A3P International Congress 2018: restitution of the key messages of the round table on the revision of Annex 1

At the 30st A3P International Congress in Biarritz (13-15 November 2018), a round table was organized on the theme of the Annex 1 (Manufacture of Sterile Medicinal Products).This round table, moderated by members of the A3P GIC Annex 1, was an opportunity to discuss with participants a number of topics that generated particular interest from industry during the public consultation phase (20 december 2017-20 march 2018) after publication of the draft by the EMA 20 December 2017.

Interest in this Round Table was also greatly motivated by the presence of Andrew Hopkins, Senior Inspector MHRA and Leader of the EMA / PICS / WHO (IWG) Working Group in charge of the revision of Annex 1.


In order to initiate the discussions, members of the A3P GIC Annex 1 (Sophie Amadio / Lilly, Julien Triquet / GSK, Eric Hurtubise / Théa Laboratories and Marc Besson / MB-GMP Compliance and leader of the GIC A3P Annex 1) have prepared a series questions around 5 themes representative of the concerns shared by industry during the public consultation phase.


1. Status of the Annex 1 review process
Prior to the exchange, Andrew Hopkins provided information on the progress of the review process and next steps leading to the release of the new version of Annex 1.
6 213 comment lines received from 140 different entities (A3P, PDA, PHSS, ISPE, LEEM, …) were reviewed individually by Mr. Hopkins and resulted in a series of comments and proposals for actions for the IWG. The draft of Annex 1 was revised taking into account the most relevant comments and returned to the IWG on October 19 2018, then returned to Andrew Hopkins on November 25 for transmission to EMA 5 December. The objective of the EMA, despite the organizational impacts related to Brexit, was to publish the final version of 1 Annex 2018.


2. Topics and issues discussed at the roundtable
a/ Supervision of the evolution of the promotional practices of the products under penalty of financial sanctions

The “global strategy for the control of contamination (particulate, microbiological, pyrogenic)” is a fundamental concept widely developed in the revision of Annex 1. The GIC A3P Annex 1 wished to clarify the implementation of this concept on industrial sites and inspectors’ expectations regarding the consolidation of information in a summary document.

Question GIC A3P Annex 1
CCs is mentioned 16 times in the update and clearly appears as a new key requirement. The new Annex 1 provides a detailed list of elements expected to be covered within a CCS. Most sites will already have many of the elements of a CCS as described by the Annex 1 but they may not be collated through a single source as appears to be the intent of this requirement. —

  • What would inspectors expect as a minimum acceptable standard document to meet the Annex 1 requirements regarding the CCS?

Andrew Hopkins response
Industrialists have already formalized numerous documents responding to the various elements constituting a CCS, but many have not yet consolidated these elements according to a global and “holistic” approach. The concept of the CCS is to get manufacturers to reflect on the relevance of their actions such as the location of exposed boxes for environmental control, the scenario of their Media Fill Tests, the design of premises and their processes … Much like the relationship between the results of clinical trials and the determination of product specifications. The development of a CCS must lead manufacturers to have a detailed knowledge of all the elements (premises, processes, equipment, components, practices, etc.) that can induce microbiological and particulate risks on the environment and products, and demonstrate how these risks are eliminated or controlled and what are the means of controlling the consequences of these risks. It is not necessary to have a single document grouping together all of these measures and justifications if they are already documented individually, but there should be a summary document making it possible to demonstrate that each site has in-depth knowledge of the processes. and risks. This summary document would make it possible to make the link between the various actions and means of measures in place by the manufacturer.
It should be remembered that the concept of CCS is not new because it has already appeared for many years in the 5 chapter of GMP, but a focus could then have been realized on the notion of chemical contamination, thus omitting interest in the concept of microbiological contamination, however, already addressed in this chapter.


b/ Quality Risk Management (QRM)

The QRM concept is also widely developed in the Annex 1 draft but could lead to some interpretations by industry and inspectors as to the application and scope of this concept. The A3P GIC Annex 1 wished to address concrete examples for which the QRM principles could be applied.

Question GIC A3P Annex 1
The draft of Annex 1 appears to be very detailed and prescriptive for some expectations but at the same time largely promotes the use of QRM principles. In addition, some sites could face limitations by design to implement some of the Annex 1 requirements.

  • Could we have some examples where Annex 1 requirements are mandatory and some examples where alternative approaches supported by sound and scientific rationale may be acceptable?
  • Would EMA encourage sites to pro-actively discuss alternative approaches with regulators before submission/implementation?

Andrew Hopkins response 
If we really understood the concept of QRMs (industry and inspectors), Appendix 1 could be summarized as follows:
“Design the processes, procedures and facilities not to contaminate the product. Design the monitoring system to detect any deleterious trend and or failure. Keep reviewing and developing as new information about your processes, procedures and designs comes to light. Keep developing as you become aware of new technological advances ”.
But it must be remembered that the 1 Appendix is ​​written for different types of products, manufacturers, for many countries in the world with different understandings of the GMP and therefore the document must be sufficiently prescriptive.
An example for which there should be no flexibility: if you find more than one CFU in your grade A environment, there is a problem regardless of your scientific justification! I read comments during the public consultation that suggested that if the identified microorganisms were not pathogenic, it was not a problem! Some requirements must remain mandatory!
The inspectors also have a lot of experience and the GMPs were written on the basis of bad practices and their consequences. When mandatory limits are specified, they rely on proven problem histories.
Another example: in the draft it is written that the clean areas must be qualified every 6 months but in fact, what we wanted to require, is that the integrity of the HEPA filters of the clean areas be checked every 6 months. This will be corrected in the revised version. Should there be a discussion around this frequency? Probably no. A frequency of 6 months is already relatively comfortable. The reason I say this is a recent example of a site that last December (2017) contacted the MHRA to report that 3 HEPA filters on 4 in its A class were unhealthy. A risk assessment was conducted with the manufacturer and the conclusion was that all the products were critical, that no batch could be recalled without out of stock, and that the production could not stop! We were therefore in a situation where the manufacturer and the agency had to release risky products. We could not rely on the results of the environmental controls, and we had 6 production months with lots at risk. Once the problem has been solved, new batches have been produced in order to substitute, as and when the lots at risk on the market. In June this year, the manufacturer informed us of new problems of integrity of the HEPA filters of their own zones with 11% of defective grade A filters, 27% in grade B and about 50% in grade C. In conclusion , wanting to extend the testing frequency of HEPA filters to 12 months seems irresponsible, at least for the most critical areas.
The inspectors have many other examples of bad experiences and this explains sometimes demanding requirements.
The fundamental point with QRMs is to demonstrate an in-depth knowledge of all the elements of the process, and this at the very level of the workshop involving the operators. If you are considering unconventional, innovative changes, it is important to meet your contacts within the agencies very quickly. Do not wait for the submission of the file, nor the first inspection. Many agencies (MHRA, EMA, FDA, …) have specific departments in charge of reviewing changes and innovations and we must have this scientific discussion well before the implementation of the industrial project. The dialogue must change between manufacturers and agencies because some manufacturers are still “cautious” at the idea of ​​discussing innovations or projects with agencies. If we want to better promote the application of QRMs, the objective is also to “educate” inspectors on technological developments because if they do not know them, they can be more conservative.



c/ Conventional Clean Areas vs Barrier Technologies

The use of barrier technologies to move the operator away from the critical zone and the product and thus significantly increase the sterility assurance level of the process is becoming more and more obvious. The A3P GIC Annex 1 wished to clarify what could be the GMP / regulatory impacts for the many sites still using conventional clean zones.

Question GIC A3P Annex 1
The development of Barrier Technologies (RABS and Isolators) is largely promoted to enhance sterility assurance for aseptic processes but there are still a number of sites operating with conventional clean rooms.

  • Would inspectors expect dedicated/enhanced risk assessments and procedures to maintain confidence with the level of Sterility Assurance provided by such facilities?
  • In the future would EMA continue to accept submission dossiers supporting the use of conventional clean rooms for aseptic processes?

Andrew Hopkins response
First of all, the requirement for manufacturers to always upgrade their installations in relation to new technologies is not new, since it appears in chapter 1 of GMPs since 2013 or perhaps even before.
Is the use of insulators the only answer? Certainly not! Because many processes are not compatible with the use of insulators, for example when you have very manual formulation processes or very small batch sizes, even a unit … The key principle is to move the operator away from the critical area and the product either by the use of adapted air flows, RABS, insulators, closed system … For us inspectors, impose the specific use of insulators or RABS would be an error because we must also allow the emergence of innovations while keeping in mind the key principle.
During the public consultation, we had a question about an old-fashioned facility that still had good environmental results; can we maintain this installation? The answer is certainly no, because the industrialist must evolve towards more modern technologies. This type of old installation is obviously at risk. The use of QRMs must first lead to good quality design of premises, processes, procedures and then design a suitable environmental control program. Do not use environmental control to justify poor design or practices.
Concerning the submission of files with conventional clean zones, first of all it should be specified that the design of the installations is not a point reviewed during the evaluation of the file by the “reviewer” but rather a question relating to GMP. So, even if your file is accepted, you can have a refusal of a GMP certificate during the inspection. I think that it will be increasingly difficult to accept installations of old design, simply because the manufacturer will not meet the requirement to update the installations in relation to new technologies. Can we change everything immediately? Certainly not, because there are still many such installations around the world.

Complementary question on the principles of harmonization between agencies (reviewers and inspectors)

  • How is the desire for harmonization between the agencies to date, particularly when we submit a file with several reviewers / rapporteurs who may not agree with each other?

Andrew Hopkins response
As the draft progresses, the PICS commits itself to train the various inspectors so as to limit the interpretations. But this training (education) is also done through bodies like A3P, PHSS, PDA that have an important role to exchange outside regulatory context on controversial issues or innovations.


d/ Integrity test of the sterilized product filter before use (PUPSIT)

The PUPSIT (Pre Use Post Sterilization Integrity Test) was the subject of many comments during the public consultation and remains a controversial subject within the industry with respect to the requirements of Annex 1. GIC Annex 1 wished to clarify the position of the IWG and in particular to know if a QRM approach is possible in specific cases that may lead to a product risk.

Question GIC A3P Annex 1
The new Annex 1 still considers that PUPSIT must be implemented but for some existing installations manipulations downstream the filter may be at risk for the product.

  • Would inspectors consider that PUPSIT is mandatory for all lines and therefore line modifications/upgrades must be carried out or would inspectors consider as acceptable if a documented risk analysis is provided?
  • — If a RA is acceptable which risk-related elements should be covered in the RA?

Andrew Hopkins response

The PUPSIT has indeed been the subject of many comments but this requirement already exists in the current version (2008) and therefore the first question is why the manufacturers have not already set up! 
From my personal point of view (does not reflect the EMA’s point of view) I think there should be opportunities for discussion. We all know that filters can be damaged or clogged: this follows physical laws. An acceptable risk analysis could be based on considerations integrating appropriate controls from the filter supplier (not only “standard” controls), transport, sterilization, if in particular it is subcontracted, associated with a good knowledge of the characteristics of the product and the filtration process. For example, if we have a process with multiple clarifying filters before the product filter, there is little chance that the solution to be filtered contains particles and therefore a low probability of clogging the filter. 
As another example, the nature and characteristics of the solution to be filtered are important decision elements. Especially if we filter the EPPI, how many liters will we have to filter before sealing the sterilizing filter? Certainly more than the production capacity of your production unit. These examples are good starting points and there are possible openings for discussion, but you must demonstrate your perfect knowledge of your process and product. I proposed this Risk Analysis approach to the IWG but some still refuse it when others do not want to hear about the PUPSIT anymore! To date, we do not know what the final position will be in the new version of Annex 1.
On the other hand, what we do not want is the use of the RA to justify that we do not want to implement the PUPSIT. If you commit this AR, you should not know the conclusion before starting the analysis.

Question from participants

  • Can you comment on the PDA initiative on clogging tests depending on the nature of the solutions and specific filtration factors?

Andrew Hopkins response
Indeed, the PDA has initiated a certain number of studies concerning the PUPSIT (“PDA / BPOG Memorandum of Understanding, Filter manufacturers joint statement, Filter blocking trial initiative”) and in particular an experiment plan consisting of simulating tests with several types of solutions in order to collect scientific information on the clogging parameters of the filters (when, what bioburden required, what level of loss of integrity can be masked by clogging, etc.). The protocol has been reviewed by the IWG. Here is a really good example on how to introduce QRM and collaboration between agencies and industry.


e/ Robustness of surface decontamination with Vaporized Hydrogen Peroxide (VH2O2)

The publication, the 20 April 2018 on the MHRA website, of a blog questioning the robustness of the VH2O2 used for the decontamination of critical indirect surfaces (VHP (Vapor Hydrogen Peroxide) Fragility) has elicited many comments within the industry and fears about maintaining this type of decontamination for existing insulator installations. This specific position of the MHRA, however, does not reflect the vision of the IWG and will not be included in Annex 1. The A3P GIC Annex 1 wished to revive the discussions in particular to clarify the possible impacts for insulators in operation for which disassembly / sterilization with steam / reassembly of bowls caps, conveyors, ramps, … can prove to risk or even impossible to by the design of the equipment.

Question GIC A3P Annex 1
MHRA has recently challenged the efficacy of Vaporized Hydrogen Peroxide for the decontamination of Indirect product contact parts (Stopper bowls, hoppers, conveyers..) and suggests that such parts be sterilized in autoclaves and then be aseptically set up once the isolator is decontaminated. Many existing isolators are not designed to allow easy dismantling of indirect product contact parts and these manipulations may generate new risks (ie loss of glove integrity due to additional manipulations…)

  • Will the revised Annex 1 include clarification on requirements for the use of decontaminating/sterilizing agents such as VH2O2?
  • Would inspectors accept a risk based approach for existing isolator lines to demonstrate that all measures are in place to ensure VH2O2 robustness?
  •  Is this position on VH2O2 applicable to conventional clean rooms where stopper bowls, hoppers and conveyors are decontaminated in place?

Andrew Hopkins response
I wrote this blog on the MHRA website because I noticed a lot of problems on how to use the VH2O2 by some manufacturers and their ignorance of the “fragility” of this decontamination agent. If the VH2O2 is used under controlled conditions of temperature, humidity, duration, one can reach an efficiency of 6 log of reduction but there are also many constraints related to all these physical parameters and in particular the notion of efficiency of penetration. And what we see during inspections is that manufacturers do not understand these constraints.
For example, I inspected a site where they had a very good RABS process decontaminated at VH2O2 and for which they used a hose (used to inert storage bags) 2 meters long. According to the manufacturer, this hose was “sterilized” by the VH2O2 only by passive contact. In no case the VH2O2 cannot diffuse “passively” through a filter and through a flexible hose of such length!
Another example with a corkscrew for which the cleaning procedure consisted only of a simple rinsing with the EPPI without any consideration of the elimination of the possible residues which can decrease the effectiveness of the VH2O2. Handling of this D-grade cap bowl with a risk of contact of the surfaces with the skin and therefore the risk of leaving traces of fatty acids on these surfaces, can lead to occlusions of microorganisms and thus prevent contact with the VH2O2.
Finally, the very questionable position of Biological Indicators (BIs) located outside of plug guides (and not inside the rail)… similarly the concept of the number of BIs per position (3) on the basis of ‘a “statistical” approach which would allow us to accept a decontamination cycle with a positive BI / 2a negative remains very controversial too…. And in all these cases, we speak of “sterilization” of critical parts in contact with the primary components in contact with the product! Knowledge and understanding of the mechanisms of action of HIV2O2and its limits are clearly inadequate by industry. 
For current installations, make an AR to justify the maintenance of bad practices: NO! Because we do not want this type of RA to be reported by the industry and justifies the continuation of these practices. On the other hand, if an industrialist has a robust process with a reinforced cleaning procedure (not a simple rinsing), practices of manipulations limiting the bioburden …. He must contact his agency to study the process on a case-by-case basis and complete it if necessary before the next inspection. 
There will be no specific position on VH decontamination2O2 in Appendix 1 but the concept is already specified that the demonstration that each critical part must be in contact with the sterilizing agent.

Public reaction
We have been working since the 1990 years on insulators and we know the fragility of VH2O2 that we have already mentioned many times with the MHRA (ex MCA)! 25 years later, we continue to debate on the same subject. We also had problems with steam sterilization and it took us several decades to put this sterilization under control by controlling the quality of steam, the concept of vacuum, the design of loads …. So thank you for giving industry time to master the VH process2O2. If we apply the QRM approaches in Annex 1 for this process, the benefits far outweigh the risks when producing millions of units per aseptic process. It should also be noted that your “voice” echoes the regulatory level and that some FDA inspectors already apply your comments during inspections. The way you express your opinion on this subject can create some problems within the pharmaceutical industry and it would be useful if you had a more “positive” speech so that the industry is motivated in its search for the improvement of this. technology.

Andrew Hopkins response
It is true that we have had a lot of deaths because of a bad mastery of steam sterilization and especially in the United Kingdom. I do not want to have the same situation with the VH2O2 and if I wrote this blog, it is also to have this kind of discussions and not to continue to see hoses of several meters long “sterilized” at VH2O2. Associations such as PHSS, A3P write guides to better understand these processes. It is also necessary to involve suppliers so that the discussion is tripartite (Agency, industry and suppliers). Regarding my approach, if I had not communicated in this way, we would not have this debate and I received a number of arrogant comments but the problems related to VH2O2 are global and not limited to just one region.

Further discussion by J. Drinkwater / PHSS
Elements limiting the effectiveness of VH2O2 as mentioned, are validated by the PHSS. We have written a recommendation for the use of VH2O2 that we have shared with A3P and it appears that so far we have focused too much on the inactivation of BIs to evaluate the performance of the VH sterilization cycle.2O2(reaching up to 12 log reduction) without worrying about the condition of the surfaces to be sterilized. It is therefore necessary to consider all the physical factors having an impact on the robustness of the VH2O2and not just the inactivation of BIs.

Post Post Roundtable on VH2O2
In response to the questions raised by the Andrew Hopkins blog published on the MHRA website, the PHSS published in December 2018 a note for Guidance N ° 1 entitled: Insurance of Sterility for Container Closure in Direct Product Contact Surfaces in Aseptic Process Filling . Role of vaporized hydrogen peroxide bio-decontamination in a contamination control strategy combination Sterilization + Bio-burden control + VHP / VH2O2. Ensuring the sterility of indirect product contacting surfaces “.
This publication takes the risks identified by Andrew Hopkins, clarifies the limits of the process and suggests an action plan to reinforce the robustness of the VH process2O2 relying in particular on strengthening:

  • cleaning the equipment parts in contact with the primary components,
  • dressing practices during disassembly / assembly operations of these equipments,
  • means of control and limitations of bioburden on the equipment concerned.

The publication suggests a sterilization of clean steam equipment offline, their protection in Tyvek sav©until they are reassembled in insulators whose unidirectional air flow is kept operational and then decontamination of surfaces exposed to VH2O2.


3. Conclusions and next steps

The interactive exchanges of this round table clarified some of the concerns of the industry and to see a less directive approach on the subject of PUPSIT. Will these openings materialize in the new version of Annex 1? The answer may be known to industry during the publication of this article if the new version of Annex 1 is officially published. Moreover, according to Andrew Hopkins, a period of 6 months could be granted for the implementation of the requirements from the date of official publication but this period could be extended for certain topics in consultation with the EMA, the PICS and WHO. 

Share article

Marc BESSON, Sophie AMADIO, Julien TRIQUET & Eric HURTUBISE – GIC A3P Annex 1