The implementation of the ICH Q3D document on elemental impurities (EI) in pharmaceutical products¹ and USP general chapters Elemental Impurities – Limits² and Elemental Impurities – Procedures³ occurred for existing drug products on January 1, 2018. FDA expected compliance with ICH Q3D—and for all real purposes general chapter —for new drug products since July of 2016. In conjunction with the ICH document and USP general chapters, USP is omitting general chapter Heavy Metals from the pharmacopeia. These changes result in a new paradigm regarding the safety assessment, measurement and control of these materials. 

In an earlier Contract Pharma article⁴ we discussed some ideas regarding how to go about conducting an initial risk assessment. The result of that risk assessment is a list of elements to consider for final evaluation in the drug product. That is, are there elements other than those listed in Table 3 of USP that should be considered as part of the final risk evaluation, and do any of those elements listed in Table 3, reprinted in this article for your convenience, appear to require additional scrutiny, up to and including additional component controls or a drug product specification? This article discusses some potential next steps in going from the initial risk assessment through method development, method validation and testing to the updating of an annual report or including elemental impurity data in a new regulatory submission. We also discuss information that should be maintained at the manufacturing site in the event of an FDA inspection. Since the process for maintaining and submitting information on existing products is untested for most firms, there is not yet a consensus on what is enough information to maintain and provide, and what might be either insufficient or too much information. Below we discuss some ideas and provide potential points to consider. What is right for each manufacturer will, in the end, be a case by case decision based on the risk assessment and information in hand at the time a summary report is provided to the FDA.

Assessing the risk
ICH Q3D and USP are essentially aligned in terms of the number of elements and their limits to consider in oral, parenteral or inhalational drug products. These include, along with any element intentionally added, the Class 1 (Hg, Cd, Pb and As) and Class 2A (Ni, V, and Co) elements for oral drug products and in addition the Class 3 elements Li, Sb and Cu for parenteral drugs and all the Class 3 elements for inhalational drugs (7, 10 and 14 elements, respectively). The evaluation of the risk of these elements being present can take many forms, including information from manufacturers or suppliers of excipients or drug substances, literature information from peer reviewed journals, or actual testing of the drug product, drug substance or excipients.

The two primary ways of assessing the risk are: (1) evaluating/testing each component of the drug product, and (2) testing the drug product itself. Quite often a combination of these two approaches works best. Testing the drug product and seeing satisfactory results can demonstrate that there are no current issues. However, depending on EI levels observed and component variability, drug product testing alone may not be sufficient to ensure on-going compliance. As noted by EMA in their implementation guideline of March 2017, “Analytical data, without a risk assessment, will not be sufficient to justify to omit a specification for an element. With a risk assessment, depending on its outcome, the number of batches from which results are shown should be commensurate with the risk of the elemental impurities present.”5 Thus, a detailed risk assessment of the components of the drug product may be helpful.

Testing of the drug product
Absent sufficient data to eliminate elements from the risk assessment, the ICH Q3D document recommends the analysis of three production-scale lots or six pilot-scale lots of drug product. The actual number of lots that should be tested may be more than this depending on, for example, the number and levels of mined excipients (e.g., talc, titanium dioxide, calcium carbonate), used in the formulation. Inherent variability of the components, testing methodology and manufacturing processes all enter into this decision. FDA notes in their June, 2016 draft guidance6 that manufacturers should establish that the analytical procedures used during risk assessments possess characteristics (e.g., accuracy, precision, specificity) such that the manufacturers can be reasonably certain—the draft guidance proposes the 95% confidence level—that the measurements can be relied upon to decide whether to include routine testing of materials in the control strategy. Clearly, elements observed at levels far below the PDE require less scrutiny than elements whose levels approach or are above the 30% threshold of concern.

The level of validation should take this into account to the extent possible. FDA also notes in its draft guidance that the validation parameters of both USP general chapter and ICH Q2 (R1) should be considered when developing a validation protocol. This situation is one where validation per the parameters listed in general chapter may not be sufficient. The general chapter calls for standardization between 0.5 and 1.5 times the PDE, (or the parameter J in general chapter , which scales the PDE by the dilution factor to take into account the working range of the instrument). For the purposes of demonstrating that the elements of interest remain below the 30% threshold of concern, validation should be done to a minimum of 30% of the PDE. In order to be sure that no elements in the lots tested are approaching the control threshold and may potentially be an issue in future lots, validation at the lower end at a level such as 10% of the PDE may be advisable. This would allow the range between 10% and 30% of the PDE to serve as an “alert level” to allow action to be taken (e.g., tighter control of EI levels in the drug substance or an excipient), and provide some assurance that exceeding 30% of the PDE (an “action level” where additional controls could be necessary), is unlikely.

Excipient risk assessment
As discussed above, in addition to, or in place of testing the drug product, assessing the risk associated with each of the components in the drug product can be helpful. Controlling the EI levels of individual components can be very effective, especially in cases where the analysis of the drug product is problematic. What is lost examining the components alone is the contribution to the EI burden from the drug product manufacturing process and from the container-closure and packaging systems, which need to be assessed separately. Examination of the potential levels of impurities, (e.g., by assessing the EI burden to the drug product if each component had one metal present as the only EI component at the level of the specification), might lead to the desire for a tighter specification, or examining the likelihood of any EI ever approaching the level of the specification.

This can be done by testing various lots of the excipients, by manufacturer history with excipients, by risk assessments performed by component suppliers, or even by examination of the literature for typical levels of EI observed in pharmaceutical excipients (see Li, et al7 for typical levels in selected excipients and Jenke, et al8 for a discussion of packaging and container-closure systems).

As discussed in the ICH Q3D document, the information leading to a decision can come from many sources, only one of which is testing. Care must be taken when using review articles, however, regarding the quality of the presented data. Often the authors of a review article will present all the data found in their literature review. It is up to the user of the information in the article to ensure that the experimental conditions leading to the data are applicable to the products in question.

Virtual companies typically rely on supplier specifications in determining the acceptability of the excipients in question once a supplier has been adequately qualified. In rare cases, an excipient supplier may delete testing by USP or an equivalent method and not replace it with either a different test or a detailed risk assessment. In such a case, a decision needs to be made with regard to how to proceed, including the possibilities of the manufacturer of the drug product having the excipient tested, or even searching for a supplier that can provide either direct measurements or an acceptable risk assessment, if alternative suppliers are available.

Drug substance risk assessment
While a drug substance risk assessment is not a required element of the Q3D guideline, the assessment of the drug substance and its potential contributions to the drug product elemental impurity burden is an essential part of the drug product risk assessment. It is possible that some drug substance manufacturers will decide to cease analysis per the removal of —Heavy Metals with no plan to either continue with an in-house method of the same technique or to upgrade to an ICP-based or other element specific quantitative method for elemental impurity determination.

Lack of information from the drug substance manufacturer does not necessarily lead to the need for a routine testing strategy. Dropping and not replacing “heavy metal” testing may be acceptable to the user of the material if the drug substance manufacturer performs an appropriate risk assessment as an alternative to testing. The drug product manufacturer can also perform a drug substance evaluation with the assistance of the drug substance manufacturer, using elemental impurity questionnaires, statements confirming that there are no metals (elemental impurities) added, or by testing three lots of material. In the worst case, it can create a situation where the drug product manufacturer may need to initiate its own drug substance analysis plan if it previously accepted the material on Certificate of Analysis (CoA) from the previously-qualified vendor. In fact, for drug substance manufacturers taking this stance, there may be cases where it may be appropriate to re-evaluate the qualification status of the facility in question to determine whether the lack of elemental impurity data or a risk assessment would create an insufficiency to maintain acceptable vendor qualification.

Virtual companies that had previously relied on outgoing release from the drug substance manufacturing site and incoming CoA acceptance at the drug product manufacturer (based on vendor qualification) may find themselves facilitating the generation of information regarding the EI burden presented by the drug substance, potentially including the analysis of the material at either site (or both) to continue to ensure the quality of the drug product made with the component. This relationship management is critical, particularly in the short-term when the path forward will be developed by industry regulation interpretation and compliance initiatives from regulatory authorities.

To control or not to control – trust but verify?
After the final collection of risk review data—drug product analysis, component analysis or risk assessment, assessment of risk from container-closure systems and manufacturing processes/surfaces/water sources—there is a need to decide if further controls are necessary.  No new controls are needed if the appropriately assessed EI levels for the drug product remain below the 30% PDE level of concern threshold value. In the best case for situations where individual components are assessed, ICH Q3D option 1 can be used (USP Individual Component Option), and all the components for each lot tested meet the EI levels described in Table 3 of USP , (meeting the limit at the 10 g/day dose level for each component). 

Another potential scenario is that the EI specifications for each component are such that at the high end of the specification each component contributes a sufficiently small amount to the total EI burden so that the drug product will pass when all these levels are summed. This may indicate that the components are under good control. Conversely, it is possible that in some cases a metal specification will allow for the drug product to exceed its PDE if one metal contributed to the EI burden and was at the top end of the specification. Here it may be possible to test for specific metals and demonstrate that this is unlikely, or to rely on supplier data, published literature or other data for this assessment. In the extreme, if the risk of exceeding the PDE is high, it may be necessary to establish a new lower component specification, either via a risk assessment from the supplier or via separate testing with the establishment of a new specification.

These scenarios, and others involving component options, are further complicated with the USP omitting general chapter from the pharmacopeia. For each component it will be necessary to determine if the component manufacturer or supplier  is planning to update its test methods to another technique, drop testing, or continue testing via their current procedure (now given an in-house test method identifier). As discussed above for the drug substance case, the manufacturer can work with the various suppliers to provide sufficient information to ensure that the EI levels of the various components are acceptable. If this on-going supplier testing is not done or not adequate to demonstrate the material is under control and if an appropriate risk assessment is not in place, drug-product manufacturer initiated testing may be appropriate, and any reduced-testing programs, (based on vendor qualification history), in use at the drug product manufacturing site may need to be altered if testing will be done to replace that previously conducted by the component manufacturer. A potential, albeit unlikely outcome of these scenarios is that the component specifications may change, and it is important to include the change control process for the specification when estimating time needed to establish that the component EI levels are well controlled. We anticipate that this may be a rare case since industry experience to date indicates that, with the potential exception of some mined excipients used at high levels, EI contributions from excipients are very low.

Testing of a number of common pharmaceutical excipients7 has shown that many excipients have actual individual elemental impurity levels far below those of their total impurity specifications. Data from the literature for a particular component, or testing of several lots of a component may allow one to conclude that the component will not be an issue as long as this grade of material from this supplier continues to be purchased.

There remains a risk in this assumption, particularly for mined (mineral-based) excipients, since EI levels can vary substantially depending, for example, on the location of a mine or where in the mine the material came from (or for a plant-based excipient, perhaps the metal status of the location where it was grown). In the majority of cases the appropriate risk assessment will establish that the variability is such that, at the level at which the component is used, no routine testing is necessary. For the few cases where the levels of a component and its variability can potentially lead to exceeding the PDE, some sort of routine testing of the component may be necessary.

A re-evaluation of the risk assessment should be undertaken each time a grade of material is changed or a new supplier is qualified to ensure that the EI levels in the new material do not impact the conclusions from the initial risk assessment. Fortunately, many excipient suppliers are now performing risk assessments and providing element-specific specifications that indeed demonstrate that the levels of EI in their products are very low. This is an ideal situation that allows drug product manufacturers to make data-driven decisions on how to proceed. 

There may be situations where the level of one or more metals exceeds 30% of the PDE threshold of concern. In this case, investigation of the source of the elevated level and implementation of corresponding controls will be necessary. Depending on how close the observed level is and the variability associated with its measurements, no action may be needed. For example, if the EI level of a drug product showed 37% of the PDE with a range (including 3 sigma about the mean) of 25-45%, no additional controls should be indicated since the EI would not be expected to exceed the PDE. The closer the observed level (with accompanying variability) approaches the PDE, the greater the need to evaluate the necessity of additional controls.

In the worst case, the risk assessment and accompanying testing will show that the EI level for one or more elements exceeds the PDE and cannot reasonably be consistently reduced to the PDE (the ICH guideline states that health authorities should not request limits lower than the PDE). Depending on a number of factors, it may be acceptable to exceed the PDE (e.g., a medically-necessary drug for an acute indication or rare disease or short-term dosing). In this case, a best effort at control should be made and realistic specifications established. It is advisable to discuss this situation with FDA or other appropriate health authority prior to settling on the specification and testing requirements and including these in a regulatory submission, to ensure that there is agreement to the way forward prior to final submission of the documentation. While the ICH document provides tools to justify higher EI levels, the acceptance of any higher levels is a regulatory authority decision.

Documents to include in a regulatory submission and to have on location
As the final step, there is a need to decide what one submits to the regulatory agency and what one needs to maintain at the manufacturing location in the event of an FDA audit. Current guidance from both ICH Q3D and the FDA suggest that the following be included in regulatory dossiers (new or updates):

• Summary of the product risk assessment process used;
• Risk assessment conclusion containing a list of identified elemental impurities and observed or projected levels;
• Method validation summary and finalized test method;
• Data from representative commercial or pilot scale batches (component or drug products, as appropriate); and
• Go-forward risk-mediation plan based on the concluded risk assessment, use of the validated method, corresponding analysis results and specifications based on the entire process.

The conclusion of the product risk assessment and corresponding risk mediation plan is important. It should summarize in a logical way the reasons why the firm believes that the drug product meets the current requirements, and any controls or specifications that have been implemented to ensure that not only the current batches are acceptable, but that future batches will be acceptable throughout the drug product’s life cycle (e.g., stability, changes in excipient or drug substance manufacturer or supplier, changes in grade of materials). Demonstration that controls are in place, if needed to ensure EI control throughout the drug product’s life cycle, will be critical to the acceptance of the overall risk assessment conclusions.

Based on excipient data generated to date, FDA is expecting most of the updates it receives regarding EI control to indicate that routine testing will not be needed. Thus, for existing drug products they request that the updates to this topic be included in the annual report. This is not true if the drug product does not meet the requirements of USP or the recommendations of ICH Q3D. In this case, a discussion with the appropriate review division is necessary soon after confirming such results and prior to updating the regulatory submission.  For new products, the P.2 (Pharmaceutical Development) section of the electronic Common Technical Document may be an appropriate location for this information.
Regarding documentation to be maintained in the manufacturer’s Pharmaceutical Quality System, the following is recommended:

• Complete risk assessment document describing process, data used, data references and information needed to support the dossier summary;
• GMP-related processes to limit the inclusion of elemental impurities;
• Change management processes (defined triggers for product assessment or control strategy updates);
• Periodic review processes; and
• Original data used in the product risk assessments, quality agreements, supplier qualification, and other supporting documentation.

These materials should be readily accessible in the event of an FDA audit and should be updated regularly as agreements, suppliers or grades of materials change and as more data are generated in support of the current risk assessment conclusions.

Conclusions
At a high level, the pieces necessary to define, execute and submit a risk assessment for EI are well defined. At this point, however, sufficiently few reviews have occurred to be sure that the level of detail and the level of controls submitted to FDA will be adequate. As FDA and other regulatory authorities have the opportunity to review and comment on what they see, this process will improve and further guidance can and will be provided. It will take the right level of risk assessment, specification setting of components and, as needed, testing to ensure and to demonstrate to the regulatory agencies that the drug product meets the requirements both now and throughout its life cycle and that the control strategy and go-forward risk mediation process in place is sufficient to maintain integrity of the risk assessment program. 

References

1. ICH Q3D Guideline for Elemental Impurities.  http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q3D/Q3D_Step_4.pdf.  Accessed January 23, 2018.
2. USP General Chapter – Elemental Impurities – Limits. http://www.usp.org/chemical-medicines/elemental-impurities-updates  Accessed January 23, 2018.
3. USP General Chapter – Elemental Impurities – Procedures.  http://www.usp.org/chemical-medicines/elemental-impurities-updates   Accessed January 23, 2018.
4. Elemental Impurities:  A Virtual Company Perspective.  Contract Pharma, April, 2017.  https://www.contractpharma.com/issues/2017-04-01/view_features/elemental-impurities-a-virtual-company-perspective/ Accessed January 23, 2018.
5. Implementation Strategy of ICH Q3D Guideline http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2017/03/WC500222768.pdf.  Accessed January 23, 2018.
6. Elemental Impurities in Drug Products – Guidance for Industry https://www.fda.gov/downloads/Drugs/Guidances/UCM509432.pdf  Accessed January 23, 2018.
7. Elemental Impurities in Pharmaceutical Excipients, Li, G., Shoneker, D, Ulman, KL, Sturm, JJ, Thackery, LM,  and Kauffman, JF, J. Pharm. Sci., 104(12), 2015 pp 4197-4206.   https://www.sciencedirect.com/science/article/pii/S0022354915000775#t0010.  Accessed January 23, 2018.
8. Materials in Manufacturing and Packaging Systems as Sources of Elemental Impurities in Packaged Drug Products: A Literature Review. Jenke DR, Stults CL, Paskiet DM, Ball DJ, Nagao LM., PDA J Pharm Sci Technol. 2015 1/2;69(1):1-48.  Available at https://www.ncbi.nlm.nih.gov/pubmed/25691713 Accessed January 23, 2018.

Acknowledgements
The authors thank Dr. Mark G. Schweitzer (Novartis) for his review and comments on the manuscript.

---

Tony DeStefano is a member of the Editorial Advisory Board of Contract Pharma. He was Vice President and Sr. Vice President of General Chapters at USP from 2008-2013. He currently is a pharmaceutical industry consultant in the areas of CMC, bioanalysis and compendial issues.

Thomas Kester is currently the Director of Analytical Services at Recordati Rare Diseases, Inc. in Lebanon, NJ. He has over three decades of experience in the analytical sciences within the pharmaceutical, biological and beverage industries.