What I hope to do in this column is to go through some of the points to be considered in establishing a user specification for an excipient. However, I am going to start with some more general considerations by way of introduction, and then look at how we might set specifications for excipients.
Background
To many people, the specification defines the material. However, we all know that products made with APIs and excipients properly conforming to their specifications in all respects can sometimes fail, for reasons we do not fully understand. Sometimes the cause is human; for example, there was an error in operating the process, or taking samples. Sometimes there is no explanation, and sometimes it becomes apparent on investigation that there is some property of a component that needs to be added to the specification. This component can be the API, or it can be an excipient.
Over the past 40 years drug molecules have become more sophisticated and more potent. Manufacturing equipment has become faster, and analytical methods have become more sensitive. In general, and in this author’s opinion, some formulations appear less robust. By contrast, we can argue that excipients have largely stayed the same, but we are expecting more from them. Excipients are an important part of formulation science, and it seems logical to suggest that if some of our formulations appear to be less robust, then part of the problem may lie with excipients, and in particular how they are used. And this may be a key point in some cases.
In part this may be due to the fact that formulation science is no longer taught in pharmacy schools, and has not been taught elsewhere until the recent introduction of pharmaceutical sciences curricula. Even with the recent introduction of such curricula, is our teaching about the uses and limitations of excipients sufficient for the needs of future formulation scientists? In part, it also implies that we do not understand enough about the excipients we use in the context of the modern sophisticated molecules we have to formulate. This is a theme that has been discussed before in this series of articles, and elsewhere [1].
Excipients, are an important part of any pharmaceutical formulation, and are an important part of the Quality by Design (QbD) design space. Not all excipients will have a critical impact on the performance of a particular pharmaceutical product. For those that do have a critical impact on the performance of the pharmaceutical product, they need to be built in to the design space in some way. In order to incorporate the excipient into the design space, we need to understand what aspect of the excipient specification is responsible for its impact on the formulation, and how it varies. Sometimes it appears obvious, sometimes it is less obvious, and this is where experience and small scale experiments might be needed prior to deciding on the Design of Experiments (DoE).
What Do We Mean by Specification?
Before we can build excipient performance and variability into our DoE, we need to understand what we have. For this, we have to look at the specification; but which specification? And what do we mean by specification? If we look in the dictionary, we find that ‘specification’ means ‘something specified’ – not particularly helpful since according to the same dictionary ‘specify’ means ‘to mention or name explicitly’ [2]. A more helpful definition is presented in the Joint IPEC-PQG Guide for Good Manufacturing Practices for Pharmaceutical Excipients [3]:
‘A list of tests, references to analytical procedures and appropriate acceptance criteria that are numerical limits, ranges or other criteria for the tests described for a material’. But what are the appropriate procedures and acceptance criteria for an excipient? The answer to this question is one of the keys to successful formulation design and development in our Quality by Design (QbD) world. Once again, I must emphasize, “Know your excipients!”
Historically, excipients have mostly been sold and purchased to rather broad specifications (also referred to as the sales specification). Typically this is the specification contained in the pharmacopeial monograph (compendial specification), or something analogous for excipients that do not have a pharmacopeial monograph. However, compendial specifications were often set many years ago, and may have been deliberately set wide because process controls and test methods were less precise than they are today, or there was some variability observed and the wide limits were designed to allow for that. In a QbD world such wide limits may not always be appropriate.
Besides the sales specification there can also be other types of specification associated with an excipient; these may include: in-process specification, manufacturing specification, customer specification, end of shelf-life specification, etc. In general, the sales specification will be wider than all these other specifications. If it isn’t, then it is likely that the excipient will not routinely meet specification, and then there will probably be problems with excipient lot selection and inventory control.
The most important specification in a QbD world is the customer specification. What does the customer want? Unfortunately, the customer does not necessarily know what they want, and we end up with two extremes. There are customers who will simply accept the sales specification for the excipient and assume (hope!) the excipient performs consistently. Other customers try to over specify their excipient requirements in the mistaken belief that by tightening the specification they can somehow cancel variability. This belief is based on the misunderstanding that the variability in the product is due to the variability in the excipients. (In the author’s experience, the active pharmaceutical ingredients are often more variable from a physical perspective than excipients.) For some applications there will be no problems that can be attributed to excipients. The formulation is sufficiently robust to accommodate the variability in the API and excipients. However, for other formulations there will be unexplained failures and out-of- trend results some of which may eventually be assigned to an excipient. Then the fun begins as the customer and manufacturer try to resolve matters, and arrive at a workable specification. The emphasis is on ‘workable’ because it will inevitably be a compromise. There is inherent variability in all manufacturing and excipients are no exception. We must also remember that the scale at which most of the common excipients are manufactured, and how they are manufactured (using some form of continuous processing) mean that it is not always possible to control things to the level the customer may desire (think they need).
Setting Excipient Specifications
So how can we set good specifications that the excipient manufacturer can meet on a routine basis, and that allows the customer to have confidence that the excipient will perform satisfactorily in the manufacture, storage or use of the finished pharmaceutical product? There are two interrelated components; the inherent capability of the excipient manufacturing process, and its inherent variability, and the factors in the formulation and processing, and how they interact, in the manufacture of the pharmaceutical finished product. However, this almost becomes a circular impasse; we need to know what influences the formulation before we can begin to put together a meaningful specification for the excipient, but we need to understand the inherent variability of that particular facet of the excipient before we can build it into the DoE.
In practice, we need to understand what we are looking to achieve and how the excipients can be used, in particular any limitations on their performance (again, “Know your excipients!”). Then we can begin to make some progress. In a QbD world we are not going to pull our DoE out of the air. We are going to undertake what effectively amounts to a risk assessment to decide which particular properties of an excipient are likely to have a major influence on a particular characteristic of the pharmaceutical finished product. In other words, even before we start any practical work, we are going to consider all the attributes of all the excipients and the API and look to see if there is a reasonable likelihood that they can influence the quality target product profile (QTPP). Then we can begin to establish our DoE. It may be appropriate to carry out some limited preliminary experiments to examine the influence of a particular excipient property to see if it should be classified as a potential critical quality attribute (CQA).
It seems that there may be some confusion that everything has to be built into the DoE from the outset. That is not what the FDA requires. What is required is that we should be able to justify our formulation design space, and the DoE will obviously be a major part of the justification. However, there is nothing to stop us performing some preliminary experiments to clarify issues, but we should document what we do and include those preliminary experiments in our justification. We can also use literature reports, internal company reports and our own experience to justify the DoE, and this is where our risk assessment fits in.
In a QbD world, we formulation scientists need to get a lot smarter at setting specifications for our excipients, but we must not over-specify because that may hurt us in the longer term. There is no point in trying to impose a specification that the excipient manufacturer cannot meet on a routine basis. As has been discussed elsewhere [4], it is a disaster waiting to happen. I have discussed in previous column the perils of excipient lot selection. In order for it to work about 50% of the manufactured excipient lots should comply with the required specification.
So what is the best recourse if you find that you have a formulation problem requiring excipient lot selection that does not meet the above criterion? There are really only two options besides accepting excipient lot selection. One option is to reformulate the product to include an alternate excipient or combination of excipient and processing that can accommodate and neutralize the unacceptable variability. The second option is to look at the API because it is presumably some interaction, direct or indirect, that involves the API. Is it possible to engineer the physical form of the API in some way to allow the manufacture of a pharmaceutical finished product with more consistent performance during manufacture, on stability or after administration to the patient? Unfortunately, both options require time and resources and senior management is usually unsympathetic to such delays. This is where QbD should really help because, if done properly, it can help us to avoid such late reformulations.
Assuming that we have developed a robust formulation using an appropriate DoE, what should we include in our excipient specification? Let us look at the requirements for the API and drug product. We are concerned with four things: purity, efficacy, safety and that the product has been manufactured to the appropriate levels of good manufacturing practice (GMP). With excipients we should not be concerned with efficacy; excipients should not possess therapeutic efficacy, although they may have physiological effects. However, we are concerned with purity, safety and GMP. In addition, we are concerned with the performance of the excipient (how it works, and how consistently it works in our formulation).
Based on these considerations, we can begin to define what tests and limits need to be included in our specification for the excipient; we need to include tests that address chemical composition, tests that address the physical form of the excipient, and possibly tests that address performance. Excipient composition was discussed in a previous article in this series [1]. Understanding and monitoring excipient composition will address the issues of purity (do we have the correct chemical material?) and safety (are there undesirable toxic components present in the excipient above the recognized safe limit?) These types of tests are often those included in the main part of a pharmacopeial monograph.
The tests for physical form include those tests that are typically contained in the Labeling Section of a USP-NF monograph. These labeling requirements are generally tests that can be used to distinguish between different physical grades of a pharmaceutical excipient that are commercially available. These will relate to performance in some way, otherwise why would there be different grades? But there may be other, more relevant tests that relate to performance that are not included in the labeling section; possibly composition tests, possibly tests that are extra to the monograph. We must remember that excipient performance can relate to manufacturing, stability or in vivo release of the API from the pharmaceutical finished product.
Some excipient monographs of the European Pharmacopoeia contain non-mandatory sections relating to functionality-related characteristics (FRCs). This approach is not an option in the US. There cannot be a non-mandatory section in a USP-NF monograph. In the USP-NF, any such tests might be included in the labeling section, if appropriate.
Before we include any customer specific requirements in our monograph, we need to know what we are dealing with, and what control the excipient manufacturer has, if any, over the particular parameter. As stated above, it does not make sense to include parameters or limits in our specification that the excipient manufacturer cannot meet. Let me illustrate this by way of an example.
In most instances in oral solid dosage forms, particle size is a factor that should be investigated for most excipients. (An exception can be made for wet granulation binders that are prepared as a solution before adding to the granulator.) But before we start rushing to produce tighter particle size specifications for all our materials we need to take a step back, and look to see if the excipient manufacturer has any control over particle size. I can already hear people asking, “But why wouldn’t they?” For many materials they do, but if we think about the starches, does the starch producer control the size of the starch grains, or does Mother Nature? The answer is that the size of the starch grains is a function of the plant source (species) and the growing conditions, and will vary from season to season and region to region. When we process corn, potatoes or any other starch source to obtain the starch, all we are doing is releasing the intact starch grains, not changing their size. If we did change their size we would be likely causing irreparable damage to them, and destroying the very characteristics we are looking for in order to use the starch as an excipient. The lesson here is to only specify what can be controlled during manufacture, but perhaps to monitor what cannot be controlled.
So to get to where we want to be – a ‘workable’ specification that the manufacturer can meet and that will provide an excipient that the user can accept and rely on, we need to include sufficient tests in the specification to meet the requirements of composition and safety. We may also need to include tests that will distinguish between different physical grades of the same pharmaceutical excipient. The key question is what tests beyond this should we include? My answer would be as few as possible. There will be times when this will be necessary, but they should be exceptions rather than the norm. For example, special limits on trace components to improve stability could be considered, provided the excipient manufacturer is confident they can meet them as required. However, trying to include as many extra parameters as possible in the hopes that you can avoid some as yet undiscovered problem is, in my opinion, a pointless exercise that may cause the user more grief than benefit. The old adage, “Be careful what you ask for!” is very relevant, because if we ask for it, and put it into our application, we will have to live with it. For excipient specifications, as for many things, the ‘kiss’ principle applies (Keep it simple, stupid!). In my opinion, we should be looking to use QbD to avoid the need for too many customer-specific tests and limits in our excipient specifications.
I hope this article on excipient specifications has been helpful. Some of the ideas may be controversial to some of you. They are offered in the spirit of helping formulation scientists, and others, understand that although they do not treat disease or improve patients’ quality of life, we formulation scientists had better take our excipients seriously and recognize that they have limitations, as do excipient manufacturers. In the next article of this series I will consider new excipients.
References
1. Moreton RC, Functionality and Performance of Excipients in Quality-by- Design World Part VI: Excipient Composition. Am. Pharm Rev. (2009), 12, (7), (Nov/Dec), 48-51.
2. The Merriam-Webster Dictionary, 1974, Pocket Books division of Simon & Schuster, published by agreement with G&C Merriam and Co.
3. IPEC-PQG Good Manufacturing Practices Guide for Pharmaceutical Excipients, 2006.
4. Moreton RC, Functionality and Performance of Excipients in Qualityby- Design World Part II: Excipient Variability, QbD and Robust Formulations. Am. Pharm. Rev. (2009), 12, (2) (Mar/Apr), 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 IPECAmericas. 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 can contact the author directly at: [email protected]