inglasiasupporting global quality in the
manufacture and distribution of medicine


The EMA 2012 Guidance for Process Validation, by Dr Peter Calcott

This article was originally published on 19 February 2013, for the Master Control website and its publication GxP Lifeline

In February 2012, the Committee for Proprietary Medicinal Products(CPMP) issued the latest version of the new EU Guidance for Process Validation (PV).  At the end of October, the opportunity for consultation (comments period) closed, marking the time when the various committees review the comments for inclusion or modification of the document.  Sometime in 2013, the formalized document will be issued for use. However, generally, what is presented at this stage is often close to the final version.  So a review of the current thinking is very opportune at this time.

This document was issued about one year behind the comparable Food and Drug Administration (FDA) document and bears striking resemblance to that document but it is not identical.  The FDA document was very clearly written and has been previously reviewed in BioProcess International and Pharmaceutical Processing by the author.  In general, while the FDA document clearly states the applicability of the guidance to products, the EU version appears to create many exceptions to the implementation of a new paradigm to such an extent that implementation appears to be very complex if you want to get any value out of it.  But now down to the details.

The document, as issued, purports to be fully harmonized with ICH Q8, Q9, and Q10.  However, there is no mention of ICH Q11.  While it does give adequate reference to the three ICH guidances initially, it really does not link them in very clearly.  It describes that for already approved products, there are no new requirements imposed, almost begging the question of new requirements for new products.  And indeed, there are new requirements with respect to expectations of what submissions need to contain and for post approval inspections.  It goes on to describe that the new guidance is applicable to human as well as veterinary drugs and that it gives the industry options to adopt the new or stick with the old approaches.  However, if you chose to go with the new approach, you must not consider PV as a one-off event.  At this point they introduce the concept of a life cycle for validation and the idea of continuous verification being applicable if enhanced process development is done (read into that Quality by Design).

Surprisingly they focus on process validation for the drug product, indicating it is not necessarily applicable to active pharmaceutical ingredient (API) manufacture.  However, companies can use it for API manufacture and for biological products if they wish.  They further indicate that the new approach can be done prior to licensure or concurrently with licensure.  The key is to demonstrate process controls which have been shown at small or pilot scale are relevant at commercial scale.

Clearly this guidance is not an attempt to create a complete paradigm shift over to the new approach like the FDA document but rather a slow introduction of the new to eventually replace the old.  As such, the authors spend time going through the traditional approaches that we are all familiar with.  There are the three-batch concept (more if needed), the option of performing work at small scale so long as the scale-up factor is no more than 10 and that the strategy be included in the dossier with the reports and data available for inspection.  Really, there is no surprise in the section marked “traditional approach.”  There is a requirement for a short description of the process with a summary of the critical processing steps and control points to be monitored during validation.  The requirements for release specifications, in-process controls with acceptance criteria, and methods employed need to be defined and described.  In addition, any other non-routine testing, sampling plans, and methods for recording and analyzing data should be explained.  The proposed timeline for completion should be described.  When completed, the report should be prepared and be available for inspection.  It should comprise batch data, completed batch records, certificates of analysis and discussion and detail of any abnormal findings and events as well as conclusions.  Be prepared to submit variations for changes that are needed in your dossier if these changes or findings are significant.

The CPMP do give you the option for including continuous verification so long as you define it by some form of documented plan and preapproved acceptance criteria.

As a precursor to the new approach, the agency advises the need to describe in detail your control strategy (inputs, control points and parameters) necessary to assure that product meets your requirements (read critical quality attributes – CQA).  The inclusion of this data, while desirable for simple products, will be required for operations that are not traditional.  What constitutes non-traditional?  Biological products, non-normal manufacturing technology, pilot systems that cannot mimic commercial scale and controlled-release drugs.  I would translate that to mean anything they are worried about.  Therefore, when in doubt, a conversation with the regulators prior to submission of the MAA is a must to prevent surprises later.  The agency indicates that the number of batches (but never less than three) should be commensurate with risk imposed by the above factors as well as traditional processes such as aseptic processing.  It will be permissible to include significant amounts of relevant small scale data so long as the impact of inputs and process controls are linked clearly to the CQA’s.  This data should clearly defend the design space as appropriate and the relevance of the work done at small scale with reference to commercial large scale.

Once the work has been completed and submitted, the CPMP intimates that continuous verification can be used subsequently to assure the process outputs are controlled and that the process is stable.  How much is needed is dependent on your knowledge of the process and its control.  You are the expert. But the agency does acknowledge that understanding of previous and similar products can be introduced into the equation, as appropriate.  Your level of understanding obtained in the development phase, the level of sophistication in automation and in process controls and the complexity of the product and process will also influence how much data you will need. This needs to be clearly defined in your submission and then tested post approval.  If everything goes well you are off to a smooth start.  If you run into significant problems, expect to file adjustments as variations.  The CPMP advises that some data should be collected at large scale before commercialization which will be inspectable.  There is an option described for a traditional approach for licensure and then easing continuous verification into the picture post- licensure.  This is exactly what companies have done for legacy products in the United States when implementing the new FDA guidance. You can also use a hybrid of traditional and new approaches based on individual unit operations.

If you have developed a satisfactory scaled-down model for your commercial scale operations, then a significant amount of work can be done at the small scale without repeating it at large scale.  This represents an opportunity to establish clear relationships between inputs, control points and parameters with CQA without having to repeat all at commercial scale.

They also advise implementing an enhanced quality system as described in ICH Q10.  A critical element of the Q10 document is a robust change control process, to name just one.  While the results of the implementation of your enhanced quality system are not needed in the dossier, they warn that it will be a system that is going to be scrutinized during inspections to assure you are operating within your design space.

Now this is where the guidance starts to introduce exceptions in vast numbers. If you chose to implement the new approaches, then fine.  But if you stick with the old, then the exceptions really create more work. If your product and processes fit into the non-standard, then you may be in for more effort than you are used to.  What constitutes non-standard? Examples include:

  • Non-standard sterilization methods such as low dose gamma or terminal sterilization if non-pharmacopeia defined
  • Specialized dosage forms , such as lung powder delivery by metered dosages, suspension/emulsions or dispersed parenterals, low dose products
  • Aseptic processing
  • Lyophilization
  • Micro-encapsulation
  • Certain mixing and coating procedures especially as it applies to dosage forms
  • Processes new to the EU (but you may get around this if you have experience outside the EU with the processes)
  • Innovative delivery, micro particles, lipoparticles, nanoparticles and micelles
  • Real time release (parametric release)
  • And certain specialized processes with non-pharmaceutical traditional processes or new technologies

At a stroke of a pen, it appears that large spectra of pharmaceuticals are exceptions.  This makes it clear that any attempts on the manufacturer’s part to gain from the guidance or even stick with the old should be preceded by clear communication and discussion of your plans with the regulators.

So where does this leave you?  Well, for new products, make sure you perform your process development to a QbD standard or at least more comprehensively than previously.  With that data in hand, a meeting with the agency is a must.  Clearly lay out your plan for process validation.  Are you going to submit a traditional with or without continuous process verification?  Will you submit a full QbD or equivalent submission?  Under these circumstances you will be submitting the continuous verification model plan.  Will you submit a hybrid with some components traditional and others QbD?

With this plan clear, negotiate with the agency, assuring all the non-traditional natures of your product and process are clearly defined.  If you believe you need more than the traditional three batches, negotiate, as with the FDA, the commitment to perform the work post licensure, submitting the report at an agreed- upon time.  With agreement from the agency, begin the execution and hope that all the small- scale work predicts the large scale and everything runs smoothly.  Then you are ready to turn on your process for continuous monitoring and verification of the robustness of your process and submission. And good luck!


  1. Guideline on Process Validation (2012) – EMA/CHMP/CVMP/QWP/70278/2012-Rev1
  2. Peter Calcott (2011) How QbD and the FDA Process Validation Guidance Affect Product Development and Operations. BioProcess International 9(10) November 2011
  3. Peter Calcott (2012) FDA Process Validation Guidance has celebrated its first birthday! How is the one year old doing? Pharmaceutical processing January 2012
  4. FDA Guidance for Industry (2011) Process Validation: General Principles and Practices Jan 2011
  5. ICH Q8 2009 Pharmaceutical Development Q8 (R2)
  6. ICH Q9 2005 Quality Risk Management Q9
  7. ICH Q10 2008 Pharmaceutical Quality Systems Q10
  8. ICH Q11 2012 Development and Manufacture of Drug Substances (Chemical Entities and Biotechnological/Biological Entities) Q11


Link to original article: