Abstract
Nearly 100 individual test articles, representative of materials used in pharmaceutical applications such as packaging and devices, were extracted under exaggerated conditions and the levels of 32 metals and trace elements (Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ge, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, Zn, and Zr) were measured in the extracts. The extracting solvents included aqueous mixtures at low and high pH and an organic solvent mixture (40/60 ethanol water). The sealed vessel extractions were performed by placing an appropriate portion of the test articles and an appropriate volume of extracting solution in inert extraction vessels and exposing the extraction units (and associated extraction blanks) to defined conditions of temperature and duration. The levels of extracted target elements were measured by inductively coupled plasma atomic emission spectroscopy. The overall reporting threshold for most of the targeted elements was 0.05 μg/mL, which corresponds to 0.5 μg/g for the most commonly utilized extraction stoichiometry (1 g of material per 10 mL of extracting solvent).
The targeted elements could be classified into four major groups depending on the frequency with which they were present in the over 250 extractions reported in this study. Thirteen elements (Ag, As, Be, Cd, Co, Ge, Li, Mo, Ni, Sn, Ti, V, and Zr) were not extracted in reportable quantities from any of the test articles under any of the extraction conditions. Eight additional elements (Bi, Cr, Cu, Mn, Pb, Sb, Se, and Sr) were rarely extracted from the test articles at reportable levels, and three other elements (Ba, Fe, and P) were infrequently extracted from the test articles at reportable levels. The remaining eight elements (Al, B, Ca, Mg, Na, S, Si, and Zn) were more frequently present in the extracts in reportable quantities. These general trends in accumulation behavior were compared to compiled lists of elements of concern as impurities in pharmaceutical products.
LAY ABSTRACT: Nearly 100 individual test articles, representative of materials used in pharmaceutical applications such as packaging and devices, were extracted under exaggerated conditions, and the levels of thirty-two metals and trace elements (Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ge, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, Zn, and Zr) were measured in the extracts. The targeted elements could be classified into four major groups depending on the frequency with which they were present in the extractions reported in this study: those elements that were not extracted in reportable quantities from any of the test articles under any of the extraction conditions, those elements that were rarely extracted from the test articles at reportable levels, those elements that were infrequently extracted from the test articles at reportable levels, and those elements that were more frequently present in the extracts in reportable quantities.
Introduction
During their manufacturing, storage, and administration, pharmaceutical drug products come in contact with systems designed to facilitate these operations. For example, the drug product may contact materials present in the manufacturing suite (e.g., mixing tanks, filters, filling lines, etc.) during the manufacturing process. After manufacturing but before use, the drug products are stored, and potentially processed (e.g., terminally sterilized), in a packaging system. Finally, the drug product may be administered during use via a medical device. Although it is the case that these manufacturing, packaging, and delivery systems are constructed with materials and by processes that seek to minimize the extent to which interactions between it and the drug product can occur during contact, it is unfortunately the case that neither truly inert materials and systems nor truly benign contact conditions exist and that interactions with potential product quality impact are the rule rather than the exception.
One type of interaction that can occur between a drug product and a material that it contacts is the migration of chemical entities out of the material and into the product. The accumulation of such migratory substances in the finished drug product is of concern due to the impact that such substances could have on the finished drug product's suitability for use. Clearly, such migratory substances could adversely affect the safety and efficacy of the finished drug product, both by direct and indirect means. Thus it is both necessary and mandatory that the extent of such an interaction be ascertained and that it be established that the impact of the interaction is within acceptable limits.
One can establish the extent of the interaction between a drug product and a contact material either directly, by analyzing the drug product for material-related leachables, or indirectly (by inference), by testing the material for extractables. Given the analytical difficulty of profiling chemically complex drug products for trace quantities of unknown leachables, establishing the extractables profiles of materials is generally utilized to estimate the leachables profile and to establish target leachables to be quantified in drug products. Additionally, extractables testing may be a more useful means of exercising change control and performing quality control (QC) testing of incoming materials.
The analytical process for characterizing an extract for extractables generally involves orthogonal approaches so that the overall analytical approach encompasses the largest population of potential extractables. Most commonly such a process involves chromatographic methods for profiling organic extractables and atomic spectroscopic methods for profiling extracted elements. Considering extracted elements specifically, inductively coupled plasma atomic spectroscopy (ICP-AS) is the methodology most commonly utilized to address such extractables. At the current time, the ICP-AS analysis is accomplished by using either atomic emission (ICP-AES) or mass spectrometry (ICP-MS) detection (see, for example, References 1⇓–3).
Several publications address elemental extractables for a small set of tested materials (for example, References 4⇓⇓⇓⇓⇓⇓–11), and there are several reviews that compile extracted element data from numerous sources (for example, References 12, 13). However, there are few papers that report extracted element results obtained for a large population of test materials, obtained via a more or less consistent analytical approach, that are relevant to pharmaceutical applications. The purpose of this paper is to report the results of such an analytical characterization, as such information may be relevant to the development, qualification, registration, and ongoing QC of manufacturing systems, packaging systems, packaged drug products, and medical devices.
Specifically, nearly 100 individual test articles—representative of materials used in pharmaceutical applications, such as acrylonitrile butadiene styrene co-polymer, acrylic copolymers, glass, polycarbonate, polyethylene, polyethersulfone, poly-vinyl chloride, polymethylmethacrylate (acrylic) copolymers, polypropylene (PP), rubber elastomers, silicones, thermoplastic elastomers and miscellaneous materials—were subjected to over 250 exaggerated extractions and the levels of 32 extracted elements (Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ge, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, Zn, and Zr) were measured in the extracts. The extracting solvents included aqueous mixtures at low and high pH and an organic solvent mixture (40/60 ethanol water). The sealed vessel extractions were performed by placing an appropriate portion of the test articles and an appropriate volume of extracting solution in inert extraction vessels (Pyrex glass bottles in all cases except for high pH extracting media, which utilized Teflon) and exposing the extraction units (and associated extraction blanks) to defined conditions of temperature and duration. The levels of extracted target elements were measured by ICP-AES. This article describes the test methodology that was used, contains the test results, and considers generalizations that can be drawn from the resulting dataset.
Experimental
Test Articles
Nearly 100 individual materials were tested and are reported herein. These test articles are representative of materials used in medical applications such as drug delivery devices, drug product packaging, and drug product manufacturing systems. The test materials were typically resins or pieces of materials that would themselves be further processed to produce devices, packaging systems, or manufacturing systems. Occasionally, the test articles were materials that had been processed to produce parts (e.g., molded) or extruded to produce films. However, none of the test articles were marketed medical devices or packaging systems or manufacturing components used to produce commercial products, and not all of the tested materials are currently used in marketed products or in manufacturing systems for marketed products. Consistent with their potential applications, some of the materials were gamma-irradiated prior to testing.
Extraction
The individual test articles were extracted and the levels of 32 metals and trace elements were measured in the extracts. The extracting solvents included aqueous mixtures at low and high pH and an organic solvent mixture (40/60 ethanol water). The sealed vessel extractions were performed by placing an appropriate portion of the test articles and an appropriate volume of extracting solution in inert extraction vessels (either glass or Teflon) and exposing the extraction units (and associated extraction blanks) to defined conditions of temperature and duration. Extractions were generally performed in triplicate, and it is the mean results that are reported herein. As the intent of the testing was to establish each material's suitability for use in specific applications, the extractions performed were customized to exaggerate, in a general sense, the anticipated conditions of clinical use. Thus the extraction conditions were not rigorously standardized across all materials. Nevertheless, the extraction conditions used to generate the tested samples are generally comparable between the test articles.
Extract Analysis
The extracts and associated extraction blanks were screened for 32 targeted metals and elements (Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ge, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, Zn, and Zr) by ICP-AES. After the extraction was complete but prior to the actual analysis, the aqueous samples were fortified to contain 1% nitric acid, and the ethanol/water samples were fortified to contain 4% nitric acid. The ICP-AES responses were calibrated using working standards that were matrix-matched versus the samples and prepared in the concentration range of 0.05 to 5.0 μg/mL from primary standards traceable to NIST reference materials. The analyses were performed on Agilent model 725ES and Varian Vista Pro ICP-AES spectrometers equipped with radial plasma viewing; typical operating parameters and sample introduction components are listed in Table I. The emission wavelengths used to quantify the targeted elements are listed in Table II; the background correction technique applied to each wavelength was typically a polynomial-fitted correction; analyte integration was accomplished by assigning 1 point per pixel. In the case of the ethanol/water extracts, the carbon interference precludes accurate determinations at wavelengths below 200 nm (see Table I) at the 0.05 or 0.1 μg/mL concentration level. For these wavelengths, spectral de-convolution was required to distinguish the analyte's spectrum from the background carbon band emission.
General Operating Parameters for ICP-AES Analysis
Target Elements and Their Primary Emission Wavelengths
Internal standard correction was used to compensate for instrument drift and slight differences between sample and standard matrices. A 2 μg/mL yttrium internal standard prepared in the working standard matrix was added continuously on-line throughout the analysis sequence.
All analyses performed in support of this work included rigorous system suitability testing. Critical performance characteristics were measured and verified against appropriate acceptance criteria (Table III). A laboratory control (LC) sample was prepared from a different stock source than the stock used for the working standard solutions. The target element concentration of the LC is within the working standard concentration range but is not identical for each element. The LC served as a QC check of the standardization of the wavelengths as well as a measure of instrument stability. Additionally, a supplemented sample (SS) was prepared for each extraction matrix by supplementing a 10 mL aliquot of an extraction blank with 1.0 μg/mL of all target elements. The SS was prepared and analyzed to demonstrate the accuracy of the method for determining the target elements in the extracting solution.
System Suitability Requirements
Results and Discussion
Thirty-two elements—Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ge, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Ti, V, Zn, and Zr—were targeted for quantitation for several reasons, including
Potential safety impact
Potential product impact
Probability of being present based on test article composition
The reporting thresholds for the extracted amounts of the targeted analytes were generally 0.05 μg/mL for all elements except a few (e.g., Ge and Li) that had a reporting threshold of 0.10 μg/mL. This difference occurs because the absolute value of the reporting threshold is determined at each test interval and there is some interval-to-interval, as well as element-to element, variation in method performance. The reporting threshold is not necessarily the method's limit of quantitation but reflects that concentration at which one can reliably discern a difference between the level of an analyte in an extract and its level in an associated extraction blank. This distinction between a practical and differential quantitation limit (QL) versus an absolute QL is relevant, as the amount of the analyte extracted from a test article is the difference between two measurements (amounts of the analyte in the extract and in the extraction blank) and not the result of a single measurement. This situation is illustrated via the following example. In this example, the test method has a QL of 0.01 μg/mL; at this QL the accuracy is 100 ± 50% and the precision is percent relative standard deviation (%RSD) = 25%. An extract and an extraction blank are tested, and the test results are extract = 0.04 μg/mL extraction blank = 0.02 μg/mL. Both the measurements are above the QL, and the difference (0.02 μg/mL) is also above the QL. However, one cannot claim, with confidence, that the extracted concentration is 0.02 μg/mL because the uncertainty in the individual results is greater than their difference.
The extraction stoichiometry (extracted material weight per unit volume of extracting solution) varied somewhat for each material tested because the intent of the testing was to establish each material's extractables properties under conditions similar to its clinical use. Thus materials that are used in different clinical settings were extracted in a different manner. Nevertheless, a stoichiometry of 1 g of material per 10 mL of extraction solvent was most commonly used. In this circumstance, the reporting threshold of 0.05 μg/mL on a solution basis corresponds to a reporting threshold of 0.5 μg/g on a material basis. Values less than the reporting threshold are contained in this paper but are considered to be concentration estimates that most correctly indicate that the particular element was detectable as an extractable in a particular circumstance.
The extracted metals and trace elements results are grouped in terms of the tested material type and are summarized in Tables IV through XVI. In general, the 32 targeted elements can be classified into four groups as a function of how frequently they were present in the extracts in reportable quantities. Thirteen of the targeted elements, Ag, As, Be, Cd, Co, Ge, Li, Mo, Ni, Sn, Ti, V, and Zr, were not extracted from any of the tested materials in measurable quantities. Eight other targeted elements, Bi, Cu, Cr, Mn, Pb, Sb, Se, and Sr, were rarely extracted in measurable quantities and were sporadically present in only a few extracts at levels near the reporting thresholds. Three additional elements, Ba, Fe, and P, were infrequently extracted in measurable quantities, meaning that while they were more prevalent than the sporadic elements, they were still present in less than 5% of all the tested extracts. These elements were generally extracted from certain types of test materials and not universally extracted from many different types of materials. Thus Ba was generally extracted from glass, PP, and acrylic copolymer; Fe was generally extracted from stainless steel; and P was generally extracted from materials, such as PP, silicone, and thermoplastic elastomers, that typically include phosphite-based anti-oxidants.
Levels of Extracted Elements, Acrylonitrile Butadiene Styrene (ABS) Copolymer
Levels of Extracted Elements, Acrylic Copolymers
Levels of Extracted Elements, Glass
Levels of Extracted Elements, Polycarbonate (PC)
Levels of Extracted Elements, Polyethylene (PE)
Levels of Extracted Elements, Polyethylene (PE), Continued
Levels of Extracted Elements, Polyethersulfone (PES)
Levels of Extracted Elements, Plasticized Poly-vinyl Chloride (PVC)
Levels of Extracted Elements, Plasticized Poly-vinyl Chloride (PVC), Continued
Levels of Extracted Elements, Polymethylmethacrylate Acrylic (PMMA)
Levels of Extracted Elements, Polypropylene (PP)
Levels of Extracted Elements, Polypropylene (PP), Continued
Levels of Extracted Elements, Rubber Elastomers
Levels of Extracted Elements, Rubber Elastomers, Continued
Levels of Extracted Elements, Rubber Elastomers, Continued
Levels of Extracted Elements, Silicone (SI)
Levels of Extracted Elements, Thermoplastic Elastomer
Levels of Extracted Elements, Miscellaneous Materials
The remaining eight elements, Al, B, Ca, Mg, Na, S, Si, and Zn, were those elements that were most commonly extracted in measurable quantities. None of these elements were extracted in measurable quantities from all the individual materials in each material group. The most prevalent extracted elements were Si and Na, which were extracted from individual materials across almost all the material groups but not all the individual materials within each group. Silicon was extracted in several materials, such as glass and silicone rubber, at concentrations well above 1 μg/mL and in many other individual materials in lesser amounts. The levels of extracted silicon exhibited two trends as function of the extraction solvent. For certain material groups, most notably acrylonitrile butadiene styrene (ABS), glass, rubber elastomer, and acrylic copolymers, the levels of extracted silicon were highest in the high pH extracts, which is consistent with an inorganic source of the silicon. For other material groups, most notably polymethylmethacrylate acrylic (PMMA) and silicone rubber, the levels of extracted silicon are highest in the ethanol/water extracts, consistent with an organic source of the silicon. Sodium was extracted in varying amount from many of the individual materials in all the material groups. Considering the other, less frequently encountered elements, zinc was generally extracted from the poly-vinyl chloride (PVC) and rubber elastomer materials in relatively higher amounts and only sporadically at much lower amounts from other individual materials. Sulfur was generally extracted from the rubber elastomers at relatively higher levels, from certain PP materials at much lower levels, and sporadically at lower levels in individual materials in the other groups. Calcium was extracted from many individual PVC, glass, PP, silicone, and rubber elastomer materials and more sporadically from materials in the other groups. Magnesium was primarily extracted from PVC and ABS materials; aluminum was more commonly extracted from the glass, PP, and rubber elastomer materials; and boron was most commonly extracted from the PVC, PP, silicone, and rubber elastomer materials.
Given the diversity of the test materials and the extraction conditions, it is difficult to elucidate universal trends in the extractables data. Nevertheless, two trends relating the concentration of certain extractables and the conditions of extraction, specifically extracting solution composition, could be discerned in the data. For example, the amounts of extractable metals and alkaline earths were generally higher in the low pH extracts, suggesting that ion exchange is an important extraction mechanism for these extractables. Additionally, the levels of extracted sulfur were generally higher in the pH 9 extracts, suggesting that one or more test article additives contained acidic sulfur moieties.
One way to utilize the results summarized in this article is to compare the test results with existing and/or developing requirements or recommendations for the composition of plastic materials used in pharmaceutical applications. Thus, for example, monographs for individual plastic materials that appear in the European Pharmacopeia (Ph. Eur.) prescribe the extraction of the materials and testing of the extracts for specified elements (14). As the extraction methods used to characterize the materials reported in this manuscript differ significantly from those in the Ph. Eur. (for example, the Ph. Eur. tests typically involve a high temperature, reflux-type of extraction with an acidic medium), it is not appropriate to rigorously compare the concentration results reported herein versus the Ph. Eur. elemental limits. However, one can qualitatively compare the list of Ph. Eur.–specified metals with those which were, or were not, detectable in the materials summarized in this paper. For instance, the various Ph. Eur. monographs on PVC materials used in a variety of applications target Ba, Cd, Ca, Sn, and Zn. Of these five targets, only Ca and Zn were present in detectable quantities in the PVC materials reported here. Similarly, the Ph. Eur. monograph for polyethylene terephthalate (PET) targets Al, Sb, Ba, Co, Ge, Mn, Ti, and Zn, none of which were extracted in reportable quantities from the PET materials reported herein. Considering polyethylene (PE) materials, the relevant Ph. Eur. monographs target Al, Cr, Ti, V, Zn, Zr, none of which were extracted in reportable quantities from the PE materials reported herein. Lastly, the Ph. Eur. monograph for PP targets Al, Cr, Ti, V, and Zn; of these targets, only Al and Zn were extracted in reportable quantities from the PP materials reported herein.
Similarly, one can compare the extractable elements data summarized herein to requirements and/or specifications for elemental or metallic impurities in drug products. Such a comparison is relevant because materials used in packaging are a potential source of elemental impurities in drug products. While a quantitative comparison between the results summarized in this paper and the specifications or limits for individual elemental impurities is not appropriate because the extractions performed on the materials that are contained in this study were not designed to facilitate such a quantitative comparison, one can qualitatively compare those elements that have limits or specifications with those that were extracted in reportable levels for the materials contained in this paper. For example, EMEA/CHMP/SWP/4446/2000, Guideline on the Specification Limits for Residues of Metal Catalysts or Metal Reagents (15), classifies 14 metal residues in terms of their safety concern and contains concentration limits for those residues, including Pt, Pd, Ir, Rh, Ru, Os, Mo, Ni, Cr, and V as metals of significant safety concern, Cu and Mn as metals with low safety concern, and Fe and Zn as metals with minimal safety concern. Of these 14 specified metals, Pt, Pd, Ir, Rh, Ru and Os were not targeted in the analyses reported herein and thus no comparison can be made. However, it is noted that three of the four remaining metals of significant safety concern, Mo, Ni, and V, were not extracted in reportable quantities from any of the materials contained in this study. The fourth remaining metal of significant safety concern, Cr, was rarely extracted from the materials in reportable quantities, and when it was extracted at reportable levels, the levels were typically low, less than 1 μg/g. Both metals with low safety concern, Cu and Mn, were rarely extracted from the materials included in this study in reportable quantities, and when the extracted amounts were reportable they were low, typically less than 0.1 μg/g. Considering the two metals of minimal safety concern, Fe was infrequently extracted from the materials in reportable levels. On the other hand, the remaining metal of interest, Zn, was one of the more commonly encountered extractable metals and was extracted from certain materials (for example, certain rubber elastomers) in quantities significantly higher than 1 μg/mL.
A similar appraisal can be made between the results reported in this study and the elemental impurities in pharmaceuticals that are listed in USP General Chapter <232> Elemental Imputrities—Limits (16). Those elemental impurities that appear in both the USP general chapter and the EMEA guideline are discussed in the preceding paragraph. Elements unique to USP General Chapter <232> include As, Cd, Hg, and Pb. Considering these elements specifically, Hg was not targeted in the analyses reported in this article and thus no comparison can be made. Both As and Cd were not extracted from any of the materials reported in this article in reportable quantities. While Pb was reported as being rarely extracted from the materials reported in this study (and at low levels when it was extracted), the occurrences of reportable extracted lead were limited to lead glass materials.
The essence of this comparative discussion is captured and summarized in Table XVII.
Qualitative Comparison of the Reported Findings with Specifications for Metallic or Elemental Impurities in Drug Products
In the final analysis, information such as that summarized in this paper may be useful in establishing requirements related to the presence of extractable elemental impurities in manufacturing systems, packaging systems, medical devices, and their associated materials of construction. Effective, achievable, and practical requirements are obtained by balancing the objectives of such requirements (e.g., producing safe and effective pharmaceuticals) with full knowledge of the types, levels, and frequencies of extractable elemental impurities that exist in packaging systems, devices, and their associated materials of construction.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
- © PDA, Inc. 2013
