Abstract
This paper describes an approach of extractables determination and gives information on extractables profiles for gamma-sterilized single-use bags with polyethylene inner contact surfaces from five different suppliers. Four extraction solvents were chosen to capture a broad spectrum of extractables. An 80% ethanol extraction was used to extract compounds that represent the bag resin and the organic additives used to stabilize or process the polymer films which would not normally be water-soluble. Extractions with1 M HCl extract, 1 M NaOH extract, and 1% polysorbate 80 were used to bracket potential leachables in biopharmaceutical process fluids. The objective of this study was to obtain extractables data from different bags under identical test conditions. All the bags had a nominal capacity of 5 L, were gamma-irradiated prior to testing, and were tested without modification except that connectors, if any, were removed prior to filling. They were extracted at 40 °C for 30 days. Extractables from all bag extracts were identified and the concentration estimated using headspace gas chromatography–mass spectrometry and flame ionization detection for volatile compounds and for semi-volatile compounds, and liquid chromatography–mass spectrometry for targeted compounds. Metals and other elements were detected and quantified by inductively coupled plasma mass spectrometry analysis. The results showed a variety of extractables, some of which are not related to the inner polyethylene contact layer. Detected organic compounds included oligomers from polyolefins, additives and their degradation products, and oligomers from the fill tubing. The concentrations of extractables were in the range of parts-per-billion to parts-per-million per bag under the applied extraction conditions. Toxicological effects of the extractables are not addressed in this paper.
LAY ABSTRACT: Extractables and leachables characterization supports the validation and the use of single-use bags in the biopharmaceutical manufacturing process. This paper describes an approach for the identification and quantification of extractable substances for five commercially available single-use bags from different suppliers under identical analytical conditions. Four test formulations were used for the extraction, and extractables were analyzed with appropriately qualified analytical techniques, allowing for the detection of a broad range of released chemical compounds. Polymer additives such as antioxidants and processing aids and their degradation products were found to be the source of most of the extracted compounds. The concentration of extractables ranged from parts-per-billion to parts-per-million under the applied extraction conditions.
Introduction
Single-use bags are designed for the preparation, storage, mixing, freezing, transportation, formulation, and filling of biopharmaceutical solutions. Media and solutions used in biopharmaceutical formulations can include simple/common buffers with broad pH range, mixtures of aqueous-based organic solvents, highly complex cell culture media, buffers containing proteins, monoclonal antibodies with stabilization agents such as Tween® or Triton®, other drug substances, and so forth.
The mechanical suitability and biological safety of bags are generally assessed by the supplier with physical, mechanical, chemical, and biocompatibility tests, for example, USP <661>, <87>, and <88> testing (1).
No universal standard for extractable testing exists. The Bio-Process Systems Alliance (BPSA) (2) has recommended a program for the qualification of product contact manufacturing materials with regard to extractables and leachables. The Parenteral Drug Association (PDA) (3) and the Product Quality Research Institute (PQRI) (4) have made substantial progress on programs to qualify final containers and closures, as well as metered dose inhalers, for extractables and leachables. Recommendations from the European Medicines Agency (EMA) (5) (guideline on plastic immediate packaging material) and U.S. Food and Drug Administration (FDA) (6) (container closure systems for packaging human drugs and proteins) focus on the final container and not the processing materials. Furthermore, regulatory guidance that addresses product contact equipment (7, 8) does not specifically include discussion of extractables or leachables. A rigorous extractables test program will include evaluation of the physical parameters that influence the outcome, such as temperature, contact time, surface area, and extractions volume. Then, a sensible combination of analytical instrumentation and methods should allow for the detection, identification, and quantitative estimation of most of the extractable compounds (and their degradation products) from single-use bag materials. Without consistent regulatory guidance, single-use bag suppliers provide extractables data using a variety of extraction conditions with a varying level scientific rigor.
What is commonly called plastic material is typically composed of polymers and additives. The manufacturing of polymeric films is carried out by extruding the raw materials. To allow a robust transformation of the resin granulates in films and to aid further manufacturing to finished products, additives are needed to adjust the characteristics of the resin (9). These additives have the purpose to ensure usability of the material during the different manufacturing steps as well as to ensure performance of the film structure as such. Indeed, very few polymers can be used without additives.
In particular, in medical applications most polymers require the use of additives in order to achieve the targeted properties for a given application: polyvinylchloride (PVC) is softened by plasticizers to achieve flexibility, antioxidants are added to polypropylene to withstand sterilization by gamma irradiation, antistatic agents might be added to polyethylene (PE) bags used for powder containment and transfer (10).
The most widely used additives in film applications are antioxidants, slipping agents, and antiblocking agents. Antioxidants are organic compounds that slow down the oxidative degradation of polymers. Primary antioxidants protect the plastic in the final application and give the final product resistance to aging so that it maintains its properties throughout its shelf life. Secondary antioxidants are needed during the processing of plastics.
Slipping and anti-blocking agents are added to the polymer during the extrusion process. They are very often used together in order to improve processing behaviour and the end-performance of polymers (10).
Many additives will degrade into by-products. This can occur during the resin-to-film extrusion step, during the film-to-bag transformation steps or during gamma irradiation. These additives and by-products are very often identified and reported in extractable studies. Additives are mostly non-volatile, whereas their by-products might be non-volatile or semi-volatile or, even volatile if the degradation pathway leads to small molecules.
However, the additives and their degradation products are not the only source of potential extractable substances. Indeed, during the polymerization process itself chemical substances such as polymerization initiators and solvents are added to the monomers to control the polymerization process. These compounds may also be found at trace level in the extractable analysis.
Films for single-use bags are made of multiple layers that may be made of polyethylene (PE), ethylene vinyl alcohol (EVOH), polyethylene terephthalate (PET), polyamide (PA), or ethylene vinyl acetate (EVA). In addition to the bag itself, tubing for filling—which may be made of a thermal plastic elastomer (TPE), EVA, silicone, or even PVC—is either welded directly to the bag or connected to a port welded on the bag. Connectors welded to the bag are mainly made of polyolefins.
The gamma sterilization of single-use bag systems changes the nature of the extractables profile by initiating chemical reactions leading to either an increase or a decrease in molecular weight of polymers (11, 12), by the formation of oxygen containing organic molecules such as alcohols, aldehydes, ketones, or acids (13, 14), and by the oxidation of the organic additives. Accordingly, the time between gamma irradiation of the bag and the extraction and analysis can also change the extractables profile (15, 16).
As a result of the described complexity, a huge variety of potential extractable compounds can be expected. This calls for the application of test formulations with different chemical characteristics and different analytical tools to obtain extractables data for the evaluation of single-use equipment.
Rationale for the Extraction Media
During biopharmaceutical manufacturing, single-use bags are typically in contact with aqueous-based media, which may contain varying levels of organic compounds. Within the scope of this study, a reasonable worst-case approach was chosen to bracket the typical biopharmaceutical process solutions using four media, 80% ethanol, 1 M HCl, 1M NaOH, and 1% polysorbate 80.
Ethanol 80% represents a worst-case organic content of a biopharmaceutical formulation. It will preferentially extract the organic additives and their degradation products compared to more aqueous extraction media and therefore may give an indication of the additive package that was used during the manufacture of the bag. It is expected that number and concentration of extractables detected in an 80% ethanol extract will be exaggerated compared to a leachable simulation under normal worst-case operating conditions.
1 M HCl and 1 M NaOH bracket the broad pH range of buffered media, acids and bases used in biopharmaceutical and biotechnology processes and therefore will show the influence of extreme pH on the extractable profile. Low pH leads to the extraction of compounds such as organic acids and phenols. High pH promotes the dissolution of basic extractables such as amides.
Polysorbate 80 is a surfactant commonly used to increase the solubility of many proteins and monoclonal antibodies in biopharmaceutcial formulations. Polysorbate 80, like most surfactants, has both hydrophilic and hydrophobic moieties that can solubilize organic extractables that would not normally be detected in an aqueous-based formulation. It also lowers the surface tension of water, allowing the bag surface to be more easily wet.
Materials and Methods
Materials
All the bags were nominally 5 L in volume and were procured from the manufacturers or from laboratory catalogs. All the bags came with tubing attached to one or more ports on the bag. They were tested without modification except that connectors, if any, were removed prior to testing. A detailed description of the film layers is given in Table I.
Film Description of the Single-Use Bag Systems
Preparation of Extraction Media
The extraction media were prepared using the following reagents: 200 proof ethanol (Pharmco Aaper), hydrochloric acid (EMD Millipore), sodium hydroxide (J. T. Baker), polysorbate 80 (Mallinckrodt Baker), and high-purity water (Burdick & Jackson).
The 80% ethanol solution was prepared by mixing ethanol and high purity water, with the following proportion 80% (v/v) of ethanol. The 1 N HCl and 1 N NaOH were prepared in high-purity water to achieve solutions with nominal pH values of 0.2 and 13.5, respectively. The pH of the solutions was confirmed after preparation. The 1% (v/v) polysorbate 80 was prepared by adding suitable volume of polysorbate 80 to high-purity water.
Extractions
The 5 L bags were filled via one of the tubings that were prepackaged with the bag. Extractions were accomplished by a static soaking method after filling each of the bags with 2.5 L of the appropriate test formulation. Air was removed and each test bag was then sealed using the tubing clamps.
A sample blank (negative control) was prepared using approximately 0.6 L of each of the four test solvents in sealed glass containers exposed to the same extraction conditions, but without contact with a test article. A second sample blank for the 1 M NaOH extraction was placed in a vessel constructed of polytetrafluoroethylene (PTFE). This sample was used for inductively coupled plasma mass spectrometry (ICP-MS) analysis.
Each extraction was performed for 1 month (30 days) at 40 ± 3 °C in suitable incubators. After the extractions were completed, each test article was inverted three times to ensure mixing and each extraction sample was transferred to glass storage vessels. Samples for ICP-MS analysis were collected immediately in appropriate plastic storage vessels. Samples for headspace analysis were collected immediately in appropriate glass vials. Samples for all other analyses were taken from the corresponding extraction sample and the sample blanks.
Analytical Methods
The analytical techniques were chosen to ensure that the majority of the extractables could be detected. All of the bags have a PE fluid contact layer. PE is hydrophobic and non-polar. Additives used for PE also tend to be hydrophobic and non-polar (17). So analytical methods were chosen appropriately, as shown in Table II. The core orthogonal methods of headspace gas chromatography–mass spectrometry and flame ionization detection (GC-MS/FID), GC-MS/FID, and high-performance liquid chromatography with UV and mass spectrometry detection (HPLC-UV-MS) were used to detected volatiles, semi-volatiles and non-volatiles, respectively. These methods as applied here are valid for organic non-polar molecules. To minimize the possibility of polar molecules not being detected, the NaOH extract was also analzyed by GC-MS/FID after the more polar compounds were derivatized. Identifications were made, when possible, using the MS spectra of detected peaks. Quantitation was estimated using either internal or external standards. Not all given techniques can be applied for each extraction medium due to chemical nature (Table II).
Overview of Analyses for the Extractables Study
HPLC-UV-MS
All samples from each extraction were analyzed by HPLC-UV-MS (Agilent 1100 or 1200 series HPLC system equipped with UV/Vis DAD and an Agilent Single Quad Mass Spectrometer using a C18 Nucleosil column; flow rate 1 mL/min; sample volume 20 μL; column temperature 40 °C). Electrospray ionization (positive and negative mode) with a range of 80–2000 M/z was used. The 1% polysorbate 80 sample was diluted with high-purity water to 0.1% polysorbate 80 to allow the analysis. The mobile phase was an acetonitrile/water gradient. The pH of the 1 M NaOH samples was lowered to approximately pH 5. The pH of the 1 M HCl samples was raised to approximately pH 5.
The polysorbate 80, NaOH, and HCl samples were then prepared by solid phase extraction (SPE) prior to analysis using C18 SPE Columns (3M). The columns were prepared by wetting with methanol and water; 2 mL of each sample was pipetted and drawn through the columns using a mild vacuum. The columns were then eluted with 1 mL of a 50:50 methanol/acetonitrile solution followed by 1 mL of methanol. The resulting eluate was then transferred to an HPLC vial and injected into the HPLC. SPE is an established and proven sample preparation method to reduce analytical interference of sample matrices within the scope of extractables/leachables analysis. In this study no additional recovery studies were performed. The 80% ethanol samples did not require sample preparation and were directly analyzed. New and altered peaks detected in the extraction samples but not detected in the sample blanks represented extractables. If any of the target compounds were detected in the extraction sample, they were quantified based on the external standard. The concentrations reported are for the original extracts.
In addition, certain potential extractables were targeted for detection using the extracted ion mode with the MS detector (Table III). The list includes additives given in the European Pharmacopoeia (18) and which can be expected in organic extracts from bags. Additionally, substances of very high concern were selected, for example, as given in the REACH (Registration Evaluation Authorisation and Restriction of Chemicals) legislation list (19).
Target Extractables for HPLC-UV-MS
GC-MS-FID
Because aqueous samples cannot be directly injected, the samples and controls were extracted into dichloromethane (DCM). Starting sample volumes were 50 mL for the 1 N Na OH, 1 N HCl, and 1% polysorbate 80. Starting sample volume was 10 mL for the 80% ethanol. The 80% ethanol sample was diluted to 40% ethanol with high-purity water to allow for phase separation with the DCM. All samples were first adjusted to pH ≈2 with 5 N HCl made from HCl (EMD Millipore) and high-purity water except for the 1 N HCl sample, which was adjusted using 1 N NaOH made from NaOH (J.T. Baker and high-purity water). The samples were extracted twice with 10 mL of DCM (Burdick and Jackson). The pH of the samples was then adjusted to pH ≈10 and extracted twice more with 10 mL of DCM. The samples were decanted (precipitates, when present, were allowed to settle) then evaporated to a volume of 1 mL thereby achieving concentration factors of 50× for the 1 N NaOH, 1 N HCl, and 1% polysorbate 80 samples, and 10× for the 80% ethanol sample. All samples from each extraction were spiked to achieve a 10 μg/mL concentration of toluene-d8 (the internal standard) to facilitate quantitative estimation of potential extractables. Spiking after the sample preparation for a quantitative estimate assumes that all the detected extractables partitioned into the DCM. This is an assumption that is consistent with the intent of quantitation to be an estimate and not an absolute value.
To increase the detection of the more polar alcohols and carboxylic acids, the 1 N NaOH samples were also derivatized prior to analysis by GC-MS/FID. To 0.5 mL of the sample, after extraction with DCM and concentration, was added 10 μL of Bis(trimethylsilyl)trifluoroacetamide (BTSFA, Supelco). The samples were placed at 30 °C for 1 h and then analyzed.
A GC-MS/FID (Agilent Mass Selective Detector and Flame Ionization Detector; DB-5MS capillary column; injection port temperature: 280 °C; temperature gradient program: initial temp: 30 °C, final temperature: 310 °C; carrier gas: helium; sample injection volume: 1 μL; Splitless). The mass spectrometer was set to an ion mass range of 29 to 650 M/z. The quantitation estimate is performed with the concentrated sample using the peak response in the FID chromatogram compared to the internal standard. New and altered peaks detected in the extraction samples but not detected in the sample blanks represented extractables. The concentrations reported are for the original extracts. Identifications were based on database matches.
Headspace GC-MS-FID Analysis
Samples from the 1 M NaOH, 1 M HCl, and 1% polysorbate 80 extractions analyzed by Headspace GC-MS/FID analysis (Agilent Mass Selective Detector and Flame Ionization Detector; DB-5MS capillary column; injection port temperature: 220 °C; temperature gradient program: initial temperature: 40 °C, final temperature: 280 °C; carrier gas: helium; sample injection volume: 1 mL; split ratio ∼1:20; ion mass range of 28–550 M/z). New and altered peaks detected in the extraction samples but not detected in the sample blanks represented extractables.
The samples were spiked to achieve a 0.1 μg/mL concentration of toluene-d8 (the internal standard). Toluene-d8 has a boiling point of 110 °C, which is approximately midrange of the compounds that are normally detected by headspace GC-MS and is used to give an indication of quantity. However, it is probable that detected extractables with lower vapor pressures are over-represented while those with higher vapor pressure are under-represented.
ICP-MS Analysis
The samples were analyzed by Agilent 7700x Mass Selective Detector or equivalent. Prior to analysis the 80% ethanol samples were diluted 1:10, the 1% polysorbate 80 samples were diluted 1:50 and underwent microwave digestion, and samples of 1 M NaOH were diluted 1:20 to allow the analysis. The 1 M HCl samples were not diluted prior to analysis. The 1% polysorbate 80 sample digestion was performed with an Ethos EZ Microwave Digestion System (Milestone) with an SK-10 high pressure rotor. The samples were diluted with 18 mΩ water to which was added hydrogen peroxide (BDH, UK) and digested in a closed PTFE vessel. Elements analyzed were aluminum, antimony, arsenic, barium, bismuth, boron, cadmium, calcium, chromium, cobalt, copper, iron, lead, lithium, magnesium, manganese, molybdenum, nickel, platinum, potassium, silicon, silver, sodium, strontium, tin, titanium, vanadium, zinc.
The concentration of each detected element was determined using a calibration curve. For the 1M HCl, 1M NaOH, and 1% polysorbate 80 samples, calibration standards were prepared in dilute acid from 1 to 100 ppb. For the 80% ethanol samples, calibration standards were prepared in 8% ethanol/water from 1 to 100 ppb. Results of the extraction samples were compared to the results from the sample blanks. Differences in the detected quantity of an element in the extraction sample when compared to the sample blank were noted. The reported value is for the undiluted samples.
Total Organic Carbon (TOC) Analysis
Samples from the 1 M NaOH and 1 M HCl extractions were analyzed for TOC using an Aurora 1030W TOC system. The NaOH samples and controls were brought to neutral pH by the addition of phosphoric acid (EMD Millipore) because of a limitation in how much chloride could be in a sample. The pH was then lowered to pH = 2 with HCl and purged with nitrogen to remove the inorganic carbon. The 1 N HCl sample was diluted 10× with high-purity water to reduce the chloride concentration. The TOC was then measured using a persulfate method with an inorganic carbon purge after the sample was lowered to approximately pH = 2 with HCl. To ensure that the TOC measurements was accurate, the 1 N NaOH and 1 N HCl sample blanks were spiked with 5 ppm TOC standard and the recoveries were measured to be 98.8% and 98.6%, respectively. The TOC present in each sample blank was subtracted from the TOC in the corresponding bag extraction samples. The difference in TOC values between the sample blanks and extraction samples were calculated.
Results
The extractable results of each bage are summarized by extracting media in the related tables. A detailed discussion regarding the compounds identified follows.
All extracts were analyzed by multiple analytical tools. The analytical detection method for each compound is noted in the tables. Concentration estimates are provided as described in the analytical methods section of this paper. The concentrations reported are for the original extracts.
It is shown in Table IV that given the same extraction conditions, the higher estimated concentrations of organic compounds were found in the ethanol 80% extract, whereas 1 M HCl, 1 M NaOH, and 1% polysorbate 80 extracts contained small amounts of organic compounds. Most of the organic acidic compounds were detected in the 1M NaOH extract, especially after derivatization. Additionally, derivates of phenol were extracted in the caustic medium. It was also seen that several oxygen-containing organic compounds (e.g., butanol, hexanal) were detected in the caustic and acidic media.
Extractable Results: Summary of Ethanol 80%
The results are discussed below in detail. In Tables IV to Table VII, the term bag means single-use bag system including tubing.
Extractable Results: Summary of 1 M NaOH
Extractable Results: Summary of 1 M HCl
Extractable Results: Summary of 1% Polysorbate 80
Caprolactam was present in the 80% ethanol extracts of Bags B, C, and D at a concentration between 1.0 and 3.5 mg/L. A likely source of caprolactam could be polyamide (PA), which is a material used in films and connectors (20). Lower concentrations of caprolactam were also detected in the 1 M HCl extract of Bags A and B and in the 1% polysorbate 80 extract of Bag B.
Stearic acid was identified as an extractable from all tested bags with the highest concentration levels in the 80% ethanol extracts, whereas lower levels of the compound were found in the 1 M NaOH extracts of Bags B, C, D, and E. Stearic acid salts, which can be used in PE as a lubricant (21), were present in the 80% ethanol extracts ranging from 200 μg/L in the Bag C extract to 1.8 mg/L in the Bag E extract.
Octadecyl 3-[3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl]propanoate [European Pharmacopoeia (EP) <3.1.13> additive #11], which is a common phenolic antioxidant for polyolefins (22), was primarily detected in the 80% ethanol extract of Bag A at a concentration of 38.0 mg/L.
Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite (EP <3.1.13> additive #12) which is also a common phenolic antioxidant for polyolefins (23), was detected in the 80% ethanol extracts of Bags A, B, C, and D with a maximum concentration of 140.0 mg/L in the Bag A extract.
In the literature (24) it is described that, like Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite (EP <3.1.13> additive #12) or Octadecyl 3-[3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl]propanoate (EP <3.1.13> additive #11), Methanetetryltetramethyl tetrakis[3-(3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl)propanoate (EP <3.1.13> additive #09) can also be used as an antioxidant for polyolefins. Methanetetryltetramethyl tetrakis[3-(3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl)propanoate was detected on the order of 100 μg/L in the ethanol 80% extract of Bags B and C.
Another identified substance was 3,5-di-t-butyl-4-hydroxyphenylpropionic acid, which is a degradation of Methanetetryltetramethyl tetrakis[3-(3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl)propanoate (25) and was found on the order of 100 μg/L in the ethanol 80% extract of Bag C.
In the 80% ethanol extracts of Bags B, C, and D, 3,9-bis(octadecyloxy)-2,4,8,10-tetraoxa-3,9-diphosphaspiro[5.5]undecane (EP <3.1.13> additive #14) was detected in the range of 100 to 400 μg/L. This compound can be added to polyolefins for its antioxidant and UV stabilizer properties (26).
In all of the extracts, hydrocarbon compounds, mostly branched chains of PE, were detected. Depending on the quality of spectral data available these compounds were either identified as specific compounds (e.g., decane) or simply as hydrocarbons. There are two possible reasons for the observation of hydrocarbons in the extracts. One is that incomplete polymerization of the polyolefins can lead to the formation of oligomers, and the second is that gamma irradiation can lead to fragmentation of the polyolefin chains of the PE material.
Four out of five 80% ethanol bag extracts contained siloxanes, with a maximum concentration of 10 μg/mL in the Bag D extract. Bag D was assembled with silicone tubing and filled with the ethanol 80% medium via this tubing, which is assumed to be the source of the siloxanes. To elicit further credence to this assumption is it noted that siloxane is not detected at a high level in the caustic medium in Bag D, which was filled via TPE tubing. The ethanol 80% extract of Bag B shows the second highest level of siloxanes with 290 μg/L. The supplier information of this bag showed that TPE tubing or silicone tubing is used. Bags A, C, and E were assembled with TPE tubing (27).
In the ethanol 80% extracts of Bags A, B, C, and D 1,3-di-tert-butyl benzene was identified. This substance is known as a degradation product of Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite (15, 28). The concentration range in the extracts is from 1.2 mg/L for Bag D to 3.2 mg/L for Bag A. Low concentration levels were also found in the 1% polysorbate 80 extract of the same bags.
Another degradation product of Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite is 2,4-di-tert-butylphenol (28, 29). This compound was detected in four ethanol 80% extracts (Bags A, B, C, and D) with a concentration range from 89 to 440 μg/L. Lower concentration levels were also found in the 1 M NaOH, 1 M HCl, and 1% polysorbate 80 extract.
Low concentrations (<58 μg/L) of 4-ethoxy-benzoic acid ethyl ester were detected in the 80% ethanol extract of Bag B, C, D, and E. In European Union legislation concerning plastic materials that come into contact with food, this compound is described as an additive or polymer production aid (30).
In all 1 M NaOH and 1 M HCl extracts, oxygen-containing organic compounds were identified (e.g., aldehyde such as hexanal in Bag A, C, and D, and ketone such as 3,3-Di-methyl-2-butanone in all bags). These compound classes can be expected from polyolefin material after gamma irradiation because irradiation leads to oxidation products coming from the polyolefin itself (13).
Acetophenone was identified in the 1 M HCl Bag B solution. According to U.S. Code of Federal Regulations (CFR) 21 § 172.515 it can be used as synthetic flavoring substance or adjuvant.
Butylated hydroxytoluene was only found in the Bag B HCl extract at the concentration of 13 μg/L. It is used as an antioxidant in plastics (31).
2-Butoxyethanol was found in the NaOH extract of Bag B at a concentration of <10 μg/L. It can be used as solvent for synthetic resins (32) or it could be a degradation of polymer after gamma irradiation.
2-(2-Ethoxyethoxy)ethanol was detected in the 1 M NaOH extracts of Bags B, C, and D at approximately 12 μg/L. It is used in the manufacturing of plasticizers (33).
Benzoic acid was found in the NaOH extracts of Bags A, C, and E at concentrations <10 μg/L. It can be used as a plasticizer in plastics (34).
Extracts of all bags showed unidentified compounds. It may not be possible to confidently identify an extractable compound for these following reasons: If an extractable compound is detected at a low concentration then the quality of spectral data may not be sufficient to provide identity information. Excess background noise created by the sample matrix can also affect spectral quality. In addition, there may be isomers of the same compound that may not be readily differentiated. Finally some molecules such as oligomers may have many possible structures and the specific one cannot be determined.
In this extractable study of five bags from different suppliers, six additives mentioned in the EP were detected, whereas no substances on the REACH list were found.
Traces of chlorobenzene were found in the caustic extracts of Bags A, B, and D in the concentration <10 μg/L. From information found in the literature, no specific conclusion regarding the origin of this compound can be given. However, chlorobenzene can be used as an adhesive in food packaging (CFR 21 § 175.105). It is also used in the production of polycarbonate resins (CFR 21 § 177.1580) and polysulfone resins (CFR 21 § 177.1655), which could be used for components.
Table VIII summarizes the trace elements detected in all extraction media identified by ICP-MS.10 BBoron was detected in the ethanol 80% extracts of Bags A, B, and D with a maximum concentration of 200 μg/L. A potential source of boron is boric acid, which according to the literature can be used as stabilizer for food-grade polymers (35, 36).
Summary of the Extracted Metals for Four Different Extraction Media
Results Summary of TOC Analysis
Silicon was primarily found in the ethanol 80% extract of Bag D with a concentration of 9.6 mg/L. As previously mentioned, Bag D was assembled with silicone tubing and filled with the ethanol 80% via this tubing, which is probably the source of silicon.
The ethanol 80% extract of Bags B contained vanadium with a concentration of 28 μg/L. Vanadium can be constituent in alloys.
Aluminum was detected in the caustic extract of Bag B with a concentration of 320 μg/L. Aluminum can be used as catalyst in the synthesis of PE, EVA, and PP (37).
Zinc was found both in the ethanol 80% extracts of Bags B, C, and E and hydrochloric acid extract of Bag E with a concentration up to 10 μg/L. Zinc could originate from zinc stearate or zinc oxide additives that are used, respectively, as a stabilizer and an anti-blocking agent for food-grade polymers (38).
Chromium (5 μg/L) and iron (57 μg/L) were found in the 1% polysorbate 80 extract of Bag A. A potential source of both metals could be contact of the bag material with stainless steel. Additionally, chromium can serve as a catalyst within a chemical complex for polyolefin synthesis (37).
For the other extracted metals, no rational source could be found in the literature.
The TOC reconciliation is subject of ongoing discussion in publications (39). The presented TOC determination is quantitative. For the calculation of carbon recovery, assumptions for the calculation of carbon content of the idenitified substances have to be made. Moreover, the quantitation done by GC-MS/FID and LC-MS is an estimation based on internal standards. GC analysis is limited to molecules smaller than approximately 600 g/mol. Therefore, no TOC reconciliation will be discussed here.
TOC is an established scouting method that can be used to compare extractables/leachables contribution from different bags based on the carbon content. The TOC values of the HCl extract are for all bags in the comparable concentration range from 1.7 to 2.1 mg/L carbon. For the NaOH extracts, higher differences in concentrations of carbon can be seen, for example, Bag C contains 1.3 mg/L carbon and Bag E 3.9 mg/L.
Discussion
The goals of this study were to present a procedure for extractables testing of single-use bag systems and to provide end-users with a database of potential extractables from such systems. Bags from five different suppliers were tested using identical extraction conditions. The results presented in this study provide qualitative extractables data for each bag exposed to extracting solutions that are considered worst case compared to typical biopharmaceutical process solutions. Quantitative estimations are given for detected compounds. For all bags, the major extractable compounds identified can be classified as stabilizer, degradation products of stabilizers, and polyolefins-related compounds. Moreover, some compounds have been determined to originate from tubing.
80% ethanol is considered a worse-case solvent with respect to the concentration of organic compounds present in a typical biopharmaceutical solution, and as a result would have a greater extraction capability. It was used to characterize the materials because it makes polymer chains swell and promotes the extraction of even large additive molecules, whereas aqueous solutions dissolve compounds which are on the surface of the material, resulting in smaller hydrophilic extractable molecules.
The results clearly show that different classes of extractables were extracted with the 80% ethanol solution compared to the aqueous solutions. In the 80% ethanol extracts, large stabilizer molecules (e.g., Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite and Octadecyl 3-[3,5-bis(1,1-dimethylethyl)-4-hydroxyphenyl]propanoate) were detected, which were not found in the aqueous extracts. The compound extracted with the highest concentration in the 80% ethanol extracts was Tris[2,4-bis(1,1dimethylethyl)phenyl]phosphite (molecular weight: 646.9 g/mol), which was present at a maximum concentration of 143300 μg/L. In the aqueous extracts the most released compound was stearic acid (molecular weight: 248.5 g/mol), which was present at a maximum concentration of 140 μg/L.
In addition to the groups of chemical compounds representing degradation products of antioxidants and other polymer additives (e.g., 2,4 bi-tert-butylphenol), degradation products of polyolefins due to gamma irradiation of the bags were also detected. These include small oxygen-containing molecules detected in all extracts such as aldehydes, ketones, carboxylic acids and include as well hydrocarbons detected in the 80% ethanol and 1% polysorbate 80 extracts.
Within the scope of this extractables study, six additives listed in the EP were detected, whereas none of the identified substances are present on the REACH list of substances of very high concern.
In addition to substances known to originate from the resin, unexpected substances (such as caprolactam, when polyamide is not part of the film structure) were detected in some of the extracts. The presence of such unexpected substances should be pursued by the single-use manufacturer to determine possible sources. Suggestions of sources to investigate are packaging materials, the bag manufacturing process, and connector materials, for example. This investigation was out of the scope of this study.
The manufacturing of plastics for single-use bag systems naturally leads to the degradation of polymers and their additives. During the manufacturing process that transforms resin to film, polymer degradation products such as oligomers and free side group chains are formed. At the same time, additives such as antioxidants, while playing their protective role, will transform or degrade. In general, it has to be considered that polymer stabilization is a dynamic process which may lead to extractables that are not initially expected if only the raw materials (resins) are evaluated with regard to potential extractables.
Other unexpected substances can be caused by gamma irradiation. Gamma irradiation creates reactive species that can interact with oxygen and moisture and lead to the formation of volatile compounds. These volatiles were primary and secondary oxidation products including aldehydes, ketones, alcohols, and carboxylic acids (40). One question that arises is whether the extractables profile is dependent on the length of time since the test article was gamma-irradiated due to the continued action of the active species. The design of this study did not take this variable into account; therefore, the results cannot be used to evaluate the “development” of extractables in the irradiated bag systems over time. It is expected that the extractables profile would change depending on the length of time since gamma irradiation was performed due to the transformation of additives in the polymer matrix as a consequence of the dynamics of the polymer stabilization process. Even if the extractables profile changes, the impact on the concentrations may be considered as low. However, this should be further investigated at process-related bag usage conditions.
In summary, the extractables data provided by single-use-bag system suppliers should cover all potential sources of extractables including resins, processing aids, and degradation products. To provide supportive extractables data the manufacturer has the obligation to manufacture products that have a consistent extractables profile. This is accomplished by careful selection of raw materials based on defined specifications, robust manufacturing processes, appropriate quality and manufacturing controls, and verified cleaning procedures.
In single-use-bag systems, the solution storage container (i.e., bag) is the component that has the longest contact time with the biopharmaceutical process solution. However, transfer lines and tubing may also introduce extractable substances into the biopharmaceutical process solution. The overall extractables profile of single-use bag assemblies can be influenced by the tubing even if the contact time is brief and high-quality tubing is used.
Within the scope of this study, the quantities of detected extractables were estimated using an internal standard of known concentration. In the case of the aqueous extraction solutions, the resulting concentrations were typically near the detection limit of the analytical methods. Even if qualified state-of-the-art analytical methodologies are applied, trace levels of extractables can be missed by physicochemical analysis. Depending on the specific application of the bags, it may be necessary to perform additional investigations, such as biological testing systems (cell cultures).
Specific and quantitative extractables information should be provided by the supplier, to support the toxicological and safety assessment that the end-user has to perform for each biopharmaceutical process solution. Based on the user's risk assessment and application (e.g., process step, proximity to final product), the evaluation of the supplier's extractables data can be appropriate to confirm suitability. In case of critical applications such as long-term storage of a drug substance or drug product stored in plastic containers, leachables studies should be performed as recommended in relevant EMA (5) and FDA (6) guidelines.
Conclusion
Based on the standardized procedures described in this study, commercially available, single-use bag systems from five different suppliers were investigated with regard to extractable substances. Test solutions bracketing the chemical characteristics of typical biopharmaceutical process solutions were used. Additionally, a stronger solvent was used for material characterization. For all bags, the primary extractables were classified as polymer additives and their degradation products or polyolefins and their by-products. Additionally, compounds from tubing materials were detected. This underscores the importance of having a testing of complete single-use bag assemblies, either through a component approach or a final bag assembly approach. An extensive database of potential extractables is provided to the users of single-use bag systems. The concentration of identified extractables ranged from parts-per-billion to parts-per-million for the extraction conditions used in the study. The concentration of particular identified compound varies from one to another bag.
In the case of critical applications (e.g., processes near the final drug product stage) or the use of more severe process conditions than represented in this study, product- and process-specific studies should be initiated. The end-user has to do a risk analysis and to provide a toxicological evaluation of the data with respect to their pharmaceutical products and the risk to patients.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
- © PDA, Inc. 2014
References
Jump to section
Related Articles
Cited By...
- Materials in Manufacturing and Packaging Systems as Sources of Elemental Impurities in Packaged Drug Products: An Updated Literature Review
- Using Extractables Data of Sterile Filter Components for Scaling Calculations
- Investigation of Air-Liquid Interface Rings in Buffer Preparation Vessels: the Role of Slip Agents