Abstract
Single-use systems for manufacturing biopharmaceuticals can include filter capsules, connectors, tubing, and polymeric film biocontainers. In order to tackle the variety of extractable compounds from these fairly complex systems, we first studied such systems' representative components, and then examined an entire single-use system comprised of filtesr, connectors, tubing, and biocontainers. This approach greatly simplifies the identification of the extractable compounds from the whole system. The test design was based on common, actual processes conducted under worst-case conditions.
Part 1 of this series of papers describes a systematic study of extractables from two components, a sterile connector and a capsule filter, in water and ethanol as model solvent extractants. The complete extractables results were obtained using a combination of qualified analytical methods. The results indicated that the potential for the connector and the capsule filter to release leachable materials in significant amounts into the chemically compatible drug product is very low, taking into account the less vigorous conditions in most processes and dilution effects when compared to the water and ethanol extraction conditions reported here. Application of study results is discussed.
1. Introduction
Single-use technology can offer distinct advantages over re-usable, cleanable systems in biopharmaceutical manufacturing. It minimizes the risks of cross-contamination and operator error in drug processing. Moreover, it can shorten processing turnaround time and reduce processing costs for new drugs (1–3).
Integrated, single-use filtration and storage systems are typically comprised of a capsule filter and polymeric film biocontainer(s) coupled with tubing, adaptors, and sterile connectors. Since all of these components are made from organic polymers or elastomers, extractables and leachable compounds have become a primary concern when adopting the technology (4). This is not surprising in view of the need to ensure that safety and regulatory requirements are satisfied. The Food and Drug Administration (FDA) regulation on Current Good Manufacturing Practice (cGMP) of Finished Pharmaceuticals (21 CFR Part 211) is applicable to process equipment including single-use systems. The section, Process Equipment 211.65(a) states, “Equipment shall be constructed so that surfaces that contact components, in-process materials, or drug products shall not be reactive, additive, or absorptive so as to alter the safety, identity, strength, quality, or purity of the drug product beyond the official or other established requirement” (5). The European Commission's Good Manufacturing Practices, Medicinal Products for Human and Veterinary Use (GMP) (6), the European Medicines Evaluation Agency (7), and the International Conference on Harmonization (8) have a similar regulatory guideline. The Parenteral Drug Association (9) has the general technical guidance on sterilizing filters. However, these guidelines are not specific in terms of the tactics required to meet the requirements above.
The US FDA Guidance on Container Closure Systems for Packaging Human Drugs and Biologics (10) addresses leachables as chemicals that actually migrate into the final drug product in the final container/closure system (although this is not the case for process equipment). Based on their presence in the final dosage, they must be identified, quantified, and assessed on their toxicity. Due to the complexity of drug products and their interferences with analytical methods, assessing leachables in protein-containing biotherapeutics is often a daunting task. To facilitate assessment of leachables in the final drug product, extractables studies from all contacting surfaces must be performed. This is usually carried out during a process validation program, where a process- and product-specific extractables test is performed to validate fluid-contacting process equipment (including single-use systems). Because single-use systems contain multiple components with diverse materials of construction which may be supplied by different vendors, the evaluation of extractables from single-use systems appears to be challenging. We first studied system components, and then examined the integrated system, in order to solve this problem. This systematic study generated an extractables data library that can help to speed process validation and to facilitate leachables assessment in final drug product.
This paper (Part 1) presents a study of extractables from two representative components, Pall Kleenpak™ connector and Kleenpak capsule filter, in water and ethanol as recommended by the Bio-Process Systems Alliance (11). Many extractables studies have been performed on filters (12–15). Historically, non-volatile residue (NVR) measurement and Fourier transform infrared spectroscopy (FTIR) analysis have been suitable because filter extractables are generally non-volatile in nature after steam sterilization, and these compounds can be captured in the NVR measurement after a low-boiling-point solvent is evaporated.
While the method continues to be suitable for approximating quantitative estimates of extractables from filters after steam sterilization, more analytical methods are needed for characterizing extractables from single-use systems. This is because single-use systems are typically pre-sterilized by gamma irradiation, which can generate trace amounts of volatile and semi-volatile compounds (16). Therefore, a gas chromatography/mass spectrometry (GC/MS) method was incorporated. The bulk properties of water extracts, total organic carbon (TOC), pH, and conductivity were measured. Ion chromatography (IC) was used to analyze acetate and formate. Some semi-volatile and non-volatile compounds were separated and identified using high-performance liquid chromatography/ultraviolet/mass spectrometry (HPLC/UV/MS). HPLC/UV method alone was not sufficient to detect and identify extractables. The metallic extractables were efficiently measured using inductively coupled plasma/mass spectrometry (ICP/MS). The significance of these studies and the applicability of the study method to process validation of single-use systems are also discussed.
2. Materials and Methods
2.1. Test Materials and Extraction Method
Extractables studies were applied to two representative components used in Pall's Allegro™ single-use systems: Kleenpak™ sterile connectors (P/N KPCHT02, Pall Life Sciences, East Hills, NY) and Kleenpak capsule filter assemblies with hydrophilic polyvinylidenefluoride (PVDF) membrane (1500 cm2 effective filtration area, P/N KA3DFLP1G, Pall Life Sciences, East Hills, NY). The Kleenpak connector is a two-piece connector system used for making sterile connections between two dry, sterile pathways. Once the connection is completed, only one material, polycarbonate, is subject to contact with the process fluid in the flow path. The connectors were tested after the connection was completed. Due to the very small fluid-contact surface area of each connector, a set of four connectors in a series linked by short lengths of polytetrafluoroethylene (PTFE) tubing were used in each test. The test filter capsule incorporates a 0.2-μm-rated Fluorodyne® II sterilizing-grade PVDF membrane filter with two major materials in contact with the process fluid: 1) polypropylene for the membrane support/drainage layers and capsule hardware, and 2) PVDF, surface-modified with a hydrophilic acrylate copolymer, for the membrane itself. Duplicate tests were performed for each solvent with each test article. Prior to the testing, the test articles were gamma-irradiated at 50 kGy, which represented an extreme limit of radiation dose used for product sterilization of single-use systems.
According to recommendations from the BioProcess Systems Alliance (11), two model solvents, water and ethanol, were chosen for the study. These solvents were chosen because most applications in bioprocesses involved buffers and aqueous solutions with low concentration of organic solvents or surfactants. Some buffer solutions have high or low pH values. Therefore, solvents with high or low pH should also be used to model the effect of pH on the extractables. These tests are in progress and the data is not available at this time. Deionized (DI) water (with 18 Megohm-cm resistivity) was obtained from a Pall Corporation DI water plant (Port Washington, NY), and ethanol (200 proof, HPLC-grade) was purchased from Sigma Aldrich (St. Louis, MO).
The recirculating system with a PTFE diaphragm pump, PTFE tubing, and a glass reservoir was used to eliminate or minimize the contribution of extractables from the test system. A negative control study was performed for each test using the recirculating system with the test solvent but without the test article. The extraction time was 24 h for water and 4 h for ethanol. All extraction tests were performed at ambient temperature of 21 ± 1 °C. The solvent volume for each test was 1.5 L. The schematic of the extractables test systems is illustrated in a publication (17).
2.2. Analytical Methods
Table I lists the analytical methods used to evaluate the extracts. TOC is a bulk method for assessing all organic molecules in the form of mass of carbon. The TOC test was performed using a Tekmar-Dohrmann (Mason, OH) Pheonix 8000 analyzer. Measurement of pH indicates change of acidity/alkalinity due to extractables. The pH meter (SP70P) was obtained from VWR International, LLC (West Chester, PA). Conductivity was measured to assess the ability of conducting electric current resulting from the presence of ions. The conductivity meter was obtained from Amber Science (Eugene, Oregon). Ion chromatography effectively detects acetate and formate. The instrument used was a Dionex (Bannockburn, IL) model ISC-2000 RFIC with an IonPac AS17 analytical column (4 × 250 mm) and an IonPac AG17 guard column (4 × 50 mm) at a flow rate of 1.0 mL/min. A gradient mobile phase of 0.3–60 mM KOH was prepared by the ICS-2000 Eluent Generator. The injection volume was 100 μL. Column temperature was regulated at 30.0 °C.
Analytical Methods Used for Assessment of Extractables
The TOC, pH, conductivity, and IC tests were performed for water extracts only, as they are not applicable to organic solvents.
Total amount of non-volatile residue (NVR) was measured gravimetrically by evaporating the test solvent. Because the volatile extractables that can be quantified using GC/MS are a miniscule portion of total extractables by weight, NVR represents the best quantitative estimate of total extractables. All individual chromatographic methods would otherwise leave some compounds undetected, especially higher-molecular-weight compounds including oligomers of polymers used as the materials of construction of single-use systems, and oligomers of polymers used as functional additives, such as wetting agents for the filter membrane. The NVR was then qualitatively analyzed using FTIR spectroscopy (Nicolet 6700), which analyzed functional groups present. The FTIR results were further compared with spectra of standards in the library in OMNIC software from Thermo Nicolet (Madison, WI) and the Pall internal customized library.
The liquid extracts were also subjected to UV spectroscopy analysis without any pretreatment. The UV spectrometer was obtained from Shimadzu (Columbia, MD). The purpose of the UV scan was to find out if the extracts contained detectable amount of chromophore-containing compounds, such as aromatics.
Volatile organic compounds were detected by headspace GC/MS using liquid extracts without any pretreatment for analysis, via the Shimadzu QP2010 instrument. The GC column was DB-624 from Agilent (Santa Clara, CA), 60 m × 0.25 mm × 1.4 μm. The shaker and syringe temperatures were 70 and 90 °C, respectively. The ion source temperature was 240 °C. The mass spectra scan range was set from 35 to 650 m/z.
Semi-volatile compounds were analyzed by direct injection GC/MS, using the Shimadzu QP2010 instrument. The GC column was DB-1 from Agilent (Santa Clara, CA), 60 m × 0.25 mm × 0.25 μm. The ethanol samples were injected without any pretreatment with other solvent, but the water samples were subjected to solvent exchange to predominantly acetonitrile before injection. The ion source temperature was 240 °C. The mass spectra scan range was set from 35 to 650 m/z.
Derivatization GC/MS was used to determine if the extracts contained any organic acids, such as fatty acids. These organic acids are not sensitive enough to be studied via direct injection GC/MS because they are highly polar and are often adsorbed by common GC columns. The fatty acids were then derivatized using BF3/1-butanol as a derivatization agent to convert to their butyl esters, followed by extraction with n-hexane (18–20). Palmitic acid-d31, used as an internal standard for the derivatization and extraction, was added to all samples and calibration standards. After the derivatization and extraction steps, 2 μL of the final n-hexane solution was injected into the GC/MS for analysis. A calibration curve based on the authentic standard with a linear regression was established for a prospective compound, using the Shimadzu QP2010 instrument. The GC column was DB-Wax from Agilent, 60 m × 0.25 mm × 0.25 μm. The ion source temperature was 240 °C. The mass spectra scan range was set from 35 to 550 m/z.
An Agilent model 1100 HPLC system was used for LC/UV/MS analysis, which included an automatic liquid sampler, binary pump, column compartment, and diode array detector (DAD) coupled with an Agilent mass selector detector. Reversed-phase liquid chromatographic separation was performed on an Agilent Hypersil ODS column (4.6 mm × 150 mm, particle size 5 μm) with the following mobile phase gradient: starting with 50% A (USP purified water) and 50% B (acetonitrile) and increasing solvent B to 100% in 11 min, holding the condition until 25 min, returning to 50% A and 50% B in 5 min, and then holding the condition for 5 min. Different mobile phase gradients (started with up to 90% A) and different mobile phases (methanol/water, methanol/isopropyl alcohol/water, and others) were also used to optimize the separation and ionization. The injection volume was 20 μL and the mobile phase flow rate was 1.00 mL/min. UV detection was set at 200 and 280 nm. Both atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) operating in both positive and negative polarity modes were employed for ionization. The mass spectra scan range was set from 100 to 2000 m/z for both APCI and ESI methods.
ICP/MS was used to analyze extracted metals. All test solutions were acidified with 2% HNO3 solution and analyzed via ICP/MS. The common metals, including heavy elements, analyzed using ICP/MS (Agilent 7500cs) were Silver (Ag), Aluminum (Al), Barium (Ba), Boron (B), Calcium (Ca), Cobalt (Co), Chromium (Cr), Copper (Cu), Iron (Fe), Potassium (K), Lithium (Li), Magnesium (Mg), Manganese (Mn), Sodium (Na), Nickel (Ni), Lead (Pb), Tin (Sn), Titanium (Ti), and Zinc (Zn). The B, Sn, and Ti elements were analyzed under the hot plasma condition, while the others were tested under the cool plasma condition.
3. Results and Discussion
3.1. TOC, pH, Conductivity, and Ion Chromatography
Table II lists the TOC, pH, and conductivity results of water extracts from the sterile connectors and filter capsules.
Average TOC, pH, and Conductivity Results of Water Extractsa
The results from the Kleenpak connector extracts were not significantly different from the negative control values. Considering that each set of test samples was a pool of four connectors, each Kleenpak connector will not cause significant changes in TOC, pH, and conductivity.
Compared to the negative control, the Kleenpak filter extracts in water had a slightly lower pH. This was probably due to the trace amount of fatty acids identified in Section 3.3.3. Since only 1.5 L of water was used for extraction testing, the actual process fluid would dilute the extractables significantly, thus, the effect of extractables on pH would be minimum during actual process use.
The acetate and formate results are listed in Table III. Acetate and formate were below detection limit for the connector extracts. Low-molecular-weight organic acids, specifically formic and acetic acid, can be formed during the gamma irradiation of polyolefin material (21). The detected levels of such acids from the filter capsule were extremely low (parts per billion range); these acids might be derived from polypropylene material used in the construction of the filter components.
Average Acetate and Formate Results of Water Extracts
3.2. Non-Volatile Residue (NVR) and Fourier Transform Infrared (FTIR) Spectroscopy
Half the volume of the extracts was evaporated down and then dried at 60 °C until constant weight was obtained. The weight measured was scaled up to the full extraction volume. Table IV lists the NVR results. The amount of non-volatile residue from a set of four Kleenpak connectors was below detection limit. The Kleenpak filter capsule NVR from aqueous extracts was very low, with a mean value of 5.5 mg. As expected, the filter capsule NVR from the ethanol extracts was higher (mean value of 56 mg) and in the range typically reported for high area filters of this type. As many bioprocessing fluids are aqueous solutions containing less than 30% organic constituents, ethanol is one of the extreme worst-case model solvents used for typical bioprocessing fluids. The flush step will lower the amount of extractables significantly. Even without flush, the NVR will be diluted in process fluid. For example, if the volume of a process fluid is 150 L, the concentration of total NVR (56 mg) will be lower than 0.5 ppm.
Average Non-Volatile Residues (NVR)
The KBr (potassium bromide) pellets were made with the NVRs for FTIR analysis. The FTIR spectra of filter capsule NVR from water and ethanol extracts are shown in Figure 1. The reference IR spectra of an acrylate copolymer, and polypropylene were included for comparison purposes. The majority of NVR from ethanol extract was from oligomers of acrylate copolymer (3450, 2800–3000, 1730, 1270, 1130–1185 cm−1) used to render the PVDF membrane hydrophilic, as well as from oligomers of polypropylene (2800–3000, 1460, and 1377 cm−1) used to construct membrane support/drainage layer and the filter hardware. The IR spectrum from water extract only showed the presence of oligomers from acrylate copolymer.
FTIR spectra of water and ethanol extract NVR, acrylate copolymer, and polypropylene.
3.3. GC/MS
3.3.1. Headspace GC/MS:
No peaks were detected from both the water and ethanol extracts from the sterile connectors using headspace GC/MS. 2-methyl-2-propanol was used as a standard to assess the limit of quantification (LOQ) and limit of detection (LOD). The LOQ was obtained based on the concentration at which the standard deviation achieved 15% or less from 12 injections of the standard. The LOD was determined by the minimum level at which the standard can be reliably detected. The LOD and LOQ for 2-methyl-2-propanol were 0.001 and 0.01 ppm, respectively.
No peaks were detected from the water extracts from the filter capsule using headspace GC/MS. Only one compound, acetal, was found in ethanol extracts from filter capsule, with average concentration of 0.850 ppm.
3.3.2. Direct Injection GC/MS:
The direct injection GC/MS analyses of the water and ethanol extracts from sterile connectors did not detect any semi-volatile-compound peaks. Using lauryl acrylate as a standard, the LOQ and LOD of the GC/MS analysis were 0.100 and 0.025 ppm, respectively. In addition, the water extracts were concentrated 50-fold and the subsequent concentrates did not generate any peaks. Bisphenol-A, the primary building block for polycarbonate, was not detected from either water extract or ethanol extract under the test conditions.
The direct injection GC/MS analyses of the water extracts from filter capsule (after solvent exchange) indicated that the extracts did not contain any semi-volatile compounds in detectable level. However, several small peaks were observed from GC/MS total ion chromatogram (TIC) of ethanol extracts from the filter capsule. In order to increase the intensity of the peaks to facilitate identification and quantitation, the ethanol extracts were concentrated 10-fold by evaporating part of ethanol under nitrogen at about 40 °C. The concentrated samples (10×) were then injected to the GC/MS. Figure 2 shows the GC/MS TICs from the original filter extract and the concentrated sample.
GC/MS total ion chromatograms of filter ethanol extract and 10× filter ethanol extract. Note: the y-axes scales are different.
Based on the peaks' mass spectra and their matches with the standard mass spectra in the Wiley library, the most probable compounds were identified, and authentic standards were obtained from commercial suppliers. These authentic standards were then used to make the standard solutions in ethanol and analyzed according to the same procedure for the filter extracts. If the retention times and the mass spectra between the peak (of unknown) and the standard matched, then the standard was assigned to the peak. The standard was then used to make a calibration curve, perform a one-point verification of the curve, generate LOD and LOQ, and analyze concentration of the corresponding peak based on the calibration curve. The results are listed in Table V.
Identified Compounds in Ethanol Extracts from Filter Capsule Using GC/MS
Three peaks at retention times of 17.714, 29.385, and 32.701 min were tentatively identified, based on the mass spectra, as isomers of 4-hydroxy-4-methyl-2-pentanone, lauryl alcohol, and decyl propanoate, respectively. The semi-quantitation of these peaks indicated that their concentrations were below 0.250 ppm.
To qualify the concentration step, a solution containing 0.5 ppm of 1,3-di-tertbutylbenzene, 1-dodecanol, 2,4-di-tertbutylphenol, lauryl acetate, lauryl acrylate, 2,2,4,4,6,8,8 heptamethylnonane (internal standard), and 5 ppm of 2-ethylhexanoic acid was made and subjected to the same concentration step as the ethanol extracts. The concentrated sample was analyzed using GC/MS, and the results are listed in Table VI.
Recovery of the Standard Compounds during Concentration Step
The recovery of all the compounds except 2-ethylhexanoic acid was found to be from 99 to 109%. The relatively lower recovery of 2-ethylhexanoic acid was understandable due to its high polarity and its possible adsorption on the column.
3.3.3. Derivatization GC/MS:
The derivatization method was extremely sensitive. Only benzoic acid (the limit of detection was 0.001 ppm and the limit of quantitation was 0.025 ppm) was detected with very low concentrations of 0.039 and 0.047 ppm in water and ethanol extracts from sterile connector extractions, respectively.
The derivatization GC/MS results from filter capsule extracts are listed in Tables VII and VIII. Several organic acids in trace amount were detected. Figure 3 shows a GC/MS total ion chromatogram of a derivatized ethanol extract from filter capsule.
Average Concentrations of Organic Acids in Water Extracts from Filter
Average Concentrations of Organic Acids in Ethanol Extracts from Filterab
GC/MS total ion chromatogram of derivatized ethanol extract from filter capsule (see Table VIII for identification).
3.4. HPLC/UV/MS
Water and ethanol extracts were analyzed by HPLC/UV/MS using diode array detection at 200 and 280 nm, and operating the MS using both APCI and ESI in positive and negative polarity modes. The practical identification and quantification limit for the analytical methods utilized was 0.2 ppm.
For both water and ethanol extracts from sterile connectors, no extractables' peaks were observed. The water extracts from filter capsule did not generate any peaks that were above the practical identification and quantification limit.
Three compounds in the ethanol extracts from filter capsule were detected and quantified. Figure 4 shows the LC/UV chromatograms from the negative control and ethanol extract at 200 nm. The peaks were identified by matching mass spectra and retention times. The mass spectrum of the compound corresponding to the peak at 6.175 min in Figure 4 matched that of the 2,4-di-tert-butylphenol in a standard mixture in APCI-negative mode. The retention times of the compound and the 2,4-di-tert-butylphenol in a standard mixture also matched. Therefore, the peak at 6.175 min in Figure 4 was assigned as 2,4-di-tert-butylphenol. Similarly, the mass spectrum of the compound corresponding to the peak at 14.021 min in Figure 4 matched the spectrum of the oxidized Irgafos® 168 in a standard mixture in APCI-positive mode. The retention times of the compound and the oxidized Irgafos 168 in a standard mixture also matched. Therefore, the peak at 14.021 min in Figure 4 was assigned as oxidized Irgafos 168.
HPLC/UV chromatograms of the negative control (top) and the filter ethanol extract (bottom).
The peak at ∼8.8 min was not ionized under the initial LC/MS conditions. To increase the signal intensity, the ethanol extracts were concentrated 10-fold, and the 10× samples were analyzed again. The peak was not ionized either. Change of mobile phase component from acetonitrile to ammonium acetate or 0.1% TFA (trifluoroacetic acid) did not result in ionization of this peak either. This peak was then collected using a micro-fractionator. The collected solution was extracted with n-hexane. The n-hexane phase was then injected into GC/MS, and the resulting peak had mass spectrum matching with that of 1,3-di-tert-butylbenzene (Figure 5). A standard solution with 12 ppm 1,3-di-tert-butylbenzene in ethanol was made and analyzed using LC/UV/MS. The results of the standard solution and the 10× filter ethanol extract were overlaid in Figure 6, which indicated the matching of the retention times. The UV scan of the two solutions generated identical patterns. These results confirmed the peak was indeed 1,3-di-tert-butylbenzene.
Mass spectrum of the peak of interest (top) and the library mass spectrum of 1,3-di-tert-butylbenzene (bottom).
HPLC/UV chromatograms of 10× ethanol extract (dotted line) and standard 1,3-di-tert-butylbenzene solution (solid line).
Both 2,4-di-tert-butylphenol and 1,3-di-tert-butylbenzene were identified and quantified during the direct injection GC/MS analysis. Oxidized Irgafos 168 was only identified using LC/UV/MS. It was quantified against an oxidized Irgafos 168 standard at 3 ppm, and the concentration of oxidized Irgafos 168 in the ethanol filter extract was found to be lower than 0.2 ppm. The compounds, 2,4-di-tert-butylphenol and 1,3-di-tert-butylbenzene have also been identified as radiolysis products of the antioxidant, oxidized Irgafos 168 (22, 23).
The results from LC/MS analysis in ESI mode on ethanol filter extract also indicated the presence of trace amounts of 2-ethylhexanoic acid, lauric acid, myristic acid, palmitic acid, and stearic acid. Since these compounds were already identified and quantified using derivatization GC/MS described in Section 3.2.3, there was no need to duplicate the efforts.
3.5 ICP/MS
The water samples were analyzed directly, while the ethanol samples were diluted ten-fold before they were introduced to the ICP/MS system. LOD of elements was determined by analyzing a known concentration of standard seven times and calculated using the following formula:
Table IX lists LOD, recovery, coefficient of determination (r2), and element concentrations from water and ethanol extracts from the sterile connector.
Average ICP/MS Results from Water and Ethanol Extracts from One Connector
Analysis for 19 elements by ICP/MS showed 15 metals were below detection limits for the extracts from the sterile connector. Only B, Ca, Na, and Zn were higher than the detection limit, but less than 1.1 ppb.
Table X lists LOD, recovery, coefficient of determination (r2), and element concentrations from water and ethanol extracts from filter capsule. Only Na, Mg, Al, K, Ca, Cu, and Zn were detected. All had concentrations in the parts-per-billion range with most of them in single-digit parts-per-billion level.
Average ICP/MS Results from Water and Ethanol Extracts from Filter Capsule
3.6. Conclusions
Table XI lists the overview of the results of water and ethanol extracts of gamma irradiated Kleenpak sterile connector and Kleenpak filter capsule with Flurodyne II 0.2-μm rated membrane (1500 cm2 effective filtration area). The qualitative and quantitative data strongly indicate that the potential for the Kleenpak connector to release leachable materials into the drug product is extremely low; taking into account the less vigorous conditions in most processes and dilution in larger volumes.
Overview of Results of Water and Ethanol Extracts From Both Kleenpak Connector and Kleenpak Filter Capsule
The concentrations of each detected and identified extractable compounds from the Kleenpak filter capsule into water and ethanol were very low, less than 0.5 and 2.3 ppm, respectively. Since only 1.5 L of solvent was used for the testing, the actual concentration of each individual compound in larger volume of process fluid, e.g. 150 L, would be less than 0.005 and 0.023 ppm. The amounts of NVR from water and ethanol extracts were in single- and double-digit milligram levels. If 150 L of process fluid was in contact with the filter capsule even without pre-flush, the concentration of total NVR would be less than 0.5 ppm.
There is no single analytical method that can identify and quantify all the extractable compounds. A combination of analytical methods listed in Table I help to solve the extractables puzzle. The approach of testing components first and then the system was proven to be effective. The detailed analytical results obtained in this study contribute to the establishment of an extractables data library for single-use systems.
4. Summary
The identities of the compounds detected are consistent with the basic materials of construction, which have been assessed for biological safety during product qualification (24, 25) by industry standard tests such as USP 〈88〉 Biological Reactivity, in vivo, for Class VI Plastics (26). The materials of construction are also included in a product Drug Master File submitted to the FDA.
The study of the two representative components of single-use system indicates that volatile and semi-volatile compounds can be identified and quantified using headspace and direct injection GC/MS, respectively, and organic acids by derivatization GC/MS. LC/UV/MS is effective in detection and identification of antioxidants and their degradation products. HPLC/UV alone is not sufficient for the detection and identification of extractables. The amount of non-volatile residues (containing majority of semi-volatile compounds and all non-volatile compounds) can be quantified by NVR measurement, and the NVR can be qualitatively evaluated by FTIR with the aid of standards. For filters made of polymeric materials and having high surface area, NVR measurement gives the best quantitative estimate of semi-volatile and non-volatile compounds, which are the majority of total extractables.
Combining NVR measurement and GC/MS results on volatile and semi-volatile compounds measurement, total extractables can be quantified. ICP/MS can effectively measure metallic extractables. These extractable compounds from the two components of single-use systems are in trace amounts even under exaggerated extraction conditions. Due to the complexity of extractables, no one analytical method can solve the puzzle alone. This systematic study approach can be applied to process- and product-specific extractables evaluation as part of the validation program for single-use systems.
Acknowledgements
We acknowledge Ronald Sauro, Yanxin Luo, Matt Connor, Neil Pothier (Chemic Labs), and Dave Wells (Chemic Labs) for conducting part of the analytical work; Jerold Martin, Helene Pora, and Janet Mathus for their constructive review of the manuscript; and Morven McAlister and Robert Dickstein for their constant support.
Footnotes
- © PDA, Inc. 2009
In This Issue
Jump to section
Related Articles
Cited By...
- Low Leachable Container System Consisting of a Polymer-Based Syringe with Chlorinated Isoprene Isobutene Rubber Plunger Stopper
- Materials in Manufacturing and Packaging Systems as Sources of Elemental Impurities in Packaged Drug Products: A Literature Review
- A Compilation of Metals and Trace Elements Extracted from Materials Relevant to Pharmaceutical Applications such as Packaging Systems and Devices