Abstract
A comprehensive digestive approach for determining the extractable and leachable metals in pharmaceutical products by inductively-coupled plasma is investigated. This study examines several acid digestion strategies for packaging materials, containers, and formulated products for complete trace metals analysis. Packaging materials, a food product, and a simulated drug product are evaluated for leachable metals by stressing the materials under accelerated stability conditions. Trace metal profiles of 64 elements for these materials are reported.
Introduction
Given the current regulatory initiatives for Quality by Design to build quality into a pharmaceutical product from the beginning, the area of extractables and leachables has focused primarily on the organic side, where pharmaceutical products are evaluated for organic components that may migrate to the product. Safety concerns, thresholds, and the establishment of product quality attributes, with regard to extractables and leachables are typically evaluated based on data collected or identified by either gas or liquid chromatography coupled with mass spectrometry (GC-MS, LC-MS). This paper investigates the equally important area of inorganic analysis with regard to these concepts. Metals are known to have either a toxic effect based on the nature of the element, such as aluminum (Al), cadmium (Cd), chromium (Cr), copper (Cu), lead (Pb), manganese (Mn), and zinc (Zn) (1) or they contribute to interactions between active pharmaceutical ingredients (APIs) and excipients, such as iron (Fe), Zn, Cu, and Mn (2).
The classical approach to controlled extractables analysis is to expose the material to a specific solvent such as water, organic solvents, or buffers (3–5) for a period of time at an elevated temperature and measure the content of the extractable. Previous trace metals analysis on such things as rubber closures, plastic resins, and other pharmaceutical packaging materials has only been reported using either an extractive technique or a single selective type of acid digestion (e.g., microwave) (6, 7). While these approaches are useful for determining extractables, trace metals analysis can be limited in each of these cases by the quantitation limit (QL) of each metal, the volume and nature of solvent used, and the inefficiency of a solid/liquid extraction. Certain metals at low levels may not be detected, so the leaching may potentially go unnoticed.
The comprehensive digestive approach proposed here involves a series of complete acid digestion techniques of the packaging or container material which would provide the laboratory several options in the analysis of these materials. Sample solutions are then analyzed by either inductively-coupled plasma mass spectrometry (ICP-MS) or inductively-coupled plasma atomic emission spectrometry (ICP-AES). The ICP-MS can provide reduced sample preparation time, increased sensitivity, QL, and a complete metals profile for elements not typically measured by ICP-AES. A decision tree (Figure 1) was developed for the appropriate type of digestion and acid mixture for several different materials and some common pharmaceutical products on the basis of the nature of these materials and potential metals. Exposure of the packaging material to drug product at accelerated stability conditions (e.g., 40 °C/75% relative humidity, RH) provides information into any leachable that has migrated from the packaging into the drug formulation.
A comprehensive method of evaluating the entire metals profile of the packaging or container material provides a more effective means for determining the potential metal extractables over current limited extraction techniques. The concept of using a comprehensive method of analysis for trace metals is based upon evaluating the material of interest for all potential contaminates. Until recently, trace metal exposure limits and toxicity data were reported in literature relative to such things as drinking water or as an air contaminate (8–12); however, recent regulatory stimuli (13) and European guidelines (14) have published limits for up to 31 elements of interest in oral and parental drug materials. The experiments described in this article were designed to demonstrate that typical packaging materials may contain trace metals at or above the exposure limits and that these trace metals may go undetected based upon the type and nature of extraction solvent used or the incorrect selection of metals to screen or monitor.
Materials and Methods
Reagents used were all trace metals-grade or equivalent without further treatment and were purchased from Fisher Scientific (Fair Lawn, NJ). Samples analyzed for this study were readily available as part of our laboratory inventory of supplies: Type I glass vials, high-density polyethylene (HDPE) bottles, polyester bottles, polycarbonate bottles, and rubber lyophilization stoppers. These had been previously purchased from several different laboratory vendors such as Fisher Scientific and VWR International (West Chester, PA).
A pre-filled syringe placebo formulation was prepared in-house based upon a commercial formulation published in the Physicians' Desk Reference (15). The placebo was prepared as follows (final concentration): 0.9% sodium chloride, 0.8% Polysorbate 80, and 0.1% sodium phosphate (diluted to volume with water). The final solution was brought to a pH 4 with hydrochloric acid. The syringe was manually filled with placebo and placed at an elevated temperature condition (50 °C) for 1 week. The syringe was placed in an inverted position to assure that no formulation came into contact with the attached needle. This was done in such a way that only a metals profile of the syringe system and formulation—without any contribution from the needle—could be determined for this discussion and study.
A sample of the ready-to-eat baby food (peach puree) was purchased at a local supermarket and stored at room temperature and at 40 °C in an oven for a period of 3 months. The baby food was packaged in an unspecified type container. This particular product was selected for this study to simulate specific conditions that exist within some pharmaceutical products and container/closure systems (e.g., buffer and salt concentrations, pH). The use of this and other non-drug product or laboratory prepared formulations was based primarily on the fact that, as a contract analytical laboratory, the purchase or use of specific prescription pharmaceutical products meeting the criteria for buffer, salt, and pH conditions was not feasible for this study. Therefore, over-the-counter items that could be easily purchased or prepared within our laboratories were used in this study.
Inductively-Coupled Plasma (ICP) Instrumentation and Conditions
The ICP-MS analysis was performed using a PerkinElmer ELAN 9000 (PerkinElmer Life and Analytical Sciences, Inc. Waltham, MA). Trace metal scans for 64 elements were monitored at various isotopes for each element, when possible.
The ICP-AES analysis was performed using a PerkinElmer Optima 5300 DV (PerkinElmer Life and Analytical Sciences, Inc. Waltham, MA). Trace metal scans for 64 elements were monitored at various wavelengths for each element, when possible.
Standard and Sample Preparation
Custom ICP stock standard preparations containing select element mixtures of the 64 elements were obtained from Inorganic Ventures (Lakewood, NJ). Serial dilutions were made from these stock standards to reach working standard levels at 25 ppb and 100 ppb in 2% nitric acid and 1 ppm and 4 ppm in 10% nitric acid for the ICP-MS and ICP-AES, respectively. Both the ICP-MS and ICP-AES were calibrated using the working standards to prepare a calibration curve. The samples were analyzed and the concentration of each metal was then calculated based upon this calibration. A method blank was prepared for each type of digestion and was subtracted from the appropriate samples during analysis. Samples were digested using the appropriate method depending upon the elements of interest and relative completeness of the digestion. At the conclusion of the digestion, the solution was evaluated for completeness of the digestion based primarily on color or clarity of the solution (e.g., deep yellow or brown solution, undigested material, or particulates). When necessary, additional acid was added to provide a colorless, light-yellow, or straw-colored—and particulate-free—final solution. The sample solution was allowed to cool, quantitatively transferred, and diluted to final volume with reagent-grade water. In the event that particulates or precipitation occurred after dilution to final volume, an alternate digestion technique would have been considered.
The following set of digestion parameters and conditions were developed within our laboratories based upon experience with thousands of different sample types relative to the work performed over time within a contract analytical laboratory. The concept of the decision tree approach began with attempts to find a single-acid solution that would quickly dissolve most APIs. Initially, we evaluated using 1:1 nitric acid/aqueous solutions to dissolve the APIs. While this solution worked effectively on many of these compounds, the net effect was that the acid concentration caused a rapid deterioration of the nickel cones within our Perkin-Elmer 9000 ICP-MS. The concept of an aqueous preparation or use of a microwave digestion technique was cited above. A recent United States Pharmacopeia (USP) stimuli (16) to use ICP techniques to replace the current USP 〈231〉 Heavy Metals also cites the use of aqueous acid solutions, organic solvents, and microwave digestion as suitable sample preparation techniques and provides a limited decision tree approach to sample preparation. Unfortunately, this process is limited only to drug substance, drug product, and excipients. These processes do not include the option to use open-vessel digestion using low temperature (single acid) or high temperature (three acids), nor do they provide specific occasions when one method may be preferred over another. In the case of the high-temperature digestion, the analysis can be both quantitative and semi-quantitative for specific elements. Open-vessel digestion at high temperatures is not suitable for the quantitative analysis of volatile elements such as boron (B), mercury (Hg), arsenic (As), and ruthenium (Ru). It can, however, be useful for the semi-quantitative analysis of these elements on the basis of the complexity of the sample. The use of the three-acid digestion method can therefore be very effective on difficult samples (e.g., rubber stoppers, plastics), where only a semi-quantitative scan of all the elements is needed to provide an initial profile.
The decision tree (Figure 1) proposed here provides a step-wise process for reviewing specific trace metal analysis (e.g., USP heavy metals, volatile elements) and then identifies potential digestion techniques to be used to prepare samples for analysis.
Single-Acid Digestion
Baby food puree was prepared by digesting approximately 100 mg of the sample with 1 mL of concentrated nitric acid heated to 100 °C for approximately 30 min. The solution was diluted to final volume with reagent-grade water (10 mL).
Two-Acid Digestions: Microwave
The pre-filled syringe placebo solution was prepared by digesting approximately 500 mg of the sample with 1mL each of concentrated nitric acid and sulfuric acid using a CEM MARS™ microwave digestion system (Matthews, NC) and closed-digestion vessel. The solution was diluted to final volume with reagent-grade water (50 mL).
Type I glass vials, weighing approximately 5 g each, were digested using a 1:1 mixture of concentrated nitric acid and hydrofluoric acid (HF) and the microwave digestion system at 180 psig. The digested solution was not neutralized prior to dilution to final volume (50 mL).
Three-Acid Digestions: Open-Vessel
Rubber stoppers, HDPE bottles, polyester bottles, and the unspecified packaging container were all prepared similarly by digesting 100 mg of material in a mixture of concentrated nitric, sulfuric, and perchloric acids in an open-glass digestion vessel heated to reflux for approximately 15 min. The solution was diluted to final volume with reagent-grade water (10 mL).
The pre-filled syringe (barrel and plunger tip) was completely digested in a mixture of concentrated nitric, sulfuric, and perchloric acids in an open-glass digestion vessel heated to reflux for approximately 15 min. The digest was then monitored at the conclusion of each cycle for any remaining particulates. Additional acid was added when necessary and the heating cycle repeated until the syringe/plunger tip was completely digested. The resulting solution was filtered prior to analysis as a precaution to assure that the ICP introduction system did not clog.
Results and Discussion
The comprehensive approach (complete digestion) lends itself to a vast and diverse set of packaging materials, pharmaceutical products, and trace metals. Current USP general tests and assays such as USP 〈381〉 Elastomeric Closures for Injections, USP 〈660〉 Containers—Glass, and USP 〈661〉 Containers—Plastics do not accurately address the entire trace metals profile. Current USP stimuli to USP 〈231〉 Heavy Metals, cited previously, only addresses 31 elements relative to the 64 elements discussed within this paper. In the case of elastomeric closures, samples are extracted by autoclave or by other prescribed solvents. The resulting solution is then measured for turbidity, pH, heavy metals such as lead (Pb), and total extractables based on residue weight. Glass and plastic containers are evaluated for heavy metals such as lead (Pb) and non-volatile residues. These procedures are limited in their selectivity of specific trace metals and in their effectiveness, in terms of the overall extraction efficiency. This data cannot provide the same level of information as a complete trace metals profile. Simply extracting materials with a series of solvents may not provide the entire profile of trace metals. Using this comprehensive approach, however, a formulator could evaluate or screen a series of packaging or container materials for total metals content. This knowledge is important because formulation decisions can be made regarding the compatibility of the packaging or container material with the product based upon known or potential interactions.
The digestion decision tree (Figure 1) illustrates a systematic approach to digesting and analyzing various types of materials for trace metals using simple to complex acid digestions. The decision tree provides the analyst several options as to how to proceed with the sample preparation based upon the metals of interest or the nature of the material. Compounds that contain fluorine are digested using a polymer–microwave vessel based on the potential generation of HF acid during the digestion. Samples containing volatile elements, such as arsenic and mercury, can be analyzed using a simple single-acid digestion when the sample either dissolves readily in nitric acid or digests completely at low temperatures (e.g., nitric acid heated to 100 °C). Alternatively, volatile elements could be digested by microwave in a closed vessel. Finally, samples that are not fully digested by either the single-acid or two-acid methods are then digested using a strong oxidizing three-acid mixture. In each case, other oxidizing reagents, such as HF or hydrogen peroxide, can be added to each digestion to aid in the complete digestion of the material or stabilization of the elements in the final digested solution.
In the first examples provided, a sample of a Type I glass vial (Table I) and a butyl rubber serum stopper (Table II) were evaluated using this decision tree approach. Clearly, glass and rubber cannot simply be dissolved by a single acid. An attempt was made to digest the butyl rubber using nitric acid; however, the digestion was not effective, with little or no degradation of the material after several attempts. These materials were then completely acid-digested as described in an earlier section and analyzed semi-quantitatively by ICP-MS.
The data analysis revealed significant levels of elements, including aluminum (Al), boron (B), barium (Ba), calcium (Ca), iron (Fe), potassium (K), magnesium (Mg), sodium (Na), silicon (Si), strontium (Sr), titanium (Ti), zinc (Zn), and zirconium (Zr).
Using this type of comprehensive metals profile, selected metals of interest can be monitored, and the tedious process of tracking the entire series of 64 various metals can be avoided. An example of this is provided in the study of a laboratory prepared pre-filled syringe drug product (Table III). After completely digesting both the syringe and a non-exposed placebo drug formulation (control), an initial profile of both was determined by performing a semi-quantitative ICP-AES scan, with a select few elements identified for further monitoring. The QL and detection limit (DL) for each of the elements was estimated using the standard deviations of replicate measurements of the method blank (Tables IV–V). Slight variations in the DL and QL can be observed in the tables. These differences, in the cases of elements such as boron (B) and arsenic (As), may be attributed to open digestion in glass-digestion vessels and the volatility of boron and arsenic at high digestion temperatures. Other slight changes are most likely due to the effect of the composition of the acid matrix relative to the specific elements and the ICP conditions used.
Most importantly, this semi-quantitative scan was useful in eliminating 75% of the elements which had similar syringe and control levels. Consequently, the stressed placebo formulation was then digested using a microwave digestion technique as previously described. This data (Table VI) indicated that only aluminum (Al) appeared to leach from the syringe into the placebo drug product formulation. Based on this information, the list of elements for this product was reduced even further and only 1 element of interest required validation.
An element-specific validation was conducted in accordance with ICH Q2(R1) Guidelines (17). It was shown to provide accuracy and precision for aluminum using method conditions and a digestion technique as shown in the work above (Table VII).
The semi-quantitative methodology applied above for the ICP-AES analysis has been demonstrated for accuracy and precision (Table VIII) within our laboratories. The microwave method was qualified, using caffeine spiked at approximately 200 ppb of each element in solution. In a few cases (e.g., Ca, P, Si, and Se), the spike recoveries exceeded the acceptance criteria of between 80 to 120% recovery. All elements had precision of < 10% relative standard deviation (% RSD), with most elements at or below 4% RSD. This acceptance range for recovery is within the recently published stimuli to the USP for USP 〈231〉 Heavy Metals (18). In the cases where elements exceed the acceptance range, studies with spiked samples of the actual drug product or material were performed to determine any specific matrix effects or recovery issues.
The classic approach to extractables and leachables may be inappropriate for use because some metals have limited solubility in water or other solvent systems. Therefore, the proposed comprehensive approach has an advantage over the classic approach. Metals such as palladium (Pd), platinum (Pt), and tin (Sn) are used extensively as catalysts. The inorganometallic forms of these elements, however, are relatively non-toxic versus the more soluble salt forms, or organometallic forms. Depending on the solvent system selected, extraction of these metals may not occur. This is not to say that the extraction of materials with different solvent systems is not useful. In fact, these experiments indicate levels of metals that can potentially leach into a specific pharmaceutical product. In most cases, however, the specific metals and initial levels contained within the packaging or container materials are unknown. This type of advanced knowledge regarding potential elements and levels might be useful in setting vendor specifications for packaging or container materials.
Examples of this concept are found in digestion versus extraction studies of HDPE bottles (Table IX) and polyester/polycarbonate bottles (Table X). For example, a significant amount of titanium (Ti) is present in the HDPE-digested bottle (Table IX), yet none of the three extraction solvents indicated levels of titanium (Ti). Also, levels of aluminum (Al) and zinc (Zn) are reported in this material. The digestive approach presents the level of aluminum found in this type of container well below both the oral and parenteral daily limits discussed above. Therefore, the need to perform either the extractable or leachable study may not be necessary to the drug product formulator when using this approach. In the case of zinc, however, the digestive value reported does exceed the daily limit, and so both an extractable and leachable study of this material can be considered necessary.
Similarly, levels of antimony (Sb), calcium (Ca), zinc (Zn), and titanium (Ti) were reported when the materials were digested in the laminate bottles (Table X). The extraction buffer (2% citric acid), while providing similar values for several elements, may not provide a clear indication of the potential for antimony (Sb) to be present in future product formulations when using this type of polyester bottle. Calcium (which does not present a toxicity issue) may complex with the product, consequently changing the solubility parameters, stability, and potentially affecting efficacy.
Another example of this approach is provided using a commercially available baby food product containing citric acid and packaged in an unspecified type container (Table XI). The baby food puree was selected for this study because it contained a significant amount of citric acid, which has been previously shown in the discussion above to be effective in the extraction of metals from packaging materials. The product was stressed for 3 months at 40 °C and analyzed for metals content by ICP-MS. The data indicate the potential leaching of some elements to the puree from the container; however, the container trace metal content is well below the daily limits for specific elements as previously discussed. Again, using the comprehensive approach, a product formulator can decide on the basis of this type of early metals information not to perform extractable or leachable studies when selecting packaging materials.
A limited accuracy study (Table XII) for the ICP-MS was conducted using the single-acid digestion technique (low temperature) on a few select elements that represent the typical 10 USP heavy metals and five additional elements that are under consideration based upon the USP stimuli discussed above. The accuracy was determined based on spiked method blanks at levels of approximately 1 ppm in the sample (2 ppb for each element in solution). The DL and QL estimates were taken from the standard deviation from a 5-ppb standard solution of each element.
Conclusion
This work demonstrates that taking a comprehensive approach toward determining trace metals in a packaging or container material can be a useful tool when establishing good Quality by Design practices and design space for extractables and leachables. The results demonstrate that determining total trace metals instead of controlled extractables can be useful when evaluating product safety. Specifically regarding pharmaceutical products and closure systems, the total metals content can be established and monitored over time to determine if specific metals exceed pre-established limits or quality attributes. This comprehensive approach can also be useful in evaluating these types of products for the establishment of quality attributes during the early stages of product formulation. Finally, this work shows that specific metals present in some packaging and container materials can leach into various products.
Acknowledgements
The authors thank the following individuals for their help in preparing samples and analysis, and for reviewing and further assisting in the completion of this paper: M. Stecker, G. Graves, and D. Lee.
Footnotes
- © PDA, Inc. 2009