Extractables are derived from a variety of sources and exhibit extensive chemical diversity. Primary sources of extractables include:

As noted above, the chemical diversity of extractables is significant. For example, the chemical additive category antioxidants includes hindered phenols, secondary aromatic amines, hindered amines, organosulfur compounds, organophosphorus compounds, and other chemical classes.

Extraction is a process of treating a material with a solvent to remove soluble substances. Extraction is a complex process influenced by time, temperature, surface area to volume ratio (i.e., stoichiometry), extraction medium, and the phase equilibrium of the material (3). There are two reasons why an extraction study is the necessary and appropriate means of accomplishing the various objectives of an extractables assessment. First, there is no other viable analytical alternative. Characterizing a material for potential leachables in its natural solid state is a goal of modern analytical chemistry rather than an accomplishment. Second, even if a direct characterization could be accomplished, it would at best only establish the identities and levels of chemical entities present in the material, and not assess the leaching characteristics of these chemical entities. A compositional assessment does not take into account any chemical reactions that can alter the molecular structures of potential leachables over the dosage form's lifecycle. For example, in the case of phenolic antioxidants that are leached by an aqueous drug product, hydrolysis and oxidation products can accumulate in the drug product. The only viable means for producing data related to leaching is to use a process such as laboratory extraction that is mechanistically similar to leaching.

The design of an extraction study is dictated by the purpose of the extractables assessment and the question(s) being asked, as well as the available information regarding the chemical composition of the test article(s) to be extracted. Extraction studies can be designed to answer questions such as:

Addressing each of these questions requires a particular set of parameters, such as extraction time, extraction temperature, extracting solvent, extraction technique, sample surface area to extracting solvent volume ratio, etc. Clearly, the intent of the extractables assessment must be established before the study design is finalized. For example, if the purpose of the extraction study is to simulate a worst-case leachables profile, then the study can be termed a simulation study that should produce an extract that:

The concept of a simulation study is addressed in greater detail in Assessment of Drug Product Leachables Associated with Pharmaceutical Packaging/Delivery Systems 1664. The means by which an extraction process is accomplished are reflected in the juxtaposition of several experimental parameters including:

Extraction processes have been described as “accelerated”, “aggressive”, “exhaustive”, “vigorous”, “harsh”, and so on, and for medical device studies certain of these terms have been defined (7). In general, extraction processes should allow completion in a reasonable time frame but should not be so aggressive that they alter the qualitative and/or quantitative nature of the resulting extractables profile. The most aggressive extraction conditions are reserved for the quantitative determination of chemical additive contents in components and materials. Because such studies are intended to quantitate specific known chemical additives and not to simulate a drug product leachables profile, it is acceptable to use extraction conditions which disrupt or dissolve the component or material being extracted, and thus to alter the resulting extractables profile, while recovering the target additive(s) without loss or chemical decomposition.

Of all the parameters involved in generating the extract, the extracting medium is the most critical because it is the extracting medium that accomplishes the extraction, and all other parameters merely facilitate the extraction. Establishing and justifying the extracting medium (or media) is both straightforward strategically and complex tactically. Strategically, if the purpose of a particular extraction study is, for example, to simulate a worst-case leachables profile, then the ideal situation is for the extracting solvent to have a similar or greater propensity to extract substances as the formulation, thus obtaining a similar qualitative and quantitative extractables profile. This is clearly stated in regulatory guidances and best practice recommendations (1, 4–6). Therefore, the most logical tactic for this simulation study is to use the formulation itself as the extracting medium and in the absence of complicating factors, such an approach is recommended. However, in certain cases the use of the formulation as an extracting medium complicates extract characterization to such an extent that it is impractical. The various guidances and recommendations suggest that if the use of the drug product as the extracting solvent is not feasible, then the drug product vehicle, or placebo, could be used as an effective extracting medium. This recommendation is derived from the fact that the drug substance itself does not typically create the “leaching power” of a drug product but rather that it is the formulation's ingredients (drug product vehicle) that establish the drug product's ability to leach substances from a contacted material.

When circumstances require that an extraction study must be accomplished with a simulating solvent(s), it is necessary to establish and justify the composition of these solvents. In order to accomplish this objective, one must consider all the physicochemical characteristics of a formulation and/or simulating solvent that influence its “extracting power”. In certain circumstances, the formulation is sufficiently simple that the critical characteristics can be readily delineated and simulated. For example, the extracting power of polar aqueous drug products consisting of soluble ingredients (such as an injectable with a drug substance, buffers, and diluent) is, for organic extractables, driven primarily by drug product pH. In such a circumstance, simulating the drug product pH with an analytically viable buffer system for the extraction study may be appropriate and justifiable. For inorganic extractables, utilization of a simulating solvent having similar metal-chelating properties as the drug product vehicle may also be appropriate and justifiable. It may also be the case that largely non-polar drug products can be readily simulated with analytically expedient organic solvents. For example, chlorofluorocarbon and hydrofluoroalkane propellants used in metered dose inhalers (MDIs) can be simulated with dichloromethane as an extracting solvent and isopropanol can be used to simulate ethanol, a common co-solvent in MDI formulations.

Many drug products are compositionally intermediate between the polar and nonpolar examples just discussed. Examples of such products include “aqueous” drug products that contain stabilizers, solubilizing agents, chelating agents and buffers, lipid-containing products, and biotechnology products containing proteins, peptides, and blood-derived products. Such products have a characteristic polarity which establishes their “extracting power”. Thus, an appropriate simulating solvent will have a polarity that matches that of the drug product. Binary mixtures of miscible solvents (such as alcohol/water) have been utilized as simulating solvents for these types of drugs product.

It may be that a single simulating solvent cannot be established and justified for a specific drug product, or that the focus of the extractables assessment is a material or system that will be characterized for use with multiple, compositionally diverse drug products. In such circumstances, the drug product's ability to leach chemical entities from a packaging system can be established based on the use of multiple extracting solvents, each of which addresses one (or more) of the extracting “mechanisms” that are relevant to the drug product (or drug products) under investigation. The use of multiple solvents is consistent with industry-driven best-practice recommendations for drug products that have a relatively high risk of dosage form interaction with the packaging system and a relatively high safety risk relative to the route of administration (e.g., inhalation aerosols and solutions, injectables and injectable suspensions) (1). Therefore, the use of multiple solvents (or extracting media) with different polarities, pH, ionic strength, or extracting powers, is recommended for high risk dosage form packaging system components and materials requiring extraction studies in order to simulate a drug product leachables profiles (see Table 1). If the goal of an extractables assessment is materials characterization, then simulating a drug product vehicle is both unnecessary and undesirable since this goal requires qualitatively and quantitatively efficient extractions. Such extractions generally are only achieved with relatively powerful organic solvent systems capable of softening, swelling, or dissolving the material's polymer matrix and releasing quantitative levels of additives and other chemical entities.

If the goal of an extractables assessment is materials characterization, then simulating a drug product vehicle is both unnecessary and undesirable since these goals require qualitatively and quantitatively efficient extractions. Such extractions generally are only achieved with relatively powerful organic solvent systems capable of softening, swelling, or dissolving the material's polymer matrix and releasing quantitative levels of additives and other chemical entities.

Extraction time and temperature are critical factors in the extraction process. Although the nature of the extraction solvent establishes the magnitude of the extraction (i.e., the amount of substances that can be extracted from a material at equilibrium), the combination of extraction time and temperature establishes the magnitude of the driving force and the degree to which equilibrium is actually achieved. In a simulating extraction study the purpose of elevated temperature is to increase the extraction rate, so that a short experimental time may simulate longer leaching times (1, 5–9).

Because extraction is a diffusion process, the relationship between the diffusion rate and temperature can be expressed empirically by the Arrhenius equation. The mathematics involved in a process that is driven by Arrhenius kinetics have been established in ASTM F1980-07 (2011) Standard Guide for Accelerated Aging of Sterile Barrier Systems for Medical Devices (10), which may be a useful guide for establishing accelerated contact conditions. As with all such models, the proper use of this model requires an understanding of the model's basis and essential principles, assumptions, and limitations (2).

Extractables profiles obtained with a given extracting medium and extraction technique can and should be monitored for equilibrium or the attainment of asymptotic levels of extractables (see Figure 1).

Extraction stoichiometry considers the physical mass and/or surface area of the test article relative to the volume of the extracting medium, and the actual physical state of the material when it is extracted. Extraction stoichiometry can be manipulated to facilitate production of a more concentrated extract. For example, consider the case of a rubber stopper for a vial that contains 5 mL of a liquid drug product. A more concentrated extract than the drug product (i.e., an extract that contains higher levels of extractables than the leachables level in the drug product) could be produced by extracting 20 stoppers in 200 mL of extracting solvent. Another aspect of extraction stoichiometry is the physical state of the test article. It is not uncommon that components or materials are cut, opened, ground, or otherwise altered in size or configuration prior to being extracted. For inhomogeneous or layered materials such as film laminates, the process of cutting or grinding prior to extraction may alter the extractables profile as it may provide a means for the extracting solvent to come into contact with (and thus more effectively extract) materials (layers) that are shielded from contact with the solution under normal conditions of use. One can argue that the use of such sized material further facilitates the extraction process, however it is possible for sizing to reveal extractables that might not appear as leachables. Nevertheless, some sizing of components or materials before extraction can be useful in certain situations and for certain purposes, including: (a) reducing sample-to-sample variability by the consistent preparation of ground homogeneous polymeric material; (b) reducing the sizes of large packaging components to allow use of standard laboratory glassware for extraction studies; (c) increasing the surface area of a packaging component or material test article (e.g., via extruding, pressing, or grinding) in order to increase extraction efficiency. In any event, careful consideration should be given to the effect of physical sizing of test articles on the extractables profile before such sizing methods are employed in extraction studies.

For extractables assessments involving components or materials whose chemical ingredients are known based on information from the supplier or fabricator, analysts can manipulate extraction stoichiometry based on the known levels of chemical additives and the known sensitivity of the analytical technique(s) that will be used to characterize the extract. For example, consider the formulation for a peroxide-cured ethylene-propylene-diene-monomer (i.e., EPDM) gasket from an MDI valve shown in Table 2.

Such information, when available from component and material suppliers, can be useful in designing an extraction study.

Analysts can also base the extraction stoichiometry on established safety thresholds for leachables. For example, an exposure of 0.15 µg/day total daily intake for an individual organic leachable has been proposed as an SCT for inhalation drug products, also termed orally inhaled and nasal drug products (5). Leachables present at or above the SCT, in an MDI for example, should be analytically and toxicologically evaluated, suggesting that extractables assessment also should be guided with the SCT in mind. The application of thresholds such as the SCT and AET to leachables assessments is discussed in greater detail in Assessment of Drug Product Leachables Associated with Pharmaceutical Packaging/Delivery Systems 1664.

In summary, extraction stoichiometry (and thus the “sensitivity” of an extraction study) can be based on:

An extraction can be accomplished in a variety of ways. It is necessary that the means of performing the extraction match the objectives of the extractables assessment. Common laboratory extraction techniques include:

Each of these extraction mechanisms/techniques has its own unique advantages and limitations. For example, reflux extraction is very efficient, but may be too harsh for certain applications and can lead to thermal decomposition of certain organic extractables; the extracting power of sonication can be difficult to control; and because of its relatively high boiling point, water performs poorly in reflux and Soxhlet but well in a sealed vessel.

If the goal of the extractables assessment is identification and quantitation of the chemical additive content of a component or material, it is typical to use extraction techniques and processes that soften, swell, or dissolve (or in the case of inorganic extractables, digest) the component or material, thereby releasing quantitative amounts of chemical additives for analysis.

Not all drug product or material-contact situations are solution mediated and not all issues related to leaching of material-derived entities involve a solution phase. For example, doses of inhalation powder contained in a capsule or blister pack for use in a dry powder inhaler may have volatiles leached from the capsule or blister material, or by specialty surface additives such as mold-release agents; a solid oral dosage form could contain volatile leachables derived from the adhesive of a paper label affixed to the plastic bottle that contains the dosage form; and inhalation solutions packaged in low-density polyethylene containers could contain volatile migrants from tertiary packaging or auxiliary components such as wooden shipping pallets. In the latter two cases, chemical entities can migrate through the plastic containers and volatilize into the airspace, subsequently accumulating as migrants in the dosage forms.

Extraction techniques specifically designed for application to volatile organic compounds are usually directly coupled to analytical instruments. These extraction techniques include headspace analysis (as headspace gas chromatography; HD/GC), direct thermal desorption (usually coupled to gas chromatography; TD/GC), and thermogravimetric analysis (TGA/GC).

Once an extract has been generated, the next objective is to perform a thorough chemical characterization of the extract. Setting a threshold (as described above), which is a specified level of an individual extracted chemical entity which requires characterization, can be based on safety considerations such as the SCT; functional considerations including nominal levels of known chemical additives in the formulation of an extracted component or material; or technological considerations such as the known or determined sensitivity of an analytical technology, instrument, or method. The extract characterization phase of the extraction study must enable the realization of the overall goals of the extractables assessment.

The ultimate objective of thorough extract characterization as defined above cannot be realized in all cases, even when state-of-the-art analytical chemistry is practiced with best available skill and diligence. It is a reality that there is no analytical technique or combination of analytical techniques that is capable of the discovery, identification, and quantitation of any and all organic and inorganic extractable chemical entities known to science. In some cases, authentic reference compounds for organic extractables may not be available for confirmation of identifications, or for quantitative instrument calibration. Thus, the practical objective of extract characterization must therefore be an exercise of due diligence in the discovery, identification, and quantitation to a reasonable degree of scientific certainty of all individual extractable chemical entities present in an extract above a specified level or threshold.

Extracts can often be analyzed directly without significant preparation or concentration. Many organic solvent extracts (e.g., dichloromethane, ethyl acetate, hexane) can be directly injected into a gas chromatograph, while others (e.g., methanol, ethanol, isopropanol) are either too reactive in the heated GC injection port or too high boiling. Organic solvent extracts with inappropriate physical/chemical properties for direct analysis by GC can be switched to more appropriate solvents. Certain extractables, such as fatty acids (e.g., palmitic acid, stearic acid) perform better in gas chromatographic analysis when they are derivatized to either methyl esters or trimethylsilyl esters.

It is usually considered inappropriate to directly analyze aqueous extracts by gas chromatography, due to the reactivity and high boiling point of water. In addition, pH-buffered aqueous extracts contain nonvolatile salts which are not suitable for GC injection. Aqueous extracts are typically back extracted with an organic solvent to remove organic extractables from the water, with the resulting organic extract being injected into the GC. Unlike GC, liquid chromatography (HPLC, LC/MS) is perfectly suited to the direct analysis of aqueous extracts, since most HPLC methods include water and water-miscible mobile phases. Water-immiscible organic solvents (e.g., hexane) cannot be injected onto these reversed-phase HPLC systems, so these must be dried and the resulting extractable residue taken up in a solvent suitable for HPLC (e.g., acetonitrile, methanol, or mixtures of these with water).

Organic or aqueous extracts with insufficient levels of extractables for analysis can be concentrated by various techniques. Many organic solvents can be dried down under inert gas, a rotary evaporator, or a Kuderna-Danish concentrator. Aqueous extracts can be lyophilized, concentrated under vacuum, or back extracted into an organic solvent which is then further concentrated.

The final concentration at which an extract is analyzed depends on the goal(s) of the extractables assessment and the inherent sensitivities of the analytical techniques applied. A good “rule of thumb” is that in order to accomplish a complete structural analysis of an unknown extractable, a GC/MS requires approximately 5 ng injected into the instrument. This suggests a concentration in the injected extract of 5 ng/µL or 5 µg/mL. In a 200-mL dichloromethane extract, this converts to a total of 1 mg of this particular extractable recovered from the extracted test article. If this analyte concentration is insufficient to meet the goal of the extractables assessment, then the following parameters can be optimized:

The completeness of an extractables assessment can only be judged against the overall goals of the assessment. For example, an extractables assessment accomplished solely for materials characterization might include one extracting solvent, one extraction technique, and one set of extraction conditions; along with a materials-based threshold (e.g., 10 ppm w/w). Such an extractables assessment might be considered complete if all extractables above the defined threshold were identified to the confident level (defined above) and quantitated. For an extractables assessment designed to establish a rigorous leachables–extractables correlation for a high risk drug product, where a challenging safety threshold might apply (e.g., 0.15 µg/day; see reference 5), good scientific practice and due diligence requires the following:

In this case, the extractables assessment might be considered complete if all extractables above the defined threshold were identified to at least the confident level, quantitated, and correlated both qualitatively and quantitatively with drug product leachables data (if available) and the known ingredients in the packaging system, packaging component(s), or material(s) of construction.

It should be noted that limited extractables assessments with relatively narrow goals can be accomplished to required completeness with a relatively focused effort. For example, extraction studies designed to quantitate the levels of specific chemical additives in specific packaging components/materials can be done with specified extraction parameters and analytical methods (see Containers—Plastics 661). The reader is also referred to various sources which describe extractables assessments and extraction studies for pharmaceutical applications (2, 5, 6), as well as other general sources which refer to extractables assessments for medical devices (9) and food contact (11).

Reference is also made to compendial chapters in this Pharmacopeia which include extraction studies with specific goals and purposes:

As stated previously, extraction studies are usually designed to produce extractables profiles, which are qualitative and/or quantitative analytical representations of the extractables content of a particular extracting medium. To illustrate the concept of an extractables profile, the following is an example of an extractables study performed for the purpose of material characterization. Extractables profiles are commonly produced by analysis of laboratory extracts by instrumental chromatographic techniques. Figure 2 and Figure 3 show example extractables profiles (GC/MS and HPLC/UV, respectively) from a hexane Soxhlet extract of a cyclic olefin copolymer (COC) material of construction (12). COC materials are used in the fabrication of pre-filled syringes, vials for small-volume parenterals, and bags for large-volume parenterals. The extracts were generated by subjecting approximately 5 g of suitably sized material to 16 hours of Soxhlet extraction with 125 mL of hexane. The resulting extract was spiked with internal standards (details not relevant to this discussion) and was analyzed directly by GC/MS (see Table 4). For HPLC/UV analysis, an aliquot of the hexane extract was reduced in volume and diluted with methanol prior to analysis (see Table 5). Since the purpose of this extractables assessment was materials characterization, extraction study parameters were adjusted relative to an extractables identification threshold of 10 ppm (µg/g). It is clear from these chromatograms that the techniques used are both complementary and orthogonal, illustrating the concept that multiple analytical methods are required to typically elucidate the complete extractable profile.