Identification and Qualification of Genotoxic Impurities as Leachables in Drug Products

A potential source of genotoxic impurities (GTIs) arises from chemicals that migrate (leach) from the materials used in packaging the components of the container closure system (CCS) Process-related impurities often have structural similarities to active pharmaceutical ingredients (APIs) and therefore may possess a similar

161 Introduction 447 162 Container Closure System Evaluation Process 448 163 Compendial Testing: United States Pharmacopoeia Biological

Reactivity Tests 451 164 PQRI Best Practices 451

1641 Safety Concern Threshold 451 1642 Analytical Evaluation Threshold 453

165 Analytical Testing Strategies 454 166 Genotoxic Impurities: General Principles 455

1661 Genotoxic Impurity: Definition 456 167 Genotoxic Impurity Hazard Identification 456 168 Establishing Acceptable Daily Intake: Lifetime 458

1681 Consideration for Higher Limits 459 1682 Assessing the Safety of Extractables 460 1683 Assessing the Safety of Leachables 462 1684 Case Study 462

169 Conclusion 463 References 464

genotoxic profile to the APIs Alternatively, leachables have no structural similarity to and typically have toxicity and genotoxicity potential significantly different from the APIs Specific to genotoxicity, chemicals that leach into drug products that are known mutagens and/or human carcinogens (eg, polynuclear aromatics [PNAs] and nitrosamines) may require more conservative levels of control than process-related drug impurities in the final drug product For these reasons, leachables are considered out of the scope of the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) impurity guideline Q3B(R2) for drug products1

Although there is limited regulatory guidance on how to qualify leachable GTIs in drug products, not all of them carry the same risks for leachable toxicity Current guidance from the US Food and Drug Administration (USFDA) provides a table of drug products ranging from high (eg, orally inhaled and parenteral drug products) to low concern (eg, oral tablets) for risk of leachable toxicity as well as general guidance on basic risk assessment2 In lieu of having detailed regulatory guidance, the Product Quality Research Institute (PQRI) developed and published best practice recommendations for the identification and qualification of leachables in orally inhaled and nasal drug products (OINDPs) that established a safety concern threshold (SCT), which is based on the carcinogenic potential of leachables In turn, the SCT is used to calculate an analytical evaluation threshold, which determines the analytical limit for the detection and identification of leachables in OINDPs3 The  concepts developed for OINDPs are being used to develop specific recommendations of parenteral and ophthalmic drug products (PODPs)4 The use of available guidance (official and draft) and the PQRI recommendations are the basis for the identification and qualification of leachable GTIs in drug products discussed in this chapter

The CCS of a drug product typically refers to both its primary and its secondary packaging materials The former include materials that come in direct contact with the drug product (eg, plastic or glass vials, single-use injection system, and metereddose inhaler), whereas the latter are constituents such as labels and secondary/tertiary packaging (eg, fiberboard box, shrink-wrap, multilaminate pouch, and shipping pallets) that directly or indirectly come into contact with the primary packaging As a general rule, extractable and leachable studies performed on the components of a CCS should be as inclusive as possible In following this recommendation, the individual components are first subjected to controlled extraction studies, using solvents of varying polarities tested at elevated temperatures, to determine what chemicals might migrate from the CCS into the drug product over time (ie, the extractables) Subsequently, using the drug product formulation within its CCS, the chemicals that actually do migrate into the product (ie, the leachables) are determined under more realistic shelf life conditions Importantly, extractable and leachable studies should not be initiated until the development of the CCS and the drug’s commercial formulation is well advanced, if not finalized, since a late-stage alteration in either could inadvertently introduce a novel leachable into the final product

Unlike process-related impurities that are inherent to any drug substance or drug product, available regulatory guidance concerning assessing the safety of impurities in drug products originating from the CCS is limited2 Two tables that define which drug products have the highest level of concern for leachables (Table 161) and the type of testing and/or data that will be required to qualify the leachables in a drug product (Table 162) are provided in this guidance A leachable evaluation is not typically required for drug products that are bucketed as medium-low and low-low relative to the route of administration and the likelihood of an interaction between the CCS and the drug product What is necessary to conduct a toxicological evaluation, in particular how to conduct a genotoxic assessment, is not provided in the guidance Second, the guidance places more emphasis on the toxicological evaluation of extractables and not leachables The issue with this process is that all extractables are not leachables Subsequently, and more importantly, the PQRI recommends a toxicological evaluation of leachables as these are the chemicals that will be found in the drug product3

TABLE 16.1 Examples of Packaging Concerns for Common Classes of Drug Products

The point of origin of any impurity in a drug product does not abrogate the need to assess its safety if human exposure to it is anticipated In terms of risk assessment of these chemicals, while determining the safety of leachables is intuitive, consideration of such data for extractables is also recommended In this regard, the former helps to ensure the safety of the drug product, whereas the latter increases the likelihood that the materials being selected for the CCS are unlikely to yield a problematic leachable such as a GTI Thus, concerning the individual constituents of any given CCS a coordinated strategy integrating compendial testing and the examination of safety data for both extractables and leachables has been recommended to ensure both product quality and patient safety5,6 Importantly, however, one must be cognizant of the test outcomes of compendial biological reactivity tests

TABLE 16.2 Safety Considerations for Common Classes of Drug Products

that are not designed to identify and/or manage this type of GTI issue, as described in Section 1641

It is important to note that biological reactivity tests7,8 are required by many health authorities as a part of the safety qualification process of a CCS These studies typically employ a standard extraction procedure, and the resulting mixture is tested in vitro for cytotoxicity or in vivo for lethality, irritation, or sensitization Importantly, however, these in vitro and in vivo tests have limitations, since they are not designed to specifically identify safety issues such as the possible presence of a GTI in the extract being tested Thus, although these tests still have merit, they should be performed in conjunction with controlled extraction and/or leachable studies to more fully and comprehensively characterize the safety profile of the CCS and determine the potential for GTI risk

For identification and qualification of leachables in OINDPs, the PQRI Leachables and Extractables Working Group proposed an SCT of 015 µg/day3,9 Below the SCT concern for carcinogenic or noncarcinogenic toxicity is negligible, and identification of leachables below this threshold generally would not be necessary Carcinogenicity was used as the basis for SCT because carcinogenic effects typically occur at lower levels of intake than those associated with noncarcinogenic toxicity This was previously demonstrated for orally administered compounds, including those with potent neurotoxicity, reproductive toxicity, or endocrine effects10

Unlike the threshold of toxicological concern (TTC) for GTIs addressed by the European Medicines Agency,11 the SCT for OINDPs is linked to an AET Unlike the TTC, the SCT is not a control threshold and in many cases particular leachables can be qualified above the SCT Second, the SCT differs from the TTC in that the total daily intake for leachables in OINDPs was set at 015 µg/day as opposed to 15 µg/ day for process-related impurities In this regard, the 10−6 or 1 in 1,000,000 risk level for OINDPs accounts for multiple extractables or leachable impurities, some with potential genotoxicity issues12

Third, in addition to the nitrosamines and PNAs mentioned previously, Table 163 provides examples of chemicals with known genotoxic potential that have been extracted or observed as leachables from components of the CCS These examples are provided by the PQRI PODP work team that assembled a list of approximately 600 chemicals that have been known to extract or leach from CCS components Fourth, based on an in silico analysis, about 11% of the approximately 600 chemicals have positive structural alerts for genotoxicity The in silico analysis validated that leachables can possess genotoxic potential and established the need for setting the AET at a level that will detect leachables that may have genotoxic potential (PQRI, unpublished data)

TABLE 16.3 Examples of GTIs Found in Container Closure Components

For OINDPs, using a metered-dose inhaler with a solvent-based propellant in its formulation as a worst-case example, the PQRI determined that it is reasonable to use an analytical threshold linked to the 10−6 risk level as a starting point for identification and evaluation of leachable impurities In contrast, this does not seem to be the case for PODP where many of the drug products are simple aqueous-based formulations and fewer leachables (in particular, genotoxic leachables) migrate from the CCS For this reason, the PQRI is currently evaluating the translation of the SCT concept for PODP and is developing recommendations that would set the SCT at a daily intake at 15 µg/day

As said earlier in Section 1641, it is important to understand that the SCT is not a control threshold and that higher limits of leachables, including genotoxic leachables, may be acceptable under certain circumstances (Section 1671)

For the remainder of this chapter, an SCT of 015 µg/day will be used to discuss the AET concept and assessment of leachables in drug products

The AET is derived from the SCT and presents the limit as a concentration in drug products Understanding the conversion of the SCT to the AET is necessary for the analytical chemist for practical application of the threshold concept in the laboratory The maximum dose per day of the drug product is used in this conversion-at its most basic, the AET is the SCT divided by the maximum dose per day In defining the maximum dose per day, it is important to take into account different drug product strengths, patient populations, and indications to ensure that the worst-case scenario is considered3

For example, an intravenous pain treatment may be supplied in 2 mL vials containing 100 mg/mL of the drug product as well as in 1 mL vials containing 500 mg/mL of the drug product A maximum dose of 1000 mg of the drug product could mean that five vials of lower strength (a total of 10 mL of the drug product) or two vials of higher strength (a total of 2 mL of the drug product) are dosed to a patient in a single day Either scenario must be covered in a worst-case AET calculation An SCT of 015 μg/day equates to an AET for a known genotoxic leachable of 0015 μg/mL for the lower strength product (dosed at a maximum of 10 mL/day) or 0075 μg/mL for the higher strength product (dosed at a maximum of 2 mL/day) To ensure proper control of the leachable in the drug product, analytical methodology will be needed to properly detect and quantitate the leachable down to 0015 μg/mL

It is apparent from this example that as the dose per day increases the AET decreases For a large-volume parenteral, it is not unusual for the AET to be in parts per trillion It becomes more critical for these products to have a thorough understanding of potential leachables such that analytical methodology may be targeted appropriately For example, a single ion mass spectrometric monitoring method may provide the necessary selectivity and sensitivity for a known genotoxic leachable The same method, however, will not be sufficient to show the absence of other leachables

To determine material composition, the individual components of a CCS are first subjected to controlled extraction studies, using solvents of varying polarities tested at elevated temperatures, to determine what chemicals might migrate from the CCS into the drug product over time (ie, the extractables) Subsequently, using the formulation of the drug within its CCS the chemicals that actually do migrate into the product (ie, the leachables) are determined under more realistic shelf life conditions

An extractables study can be designed to be either very aggressive, to provide maximum information about the packaging component, or less aggressive to be more predictive of leachables Both types of studies can be valuable to the analytical chemist in learning about packaging and developing appropriate leachables methods For example, a very aggressive extractables study may result in the detection of a large antioxidant such as Irganox 1010, which would not have been detected in a milder extraction The large antioxidant itself may be unlikely to leach into an aqueous formulation; however, it may have smaller breakdown products that are likely to become leachables Understanding the presence of this antioxidant in the packaging will allow the correlation of smaller antioxidant-related leachables to the extractable, even if the antioxidant itself never leaches

A thorough extractables study will include an analysis for volatile, semivolatile, and nonvolatile organic extractables as well as for inorganic extractables Sample handling should ensure that the targets of interest are preserved For example, reflux is a common extraction procedure and it is also common to concentrate extracts to ensure that a low AET is reached However, a reflux that is followed by evaporation to dryness and reconstitution will significantly impact the retention of volatile and semivolatile analytes Instead closed-vessel extraction may be more appropriate for volatiles, and any concentration steps should not be to dryness if semivolatiles are targeted The methodology at this stage is generally nonvalidated screening methods using mass spectrometric detection

A well-designed extractables strategy will need to account for the aggressiveness of the extraction conditions when considering the interface between the analytical chemist and the toxicologist An aggressive study may be more useful in learning about the packaging component and its additives but may be less useful as a direct predictor of likely leachables; at times, aggressive studies may result in the detection of so many extractables that reporting down to the AET becomes impossible It will likely be impractical for a toxicologist to assess all of the extractables in such a study; nevertheless, an assessment of the major extractables or extractables that are expected to become leachables based on their chemical properties (ie, polarity) may still be valuable In such an instance, a less aggressive extraction study, or a migration study designed to be more predictive of leachables, may be needed to bridge the extractables and the leachables work In a migration study, an aqueous/organic mixture such as water/ethanol is used to simulate a drug product’s ability to solubilize leachables3 Temperature conditions are generally milder than those in an extraction study This simulated drug product matrix facilitates sample preparation and analysis and results in a more realistic list of potential leachables

than an aggressive extraction study This data will likely be more valuable for toxicological assessment

An extraction study or a migration study that is considered predictive can be reviewed more closely for the presence of unknown extractables It is often easier to identify an analyte at the extractables stage, in a simpler matrix where the concentration is higher, than in a leachables study An unknown at 1000 times the AET in a water extract is likely a good target for identification work at that stage On the other hand, a small peak in a less predictive solvent may not be worth the analytical investment to identify at the extractables stage

The role of the extractables work is to inform the leachables targets and to look for early toxicological “red flags” The extractables work is an important piece of the overall strategy The ultimate goal is, of course, accurate detection and quantitation of leachables in a drug product These are the analytes that will result in actual patient exposure and ultimately require a complete toxicological assessment Therefore, the leachables methodology is generally validated and incorporated into a stability study to allow testing over the shelf life

Based on the extractables data, and discussions with a toxicologist, the analytical chemist develops and validates leachables methods At times, it may be difficult to reach the AET for all target leachables Depending on the leachables of interest, and the early toxicological assessment, it may be appropriate to validate an analytical method with a higher reporting threshold However, if the toxicologist has indicated an alert for a given leachable, then the sensitivity of the analytical methodology down to the AET will be critical to the success of the leachables study In extreme cases where drug product matrix complexities combined with low AETs make method development impossible, a rationale to use a placebo or a simplified drug product matrix may be justified Ideally, leachables methods will also allow the reporting of unknown or unexpected leachables Depending on the drug product matrix and AET, this will be more difficult for some products than others If the extractables profile is well understood, the extractables data may assist in a justification to exclude the reporting of unknowns in leachables methods Together, the extractables and leachables analytical work should form a comprehensive strategy that ensures that potential patient exposure to leachables is understood and reported to the toxicologist for assessment

Over the past decade, the identification, qualification, and control of API impurities known or suspected to possess genotoxic potential have been subject to increased scrutiny by pharmaceutical companies and regulatory authorities As a consequence, the topic has been the focus of several regulatory guidance documents11,13,14 and numerous peer-reviewed publications Most recently, the topic has been adopted for the development of an ICH guideline15 Although the focus of existing GTI guidance is on the identification and control of GTIs associated with the synthetic process, much of the framework that has been established is also useful to consider for GTIs that leach from pharmaceutical packaging Some aspects of this framework are reviewed in Section 1651

Given that the pharmaceutical synthesis of a drug substance requires the use of materials that are intrinsically reactive in nature, residual impurities may have the potential to react with cellular macromolecules The impurities that have the potential to directly interact with DNA, causing mutations because a mutagenic mechanism of genotoxicity is assumed to possess a linear dose-response relationship (ie, have no threshold for the effect) unless there is sufficient evidence to prove otherwise, are of particular concern As a consequence, limiting human exposures to very low levels is necessary to minimize excess cancer risk Genotoxic chemicals that are nonmutagenic typically have a mechanism of action associated with a threshold16-20 and would not be expected to pose carcinogenic risk in humans as impurities in pharmaceuticals It is important to note that there is compelling experimental evidence to indicate that threshold or sublinear dose-response relationships also exist for some mutagenic carcinogens21-26 However, for mutagenic carcinogens the burden of proof is on the pharmaceutical sponsor to provide sufficient evidence that this is the case for the specific compound of concern

The primary methods recommended for the identification of impurities with mutagenic potential are the observation of chemically alerting features and the Ames bacterial reverse mutation assay11,13,27 This approach is supported by a number of studies in which structurally alerting features and/or mutagenicity in the Ames assay have been shown to be highly predictive of nonthreshold genotoxic carcinogens28-30 Given that it is not possible to isolate and test every potential impurity in the Ames assay, it is often necessary to rely on quantitative structure-activity relationship (Q)SAR methods to predict the mutagenic potential of novel chemicals based on their chemical structure Although existing regulatory guidance does not provide any specific recommendations on what constitutes an acceptable structure-based assessment, based on the current draft of ICH M7 it is likely that two complementary in silico methods and the use of expert knowledge will be required15

Commercially available software packages such as Deductive Estimation of Risk through Existing Knowledge (DEREK) Nexus,31 MC4PC,32 and Leadscope Model Applier33 are already widely used across the pharmaceutical industry to predict the mutagenicity of impurities Each has strengths and limitations, and their performance has been extensively reviewed with no single system performing significantly better than another34-37 Two recent surveys of practical experience (including a total of 13 pharmaceutical companies) found that current practices employed for impurity assessments are highly similar38,39 All companies surveyed initiate in silico assessments by processing structures through a (Q)SAR system or systems (vendor and/or in house) to identify substructures that may confer mutagenic potential The majority of companies complement the output of an in silico system with an expert knowledge approach, such as consideration of mutagenic structure-activity relationships, the structural similarity of an impurity to the respective API, database searches

(public and/or proprietary), and in some cases consultation with medicinal chemists to better understand the impact of various substitutions on chemical reactivity More importantly, the surveys found that the current practices used to make mutagenicity predictions deliver high negative predictivity38,39 and sensitivity39 Summary results of the surveys, which support this conclusion, are presented in Table 164a and b In Table 164a, the data collected from eight companies are combined38 For 566 compounds that were predicted negative based on (Q)SAR systems alone, 94% were negative in the Ames assay For 408 compounds that were predicted negative using both (Q)SAR systems and expert evaluation, 99% were negative This demonstrates that one can have a high degree of confidence when predicting that an impurity lacks mutagenic potential based on (Q)SAR analysis Further, the level of confidence of a negative prediction is even greater when the results of the (Q)SAR analysis are subject to expert interpretation Similar findings were observed in a more recent survey of five companies (Table 164b39), with negative predictive values ranging from 86% to 100% In addition, the survey reported high sensitivity (80%–100%), which

TABLE 16.4a (Q)SAR vs. (Q)SAR Plus Expert Knowledge: Summary of Concordance

TABLE 16.4b Results for Complete Procedures as Used in Companies (DEREK) and a Second (Q)SAR Tool, Including Expert Knowledge

provides confidence in the ability of current practices to identify impurities that possess mutagenic potential

Once complete, the results of a structure-based assessment can be used to categorize impurities into one of five classifications, as defined in Table 16527 This classification system is useful as it helps to differentiate mutagenic impurities that require low-level control (classes 1 and 2) from those that are not mutagenic (classes  4 and 5) It also defines the impurities that would require an Ames test to determine their mutagenic potential (class 3)

The approach used to define an acceptable daily intake (ADI) for a mutagenic impurity requires case-by-case consideration and is dependent on the available information In cases where both positive mutagenicity and carcinogenicity data are available (class 1 impurities), one should first consider the possibility of establishing a compound-specific ADI11,14,27 To determine the appropriate methodology to use to calculate a compound-specific limit, it is necessary to first consider what is known about the mechanism of mutagenicity and of carcinogenicity as well as the shape of the dose-response curve Both the Committee for Medicinal Products for Human Use (CHMP) and the USFDA agree that for compounds with clear evidence of a threshold exposure levels without appreciable risk of carcinogenicity can be established according to the procedure outlined for class 2 solvents in the ICH Q3C guidance40 This established risk assessment approach calculates an ADI from the no-observed-effect level or the lowest-observed-effect level in the most relevant animal study For mutagenic compounds, this would ideally be a carcinogenicity study A weight adjustment factor and a number of safety factors are applied to the animal data to determine an appropriate human exposure

For compounds without sufficient experimental evidence for a threshold-related mechanism, numerous methods are available to calculate a compound-specific ADI Neither the CHMP11 nor the USFDA14 provides guidance on a particular technique One cautious approach considers the tumorigenic dose evaluated in long-term cancer bioassays from the most sensitive species and sex Linear extrapolation is performed from this dose to a dose level that attains an acceptable excess cancer risk in humans (1 × 10-5)11,13,14,27

For mutagens with unknown carcinogenic potential (class 2) or for which there is insufficient carcinogenicity data to calculate a compound-specific ADI, it is recommended that a default exposure limit of 15 µg/day be applied11,13,14,27 This limit, referred to as the TTC, represents a daily exposure level for any unstudied chemical that will not pose a significant risk of carcinogenicity or other toxic effects41-43 The TTC was derived by taking into consideration the distribution of carcinogenic potencies of over 700 compounds and is an estimate of the daily exposure for most carcinogens that would be associated with an excess cancer risk not exceeding 1 in 1,000,000 In the context of food-contact materials, the TTC limit associated with the 10-6 excess cancer risk is defined as 015 μg/day42 However, for the application of TTC to GTIs in drug substances, a ten-fold higher limit, 15 μg/day, corresponding to a 1 in 100,000 lifetime risk of cancer, is considered justified as pharmaceuticals offer a benefit to patients11,14,27 There are some highly potent classes of mutagenic carcinogens (specifically N-nitroso compounds, azoxy compounds, and aflatoxinlike compounds) that may require even lower levels of exposures to ensure negligible cancer risk42

It is worth noting that the TTC is considered to be a very conservative limit, as numerous worst-case assumptions were applied in the derivation of the limit42,44 For example, simple linear extrapolation was used starting from cancer bioassay TD50 values down to a dose associated with 1 in 1,000,000 cancer incidence An assumption associated with using this simple extrapolation method is that all biological processes involved in the generation of tumors at high dosages are linear over a 500,000-fold range In addition, none of the protective biological processes that are more likely to be effective at lower doses are taken into account Therefore, it is reasonable to assume that the estimates of excess cancer risk are significantly overestimated

The methods for the derivation of TTC and compound-specific ADIs mentioned in Section 168 are based on the premise of daily lifetime exposure to a mutagenic carcinogen for 70 years However, very few pharmaceuticals will be prescribed in this manner, with many being used for much shorter durations In addition, exposure to mutagenic impurities needs to be appropriately managed during clinical development, where depending on the drug and the stage of development exposures can range from a day to several years Regulators have acknowledged that there are circumstances that warrant consideration of higher limits for mutagenic impurities, including investigational phases of clinical development and drug indications with less than lifetime exposures Several different recommendations have been made to define specific durations of exposure and their respective ADIs13,14,27 The topic is also currently under discussion as part of the development of the ICH M7 guidance Therefore, a single, harmonized view on acceptable limits for less than lifetime exposures may be forthcoming The recommendation provided in the 2010 CHMP Q & A (R3) document for clinical development stages is provided as an illustration of what is commonly referred to as staged TTC limits (Table 166)

Allowing higher daily intakes of a mutagenic impurity for less than lifetime exposures is based on a stochastic mode of action, that is, the risk of cancer is dependent on the total cumulative dose of mutagenic carcinogens consumed over a lifetime45 In addition to less than lifetime exposures, there are other circumstances that may warrant consideration of higher limits Several examples that have been mentioned in regulatory guidances include when a drug is used to treat a life-threatening condition, when life expectancy in less than 5 years, and when the impurity is a known substance and exposure to the substance occurs at higher levels from other sources11,14

Subsequent to selecting materials for the CCS that have passed appropriate compendial testing, a safety evaluation of extractable data derived from the individual components is warranted Since the extractable studies only identify chemicals that might migrate into a final product from a CCS, any analyses of their safety profile would typically not be submitted to regulatory authorities Nevertheless, since the chemicals that actually do migrate into the final product from the CCS (ie, the leachables) are typically a subset of the extractables, a safety evaluation at this stage represents a “window of opportunity” for a potential safety issue to be identified and mitigated Thus, for the purpose of an internal decision-making process limiting the research to assess the most critical of potential safety concerns is appropriate Toward this end, assessing each extractable for the toxicological end points of mutagenicity, carcinogenicity, sensitization, irritation, and adverse reproductive effects is an example of the most critical issues to be evaluated at this stage Ideally, such an assessment should be literature based; but where little safety data are available for a particular extractable, an examination of its chemical structure for known toxicophores using in silico predictive computational platforms such as DEREK and/ or MultiCASE is quite useful and recommended As illustrated in Section 167, the use of in silico tools for prediction of mutagenic potential in particular can be very robust Collectively, using this approach the data can first be assessed qualitatively to rule out as many of the safety end points as possible If a particular extractable

TABLE 16.6 ADI Limits Based on Duration of Exposure

does have a potential liability, then a quantitative assessment should be employed, assuming the unlikely scenario that the largest amount of the chemical recovered (considering the results of all of the individual solvent systems) will occur in the leachable study In making this unrealistic assumption, if the available data can be leveraged to mitigate the risk, no further action is necessary and development should continue Importantly, as part of the mitigation process in this quantitative approach, the similarities and differences between the solvent conditions yielding a particular extractable and those that will be employed using the actual formulation of the drug in the leachable study should be considered If such differences are great, the degree of concern that the extracted chemical will present in the subsequent leachable work and become a potential safety concern is greatly diminished It is important to acknowledge that after following this approach, the rendering of a definitive safety position could still be hampered by an overall lack of data Given this scenario, the toxicologist should determine if sufficient data are available from structurally similar compounds to develop a safety position in the event that the extractable in question subsequently presents as a leachable Alternatively, addressing the data gap by conducting the appropriate in vitro or in vivo tests needed to rectify the deficiency is also an option, assuming that it is feasible to do so Otherwise, if faced with continued uncertainty or an unacceptable potential safety issue at this stage a packaging change may need to be considered in the hopes of identifying constituents that may yield a more toxicologically acceptable extractable profile

The extractable stage is the first opportunity to identify and mitigate a potential CCS-related GTI matter when developing a parenteral product A mutagenic extractable (eg, positive in the Ames test) should be duly noted, since in the subsequent leachable study it must be either absent or below the TTC of 15 μg/day (in some cases even lower for some very potent genotoxic carcinogens) to support a definitive safety position, unless the chemical has also tested negative for carcinogenicity in vivo, which would abrogate the need for such limitations If the concern for mutagenicity is solely based on a computational structural alert, it must be considered valid unless a literature-based assessment or follow-up Ames testing can prove that the in silico prediction was invalid It is important to note that control to the TTC is limited to DNA-reactive chemicals and is not inclusive of nongenotoxic animal carcinogens that elicit their effects via a threshold mechanism In such cases, the establishment of appropriate quantitative safety margins derived from the animal studies where such effects were characterized is justified Finally, prior to proceeding to the leachable studies it is very critical to consider both quality and regulatory expectations In this regard, the overall quality of the product will be impacted by the number and types of chemicals that leach into it from the CCS during the migration studies In addition, although a particular chemical may lend itself to a robust risk assessment, there may be regulatory sensitivities to its presence in a final product Thus, if assessing the extractable profile yields a potential quality or regulatory concern, it is wise to consider mitigating it at this stage if at all possible Overall, confidence at the extractable stage in matters related to safety, quality, and regulatory sensitivities greatly increases the likelihood that no issues of concern will arise when the final leachable profile is generated

Since humans will be exposed to the analytes identified in the leachables study, a comprehensive regulatory-ready risk assessment should be developed for each of them In most cases, the leachables will be a subset of previously identified extractables, but some exceptions do occur In any case, a comprehensive review of all relevant toxicology data from published and database sources should be performed for each chemical to develop a risk assessment Regarding the latter, the toxicologist should consider a wide range of different factors that will ultimately determine the safety of a particular leachable in a specific drug Thus, leachable risk assessments should always be project specific, and what is acceptable in one drug product may not be acceptable for another First and foremost, the maximum anticipated daily dose for each leachable should be calculated based on the highest concentration measured during the leachable study Duration of patient exposure should be considered, since in general chronic daily exposure to potential toxins can possibly carry a higher theoretical risk of eliciting an adverse effect than short-term exposure The route of exposure is important, since the toxicological profile of any chemical can be affected by how it is administered Patient population and the drug’s indication should also be considered when assessing potential risk In this regard, a particular group of patients (eg, pediatrics or women of childbearing potential) may be more sensitive to a leachable’s adverse effect profile, and if the latter includes the kidney as a target organ then exposing patients who may already be predisposed to renal insufficiency should be a consideration Also, if a drug is indicted to treat a lifethreatening condition such as cancer a greater degree of latitude regarding the potential safety issues for any given leachable is acceptable Overall, a risk assessment for a leachable, or any given chemical impurity, should be based on its dose and how, when, and to whom it will be administered to in the course of treating the patient with the intended pharmaceutical product

Once the potential liabilities of a particular leachable are identified, the risk assessment is developed based on the various factors previously described A leachable’s known toxicity profile in animals and humans is compared to the anticipated patient dose, and the toxicologist determines if appropriate safety margins exist to support safety If data are plentiful, this can be readily determined Alternatively, if in some instances there is a paucity of data the toxicologist can employ safety factors to account for unknowns46 Overall, in the end if there is sufficient safety for the anticipated human dose the assessment is completed

An unknown was reported in a leachable study by gas chromatography-flame ionization detection Since the analyte was unidentified, a surrogate standard was used to quantitate the analyte A carefully chosen surrogate standard will have a response that is typical of analytes quantified by the method; a conservative approach will err on the side of overestimation but practicality necessitates some assumption of similarity in response The unknown increased over the course of several months of drug product storage, crossing the AET and triggering identification work Analysis

of the samples by gas chromatography-mass spectrometry (GC-MS) resulted in a tentative identification of phenanthrene Analysis of an authentic standard confirmed the identification

The drug product was packaged in plastic vials with grey rubber stoppers The drug product was also highly formulated with surfactants and cosolvents, which may have made the product more susceptible to leachables Before the leachables study began, the vials and stoppers underwent controlled extraction studies-extraction of the packaging components in solvents of varying polarities followed by extensive analytical analysis by headspace gas chromatography, GC-MS, and highperformance liquid chromatography-mass spectrometry The extraction report was revisited, and the analyte of interest was not reported in those studies despite the fact that they were thorough, and analytes were reported down to the AET It was noted that the extraction study was conducted approximately 2 years before the leachables study was initiated and that different lots of packaging components were used

As part of the investigation, new vials and stoppers were extracted and the extracts were injected on the leachables method Phenanthrene and other PNA hydrocarbons were detected in the new extracts of the stoppers This resulted in an investigation into the stopper supplier and their supply chain An agreement was in place with the stopper supplier that any changes to their product or process would result in a notification They, in turn, also had similar agreements in place with their suppliers

The investigation eventually revealed a lot of carbon black, which was incorrectly supplied as “low PNA” carbon black even though it did not meet those requirements

The error was a labeling mix-up at the carbon black supplier, a company that also supplied colorants for materials such as tires and automobile parts Since it was not an intentional change, it had not triggered a notification to the stopper supplier

Corrective actions included additional controls and checks at the carbon black supplier, as well as a release test for PNAs for incoming lots of black colorant at the stopper supplier

Leachables from the CCS are inevitable in most, if not all, drug products Analytical identification and subsequent qualification of leachables in certain drug product dosage forms is required based on current regulatory guidances It is critical that analytical methods can identify leachables in drug products More importantly, the analytical methods must be capable of identifying leachables that have genotoxic potential

The use of the SCT and AET approach recommended by the PQRI should be able to identify leachables with genotoxic potential in drug products Based on the PQRI recommendations, the acceptable limit of a leachable GTI is ≤015 µg/day for OINDPs, and ≤15 µg/day for other drug products The ICH M7 Expert Working Group (EWG) is developing new guidance that will allow higher levels of GTIs in drug products, which are dependent on a number of factors, as discussed Based on the current draft of this guidance,15 the basic principles being developed by the EWG should be applicable to leachables in drug products