IN-USE STABILITY TESTING:

W H A T D A T A ARE REQUIRED AND

WHEN?
Scott Sutton. Brian Matthews and Danny Dunn
discuss the implications of the lack of guidance for
human pharmaceuticals in the EEA. and draw
attention to some practical limitations

This article considers current guidance on in-use stability testing in the Europcan Economic
Area (EEA). It highlights the problems arising from the absence of specific guidance for human
medicinal products and illustrates these using two examples of different product types. Finally, a
proposal is made for a guideline on in-use stability testing, specifically for human
medicinal products.

Current requirements
The Europcan Community (EC) Committee for Veterinary Medicinal Products (CVMF) has
recently adopted a guideline for in-use (open-bottle) stability testing1. There is no equivalent
No guidance on document offering guidance in the European (i.e. the EC and EEA) human medicinal product
in-use stability testing sector. However, several Committee for Proprietaly Medicinal Products (CPMP)/CVMP and
existsfor human International Standards Organisation (ISO) guidelines have recently addressed the need and
medicinal products conditions for pe&orming an in-use stability study as a part of the development process for new
medicinal products and medical devices2-'.
It is clear from these documents that chemical, physical and microbiological data arc
expected to support the proposed in-use stability period for products designed to be used on inore
than one occasion. In recent questions from regulatory authorities it has been suggested that
in-use stability studies should incorporate removal of the product from the container on a number
of occasions under in-use conditions with the product left out of its carton.
Studres desrgned by The absence of a detailed and authoritative guideline for the human pharmaceutical sector
manufacturers may not docs not assist in the design and conduct of such studies, and when manufacturers design their
necessarily be acceptable own studies it is by no means certain that the regulatoly agencies will find the data to be
to regulators completely acceptable. Furthermore, there are circumstanceswhen the generation of data from
specific in-use simulations may not add sigmficantly to the data already generated from
International Conference on Hamonisation of Technical Requirements for Registration of
Pharmaceuticals for Human Use (1CH)-compliant stability studies. There are also situations
(e.g. when small multi-dose containers are involved) where it is difficult to conduct meaningful
studies due to the limited quantities of residual product on which to perform chemical or
~nicrobiologicaltesting.

In-use stability testing: considering the problem

Two product types used This question will be explored using two different product types. At one end of the spectrum is a
to rllustrate the problems bulk pack used in a pharmacy, and packaged in a glass container. At the other end is an
ophthalmic medicinal product, packaged in low-density polyethylene (LDPE)~.

Scott Sutton, PhD is Associate Director, R&D Microbiology, AlconLaboratories, Fort Worth. Texas, USA. Brian
Mattllovs, B P h m , PhD, MRF'hannS, MBIRA is Director. EC Registration at Alcon Laboratories (UK), Hemel
Hempstead, UK. Danny L Dunn, PhD is Senior Director, R&D Analytical Chemistry, Alcon Laboratories, Fort Worth,
Texas, USA.

OCTOBER 1998 OTHE REGULATORY AFFAIRS JOURNAL

Four different aspects of the problem will be addressed:
. package type: (a critical parameter) if the container is completely filled and oxygen
impermeable then there will be no opportunity for osygen degradation to become
Different aspects
to be addressed..

. apparent during product development stability studies;

susceptibility to oxidation: if the product, or components of the product, are volatile or
oxygen labile then special concern should be reserved for the effects of multiple
aerations;
in-use period: if the proposed in-use shelf-life is beyond one month then the 28-day
compendial presewativeefficacy test does not provide a satisfactoty level of assurance as

. to the adequacy of the presewative system;

how should an in-use test be performed: signif~canttechnicaland practical issues exist
concerning the design of a test to support in-use stability.

PACKAGE TYPE AND SUSCEPTIBILITY TO OXIDATION

Guidance for in-use preservative testing is described in an EC guideline C&IP gu~dancestates
( E M E A I C V M P I ~ ~which
~ I ~ ~specifically
~) states that additional testing may be applicable to that addrtronal tests
multi-dose or topical dosage forms if a specific problem (in terms of stability of the product) has may be necessaryfor
been identified after the product has been opened. For example, tlns situation would apply to a multr-dose ov topcal
product with anactive ingredient that is subject to oxidative degradationand has been packed in a dosagefornls
completely filled oxygen-impermeablecontainer (e.g. a pharmacy bulk pack). Its glass package
provides a barrier to oxygen so that oxidative stress would not be seen during the normal stability
program. The situation would be exacerbated if the active ingredient, the antin~icrobial
preservative or any other critical formulationcomponentexists in a reducedform. A real concern
would exist about the potential for oxidative degradationof the fonnulatedproduct upon repeated
exposure to oxygen during use.
Contrast that situation with the hypothetical ophthalmic eye-drop. The most common
container resin for ophthalmic products is LDPE as described in the European Pharmacopoeia
( ~ h ~ u r )As
' . discussed before, aprimaty purpose of in-use stability testing is to establish that the
active ingredients or presen~ativeare not degraded by exposureto o'ygen while being dispensed.
LDPE is known to be readily permeable to oxygen7, and ophthalmic containers are intentionally
prepared w-itha 2-5 mL &headspace. This allows the bottle to be more 'squeezable' and helps to
ensure that drops are dispensed instead of a 'liquid stream'. In this situation any oxidative
degradation of an active ingredient or presewative due to exposure to oxygen would be readily
evident during normal stability studies which extend over 36 months or longer at storage
conditions of 25 OCI 10 'C and possibly 30 "C,at various humidity conditions. It is not clearwhat
additional stability infomiation the simulation of in-use conditions would provide for this type
of product.

IN-USE PERIOD
In these examples, the phannacy bulk pack may be used for months wlule the ophthalmic Stability studies
eye-drop normally has a standard in-use period of 28 days or fewer. The eye-drop period of use is conducted in line
in agreement withguidance document C P M P I Q W P / ~and ~ ~will
/ ~be
~ justified
~ by the chemical with ICH vequirements
stability and the preservative efficacy demonstrated during ICH-compliant stability studies. This should be sufficier?t
is noted in the introductoty text to the Pli Eur antimicrobial presemtive effectiveness test8.
which declares that the tests provide proof of 'adequate protectionfrom adverse effects that may
arise from microbial contamination or proliferation during storage and use of the preparation'.
This interpretation seems to be one of the rare instances of agreement among pharmacopoeias on
the subjcxt of antimicrobial preselvative efficacy testillgg.lO.
This interpretationis also supported in the scientific literature. Several recent studies have
correlated the activity of the presewative system with the performance of the finished product.
Two of these articles describe work involving members of the British Pharmacopoeia (BP)

OTHE REGULATORY AFFAIRS JOURNAL OCTOBER 1998

 Working Party on Antimicrobial Preservative Efficacy Tests. In the first, Davison et all1
examined a variety of used ophthalmic products, correlating the level of contamination seen in
actual use with the level of antimicrobial efficacy as determined by the compendial test. They
found that the BP criteria then in force and the then-proposed Ph Eur 'A' criteria, correlated with
low contaminationlevels in actual consumer use. The second study in this series12 extended the
evaluationto a wide range of medicinalproducts.It was found that meeting the PhEur 'A' criteria
provided protectionagainst contaminationin actual consumeruse in these different dosage forms
as well. This correlationhas also been observed with personal care products, where products that
had more efficaciouspreservative systems (as determined by compendialtests) were more highly
resistant to contaminationin a simulated patient abuse s t ~ d y ' ~ .
Productspackaged in A second concern is described by the guidance document C P M P I Q W P I I ~ which ~I~~
large volumes intended identifies a need for inclusion of an 'in-use shelf-life' for medicinal products packaged in large
for long-term use ma.v volumes and intended for a longer in-use shelf-life (section 3.3.1.1). It is also interesting to note
require additional testing that most veterinary products (covered by E M E A I C V M P I ~ ~as~ discnssedpreviously
I~~~, inthis
article) are packaged in large volumes, and so are likely to be used well beyond the 28 days
covered by the compendial antimicrobial presewative effectivenesstest.
Additional light is shed on this topic in a draft IS0 standard for the testiug of contact lens
care products which has been prepared by IS0 Technical Committee 172 Subcommittee 7
(ISOITC 172lSC 715. Although this is directed to the medical device industly. it discusses the
Gurdelme on multr-dose topic of in-use shelf-llfe to a great extent. The underlying principle is that 'the preparationmust
presevved contact lens meet the requiremeuts for an adequately preserved contact lens care product throughout its
care products grves intended discard date'. The intended discard dates are given as three months or greater due to the
addrtronal mformatron large size of the containers (360 mL).
The hypothetical bulk pharmacy package described earlier. with an extended in-use
shelf-life, should require investigation beyond the compendial antimicrobial effectivenesstest.
However, products inteuded for only 28 days in-use dating that have acceptable
presewative effectiveness (e.g meeting the 'A' criteria) are a different matter entirely. The
compendial presewative efficacy test covers the entire in-use period and is monitored for the
storage shelf-life of the product during its development. The level of challenge in such tests is in
the order of lo6 Colony Forming Units (CFU) of the index organisms per mL: far in excess of
what is likely to be encountered in an in-use situation. The Ph Eur indicates that meeting the
'A' criteria is accepted as demonstntion that the product has 'adequate protection from adverse
No additional testfng effects that may arise from microbial contamination or proliferation during storage and use'.
should be required for Therefore. for a product that meets the 'A' criteria consistently during testing, the
small volume eye-drop pharmacopoeialexperts who devised the preservative efficacytest presumably were content that
preparations meet~ngthe there was no need for additional testing. Indeed, it is not clear what additional information the
Ph Eur 'A ' criteria simulation of in-use conditions would provide.

HOW SHOULD THE TEST BE PERFORMED?

Guidance on testfng There is no guidance on the performance of this type of testing for the human medicinal product
is lacki.ng industiy. Two guidance documents do provide some direction, and these will be considered. The
cW'and ISO' documents, one for veterinary products and the other a draft standard for
medical devices (contact lens care products). are consistent in their approach. The investigatoris
to dispense the product in a manner similar to its use and then test the chemicallpl~ysical
properties and presewative effectiveness at the end of the in-use period. Is this approach
appropriate for medicinal products?
CTWP andISO gurdance In the first case, that of the pharmacy bulk pack, this is a reasonable approach. The
may be apphed to container may be entered repeatedly during a period of s e v e d months and can be expected to
pharmacy bulkpack provide a suitable volume of product for testing after the proposed in-use stability period. The
products, but not second case, the ophthalmic eye-drop, is not so simple.
ophthalmrc eye-drops

OCTOBER 1998 0THE REGULATORY AFFAIRS JOURNAL

 The frequency of use for ophthalmic eye-drops is generally given as two to four times a day.
Presumably, therefore, the simulated in-use test should cover the worst case. There is potential
for microbial contamination each time the product is opened and used, as well as potential for
oxidative degradation.
Consider the potential for oxidative degradation. Does chemicaYphysica1testing need to Sinall volume of
be done on drops removed from the container as the test proceeds, or is it sufficient to perform product and container
chemical/physical testing on the small amount of sample remainingat the end of simulated in-use poses problems
study? In either case, it will be a significant technical challenge to analyse the small volumes
available with appropriate precision and accuracy.
Consider the challenge to the preservative system. How great a challenge will there be?
Should this test be designed (as suggestedby CVMP and IS0 guidance) to simulate normal use,
and then challengedusing current compendialpresenaive testing? If that is to be the case, then a
significant logistical issue arises: as the product is intentionally packaged in a volume sufficient
for only 28 days of use (generally 5 mL/unit). at the end of the 28 days there is very little left to
test. This makes performing the test extremely challenging as the preservative test normally
requires approximately 100 mL of sample.
One option for testing the preservative system might be to perform a test in vibo to
, simulate repeated microbial challenges. Such a test has been described by Urhan et all4. They
snggest a repeated challenge: using a battery of challenge organisms, which would provide a
model for a simulatedin-use testing. By this method, the sample is challenged repeatedly over the
28-day in-use shelf-lifewitha low level ( 1 0 ~ - 1 0CFU1mL)of
~ eachorganism, and the efficacy of
the preservative system is evaluated under this simulated abuse situation. This approach has the
advantage of simulating consumer abuse of the product, and may be designed to monitor the
entire in-use shelf-life. It also allows for pooling of small volume medicinal products to increase
the sample size. This method has some drau~backs,however, as it would be impractical to use the
same inoculation level as that employed in the presewative efficacy test if this was to be
introduced four times a day over 28 days. Therefore, some compromises must be made as the
container is too small to hold that much product and inocula (and the dilution effects on the
product's preservation system would become significant anyway).
Additional questions concerning in-use stability testing protocols include:
. is it sufficientto in-use test only one lot or should multiple lots be tested;
if the product is sold in multiple sizes, do all sizes need to be in-use tested; and
Other questions ar~se..

should in-use testing be performed on fresh samples or on samples at the end of their
shelf-life?

Regulatory guidance
As indicatedpwiously, there are regulatory guideline requirements (stated or implied)for in-use Consultation on any
stability testing, but there is no regulatory guidance on how to conduct this test in the human proposedgzndance
medicinal product sector. It would be of great assistanceto manufacturers ifthe regulators would would be welcomed
consult on proposed guidance, propose data that they would like to see and define the type of b-v manzrfacturers
studies that they would find acceptable. However, in devising the guidelines it is essential that
due consideration be given to the practicalities of actually conducting the test. In particular. the
restrictions on cond&ing such tests with small-volumeliquid dosage forms should be taken into
account. In these cases, the availability of sufficient product for testing is severely limited,
especially if the recommended frequency of dosing is relatively high.
There should also be guidance on situations where in-use testing would not be considered ... and consideration
necessiuy. For example, products packaged in containers permeable to oxygen with an given to case.c where
intentional head-space should be excluded from in-use testing as any potential for the oxidative in-use testing is
degradation of an active ingredient or preservative would be readily evident during normal not necessary
stability studies. It is also suggested that fuller account should be taken of the Ph Eur's
presewative efficacy test results: no additional preservation testing should be required for

OTHE REGULATORY AFFAIRS JOURNAL OCTOBER 1998

 products which are shown to meet the Ph Eur's 'A' criteria in the normal stability tests. If
regulators consider that such tests are required, there is a special need for advice on how they
would wish the test to be conducted (e.g. the types of organismto be used. the inoculum level and
whether or not the testing should be conducted in the container).

Summary
There are situations where studies designed to address the question of in-use shelf-lie are
desirable and necessary. Current regulatoy documents from the veterinruy pharmaceutical and
medical device sectors provide guidance on how to perform this test. There is no equivalent
detailed advice in the humanphmaceutical sector. We suggestthat there are two conditions that
would prompt this investigation:
Situations where testfng .if the container is completely filled and oxygen impermeable: there will be no
would be necessary opportnnity for oxygen degradation to become apparent during product development
stability studies; if there is an ingredientin theproduct that is oxidation sensitive, then the
opening of the pack will provide the first challenge to the product; and
where the intended in-use shelf-life exceeds one month: if the proposed in-use
shelf-life is beyond one month, then the 28-day compendia1 presewative efficacy test
should be supplementedwith a simulated in-use test that lasts the duration of the intended
in-use shelf-life; this would be the case for large volume. multiple-use products with an
in-use shelf-life of several months.
Finally. the absence of a detailed guideline for the human medicinal product sector adds
The need for confusion to an already uncertain topic. Sigmficant technical issues prohibit the use of the
guidance specijkallv CVMP-based approach in small volume liquid mediciualproducts since little or no sample
on human medicinal remains at the end of a simulated consumer-use study. There is a definite and urgent need for
products is emphasised additional guidance from the European regulatoly agencies.

Proposal
The following proposal is put fonvard as the basis for a specific guideline for in-use stability in
the h m n medicinal product sector.
Before considering the provision of in-use stability data, due account should be taken of
the available information from conventional stability studies. However, results from in-use
stability testing should be supplied as p a t of a marketing authorisation application for a human
medicinal product in the following circuinstances,where:
In-use stability
testing should be
. the manufacturer's recommended period of use is longer than that recommended in the
relevant monograph of the Ph Ewl
required where ... there is aprima facie case for cheinical and/or physical testing being relevant (e.g. in
cases where the head-space gas in the container has been replaced with an inert gas and

. where interchange with atmospheric oxygen is possible);

there is aprima fade case for antimicrobial presewative efficacy testing (e.g in cases
where the antimicrobialpreservationefficacy test data for the product suggest that it will

. not consistently meet the 'A' criteria of the Ph Eur test): and
it is physically possible to simulate use of the product and have sufficient left to allow
analytical tests and/or antimicrobial presewative efficacy tests to be conducted on
the remainder.
Where the first three of the points givenabove do not apply to a particularproduct, no additional
testing should be required. When one or both of the second and third conditions apply,
two representativebatches of the product should be tested, at least one of them being towards the
end of its shelf-life. As far as possible the test conditions should simulate the practical use of the
product (volume of product removed and frequency of removal according to the instluctions for
use). The actual tests to be applied should relate only to those factors which might be of concern

OCTOBER 1998 BTHE REGULATORY AFFAIRS JOURNAL