Introduction
In this paper, I plan to discuss excipients and Quality-by-Design (QbD) in the context of the manufacture of pharmaceutical products using continuous processing. Continuous processing is not new; it has been around for many decades. It is just that the pharmaceutical industry has been even more reluctant to consider it than they have QbD. (QbD has only been around as a concept since the mid 1980s [1]; continuous processing has been around for a lot longer.) The reasons for this slowness to adopt new technologies and concepts in general are worth exploring because, in my opinion, they may provide some understanding of how long it could take for QbD to become generally accepted. (But hopefully it will be much quicker because continuous processing can bring real benefits to the pharmaceutical industry.)
The pharmaceutical industry has always hidden behind a façade that we cannot change things because the Food and Drug Administration (FDA) will not accept it. If this was ever the case, it is not so today. In my opinion the FDA has always tried to provide Regulations and Guidances that are not prescriptive, but simply set out the minimum standards the industry is expected to attain. However, the Agency rightly has an expectation that a certain standard will be achieved. Unfortunately, we have two things that work against that concept; the ‘corner-cutters’ and the ‘dinosaurs’.
There are people who will always seek to cut corners, or otherwise undermine the regulations, to maximize profits at the expense of patient safety (e.g. the generic drug scandal). The Agency has therefore responded to such events with tighter regulations and more specific Guidances. The FDA’s mandate is to protect the public health, and they must respond to such threats, and be seen to respond to them. But the Agency has also responded to strong scientific justification and withdrawn a Guidance document that was found to be flawed. They withdrew the Blend Uniformity Guidance in the light of the work carried out by the Product Quality Research Institute (PQRI) Working Group, and we now have a better, scientifically-justifiable approach.
There are some people in the pharmaceutical industry who may be likened to ‘dinosaurs’; not in age, but in thinking. They are not comfortable unless they are told exactly what to do; i.e. they do not have to think about how to justify their decisions. They prefer to work to a very detailed Regulation or Guidance. The following quote from Leonardo da Vinci (1452 – 1519) is as relevant today as it was in the great man’s day:
“Anyone who conducts an argument by appealing to authority is not using his intelligence; he is just using his memory.”
When validation was first proposed there were no detailed requirements; the wording was a lot more flexible. The pharmaceutical industry; however, went back to the FDA and asked how many batches they should make as part of a validation program. The FDA gave a number which, although statistically correct, was not well received, and we arrived at the 3-batch validation paradigm we still use today (unless we have opted for a QbD program). Statistically speaking, three batches are not particularly useful, and it can be argued that this 3-batch validation paradigm has held back the pharmaceutical industry from adopting better concepts in manufacturing (such as continuous processing). But the 3-batch validation paradigm is easily justified by the ‘dinosaurs’.
Now we have ‘Quality in the 21st Century’ which includes QbD and Process Analytical Technologies (PAT), and is arguably the most significant change in the pharmaceutical regulatory environment in 30 years, and which is designed to be flexible. Yet some sectors of industry are still reluctant to embrace it, i.e. step outside their comfort zone. The FDA is looking to reduce the regulatory burden on the pharmaceutical industry, but some parts of the industry it seems are not prepared to even consider it. Perhaps this is not so surprising given the history and our reluctance to accept change as a constant in life. It took several years for GMP to be well accepted; probably it will take just as long for QbD to be generally accepted.
However, there are also some people in the pharmaceutical industry who are prepared to seriously consider continuous processing. For example, Novartis has a project on continuous processing with the Massachusetts Institute of Technology. At least one other major pharmaceutical company also has an active project on continuous processing. Another encouraging sign is that some of the pharmaceutical equipment suppliers are offering continuous processing systems.
But without QbD, I am not sure that it will be easy to introduce continuous processing because we will not have the enhanced understanding of critical quality attributes (CQAs) of the active pharmaceutical ingredient(s) (API) and excipients, and critical process parameters (CPPs), and how they relate to the quality target product profile (QTPP) that QbD brings, and is necessary for the development and implementation of an adequate Design Space to achieve a robust continuous manufacturing process for a pharmaceutical product.
Continuous Processing
So why should we consider continuous processing? Quite simply, it is a logical extrapolation/conclusion from the combination of the QbD and PAT concepts. In addition, scale up of manufacture becomes much simpler (scale up of continuous processing means running the process for longer), operator safety is enhanced, and the opportunity for operator error is removed. With proper development of robust formulations through the application of the principles of QbD, and the requisite controls linked to PAT, there is every reason to believe that the finished product manufactured using continuous processing will be less variable than product manufactured using traditional batch processing. Mollan et al. [2], and Trout [3], among others, have reviewed the advantages of continuous processing.
The principles and details of continuous processing, as they might apply to pharmaceutical product manufacture, have been worked out in other industries, notably the food processing industry, but also the fine chemical industry. As was explained in an earlier article in this series [4], many pharmaceutical excipients are manufactured using continuous processing. It is also interesting to note that some of the unit processes and equipment used in pharmaceutical product manufacture are inherently continuous such as tablet machines, encapsulation machines, roller compaction, milling, sieving, spray drying, bottle filling, etc. Some of the current batch unit processes and equipment can be adapted to continuous processing, e.g. material dispensing, wet granulation, powder blending, liquid mixing, fluid-bed drying, film coating, etc.
Two of the key points to be resolved for any continuous processing operation are the time to achieve ‘steady state’ at startup, and when the steady state conditions are no longer maintained at shut-down. These will both directly relate to the amount of reject material at the beginning and end of a run. There may be ways to minimize this wastage using a combination of engineering design and equipment control. A third key component will be the effective integration of the different units in the equipment train with the appropriate PAT systems and control systems.
However, the transition from a batch manufacturing process to a continuous process for a pharmaceutical finished product may require that we re-think exactly how we do things so that we can get maximum benefit from continuous processing. For continuous processing we need to think about processing in a different way, and we should be asking and answering the following questions (and probably others too):
- What are we trying to achieve?
- How does the process or equipment operate, and what are the limitations of the process or equipment?
- What are the properties and limitations of the materials? (Know your excipients!)
- What else do we need to do to make it work in a continuous process?
- Can we adapt the batch process to a continuous or semicontinuous process (several smaller units operating in parallel, but staggered, referred to as a ‘multi-cell operation’)?
- Are the individual units of the equipment train matched for throughput?
- What in-process controls (e.g. PAT) can we use, and do we need?
- How can we integrate the equipment, sensors and in-process controls effectively?
By way of an example, let us consider lubrication of a tablet or capsule powder blend using magnesium stearate. Magnesium stearate is a hydrophobic boundary lubricant that can cause problems due to over-mixing (reduced tablet strength and extended disintegration and slower dissolution). In addition, the risk of magnesium stearate over-blending is increased as we scale up. With continuous processing, scale up simply means extending the run, and the size of the units in the process equipment train will be smaller than the processing units in commercial scale batch processing. Thus once we have established the continuous process, validation will be more straightforward since there will be no further scale up.
We typically add magnesium stearate to the blend after the main mixing is complete, and then further blend for a short period to disperse the lubricant. When we lubricate a powder blend, what are we looking to achieve? Is it a homogeneous mix of magnesium stearate, or a mix of magnesium stearate that is sufficient for its intended use, and how do we achieve it? We need a mix whereby there is sufficient magnesium stearate available within a unit dose to lubricate the granule, but we do not want to over mix so that we effectively coat the total granule or blend surface with a hydrophobic layer of magnesium stearate. Neither do we want the magnesium stearate to be insufficiently mixed with the other components of the formulation such that it compromises product content uniformity. Effectively we require an incomplete film of magnesium stearate on the surface of the granule or powder blend. The way we typically achieve this in batch processing is to blend in the magnesium stearate for a relatively short time. But there may be another way; we could add a much smaller amount of magnesium stearate and blend for longer. For continuous processing the latter may be the preferred option.
However, conversion to continuous processing will seldom be easy, and we must consider how we can best make it work. We must first consider what project-specific information we need to have the best chance of success. The detailed requirements will need to be assessed on an individual project basis. We would not expect to need the exact same kind of information for every project; the API, formulation and route of administration together will influence the details of the information required. However, in general terms, we will require the QTPP and a robust formulation and process.
The QTPP provides the details of the release specification for the formulation and thus sets the acceptance criteria.
A robust formulation and process can be defined as;
A formulation and process which together provide for the manufacture of the drug product, and which together are able to accommodate the normal variation in both APIs and excipients without compromising any aspect of the safety, efficacy and purity of the drug product during manufacture, stability, in vivo performance of the drug product, or any other attribute of the drug product critical to the patient’s care and well-being.
Excipients in Continuous Processing
As with any formulation and manufacturing process, excipients will be an important part of, and have a significant influence on, the design of a robust formulation and continuous manufacturing process for a medicinal product. As with batch processing, we will have to deal with the inherent variability of our excipients (as well as with the API). This will not change, and cannot change. This is why we need better understanding of our excipients, and particularly their variability and limitations.
So how do we accommodate such variability into continuous processing? There are at least two approaches:
1. Use the multi-cell approach and appropriate end-point detection to ensure the output from the particular unit process provides a consistent input to the next unit process in the process chain.
2. Design the formulation and process so that we can achieve a degree of overprocessing that still gives a satisfactory output, and that provides a consistent input to the next unit process in the process chain.
An example of the application of the multi-cell approach could be wet granulation, where three or four small high-speed mixer/granulator units are used in staggered rotation and the granulation taken to its end-point for each small granulation before being transferred to the next unit in the equipment chain for subsequent processing.
An example of deliberate over-processing would be the magnesium stearate lubrication example cited above.
However, we will need the better understanding of our excipients, because without such enhanced understanding, we will not be able to establish our design space that will allow us achieve a robust formulation and process. Without a robust formulation and process and its inherent design space, it is unlikely we will be able to establish a continuous process that is sufficiently robust for the needs of the pharmaceutical product, and ultimately the patient.
The problem of establishing the design of experiments and thus the Design Space, and particularly how to incorporate excipient variability into the design of experiments, is always going to be a factor in any QbD development program. As has been stated in a previous article in this series, if we think about what we are trying to achieve, and the opportunities QbD affords, there should be no need for the formulation developer to be looking to obtain lots of excipient at the limits of specification; they are unlikely to be available, and QbD provides us with better options [5]. As stated in the previous article, these options include:
- Alternate grades (based on the distinctions used to separate the grades on the market; and including use of a technical grade material that has a different set of specifications).
- Blending different grades.
- Fractionation of the grade (e.g. sieve fractions).
- Dilution (using some inert material).
- Using chemically different but closely related materials (e.g. polymers with different degrees and ratios of substitution).
We also have the option, in principle, to develop formulations that deliberately use two grades of an excipient, or even two different excipients, in combination to balance out variation in e.g. another component such as the API. However, it is difficult to see how this might apply in continuous processing, except possibly in a multi-cell approach. Over-processing in some way may be easier to implement in continuous processing.
Conclusion
To achieve a successful implementation of continuous processing for pharmaceutical product manufacture, as with QbD, a better understanding of the excipients we use will be a key component of that success. This understanding will probably include a better understanding of how excipient chemical composition and physical structure relate to performance. It will include a better understanding of the limitations of our excipients, and probably how these limitations relate to chemical composition and physical structure. Superimposed on all this is excipient variability, and we have to develop formulations and processes that are able to routinely accommodate the inherent variability in all our component materials, rather than try to restrict it through overly tight excipient specifications that cannot guarantee continuity of supply of our excipients, and thus products for the patient.
Continuous processing is in the future still, but the FDA has given the pharmaceutical industry the tools with the introduction of the QbD and PAT concepts. It is now up to the industry to make it work. However, we will still have to justify our implementation and our decisions and choices for the formulation, processing and equipment train using hard science-based principles. This will always be the necessary.
This was the final article in this series on Excipients in a QbD World. I hope this article and the preceding ones have provided useful information for you, and provoked useful discussion. I would like to thank the Editors and Russell Publishing for the opportunity they have provided, and you for reading these articles.
References
1. Juran JM. Juran on Quality by Design, Juran Institute, Inc., New York, NY, 1992.
2. Mollan MJ and Lodaya M. Continuous Processing in Pharmaceutical Manufacturing, http://www.ieor.berkeley. edu/~shen/ieor298/pdd/ContinuousProcessinginPharmaManufacturing.doc
3. Trout B. Next wave model, World Pharmaceutical Frontiers, March 2009, 74 – 76. http://www.worldpharmaceuticals. net/editorials/015_march09/WPF015_nextwave.pdf
4. Moreton RC. Functionality and Performance in a Quality-by-Design World: Part 1. (2009), Am. Pharm. Rev., 12, (1), 40 – 44.
5. Moreton RC. Functionality and Performance of Excipients in a Quality-by-Design World: Part 2 Excipient Variability, QbD and Robust Formulations, Am. Pharm. Rev. (2009), 12, (2), 24 27.
Dr. Moreton has over thirty years’ experience in the pharmaceutical industry. He has worked as a formulation scientist developing a variety of different dosage forms, and has experience in the design, development, scale-up, technical transfer and validation of drug products and associated analytical methods, both during clinical development and eventual transfer into commercial manufacture, and working with licensing partners and contractors. He has also worked in QA/QC, Regulatory Affairs and Technical Support in excipients and drug delivery.
He is a past Chair of the AAPS Excipients Focus Group, and of IPEC-Americas. He is a member of the International Steering Committee of the Handbook of Pharmaceutical Excipients, and of the USP Expert Committee—Excipient Monograph Content 2. He has authored and co-authored scientific papers and book chapters, and lectured extensively in the areas of excipients, drug delivery and formulation at universities, training courses and symposia in the U.S. and Europe.
Readers may contact the author directly at: [email protected]