Originally published May 1997
TECHNICAL PAPER SERIES
Elastomers are widely used in drug-delivery systems, primarily as seals or tubing. Seals range from septa on the tops of injectable drug-delivery containers to critical seals in pressurized, metered-dose delivery canisters. The primary function of seals in this context is to isolate the drug from the external environment, but they also play a key role in metering valves, syringes, and control valves for intravenous delivery systems. Natural rubbers were commonly used in the early years, to be gradually replaced by synthetic elastomeric compounds. In some cases, these synthetic elastomer formulations were "borrowed" from other applications in which their longevity and performance had been well established.
As advances in drug purification and potency have allowed considerable extension of shelf life, the long-term requirements of elastomers have been pushed significantly, creating the potential for difficulties. For example, long-term interaction of a drug with an elastomer may lead to patient complaints about odors with inhalable medications. Pressure loss through seal leakage in pressurized delivery systems can influence patients to try competing brands. Deactivation of medications by organic and inorganic constituents of elastomeric compounds can occur when a seal elastomer is considered in its generic state and a detailed review of the specific formulation is not offered by the manufacturer. Perhaps most critically, unanticipated contamination of drugs can result in serious allergic reactions.
This article presents several case studies in which elastomeric components were found to have a negative influence on the performance of pharmaceutical packaging and/or drug formulations. The forensic methods used to identify the root cause of these problems and the corrective actions implemented to prevent reoccurrences are also discussed. The case studies are offered in the hope that this information might contribute to the growing awareness of the need to consider the broader aspects of polymeric systems for drug interactions. In fact, the pharmaceutical industry has responded aggressively in addressing the potential problems discussed in this article.
All of the cited case studies involve elastomeric materials. The test methods described in the following paragraphs were used to evaluate some or all in order to help uncover the root cause of these events. Although additional test methods were also used, they are not discussed in detail since they did not provide key information in these investigations.
Gas chromatography and mass spectrometry (GC/MS) analyses were used to determine selected formulation details of the elastomeric compounds. In most cases, thermal-desorption sampling of the compounds was carried out. A Soxhlet extraction apparatus with ethanol as a solvent was used to reveal the extractable contents in a seal from one of the valve delivery systems. This method was considered a more reasonable representation of the in-service extraction, since ethanol was part of the medication formulation.
Representative sections from seal samples were examined using a scanning electron microscope (SEM). These sections were prepared by freezing small pieces of each sample in liquid nitrogen, followed by rapid mechanical sectioning to minimize plastic deformation of the sample surface. This method is useful in evaluating the quality of mixing of the elastomer, the size and morphology of the filler particles, and the bond between the fillers and the matrix. Voids, contaminants, or other manufacturing anomalies can also be detected using this method. An energy-dispersive x-ray spectrometer, which is an attachment to the SEM, was used to identify the inorganic filler particles and contaminants located within the elastomer matrices. In one case, tissue samples were analyzed using this method to determine their elemental composition.
Polarized optical microscopy was employed to determine the optical properties and physical morphology of contaminants and functional ingredients for elastomers, drug formulations, and, in one case, biopsied tissue samples. The polarizing microscope uses polarized light interaction with an optically active specimen to provide measurements of key optical properties such as refractive index, birefringence, crystallographic structure, dispersion of refractive index, and others.1,2 Fitting the polarizing microscope with a programmable thermal stage permits observations of crystalline materials at elevated temperature, allowing for precise observation of melting point, decomposition temperature, phase separation, and other characteristics.3,4 Since the combined method is based on microscopic observation, trace analyses are emphasized.
To evaluate the mechanical performance of elastomer seals in valve delivery systems, a custom-designed fixture used on a standard load frame was developed to measure the dynamic stem displacement versus load characteristics of the valve. Valve performance has a subjective aspect, in that patients come to rely on the feel of a delivery system and, since their health is at stake, are accustomed to questioning a valve that feels different from one they are accustomed to using. Each valve was actuated using a haversine waveform 200 times to simulate typical usage of a metered-dose inhaler. For each cycle, the load versus stem displacement was dynamically monitored. Analysis was performed on representative hysteresis loops at the beginning and end of the valve life, evaluating the function of the valves and seals over time in order to determine the cause of varying performance among manufacturing lots.
CASE STUDY #1: SEPTUM CONTAMINATION
Patients receiving a postoperative intramuscular analgesic were experiencing a high rate of adverse reactions including, in the worst cases, gangrenous infections. Thorough clinical testing had disclosed no such side effects, and the affected patients indicated no near-term symptoms following administration of the analgesic. In some instances, infections developed many days or even weeks following treatment. Classical pathological analyses were conducted, to no avail. After many traditional investigative avenues had been exhausted, a forensic approach was adopted to examine tissues from affected patients, comparing areas both near and away from the site at which the drug had been administered.
Thin sections were prepared from biopsied tissues. Examination with a polarizing microscope indicated that crystalline "deposits" were present within the affected areas, but absent in the nonaffected areas. Investigators first assumed that these deposits were organic compounds, either products of drug decomposition or impurities from the drug. Optical-property measurements of the deposits showed no correlation with the drug compound or with any postulated reaction products. Thermal microscopy was then used to begin the forensic approach to this problem. Heating of the tissue indicated that the crystalline material exhibited no melting point, even when it was heated well above the tissue's char temperature, indicating that the material was inorganic.
Figure 1. SEM micrograph (magnification 700*) of talc flake within gangrenous muscle tissue of patient treated with injectable analgesic.
In Figure 1, the crystalline material is shown in the tissue, as observed with an SEM. Elemental analysis identified the contaminant as an oxide containing magnesium and silicon. Subsequent application of polarized optical microscopy enabled the measurement of basic crystallographic properties of the deposit, confirming that it was talc.4 A microscopic examination of representative bulk samples of the drug failed to disclose any talc, and upon review, the possibility of talc being introduced from gloves during surgery was discounted. (Talc is a release agent that has been used for "latex" surgical gloves.) However, analysis of drug residues from the injectable vial showed clear evidence of talc contamination. Direct microscopic examination of the container components revealed that the rubber septum on the top of the container was the source of the contamination. The manufacturer had used talc as a separating agent to prevent cohesion of large quantities of the septa following the punching operation in which they were formed from a sheet of cross-linked Buna-n rubber.
CASE STUDY #2: ELASTOMER FORMULATION
During usage of a certain inhalable asthma medication, patients were complaining about a bad odor. Descriptions about the character of the odor varied significantly, from "skunklike" to "burned rubber." Initial characterization of the bulk drug compound disclosed no viable source for the off odor. All starting materials and propellants were carefully analyzed and found to be free of any compound not intended to be present. Inhalers were recovered from patients who had complained about the problem and these devices were subjected to laboratory analysis.
Methods employed to concentrate and analyze traces of accelerants in arson cases were used to examine the contents of the inhalers. These methods included cryofocusing of the gaseous phase as the inhaler was routinely discharged. Comparative GC/MS analysis between good and malodorous lots indicated the presence of a mercaptobenzothiazole compound on the bad lots, even though subsequent reanalysis of the bulk drug compound and starting materials disclosed no such compound. Only when controlled ethanol extractions from the inhaler alone were carried out was the source of the offending compound identified. Specifically, the elastomeric seals used on the metering valve were found to be the root cause. However, discussions with the valve manufacturer and elastomer compound supplier failed to immediately identify the source.
Considerable investigative work ensued, leading to the supplier of the seals. The seal manufacturer purchased the Buna-n elastomer compound from a compounding specialty supplier, provided the basic compound design, and performed the compounding in-house. Postcuring ethanol extraction of the elastomer for removal of residual cross-linking by-products was conducted at the valve manufacturer's facilities. The compound manufacturer disclosed the formulation of the elastomer, and an investigation of the raw materials was initiated.
The elastomer contained 11 ingredients, including two inorganic reinforcements, an antioxidant, a cross-linking agent and associated synergist, stearic acid, an aliphatic oil, a silicone lubricant, and others--none of which, in principle, contained the offending agent. Chemical analysis of the organic raw materials, using thermal-desorption sampling and GC/MS, identified a mercaptobenzothiazole in the Buna-n elastomer. Research into this polymer indicated that it had been developed for outdoor applications, and that the thiazole compound was intended as a UV inhibitor. In this case, the compound manufacturer had selected a base polymer to meet certain engineering requirements, not for chemical suitability during drug contact.5
CASE STUDY #3: SEAL IRREGULARITY
Pressurized medication-delivery canisters are expected to maintain a low leak rate so that patient dosage and product stability are preserved within design parameters for the shelf life of the product. Issues of excessive leak rate have proven to be a challenge to the manufacturers of such devices, and there are various root causes for the related problems.
Figure 2. SEM micrograph (magnification 90*) of inner diameter of packaging seal (punched surface).
Figure 3. Illustration showing interaction of canister seal and stem during valve actuation.
In this specific case, high leak rates were being experienced with a neoprene (chlorinated polyethylene elastomer) seal. The elastomer was reinforced with calcium carbonate and talc, the latter having been added to achieve a target Shore hardness. Although this goal was met, problems inherent to talc made its use less than desirable for this application. The talc did not form a chemical bond with the base polymer, and its high aspect ratio and lamellar structure caused it to attain a preferred orientation within the seal during processing with a calendering mill. When the hole was punched through the gasket to accommodate the stem of the device, an irregular sealing surface resulted. As shown in Figures 2 and 3, a leak path was established and a potential source for particulate contamination of the drug was formed.
Figure 4. SEM micrograph (magnification 50*) showing the surface of a molded seal.
An as-molded, filled-rubber surface from another representative seal is shown in Figure 4. In this case, a good-quality seal can be maintained since the elastomer covers the outer surface reasonably uniformly, as compared with a cut surface in which the inorganic ingredients become exposed and porosity results from separation of the particles from the polymer matrix. Further improvements can be achieved by selecting inorganic reinforcements of lower aspect ratio and controlled particle size.
CASE STUDY #4: ADDITIVE FORMULATION
A series of valves were found to have widely differing performance, based on testing at the manufacturer's site and returned products from patients. Laboratory tests of the mechanical performance of the valves were made and data compared from the beginning and end of the expected valve performance life of 200 actuations. As shown in Figure 5, the frictional forces led to significant sticking of the valve after initial actuation, indicating some type of degradation of the system that resulted in decreased lubricity of the stem-to-seal interface. The nominal formulation and properties of various lots of seals that performed well were compared with those that performed poorly, and no major differences were found.
Figure 5. Representative hysteresis loop for metered-dose inhaler valve showing stem displacement versus load.
All seals had been produced by the same manufacturer, using the same formulation and performance criteria, including tensile strength, elongation limit, and Shore hardness. Elemental analysis was used to compare good and bad lots. The only difference found was the presence of a small concentration of calcium carbonate in the lot that performed poorly. The elastomer formulator was again contacted, and presented a convincing case that no calcium carbonate had been added to the formulation. With the cooperation of the formulator, an investigation of the raw materials was made. It was discovered that, in some cases, the dicumyl peroxide cross-linking agent was of a "supported" variety, supplied predispersed in calcium carbonate, and that the particle size and impurity level of the carbonate were poorly controlled. Substitution with a pure peroxide eliminated the mechanical problems with this valve.
The formulation of elastomers is a well-established art and science that has matured considerably during the past century, with a plethora of new ingredients becoming available in the past 15 years. Since the advent of synthetic rubber and other polymers in the early 1940s, commercial elastomer formulation has followed traditional guidelines. Readily available, basic compounds have been developed for a wide range of engineering applications; the roles of major ingredients are well understood; and a mature supply network has been established. Elastomeric materials tend to be generically regarded as "Buna," "neoprene," "Viton," and so on, despite the fact that the base polymers from which these names are derived may constitute only 50% of the compound by weight. The other ingredients are typically proprietary and, in fact, may not be suitable for all uses. Elastomers for specific applications are often based on commercial elastomer formulations to which selected ingredients are added, as opposed to a fresh compound newly designed from the base ingredients.
Although it is tempting in a technologically advanced field to use sophisticated materials for medical sealing applications, it is often necessary to return to the basics and reexamine the goal in the context of the application. Using this approach, a polymer would be formulated from a minimum number of ingredients, each of which would be as safe for human contact as possible. By reducing the number of ingredients, the controls required of each supplier are greatly simplified and the potential for adverse interactions greatly reduced. Currently, raw elastomers are available with "vitamin E" antioxidants (alpha-tocopherols), and--though many customers are not aware of the fact--virtually all commercial elastomers are supplied with an antioxidant already included.
Cross-linking additives are another source of antioxidants and other chemicals. Cross-linking is a chemically challenging process, typically including decomposition of a reactive peroxide. This requires an added antioxidant to protect the polymer, while reaction products include acetophenone, cumyl alcohol, acetic acid, and others. All reaction products must be removed by postprocessing Soxhlet extraction with ethanol to avoid interaction with the drug formulation. For such cases, radiation-induced cross-linking provides a clean alternative.
Elastomer formulation for medical products cannot be conducted in a generic manner. A significant development effort is required to fine-tune a given formulation, and this recipe is typically proprietary. The end-user, though, should be cognizant of the subtleties of formulations and the roles and sources of key ingredients. The literature contains many valuable references on this subject.69
At most pharmaceutical companies, measures are being adopted to carefully monitor the composition and potential interactions of any elastomers that come into contact with drug products. In other industries, minor changes in the often unnecessarily complex elastomer formulations used for applications such as seals or gaskets can lead to failures or performance deficiencies that are very difficult to trace. In addition to controlling impurities and ingredients in the base compound, medical manufacturers must monitor all aspects of handling. The use of mold-release agents, processing aids that vary from batch to batch, and surfactants and release treatments all deserve keen attention. The importance of considering these materials as they are processed and of working directly with the vendors is paramount, and may contribute significantly to improved products offering reduced risks and more reliable performance.
1. Troger WE, Optical Determination of Rock Forming Minerals, Stuttgart, Germany, E. Schweizerbart'sche Verlagsbuchhandlung, 1979.
2. Winchell AN, The Optical Properties of Organic Compounds, New York, Academic Press, 1954.
3. Chamot E, and Mason C, Handbook of Chemical Microscopy, New York, Wiley, 1958.
4. McCrone W, and Delly J, The Particle Atlas: Edition Two, Ann Arbor, MI, Ann Arbor Science Publishers, 1973.
5. Bhowmick A, and Stephens H (eds), Handbook of Elastomers, New York, Marcel Dekker, 1988.
6. Hoffman W, Rubber Technology Handbook, New York, Hanser, 1989.
7. Schnaebel W, Polymer Degradation: Principles and Practical Applications, New York, Hanser, 1981.
8. Sekutowski D, "Inorganic Additives," in Plastics Additives and Modifiers Handbook, Edenbaum J (ed), New York, Van Nostrand Reinhold, 1992.
9. Gachter R, and Muller H, Plastics Additives Handbook, Cincinnati, Hanser Gardner, 1993.
Joseph H. Groeger is cofounder and vice president of Altran Materials Engineering (Cambridge, MA), which provides industrial consulting and failure-analysis services. He worked for 17 years at the University of Connecticut, where he taught polymer science, microscopy, and forensic science and managed a research department in materials science. He has authored numerous technical papers and textbook chapters in applied polymer science, metallurgy, and forensics.
Leslie M. Compton has worked at Altran since 1993, specializing in failure analysis and product development. She holds a BS in materials science and engineering from the Massachusetts Institute of Technology.