Page 241 - Biomedical Engineering and Design Handbook Volume 2, Applications

220 MEDICAL DEVICE DESIGN

There are several references to consult to get a better understanding of all the considerations and
to determine rational sample sizes for package system validations. These include The Handbook of
Applied Acceptance Sampling: Plans, Procedures, Principle, and an article published in MDDI
Magazine “Sample Size Selection Using a Margin of Error Approach,” by Nick Fotis and Laura Bix.
The sample size will vary, depending on several characteristics of the validation plan:

• The packaging system being evaluated
• The type of result for the test (e.g., attribute or variable data)
• The risk tolerance of the company
Since the ultimate acceptance of the package system is dependent on the condition of the sterile
barrier system, the strength and integrity tests must be performed on a statistically significant sample
size based on the confidence interval and reliability.
When considering the sample size of the shipping box (protective package), it must be under-
stood that the unit of test is the shipping box, and one shipping box holding 24 sterile barrier sys-
tems is a sample size of one when subjecting it to the performance tests (distribution simulation).
Many studies will only require a sample size of one if the shipping box contains enough sterile bar-
rier systems to obtain a statistically significant sample size for integrity testing.
In all cases the sample size will depend upon company risk policy, tester reliability and precision,
economics, and regulatory requirements.

7.9.4 Defining the Shelf-Life Study
As the FDA moves toward harmonization with the European standards through revision of its GMP
and through adoption of ISO and CEN standards, the need for guidance on the performance of accel-
erated aging protocols is crucial.
The net result of publishing expiration dates is that there must be some documented evidence
which supports the product expiration claims; thus, the need to perform shelf-life studies. However,
real-time shelf-life studies are not an alternative in a fast-changing industry which can see two to
three generations of products developed over the time it would take to document a 2-year shelf-life
claim. So, the need for accelerated aging protocols as an alternative in developing a product and
introducing it into the marketplace in a timely fashion is essential. Concurrent real-time studies must
be performed to substantiate results of accelerated studies.
Ideally, accelerated aging involves a single measurable characteristic under extreme conditions to
simulate, in a short time, the conditions the package would likely be subjected to during its desig-
nated shelf life. Some protocols rotate the packages through three environments designed to simu-
late the aging process. These conditions include high temperature and high humidity, high
temperature and low humidity, and freezing conditions. The use of humidity in aging protocols has
been discussed earlier, and caution should be taken when using high humidity for aging studies. Low
temperature is included since it has been implicated in package failure through cold creep and mate-
rial embrittlement, and packages may, in fact, be exposed to these temperatures in winter time dis-
tribution systems or in the cargo areas of aircraft. However, these conditions represent more of an
environmental challenge and do not have any influence on determining the package shelf life.
Current information found in the previously mentioned guidance documents can be reduced to
four basic principles for determining shelf-life and consequent expiration dating:
• Determine an acceptable target expiration date based on R&D data, on the likely distribution and
storage conditions that the product will encounter prior to its use, and on the company’s market-
ing strategies.
• Select the test temperature parameters that will be tested based on the physical properties of the
packaging materials (e.g., glass transition temperature).
• Conduct the testing under consistent procedures.
• Once all the testing has been completed, validate the test data.