See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257777621

The Application of Science- and Risk-Based Concepts to Drug Substance

Stability Strategies

Article  in  Journal of Pharmaceutical Innovation · December 2012

DOI: 10.1007/s12247-012-9135-9

CITATIONS READS

10 933

6 authors, including:

Stephen Colgan Jon Beaman

Pfizer Pfizer
38 PUBLICATIONS   437 CITATIONS    9 PUBLICATIONS   46 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jon Beaman on 01 August 2016.

The user has requested enhancement of the downloaded file.

J Pharm Innov
DOI 10.1007/s12247-012-9135-9
CASE REPORT
The Application of Science- and Risk-Based Concepts
to Drug Substance Stability Strategies
Stephen T. Colgan & Timothy J. Watson & Robert D. Whipple &
Roger Nosal & Jon V. Beaman & David M. De Antonis
# Springer Science+Business Media New York 2012
Abstract The International Conference on Harmonization
(ICH) has provided practical guidance on the amount and type
of drug substance stability data needed to support marketing
applications (International Conference on Harmonization
2001, 2002, 2003a, b). Additional guidance has been issued
by the World Health Organization (WHO 2009). Recent sci-
entific advances and practices have resulted in improved
scientific understanding of the chemical and physical attrib-
utes that contribute directly or indirectly to drug substance
stability. Combining this improved understanding with the
science- and risk-based approaches detailed in ICH Q8, Q9,
and Q10 allows for alternative and more scientifically
driven approaches to meet the scientific and regulatory objec-
tives for drug substance stability (International Conference on
Harmonization 2005, 2008, 2009). In this paper, proposals are
presented to more fully leverage enhanced product knowledge
to design improved stability strategies. The chemical and
physical attributes that potentially impact drug substance sta-
bility are discussed, and strategies that leverage accelerated
stability studies are presented.
Keywords Stability . Drug substance . Stability-related
quality attributes . Control strategy
Some of these concepts were presented at the AAPS Current Trends in
Stability Workshop (Washington, D.C., September 2009): Colgan S.
“The Application of Quality by Design’s Science and Risk Based
Concepts to API Stability Strategies”
S. T. Colgan (*) : T. J. Watson : R. D. Whipple : R. Nosal :
D. M. De Antonis
Pfizer Global Research and Development,
Groton, CT, USA
e-mail: stephen.t.colgan@pfizer.com
J. V. Beaman
Pfizer Global Research and Development,
Sandwich, UK
Introduction
The International Conference on Harmonization (ICH) Q8
states that pharmaceutical development should include, at a
minimum, a definition of the quality target product profile as
it relates to quality, safety, and efficacy, considering stability
and several other factors. Q8 also states that a greater prod-
uct understanding can create a basis for more flexible regu-
latory approaches, the degree of which is predicated on the
level of relevant scientific knowledge. Although Q8 covers
drug products, its principles can be applied to drug sub-
stance as well. ICH Q11 recently issued at stage 4 states that
drug substance critical quality attributes typically include
those properties or characteristics that affect identity, purity,
biological activity, and stability. In addition, changes to the
manufacturing process should be evaluated for impact on
the quality of the drug substance. The evaluation should be
based on scientific understanding of the manufacturing pro-
cess and should determine appropriate process point(s) and
testing to analyze the impact of the proposed change.
In regulatory filings, formal stability studies are designed
to evaluate quality over time under the influence of temper-
ature, humidity, and light and to facilitate the establishment
of appropriate controls, retest periods, and underwrites pack-
aging and storage recommendations. Drug substance stabil-
ity is related to inherent properties of the specific compound,
as well as its stability-related quality attributes. A stability-
related quality attribute (typically a subset of quality attrib-
utes) directly or indirectly impacts stability but may or may
not impact other considerations (such as manufacture of the
drug product). Identification, quantification, and control of
the stability-related quality attributes ensure drug substance
quality over time.
There are potential benefits for the pharmaceutical industry,
global regulators, and patients that could be realized by adapt-
ing scientific stability strategies that leverage understanding of

these chemical and physical attributes which can impact drug
substance stability. Consistent with the Q8 and Q11 principles,
a fundamental understanding of stability, coupled with risk-
based- and prospective control strategies, could reduce the
overall number and/or type of stability studies (regulatory
flexibility), while increasing confidence in the proposed drug
substance storage retest periods.
Understanding which attributes or variables that do not
directly or indirectly affect drug substance stability is as
important as well because one can justify that these can be
controlled through internal (non-regulatory) quality systems.
This approach will allow industry and regulators to focus on
the drug substance variables and attributes that have the high-
est probability of affecting stability and potentially safety and/
or efficacy as well. Examples of variables that generally do not
directly or indirectly affect drug substance stability include the
synthetic route, processes, site of manufacture, and batch size.
Here, the stability-related quality attributes are unaffected as
these variables change. In summary, the stability-related qual-
ity attributes need to be identified, understood, and controlled.
The other factors mentioned above (synthetic route, etc.) can
then be assessed to determine whether they impact the
stability-related quality attributes. This concept can be applied
during development, for marketing applications, and for post-
approval changes.
Control Strategies for Stability
A control strategy as defined in ICH Q10 is “A planned set
of controls, derived from current product and process un-
derstanding that assures process performance and product
quality. The controls can include parameters and attributes
related to drug substance and drug product materials and
components, facility and equipment operating conditions,
in-process controls, finished product specifications, and the
associated methods and frequency of monitoring and con-
trol.” Although not explicitly stated in Q10, the concept of a
control strategy can also be applied to drug substance and its
related stability.
The foundation of a comprehensive control strategy for
stability is the proper selection of a stable drug substance.
Ideally (from a stability viewpoint), the polymorph of the
drug substance should be the lowest energy (most stable)
form. The drug substance form chosen (e.g., counter ion)
may need to balance stability and other considerations such
as bioavailability and ease of manufacture of the drug prod-
uct. Project teams should then work to understand and
control the stability-related quality attributes that affect the
inherent stability of the drug substance. Although drug
substance instability can be managed downstream (in the
drug product or though protective packaging), this may not
be as effective as controlling instability by selecting the
J Pharm Innov
most suitable drug substance form and understanding and
controlling the stability-related quality attributes of the drug
substance.
Identification of Stability-Related Quality Attributes
Regulatory stability guidances are based on general princi-
ples of stability and chemical degradation, and were written
primarily to harmonize regulatory standards for marketing
applications. These guidances describe what is expected in
terms of the number of batches, time on stability, stability
conditions, and indicate that “Stability studies should in-
clude testing of those attributes of the drug substance that
are susceptible to change during storage and are likely to
influence quality, safety, and/or efficacy.” Interestingly, cur-
rent guidance does not necessarily require an understanding
of the attributes most critical to the stability. Identification of
those attributes that directly or indirectly contribute to drug
substance stability is critical to establishing the fundamental
understanding and design of a robust control strategy for
stability. Stability-related quality attributes can be catego-
rized into those with the potential to directly or indirectly
impact drug substance stability and those that are unlikely to
impact stability.
Attributes with the potential to impact drug substance
stability include ratio of polymorphs, degree of crystal-
linity, surface area, moisture content, and identity of
residual solvent and/or level. Certain impurities may
cause degradation (i.e., acids/bases, metals) and in rare
cases may improve stability (e.g., antioxidants). In many
cases however, these attributes will not alter drug substance
stability.
Attributes unlikely to impact drug substance stability
include appearance, physical attributes (i.e., flowability,
bulk density), or minor changes in the impurity profile. Note
that while an attribute like appearance is unlikely to affect
stability (batches of different color will demonstrate similar
behavior on stability), appearance is an attribute that may
change on stability. Instability usually is initially detected by
changes in degradant levels but also can be manifested by
changes in other attributes such as assay or solid form.
Process changes that result in new or increased levels of
process-related impurities usually have no effect on the
overall stability of the drug substance because these impu-
rities are chemically inert. As such, changes in the impurity
profiles should be carefully assessed relative to the specific
impurity and its reactivity with the drug substance. Al-
though each drug candidate should be evaluated on a case-
by-case basis, additional long-term stability studies to sup-
port changes to the attributes detailed above are likely to be
of low value and unlikely to provide any benefit related to
patient safety.
J Pharm Innov
As detailed earlier, changes to the synthetic route, site of
manufacture, or batch size generally do not affect drug
substance stability unless the changes result in a change in
a stability-related quality attribute. As the synthetic process
is developed and drug substance attributes change from
batch-to-batch, project teams should consider short acceler-
ated studies (vide infra) to evaluate if the change has altered
drug substance stability as a result of a change in a stability-
related quality attribute. By conducting side-by-side scien-
tifically based stability challenges as the drug substance
process evolves, a constantly expanding stability knowledge
base will be established. If the stability of the new batch is
different, a traditional stability study should be considered.
The practice of evaluating process changes with short ac-
celerated studies also can minimize the number of “low
value” traditional long-term stability programs either
for batches with equivalent stability or unacceptable
stability performance. A low-value stability program is charac-
terized as one that does not provide an improved understanding
of stability.
Current Regulatory Guidance and Expectations
The ICH guidances for stability (Q1A through E), which are
intended for commercial product registration, outline gener-
al expectations for batch selection, storage conditions, and
testing frequency for accelerated and long-term drug sub-
stance stability programs. Stability programs should include
at least three batches of drug substance manufactured to a
minimum of pilot scale by the same synthetic route, method
of manufacture, and procedure that simulates the final pro-
cess used for production batches. The overall quality of the
batches of drug substance placed on formal stability studies
should be representative of the quality of the material to be
made on a production scale. Unfortunately, interpretation of
ICH guidances varies within industry, from country-to-
country and between regulatory agencies. To avoid risk,
both industry and regulators have often adopted a conserva-
tive interpretation of these guidelines which does not focus
on the fundamentals of stability relative to the manufactur-
ing design space. Table 1 below outlines some of the differ-
ent regional interpretations of registration stability
expectations for new chemical entities (NCEs).
In addition to the expectations for stability data detailed
above, regulatory authorities on occasion have requested
supplementary confirmatory stability studies when there
have been concerns that these assessments are insufficient
to ensure quality through expiry/retest period.
While the ICH Q1A details reasonable expectations for
the type and quantity of stability data required to establish a
drug substance retest period, flexibility is allowed to encom-
pass a variety of practical considerations, and alternative
Table 1 Summary of regulatory expectations for registration stability
Regulatory Comments/experiences
governance
FDA/ICH Three pilot-scale batches by the same synthetic
route, method of manufacture and procedure
and representative of commercial route. Although
12 months is the expectation, 6 months may be
acceptable in some cases.
WHO Three pilot-scale batches by the same synthetic
route, method of manufacture and procedure and
representative of commercial route.
Japan Although not an actual requirement, there is a JP
expectation that “pilot-scale” is 10 % commercial
scale.
China Expectation that “pilot scale” is equal to amount of
API needed to manufacture 10 % of the
commercial Drug Product scale.
Brazil At least one batch should be made at the proposed
commercial site
approaches can be used when scientifically justified. Similarly,
ICH Q8, Q9, and Q10 encourage science- and risk-based
approaches which can lead to innovation and continual im-
provement. Unfortunately, these scientifically sound alterna-
tive approaches are often not fully embraced globally in lieu of
the more conservative prescriptive interpretation of the gui-
dances. The flexibility to use alternative approaches as de-
tailed in Q1A can be partially offset by different regional
interpretations of the guidelines, and supply chain logistics
can be challenging when flexibility is not globally accepted.
Traditional Industry Practices
Clinical Stability
A typical industry practice has been to place each clinical
drug substance batch manufactured using a new synthetic
route or process on both accelerated and long-term stability.
These batches are then monitored for various quality attrib-
utes (e.g., assay, purity, polymorphic form, water content,
etc.) regardless of whether or not these attributes are
stability-related or are impacted by the change in the route
or process. In many instances, this conservative approach to
selection of drug substance stability lots has led to an
unwieldy number of clinical stability programs, with little
regard to the actual scientific justification or value of the
data (Li et al., in preparation).
Registration Stability
Based on what may be considered a conservative interpre-
tation of the ICH guidance, some companies have perceived
a regulatory expectation that the drug product stability

batches be manufactured with the same drug substance
batches used for the registration drug substance stability
programs. As such, to achieve a minimum of 12 months of
drug product and drug substance stability data for three
pilot-scale batches using the proposed commercial route
and processes, companies often manufactured three consec-
utive batches in their commercial facility 18+ months prior
to product registration. Even though the material may be
used for ongoing clinical trials and commercial product
launch, there is typically a significant expense associated
with use of the commercial facility, as well as the risk of
unanticipated project termination. Additionally, the use of
three consecutive batches manufactured using the same
batches of starting materials, manufacturing equipment,
and manufacturing process does little to demonstrate stabil-
ity robustness over several years of commercial production
(Fig. 1, top). The batches produced with the same drug sub-
stance starting materials, manufactured in the same equipment,
by the same staff will be essentially identical. This approach
provides limited statistical value, limits understanding of the
stability-related quality attributes, and limits understanding of
the manufacturing design space.
Post-approval Stability
Stability studies are traditionally carried out on the first three
commercial production batches (if not part of the original
registration) and at least one batch annually thereafter to
confirm the commercial retest period. Drug substance batches
were also traditionally set up and evaluated for stability to
support site changes.
Impurity Levels
nt
e
lv
Water Content
So
Initial 3 months
Impurity Levels
nt
e
lv
Water Content
So
Retest Limit
Internal Limit
Fig. 1 Manufacturing three consecutive batches produced using the
same manufacturing parameters and ingoing materials will result in drug
substance with similar stability-related quality attributes. On stability,
these batches demonstrate essentially equivalent behavior over time
(top row). Alternatively, using developmental materials manufactured
J Pharm Innov
When the traditional stability strategies detailed above
are considered, several questions emerge:
& Are these practices actually providing useful information?
& Can we leverage our enhanced understanding when we
approach stability studies?
& Are there alternative (perhaps better) ways to meet our
stability commitments?
Stability Strategies that Leverage Enhanced Understanding
Clinical Stability
As an alternative risk-based approach to supporting clinical
programs, only those changes which affect potential
stability-related quality attributes (e.g., polymorph, surface
area) would trigger the need for a subsequent batch to be
placed on stability. Early in development, the stability-
related quality attributes may not be fully understood, and
in those instances, accelerated stability studies may be un-
dertaken to determine whether stability performance has
been affected before committing to long-term stability stud-
ies. These studies are particularly valuable to efficiently
demonstrate equivalent or superior stability performance
during the development process in lieu of or in addition to
long-term stability programs.
Although the harsh conditions used in forced degradation
conditions may generate degradants that are never seen when
the drug substance is stored under normal conditions, stress
testing and forced degradation studies can help identify po-
tential degradation products, pathways, and the intrinsic
z
6 months 9 months 12 months
3 Batches w/ Similar Material Attributes
3 Batches w/ Different Material Attributes
with different processes or ingoing materials helps to identify stability-
related quality attributes which require control. The batch represented by
the yellow ball is less stable than the other two batches and is illustrative
of a batch that has a stability-related quality attribute that requires control
J Pharm Innov
stability of the moiety. These studies also facilitate develop-
ment of the analytical procedures capable of detecting these
degradants if they actually do occur. Similar predictive stabil-
ity studies may also be used to identify and assess stability-
related quality attributes. Identification of these attributes dur-
ing the development process provides the opportunity to con-
trol these attributes as part of the overall manufacturing design
space strategy and minimize lot-to-lot variability.
Accelerated stability assessment program (ASAP) [9] is
an example of a predictive stability tool that has been used
to identify stability-related quality attributes. As an exam-
ple, for a specific drug substance process in development,
sodium thiosulfate was generated at differing levels as a
function of the crystallization process. Figure 2 illustrates
that higher levels of this impurity significantly improved the
overall stability of the drug substance and minimized for-
mation of the primary degradation product on stability. This
study revealed that the levels of sodium thiosulfate (a
stability-related quality attribute) must be controlled to min-
imize lot-to-lot variability on stability.
Although Fig. 2 illustrates how an accelerated study was
used to demonstrate apparent batch-to-batch differences in
drug substance stability performance, this represents the
exception to the usual case wherein ASAP demonstrates
similar stability behavior between batches.
Figure 3 illustrates the stability data from another ASAP
experiment for a different drug substance. In this case, differ-
ences in degradation rates from batch-to-batch are observed
under extreme conditions of high temperature and humidity.
Interestingly, using the ASAP moisture-corrected Arrhenius
equation to model the long-term stability of these same data
(Fig. 4), there is little batch-to-batch variability in the pre-
dicted review periods, regardless of changes in the manufac-
turing process.
The examples detailed above illustrated (1) batch-to-batch
variability of a relatively unstable drug substance because a
Fig. 2 Results of accelerated
stability assessment program
used to reveal lot-to-lot variabil-
ity based on experimentally
designed process changes. Lot 4
had the highest levels of sodium
% Degradation Product
thiosulfate and was the most sta-
ble. As the levels of sodium
thiosulfate decreased, instability
increased. Lot 2 had the lowest
levels of sodium thiosulfate;
30 °C/60 % RH is considered an
accelerated condition for this
drug substance because it is a
relatively unstable compound
stability-related quality attribute was not controlled (Fig. 2)
and (2) batch-to-batch variability at extreme conditions
(Fig. 3), but equivalent retest periods under normal storage
conditions (Fig. 4). A third example leveraging accelerated
stability conditions to evaluate equivalence between batches is
shown below. For this drug substance, stability is demonstrat-
ed for both batches even under extreme conditions (Tables 2
and 3). These data are consistent with a very stable drug
substance with all of the stability-related quality attributes
consistent between batches.
Using ASAP or other accelerated stability challenges,
stability issues that traditionally would have only been
detected after extended time can be rapidly detected. Chem-
ists can quickly determine what are the stability-related
quality attributes and the impact of changes to stability-
related quality attributes with process changes. This type
of information can be used during development to address
questions such as:
& Can this batch of drug substance be used to manufacture
drug product (or is a rework/reprocess of the drug sub-
stance advised)?
& Should this batch of drug substance be set up on a
formal stability program?
& With enhanced product knowledge, if subsequent drug
substance batches are placed on stability, would a leaner
stability strategy be appropriate?
With these questions in mind, the following stability
strategies are advocated for stable drug substance stability
programs to support clinical trials (if adequately justified):
& Initial retest period assignment of 18 months based on
purposeful degradation studies, in silico modeling, or
stressed conditions (70 °C/75 % RH)
& Stability on first good manufacturing practice (GMP)
batch set up prior to release, at accelerated, long term,
Lot-to-Lot Variability in
Primary Degratation at (30°C/60% RH)
14
12
Lot 1 Lot 2
Lot 3 Lot 4
10
Lot 5 Lot 6
8
6
4
2
0
Initial Day 20 Day 30 Day 44
Time at 30°C/60%RH
 J Pharm Innov

Fig. 3 Stability data from an Use of Accelerated Stability Assessment Programs to

ASAP experiment showing lot-
to-lot variability under extreme Evaluated Impact of Process on Stability Performance
3.5
conditions of high temperature
High Shear Wet Mill
and humidity. Air classified 3 Sonocrystallisation

Impurity Levels (%)

milling (blue) shows signifi- Un-milled
cantly higher impurity levels 2.5 Air Classified Milling
after storage under several
2
different conditions
1.5

0.5

70C/5%RH

80C/5%RH

80C/5%RH
50C/75%RH

60C/40%RH

70C/75%RH

70C/75%RH

80C/40%RH

80C/40%RH

80C/40%RH
60 days 60 days 40 days 30 days 60 days 30 days 60 days 20 days 40 days 60 days

Storage Conditions

and 1 condition below long term for trending across
temperatures. Reduced testing employed at initial, 3,
12, 24, and 36 months. But no testing at 6 weeks, 6, 9,
or 18 months.
& No external standard assay testing if the lot on stability
is the same as that used for the working standard.
& For route changes, if scientific justification has deter-
mined that the original batch placed on stability remains
representative of the new route, then the new batch is not
placed on stability. Otherwise, a reduced testing strategy
is prosecuted to justify the drug substance retest period
assignment.
Registration Stability
The stability studies conducted to support registration
should be information rich. Batches should be scientifically
selected as the drug substance synthetic process is devel-
oped and ideally should vary in the levels of the stability-
related quality attributes, as long as the attributes remain
within their control limits. In addition, these batches should
also vary in the levels of attributes that are not stability-
related as these studies can validate that these attributes do
not have an effect on stability. An example of this would be
to select batches for stability that vary in the level of residual
solvent. Changes in residual solvents levels generally will
not have an effect on the stability of the drug substance.
Batches should not be specifically made for stability studies
but should be scientifically selected as batches are made
during development to support clinical trials, safety studies,
site qualification, or for other uses. The registration stability
package can be further strengthened by including additional
supportive stability data on batches made with processes
that differ from the commercial process but can help support

Fig. 4 Results of ASAP Retest Period Estimates Based on Modeling of Degradation Product
statistical model showing in Batches Manufactured using Different Milling Processes
estimated retest periods based on
14
the same data presented in Fig. 3.
This clearly demonstrates that
despite significant changes in 12
Air Classified Milling
impurity profiles under extreme
High Shear Wet Mill
stability conditions, batches from 10
Retest Period (years)

all four processes are expected Sonocrystallisation

to have a similar retest period 8 Un-milled
under milder storage conditions.
If the drug substance storage 6
conditions were specified as
25 °C/60 % RH, all batches could
4
easily achieve a conservative
retest period of 5 years
2

0
25C/60% RH 30C/65% RH 40C/75% RH
Storage Condition
J Pharm Innov

Table 2 Stability results for batch # 1 that matrixing for drug substance stability will be of limited
Test 70 °C/75 %RH 80 °C/40 %RH utility and traditionally would not be considered by most
Initial 14 Days 14 Days pharmaceutical companies for registration stability. Certain-
ly, a matrixed approach would not be suitable for a com-
Appearance Off-white Off-white Off-white pound that has moved quickly through development with
powder powder powder
limited product understanding. For cases where the appli-
Assay 99.5 99.0 99.5
cant will propose supplying less than 12 months data at the
Impurities
time of the filing, a matrixed approach may not be suitable
(% area)
either. That said, for drug candidates that have been in
(a) NMT 0.05 NMT 0.05 NMT 0.05
development for long periods, for those with extensive
(b) NMT 0.05 NMT 0.05 NMT 0.05
product knowledge, or for those cases with a large number
Water content 0.24 0.18 0.21
(%) of drug substance batches that have been monitored during
PXRD Form A Form A Form A development, a matrixed strategy for the registration stabil-
ity batches should be considered.
When considering risk and risk management for registra-
proposed retest dates and the proposed drug substance tion stability studies, industry should recognize that stability is
specifications. an important part of the much larger CTD Module 3, which is
Several options for the selection of batches to be included part of the complete registration dossier. The dossier will be
in the Common Technical Document (CTD) are illustrated deemed acceptable (or not) based on the quality of the dossier
in Fig. 5. As described previously, the traditional and most and other considerations such as the safety and efficacy of the
conservative approach is located on the lower left corner of product and whether the product represents an unmet medical
the figure. Although this approach will not likely be chal- need. In general, regulatory flexibility is more likely if the
lenged by the regulators, it can be the most costly and will product falls within a few special categories such as orphan
provide little opportunity to increase stability understanding. drugs or those that fill an unmet and/or critical medical need.
The option in the lower right corner represents a very Here, the applicant may be able to negotiate options such as
aggressive regulatory strategy and may be unacceptable if providing additional stability data during the review or flexi-
the drug substance batches selected for stability are not bility on what is considered a pilot-scale batch. Worst case
representative of the commercial process. Other cases are scenarios should also be reviewed as part of the risk analysis.
presented that moderate both risk and scientific value. It is The worst case may only require that the applicant initiate
evident that there are many options that should be consid- additional stability studies or provide data as a post-approval
ered when selecting the stability batches that will be used to commitment. The applicant should determine if the worst case
support registration. The risk (of not meeting regulatory scenarios are acceptable or require mitigation and/or negotia-
standards) associated with each of the Fig. 5 options can tion before filing.
be mitigated with increased product understanding. As de-
tailed earlier in this paper, accelerated studies can be used Post-approval Stability
during development to mitigate risk by identifying the
attributes that will have an effect on stability. Post-approval, there is a global GMP expectation that at
A stability strategy that may be appropriate for some drug least one batch of drug substance is placed on stability per
candidates is matrixing. Interestingly, ICH Q1D indicates year (assuming adequate production). But if the stability-
related quality attributes are understood and if post-approval
Table 3 Stability results for batch # 2 batches meet all acceptance criteria, a traditional post-
approval stability program will not be value-added. To meet
Test 70 °C/75 %RH 80 °C/40 %RH
Initial 14 Days 14 Days a regulatory commitment for annual batches or the GMP
expectation for annual stability batches, a short ASAP study
Appearance Off-white powder Off-white powder Off-white powder may be an appropriate alternative to a traditional annual
Assay 99.5 99.0 99.5 stability program and can be proposed to the regulators.
Impurities For post-approval changes to the synthetic process, site
(% area) change, or batch size, if the stability-related quality attrib-
(a) NMT 0.05 NMT 0.05 NMT 0.05 utes are within specification, a traditional stability program
(b) 0.09 0.09 0.09 is not scientifically needed.
Water 0.18 0.20 0.20 Recognizing the current regulatory environment, it would
content
PXRD Form A Form A Form A
be overly optimistic to plan for global acceptance of all the
post-approval strategies detailed above. However, leaner or

Fig. 5 Several options for High
registration stability that vary
scientific value and regulatory
risk
Scientific
Value
Low
more nimble stability programs may be acceptable. With a
stable and well understood drug substance, the number of
stability pull points can be greatly reduced, especially dur-
ing the first year of storage. A matrixed testing protocol also
may be acceptable where all tests are not monitored at each
time point. Traditionally, the testing done at release is large-
ly duplicated on stability even though many of these tests
are not stability indicating (identification, water content), are
not stability-related quality attributes, and/or are not shelf-
life limiting attributes (assay). If the purpose of stability
studies is to confirm the drug substance retest date, then
the only attributes that needs to be evaluated on stability are
those that may limit the retest period. Use of a quality by
design approach to justify the removal of the assay from
active pharmaceutical ingredients (API) stability testing pro-
tocol has recently been proposed [10].
The lean strategies advocated above were recently
employed in several marketing applications for several
new chemical entities. Leveraging significant product un-
derstanding that was detailed in the marketing application,
regulatory relief was secured on post-approval stability
commitments in several markets:
& Six months of API stability was provided in the initial
filing and a 24-month retest period was approved.
& Appearance and purity were the only tests on both post-
approval and annual stability commitment protocols
& The first stability pull point was at 12 months and annually
thereafter.
& Only one storage condition (25 °C and 60 % RH)
& Markets that approved this included: USA, European
Union, Japan, Switzerland, and Canada. Approvals in
addition regions are pending.
Stability Risk Assessments and Regulatory Acceptance
of New Stability Strategies
New concepts that require regulatory approval will understand-
ably generate resistance from the authorities. Recognizing that
J Pharm Innov
Batches scientifically Batches scientifically
selected as the DS selected as the DS
process is developed. process is developed
Batches are but may have minor
representative of the differences from the
commercial process. commercial process.
Batches selected as the DS
process is developed. One
batch made at the
commercial site.
Clinical stability for each
new process and 3 Phase 1, 2 DS used for
commercial campaigns Registration Stability
for Registration stability
Low Regulatory Risk High
resistance is directly proportional to impact and inversely pro-
portional to trust (resistance α impact/trust), industry should
request a reasonable amount of flexibility (lowering the impact)
and work hard to clearly articulate their proposals with sound
scientific data (and hence maximizing trust). While formal risk
assessments of manufacturing processes are well established
and leveraged in quality by design (QbD) submissions, risk
assessments of stability are less well established and have been
underutilized by industry. A process for formal stability risk
assessments has been presented [11] and these were leveraged
by Pfizer for several NCE CTDs that were submitted for global
approval in 2011. While regulatory feedback varied across
regions and even within regions (for different products), a trend
towards acceptance of leaner and “alternative” stability strate-
gies is clearly evident. Interestingly, these concepts are finding
acceptance not only in the major markets but in the Emerging
Markets region as well. Day-80 feedback for a NCE under
review in the EU is an example of current thinking: [12]
“The proposed post-approval testing frequencies, stor-
age conditions, and testing methods are based on a QbD
approach and take into consideration the findings of the
above stability data. Testing done annually at 12, 24,
and 36 months will include appearance and purity. The
development formal registration stability studies con-
firm that more frequent testing and storage at the accel-
erated condition do not provide critical information
needed to establish the retest period or monitor stability
and quality of the drug substance. Assay was considered
as non-stability indicating as the level of the total impu-
rities remained consistent for the pilot long-term stabil-
ity batches while the assay values varied. In addition,
chiral purity, microbial quality, and water content were
also investigated and deemed non-stability indicating.”
A Proposed Path Forward
A continued partnership and scientific-based collaboration
between industry and the regulators is needed. Industry needs
J Pharm Innov
to identify stability-related quality attributes and the process
parameters that may have an effect on stability. With this
scientific understanding, stability studies can be better engi-
neered and low-value programs can be avoided. Alternative
stability approaches should be leveraged and proposed to
regulators as alternatives to traditional stability commitments.
Achieving global acceptance of the strategies for stability
studies detailed in this paper will take some time but will be
worth the effort because the patients will ultimately benefit.
Acknowledgments The authors would like to thank Bob Baum,
Stephane Caron, Chi-wan Chen, Dave Damon, Allan Edwards, John
Groskoph, Chuck Hoiberg, Vince McCurdy, Megan McMahon, Julian
Meekings, Greg Sluggett, Robert Timpano, Ken Waterman, Kevin
Ryan, Roger Weaver, and PhRMA Stability Expert Team.
References
1. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q1A(R2): stability testing of
new drug substances and drug products. Rockville: FDA; 2003a.
2. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q1D: bracketing and matrixing
designs for stability testing of new drug substances and drug
products. Rockville: FDA; 2002.
View publication stats
3. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q1E: evaluation of stability
data. Rockville: FDA; 2003b.
4. International Conference on Harmonization. International Conference
on Harmonization Guideline Q7A: good manufacturing practice guid-
ance for active pharmaceutical ingredients. Rockville: FDA; 2001.
5. WHO. Stability testing of active pharmaceutical ingredients and
finished pharmaceutical products. In: WHO Expert Committee on
Specifications for Pharmaceutical Preparations. 43rd report. Gene-
va, World Health Organization, 2009, Annex 2 (WHO Technical
Report Series, No. 953).
6. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q8(R2): pharmaceutical devel-
opment. Rockville: FDA; 2009.
7. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q9: quality risk management.
Rockville: FDA; 2005.
8. International Conference on Harmonization. International Confer-
ence on Harmonization Guideline Q10: pharmaceutical quality
system. Rockville: FDA; 2008.
9. Waterman KC, Carella AJ, Gumkowski MJ, Lukulay P, Macdonald
BC, Roy MC, Shamblin SL. Improved protocol and data analysis for
accelerated shelf-life estimation of solid dosage forms. Pharm Res.
2007;24(4):780–90.
10. Skrdla PJ, Wang T, Antonucci V, Dowling T, Ge Z, Ellison D, Curran
J, Mohan G, Wyvratt J. Use of a quality-by-design approach to justify
removal of the HPLC weight % assay from routine stability testing
protocols. J Pharm Biomed Anal. 2009;50:794–6.
11. Weaver R. AAPS Annual Meeting presentation stability risk
assessments with a focus on physical stability issues. 2011.
12. Co-Rapporteur Day 80 Critical Assessment Report for a New
Chemical Entity. 2012.