Q. In 2012, we asked you what is your personal vision of the E&L world in 2015 and beyond?
You replied: While the assessing the direct impact of leachables on patient safety will be a well-understood and a well-managed process in 2015, the assessment of secondary interactions will still be very much a work in progress at this time. While it is well-recognized and well-documented that such secondary interactions exist, the state of the science with respect to understanding, predicting and managing these processes is not at a point where the issue of secondary actions can be effectively managed. For example, while one can confidently predict the direct safety impact of leachable x found in drug product y, it is not possible to predict that leachable x will react with the ingredients in drug product y to either make them (a) subpotent or (b) unsafe. At the current time, such incompatibilities are addressed retroactively based on precedent. In the future, such incompatibilities will proactively be predicted based on sound scientific principles.
To what extent were you correct?
The circumstances where leachables could impact product and process attributes other than safety continue to be well-documented. In addition to drug-leachables interactions (specifically with protein and protein-like drug products), circumstances where leachables impact cell growth in bio-manufacturing have been recently documented.
These circumstances are largely unpredictable based on assessment of an extractables or leachables profile. While an expert toxicologist can assess a list of leachables and infer their safety impact (presuming the required information is available), it is more challenging to review a list of leachables and say, “yes, that leachable is going to cause my protein to aggregate” or “that leachable is going to reduce my production yield.”
However, with time and investigation, I believe this situation will improve. In the same way that there are in silico methods for assessing the toxic potential of leachables, there could be rules-based approaches for establishing a leachable’s likely interaction with certain protein domains, leading to adverse effects. Similarly, one can envision a rules-based approach that could establish a leachable’s ability to retard the growth of certain cell lines.
Armed with such tools, investigators will be able to effectively establish the probable product and process impact of leachables strictly on the basis of the leachables themselves (identity and concentration).
As the understanding of these phenomena improves, one can anticipate that such tools will be devised and discussed.
Q.When you now think about what the future of E&L could look like, what springs to mind? You previously mentioned that we know so much today that it is not necessary to be in a testing-centric paradigm moving forward, what do you mean by this?
When you know nothing about a system you are characterizing, you have to perform extensive testing. Years ago, E&L was in this situation. However, now that best demonstrated practices have been established and standardized, many E&L assessments have been performed (and to a limited extent documented in the public domain), much experience has been gained and much good thinking has occurred. Through conferences, publications, and professional societies, this experience and good thinking has been shared with a talented, dedicated and capable scientific community. This has created a situation where nowadays it is rare that one is characterizing a system that is a true “black box.” Rather, one can now approach most E&L assessments with a strong foundation of relevant knowledge. By sharing and leveraging this knowledge, laboratory investigations can be optimized and, in some cases, eliminated.
Q. E&L around manufacturing systems is a hot topic, and an accelerating trend in the pharmaceutical industry is the use of plastic components in systems used to produce an active pharmaceutical ingredient (API) or a finished drug product (FDP), how do you propose companies should select and qualify plastic components used in production systems for pharmaceutical products?
An important point here is the significant difference between selection and qualification. Selection as a process is compatible with the concept of a standard testing protocol because it is a process that involves the testing, evaluation and comparison of candidates. Selection is facilitated if all candidates are tested in the same manner and can be accomplished even if the component’s compatibility with specific process streams under specific conditions of contact has not been definitively established. One could envision a manageably-sized standard test protocol that would provide information that is useful for selection in a vast majority of circumstances.
Qualification, on the other hand, is inconsistent with the concept of a standard testing protocol as the specific purpose of qualification is to establish the component’s compatibility with a specific process stream under specific conditions of contact. A standard test protocol that would cover all potential process streams and all potential process conditions of contact could be so unmanageably large that the burden of performing the protocol could well outweigh the value in so doing.
It seems reasonable to me, therefore, to propose a staged approach as follows:
All materials of construction used in production components should pass the USP <661.1> tests currently reserved for materials used in packaging systems. At the material level, suitability assessment is the same for both packaging and production components.
In certain low risk processes, item 1 above is all the testing that is required for component selection. As the risk increases, item 1 testing is augmented by additional testing.
In high risk processes, item 1 is augmented by extractables profiling of the component using a standard testing protocol.
Medium risk processes would be supported by testing intermediate between the item 1 baseline and the item 3 extractables profiling.
A simple, universally applicable, and justified risk evaluation matrix can be developed to differentiate between the various levels of risk and thus used to drive the required level of testing.
Component selection can be based on the appropriate level of testing as defined in items 1 through 5.
Selected components are qualified (established to be suited for their intended use) based on testing that is consistent with the conditions of component use.
Generally speaking, there may be significant value added if the standard testing required for component selection is performed by the component’s vendor. Generally speaking, the customized testing for component qualification is performed by the component’s user. In certain circumstances, the conditions of standard testing and the customized conditions of use are sufficiently similar that information used for selection may also be relevant to qualification.
Q.Your presentation at the conference is on the promise and challenges of standardizing the analytical methods used to screen extracts, why do you see a need to standardize these methods in the industry?
My presentation focuses not only on standard methods but, more importantly, standard methods supported by E&L databases. Let’s be honest with ourselves. For organic extractables, the general community of E&L practitioners has adopted a screening paradigm that leverages three methodologies, headspace GC/FID/MS for volatiles, “direct injection” GC/FID/MS for semi-volatiles and LC/UV/MS for non-volatiles. When you look across methods implemented in various testing laboratories, you see that the individual methods are very similar; so similar, in fact, that one could envision that everyone could use essentially the same method (or more correctly, everyone could use very similar methods that have the same key performance characteristics). While this could lead to benefits such as getting the same profiles across laboratories, this is only part of the story. In addition to the standard methods, let’s say we have tested a large number of known extractables/leachables with these standard methods and recorded key response attributes like relative retention time, key spectral features, accurate mass, response factors, etc. Now we inject an extract, see a peak and match it to a peak in our database. If there is a match we get a confirmed identity and an accurate concentration estimate. In fact, since the method is standardized and we have shared the database with every testing laboratory, anyone who tests the same extract will get the same result.
If everyone is using standard methods supported by the same database, then why is everyone testing the same material or components for themselves? Could we envision a future where a plastic material or component is extracted by a standard extraction protocol and where the extracts are tested by standard methods (supported by universal databases) once and then the information resides in the public domain? While such information would not eliminate customized E&L testing, it would eliminate largely redundant preliminary screening activities, thus freeing up precious resources and expertise to focus on the “real” issues and on innovation.