WO2025040552A1

WO2025040552A1 - Method for sterile filtration of aqueous solutions containing albumin

Info

Publication number: WO2025040552A1
Application number: PCT/EP2024/072965
Authority: WO
Inventors: Michael BIERI; Michele ANDLER HIMMELSPACH
Original assignee: CSL Behring AG
Current assignee: CSL Behring AG
Priority date: 2023-08-18
Filing date: 2024-08-15
Publication date: 2025-02-27
Anticipated expiration: 2026-02-18
Also published as: CH720224B1

Abstract

The present invention relates to a method for sterile filtration of an albumin-containing aqueous solution (2), wherein the method comprises filtering the solution (2) with at least one filter (3) to obtain a filtrate (1), wherein the at least one filter (3) is conditioned prior to filling the filtrate (1) into at least one container (10, 10') by purging the at least one filter (3) with a quantity (5) of the albumin-containing aqueous solution (2) i) at a differential pressure across the respective filter (3) of at most 1 bar and/or ii) by means of pressure filtration, wherein the filtrate (1) is maintained in the outlet (4) of the respective filter (3) at a back pressure of at least 0.5 bar is maintained.

Description

Method for sterile filtration of aqueous solutions containing albumin

The present invention relates to a method for sterile filtration of an albumin-containing aqueous solution, as well as to an al- bumin solution obtained by a method disclosed herein. Further disclosed herein is the use of conditioned filters to reduce the opalescence of filtered aqueous solutions containing albumin.

Albumins of various types are of great importance as biochemi- cals, e.g. as components of culture media, cell culture media and as stabilisers for diagnostic agents in biology, medicine and pharmacy.

In the course of the manufacture of aqueous human albumin solu- tions, an obligatory step is the final pasteurisation of the product at 60 °C for 10 hours. This heating step was introduced to inactivate infectious agents and to eliminate or reduce the risk of transmission of viruses, such as the hepatitis B virus, that may occur despite the various purification steps, owing to possible contamination of the biological raw material. Cur- rently, all national and international pharmacopoeias require the pasteurisation of albumin solutions at 60 °C for 10 hours, it being necessary to effect this pasteurization at the abso- lutely last stage of the manufacture, namely in the final con- tainer, which is usually a vial made of pharmaceutical type 1 borosilicate glass (also known as "neutral" glass).

Notwithstanding the relative stability of albumin with regard to heat, the albumin must be protected in order to prevent gelling during pasteurisation, with the aid of suitable stabilisers. Currently, sodium caprylate and sodium acetyltryptophanate are generally used as stabilisers, either alone or in combination. Sodium mandelate may also be used on its own or in combination with sodium caprylate. The stabilisers are added in large excess relative to the albumin and, owing to their own affinity for al- bumin, are able to effectively protect it against direct dena- turation by heat. Insufficient stabilisation leads to denatura- tion of the albumin and thus to progressive aggregation of the albumin molecules and any other heat-sensitive impurities pre- sent, resulting in the formation of higher molecular weight par- ticles. These particles lead to an opalescent appearance of the aqueous albumin solutions due to light scattering, a property often referred to as opalescence. Opalescence refers to the characteristic optical effect caused by light scattering in a milky-cloudy material, which is characterised by a coloured shimmer. According to the European Pharmacopoeia, opalescence is classified from 0 to IV with 0, 3, 6, 18 or 30 nephelometric turbidity units (NTU) (European Pharmacopoeia 2016, EP9, section 2.2.1: "Clarity and degree of opalescence in liquids").

However, stability studies have shown that even when using the stabilisers mentioned above, opalescence can occur during or af- ter pasteurisation when the albumin-containing aqueous solution is filtered prior to pasteurisation. This opalescence is clearly visible in the Seidenader semi-automatic visual inspection sys- tem and also with the naked eye when a vial containing the fil- tered albumin solution is placed in front of an indirect back- light. Opalescence can occur for the most dilute albumin solu- tions, i.e. with 4% or 5% protein content, as well as for the most concentrated solutions, i.e. with 20% or 25% protein con- tent. Furthermore, it has been observed that vials showing opal- escence develop white ring deposits on the inside of the vial at the liquid-air interface after a few months, approximately after 2-3 months of storage. Although the appearance of white ring de- posits correlates with a decrease in the opalescence of the aqueous albumin solution, it is nevertheless not tolerated ac- cording to national and international pharmacopoeias.

US 5,118,794 discloses a process for stabilising human albumin solutions for therapeutic purposes for the purpose of their heat treatment in a container, wherein the method comprises the addi- tion of a surfactant in addition to the usual stabilising for- mula. However, challenges associated with surfactants have been encountered, for instance, impurities, degradation and possible triggering of adverse immune reactions.

More generally, WO 2008/045563 relates to methods for reducing the opalescence of protein solutions and to compositions of con- centrated protein solutions with reduced opalescence. The meth- ods include altering the ionic strength of the solutions such that the opalescent appearance and/or the amount of higher mo- lecular weight particles in the protein solution is reduced and/or eliminated, for example by decreasing the concentration of a salt present in the solution. In addition, this document does not contain information on the reduction of opalescence in albumin-containing solutions.

The similarity of opalescent solutions to aggregated protein so- lutions has raised concerns about the loss of protein activity and the potential to cause immunogenicity in pharmaceutical for- mulations (CLELAND, J.L., The Development of Stable Protein For- mulations: A Close Look at Protein Aggregation, Deamidation, and Oxidation, Crit. Rev. Therapeutic Drug Carrier Systems 1993, Vol. 10, No. 4, pages 307-377). Since opalescence and/or the ap- pearance of deposits during pasteurisation indicate a certain denaturation of the product and raise doubts about its good tol- erability and efficacy when administered to the patient, the de- tection of opalescence and/or deposits requires the rejection of the vials concerned during the final quality control inspection, which leads to economic losses.

The object of the present invention is to remedy these and other disadvantages of the prior art and, in particular, to provide a method for sterile filtration of an albumin-containing aqueous solution, in particular an aqueous solution containing the albu- min component of human blood. A further object of the present invention is to provide safe and stable albumin solutions, in particular containing the albumin component of human blood, in which the opalescence associated with particle formation is eliminated or at least reduced. Finally, it is a further object of the present invention to provide methods for conditioning a filter which can be used to reduce the opalescence of filtered albumin-containing aqueous solutions.

The objects are solved by the independent claims, wherein the dependent claims relate to preferred embodiments.

Accordingly, in one aspect, the present invention features a method of sterile filtration of an albumin-containing aqueous solution, in particular an aqueous solution containing the albu- min portion of human blood. The method comprises filtering the solution with at least one filter to obtain a filtrate. The method is characterised in that it further comprises a step of conditioning the filter prior to filling the filtrate into at least one container by purging the at least one filter with a quantity of the albumin-containing aqueous solution, at a dif- ferential pressure across the respective filter of at most 1 bar and/or by pressure filtration, wherein the filtrate is main- tained at a back pressure of at least 0.5 bar (0.05 MPa) at the outlet of the respective filter. In the context of the present invention, sterile filtration is a process used to remove microorganisms such as bacteria and vi- ruses from liquids or gases, ensuring that the filtered medium is essentially free from microorganism contaminants, for example at a level which meets the requirements for sterility using standard testing methods as specified in the current edition of a default standard, such as the European Pharmacopeia or the United States Pharmacopeia, and which is suitable for sterile applications .

Preferably, in the context of the present invention, purging means wetting the filter, for example by rinsing.

Visual inspections on a visual inspection apparatus, as de- scribed in more detail below, have shown that the filtered albu- min solutions thus obtained are clear and have an opalescence which is reduced by at least one class compared to a filtrate obtained without the conditioned filter described herein. In particular, the filtered albumin solutions thus obtained are clear and have an opalescence of at most class I, classified with 3 nephelometric turbidity units (NTU).

In order that the present invention can be better understood, certain terms are first defined. Further definitions can be found in the detailed description.

In the context of the present invention, the term "albumin-con- taining aqueous solution" refers to any composition comprising water and at least albumin as the major protein component. More particularly, the term "albumin-containing aqueous solution" re- fers to any aqueous composition in which albumin comprises more than 80% of all proteins in the composition and which is in- tended for clinical treatment. Albumin-containing aqueous solu- tions as used in the context of the present invention are de- fined in particular in the European Pharmacopoeia under the ti- tle "human albumin solution".

The human albumin which is the subject of the present invention is obtained in particular by extraction and purification by any suitable method from a human albumin source or also by culturing animal or plant cell cultures, bacteria or yeasts which have been transformed to produce human albumin by genetic engineering methods as known to the skilled person.

The at least one filter used in the method described herein is a sterile filter. The sterile filter can be a double-layered ster- ile filter equipped with two membranes with the same or differ- ent pore sizes. In the context of the present invention, a ster- ile filter is a filter for removing microorganisms such as bac- teria and viruses from liquids or gases, ensuring that the fil- tered medium remains essentially free from microbial contami- nants for example, at a level which meets the requirements of sterility measured using standard testing methods as specified in the current edition of a default standard, such as the Euro- pean Pharmacopeia or the United States Pharmacopeia, and which is suitable for sterile applications. Other terms for 'sterile filter' include 'sterilising filter'.

By "prior to filling the filtrate into at least one container" is meant that only the filtrate obtained by filtering an albu- min-containing aqueous solution through a conditioned filter is transferred into the one or more containers. In other words, ac- cording to the invention, the filling process is started only after the conditioning of the at least one filter described herein, so that the filtrate in all containers has the advanta- geous properties with respect to long-term stability and, in particular, reduced opalescence.

"NTU" stands for "Nephelometric Turbidity Unit" and is the unit used to measure the turbidity of a liquid or the presence of suspended solids in a liquid. The higher the concentration of suspended solids in the liquid, the higher the turbidity. Stabi- lised formazine turbidity standards for opalescence are availa- ble, for example, under the trade name StablCal® from Hach Lange GmbH (Switzerland). The relationship between NTU and suspended solids is as follows: 1 mg/1 (ppm) corresponds to 3 NTU. The turbidity is measured by 90° light scattering.

The inventors of the present invention have found that filtra- tion prior to pasteurisation has a decisive influence on the particle formation leading to opalescence during and/or after pasteurisation, as described in Example 1 below. This is sur- prising as filtration studies carried out by the inventors con- firmed that particles responsible for opalescence are in the size range between 450 nm and 800 nm. However, such particles should have been separated during filtration by the sterile fil- ters used. Even more surprising was the finding that particles in the range between 450 nm and 800 nm were detected in both opalescent and non-opalescent samples. It has now been surpris- ingly found that it is a combination of filtration and pasteuri- sation that causes opalescence, wherein the aggregates responsi- ble for opalescence form after filtration during and/or after pasteurisation. By conditioning the at least one sterile filter with albumin prior to filling the filtrate into the final con- tainer (s), e.g. vials, the number and size of said particles can be significantly reduced. As a result, it was found that the oc- currence of opalescence in the filtered albumin solutions (here- inafter also referred to simply as "filtrate") is inhibited or at least reduced. In addition, the formation of the white de- posit, which can be observed after some time at the liquid-air interface, can also be avoided by the method disclosed herein, which means that the packaging - and thus also the production - of filtered albumin solutions can be carried out more effi- ciently by the method disclosed herein, as fewer vials have to be sorted out in quality control.

In embodiments, the at least one filter is conditioned by purg- ing the at least one filter with a quantity of the albumin-con- taining aqueous solution by means of pressure filtration and maintaining the filtrate in the outlet of the respective filter at a back pressure of about 1 bar. At this back pressure, a rel- atively rapid and complete saturation of the filter with albumin can be achieved without subjecting the filter and/or the fil- trate to excessive stress which could damage the filter or ren- der the filtered albumin solution, i.e. the product (filtrate), unusable.

In other embodiments, the at least one filter is conditioned by purging the at least one filter with a quantity of the albumin- containing aqueous solution by pressure filtration with an inlet pressure at the inlet of the respective filter of about 2 bar and maintaining the filtrate at the outlet of the respective filter with a back pressure of about 1 bar.

In embodiments, the quantity of the albumin-containing aqueous solution used for purging the at least one filter is between about 2.5 litres and about 11 litres per square metre of filter area of the at least one filter. In certain embodiments, the quantity of albumin-containing aqueous solution used to purge the at least one filter is about 10 litres for a filter area of the at least one filter of 1 square metre. Such a volume ensures good purging and saturation of the respective filter with albu- min.

In embodiments, the filtrate is recirculated after passing through the at least one filter. In this way, the consumption of albumin-containing aqueous solution can be reduced and the effi- ciency of the method can be further increased.

In embodiments in which the at least one filter is conditioned by flushing the at least one filter with an quantity of the al- bumin-containing aqueous solution at a differential pressure across the respective filter of at most 1 bar, no back pressure is applied at the outlet of the respective filter. Under these conditions, it was surprisingly found that the albumin solution of the filtrate has a particularly low opalescence, e.g. an opalescence class I classified as 3 NTU, if any.

In embodiments in which the at least one filter is conditioned by purging the at least one filter with an quantity of the albu- min-containing aqueous solution at a differential pressure across the respective filter of at most 1 bar, the purging of the at least one filter with the quantity of the albumin-con- taining aqueous solution is carried out at a flow rate through the respective filter of about 2 to about 4 litres per minute per square metre of filter area of the at least one filter. Preferably, the at least one filter is purged with the quantity of albumin-containing aqueous solution at a flow rate through the respective filter of approximately 3 litres per minute per square metre of filter surface area of the at least one filter. In embodiments, the at least one filter is rinsed with ultrapure water (water for infusion) before the respective filter is purged with the quantity of albumin-containing aqueous solution. Preferably, the volume of ultrapure water used to rinse the at least one filter is about 1 litre to about 11 litres per square metre of filter area of the at least one filter. Preferably, the volume of ultrapure water used to rinse the at least one filter is about 1 to 10 litres per square metre of filter area of the at least one filter. Such a quantity ensures good rinsing of the respective filters.

In embodiments, the method further comprises a step of pasteur- ising the filtrate. In particular, the pasteurisation of the filtrate is carried out at 60 °C for 10 h after filtration. Un- der these conditions, germs and viruses that have entered or are present in the filtered albumin solution despite upstream safety precautions are reliably inactivated and rendered harmless. Preferably, the filtrate is pasteurised in the at least one con- tainer. This eliminates the risk of recontamination of the fil- tered albumin solution by transferring it again.

In embodiments, the filtrate is successively filled into sterile vials, which are then sealed with a stopper.

In embodiments, the method further comprises a step of pasteur- ising the filtrate as described herein and a step of incubating the pasteurised filtrate to obtain an incubated filtrate. In particular, the step of incubating the pasteurised filtrate is carried out over a period of about 10 to 20 days, preferably for about 15 to 20 days. Preferably, the step of incubating the pas- teurised filtrate is carried out for 15 or 16 days. The incu- bated filtrate can be stored. The storage may, for example, be for a predetermined period of time. In one embodiment, the step of incubating the pasteurised fil- trate is carried out at a temperature between about 25 and about

35 °C, preferably at a temperature between about 29 and about 32 °C .

In embodiments in which the at least one filter is conditioned by purging the at least one filter with a quantity of the albu- min-containing aqueous solution by means of pressure filtration and the filtrate in the outlet of the respective filter is main- tained at a back pressure of at least 0.5 bar, the back pressure is monitored at least during the conditioning of the respective filter.

In embodiments in which the at least one filter is conditioned by purging the at least one filter with a quantity of the albu- min-containing aqueous solution by means of pressure filtration and the filtrate in the outlet of the respective filter is main- tained at a back pressure of at least 0.5 bar, the purging of the respective filter with the quantity of the albumin-contain- ing aqueous solution is carried out at a flow rate through the respective filter of about 1 litre to about 3 litres per minute per square metre of filter area of the at least one filter.

Preferably, the respective filter is purged with the quantity of albumin-containing aqueous solution at a flow rate through the respective filter of approximately 2 litres per minute per square metre of filter surface area of the at least one filter.

In embodiments, the albumin is selected from human serum albu- min, bovine serum albumin, egg albumin and recombinant human se- rum albumin. In embodiments, the albumin concentration in the albumin-con- taining aqueous solution and/or filtrate is between 19 % and 30 %, based on the total weight of the albumin-containing aqueous solution and/or filtrate, respectively. This is done prior to and/or after filtration. Alternatively, the albumin concentra- tion in the albumin-containing aqueous solution and/or in the filtrate before and/or after filtration is between 3 % and 6 %, based on the total weight of the albumin-containing aqueous so- lution and/or filtrate, respectively. Again, this is done prior to and/or after filtration.

In embodiments, the method further comprises a step of pasteur- ising the filtrate to obtain a pasteurised filtrate and a step of evaluating the opalescence of the pasteurised filtrate. Pref- erably, the opalescence of the filtrate is evaluated by measur- ing turbidity or evaluating a change in the formation of higher molecular weight particles.

The term "higher molecular weight particles" refers to an asso- ciation of at least two molecules. In certain embodiments, the molecules are lipids and/or proteins, e.g. albumins, wherein the protein association results in the formation of higher order ag- gregates of monomeric protein. The association can result from non-covalent (e.g. electrostatic, van der Waals) protein-protein interactions. The proteins can be identical or different. The higher molecular weight particles typically have a molecular weight of about 10⁴ Da or higher, typically about 10⁶ Da or higher. The weight average molecular weight of the aggregated molecules in the solution can be detected, for example, by one or more of the following methods: Light scattering techniques such as static and/or dynamic light scattering, or asymmetric flow field flux fractionation. Details of the above measurement methods are described in more detail below. In addition to the step of evaluating the opalescence of the pasteurised filtrate, the opalescence of the samples after pas- teurisation and incubation can be determined. Accordingly, in embodiments, the method further comprises a step of pasteurising the filtrate, in particular as described herein, to obtain a pasteurised filtrate, a step of incubating the pasteurised fil- trate to obtain an incubated filtrate, and a step of evaluating the opalescence of the incubated filtrate.

In particular, the opalescence may be evaluated immediately af- ter pasteurisation and/or incubation. However, the step of eval- uating the opalescence of the pasteurised and optionally incu- bated filtrate can also be carried out after the corresponding filtrate has been stored for about one day to several years, in particular for about five years. Preferably, the step of evalu- ating the opalescence of the pasteurised and optionally incu- bated filtrate is carried out after the corresponding filtrate has been stored for about 1 to 6 months, preferably after the corresponding filtrate has been stored for about 1 to 3 months.

The opalescence of the incubated filtrate can be assessed at the temperature at which the incubation is carried out, in particu- lar at a temperature of about 29 °C to about 31 °C. Alterna- tively or additionally, the opalescence can also be evaluated at the temperature at which the incubated filtrate is stored. This temperature may depend on the product in question and may, for example, be between about 2°C and about 8°C or ambient tempera- ture. It is also conceivable that the opalescence of the incu- bated filtrate is evaluated neither at the incubation tempera- ture nor at the storage temperature, but at room temperature in a laboratory. The term "room temperature" is understood by the skilled person and can in particular mean a temperature of about 20 °C.

Accordingly, in embodiments in which the method further com- prises a step of evaluating the opalescence of the incubated filtrate, the opalescence of the incubated filtrate is evaluated at a temperature of 3≤5 °C. Preferably, the opalescence of the incubated filtrate is evaluated at a temperature of 3≤0 °C, more preferably at a temperature of 2≤5 °C.

In certain cases, the opalescence of the incubated filtrate is evaluated at a temperature of approximately 20 °C.

In embodiments, the opalescence of the incubated filtrate is evaluated at a temperature between about 2 °C and about 8 °C.

In embodiments, the opalescence of the filtrate and/or the incu- bated filtrate is assessed by one or more of the following meth- ods: visual inspection, dynamic light scattering and tunable re- sistive pulse sensing (TRPS).

In embodiments in which the method further comprises a step of evaluating the opalescence of the filtrate and/or the incubated filtrate, the filtrate and/or the incubated filtrate has/have an opalescence that is reduced by at least one class when compared to a filtrate of the same albumin-containing aqueous solution that has not been filtered through the conditioned filter. Pref- erably, the pasteurised filtrate and/or the incubated filtrate has/have at most a class I opalescence, classified as 3 NTU. Even more preferably, the filtrate and/or the incubated fil- trate, when assessed 2 months after filtration, exhibits at most Class I opalescence, classified as 3 NTU. In embodiments, the albumin-containing aqueous solution is ap- plied to the at least one filter at a pressure of approximately 1 bar. At this pressure, a good throughput can be achieved with- out overloading the filter or the albumin-containing aqueous so- lution.

In embodiments, the at least one filter has a filter area of about 0.01 to about 0.8 m² . Preferably, the at least one filter has a filter area of about 0.4 m² . With this filter surface area, a relatively large volume, as occurs in the production of fil- tered albumin solutions in the pharmaceutical industry, can be filtered in an acceptable time.

Preferably, the pore size of the at least one filter is approxi- mately 0.22 pm or less. If the at least one filter is a double- layer filter equipped with two membranes, the pore size of the membrane arranged first in the flow direction of the filter can have a larger pore size than the membrane arranged downstream in the flow direction, i.e. the membrane of the sterile filter. In particular, the first of the two membranes arranged in the flow direction of the double-layer filter may have a pore size of about 0.45 pm or smaller, and the pore size of the other mem- brane arranged downstream in the flow direction of the double- layer filter, i.e. the membrane of the sterile filter, may be about 0.22 pm or smaller. Filters or membranes with such pore sizes are generally readily available on the market. In particu- lar, the pore size of the filter can be 0.22 pm or 0.20 pm. Mem- branes with such a pore size enable the desired sterilisation effect without excessively restricting the flow through the fil- ter. In embodiments, the at least one filter is a polyethersulfone (PES) membrane filter or a nylon filter. Such filters are suita- ble for use in the pharmaceutical industry and are generally readily available.

In embodiments, prior to filtering the albumin-containing aque- ous solution with the at least one filter, the method further comprises a step of filtering the albumin-containing aqueous so- lution with a pre-filter, wherein the pore size of the pre-fil- ter is about 0.45 pm or smaller.

The amount of stabiliser added to the albumin-containing aqueous solution is proportional to the albumin concentration of the so- lution in question. In embodiments, the albumin-containing aque- ous solution is stabilised with about 0.08 mM of at least one stabiliser selected from sodium N-acetyltryptophanate and sodium caprylate per gram of albumin. This amount can effectively pro- tect the albumin molecules from heat denaturation during pas- teurisation and reduce the formation of aggregates to achieve stability over a desired period of time, in particular over the shelf life of the final albumin product, e.g. for at least 1 month, preferably for at least 3 months, more preferably for at least 6 months, most preferably for one year or longer.

In a further aspect, the present invention relates to a filtered albumin solution obtained by the method of sterile filtration disclosed herein, in particular a pharmaceutical composition comprising the filtered albumin solution. Such an albumin solu- tion is characterised by lower opalescence compared to a compa- rable albumin solution which has been filtered through one or more filters, but which has not been conditioned as described herein. In a further aspect, the present invention relates to the use of a filter which has been conditioned by purging with an quantity of an albumin-containing aqueous solution either at a differen- tial pressure across the filter of at most 1 bar or by pressure filtration and holding the filtrate in the outlet of the filter at a back pressure of at least 0.5 bar, for reducing the opales- cence of the filtered albumin-containing aqueous solution.

Some embodiments of the present invention are described in more detail with reference to the accompanying figures, wherein iden- tical reference signs are used to indicate identical or corre- sponding elements. The various views and illustrations of the embodiments shown in the figures are schematic representations:

Figure 1: Schematic illustration of a visual inspection box used to test for opalescence;

Figure 2: Schematic illustration of the sterile filtration set-up;

Figure 3: Summary of the results of the visual inspection for examples 1-3;

Figure 4: Image showing the opalescence of the vials from examples 1-3;

Figure 5: Image showing white deposit rings of the vials from examples 1-3;

Figure 6: Summary of the optical density results for exam- pies 1-3;

Figure 7: Summary of the turbidity results for examples 1-3; Figure 8: Schematic illustration of filtration with applied back pressure;

Figure 9: Schematic illustration of filtration with reduced pressure;

Instrumental methods

A visual control box 20 with black inner walls 21 and a white light source 22 mounted in the upper part of the box was used to examine samples for opalescence and to determine the degree of opalescence. The sample containers 10, 10' were placed in front of a slit 23 in the box so that the containers 10, 10' were ex- posed to indirect light, as shown schematically in Figure 1. The opalescence was classified according to the European Pharmaco- poeia 11.2, 2023, "2.2.1 Clarity and degree of opalescence of liquids" in class 0 - IV (corresponding to 0, 3, 6, 18 and 30 NTU, respectively).

Dynamic light scattering (DLS) was measured with a Malvern Zetasizer nano zs to determine the particle size distribution (non-quantitative) in the size range of 0.4 - 7000 nm, the mean Z-average (Z-Ave) and the optical density (mean derived count rate (DCR), in kcps). The Z-Ave is the harmonic mean of the in- tensity-weighted hydrodynamic diameter, which was previously used as a reference value for the evaluation of albumin opales- cence. The mean derived count rate (DCR) gives an indication of the number of particles. Samples of opalescent and clear fil- tered albumin solutions were measured undiluted in forward (12.8°) and backward (173°) scattering mode at 20 °C. The turbidity was determined by 90° light scattering. The re- sults are given in nephelometric turbidity units (NTUs).

Generation of opalescent starting material

In order to provide a starting material under controlled condi- tions that leads to pronounced opalescence and the formation of white deposit rings after the sterile filtration known from prior art, the 20 % albumin diafiltrate prepared from the Kistler/Nitschmann precipitate C was concentrated and formulated with the stabilisers tryptophan and caprylic acid (20 mM each) to albumin 25 %. This bulk albumin solution was aerated by blow- ing compressed air directly into the solution with a nozzle, causing small air bubbles to form. The flow rate was adjusted so that bubbles formed continuously. To avoid microbial contamina- tion, a 0.22 pm filter was connected to the inlet of the con- tainer with the bulk albumin solution. Aeration was carried out over a period of 6 hours. The starting material obtained in this way was used as an albumin-containing aqueous solution in the comparative experiments.

Example 1

The reference samples were prepared by filtering the starting material as described above without conditioning the filter. A schematic representation of the sterile filtration set-up is shown in Figure 2. The albumin-containing aqueous solution 2 was filtered using a pump (not shown), which provided a pressure of

2 bar without applying a back pressure to the outlet 4 of the filter 3. The filter 3 used was a disposable cartridge with a pore size of 0.2 pm and a sterile filter area of 0.015 m² , which contained the same material and the same pleated filter membrane as the 10-inch filters normally used on a production scale. The filter 3 was first rinsed with 260 mL ultrapure water (not shown) and then purged with 150 mL albumin-containing aqueous solution 2. The filtrate 1 of the albumin solution was succes- sively collected in 2050 mL vials 10, which were immediately sealed with a stopper (not shown). The vials 10 were then pas- teurised at 60 °C for 10 hours and incubated at 30 °C for 15 days. The resulting 20 samples served as a reference to which the samples obtained by the improved method according to the present invention were compared.

Visual inspection of the reference sample vials in the visual inspection box confirmed opalescence (Class ITT) for each vial after pasteurisation and incubation, as shown in the results summarised in Table 1 and Figure 3, wherein statistical signifi- cance was calculated using a two-tailed t-test with Sigma Plot 12.0 software. An example of a vial with opalescence (class ITT) is also shown in Figure 4 (middle image; Ex. 1).

Moreover, the vials containing the reference samples were visu- ally assessed for white rings on the inside of the vials at the liquid-air interface two months after their incubation, and in- deed white rings were observed in the reference sample vials, as indicated by the white arrow in Figure 5 (centre image; Ex. 1).

The optical density measured by dynamic light scattering (DLS) was highest for the reference samples, as shown in Figure 6. This indicates a comparatively large particle size and a high particle content. Accordingly, the reference samples were also the most turbid, as Figure 7 shows. Again, the statistical sig- nificances in Figures 6 and 7 were calculated using the two- sided t-test of the Sigma Plot 12.0 software. Example 2

In one embodiment of the improved method described here for re- ducing the opalescence of filtered albumin solutions, 20 samples were prepared as described in Example 1, with the only differ- ence being that a back pressure of 1 bar was applied to the fil- ter during rinsing with water and albumin, wherein the direction of flow through the filter 3 is indicated by arrows, as shown schematically in Figure 8. The circled numbers indicate the pressure values before and after filter 3 in bar.

Visual inspection of the vials obtained by saturating the filter with back pressure during water and albumin rinsing revealed that these samples showed a reduction in opalescence from class III (reference) to class (II), classified according to the Euro- pean Pharmacopoeia 2016 (EP9, section 2.2.1: Clarity and degree of opalescence in liquids), as shown in the results summarised in Table 1 and Figure 3, wherein statistical significance was calculated using a two-tailed t-test with Sigma Plot 12.0 soft- ware. An exemplary vial with opalescence (Class II) is also shown in Figure 4 (left image; Ex. 2). None of the vials from example 2 showed white rings on visual inspection two months af- ter incubation, as shown in Figure 5 (left image; example 2).

Furthermore, saturation of the sterile filter by back pressure during rinsing also reduced the optical density and turbidity compared to the reference samples, as shown in Figures 6 and 7. Again, statistical significance was calculated using a two- tailed t-test with Sigma Plot 12.0 software. Example 3

In another embodiment of the improved method described herein for reducing the opalescence of filtered albumin solutions, 20 samples were obtained as described in Example 1, wherein the only difference this time consisted of rinsing with the albumin- containing aqueous solution at a reduced pressure of 1 bar (in- stead of 2 bar), as shown schematically in Figure 9, wherein the flow direction through the filter 3 is indicated by arrows and the pressure values before and after the filter 3 in bar, re- spectively, are indicated by circled numbers.

Visual inspection of vials obtained by saturating the filter with reduced pressure or slower albumin rinsing showed that these samples exhibited a reduction in opalescence from Class III (reference) to Class I or 0 (with 5 exceptions), as shown in the results summarised in Table 1 and Figure 3, wherein statis- tical significance was calculated using a two-tailed t-test with Sigma Plot 12.0 software. An exemplary vial showing only Class I opalescence, which does not correspond to opalescence according to the European Pharmacopoeia 2.2.1 "Clarity and degree of opal- escence of liquids", is also shown in Figure 4 (right-hand im- age; Ex. 3). As with the sample containers from example 2, none of the vials from example 3 showed white rings on visual inspec- tion two months after incubation, as shown in Figure 5 (right- hand image; example 3).

Saturation of the sterile filter using reduced pressure, i.e. with a slower flow rate during rinsing, also led to a reduction in optical density and turbidity compared to the reference sam- ples, as shown in Figures 6 and 7. Again, statistical signifi- cance was calculated using a two-tailed t-test with Sigma Plot 12.0 software. Table 1

Claims

1. A method of sterile filtration of an albumin-containing aqueous solution (2), wherein the method comprises filter- ing the solution (2) with at least one filter (3) to obtain a filtrate (1), characterised in that, prior to filling the filtrate (1) in at least one container (10, 10'), the at least one filter (3) is conditioned by purging the at least one filter (3) with a quantity (5) of the albumin-contain- ing aqueous solution (2) i) at a differential pressure across the respective filter (3) of no more than 1 bar; and/or ii) by means of pressure filtration, wherein the filtrate (1) is maintained at a back pressure of at least 0.5 bar at the outlet (4) of the respective filter (3).

2. Method according to claim 1, wherein the back pressure at the outlet (4) of the respective filter (3) is about 1 bar.

3. The method according to claim 1 or 2, wherein the quantity

(5) of the albumin-containing aqueous solution (2) used for purging the at least one filter (3) is between about 2.5 and about 11 litres per square metre of filter area of the at least one filter (3), preferably approximately 10 litres per square metre of filter area of the at least one filter

(3).

4. Method according to one of the preceding claims, character- ised in that the at least one filter (3) is rinsed with ul- trapure water (6) before the respective filter (3) is purged with the quantity (5) of the albumin-containing aqueous so- lution (2).

5. The method according to claim 4, wherein the volume of ul- trapure water (6) used to rinse the at least one filter (3) is about 1 to 10 litres per square metre of filter area of the at least one filter (3).

6. The method according to any one of the preceding claims, wherein prior to and/or after filtration, the albumin con- centration in the albumin-containing aqueous solution (2) and/or in the filtrate (1) is between 19 % and 30 %, based on the total weight of the albumin-containing aqueous solu- tion (2) and/or the filtrate (1), respectively.

7. The method according to any one of claims 1 to 5, wherein prior to and/or after filtration, the albumin concentration in the albumin-containing aqueous solution (2) and/or in the filtrate (1) is between 3 % and 6 %, based on the total weight of the albumin-containing aqueous solution (2) and/or the filtrate (1), respectively.

8. The method according to any one of the preceding claims, wherein the method further comprises a step of pasteurising the filtrate (1) to obtain a pasteurised filtrate (7) and a step of evaluating the opalescence of the pasteurised fil- trate (7).

9. The method according to any one of the preceding claims, wherein the method further comprises a step of pasteurising the filtrate (1) to obtain a pasteurised filtrate (7), a step of incubating the pasteurised filtrate (7) to obtain an incubated filtrate (8), and a step of evaluating the opalescence of the incubated filtrate (8).

10. The method according to claim 8 or 9, wherein the opales- cence of the pasteurised filtrate (7) or of the incubated filtrate (8) is evaluated at a temperature of 3≤5 °C, pref- erably at a temperature of 3≤0 °C, even more preferably at a temperature of 2≤5 °C.

11. The method according to claim 10, wherein the opalescence of the incubated filtrate (8) is evaluated at a temperature of about 20 °C.

12. The method according to claim 10, wherein the opalescence of the incubated filtrate (8) is evaluated at a temperature of between about 2 to 8 °C.

13. The method according to any one of claims 8 to 12, wherein the opalescence of the pasteurised filtrate (7) and/or of the incubated filtrate (8) is evaluated by one or more of visual inspection, dynamic light scattering, and tunable resistive pulse sensing.

14. The method according to any one of claims 8 to 13, wherein the pasteurised filtrate (7) and/or the incubated filtrate (8) has/have an opalescence which is reduced by at least one class when compared to a filtrate of the same albumin- containing aqueous solution (2) that has not been filtered through the conditioned filter, preferably the pasteurised filtrate (7) and/or the incubated filtrate (8) has/have at most class I opalescence, classified with 3 NTU.

15. Method according to one of the preceding claims, character- ised in that the albumin-containing aqueous solution (2) is applied to the at least one filter (3) at a pressure of about 1 bar.

16. Method according to one of the preceding claims, wherein the at least one filter (3) has a filter area of about 0.01 to about 0.8 m² , preferably the at least one filter (3) has a filter area of about 0.4 m² .

17. The method according to any one of the preceding claims, wherein the pore size of the at least one filter (3) is about 0.22 pm or smaller, preferably about 0.22 pm or about 0.20 pm .

18. The method according to any one of the preceding claims, wherein prior to filtering the albumin-containing aqueous solution (2) with the at least one filter (3), the method further comprises a step of filtering the albumin-contain- ing aqueous solution (2) with a pre-filter (9), wherein the pore size of the pre-filter (9) is about 0.45 pm or smaller.