OR WAIT 15 SECS
Jin Wang is manager in manufacturing sciences, product supply biotech at Bayer HealthCare
Boehringer Ingelheim, 900 Ridgebury Road, Ridgefield, CT 06877, USA
Pfizer Inc., Eastern Point Road, Groton, CT 06340, USA
Janssen Pharmaceutica R&D, a Division of Janssen Pharmaceutica NV, Turnhoutseweg 30, 2340 Beerse, Belgium
AstraZeneca, Silk Road Business Park, Macclesfield, Cheshire, SK10 2NA, UK
Abbvie, 1 N Waukegan Rd, North Chicago, IL 60064, USA
Eli Lilly & Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
Yan Wu is principal scientist at Merck & Co., Inc.
Hanlin Li is associate director at Vertex Pharmaceuticals.
A published regulatory template sharing best practices for filing RBPS data would benefit the industry and regulatory reviewers by enabling a consistent presentation of predictive data and conclusions.
The science of stability has significantly evolved since the advent of International Council for Harmonization (ICH) Q1A(R2) (1). Improved modeling tools coupled with appropriately tailored protocols have enabled similar or better stability predictions within accelerated timeframes, when compared to a more traditional ICH approach (2–4). These tools provide increased understanding of attributes that influence drug substance and product stability instead of following the traditional ICH approach, which simply demonstrates stability in an empirical manner. These contemporary tools and approaches are well aligned with the science and risk-based approaches detailed in ICH Q8–Q11 (5–8) and have been termed risk-based predictive stability (RBPS).
Companies are utilizing these RBPS tools to better enable development of medicines (9). In 2015, the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ) launched a working group to focus on the use of RBPS tools to optimize pharmaceutical development. The working group has approximately 50 members from 18 companies across the pharmaceutical industry. The working group conducted a survey of the industry to understand sponsor companies’ experiences using RBPS tools (10). The survey was highly informative and indicated that RBPS tools were being utilized in a variety of applications across the development continuum. A key learning was that of all of the companies utilizing RBPS tools, approximately 55% of them were leveraging the data in a regulatory capacity.
Over the course of working group discussions, it was determined that utilization of RBPS data was used in excess of 100 submissions by the working group companies. A selection of case studies that discuss the regulatory feedback on these submissions will be published in the near future. During the course of discussions within the RBPS working group, it was concluded that a published regulatory template sharing best practices for filing RBPS data would benefit the industry and regulatory reviewers by enabling a consistent presentation of predictive data and conclusions. The majority (85%) of survey respondents confirmed that a template would benefit the industry. This template could help companies standardize on key elements that should be included when filing RBPS data in Module 3 stability sections (i.e., S.7 and P.8) of regulatory submissions. The recommendations within this manuscript for presenting RBPS data in a regulatory submission are based on industry early adopter experience and are intended to be used in setting shelf-life for drug substance or drug product that is used to support clinical development. The term ‘shelf-life’ is used throughout this manuscript, but the terminology will vary for drug substance (re-test) and from company to company (e.g., clinical use period).
This manuscript consists of two sections. Section I is a high-level outline of the key elements for a RBPS filing section. Section II provides a specific example of how a RBPS filing may look.
The following elements (Table I) should be considered when filing a RBPS data package to support an initial shelf-life for drug substance or drug product.
The sponsor should also describe the assumptions as context for its modeling approach and assess the impact of these assumptions on the study results and interpretation.
Each section is described further as follows.
Introduction. A discussion of the stability risk assessment, along with a justification of the chosen potential shelf-life limiting attributes (SLLA[s]), should be included in this section. All potential SLLAs should be considered, including both physical and chemical attributes.
Utilization of RBPS leverages advanced modeling approaches of data that have been generated under a variety of stress conditions. Typically, there are a few key quality attributes that are shelf-life limiting, such as a degradation product.
Based on the stability risk assessment, the rationale for the choice of which attributes were modeled as shelf-life limiting attributes should be discussed and justified.
Description of the model used. Provide a description of the model used, along with appropriate literature references, as applicable. A description of the software that is used should also be included. Additionally, any assumptions regarding packaging (e.g., material type, moisture permeability, or moisture vapor transmission rate) should be detailed if they are used to support modeling.
Discussion of experimental design. Provide the experimental conditions (e.g., temperature/relative humidity and time points) that were used for the study in tabular format. A discussion may be included on how the storage conditions were selected, especially if they were driven by particular physiochemical properties of the drug substance and/or drug product formulation components. In some cases, the samples assessed may be a different formulation than the clinical formulation, where the excipient-to-active ratio may be worst case. In this case, include a discussion of why the samples used for the study were ‘worst case’ to maximize possible degradation.
Also discuss why the studied container closure was selected (e.g., open containers allowing for better correlation with the impact of humidity).
Provide a summary of what shelflife limiting attributes were evaluated after storage (e.g., degradation product X, appearance). Address any differences in analytical procedures used from those provided in the Analytical Procedures sections of the regulatory filing, if applicable.
Discussion of results. Provide a detailed discussion and interpretation of the results. Specifically discuss the shelf-life limiting attribute(s) (e.g., degradation product x) and how this was modeled to set a shelf life for the drug. A discussion/explanation of any other changes (e.g., appearance) would be appropriate as well.
Long-term stability program. The planned long-term stability commitment should be discussed. The study design may be supported by RBPS results. Based on the understanding of the modeling, this could encompass a variety of approaches. These approaches could include ICH-like testing, reduced time points, reduced conditions, and/or contingency storage.
Conclusion. Provide a conclusion to indicate the shelf-life that is supported by the modeling data. Where applicable, outline how extensions to the initial shelf-life will be assigned.
The purpose of this section is to provide an example of a RBPS filing for a first-in-human (FIH) study. The example given below is for a small-molecule, solid oral dosage form and could be included as part of P.8.1 within the clinical application. It may be adapted for other small- molecule formulation types and drug substances as relevant.
This example is based on the Accelerated Stability Assessment Program (ASAP) model. Other models or software packages may be used as appropriate.
Introduction. Based on a stability risk assessment, it was concluded that drug product Degradant A is expected to be the SLLA. The drug product is designed as an immediate-release capsule. Dissolution was not modeled, because it is not expected to be a SLLA. This is based on the fact that drug product, when exposed to accelerated conditions, did not show any meaningful changes in dissolution profiles. Drug product assay is also not expected to be a SLLA because the drug product degradant limits (i.e., not more than 0.5%) are set such that they would fail before a significant change in assay would be observed.
The ASAP approach was used to develop an in-depth understanding of the chemical stability performance of the drug product as a function of temperature and relative humidity. This understanding was used to determine appropriate packaging (as described in Section P.7 of the regulatory filing) and storage conditions, to predict an initial shelf life for the clinical drug product, and to determine the SLLAs to be included in the confirmatory long-term stability protocol on a representative batch of drug product.
The RBPS will be supplemented by a confirmatory study that includes long-term storage conditions and traditional accelerated storage conditions (40°C/75%RH); this confirmatory stability study has been initiated for a representative batch. [Include specific information such as lot number, manufacturing scale, etc. Also include a justification as to why it is considered to be representative.] The accelerated and long-term data from this batch, when available, will be used to confirm the predictions of the model and to support further shelf-life extensions.
Description of the model used. The ASAP approach was used. This is a statistically designed RBPS program based on the modified Arrhenius equation. The design of the predictive study is based on literature that demonstrates the modeling of observed degradation of solid oral-dosage forms (2,3). Short studies were conducted on open-dish samples of the representative batch of drug product at elevated temperatures over a range of humidity conditions with the goal of reaching the specification limit for the identified SLLA at each condition as detailed as follows. Humidity determines water activity in the drug product and, therefore, can have a significant effect on reaction rates in solid drug products, even for reactions which themselves do not involve water.
The humidity-corrected Arrhenius equation (Equation 1) reflects both the influence of the temperature and the influence of moisture on the kinetics of the degradation product formation.
The resulting open-dish data were fit to a humidity-corrected Arrhenius equation using ASAPprime Version 5.0 [alternative commercial or inhouse software may be used]:
ln k = ln A –Ea/RT + B(RH)
Where k is the degradation rate, A is the Arrhenius collision frequency, Ea is the activation energy for the chemical reaction, R is the gas constant, T is the temperature in Kelvin, B is a humidity sensitivity constant, and RH is relative humidity.
The moisture sorption isotherm for the drug product was determined, and the moisture permeability of the primary package was determined based on literature data. This information was used to estimate the dynamic water activity in the packaged drug product as a function of time at the proposed storage condition.
The model derived from fitting the data to Equation 1 was then used to calculate the expected value and the upper and lower 95% confidence limits for the SLLA as a function of time in the selected package at the long-term storage condition.
Discussion of experimental conditions. A representative drug product lot for compound X was stored in an open dish configuration at the temperatures and humidity conditions outlined in Table II. The exposed samples were then tested for degradants by the high-performance liquid chromatography (HPLC) procedure that is described in P.5.2 of the regulatory filing. Other potential SLLAs were also tested over the conditions studied. These included physical appearance and dissolution. [If the analytical methodology used differs from that provided in P.5.2, provide further explanation.]
Discussion of results. The data collected indicate that Degradant A will be the SLLA at long-term storage conditions of 25 °C/60%RH. Levels of Degradant A ranged from 0.00% to 1.00% (Table II).
These data were fit to the modified Arrhenius equation (Equation 1). All stability attributes are expected to remain within specification limits for at least 12 months. Modeling predictions for the shelf-life limiting attribute, Degradant A, are included in Table III.
The results are plotted in Figure 1 and shown in Table IV.
The modeled data presented in Figure 1 are based on an assumption that the degradation kinetics to the specification limit is occurring in a linear fashion. For this degradation pathway-and based on the ASAP study data-this assumption is consistent with the chemistry and stability knowledge of the drug substance, stability knowledge of the drug product at this stage of development, purposeful degradation, and literature. The data may also be based on drug substance and drug product knowledge gained to date.
Other non-SLLA drug product attributes were tested following exposure to the conditions studied. This included physical appearance and dissolution. None of the data showed a meaningful change in those attributes. [If no degradation is observed during an ASAP study with conditions such as 70 °C at both high and low humidity for at least three weeks, an initial shelf life of a minimum of 12 months is deemed to be appropriate (11).]
Long-term (confirmatory) stability program. The initial shelf life is based on the ASAP study. Subsequent shelf-life extensions will be supported using a long-term stability study. The identified SLLA will be studied as well as assay, physical appearance, impurities, and dissolution. The protocol for the long-term stability study is provided in section P.8.1 of the regulatory filing.
As additional long-term stability data become available, they will be assessed against the same acceptance criteria and reviewed against the modeling predictions. The shelf life may be extended as these additional long-term data become available. The shelf life will not be extended beyond the last time point as outlined in the long-term stability protocol.
Following each long-term stability time point, the results are reviewed to confirm that the acceptance criteria are met and to monitor for trends and unexpected test results. Trending is conducted to confirm the extrapolation of the shelf-life remains appropriate.
An amendment will be submitted if there is any change in the storage condition or packaging configuration of the investigational medicinal product during the clinical trial.
On the basis of additional long-term stability data for the representative batch, the shelf life will be extended without submitting a substantial amendment, unless stated otherwise in applicable regulations. The specifications and recommended storage conditions will remain the same.
Conclusion. An initial shelf life of 12 months when stored at or below 30 °C has been established.
The industry survey on use of RBPS tools indicated that more than half of the companies surveyed use the data from RBPS studies in their regulatory submissions. As outlined within this article, an effective application of RBPS within regulatory submissions is to support an initial shelf-life for an early development formulation. Companies have been using this science-based approach for several years.
The stability understanding gained from a well-designed RBPS study generally exceeds knowledge gained from a three-month time point data at the long-term storage and traditional accelerated conditions (40 °C/75% RH). Therefore, shelf-life predictions supported by RBPS studies are considered conservative, because typically, the predicted shelf-life limiting attribute will not breach the acceptance limit until well beyond the assigned shelflife.
Per current clinical guidelines published by the European Medicines Agency (EMA) (12), three months of long-term data may be used to set a 12- month clinical shelf-life. Given the increased knowledge obtained on potential degradation from predictive tools, a similar initial clinical shelf-life may be justified if the RBPS data support it. Additionally, companies are maintaining a conservative approach past the 12-month initial clinical shelf-life by basing further extensions on long-term data.
Use of this approach can reduce clinical start timelines by months, resulting in potentially life-saving therapies entering the clinic faster. The template outline provided within should provide others wishing to implement a similar strategy with a good starting point. A publication of industry case studies with regulatory feedback is currently under preparation. Industry continues to seek collaborative inter-action and is open to consultation with regulatory agencies to jointly integrate RBPS tools to support shelf-life assignments.
1. ICH, Q1A(R2) Stability Testing of New Drug Substances and Products, Step 4
version (ICH, 2003).
2. K.C. Waterman and R.C Adami, International Journal of Pharmaceutics 293 (1-2), 101-125 (2005).
3. K. Waterman, et al., Pharmaceutical Research 24 (4), 780-790 (2007).
4. A. Oliva, J.B. Farina, and M. Llabres, Talanta, 94, 158-166 (2012).
5. ICH, Q8(R2) Pharmaceutical Development, Step 4 version (ICH, 2009).
6. ICH, Q9 Quality Risk Management, Step 4 version (ICH, 2005).
7. ICH, Q10 Pharmaceutical Quality System, Step 4 version (ICH, 2008).
8. ICH, Q11 Development and Manufacture of Drug Substances (Chemical Entities
and Biotechnological/Biological Entities), Step 4 version (ICH, 2012).
9. A.L. Freed, E. Clement, and R. Timpano, Regulatory Rapporteur, 11 (7/8),
10. H. Williams, et al., Pharmaceutical Technology, 41(3), 52-7 (2017).
11. Q. Chan Li, et al., J Pharm Innov, 7, 214-224 (2012).
12. EMA, Guideline on the Requirements for the Chemical and Pharmaceutical
Quality Documentation Concerning Investigational Medicinal Products in Clinical Trials (EMA/CHMP/ QWP/545525/2017).
Authors’ Note: This paper was developed with the support of the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ, www.iqconsortium.org). IQ is a not-for-profit organization of pharmaceutical and biotechnology companies with a mission of advancing science and technology to augment the capability of member companies to develop transformational solutions that benefit patients, regulators, and the broader research and development community.
Vol. 42, No. 8
When referring to this article, please cite it as D. Stephens et al., "Risk-Based Predictive Stability for Pharmaceutical Development–A Proposed Regulatory Template," Pharmaceutical Technology 42 (8) 2018.