Applying QbD to Excipient Formulation and Development

This article is part of PharmTech's supplement "Solid Dosage and Excipients 2010."

Since the concept of quality by design (QbD) emerged nearly a decade ago, the pharmaceutical industry has been trying to understand and put into practice a more flexible and risk-based approach to manufacturing. Although the introduction of US Food and Drug Administration guidances and International Conference on Harmonization (ICH) guidelines—ICH Q 8, 9, and 10—have provided some details on how QbD can be applied, many questions remain. The development and use of excipients in drug-product formulations has posed several challenges, including understanding excipient variability and functionality when implementing a QbD-based approach.

To provide some answers and potential solutions to these common challenges, Pharmaceutical Technology gained input from the following industry experts: William F. Busch of Dow Wolff Cellulosics–Pharmaceutical Excipients, a business unit of Dow Chemical (Midland, MI), with Tom Frazier, product market manager for Carbowax Sentry, a product line of Dow Chemical; Stuart Porter, senior director of pharmaceutical research and development at International Specialty Products (ISP) in Wayne, NJ; and John Langridge, manager of research and development/quality control at DMV-Fonterrra Excipients (DFE) in Goch, Germany. They provide examples and advice on excipient selection, regulatory expectations, supply-chain considerations, and more.

Excipient selection

PharmTech: Can you outline how QbD may be applied in excipient selection and understanding the functionality of the excipient in the formulation?

Porter (ISP): QbD comprises many steps that typically involve the following:

• Examining all of the formulation and process parameters, typically by employing a statistical design-of-experiments (DoE) approach to assess the impact of each parameter and the interaction between parameters on the final product. The outcome of this approach would be to facilitate effective formulation and process optimization.

• Conducting a risk-assessment analysis of the formulation and the manufacturing process to determine which parameters have the greatest impact on product quality and characteristics. This analysis involves examining all parameters that are likely to have an effect, determining what that effect will be, what the likelihood that an event that could have an effect will occur, how frequent that occurrence might be, and how likely is it that such an event will be observed. For each of these characteristics, a rating system is used to apply a total score to each parameter. Parameters that receive the highest score are deemed to pose the highest risk and should be be investigated to reduce that risk to acceptable levels.

• Based on the results of the DoE and risk assessment, a design space can be created that will define the operating ranges (both formulation and process) within which a product can be manufactured and be expected to consistently meet the required quality standards and performance attributes.

Ultimately, when this process is established, the goal is to minimize reliance on conventional in-process and post-manufacturing quality-control testing of samples by implementing an effective process-analytical-technology (PAT) strategy that enables critical process parameters (CPPs) to be monitored. This process uses tools that have feed-back and feed-forward capabilities, so that operation within the required design space is assured.

In terms of how QbD implementation might affect excipient selection, there are two key considerations. First, during the application of statistical DoE, it is likely that the impact (in quantitative and functionality terms) of the excipient will be determined. As a result of this analysis, selection of a particular excipient may well rest on how sensitive the formulation (and the ultimate manufacturing process) is to potential variation in that excipient. Such sensitivity should be considered in terms of the range of excipient content that can be tolerated without having a demonstrably negative effect on the formulation, and in terms of the likely impact of batch-to-batch variability in that excipient. Potential negative effects should be considered with regard to how the excipient affects functionality of the dosage form (e.g., drug-release characteristics) as well as the chemical stability of the active pharmaceutical ingredient (API) in the dosage form.

Second, during the risk-assessment stage, excipient selection may hinge on the size of that excipient's risk factor. For an excipient supplier, complying with the demands of a pharmaceutical manufacturer during this processes can be troublesome. The reason for this is that a pharmaceutical manufacturer will often ask for samples of a particular excipient that represent the extremes of the specification range. Although an excipient supplier will, out of necessity, set a range of specifications for a particular material, that material is often manufactured to tolerances that are much tighter than those depicted by the specification ranges. Production processes are then designed to target the middle of that range to avoid out-of specification production. Hence, it may not be possible to supply samples that represent the extremes of the specification ranges.

Langridge (DFE): We see an increasing trend in customers' application of QbD to excipients. In terms of excipient selection, there is more interest in matching excipients to processes in addition to traditional excipient selection using preformulation and excipient compatibility studies. Examples include selection of materials with good recompacting properties for dry-granulation processes (e.g., roller compaction) and selection of materials with low sensitivity to lubricants and to compaction speed for scale-up.

With regard to understanding functionality, there is a definite trend in testing the ranges of excipient specifications during the development process. This, of course, depends on the criticality of the excipient for correct performance. Although not directly relevant to oral solid dosage forms, QbD studies are always applied to carrier lactose for dry-powder inhalers, to the extent that several customers now frequently request commercial batches, not just trial batches, targeted to a particular point in the overall particle-size specification.

Additionally, we see more interest in the understanding of how nonspecified properties (i.e., not specified on a certificate of analysis, or CoA) of an excipient may affect processing. One example is the role of amorphous lactose in the compaction of spray-dried lactose, and the proper packaging and environmental controls needed to maintain the amorphous lactose. Another example is the role of viscosity of polymeric disintegrants and the effect of viscosity on granulation and disintegration.

Critical process parameters

PharmTech: What CPPs are affected by the choice of the excipient and the interaction of the excipient in the formulation for a solid dosage product?

Porter (ISP): CPPs may well vary from process to process and product to product. Thus, what a CPP is for one product may not be for another product. In addition, what may be deemed to be CPPs for the manufacture of an oral dosage form that is presented as an immediate-release powder-filled capsule product is likely to be entirely different when manufacturing a film-coated, sustained-release matrix tablet. If the criteria for excipient selection are scientifically valid, then the choice of a particular excipient may have no impact on CPPs. Under other circumstances, selection of a particular excipient may affect several CPPs.

Langridge (DFE): There are some very obvious examples. Of course, one should always use excipients designed for direct compression (DC) in a DC process. One example we experienced regarding what can go wrong when two excipients interact in a process is the failure of a DC formulation to compress on scale-up. The root cause in this instance was found to be unintentional over-mixing of the lubricant (magnesium stearate) with sieved lactose in the rotary feed-frame of a tablet machine. An open feed-frame had been used in development.

In terms of material specification, we have examples of differing disintegration time of tablets when different sources of croscarmellose sodium were used. The croscarmellose had a different particle-size distribution that was not noted on the CoA.

Design space

PharmTech: What should be taken into consideration when devising the design space for a solid dosage formulation?

Busch/Frazier (Dow): Dow believes there are two key considerations when thinking about design space. The second one is what most folks think of when they hear the words 'design space,' which we define as 'process design space.' What are the potential interactions between the various unit operations in the drug manufacturer's process and the drug dosage formulation that control the final product efficacy? This thinking assumes that the formulation is fixed and that the only way to handle formulation variability, which may occur via any of the formulation components (e.g., the API or excipients), is through process changes that are well understood and documented to 'correct' for formulation variability.

For example, when using a formulation with polyethylene oxide as a hydrophilic matrix-tablet rate controlling polymer, if changes in molecular weight occur between one lot and the next, one may have the flexibility in the process to change the compression force during the tableting operation to attempt to compensate for excipient variability. However, Dow believes that another, and in fact the first, important aspect of design space is the formulation design space, which is the understanding of potential interactions between the formulation components, and how individual component variability affects the finished dosage-form's efficacy.

If one understands the function of the formulation components, it may be possible to reduce or minimize finished product variability by making slight changes to the formulation. Using the example above, with a well-documented design space, if polyethylene oxide is being used to control the dissolution profile of the API and if a lot of polyethylene oxide with a lower molecular weight is being used compared to the previous lot which had a higher molecular weight, one may have the ability to increase the percentage of the polyethylene oxide in the finished tablet to compensate for the lower-molecular-weight material being used.

A second method might involve using a blend of two grades of polyethylene oxide to achieve a given functionality. If one of the grades shifts in molecular weight, then its percentage in the formulation could be changed to compensate for this shift, thus resulting in a formulation with consistent efficacy but inconsistent composition. The benefit of formulation flexibility is that it does not force the drug manufacturer to rely solely on the process to smooth out the variability inherent in the formulation. Slight changes in formulation and process can work together to minimize the variability inherent in the formulation (1). A second example is colored-tablet coating formulations, where slight variations may be made in the formulation to achieve a consistent finished tablet color due to variability in the various color components used to make the master batch.

Langridge (DFE): One should consider the risk analysis, which needs to include factors that may affect product performance outcomes (e.g., critical tablet properties such as dissolution rate and stability) and product processing ability (e.g., the ability to perform high-speed compaction). As an example of the former, we know from work in our laboratory that the right combination of DC filler-binder and disintegrant is vital to maintain stability of disintegration, and thus, dissolution. Dibasic calcium phosphate dihydrate has the potential to dehydrate under accelerated stability conditions with consequent changes in disintegration. It is essential that the right disintegrant is selected in formulations containing high levels of this filler to prevent adverse stability findings. Excipient companies can help by providing formulators with enough information about the criticality of excipient interactions to enable formulators to make a well-informed risk analysis and to construct the best design space.

Process understanding

PharmTech: In your experience, how has using a QbD approach improved process understanding? What are the advantages and disadvantages of applying QbD to solid dosage formulations?

Porter (ISP): In addition to the answer provided to the first question, in general, QbD is a concept that requires a pharmaceutical manufacturer to design a product on the basis of employing tools (e.g., statistical DoE) that provide a scientifically valid framework for identifying all key process and formulation parameters, and thus facilitate effective process optimization. That process, coupled with the implementation of an effective risk-assessment strategy ensures that factors that impact product quality (in its broadest sense) are clearly identified. When conducted effectively, such an approach will provide levels of understanding of the process that were never possible before, especially given that the critical information is developed using tools that are more scientifically valid than the empirical approaches (often based on personal bias) used in the past.

If used effectively, the implementation of QbD will ensure that every batch of manufactured product conforms to specification. In addition, the need to rework or recall batches as a result of failure to meet specification will potentially become a thing of the past.

In terms of disadvantages of QbD implementation, certainly, upstream costs such as those associated with formulation and process development, are likely to increase, but these increases are likely to be more than offset by downstream savings because rejected batches and analysis of cause of failure, can be eliminated. The timeframe for product development is likely to increase as well, which may be somewhat problematic for generic-drug manufacturers who are already working under severe time constraints. The implementation of QbD could also complicate matters when product development is outsourced. In these instances, those responsible for product and process development are farther removed from those ultimately responsible for product manufacture.

Langridge (DFE): The biggest lesson to be learned is that studies can find the points at which a formulation will fail, and a prudent formulator will place his product well away from the edge-of-failure. I have two examples.

The first instance involved a desire to minimize the amount of magnesium stearate in a formulation because of the well-known risk of overlubrication. The lubricant level selected from development trials was insufficient to prevent the slow build-up of a film on the punches during long tableting runs, thereby, resulting in sticking. We learned to assess the lubricant level in an experimental design and to use an amount well above the level that appears to prevent sticking in small-scale trials. This approach provides assurance against scale-up problems.

A second example focused on minimizing the coating time for an enteric-coated tablet by applying the thinnest coat consistent with enteric protection. Periodic failures occurred during production, and the sampling size for disintegration testing, based on sampling by attributes, had to be increased enormously with consequent loss of yield. We had to reprocess several batches. The net result was that the coating process was worse than it would have been if the product had originally been placed away from the edge-of-failure with a slightly thicker coat.

It's hard to identify the disadvantages of applying QbD to oral solid dosage forms in principle. For innovative companies, however, it is possible that new drug availability constraints can limit the extent of experimental trials that can be performed.

Suppy-chain challenges

PharmTech: Supply-chain integrity for pharmaceutical ingredients is of great importance to the industry. In terms of excipient supply, how have supply-chain practices (e.g., audits, vendor selection, and qualification) evolved to meet this challenge?

Busch/Frazier (Dow): Dow has been a member of the International Pharmaceutical Excipients Council of the Americas (IPEC–Americas) since its inception, and we use IPEC's excipient guidelines in the manufacture of our excipients. These guidelines include aspects concerning composition, stability, CoAs, quality agreements, good manufacturing practices, and notification of change practices.

At Dow, we also conduct regular, customer audits at our various excipient-manufacturing sites, done under the mantle of transparency, continuous improvement, and customer access. Regularly scheduled audits, including multiple customers, cuts down on the number of tailored audits required.

Our vendor selection is also quite rigorous. As a global supplier, we require our vendors to have globally acceptable capabilities. This requirement helps us to use recognized, global leaders that have the reach and technical expertise we require to support their products in our operations . In terms of qualifications in the last year, none of our suppliers have changed significantly.

Porter (ISP): Audits of ingredient suppliers are not new. It is likely, however, that audits have become more rigorous in recent years. In a similar vein, vendor selection and qualification practices have evolved over a number of years, and the requirements have become more rigorous over time.

One area receiving more attention lately is the use of third-party audits to verify whether excipient suppliers comply with suitable quality standards. Although excipient suppliers value audits as opportunities to improve quality practices, many suppliers have several hundred customers, and it is simply not practical for every or even most customers to conduct an audit. However, this is in conflict with regulatory and internal pharmaceutical company requirements. Thus, establishment and acceptance of third-party audits are starting to emerge.

In terms of maintaining security of the supply chain, pharmaceutical companies appear keen to ensure that more than one vendor is qualified for each and every raw material used.

Langridge (DFE): Audits increase annually. There is a desire among many companies to qualify more than one supplier of excipients in case of supply failure. This trend tends to imply that companies are starting to test specifications that encompass a range of suppliers rather than limit tests to a single supplier's specification.

Regulatory considerations

PharmTech: From an industry perspective, what would you identify as the key regulatory considerations with regard to excipient manufacture and use?

Busch/Frazier (Dow): Geographic and regional differences that need to be matched with a global selling proposition present challenges for excipient manufacturers, particularly in emerging geographies where customer requirements are quickly evolving to mimic Western standards. For example, many Indian companies are quickly changing to meet European standards. Other issues that are of importance include the need for regional regulatory expertise. An ever changing supplier-customer mix will make reputation more important.

The emergence of the IPEC Federation cannot be understated (see more details on the federation). The federation will coordinate and harmonize the activities of the various individual IPEC organizations. The relatively recent harmonization of the monographs between the US, Europe, and Japan for certain excipients such as hypromellose, has been a tremendous benefit to the industry as well. Harmonization will reduced redundant testing and make it much easier to provide products on a global scale. The IPEC Federation will similarly benefit the industry by issuing guidance documents that are global in nature and thereby prevent (or at least minimize) geographic variability.

With regard to any functionality guidance, in the United States functionality tests are part of the US Pharmacopeia general chapters. This location for such tests is critically important because excipients can be used in a variety of applications, each of which may have certain functional needs. The end user can select those tests appropriate to their application, and omit those tests that are not applicable. To require functionality testing in the monograph forces all tests to be run on the product, some of which may have no bearing on the performance of the excipient for its specific intended application. This extra testing causes unnecessary work for the excipient supplier and passes additional costs to the drug manufacturer.

If functionality cannot be predicted by any of the various tests required by the monograph, additional tests that can predict performance in the specific application can be jointly developed by the excipient supplier and user through a quality agreement.

Porter (ISP): Although not a recent event, the requirement that excipient manufactures comply with some semblance of good manufacturing practice has become more rigorous over time. This requirement has led excipient manufacturers to adopt many of the practices that have long-been used by pharmaceutical manufacturers (e.g., implementation of change-control procedures, conducting process validation programs during process start-up and when significant process changes are to be implemented or when manufacture of a product is moving to a different manufacturing site).

Excipient functionality is a thorny subject because it requires that meaningful functionality tests are identified—these tests also must be simple and inexpensive to implement and operate. The result, quite often, is that the test has little resemblance to a true functionality test, but rather produces data that hint at, rather than confirm, excipient functionality.

Langridge (DFE): There is more desire to understand the role of excipient properties from regulators and industry. The question is whether the 'functionality' aspects of excipients should be addressed officially in pharmacopeias or on a case-by-case basis (e.g., in a new drug application or other regulatory submission). Key properties of an excipient for function in one formulation may be irrelevant in another formulation, even if the dosage form is the same. Therefore, official measures of functionality can only be framed in the very broadest terms. In the European Pharmacopoeia, some functionality-related characteristics (FRCs) are marked as nonmandatory, but this listing does not prevent excipient users from expecting these to be certified and controlled, even though the FRC may have no effect on the formulation.

To take a previous example, amorphous content is important in the compaction of spray-dried lactose, but it may not be of consequence if the tablet formulation contains other compactable components. To take a second example, the development of two monographs for lactose for inhalation may well result in a loss of functionality for some users because the extra limitations of the monographs restrict the lactose choices available to dry-powder inhaler formulators.

The downside of determining excipient functionality on a case-by-case basis is that for some products, this can lead to the development of tailored grades for different customer products, which may be a huge hurdle for some excipient manufacturers. Those excipient manufacturers willing to tailor grades will expect, and indeed require, substantially higher prices from the purchaser. In the end, it is the purchaser who must decide whether the cost of a tailored excipient is worth the security it brings for their product.

Reference

1. J. L'Hote-Gaston et al., Pharm. Technol. 33 (12), 36–41 (2009).