the manufacture of peritoneal dialysis solutions, or in peritoneal dialysis solutions as such: In its international patent application WO 2009 / 117302, BAXTER describes the removal of contaminating substances by passing over an ion exchange resin.

[0054] It is regrettable, however, that this technology is only partially effective, the inventors themselves acknowledging that the purpose of this resin adsorption is to "reduce" the risk of contamination, as it only retains a "portion" of the microbial contaminants present in both the glucose polymers and the final solution for peritoneal dialysis.

[0055] It is even recommended to pass glucose polymers or peritoneal dialysis solution through several ion exchange resins in series, which constitutes a particularly heavy decontamination process.

[0056] A peptidoglycan detection test is also implemented (IL6 analysis) to verify the safety of the solutions thus obtained.

[0057] In its international patent application WO 2009 / 117303, a method for detecting peptidoglycans in glucose polymers or directly in the peritoneal dialysis preparation is proposed, comprising determining the response to IL-6 established as a function of a pro-inflammatory substance concentration taken as a reference, and establishing a dose / response curve of the response to IL-6 in relation to various pro-inflammatory substance concentrations.

[0058] This method therefore constitutes an improvement on the detection method developed in patent application EP 1,720,999.

[0059] However, this method requires the use of IL6-producing cells isolated from human subjects hypersensitive to particular pro-inflammatory substances, in this case peptidoglycans.

[0060] Once again, this diagnostic method allows for the rejection of contaminated batches with high sensitivity, but does not allow for the prevention of said contamination.

[0061] In its international patent application WO 2009 / 117304, it is a question of filtering the glucose polymer solution or the preparation for peritoneal dialysis on an ultrafiltration membrane having a cut-off threshold of 30 kDa.

[0062] BAXTER would thus have established that the size of contaminants between 30 and 100 kDa would be more likely to cause aseptic peritonitis, while smaller size contaminants would have no adverse effect.

[0063] The process described by BAXTER in this patent application therefore consists of filtering any glucose polymer solution or preparation intended for peritoneal dialysis on this ultrafiltration membrane and testing, using the same detection test as that described in patent application WO 2009 / 117303, the retentate and the filtrate thus obtained.

[0064] The principle is as follows: Glucose polymer solutions are eliminated if, after ultrafiltration, the retentate causes a pro-inflammatory reaction and the filtrate does not.

[0065] However, glucose polymer solutions will be retained if the retentate and filtrate cause a pro-inflammatory reaction and the response of the filtrate is greater than that of the retentate.

[0066] If both the retentate and the filtrate elicit a pro-inflammatory response, and the filtrate response is less than that of the retentate, the nature of the filtrate response is examined: If the response of the filtrate is 50% or more equal to that of the retentate, the solutions are kept; if the response of the filtrate is less than 50% of that of the retentate, the solutions are rejected.

[0067] This method is not satisfactory, however, because it is based on the assumption that small fragments are not harmless, and therefore no technique is proposed to eliminate them.

[0068] In its patent application 2009 / 117558, BAXTER acts through enzymes that degrade the cell wall components of microbial contaminants.

[0069] These enzymes, such as lysozyme, used in soluble or immobilized form, are implemented alone or in combination with other enzymes for the degradation of microbial contaminants, and the glucose polymer or dialysis preparation thus treated is tested for its response to cytokines.

[0070] This enzymatic treatment on the final product or on the dialysis preparation will then actually be prescribed to patients if no response to cytokines is detected.

[0071] However, once again, nothing is proposed to eliminate the debris from the hydrolyzed cell wall components, nor to eliminate the substances that naturally contaminate the enzymatic preparations used.

[0072] From all the above, it follows that no technical solution for securing glucose polymers intended for the preparation of peritoneal dialysis solution, or for securing the peritoneal dialysis solution itself, is completely satisfactory.

[0073] It is to the credit of the Applicant Company that it has found that the solution must be based primarily on quality control of the raw material used for the preparation of glucose polymers intended for the preparation of peritoneal dialysis solution, i.e. to propose a simple and effective process for eliminating contaminants from starch hydrolysates as such.

[0074] In other words, it is about addressing the contamination problem at its source, and not just on the final products or on the final solution for peritoneal dialysis.

[0075] The present invention therefore relates to a process for preparing glucose polymers for the manufacture of peritoneal dialysis solutions, a process which includes the decontamination of starch hydrolysates from which said glucose polymers will be prepared.

[0076] This decontamination aims to guarantee the production of starch hydrolysates that do not contain residual contaminants, i.e., guaranteeing levels less than or equal to the following values: For live microorganisms: total mesophilic flora: < 50 / g; molds and yeasts: < 15 / g; acidothermophilic Bacillus: < 10 / g. For endotoxins (LPS) and β-glucans, via the LAL test (endpoint gel clotting method) using reagents manufactured by CHARLES RIVER-ENDOSAFE (LAL lysate with a sensitivity of 0.015 EU / ml, ref. OR15015, and CSE endotoxins 500 ng or 10 ng per vial, ref. E110 or E120): ≤ 0.6 EU / g for peptidoglycans and β-glucans via a high sensitivity test developed by the Applicant company: < 8 ng / g of glucose polymers.

[0077] By "high sensitivity test developed and validated by the Applicant company", we mean a test developed and validated by the Applicant company, by adapting the SLP-HS single set kit ref. 293-58301 manufactured and marketed by WAKO Pure Chemical Industries Ltd.

[0078] This test involves using the so-called "SLP-HS" (Silkworm Larvae Plasma-High sensitivity) reagent, prepared from silkworm plasma, capable of: react with peptidoglycans and β-glucans contained in a 5% glucose polymer solution in water (e.g., special water for LAL testing), induce a serine-protease cascade reaction and detect and / or quantify said peptidoglycans and β-glucans using a TOXINOMETER tube reader manufactured and marketed by WAKO Pure Chemical Industries Ltd at very low thresholds, i.e., a limit of detection (LOD) at a threshold of approximately 0.05 ng / ml (i.e., 1 ng / g of glucose polymer) and a limit of quantification (LOQ) at a threshold of approximately 0.15 ng / ml (3 ng / g of glucose polymer). (LD and LOQ determined in the tested glucose polymer product)

[0079] More specifically, the SLP-HP test consists of: Prepare the glucose polymer to be tested in a 5% solution in suitable quality water (e.g., special water for LAL testing), and prepare a peptidoglycan standard curve in water over the application range of 0.04 to 2.5 ng / ml (target values) using the peptidoglycan standard (extract of Staphylococcus aureus) Using the SLP-HS single set kit for establishing a calibration line (linear regression on a logarithmic scale Ta = f(PG content)), introduce 100 µl of the prepared solution to be tested into the HS-SLP tube after reconstitution by adding 100 µl of the diluent (supplied in the previously mentioned kit), introduce the SLP-HS tube into the incubation well of the TOXINOMETER tube reader (WAKO Pure Chemical Ltd) thermostated at 30°C and parameterized according to the conditions prescribed by the manufacturer, the PG content of the solution to be tested being calculated using the established calibration line.

[0080] The result is expressed in ng / ml of the 5% solution tested and then in ng / g of glucose polymer.

[0081] The invention relates to a process for decontaminating a starch hydrolysate, for the preparation of glucose polymers intended for peritoneal dialysis, characterized in that it comprises the following steps: 1) prepare a starch hydrolysate by enzymatic or chemical hydrolysis of starch to obtain a Dextrose equivalent (DE) of less than 20, or by acid hydrolysis of a waxy starch milk to an DE between 8 and 15, 2) filter said starch hydrolysate to remove any contaminant the size of a microorganism such as yeasts, molds or bacteria, either by means of membrane filtration with a pore diameter of 0.22 µm, optionally preceded by membrane prefiltration with a pore diameter of 0.45 µm, or by means of ultrafiltration on a membrane with a cut-off threshold of 300,000 Da, 3) treat said starch hydrolysate thus freed from contaminating microorganisms with an enzyme for the degradation of the polysaccharide constituents of cell walls, laminarinase, laminarinase being introduced into the starch hydrolysates at a concentration of 0.001‰ to 1% on a dry weight basis of starch hydrolysates, and being implemented at 10% dry matter at a temperature of 50°C, at a pH of 4.6, for 5 to 24 hours, 4) ultrafilter the starch hydrolysate thus enzymatically treated with a membrane having a cut-off threshold of 20,000 Da to 50,000 Da, 5) treat the ultrafiltration permeate containing the starch hydrolysate thus obtained on high adsorption capacity activated carbon, 6) collect the starch hydrolysate thus decontaminated.

[0082] In the first step of this process according to the invention, the starch hydrolysate can be prepared: either by classical hydrolysis by enzymatic or chemical means of starch from various plant sources such as wheat, corn, potato, pea, rice, cassava... in order to achieve a Dextrose equivalent (DE) of less than 20, characteristic of maltodextrins, or by acid hydrolysis of a waxy starch milk up to a DE between 8 and 15 possibly completed by enzymatic hydrolysis using bacterial alpha-amylase up to a DE between 11 and 18, according to the teaching of patent EP 667.356 mentioned above.

[0083] To test the robustness of its decontamination process, as will be exemplified below, the Applicant company has access to hydrolysates of corn starch or maltodextrins that are artificially or naturally contaminated: by the 4 categories of contaminants presented above, by peptidoglycans (in free or bound form, i.e. on the surface of living cells) with or without endotoxins, by β-glucans and peptidoglycans, with or without endotoxins.

[0084] In the second step of this process according to the invention, microorganisms likely to contaminate said starch hydrolysates, i.e. yeasts, molds and bacteria, and in particular acid-thermophilic bacteria of type 1, are eliminated. Alicyclobacillus acidocaldarius, their size being greater than the diameters of the filtration pores.

[0085] Any method known to a person skilled in the art is usable, but the Applicant recommends implementing: either a sterilizing filtration which consists mainly of a membrane filtration with a pore diameter of 0.22 µm, preceded where appropriate by a membrane prefiltration with a pore diameter of 0.45 µm.

[0086] Filtration is achieved using several cartridge filters inserted into a vertical housing towards which the starch hydrolysate is directed. These cartridge filters are supplied by PALL or MILLIPORE. The cartridge sizes can be 10, 20, or 30 inches, and the number of cartridges installed provides sufficient filtration surface area to allow for a product flow rate between 1 and 20 L / min / m².

[0087] These cartridge filters are designed to withstand continuous operation at high temperatures, around 75°C, and to handle the aforementioned flow rate for over 700 hours. Operating at 75°C helps to limit microbiological growth, particularly of thermophilic flora.

[0088] Their temperature resistance also allows for sterilization before use. This sterilization process involves passing 2 bar steam through the casing for 20 minutes. This is followed by rinsing with purified water for 5 minutes.

[0089] These filters also possess resistance capabilities to certain chemicals used for equipment cleaning, including peracetic acid at a concentration of 5%.

[0090] An integrity test can be performed on these cartridges using an integrity test tool such as the one offered by MILLIPORE. This integrity test is carried out when the cartridges are installed to verify their correct assembly. The test is then performed before each cleaning of the equipment and finally before disassembly to validate their proper functioning during the production phase.

[0091] The operating pressure difference (ΔP) of these filters must not exceed 2 bar to guarantee their integrity. If it does, these filters must be replaced with new ones. i.e., ultrafiltration on a membrane with a cut-off threshold of 300,000 Da.

[0092] The cutoff threshold is thus chosen in such a way as to retain any potential cellular contaminants in the retentate.

[0093] The cut-off threshold allows the retention of microorganisms, i.e. yeasts, molds and bacteria, and in particular acidothermophilic bacteria of the type Alicyclobacillus acidocaldarius.

[0094] The filter surface area is determined based on the nature of the fluid and the flow rate to be treated.

[0095] Ultrafiltration membranes can be ceramic or organic. These two types of membranes have different resistances to temperature and chemicals; ceramic membranes are preferred as they can operate at temperatures above 75°C.

[0096] Their temperature resistance allows for steam sterilization before use. This sterilization process involves passing 2 bar steam through the casing for 20 minutes. This is followed by rinsing with purified water for 20 minutes.

[0097] It is also advisable to work at a temperature of around 75°C to avoid any microbiological development.

[0098] These filters also possess resistance capabilities to certain chemicals used for equipment cleaning, including peracetic acid at a concentration of 5‰ and caustic soda at a concentration of 1%.

[0099] The pressure of the starch hydrolysate feed is between 2 and 20 bar and is regulated by the pump supplying this module. If the maximum pressure is reached but the starch hydrolysate flow rate is too low, the membranes should be cleaned with caustic soda to restore their full efficiency.

[0100] Monitoring the reduction in contaminant levels can be analyzed by taking periodic samples from the permeate.

[0101] Effective elimination of microorganisms will be considered if the samples taken from the filtrate show a level below the quantification threshold of conventional assay methods.

[0102] In the third step of this process according to the invention, the said starch hydrolysate, thus freed from contaminating microorganisms, is treated by a degrading enzyme of the polysaccharide constituents of cell walls, laminarinase.

[0103] After the elimination of living microorganisms, this step therefore makes it possible to eliminate residual toxic molecules (released by microorganisms or still present in the debris of cell membranes).

[0104] This enzyme for the degradation of the polysaccharide constituents of cell walls is chosen so as to break the membranes of the two major contaminating germs of starch hydrolysates (yeasts and Gram (+) bacteria) which could have escaped the previous filtration step, and especially to hydrolyze their β-glucans and peptidoglycans which may be present in free form in said starch hydrolysates.

[0105] Laminarinase exhibits endo-β-1,3 glucanase activity (EC 3.2.1.6), which destroys the cell wall of yeasts and some molds.

[0106] Lysozyme, an acid hydrolase (EC 3.2.1.17), destroys the bacterial cell wall of Gram-positive bacteria by catalyzing the hydrolysis of its constituent glycosaminoglycans. It hydrolyzes the covalent bonds between N-acetylmuramic acid and the 4th carbon atom of the N-acetylglucosamine in peptidoglycans. The peptidoglycan "skeleton" of the bacterial cell wall is, in fact, a copolymer composed of alternating covalent links between these two molecules, to which peptides are attached that bridge the polymer chains.

[0107] Laminarinase is introduced into starch hydrolysates at a concentration of 0.001‰ to 1% on a dry weight basis of starch hydrolysates, preferably between 0.1% and 0.5%.

[0108] Laminarinase is applied to 10% dry matter starch hydrolysates at a temperature of 50°C, at a pH of 4.6, for 5 to 24 hours, preferably for 20 hours, as will be exemplified below.

[0109] Lysozyme, on the other hand, can be used on starch hydrolysates with 10% dry matter at a temperature of 37°C, at a pH of 7, for 5 to 24 hours, preferably for 20 hours, as will be exemplified below.

[0110] In the fourth step of this process according to the invention, tangential ultrafiltration of the starch hydrolysate thus enzymatically treated is carried out.

[0111] The ultrafiltration membrane cutoff threshold is chosen to retain any degradation products of β-glucans and peptidoglycans in the retentate and, above all, to retain the enzymes used in the previous step.

[0112] The ultrafiltration membrane then has a cut-off threshold between 20,000 and 50,000 daltons, preferably in the order of 30,000 daltons.

[0113] In the fifth step of this process according to the invention, the starch hydrolysate thus obtained, i.e. the permeate or filtrate, is passed over high adsorption capacity activated carbon.

[0114] The Applicant recommends that particular care be taken in the implementation of this activated carbon treatment step, which constitutes a key step in this first variant of the process according to the invention.

[0115] The Applicant company has indeed found that the choice of an appropriate quality of activated carbon to be used in this finishing step conditions the near absence of contaminants likely to be detected in the starch hydrolysates thus treated ("near" meaning at concentrations well below the threshold of quantification of conventional assay methods).

[0116] It is known to those skilled in the art that activated carbon is an inert, adsorbent carbonaceous material consisting of a porous network developing a considerable surface area of ​​up to 1,500 m² per gram of material (pore diameters varying between 4 and 100,000 Angstroms).

[0117] To be effective, activated carbon must therefore have a certain internal surface area developed within the porous structure on which the impurities that we wish to eliminate will adsorb.

[0118] To model this adsorption capacity, a person skilled in the art usually describes the adsorption process using an adsorption isotherm (Langmuir isotherm, Langmuir-Hinshelwood isotherm, BET isotherm or Freundlich isotherm).

[0119] The adsorption isotherm is a curve that represents the amount of impurities adsorbed per unit mass of activated carbon as a function of the residual concentration of impurities in solution.

[0120] To construct this isotherm, known and increasing quantities of activated carbon are introduced into a volume of solution to be treated, and after a given contact time, the residual concentration of adsorbed impurities is measured.

[0121] The Applicant company therefore recommends, as a preamble to any implementation of activated carbon intended here to complete the decontamination of starch hydrolysates intended for the preparation of glucose polymers intended for peritoneal dialysis (fifth finishing step), to precisely choose the quality of the activated carbon by modeling its adsorption isotherm.

[0122] A preliminary test is thus carried out to choose the most effective activated carbon, a test which consists of tracing the Freundlich adsorption isotherm for different qualities of commercially available activated carbon (or even testing different batches of the same type of activated carbon), with regard to their capacity to adsorb known quantities of contaminants such as bacterial endotoxins and peptidoglycans which are taken here as references.

[0123] The aim is therefore to carry out a precise qualification of all batches intended to be used in this key finishing stage of starch hydrolysates.

[0124] The Applicant recommends testing more specifically here the "meso" type quality, i.e. presenting a high methylene blue index, or the "micro" type quality, i.e. presenting a high iodine index.

[0125] Thus, the method may include a preliminary step of determining the adsorption isotherm of several activated carbons, in particular one of the "meso" type and one of the "micro" type, and selecting the most suitable activated carbon.

[0126] The adsorption isotherm, which is classically used to determine the specific surface area of ​​a given activated carbon, will therefore be used here to help in choosing the activated carbon that will: to adsorb all the degradation products of β-glucans and peptidoglycans generated in the previous steps, regardless of their size, and especially to adsorb the endotoxins likely to contaminate the starch hydrolysates, and also the endotoxins brought by the enzymes in use (unfortunately, no commercially available industrial preparation of these enzymes is free of them).

[0127] As will be exemplified below, the procedure for determining the Freundlich adsorption isotherm recommended by the Applicant consists of mixing the incoming starch hydrolysate with each batch of activated carbon (5 increasing doses ranging from 0.125% to 2% on the dry matter to be treated, i.e. 0.125%; 0.25%; 0.5%; 1% and 2%) at a temperature of 75°C for one hour.

[0128] The temperature of 75°C is the classic implementation temperature of commercially available activated carbons, validated for adsorption kinetics.

[0129] The duration of one hour of contact time was determined in such a way as to eliminate the kinetic constraint with respect to the thermodynamics of the system.

[0130] As for the pH, it is set at a value of 4.5, because the Applicant company determined that in order to ensure optimal adsorption of peptidoglycan impurities and endotoxins, it was necessary to first condition the starch hydrolysate sample to an acidic pH value, i.e. of the order of 4.5.

[0131] As will be exemplified below, the quality of the activated carbon for this fifth finishing stage was identified as a "micro" type activated carbon with a high iodine index.

[0132] A NORIT ®< SX + type activated carbon, marketed by the NORIT company, is perfectly suited to this quality.

[0133] In the final step of this process according to the invention, the resulting starch hydrolysate is then collected, and its contaminant content is analyzed.

[0134] The invention will be better understood with the help of the following examples, which are intended to be illustrative and not limiting. Reference example 1 : Preparation of starch hydrolysate in accordance with the teaching of patent EP 667.356

[0135] The raw material for obtaining the glucose polymers according to the invention is produced from waxy maize starch in the following manner: cleaning the corn to keep only whole kernels, soaking the cleaned corn in lactic acid to soften the kernels, wet milling, then separating the different components, i.e.germ, cellulose envelope, proteins and starch, cleaning of the starch by counter-current filtration with disinfected water to purify the starch both physico-chemically and bacteriologically, centrifugation and drying of the starch, suspension of the starch in disinfected water to a final dry matter content of 40% and a temperature of 45°C to 50°C, acidification of the starch suspension by the addition of HCl to a pH < 2, and increasing the temperature to 115 to 120°C for 6 to 8 minutes, flocculation of proteins and fats at this pH, neutralization of the suspension to pH 5, filtration of the suspension through diatomaceous earth (to retain residual proteins, fats and cellulose), demineralization on strong cationic resin and weak anionic resin, decolorization on standard activated carbon, atomization of the concentrated solution in an MSD-type atomizer marketed by the company NIRO. .

[0136] Particularly contaminated batches, referenced "A-5250" and "B-3063", were chosen separately to test the validity of the processes according to the invention (more or less difficult to decontaminate depending on the nature and size of these different contaminants: peptidoglycans, β-glucans and endotoxins).

[0137] Table I below presents the nature of the contaminants identified in these different batches.

[0138] For comparison purposes, the nature of the contaminants that can be found in six batches of commercial “C” maltodextrins is also presented. Tableau I. Starch hydrolysates Alicyclo bacillus acidocaldarius (Germ name south 1 g) Yeast (Number of germs per 1 g) Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Distribution of β-glucans / Endotoxins (%) Measurement methods LAL Test High sensitivity test Modified LAL Test lot “A-5250” no. 1 0 < 5 9, 6 2.462 90 / 10 lot “A-5250” no. 2 0 < 5 19,2 2.576 90 / 10 lot “A-5250” no. 3 0 < 5 9, 6 1.882 90 / 10 lot “A-5250” no. 4 0 < 5 4, 8 3.166 90 / 10 lot “A-5250” no. 5 0 < 5 9, 6 2.277 90 / 10 lot “A-5250” no. 6 0 < 5 9, 6 3.540 90 / 10 Lot “B-3063” No. 1 150 < 5 2,4 36.029 0 / 100 Lot “B-3063” No. 2 150 < 5 2,4 25.029 0 / 100 Lot “B-3063” No. 3 150 < 5 9, 6 27.260 0 / 100 Lot “B-3063” No. 4 150 < 5 1,2 10.241 0 / 100 Lot “B-3063” No. 5 150 < 5 2,4 55.691 0 / 100 Lot “B-3063” No. 6 160 < 5 2,4 46.367 0 / 100 Lot “C” No. 1 2 < 5 9, 6 13.660 75 / 25 Lot “C” No. 2 2 < 5 3, 6 9.374 75 / 25 Lot “C” No. 3 2 < 5 9, 6 10.556 75 / 25 Lot “C” No. 4 2 < 5 4, 8 9.353 75 / 25 Lot “C” No. 5 2 < 5 9, 6 11.920 75 / 25 Lot “C” No. 6 2 < 5 9, 6 16.288 75 / 25

[0139] These different batches therefore offer three possible contamination options: 6 batches “A-5250” contaminated mainly by peptidoglycans (free polysaccharides) and β glucans, 6 batches “B-3063” contaminated mainly by peptidoglycans (living cells or free polysaccharides) and endotoxins, without any trace of β glucans, 6 batches “C” whose contamination profile is broad spectrum, i.e. containing all categories of contaminants. Reference Example 2: Selection of activated carbon grades based on the nature and quantity of β-glucan and peptidoglycan contaminants likely to be found in starch hydrolysates from lots A-5250 and B-3063 and a commercial maltodextrin C

[0140] As mentioned above, the activated charcoal treatment steps are mainly implemented to adsorb cellular debris such as endotoxins, β-glucans and peptidoglycans.

[0141] The choice of activated carbon quality is based on the analysis of Freundlich isotherms plotted for each activated carbon directly on the starch hydrolysate to be decontaminated.

[0142] As will be demonstrated below, the choice of activated carbon will depend directly on the contaminant load (in nature and quantity) present in the starch hydrolysates to be treated.

[0143] As indicated above, the procedure for determining the Freundlich adsorption isotherm consists of mixing the starch hydrolysate (at 10% dry matter) with each type of activated carbon (5 increasing doses ranging from 0.125% to 2% on the dry matter to be treated, i.e. 0.125%; 0.25%; 0.5%; 1% and 2%) at a temperature of 75°C for one hour at a pH of 4.5.

[0144] The relative effectiveness of "meso" and "micro" type activated carbons is then determined according to the nature of the contaminants in each batch by measuring the quantity of impurities adsorbed per unit mass of activated carbon as a function of the residual concentration (after one hour of reaction) of the impurities in solution. 2.1 Determination of the isothermal of Freundlich activated carbons of type NORIT SX + and ENO-PC versus two lots A- 5250

[0145] All batches A have a similar profile: contaminated mainly by peptidoglycans (polysaccharides in free form) and β glucans.

[0146] Table II below shows, for 5 increasing doses of activated charcoal, the residual levels of β glucans and residual peptidoglycans of batches A-5250 No. 1 and No. 2 of example 1, after treatment with the two grades of activated charcoal. Table II. Lot #1 NORIT SX + Lot #2 ENO-PC Dose of activated charcoal LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) 0 9, 6 2.462 19,2 2.576 0,125 1,2 555 0,3 416 0,25 0,3 127 0,3 156 0,5 0,3 10 < 0,3 113 1 0,3 < 1 < 0,3 13 2 < 0,3 < 1 < 0,3 < 1

[0147] If we analyze the adsorption curves for peptidoglycans (responsible for aseptic peritonitis), taken here as a reference, we observe that: The adsorption curve of NORIT SX+ activated charcoal, plotting y = Co − C dose de noir = f C où Co = initial contaminant dose C = dose of residual contaminants translates to the mathematical equation:

[0148] The adsorption curve of ENO-PC is represented by the following equation: Y = 1045 . x 0 , 4156 with an r 2< (correlation coefficient) of 0.9190.

[0149] The linearity of these two curves taken separately accurately reflects the adsorption efficiency of peptidoglycans by these two grades of activated carbon.

[0150] But comparing the two equations, it appears that: For a dose "x" of peptidoglycans, for example 2000 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 86%; while for a dose "x" of peptidoglycans, for example 10 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 180%.

[0151] It follows that the ENO-PC "mesopore" quality is well suited to the adsorption of large quantities of peptidoglycan contaminants, and that the NORIT SX + "micropore" quality is well suited to the adsorption of small quantities of residual peptidoglycans.

[0152] This result allows, for lot A-5250, to define the order of implementation of activated carbons in the decontamination steps of this particular lot of starch hydrolysate: a treatment step with ENO-PC activated carbon followed by a treatment step with NORIT SX + activated carbon. 2.2 Determination of Freundlich isotherms of NORIT SX+ and ENO-PC type activated carbons versus two batches B-3063

[0153] All batch B samples have a similar profile: contaminated mainly by peptidoglycans and endotoxins.

[0154] Table III below shows, for 5 increasing doses of activated charcoal, the residual levels of β glucans and residual peptidoglycans of lot B-3063 No. 1 and No. 2 of example 1, after treatment with the two grades of activated charcoal. Tableau III. Lot #1 NORIT SX + Lot #2 ENO-PC Dose of activated charcoal LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) 0 2,4 36.029 2,4 25.029 0,125 0,5 140 0,3 928 0,25 < 0,3 50 < 0,3 300 0,5 < 0,3 10 < 0,3 83 1 < 0,3 < 1 < 0,3 10 2 < 0,3 < 1 < 0,3 5

[0155] Analyzing the adsorption curves for peptidoglycans (responsible for aseptic peritonitis) reveals that: The adsorption curve of NORIT SX+ activated charcoal is represented by the following mathematical equation:

[0156] The adsorption curve of ENO-PC is represented by the following equation: Y = 6543 . x 0 , 4849 with an r 2< (correlation coefficient) of 0.9813.

[0157] The linearity of these two curves taken separately accurately reflects the adsorption efficiency of peptidoglycans by these two grades of activated carbon.

[0158] But comparing the two equations, it appears that: For a dose "x" of peptidoglycans, for example of 2000 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 279%; while for a dose "x" of peptidoglycans, for example of 10 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 417%.

[0159] It follows that the ENO-PC "mesopore" grade is not suitable for the adsorption of peptidoglycans from this B-3063 batch, and that the NORIT SX+ "micropore" grade can be used here for the adsorption of small and large quantities of peptidoglycans.

[0160] This result allows, for batch B-3063, to define the order of implementation of activated carbons in the decontamination steps of this particular batch of starch hydrolysate: here two NORIT SX + type activated carbon steps in series. 2.3 Determination of Freundlich isotherms of NORIT SX+ and ENO-PC type activated carbons versus two batches C

[0161] The following table IV shows, for 5 increasing doses of activated charcoal, the residual levels of β glucans / endotoxins and residual peptidoglycans of batches C no. 1 and no. 2 of example 1, after treatment with the two qualities of activated charcoal. Table IV Lot #1 NORIT SX + Lot #2 ENO-PC Dose of activated charcoal LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) LAL detection method (EU / g) High sensitivity peptidoglycan test (ng / g) 0 9, 6 13.660 9, 6 9.374 0,125 1,2 775 1,2 896 0,25 < 0,3 190 0, 6 369 0,5 < 0,3 25 < 0,3 106 1 < 0,3 10 < 0,3 26 2 < 0,3 < 1 < 0,3 28

[0162] Analyzing the adsorption curves for peptidoglycans (responsible for aseptic peritonitis) reveals that: The adsorption curve of NORIT SX+ activated charcoal is represented by the following mathematical equation:

[0163] The adsorption curve of ENO-PC is represented by the following equation: Y = 783 , 97 . x 0 , 6559 with an r 2< (correlation coefficient) of 0.9353.

[0164] The linearity of these two curves taken separately accurately reflects the adsorption efficiency of peptidoglycans by these two grades of activated carbon.

[0165] But comparing the two equations, it appears that: For a dose "x" of peptidoglycans, for example of 2000 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 128%; while for a dose "x" of peptidoglycans, for example of 10 ng / g, the relative effectiveness of the quality of NORIT / ENO-PC activated carbons is 2349%.

[0166] It follows that the ENO-PC "mesopore" quality is not suitable for the adsorption of peptidoglycans from these C batches, and that the NORIT SX + "micropore" quality can be used here for the adsorption of small and large quantities of peptidoglycans. Example 3: Process for decontaminating starch hydrolysate using the lysozyme (reference) or laminarase (invention), followed by an ultrafiltration stage and then activated carbon

[0167] We choose here lots A-5250 #3 and #5, lots B-3063 #3 and #5, and lots of commercial maltodextrins C #3 and #5.

[0168] While it is relatively easy to eliminate microorganisms of the type Alicyclobacillus acidocaldarius and yeasts on a 0.2 µm microfiltration module or on an ultrafiltration module with a cut-off threshold of 300 kDa under the conditions explained above, the elimination of endotoxins, β-glucans and / or peptidoglycans requires the implementation of a combination of very specific treatment steps.

[0169] As described above, this combination of steps consists of: by using specific enzymes such as lysozyme or laminarinase, to degrade endotoxins, β-glucans and peptidoglycans, in order to significantly reduce their size, by the ultrafiltration step, to eliminate the enzymatic fraction and to retain the impurities brought in by the industrial enzymes themselves, by the final activated carbon treatment step, to adsorb all residual small debris.

[0170] With regard to this last activated carbon treatment step, the Freundlich isotherm curves as established in example 2 teach that the NORIT SX + quality is the one to be implemented.

[0171] Table V below presents the measurement of residual β-glucans and peptidoglycans contaminants after treatment of batches A-5250: to lysozyme (enzyme marketed by the company FUKA, with an activity of 70,000 U / mg), or to laminarinase (enzyme marketed by the company SIGMA under the brand name CYTOHELICASE ®<, with an activity of 1 U / mg).

[0172] It is important to note that these enzyme preparations are not free of contaminants.

[0173] It is thus determined that: FUKA lysozyme has an endotoxin level > 9.6 EU / ml and a petidoglycan level of 31000 ng / g, SIGMA CYTOHELICASE ® has an endotoxin level > 9.6 EU / ml and a petidoglycan level of 38000 ng / g. Tableau V. Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans lot “A-5250” no. 3 9, 6 1.882 1) Lysozyme (0.1% dose) 20 h incubation 9, 6 1.694 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 9, 6 639 3) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 217 lot “A-5250” no. 5 9,6 2.277 1) CYTOHELICASE ® < dose of 0.1 ‰ After 20 h of incubation 38,4 191 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 0,6 < 0,3 3) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 < 20

[0174] The results show that for lot A-5250 which will be recovered in the activated carbon treated ultrafiltration permeate, the laminarinase pretreatment is much more effective than the lysozyme treatment in removing β-glucans and peptidoglycans, in the sense that the lysozyme treatment has no effect at all on lot A-5250.

[0175] The following Table VI presents the measurement of residual endotoxin and peptidoglycan contaminants after treatment of batches B-3063. Table VI Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans Lot “B-3063” No. 3 9,6 27.260 1) Lysozyme (0.1% dose) 20 h incubation 2,4 4.467 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 1,2 30 3) NORIT SX+ activated carbon treatment 0,5 % On the ultrafiltration permeate < 0,3 < 20 lot “B-3063” no. 5 2,4 55.691 1) CYTOHELICASE ® < dose of 0.1 ‰ After 20 h of incubation 4, 8 185 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate < 0,3 < 20 3) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 < 20

[0176] The results show that for batch B-3063 which will be recovered in the activated carbon treated ultrafiltration permeate, pretreatment with laminarinase is much more effective than that with lysozyme in removing peptidoglycans.

[0177] Table VII below presents the measurement of residual β-glucans / endotoxins and peptidoglycans contaminants after treatment of batches C. Table VII Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans lot “C” no. 3 9, 6 10.556 1) Lysozyme (0.1% dose) 20 h incubation 9, 6 17.968 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 9, 6 534 3) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 20 lot “C” no. 5 9, 6 11.920 1) CYTOHELICASE ® < dose of 0.1 ‰ After 20 h of incubation 4,8 1293 2) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate < 0,3 76 3) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 < 20

[0178] The results show that for lot C which will be recovered in the activated carbon treated ultrafiltration permeate, the laminarinase pretreatment is much more effective than the lysozyme treatment in removing endotoxins / β-glucans and peptidoglycans, in the sense that the lysozyme treatment has no effect at all on lot A-5250.

[0179] These results demonstrate that the choice of cell wall degradation enzyme is not insignificant here and is indeed dependent on the nature of the contaminants present in these different batches.

[0180] Treatment with CYTOHELICASE ®< will clearly be preferable to that with lysozyme. Reference example 4: Treatment by ultrafiltration followed by activated carbon

[0181] As will be demonstrated, in certain cases, treatment by ultrafiltration alone can achieve a completely satisfactory reduction of endotoxins, β-glucans and peptidoglycans.

[0182] The process is conducted in the same way as described in example 3, but without the use of enzymes (therefore without modification of pH and temperature). 4.1 Processing on batch A-5250

[0183] Table VIII shows the residual β-glucan and peptidoglycan contaminant content after each step, for lot A-5250 #4. Table VIII. Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans Lot “A-5250” No. 4 4,8 3.166 1) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 4,8 883 2) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 < 20

[0184] The combination of ultrafiltration and activated carbon finishing steps ensures decontamination of the starch hydrolysates from batch A-5250.

[0185] If we compare with the values ​​of the process in Table V (excluding values ​​obtained by the lysozyme treatment) we note however that the reduction in β-glucans and peptidoglycans is better if the treatment with CYTOHELICASE ®< is carried out before the ultrafiltration step, even if, after the finishing treatment with NORIT SX + activated carbon, the results are identical.

[0186] This pre-treatment with CYTOHELICASE ®< is therefore recommended to optimize the effectiveness of the decontamination treatment (capability of the decontamination process according to the invention, i.e. its ability to produce precisely and repeatably starch hydrolysates free of contaminants). 4.2 Processing on batch B-3063

[0187] Table IX shows the residual contaminant levels of endotoxins and peptidoglycans after each step, for lot B-3063 #4. Tableau IX. Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans Lot “B-3063” No. 4 1,2 10.241 1) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 1,2 < 20 2) NORIT SX+ activated carbon treatment at 0.5% On the ultrafiltration permeate < 0,3 < 20

[0188] The combination of ultrafiltration and activated carbon finishing steps ensures decontamination of the starch hydrolysates from batch B-3063.

[0189] If we compare with the values ​​of the process in Table VI (excluding values ​​obtained by the lysozyme treatment) we note however that the reduction of endotoxins and peptidoglycans is also better if the treatment with CYTOHELICASE ®< is carried out before the ultrafiltration step, even if, after the finishing treatment with NORIT SX + activated carbon, the results are identical. 4.3 Batch treatment C

[0190] Table X shows the residual contaminant levels of endotoxins and peptidoglycans after each step, for batch C no. 4. Tableau X. Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) Measurement methods LAL Test High sensitivity test & peptidoglycans Lot C n°4 4, 8 9.353 1) UF treatment 30 kDa Volumetric Concentration Factor = 10 permeate 38, 4 340 2) NORIT SX+ activated carbon treatment 0,5 % On the ultrafiltration permeate < 0,3 < 20

[0191] The combination of ultrafiltration and activated carbon finishing steps ensures decontamination of the starch hydrolysates from batch C.

[0192] If we compare with the values ​​of the process in Table VII (excluding values ​​obtained by the lysozyme treatment) we note however that the reduction of endotoxins / β-glucans and peptidoglycans is also better if the treatment with CYTOHELICASE ®< is carried out before the ultrafiltration step, even if, after the finishing treatment with NORIT SX + activated carbon, the results are identical. Example 5: Process for decontaminating starch hydrolysate using a series of activated carbon steps.

[0193] The lots “A-5250” No. 6; “B-3063” No. 6 and “C” No. 6, having been cleared of their contaminating germs in the manner described in Example 3, are subjected to the following treatments. 5.1. Two treatments in series with activated carbon of the same quality.

[0194] This involves applying a double treatment with the same activated carbon, in this case NORIT SX + at 0.25% on dry, characteristic of a "micropore" type treatment.

[0195] Table XI below presents the results obtained. As a control, a single treatment with activated charcoal and a double dose of NORIT SX+ was also carried out. Tableau XI Starch hydrolysate Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) lot “A-5250” no. 6 departure 9, 6 3.540 1st< TN 2,4 593 2nd< TN < 0,3 < 10 TN "double dose" 0 60 Lot “B-3063” No. 6 departure 2,4 46.367 1st< TN < 0,3 223 2nd< TN < 0,3 < 10 TN "double dose" < 0,3 76 Lot “C” No. 6 departure 9, 6 16.288 1st< TN < 0,3 80 2nd< TN < 0,3 < 10 TN "double dose" < 0,3 14

[0196] It is clear that a two-step activated charcoal treatment is more effective than a single treatment with a double dose, especially for the elimination of peptidoglycans. 5.2 Two treatments in series with activated carbon of different qualities.

[0197] First treatment: Chemical black: NORIT ENO-PC at 0.25% of the "mesopore" type; Second finishing treatment: NORIT SX + at 0.25% of the "micropore" type.

[0198] Table XII below presents the results obtained. As a control, a single treatment with activated carbon using a mixture of the two grades of activated carbon was also carried out (referenced TN "mix"). Table XII Starch hydrolysate Endotoxins + β-glucans (EU / g) Peptidoglycans (ng / g) lot “A-5250” departure 9, 6 3.540 1st< TN 0,6 299 2nd< TN < 0,3 < 10 TN "mix" < 0,3 37 Lot “B-3063” Departure 2,4 46.367 1st< TN < 0,3 1096 2nd< TN < 0,3 < 10 TN "mix" < 0,3 91 Lot "C" Departure 9, 6 16.228 1st< TN 1,2 777 2nd< TN < 0,3 11 TN "mix" < 0,3 68

[0199] It also appears here that a two-step treatment is more effective than a single-step treatment, mixing the two qualities of activated charcoal.

[0200] For lot A-5250, it appears that the first mesopore treatment allows a more effective removal than the micropore treatment, ensuring a micropore treatment finish which guarantees a contaminant level below the quantification threshold of classical assay methods (in accordance with the teaching of example 2).

[0201] For batches 3063 and C, as found in example 2: a double treatment with the "micropore" quality NORIT SX + is the most effective.

Claims

1. A method for decontaminating starch hydrolysates from which glucose polymers for producing peritoneal dialysis solutions will be prepared, characterized in that it comprises the following steps: 1) preparing a starch hydrolysate prepared by enzymatic or chemical hydrolysis of starch so as to achieve a dextrose equivalent (DE) of less than 20, or prepared by acid hydrolysis of a waxy starch milk to give a DE between 8 and 15, 2) filtering said starch hydrolysate so as to remove any contaminant having the size of a microorganism of yeast, mold or bacteria type, either carried out by means of a membrane filtration where the pore diameter is 0,22 µm, optionally preceded by a membrane filtration where the pore diameter is 0,45 µm, or carried out by means of a membrane of ultrafiltration where the cut-off threshold is 300.000 Da 3) treating said starch hydrolysate which the contaminating microorganisms have thus been removed by an enzyme for degrading the cell wall polysaccharide constituents, the laminarinase, the laminarinase being introduced into the starch hydrolysates at the concentration of 0,001 ‰ to 1% on a dry weight basis of starch hydrolysates, and being used at 10% of dry mass at a temperature of 50°C, at a pH of 4,6, for 5 to 24 hours, 4) ultrafiltering the starch hydrolysate thus enzymatically treated with a membrane where the cut-off threshold is from 20.000 Da to 50.000 Da, 5) treating the resulting ultrafiltration containing the starch hydrolysate on activated carbon with a high adsorption capacity, 6) collecting the starch hydrolysate thus decontaminated.

2. The method as claimed in claim 1, characterized in that the ultrafiltration of step 4) consists of a membrane of ultrafiltration where the cut-off threshold is 30.000 Da.

3. The method as claimed in claim 1 or 2, characterized in that step 5) of treating with activated carbon with a high adsorption capacity consists of a quality of activated carbon of "micropore" type.