Moving Forward towards Standardized Analytical Methods for Extractables and Leachables Profiling Studies

Problem Statement

Strategically, the process of leachables and extractables profiling is well understood and standardized, as all practitioners of extractables and leachables profiling use the same general approach. Specifically, extractables and leachables profiling involves two steps: generating the test sample and screening the test sample. For leachables, the test sample is the drug product, properly aged within its conditions of contact. For extractables, the test sample is an extract, generated by extracting the test article with a solvent under defined conditions of contact. In either case, the test sample is screened by appropriate analytical methods to discover, identify, and quantify the substances of interest.

Considering analytical screening for organic extractables/leachables, there is a certain level of standardization in the tactics used. It is generally accepted that a proper screening approach for organics includes the following orthogonal techniques:

• Gas chromatography with headspace sampling (headspace GC) for volatiles,

• “Direct injection” gas chromatography with flame ionization and mass spectrometric detection (GC/FID/MS) for semi-volatiles, and

• Direct injection liquid chromatography with ultraviolet (UV) absorption and mass spectrometric detection (LC/UV/MS) for non-volatiles.

Furthermore, there is some level of standardization in the specific methods used for each one of these techniques. For example, most LC screening methods are based on a reversed-phase separation accomplished with a C18 stationary phase and a mobile phase gradient consisting of an aqueous buffer and one or more organic solvents such as methanol and acetonitrile. Similarly, most GC methods employed in screening use capillary columns packed with a common stationary phase (USP designation G27, 5% phenyl, 95% dimethylpolysiloxane high temperature) and similar oven temperature profiles.

It is at the next level of tactical detail, operational conditions, where standardization is lost, as the specific conditions used by individual practitioners begin to diverge rather significantly. For example, one practitioner might use an LC column of certain dimensions packed with a C18 stationary phase from a certain vendor, while another practitioner might use a column with different dimensions and a different stationary phase that still fits the general C18 classification.

If two practitioners use generally similar methods with different operating conditions to generate a chromatogram (which is the basis of an extractables or leachables profile), their chromatograms might differ, suggesting that their extractables (or leachables) profile could differ. Furthermore, a profile generated by one practitioner is “owned” by that practitioner, and it is rare that profiles are shared between practitioners. These and other circumstances drive inefficiencies and inequities that are inherent in the present system where “everybody tests the same test articles with their own methods”, and it is eliminating these inefficiencies and inequities that drives the desire to standardize analytical methods and processes.

The Challenge of Standardization

Despite the substantial gains in productivity, consistency, and quality that could be derived from standard screening methods, there is little enthusiasm for developing, justifying, and qualifying such standard methods in the scientific and/or regulatory community. This situation arises due to the complexity of task, the inevitable practical issues (such as who has the authority and ability to establish standard methods, who has the authority to encourage or mandate the use of standardized methods, who will maintain the standards so that they continue to reflect the best scientific practices, etc.), the perception that “it is not possible to standardize methods because every application is different”, the fear that “the standardized method might not be the best method”, and the associated business and intellectual property implications.

I will consider the “who” questions later. Considering perceptions and fears, there are two realities that mitigate these concerns. Considering the fear that the standardized methods might not be the best available methods, I suggest that there is no best method and if there was a best method, it would be only incrementally better than the others. In this circumstance it is inappropriate to hold up progress, reflected in the adoption of standard methods, while we search (in vain) for perfect methods. Given sufficient expertise and resources, a team of analytical professionals could devise, justify, and qualify standard analytical methods that are “suitable for their intended purpose”. Considering the perception that it is not possible to standardize methods, a review of the scientific literature dealing with analytical methods for screening reveals that the methods differ in subtleties and agree in generalities. This suggests that not only that standardized methods can be identified and qualified, but that so doing is a process of optimization of existing methodologies as opposed to developing new methodologies. In fact, the pursuit of good science has been a standardizing influence on the industry, and the search for good science by competent analytical scientists has led those scientists to largely the same methods. Bad science, reflected in ineffective or inadequate methods, has been largely eliminated from the landscape, as the practice of bad science leads to undesirable outcomes, which in turn eliminates the bad science practices.

Nevertheless, a significant practical issue must be addressed to realize the full benefits of standardized methodology. Even when a standard method is implemented, a standard outcome might not be produced. Although standard methods specify specific operating conditions, they do not specify specific instruments or instrument systems, nor do they specify exact laboratory conditions. Thus when standard methods are implemented on different instruments and in different laboratories, different results might be produced. For chromatographic methods, retention times and peak responses (size and shape) can vary depending on the instrument system used, even when the instrument systems are operated using the same standard conditions. Thus the significant challenge in standardization is less about establishing the standard methods themselves and more about establishing those system suitability tests (and acceptance criteria) that are sufficiently robust that they ensure that the chromatographic outcomes produced are reproducible and recognizable each time the standard method is implemented.

The Full Promise of Standardization: Profiling Enabled By a Knowledge Set

The benefits of standardization notwithstanding, test method standardization is not the final objective; rather, it is the foundation for something more powerful. Ultimately, standardized methods produce data, while decisions about product safety and quality require information. Standardized methods provide a stable and robust platform for generating data. Standardization's full promise is realized when a means of converting data into information is built on top of this stable platform. For example, consider the chromatographic analysis of an extract. The resulting chromatogram contains peaks, attributable to individual extractables, that have certain properties including retention time, magnitude of response, and nature of response. While this is important data, it is relatively useless as information in the context of safety and quality impact assessment, which requires the identity and concentrations of the relevant individual substances. To wit, the process of safety and quality assessment consists of three steps:

1. Use of analytical testing to produce data about individual extractables and leachables.

2. Conversion of that data into information (identity and concentration) that is required input into a safety or quality assessment.

3. Use of the required information to perform the assessment.

Standardized test methods provide a robust and stable platform for generating high-quality data (step 1). While this is significant, the platform's full value is not that it exists but rather what it enables. Thus the full value of the standardized methods is that they are the foundation for the means of converting data to information and information to assessment.

Although the entire population of extractables and leachables is not known (and expands as new materials are used in pharmaceutical applications), comprehensive lists of the more frequently observed extractables and leachables, some containing 500 or more compounds, have been reported. If reference standards could be obtained for these documented substances, these reference standards could be analyzed by the standardized methods, creating the data, such as retention time, magnitude of the response and nature of the response, that is necessary to identify and quantify these same substances when they are revealed in extracts or drug products that are analyzed by the standard methods. Thus the knowledge set created by injecting the reference standards and recording their chromatographic responses provides the means for identifying and quantifying the compounds responsible for peaks in extract or product chromatograms. Because this analysis is performed versus an authentic standard, identifications are verified and concentrations are highly accurate. Furthermore, when such a knowledge set is directly linked to a chromatography data collection system by a knowledge management process, the activities of identification and quantification can be performed automatically (by a fixed and reproducible process) as opposed to manually (in a less fully defined manner with a more variable outcome).

The difference between a standard method and a standard method supported by a knowledge set is the difference between the related concepts of screening and profiling. In screening, one obtains an analytical response and attempts to link the response to a specific compound at a specific concentration. In screening the general outcome is a tentative identification (for example, a match to a reference spectral library) and a concentration estimate (obtained, for example, using an internal standard's response). In profiling, the same analytical response is compared to a “target list” of known extractables generated by the analysis of reference standards. If the response can be linked to one of the listed targets, then the identity obtained is verified (or confirmed) and the concentration obtained is more accurate than an estimate, as it is based on the normalized response of an internal standard.

Although creating a standard of science based on standard test methods tied to a knowledge set and supported by a knowledge management process as an efficient and robust means of consistently identifying and quantifying extractables and leachables is a significant goal itself, the essential concept can be extended further. For example, consider the third step of the safety assessment process: use of the required information to perform the assessment. Although the following is a considerable oversimplification of actual toxicological assessment, in essence this process involves the following:

1. Guided by the compound's identity, an expert toxicologist collects and reviews the compound's available toxicological data to establish a permissible exposure (PE).

2. Based on the compound's concentration in the extract/drug product and considering the dosing of the drug product (and the stoichiometry of the extraction process as appropriate for extractables), one calculates the actual (or expected maximum) exposure (AE).

3. The PE and the AE are compared to establish the potential safety impact.

Thus the primary contribution of the toxicologist is defining the PE. If the knowledge set contains the PE, then it is trivial for the knowledge management system to establish the AE (based on inputs that specify the product dosing and the leachable's concentration in the drug product), compare the PE and AE, and “perform” the toxicological safety assessment (calculating the PE/AE ratio and comparing that ratio to an acceptance criterion).

Lastly, one could envision a knowledge set that includes, in addition to analytical and safety data, information relevant to the genesis of the extractable/leachable. For example, including a description of the source material in a searchable knowledge management system would provide the means of querying the system with questions such as “In what materials do I see extractable x?” or “If I have material y, what can I anticipate the extractables profile to be?”

The Curse of Standardization: The Blind Application of Standardized Methods

Something about the word “standardized” can lead to irrational and unsubstantiated expectations, as if the standardization of a methodology somehow makes it perfect. In fact, method standardization produces a method that is generally suited for its intended use, not perfectly suited for that use. Blindly using a standard method in all applications without proper review of the method's output is a problem just waiting to happen because so doing fails to recognize imperfections in, and shortcoming of, the standard method. In the context of extractables and leachables profiling, critical imperfections in standardized methods include their lack of universal applicability and their lack of comprehensive response.

Due to the chemical diversity of drug products to be profiled for leachables and of extracts to be profiled for extractables, there will be drug products and extracts for which the standard method will not produce an acceptable outcome. While some such problems will surface during robust method qualification, it is likely that not all problems will be captured in this manner. Users of a standard method must therefore be diligent in their time of use assessment of the applicability of the standard method in a particular circumstance and must not fall victim to the overly simplistic thinking that “if the method passed its system suitability assessment, then everything must be OK”. Close and purposeful examination of the method's output is necessary to confirm that a standard method is applicable under a specific set of circumstances.

As the analytical profiling uses complementary and mutually reinforcing standardized test methods, it is relatively easy to assume that the suite of standard methods is sufficiently comprehensive that it produces a “real”, unique, and/or useful response for every possible extractable or leachable. Blind application of the standard methods leads to the assumption that the profile generated by the standard methods is the test article's full and complete profile. Despite the general use of the three complementary techniques discussed previously for extractables and leachables profiles, it is well known and well documented that these appropriate and potentially “standardizable” methods do not capture all the extractables or leachables that may be relevant in certain situations. Thus a necessary step in the analytical process of generating an extractables or leachables profile is considering the possibility that the profile revealed by the standard test method(s) is incomplete or inaccurate.

Lastly, I previously discussed a proposal for using knowledge management to “perform” the mechanics of a toxicological safety assessment. Because that proposal was based on an oversimplified generalization of the toxicological safety assessment process, the proposal facilitates, but does not replace or minimize, the necessary expert review that is the cornerstone of a robust and rigorous assessment.

The Question of Who Will Be the Hero

The effort required to establish, justify, qualify, road-test, maintain, and enforce the standard methods (coupled with a knowledge set and knowledge management process) that support extractables and/or leachables profiling is substantial and is surely the major reason why the benefits of standardization have not been realized to date. Bringing standard methods into the mainstream of extractables and leachables profiling will require a heroic effort on the part of the scientific community, individuals and organizations alike, who are united in their mutual concern for the welfare of people who rely on those life-enhancing and life-sustaining drug products whose registration and ongoing marketing is supported by extractables and/or leachables assessment. The pivotal question is “which individuals and organizations will step forward and make this concept a reality?” Potential answers include (but are not limited to):

• Agencies with regulatory responsibility for safe and effective packaging, manufacturing systems or administration devices (e.g., the FDA);

• Standards-setting organizations responsible for drug product quality (e.g., USP, ASTM, NIST);

• Industry groups and consortia whose mission touches the area of extractables/leachables assessment (PDA, ELSIE, IPAC/RS, BPSA, BPOG);

• Suppliers of materials, components, and systems used in packaging, drug administration, and/or drug manufacturing;

• Packaged drug product manufacturers and vendors;

• Academic and academic-related research organizations;

• Contract research organizations that provide services related to extractables and/or leachables testing; and

• Individual, “interested party” contributors.

Considering the scope and magnitude of the effort required to establish, maintain, and enforce the standard methods, it is likely that the creation of such standards will require a collaborative effort among these various participants. For example, it might be a reasonable expectation that

1. Regulators would support the use of science-based standard methods by their review practices,

2. A standard-setting organization would house and maintain the standard;

3. Industry consortia would collaboratively direct and coordinate the various studies required to establish the standard and populate the knowledge set;

4. Commercial organizations would support safe and effective packaging, administration devices, and manufacturing systems by providing the resources required to perform the necessary studies;

5. Commercial organizations with the requisite capabilities would create and maintain the infrastructure required to populate and use the knowledge set; and

6. Individual contributors would perform the necessary activities of testing, generating, and reviewing analytical results, identifying reliable results, using the reliable results to define the standard methods, and establish the scientific basis of the knowledge base.

In closing, I note that this vision for the extractables and leachables community of science is not the imagination of a single individual but rather is an aspiration shared by many who are active in this field. To a certain extent this concept has already been embraced by certain of the potential participants noted previously, and several organizations have made substantiate progress towards realizing essential portions of this concept. However, I know of no organization that has realized this vision in its entirety, and I know of no organization (or team of organizations) that is working to put the realization of the concept into the public domain so that it can become a universal tool and a standard of practice.

There is a consensus within the pharmaceutical industry that says that when it comes to the means of safety assessing packaging, devices, and manufacturing system components, the industry will collaborate and not compete. This consensus is realized in industry consortia formed for this exact purpose and in standards-setting organizations staffed with volunteers dedicated to doing what is right for patients and what is right by science. This consensus is embodied in the considerable efforts of these participants as they work, either individually or organizationally, to establish standards and standard practices that effectively and efficiently produce the greatest possible benefit with the lowest reasonable effort. The ability of these efforts to produce tangible and meaningful outcomes is reflected in standards, recommendations, and other similar documents that capture and communicate best demonstrated science and practices, such as the USP informational chapters on extractables and leachables (<1663> and <1664>, respectively). Considering this foundation of successful collaboration, is it really so outlandish to propose that the community as a whole come together and collectively accept the challenge of method standardization?

If the scientific community purposefully addresses method standardization in extractables and leachables testing, it should do so to produce the greatest benefit. As noted herein, that greatest benefit is derived when a knowledge set and a knowledge management tool is built on top of the stable foundation provided by the standard methods. Moreover, the greatest benefit is derived when the standard methods, knowledge set, and knowledge management tool are universally available to all practitioners and they are universally supported by those who are responsible for the safety and quality of drug products. As this goal of the universal application of a process that involves knowledge management supported by standard methods has not been realized, there is still a considerable job ahead of us. Let's get to it and get this job done, for the benefit of our industry and the benefit of our patients.