Accelerated Predictive Stability (APS): Fundamentals and Pharmaceutical Industry Practices

Accelerated Predictive Stability (APS): Fundamentals and Pharmaceutical Industry Practices provides coverage of both the fundamental principles and pharmaceutical industry applications of the APS approach. Fundamental chapters explain the scientific basis of the APS approach, while case study chapters from many innovative pharmaceutical companies provide a thorough overview of the current status of APS applications in the pharmaceutical industry. In addition, up-to-date experiences in utilizing APS data for regulatory submissions in many regions and countries highlight the potential of APS in support of registration stability testing for certain regulatory submissions.

This book provides high level strategies for the successful implementation of APS in a pharmaceutical company. It offers scientists and regulators a comprehensive resource on how the pharmaceutical industry can enhance their understanding of a product’s stability and predict drug expiry more accurately and quickly.

• Provides a comprehensive, one-stop-shop resource for accelerated predictive stability (APS)
• Presents the scientific basis of different APS models
• Includes the applications and utilities of APS that are demonstrated through numerous case studies
• Covers up-to-date regulatory experience

• Language: English
• Publisher: Academic Press
• Release date: May 14, 2018
• ISBN: 9780128027851

Book preview

MVTR)

Part I

General Chapters

Chapter 1

Accelerated Predictive Stability: An Introduction

Fenghe Qiu    Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, CT, United States

Abstract

Accelerated predictive stability (APS) approaches, such as accelerated stability assessment program (ASAP) and accelerated stability modeling (ASM), are emerging science and risk-based stability programs that have been successfully implemented in a number of pharmaceutical companies for a wide range of applications. This chapter provides a brief introduction of the APS approach and a high-level overview on the current status of APS applications in the pharmaceutical industry in the full drug development cycle with a focus on applications in the clinical development stages.

Keywords

Accelerated predictive stability (APS); Accelerated stability assessment program (ASAP); Accelerated stability modeling (ASM); Shelf life; Change assessment; Stability-related material attribute (SRMA); Shelf life-limiting attribute (SLLA); Science and risk-based approach

Chapter Outline

1.Historical Context of Accelerated Predictive Stability (APS)

2.General Principles of APS

3.Extent of APS Applications

4.Areas of APS Applications

4.1Applications During Clinical Development

4.2Applications for Registration

4.3Applications for Postapproval Changes

5.Conclusions

References

Further Reading

1 Historical Context of Accelerated Predictive Stability (APS)

The ICH stability approach was established between the mid-1990s and early 2000s marked by the publication of a series of stability guidelines (ICH, 1996a,b; 2002; 2003a,b). Under the ICH formal stability paradigm, the basic storage conditions are defined according to the climate zone of a pharmaceutical market. Four climate zones are distinguished in the world by their characteristic prevalent annual climatic conditions. As the ICH approach is designed mainly to confirm, not so much to predict, the retest period or shelf life, the entire proposed retest period/shelf life must be eventually covered by the long-term stability testing. Limited extrapolation is permitted at the time of submission under certain conditions, for example, if the long-term data and accelerated data both show little or no change over time, the proposed retest period or shelf life can be up to twice, but not more than 12 months beyond, the period covered by long-term data (ICH Q1E, 2003b). Although the ICH approach is intended for commercial registration of new drugs, in practice, regulators often expect ICH or ICH-like (same as ICH approach in storage conditions but may vary in minimum storage time and/or time points) stability testing for clinical trial applications partly because there is no specific harmonized ICH stability guidelines for clinical trial applications. It may have been an unintended consequence, but this approach also has trickled down to other development stability testing practices not intended for establishing a retest period or shelf life, such as excipient compatibility studies, formulation selections, packaging selections, etc., regardless whether scientifically it is the most appropriate approach or not. Such proliferation of the unintended use of the ICH harmonized approach is not always beneficial for some development stability testing, particularly for those in the clinical development stages, because it takes too long to reach the end point, thus is often found to be in the critical path in the fast-paced pharmaceutical development, and is not predictive.

On the other hand, despite the statement alternative approaches can be used when there are scientifically justifiable reasons for marketing application in ICH Q1A (R2), alternative stability approaches have not yet been fully explored by either the regulators or the pharmaceutical industry for marketing registration partly because there are few details on what constitutes a regulatory acceptable alternative stability approach in the guideline. In the past decade, use of alternative approaches has been reported by a number of companies in drug development programs and IMPD/IND submissions as the primary means for assigning drug substance initial retest period or/and clinical material use period based on the knowledge of drug degradation kinetics obtained under non-ICH stability conditions (Li, 2012; Freed et al., 2014).

Kinetic study of chemical degradation of solid pharmaceutical materials has long been an interest of research and many models and methodologies have been reported (Hirota et al., 1968; Vyazovkin and Wight, 1977; Rodante et al., 2002; Khawam and Flanagan, 2005); however, most of these studies were carried out in academic research institutes and the methodologies were rarely applied on a routine basis in pharmaceutical company's drug development programs due to the many limitations of the methodologies for pharmaceutical solids and the expertise involved in conducting such studies. In 2007, Waterman and coworkers (Waterman et al., 2007) described an approach termed accelerated stability assessment program (ASAP) for stability modeling of pharmaceutical materials. ASAP was developed on the basis of the so-called moisture modified Arrhenius Equation, which was initially demonstrated by Genton and Kesselring (1977) and the isoconversional model-free approach (Vyazovkin and Wight, 1977) and provided a practical protocol that can be carried out as a routine stability test in a regular pharmaceutical analysis laboratory. Encouraged by the successful applications of ASAP within Pfizer (Waterman et al., 2007; Colgan et al., 2012; Freed et al., 2014) and the commercial availability of ASAPprime, a software developed by FreeThink Technologies (Branford, CT) specifically for supporting ASAP studies, there has been significantly increased interest in recent years in ASAP adoption in the pharmaceutical industry. One of the benefits of ASAPprime is that it allows a regular pharmaceutical analytical scientist without formal training in statistics to perform the ASAP stability modeling, albeit ASAP modeling can also be performed using other commercially available software as demonstrated by Fu et al. (2015). Similar approaches were reportedly used by other companies internally too, for example, Clancy and coworkers recently published the kinetic model development of the so-called GSK approach (Clancy et al., 2017) termed as accelerated stability modeling (ASM), but those approaches have not yet been as well published in the public domain as ASAP.

2 General Principles of APS

In the context of this book, we define the alternative approaches such as ASAP and ASM that are intended to predict the long-term stability based on accelerated short-term stability testing as the accelerated predictive stability (APS) approach. In contrast with the ICH approach, APS emphasizes speed and agility (accelerated), and science and risk-based prediction (predictive) and is typically characterized by the following elements:

•Multiple storage conditions with appropriate elevated temperatures and a wide range of relative humidity conditions on open dish samples are typically used; actual conditions are tailored, based on the physical and chemical stability of the drug to ensure the mechanism of degradation at long-term storage condition would not change at the APS conditions.

•Storage time points are flexible depending on the storage conditions and the physical and chemical stability of the drug, typically from few days to a few weeks. Time points can be from one to several depending on the intended application.

•Study end points are not fixed in time, but rather are linked to the specification limit of the SLLA, for example, degradation product or assay loss. As such, the storage duration is shorter at a harsher condition and longer at a milder condition in order to generate similar level of degradation.

•Degradation is modeled by fitting the input data (e.g., the value of the SLLA, T, RH, and time if applicable) to the modified Arrhenius or an empirical equation based on either a model-free (e.g., ASAP) or a specific kinetic model (e.g., ASM)

•Statistical analysis is performed to assess the model's goodness of fit.

•Stability at long-term storage condition is predicted for rank order, or if appropriate, retest period or shelf life using the established APS model based on the data from APS storage conditions.

•The APS model should be verified using real-time stability (ICH) data for registration or postapproval applications.

In many ways, APS studies can be considered as bridging the gap between wholly empirical or data-driven approaches and mechanistic modeling approaches. For example, a purely data-driven approach may attempt to model the effects of temperature using temperature input as Celsius (°C), whereas APS approaches build on underlying principles and model the effects of temperature using reciprocal absolute temperature (K− 1). On the other hand, APS approaches are not mechanistic models; a full mechanistic understanding of the stability of a pharmaceutical product is very rarely attempted since pharmaceutical stability is an emergent property, that is, many factors contribute to the overall stability behavior. While mechanistic models have an important role in understanding pharmaceutical stability, they are unlikely to be sufficiently accurate and reliable for predicting pharmaceutical stability. Conversely, APS studies have shown that it is possible to obtain reliable predictions without a full mechanistic understanding of pharmaceutical stability.

Over recent years, it has become evident that APS can play a key role in predicting, understanding, and ultimately controlling pharmaceutical stability. This chapter provides readers with a brief overview of the extent and areas of applications of APS in the pharmaceutical industry with a focus on the areas of applications. Details on the fundamental aspects of these approaches and case studies of their industry applications are covered by other general and case study chapters in this book and thus will not be discussed in this chapter.

3 Extent of APS Applications

Although there are increasing interests in the APS approaches in recent years, it is safe to say most of the pharmaceutical companies are using the ICH or ICH-like approach for development stability testing besides the formal registration stability. So how widely are APS approaches currently utilized by the pharmaceutical industry? According to the website of FreeThink Technologies, the provider of ASAPprime, 18 of the top 20 pharmaceutical companies licensed ASAPprime (FreeThink, 2017), suggesting at least most major pharmaceutical companies have already implemented ASAP to some extent. This is evidently supported by a recent industry survey conducted by the risk-based predictive stability (RBPS, which is equivalent to APS from all practical purposes of this book) working group, an industry working group under the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ). When asked Has your company used RBPS to enable development or post approval changes among the 33 member companies, out of the 18 companies responded, 16 companies reported use of RBPS approaches in their companies for various purposes, while other companies responded RBPS approach was not used due to lack of knowledge or because RBPS approach has not been considered (Williams et al., 2017). With regard to the types of RBPS approaches used, most of the companies adopted the ASAP approach and used the commercial ASAPprime software for data modeling, while two other companies reported that comparable in-house APS approaches were used. Furthermore, the survey also revealed encouraging signs of increased regulatory familiarity and acceptance of APS; at least 10 companies among the IQ members surveyed have already included ASAP results in regulatory submissions and such submissions have been accepted by over 20 countries around the world including all major pharmaceutical markets (Williams et al., 2017).

4 Areas of APS Applications

APS can be designed to suit a wide range of materials and dosage forms, including powder materials (API, intermediates, reference materials, powders for oral solutions), liquids (Badejo et al., 2016) and suspension, lyophilized product, semisolid and solid oral-dosage forms (tablets, chewable tablets, capsules, and soft gelatin capsules), and transdermal patches (Ainampudi et al., 2016). Practical considerations in experimental design are discussed in Chapter 4 of this book.

Based on literature review and the IQ survey results, the overwhelming majority of applications of APS so far are on small-molecule drug (including new molecular entities, generic, OTC, and consumer health product) chemical stability modeling and prediction (Williams et al., 2017). Chemical degradation of small-molecule drug involves relatively straightforward bond cleavage or formation of the active pharmaceutical ingredient, which in principle follows the Arrhenius kinetics, as discussed in Chapter 3 of the book. In contrast to the ICH stability approach (fixed temperature and relative humidity conditions and standardized time points typically from months to years), to take advantage of both temperature and moisture effects on the drug degradation kinetics, APS utilizes more severe yet still appropriate storage conditions, thus shorter storage times, based on the physicochemical properties of the product, to ensure valid extrapolation can be made from APS storage conditions to the long-term storage condition.

APS has also been used in prediction of product physical changes such as color, hardness, disintegration, and dissolutions and form change by a smaller group of companies (Williams et al., 2017). As an example, Lin et al. recently reported a case study on prediction of the changes in drug dissolution from an immediate-release tablet containing two active pharmaceutical ingredients using ASAP (Li et al., 2016).

In general, as shown in Fig. 1, APS can be applied in lieu of ICH-like stability testing or stress testing when stability assessment is required in the small-molecule drug development life cycle for a variety of purposes, including but not limited to the following areas.

Fig. 1 Areas of APS applications for small-molecule drug development.

4.1 Applications During Clinical Development

The primary purpose of stability testing during clinical development is to understand the intrinsic chemical and physical stability of the API, identify stability-related material attributes (SRMAs, e.g., polymorph, PSD, water content) and shelf life-limiting attributes (SLLAs, e.g., degradation product, assay loss, or dissolution) and support the development of a formulation with feasible shelf life for clinical trials and regulatory submissions. As the pace of development at this stage is fast and changes are very common, APS application in clinical development is most beneficial due to its inherent advantage over the conventional ICH stability approach in speed and accuracy. This is reflected in the findings from the IQ RBPS survey, most companies utilizing APS responded applications in this space (Williams et al., 2017).

4.1.1 Fast comparative stability assessment of API salt form and polymorph

One of the challenges in early development is to select a stable API form for further development. The typical process is to screen a large number of API salt forms and/or polymorphs manually or using high throughput technology, followed by a series of chemical and physical testing including stability testing on a number of selected promising forms. APS can be used to replace the typical high temperature stress testing and provides a more complete and accurate understanding of the API stability with comparable workload and turnaround time. For example, the observed levels of degradation of two solid forms with different activation energies could be reversed at high temperature vs. those at the long-term storage temperature, thus potentially resulting in mistakenly selecting the relatively unstable form if a one-temperature stress testing approach is used, because the more stable form at long-term storage temperature could exhibit higher level of degradation at high temperature due to a higher activation energy. The use of APS will prevent such mistake from happening.

4.1.2 Fast stability rank order of prototype formulations

Formulation screening is one of the top three APS application areas during clinical development according to the IQ survey responders (Williams et al., 2017). During the development of clinical trial formulations, a number of prototypes are typically selected initially based on excipient compatibility studies and are subject to physical, chemical, and may be also in vivo testing for comparison of their stability and bioavailability. In this process, stability testing is often in the critical path for decision making as the 3-month ICH accelerated stability at 40°C/75% RH is the gold standard for this purpose. Alternatively, some companies might use high-temperature stress testing (e.g., 60°C for a month) to save time. Nonetheless, both approaches have the same limitation as discussed for using one temperature testing and overlook the moisture effect to the formulation stability. Hence, APS is a much better predictive tool for formulation prototype rank ordering.

4.1.3 Support API and formulation process development

In early development, the API synthetic process is still evolving. API batches for nonclinical or clinical supplies may be manufactured using different routes of synthesis or isolation methods; consequently, the resulting API may have varying material attributes such as polymorph, PSD, water content, or impurity profile. Those material attributes may or may not be potentially stability related, that is, stability-related material attributes (SRMAs). APS assessment in lieu of ICH-like stability testing on API batches manufactured before and after the changes provides quicker but enhanced understanding on the impact of process changes to the API stability, and the knowledge of the SRMAs. Non-SRMAs do not have to be considered for selecting a representative API batch for formal stability testing for regulatory purposes.

Similarly, owing to its flexibility and agility, APS is powerful in assessing stability impact of changes or process optimization during formulation development, for example, assessing the impact of milling, mixing, granulating processes and techniques.

4.1.4 Packaging selection

Packaging development is an integral part of the pharmaceutical development as the shelf life of the drug product is determined by the stability of the drug within its packaging. Using solid oral formulation as an example, common packaging includes bottles and blisters made from a variety of materials with variable moisture permeability. The considerations for selecting an appropriate type and grade of packaging should be made to strike a balance between the need to protect the drug to achieve the desired shelf life and cost. When necessary, desiccant can be included in the packaging for additional protection to a moisture-sensitive drug. Conventional practice for packaging screening is to set up comparative stability testing with the various packaged products at ICH accelerated stability conditions for a minimum of 3 months, which not only take a long time to obtain the data, but also require significant amount of API and resources. In contrast, APS does not rely on parallel stability testing of the drug in different packaging configurations in order to evaluate the packaging performance; rather, APS is only conducted on one set of product samples to establish the drug degradation kinetics at open dish conditions. This provides quantitative information on the effects of humidity. Therefore, the stability of the drug in the proposed packaging is modeled in silico by taking into consideration the moisture vapor transmission rate (MVTR) of each proposed packaging material and the product moisture sorption isotherm, thus avoiding the tedious and material-consuming parallel ICH stability testing. For early clinical development when the shelf life requirement is relatively short and the fit-for-purpose approach is more tolerated, the MVTR values of proposed packaging materials found in literature or commercial software database (e.g., ASAPprime) and the estimated product moisture sorption isotherm from combining the individual moisture sorption isotherm of the formulation components may be utilized. With the advancement of the clinical development, additional effort and investment are justified to experimentally determine the product-specific moisture sorption isotherm and material-specific MVTR to ensure the ultimate accuracy of the APS model prediction for the shelf life of packaged product (Waterman and MacDonald, 2010; Forcinio, 2015).

4.1.5 Excursion evaluation

When temperature excursion occurs during transportation or storage, its impact to the shelf life of the clinical supply can be a huge concern. As stated in ICH Q1A (R2) guideline, Excursions that exceed the defined tolerances in temperature and relative humidity for more than 24 hours should be described in the study report and their effect assessed. The conventional approach for thermal excursion assessment is using the mean kinetic temperature (MKT). This approach is based on two assumptions, (1) the degradation follows the simple linear Arrhenius model, (2) the activation energy of the drug degradation is approximated to be 19.87 kcal/mol (83.144 kJ/mol) (Grimm, 1993). In reality, both assumptions are far from accurate. The activation energies of drug degradation vary significantly depending on the structure of the drug molecule (Baertschi et al., 2011). The APS approach models both temperature and moisture excursions and uses the actual activation energy of the product; therefore, a much more accurate assessment on the impact of excursions can be achieved.

4.1.6 Predict drug substance retest period and clinical supply use period

Utilizing the APS prediction in lieu of ICH stability testing as the primary means to assign the initial API retest period and initial use period/shelf life for the clinical supply (often with a commitment of placing the clinical supply on a concurrent/confirmatory ICH stability study) for CTA submissions is one of the top three APS applications among all the IQ survey responders (Williams et al., 2017). This approach is attractive first of all because it can expedite the IMPD/IND submission timeline by at least 2–3 months without compromising the quality of the submission and patient safety, and secondly because regulatory agencies are more flexible in accepting non-ICH alternative approaches for clinical trial applications. Recently, Freed et al. (2014) published some of Pfizer's successful examples that included the use of APS for initial retest period/shelf life assignment for clinical trial applications. The IQ survey reveals that at least 10 companies among the survey responders have submitted retest period/shelf life primarily based on or supported by APS prediction (Williams et al., 2017). In mitigating the potential regulatory risk, most of these companies do exercise caution when assigning such an initial retest period/shelf life by limiting it to 1–2 year maximum; even the data suggest a very stable product. So far, over 20 countries have reportedly accepted such initial clinical shelf life based primarily on APS data without ICH stability or in some cases ICH data are submitted up on agency's request for confirmation during the review cycle. It is worthwhile to mention that a few countries are known to decline such submissions based on some early APS adopters’ experiences if without ICH data as supportive data (Williams et al., 2017). APS can also be utilized in combination with ICH stability data to justify a shelf life assignment longer than the period confirmed by the ICH stability data at the time of submission, for example, APS plus 3-month accelerated stability data to support a 24-month use period. A recent successful APS submission, where a 3-year shelf life based on ASAP prediction plus 3-month real-time data was accepted by multiple countries for a phase 1 submission (http://freethinktech.com/industry-focus-2/pharma-success-stories-page/), highlights the potential of this approach to expedite clinical timelines.

For all late-stage clinical development applications, the APS model should be validated/verified using the long-term ICH stability data, which should have obtained already during earlier development stages.

When APS is intended to be used as the primary data to justify a retest period/use period, it is advisable to plan a fallback strategy, for example, to make 3-month ICH stability data available at the time to respond to the deficiency letter, to avoid potential clinical trial delay just in case the agencies decline the APS justification. Even in this scenario, the initial CTA submission can still be expedited by a couple of months by submitting the APS-backed shelf life in the initial submission.

4.1.7 Other applications

Other potential areas of APS application include supporting shelf life of bulk drug and modified comparator, fast stability assessment for in-depth understanding or stability problem solving, for example, stability OOS investigation, kinetic investigation of genotoxic degradation product, mass balance investigation. Recently, suitability of APS for in-use stability has also been demonstrated (Remmelgas et al., 2013; Waterman et al., 2016).

4.2 Applications for Registration

For registration, formal stability data on three batches of API or product representative of the commercial batches according to the ICH stability protocol are typically expected by the health authorities. Although ICH Q1A (R2) (ICH, 2003a) states alternative storage conditions can be used if justified, in reality, few companies would be willing to risk a delay of the marketing approval by using nonconventional stability approaches at this stage, even if it is scientifically justified. As a result, APS is not recommended to be used as a primary tool for supporting the API retest period or the drug product shelf life for registration. Nonetheless, APS data can still be submitted in NDA/MAA as supportive data. As revealed by the IQ RBPS survey, five companies successfully submitted APS data in marketing applications as supporting data for packaging selection, retest period/shelf life, excipient compatibility, specification and demonstrating batch equivalence, etc. The APS data were typically presented alongside the formal ICH stability data (Williams et al.,