“In order to minimise the risk of a serious medical hazard due to cross-contamination, dedicated and self-contained facilities must be available for the production of particular medicinal products, such as highly sensitising materials (eg. penicillins) or biological preparations (eg. from live microorganisms). The production of certain additional products, such as certain antibiotics, certain hormones, certain cytotoxics, certain highly active drugs and non-medicinal products should not be conducted in the same facilities. . . ”
This wording does not provide any clarity on which of these types of compounds fit the requirement. In the most conservative approach many companies (and regulators) have defaulted to the mode that all of these compounds should be in dedicated facilities. It also should be noted that additional confusion arises with how “hormones,” “cytotoxic,” and “highly active” are defined. These terms do not have standard industry definitions.
After several hours of successful discussion around these documents, it was clear the EMA was willing to work with industry to develop a good solution. A major decision was reached that the toxicological guidance document would only address the limit setting method, and not address which situations may or may not require dedicated facilities, whereas the GMP updates would address how the limits set by the toxicological guidance document would be used to determine the need for dedicated facilities or other risk control methods to manage the risk of cross-contamination. Other areas discussed included:
- When available human data and chemical-specific safety/adjustment/uncertainty factors should be used to determine the limits
- That the ICH M7 method would be used to determine the threshold for compounds with genotoxic properties where there is insufficient data to determine a chemical specific limit
- That additional detail is required in setting limits for investigational medicinal products (IMPs) and biological medicines where they can easily be deactivated
- And that more discussion is required for the application of the toxicological guidance to veterinary products
In 2005 the EMA published a concept paper on the need to update GMPs to provide better clarity on the requirements where it identified compounds and classes of compounds that it felt should be in dedicated facilities. At that time industry responded that a science- and risk-based approach should be used to determine the requirements.
Around the same time a group of industry professionals along with the FDA began work on ISPE’s Risk-MaPP Baseline® Guide. ISPE’s Risk-MaPP Baseline® Guide outlines a scientific risk-based approach, based on ICH Q9, to manage the risk of cross-contamination in order to achieve and maintain an appropriate balance between product quality and operator safety. This document was published by ISPE in 2010.
In 2011 EMA issued another concept paper stating the need for a toxicological tool. Specifically the document states:
“Currently toxicological data are not always used in establishing limits for cross-contamination. In some cases arbitrary limits such as 1/1000th of the lowest clinical dose or 10 ppm are used as limits for cleaning validation. These limits do not take account of the available pharmacological/toxicological data and possible duration of exposure and may be too restrictive or not restrictive enough. A more scientific approach based on current available pharmacological and toxicological information is required to establish threshold values to be used as part of the overall Quality Risk Management in shared facilities.”
As a result of the above concept paper, a toxicological tool titled, Setting health-based exposure limits for risk identification in the manufacture of different medicinal products in shared facilities, was published in January 2013 for industry comment. The PDE (permitted daily exposure) as outlined in ICH Q3C Impurities: Guideline for Residual Solvents serves as the basis for the proposed tool. While this methodology is very similar to the ADE (acceptable daily exposure) approach as outlined in ISPE’s Risk-MaPP Baseline® Guide, the selection of factors — used to adjust the No-Observed-Adverse-Effect Level (NOAEL) to a daily dose that is unlikely to cause an adverse effect if an individual is exposed, by any route, at or below this dose every day for a lifetime — is where the two methods appear to vary. The PDE approach was developed for solvents and is also being used for elemental impurities where very little human data are available and therefore rely more heavily on factors to address the uncertainties. The ADE approach is based on the ADI (acceptable daily intake) approach used for many years in the food industry and medicines used for animals in the food chain. This approach recognizes the vast human data available for the drug products obtained through clinical trials and actual commercial use. This allows the adjustment/uncertainty factors to be more reflective of the dataset available.
Several industry stakeholder groups provided comments on the draft toxicological tool. These comments ranged from allowing more flexibility in the limit setting method to excluding API manufacture, investigational materials manufacture and veterinary products, with several industry stakeholders requesting the use of ADEs and the method outlined in ISPE’s Risk-MaPP.
At the same time (January 2013) the EU issued proposed revised text for managing the risk of cross-contamination in Chapters 3 and 5 of the GMPs. The revised text states:
“Cross-contamination should be avoided for all products by appropriate design and operation of manufacturing facilities. The measures to prevent cross-contamination should be commensurate with the risks. Quality Risk Management principles should be used to assess and control the risks. Risk assessment should include among other parameters a toxicological evaluation of the products being manufactured (see Guideline on setting health based exposure limits for use in risk identification in the manufacture of different medicinal products in shared facilities).
Dedicated facilities are required for manufacturing when a medicinal product presents a risk:
a) Which cannot be adequately controlled by operational and/or technical measures or
b) Scientific data does not support threshold values (e.g. allergenic potential from highly sensitising materials such as beta lactams) or
c) Threshold values derived from the toxicological evaluation are below the levels of detection”
Chapter 5 states:
“A toxicological evaluation should be the basis for the establishment of threshold values in relation to the products manufactured (see Guideline on setting health-based exposure limits for use in risk identification in the manufacture of different medicinal products in shared facilities). Where the toxicological evaluation supports a threshold value, this should be used as an input parameter in risk assessment. A Quality Risk Management approach should be used based upon this toxicological evaluation and the potential cross-contamination risks presented by the products manufactured. Factors including facility/equipment design, personnel flow, physico-chemical characteristics of the active substance, process characteristics, cleaning processes and analytical capabilities relative to the threshold values for products should also be taken into account. The outcome of the Quality Risk Management process should be the basis for determining the necessity for and extent to which equipment and facilities should be dedicated to a particular product or product family. This may range from dedicating specific product contact parts to dedication of the entire manufacturing facility. It may be acceptable to confine manufacturing activities to a segregated, self-contained production area within a multiproduct facility, where justified.”
The proposed wording suggests the application of a risk management process covering all products to be produced in a shared facility to confirm that the risk of cross-contamination is managed to acceptable levels. Additional updates in Chapter 5 provide suggestions for possible controls to manage the risk of cross-contamination.
While these updates to the GMPs seemed to be in line with industry’s comments on the 2005 EMA concept paper, some of industry now feels these updates are too much of a burden and no real value to patients. Industry’s concern with the proposed GMP updates focuses on implementation. Some in industry, especially EFPIA (the European Federation of Pharmaceutical Industries and Associations), question the need to apply this approach to legacy/all products. EFPIA contends that these products are currently being made safely and as long as there are no major changes to the process or facility, the existing products should be excluded from the requirement to have a toxicological evaluation. They argue that the 1/1000th of a low clinical dose (LCD) and 10 ppm limits for cleaning are safe, since much effort is expended in determining a safe clinical dose. EFPIA would like only “certain” products to follow the proposed EMA risk management process.
These objections were verbalized at the recent EMA workshop with industry representatives to review comments received on both the toxicological guidance and the proposed GMP updates. These arguments are flawed for several reasons.
There continues to be confusion between risk and hazard. Hazard is a measure of the harm, while risk reflects exposure to the hazard and is the likelihood that harm will occur. The ADE or PDE are measures of harm or hazard indicators. These values represent how harmful the chemical is to the patient. To base the need to follow the EMA’s proposed process solely on the value of the ADE or PDE is not risk-based; it is hazard-based. In order to assess high risk several other factors must be considered to determine the likelihood of harm due to potential exposure to the hazard. These factors include but are not limited to the products presentation (API, oral solid dose, parenteral, etc.), the process (open or closed?), the equipment (common or dedicated?), processing volumes and the actual properties of the API. When these factors are taken into consideration, the high hazard product is often not the high-risk product. This is because high hazard products are usually produced in a contained way, mainly to protect the operators, in smaller quantities, and the final product has a low quantity of API.
But if we consider the other products in the low and medium hazard range the exposure potential due to the processing parameters increases the risk. Many of these products are still produced globally in open processing equipment, they have a high percentage of API per dose, share equipment and are produced in large volumes. In addition many of these products are taken daily over a lifetime such as diabetes, cardiovascular and anti-psychotic drugs. The risk that a patient could receive a cross-contaminated product is higher with these types of products even though much of the industry is focused on the high hazard drugs (hormones, cancer medicines, etc.).
Cleaning or retention is not the only method by which cross-contamination can occur. Mix-up, mechanical transfer (non-product contact surfaces) and airborne transfer are also potential modes of cross-contamination. Mechanical and airborne transfer need to be assessed as potential exposure routes, especially for low and medium hazard products manufactured in open processes. For high hazard compounds, operations tend to use closed or contained systems to carefully manage mechanical and airborne transfer, mostly to protect the operator.
It is true that for approximately 90% of products, the traditional cleaning limits of 1/1000th LCD and/or 10 ppm will be lower (much lower) than when determined using the ADE or PDE. It is important to remember that both ADE and PDE are values that represent a safe dose that patients can take for a lifetime without an adverse effect; this is a very conservative value. Many in industry and even the regulators have difficulty in allowing the cleaning limits to rise.
It seems counterintuitive to producing safe products. In fact, using the ADE or PDE as the basis for setting cleaning limits have less risk than maintaining an artificially low limit, because of the large margin of safety provided between the actual data and the acceptance limit. A hierarchy of limits can be used where the normal range is set at the process control limit with the acceptance limit set at the ADE/PDE derived value. For added comfort, action and alert limits can also be set between the process control and acceptance limits. It should be noted that failure investigations to satisfy regulatory requirements would only be necessary if the acceptance limit was exceeded. Exceeding the acceptance limits is more likely when the limits are set artificially low and are in the same range as the actual results. With the use of the hierarchy of limits approach there should be sufficient warning so that a failure investigation could be a very rare event.
In some cases the values determined with the ADE/PDE will be very large and many argue that it is not appropriate to “allow” carryover of one material into another up to this limit. They are correct. But all GMPs require visually clean equipment and facilities. So the real upper limit for the cleaning limits is visually clean. Even if a “safe” value is calculated above visually clean, the GMPs are clear in the requirement for being visually clean and this would be the overriding limit in these cases.
Many facilities that produce low and medium hazard compounds manually wash the shared equipment between different products and only use a visual inspection to determine if the equipment is adequately clean. While these companies have validated the cleaning process to show that the limits can be met, the manual cleaning process is not robustly repeatable to actually ensure the limits are being met after each and every cleaning. Often the cleaning limits set are below the visual acuity of the operators (especially when set by 1/1000th LCD or 10 ppm) or more often the company does not actually know what the visual acuity of its operators are. If the limits are below visual acuity, visual inspection is not an adequate method to determine if the equipment is clean enough. This means that a chemical analysis either with swab or rinse samples is needed. But if the limit set using the ADE/PDE brings the value into the visual acuity range by updating the acceptance limits, then chemical analysis can be avoided, which will save both time and money.
It should be understood that the therapeutic effect in one product could be considered an adverse effect as a cross-
contaminate in another product. For example, if a patient has low blood pressure and takes a diabetes medicine that is produced in the same facility as blood pressure-lowering medicines, this patient could suffer an adverse effect if their diabetes medicine is cross-contaminated with the blood pressure medicine. To rely solely on the clinical dose with default safety factors (such as 1/1000th) may not adequately protect patients.
Another argument for the exclusion of existing products in the updated documents is that products made for 20 or more years using traditional cleaning limits cannot be unsafe. This argument is akin to refusing to wear seat belts in a car because cars did not have them 20 years ago! Part of the recent Process Validation Guidelines — both FDA and EMA — is a requirement for continuous improvement. Understanding how toxicological and pharmacological data can be used to set appropriate limits to manage the risk of cross-contamination fits nicely into the process improvement arena. In addition, ADE/PDE is a direct indicator that specifically links scientific knowledge to the safety of the patient as required by ICH Q9.
The EMA workshop clearly showed that industry is at a crossroads. Some see the value in using a science- and risk-based approach to support patient safety for all products, while others are holding on to old paradigms under the guise of possible drug shortages and increase in cost of medicines. In reality those holding onto the old paradigms lack a true understanding of how this risk-based approach can make a facility more efficient with less quality issues resulting in downtime while producing safer medicines for patients. Those holding onto to the old paradigms also cannot see the savings presented in running a true risk-based system but rather just see the extra effort upfront to modify the quality systems, obtain ADEs/PDEs and perform possible revalidation work that may be required.
Due to these divergent views the EMA requested recommendations from industry for a screening tool to determine which products or risk situations will require the use of the toxicological assessment.
So while other areas of pharma manufacturing are embracing science- and risk-based approaches, the some in industry are back trying to define “certain” in terms of where dedicated facilities may be required.
Stephanie A. Wilkins, PE, is president of PharmaConsult US, Inc. She can be reached at firstname.lastname@example.org.
Photo courtesy of SAFC.