a;
.. “E?
z 3..

\. .

V I . ‘ .
Ami .Nﬁ. .nm‘ $0. a ‘
emmiawwﬂuwm‘uwmm
.
..

 

...A xv .
Fair}

«13. I
1.213

A

LU

m

 

Mljn

 

:4 V . .. . . ,ﬁhﬁﬂwﬁa.

This is to certify that the
dissertation entitled

THE EFFECT OF PRESSURE DIFFERENTIAL ON
MICROBIAL PENETRATION OF A STERILE MEDICAL
DEVICE TRAY

presented by

JANE ERIN SEVERIN

has been accepted towards fulﬁllment
of the requirements for the

degree In Packaging

 

vim {54g

yajor Professor‘s Siﬁiature
MS“ Mb

Date

 

MSU Is an AMnnative ActiorVEqual Opportunity Institution

— —-—.—._.—— ﬁ- -— -—__.-.——r

LIBRARY
MichIgan State
University

 

 

 

 

-—-_.—.-.—n-n--o-o--o-o-o-u—o-o-u--o-o-o-o-o-.-¢-.—g-a-n—u—.-o—---..—._._.—o-o-n-o-o-o-o-u-c-o-o-o-o-u-o-o-o-o—o-o-o-o-I-

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

AUGBtZZSW,

 

J‘llil 1 4 20M

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/CIRC/DateDue.indd-p.1

 

 

THE EFFECT OF PRESSURE DIFFERENTIAL ON MICROBIAL
PENETRATION OF A STERILE MEDICAL DEVICE TRAY

By

Jane Erin Severin

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
School of Packaging

2006

ABSTRACT

THE EFFECT OF PRESSURE DIFFERENTIAL ON MICROBIAL
PENETRATION OF A STERILE MEDICAL DEVICE TRAY

By
Jane Erin Sever-in

A developmental need of the medical device industry is a body of knowledge
regarding the required level of sensitivity for package integrity tests. As a result of this
need, the objective of this research was twofold: (l) to create and verify a new
methodology for whole-package microbial testing for medical device packages and (2) to
examine the effect of pressure differential on microbial ingress in a sterile medical device
tray.

The new method was a clear improvement over existing techniques because it
reduced the chance of false positives that were commonly observed in the previous tests
of integrity. The method involved aseptically ﬁlling sealed, sterile device packages with
a known volume of an appropriate grth medium, exposing the packages to a microbial
challenge, incubating the package and inspecting it for growth.

After creating and reﬁning the test technique, ﬁirther work explored the impact of
pressure differential and a limited number of hole sizes (0, 10, and 100 micron) on
microbial penetration. The speciﬁc pressure differential examined simulates the descent
of an aircraft ﬁom 8000 feet or a ground shipment descending from 8000 feet. While air
transportation is a common mode for medical device distribution, it is important to
recognize that packages are routinely exposed to pressure differentials in all elements of

handling.

Pressure differential and hole size had a statistically signiﬁcant effect on the
microbial penetration of the sterile medical device test tray used in this study. The
critical penetration threshold is less than 100 microns. These results and the development
and veriﬁcation of a whole package microbial challenge test method provide useful
information with which to build future research that seeks to determine the required level

of sensitivity for packaging integrity tests for sterile packaging.

C0pyright by
JANE ERIN SEVERIN
2006 -

This Dissertation is dedicated to my husband, Paul, and our children, Brenna and
Jack, in appreciation for their sacriﬁce, and for their love and support

ACKNOWLEDGMENTS

I would like to express my appreciation to my research guidance committee.
Dr. Laura Bix, my major professor, for her knowledge, for being a great advisor, for
always being available, for always being positive, for her great patience and for having
conﬁdence in me. I would like to express gratitude to Dr. John Linz, from the
Department of Food Science and Human Nutrition, for the great amount of time he
committed to ensm'ing that I understood the microbial procedures and concepts to
successfully perform this work. He also provided his laboratory for use for close to two
years. I thank him for his knowledge, support and patience. I would also like to express
gratitude to Dr. Dennis Gilliland, ﬁom the Department of Statistics and Probability, for
the great deal of time be invested in the study design and results analysis, but more
importantly for his patience and ability to teach the concepts of statistical analysis. I
would also like to express gratitude to Dr. Gary Burgess for his tremendous contribution
concerning the test methods employed in this research, as well as his ability to interpret
and understand the results. I would like to thank Dr. Hugh Lockhart for sharing his
wisdom, providing guidance, and for his tremendous support throughout this endeavor.
All of their contributions were critical to the completion of my research.

I would like to acknowledge the contributions of the following companies (in
alphabetical order): Amcor Flexibles for providing the Heat Seal Coated PTH-Ol7 1073B
Tyvek lid stock, Becton Dickinson, for their donation of a very large supply of syringes
and needles, Cardinal Health for providing the EtO sterilization for the sealed empty trays
and for a large supply of disposable exam gloves, Dupont, Lenox Laser, for providing the

laser drilling of the test trays, Medtronic for supplying the ATCC Strain of E. coli used,

vi

Mocon for providing the self sealing septum material, Oliver Products for providing a
large supply of corrugated shipping cases, Perfecseal for providing the PETG formed
trays, SenCorp for providing a signiﬁcant discount and technical assistance for the
MD 2420 heat sealer, and Smith and Nephew for providing radiation sterilization of the
trays. Additionally, I wish to thank the following undergraduate technicians:
Kristi Radakovic and Rebecca Schaeff for their work in the laboratory, and
John Paul Severin for his documentation support.

I would like to express special appreciation to my family for their support, love, and
encouragement. My husband, my best friend, also provided technical support, but more
importantly ﬁlled in for me with our children, Brenna and Jack, when I was busy at
school.

Finally, I would like to acknowledge my mother, my best friend and supporter
throughout my life. As I complete my doctoral degree, she suffers from the late stages of

Alzheimer’s Disease. I know that on some level she is aware of this accomplishment and

is very proud.

vii

TABLE OF CONTENTS

 

 

 

 

 

 

LIST OF TABLES x

LIST OF FIGURES xii

CHAPTER 1

INTRODUCTION 1
The Legislation and Regulation of Medical Devices .................................................. 1
The State of the Industry ............................................................................................. 2
Issues of Health and Package Integrity ....................................................................... 3
Package Integrity Testing ........................................................................................... 5
Basis for Developing a New Method .......................................................................... 7
Primary Research- Applying the New Methodology ................................................ 14

CHAPTER 2

LITERATURE REVIEW 17
Historical Perspective on Testing ............................................................................. 17
Varying Objectives ................................................................................................... 42
Varying Results ......................................................................................................... 43
Variables ................................................................................................................... 48
Conclusion ................................................................................................................ 5 7

CHAPTER 3

MATERIALS AND METHODS ‘8
Materials ................................................................................................................... 58
Services ..................................................................................................................... 61
Equipment ................................................................................................................. 62
Methods ..................................................................................................................... 64

CHAPTER 4

EXPERIMENTAL DESIGN 91
Variables ................................................................................................................... 91
Constants ................................................................................................................... 91
Response Variable .................................................................................................... 91
Justiﬁcation of Design .............................................................................................. 91
Variables Details ....................................................................................................... 92
Constants Details ...................................................................................................... 99
Test Groups ............................................................................................................. 103
Statistical Design .................................................................................................... 103

viii

CHAPTER 5

 

 

 

RESULTS 106
Results Summary .................................................................................................... 1 15
Statistical Analyses ................................................................................................. 1 19
Conclusion .............................................................................................................. 123

CHAPTER 6

DISCUSSION 124
Recommendations for Future Research .................................................................. 132

APPENDIX I 134

BIBLIOGRAPHY 149

 

LIST OF TABLES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 1- Classiﬁcation of Medical Devices (Rosen, 2004) 2
Table 2 - Materials Examined by Placencia et al (1986) 22
Table 3 - Physical Variables in Published Research 44
Table 4 — Microbiological Variables in Published Research 47
Table 5 - Percent of Flexible Pouches Found to have Contamination at Varying
Challenge Concentrations (Anema and Schram-1979) ‘0
Table 6- Percent of Flexible Pouches Found to have Contamination at Varying
Exposure Times (Anema and Schram—l979) ‘0
Table 7 - Colony Forming Unit Quantiﬁcation From Growth Curve Study

(Units CFUlsqin) 69
Table 8 - Microbial Dose Delivery 72
Table 9 — Microbial Dilution Volume/Resultant Concentration 86
Table 10 - Tray Lidstock Deﬂection 97
Table 11 — Trial One - 106
Table 12 - Trial Two 107
Table 13 - Trial Three (Replicate One) 108
Table 14 — Trial Four (Replicate Two) - 108
Table 15 — Trial One Results 109
Table 16 - Trial Two Results 1 10
Table 17 - Trial Three Results 1 11
Table 18 - Trial Four Results 112
Table 19 — Petri Dish Growth Quantiﬁcation Results 114

 

Table 20 - Aggregated Results of Positive Responses for n = 136 Trays ............... 123

Table 21 - Description of Test Groups 136

 

Table 22 - Test Group Procedure Description 138

 

Table 23 - Test Group 1 - Aseptic Technique with NO Alcohol Swabbing Through
Contaminated Fields with Varying Concentrations of Microbes (Steps 5 and 8

in Figure 22 are Eliminated but Injection Occurs Through a Contaminated

Field) 142

 

Table 24 - Test Group 2 - Aseptic Technique WITH Alcohol Swabbing Through
Contaminated Fields with Varying Concentrations of Microbes (Figure 22
Procedure Step 8 is Eliminated but Injection Occurs Through a Contaminated
Field) 143

 

Table 25 - Test Group 3 (CONTROL) - Full Aseptic Technique Through a Non-
Contaminated Field (Figure 22 Procedure Step 8 is Eliminated) 144

 

Table 26 - Test Group 4 (POSITIVE CONTROL) - Inject Through Varying
Contaminated Fields with no Self-Sealing Septum or Aseptic Technique

(Positive Control) (Steps 1-5 and 8 are Eliminated from Figure 22 Injection
Occurs Through Contaminated Field) 145

 

xi

LIST OF FIGURES

Figure l - Growth in Various Segments of the Medical Device Industry—Obtained

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

From http://www.fda.gov/cdrh/present/fdli05/index_ﬁles/images/image56.png 1
Figure 2 - Traditional Whole-package Microbial Challenge Test 11
Figure 3 - Escherichi coli K-12; ATCC 29181 - SEM Photograph <9
Figure 4 - Test Tray #9
Figure 5 - B. Braun Vented Needle System 60
Figure 6 - Racking System Inside Aerosol Cabinet 63
Figure 7 - Mounted Canisters on Aerosol Cabinet 63
Figure 8 — Cabinet Test Conﬁguration - Top View 66
Figure 9 - 102 Concentration 36 Hour Growth 70
Figure 10 - 104 Concentration 36 Hour Growth 70
Figure 11 - 10‘ Concentration 36 Hour Growth 71
Figure 12 - Compilation of Growth Curves for 3 Microbe Concentrations .............. 71
Figure 13 - Petri Dish Exhibiting Microbial Growth 74
Figure 14 - Jig Designed for Uniform Introduction of Media

to Test Trays 84
Figure 15 - Top View of Test Trays Loaded in Jig 84
Figure 16 - Bottom View of Trays Loaded in Jig 85
Figure 17 - Tray Exhibiting Growth on Quebec Colony Counter 88
Figure 18 - B-D Enterotuben'll - Identiﬁcation System for Enterobacten'aceae ...... 90
Figure 19 - Petri Dish from a Trial Exhibiting Growth 1 15
Figure 20 - Test Tray Exhibiting Microbial Growth, Trial Three 116

 

Figure 21 - Enterotube'm II Results 118

 

xii

Figure 22 — Proposed Whole Package Microbial Challenge Test through
Aseptic Introduction of Sterile Growth Medium 134

 

Figure 23 - Uncoated PETG Tray Provided by Perfecseal Sealed with 10733
Heat Seal Coated Tyvekc’ Provided by Amcor Flexibles 135

 

Figure 24 - Tray Exhibiting Growth as Seen Using the Quebec Colony Counter ...l4l

Figure 25 - Logistic Regression Estimates for Groups 1, 2, and 4
(microbial concentrations for Group 3 did not vary) 147

 

xiii

CHAPTER 1

INTRODUCTION

The Legislation and Regulation of Medical Devices

Medical devices range from simple items, like tongue depressors and bedpans, to
complex pieces of equipment, such as programmable pacemakers with micro-chip
technology. A device is deﬁned by section 201(h) of the Federal Food, Drug, and

Cosmetic Act (FD&C Act) as:

"an instrument, apparatus, implement, machine, contrivance, implant, in vitro reagent

or other similar or related article, including a component part, or accessory which is:

. in the official National Formulary, or the United States Pharmacopoeia, or any

supplement to them,

. intended for use in the diagnosis of disease or other conditions, or in the cure,

mitigation, treatment, or prevention of disease, in man or other animals, or,

. intended to affect the structure or any function of the body of man or other
animals, and which does not achieve any of it's primary intended purposes
through chemical action within or on the body of man or other animals and which
is not dependent upon being metabolized for the achievement of any of its
primary intended purposes." (U .8. Food and Drug Administration. Federal Food

Drug and Cosmetic Act (F D&C), 193 8).

If a product is labeled, promoted, or used in a manner that meets this deﬁnition, it will

be regulated by the Food and Drug Administration (FDA) as a medical device. As such,

it is subject to regulatory control. The level of regulatory control is dependent on the
device’s inherent risk, and its intended use and indications for use

(http://www.fda.gov/Cdrh/dexadVice/3 l 3 .hlml, 2005).

 

The level of scrutiny employed by the FDA falls into three broad regulatory classes:
Class I devices represent the lowest level of patient risk; these devices require only the
general controls dictated by the FD&C Act of 1938. Classes 11 and III require not only
the general controls, but also an enhanced level of regulatory scrutiny dictated by the
1976 Medical Device Amendment (Code of Federal Regulations, Title 21, Volume 8,

Part 814-Pre-Market Approval of Medical Devices Amendment, 1976) (See Table 1)

 

 

 

 

 

 

 

 

 

 

 

 

(Rosen, 2004).
Table 1- Classiﬁcation of Medical Devices (Rosen, 2004)
‘ 0
Devrce Level Of Control Risk Level A Of: All Examples
Class Devrces
Dental Floss
0 a
Class I General Controls Low 46 A) Arm Slings
General Controls and o Facial Implants,
Class H Special Controls Moderate 47 /° Catheters
Class General Controls and Moderate to 7V Pace Makers,
III Premarket Approval High ° Vascular Grafts
The State of the Industry

The medical device industry, while considerably smaller in sales dollars than its
pharmaceutical counterpart, is a growth market. In 2004, sales were reported at $100
billion and grew at a 9 percent rate per year, while the pharmaceutical industry had
annual sales of $500 billion and grew at a rate of 7 percent (Rosen, 2004;

Trombetta, 2005). Growth has been the hallmark of many major segments of the medical

device industry (Figure l).

16000
14000

12000 Ophtthnc
10000 Beuomemcm
I X-Ray

8000 Denun

Surgmal
lnstrunients
4000 Dmgnosncs

6000

2000

0
1998 1999 2000 2001 2002 2003 2004

 

Figure l - Growth in Various Segments of the Medical Device Industry — Obtained
From http://www.fda.govlcdrh/present/fdlioS/index_ﬁles/images/image56.png

The following are key factors driving the growth of the industry (Rosen, 2004):
l. ‘The accelerated aging of the United States, European, and Japanese populations
(particularly age groups over 65 years).’
2. ‘The accelerated integration of drugs and devices.’
3. ‘The interface of nanotechnology and medical device industry.’ This facilitates
the reduction in size of many devices.
4. ‘The expanded use of diagnostics to customize therapeutic regimens on a patient
level and to aid in the prediction of disease onset’.
Issues of Health and Package Integrity
Producing medical devices that maintain sterility until they reach the end-user is a
persistent challenge for industry. According to ﬁgures compiled by the FDA, packaging
accounted for 42.5 percent of sterility-related recalls in 1986, and 40 percent of sterility-
related recalls in 1987 (Bryant, 1988). These ﬁgures prompted additional FDA scrutiny

of manufacturing processes and resulted in the implementation of additional controls

(Food and Drug Administration, Center for Devices and Radiologic Health, QSR Manual,
Section 820.70). While these measures have resulted in a reduction in the number of
recalls due to packaging sterility (Bryant, 1988), debates regarding what size of defect
allows microbes to penetrate the package and result in a breach of sterility remain. With
the risks of failure being potential loss of human life and staggering negative economic
impacts, these issues will likely remain very visible.

Consequently, those working in the medical industry face a daily balancing act
between patient safety and cost competitiveness. Patient safety comes under increasing
scrutiny as the frequency of nosocomial (hospital acquired) infections rise. The Centers
for Disease Control and Prevention (CDC) estimate that approximately 90,000 to 100,000
patients die annually as the result of nosocomial infections. This represents the 4til
leading cause of death in the United States according to Littell (2004), and the cost to
treat the nearly 2 million occurrences is estimated to be $11 billion per year (Schierholz
and Beuth, 2001) (Littell, 2004). Such statistics have resulted in programs aimed at
reducing infection rates and brought everything ﬁom the genetics of microbes to common
hygiene practices under scrutiny (Centers for Disease Control and Prevention [CDC],
2002). Adding to the concern surrounding infections is the indication that microbes are
becoming more resistant to both processes and drugs (Berens, 2002; Ehrman, 1998;
Rubin et al., 1999) that are intended to kill them.

Dr. Michael Burday, PhD, Director of the Division of Clinical Microbiology at the
University of Medicine and Dentistry of New Jersey (UMDNJ)-University Hospital,
states that the ‘rate of infection is likely underreported due to the difﬁculty in tracking

patients and reduced hospital stays.’ Additionally, the types of organisms involved and

the causes of these infections are numerous, further complicating the tracking of
nosocomial infections (Littell, 2004).

While the number of sterile medical devices that have become contaminated prior to
use is unknown, it is known that devices contaminated with bacteria have the potential to
cause harm (Littell, 2004). The more invasive a device, or the higher its regulatory
classiﬁcation (Table 1), the greater the risk. This fact was borne out in a study published
in 1978 that estimated that 45 percent of all nosocomial infections were due to implant
devices (Stamm, 1978). These increased rates of infection are likely due to the fact that
invasive devices impair the patient’s defense mechanisms and, when contaminated, can
result in resistant chronic infection or tissue necrosis (Schierhola and Beuth, 2001).
Package Integrity Testing

As a result of these realities, package integrity or the ability of the package to prevent
microbial penetration, especially when medical devices are involved, is increasingly
important. Device manufacturers must package devices so they withstand “the rigors of
shipping and arrive at the point of use in a safe and functional condition”
(Spitzley, 2002). In addition, they must employ testing protocols that provide evidence
that package integrity will be maintained from the sterilization process until delivery to
the patient.

Current tests tend to be destructive and, therefore, costly. Only a fraction of the
packages manufactured can be analyzed when destructive techniques are employed.
Innovative integrity tests are needed to allow for increased sample sizes without

increased destruction of packages.

New tests are emerging in response to this need. Various techniques are employed in
the non-destructive detection of package defects, including: trace gas sensing,
pressure/vacuum decay, mass-ﬂow, and acoustic micro imaging. The greatly improved
sensitivities of these new techniques (one system produced by Alcatel is touted to detect
hole sizes of 100 A [Leventon, 2001]) concern some in the device industry
(Leventon, 2001; Spitzley, 2002). “As device packagers consider the alternatives, how
do they know how much sensitivity is enough and how much is overkill?” (Leventon,
2001)

The required testing sensitivity level, in terms of the size of defect, that results in
microbial contamination has been examined many times. The published limits ﬂuctuate
widely and, as a result, medical device manufacturers’ current test protocols also vary
greatly in terms of acceptance criteria (Gilliland et al, 2004). Without a pressure
differential, published values for the sensitivity level vary ﬁ'om “between 0.2 micron to
0.4 microns” (Howard et al, 1980) to suggesting that it is likely that the critical leak size
exceeds 10 microns (Lampi, 1980). Others indicate that the critical value is “less than
10 microns” (Blakistone et al., 1996; Harper et al., 1995; Hurme et al., 1997; Keller et al.,
1996 and Keller, 1998), “10 or 5 microns” (Chen et al., 1991), “considerably larger than
1 micron” (Lake et al, 1985), 1 micron “under certain conditions” (McEldowney et al.,
1990; Placencia et al., 1986), 11 microns (Lampi, 1980), all the way up to 22 microns

(Gilchrist et al., 1989).

It is important to understand the variations in experimental methods that could help
explain the observed ranges of values reported in the literature. These include:
0 duration of time that the package is exposed to the microbes
' Package geometry
0 porosity of package components
0 thickness of package components
0 rigidity of the package at test
0 method of microbial challenge (immersion, aerosol, talc and dynamic, or static)
0 pressure differential across the sterile barrier
0 types, sizes, states, and concentrations of the challenge organisms
0 types of package defects and the methods of generating, and measuring, the defects.
Basis for Developing a New Method

The research conducted and presented in this dissertation provides answers to one
speciﬁc aspect of the question posed above by Leventon (2001), regarding the level of
sensitivity required for package integrity tests. That speciﬁc aspect is the determination
of ‘what effect does pressure differential, on the inside and outside of the package, have
on microbial penetration of a rigid, non-porous tray intended to contain a sterile medical
device?’ The speciﬁc pressure differential examined simulates the descent of an aircraft
from 8,000 feet, or a ground shipment descending from 8,000 feet. Two experiments
were conducted to answer this question. Each of the experiments was conducted using a
tray typical for the device industry.

It is understood that pressure differential is only one of a variety of factors that have

the potential to inﬂuence microbial penetration, but it is believed that keeping the design

of the experiment as simple as possible is the best approach, and provides useful
information to build future research.

The value of this research is great. One of the major reasons for packaging-related
recalls in the device industry is listed by the FDA as, “Defective packaging: Potential for
breach in sterility” (Food and Drug Administration, Center for Devices and Radiologic
Health, QSR Manual, Section 820.70).

“Potential” becomes the operative term. If a manufacturer ﬁnds a defect, they
remove the packaged product from distribution. It is removed, not because it has a
known breach of integrity, but because there is a potential for it. As a safe, but costly,
approach, when a defect is detected, the entire lot is likely to be pulled.

These types of recalls have many root causes including, but not limited to, machinery
issues, material concerns, and the rigors of distribution. Little is known about precisely
what size defect will allow for a breach of sterility, and what size poses no threat to the
patient. As discussed earlier, there is a great deal of variation in results ﬁom previous

research that contribute to confusion in the indusu'y.

The ultimate outcomes of the body of work, which includes this research and research
to follow, are to enhance patient safety and to reduce packaging costs. Answers
regarding “what effect does pressure differential have on microbial penetration of a rigid,
non-porous tray intended to contain a sterile medical device” help to ensure that:

0 patient safety is maximized

o informed decisions can be made in the event of a potential recall situation

0 a sensitivity benchmark is established for integrity testers

0 costs are minimized (packages and products are not thrown out unnecessarily)

The maintenance of a sterile barrier, or package integrity, for packages that contain
medical devices is paramount because package integrity is of critical importance to
patient well-being. Package integrity is deﬁned in ASTM F1327-98 as ‘the physical
capability of a given package to protect its contents with the desired level of protection
over a deﬁned period of service; for example, as a barrier to physical, microbiological, or
chemical challenges’. ISO 11607-1 deﬁnes package integrity as the ‘unimpaired physical
condition of a ﬁnal package.’

Package integrity in the device industry is generally tested in one of two ways:
physical tests and microbial challenge tests, also called biotests.

Physical tests typically look for defects in the package, which may permit the passage
of microorganisms. Physical tests include: visual inspection, destructive and non-
destructive (ASTM F1886), dye penetration, destructive (ASTM F1929), bubble
immersion, destructive (ASTM D3078, ASTM D4991, ASTM F2096), trace gas sensing

techniques, non-destructive (ASTM F2391) for helium and (ASTM F2228) for carbon

dioxide, pressure decay, non-destructive (ASTM F2095) and vacuum decay, non-
destructive (ASTM F2338), and acoustic microimaging, non-destructive.

Microbial challenges are based on the principle that ‘a leaky container will allow
bacteria to contaminate the container and its contents’ (Marcy, 1992). Although there is a
standard for challenging packaging materials with biological organisms, (ASTM F1608-
00 ‘Standard Test Method for Microbial Ranking of Porous Packaging Materials
[Exposure Chamber Method])’, no ASTM standard currently exists for challenging an
entire package with microbes.

As previously discussed, medical device packages must “withstand the rigors of
shipping and arrive at the point of use in a safe and functional condition”
(Spitzley, 2002). Professionals in the device industry must demonstrate, with a very high
degree of conﬁdence using objective evidence, that the integrity of medical device
packages will be maintained throughout storage, handling, and distribution. This process
is referred to as package validation (Global Harmonization Task Force, Process
Validation Guidance, 1999).

In the past, validation typically utilized a whole package microbial challenge test that
subjected the device package to standardized tests that simulated distribution and then
challenged the packages with microbes via talc, aerosol, or immersion methods. After
these steps, the outside of the package was disinfected thoroughly, and the package was
opened and swabbed on the inside after which time the swab was cultured, incubated and
examined for growth (Figure 2). However, this approach drew much criticism from the

industry (Hansen, Jones et al., 1993; Spitzley, 1994; Scholla, Sinclair et al., 1995).

10

Step 1 Step 2 Step 3 Step 4 Step 5

   

   

   

, Sealed, sterlized Outside of trays are
59"“- §°""z°° trays are thoroughly disinfected Packages are Swarm are culued,
trays are abjected ”0".”de with (mumpme opened and the Incubded and
to starliardlzed microbes (typically dishfacﬂon can re 5m conte rts are examined for growth
teststhd slmrlate aerosol ortat) In {9139 posmes) swabbed
sh'pplng am
handing

Figure 2 - Traditional Whole-package Microbial Challenge Test

This criticism was brought to light in a study conducted by the Health Industry
Manufacturers Association’s (HIMA) Sterile Package Working Group (Hansen, Jones et
al., 1995). The study compared whole package microbial challenge testing to two
physical tests in their ability to detect channels in the seals of packages. The two tests
that HIMA compared with the whole package microbial challenge test were dye
penetration and visual inspection.

When conducting the dye penetration test per ASTM F1929-98 (2004), a solution
consisting of a wetting agent and an indicating dye is injected into the product-ﬁlled
package. The orientation of the package is adjusted so that the dye is allowed to contact
the seal edge for a minimum of 5 seconds and a maximum of 20 seconds. This is a
simple, easy way to test for the presence of channels, a type of defect found in seals.
Visual inspection is also a standardized test method (ASTM F1886), which can employ

light and magniﬁcation in order to examine packages for defects.

The HIMA group rejected the null hypothesis, ﬁnding a signiﬁcant difference existed
between the physical methods and the whole package microbial test’s ability to detect
seal defects in sterile medical device packages. The physical test methods correctly
identiﬁed all samples containing defects, and did not report any false positives. The
whole package microbial challenge test “resulted in correctly positive ﬁndings [sic] of the
test organism in only two instances. Equally important, the microbial challenge test also
resulted in a false positive in one control sample and positive ﬁndings of organisms other
than B. subtilis [the test organism] in four other instances” (Hansen, Jones et al., 1995).

The HIMA group criticized whole package microbial testing for a number of reasons.
They indicated that the methods used in whole package microbial challenge testing were
not reproducible or reliable, and that sophisticated personnel and facilities were needed to
conduct testing of this type. They also indicated concerns about inadequate disinfection
of the package after it had been exposed to microbes (Figure 2, Step 3). The group
stated,

“. . .the exterior of the package should be disinfected to reduce the probability

of obtaining a false positive dining sterility testing. What is an appropriate

level of disinfection? Should a physical or chemical disinfection process be

used? If a physical method, such as ultraviolet light is chosen, are the

packaging materials UV transparent? The use of UV light may actually

reduce the microbial population inside the package if the ﬁlm component

allows UV transmission. If a chemical disinfectant is chosen, does the liquid

permeate the porous packaging material and does the material have sufﬁcient

12

wet strength to withstand the chemical disinfection?”

(Hansen, Jones et al., 1995)

The group also took issue with Steps 4 and 5 of the process (Figure 2, Steps 4 and 5):

“. . .microorganisms that have entered the package must be recovered and
cultured. If the surface area of the device is relatively small compared to
the surface area inside the package, should additional techniques be used
to improve recovery of microbes that have entered the package?”
(Hansen, Jones et al., 1995)

Experiment One, New Method Development - Can we aseptically ﬁll packages with a
microbial growth medium without introducing microbes?

Experiment One is an alternative to the traditional method evaluated by the HIMA
group, to more effectively measure the microbial penetration into packages. In this study,
sealed, sterile device packages were aseptically ﬁlled with a known volume of sterile
agar, exposed to an aerosol of microbes and then incubated to see if any microbial growth
takes place. This reduced the probability of false positives, as packages are not opened to
detect the entry of microbial cells. This technique also eliminated the need to disinfect
the outside of the packages, another of the criticisms of the HIMA group.

In order to employ this technique to examine the impact of pressure differential on
microbial penetration, described later in this dissertation, the new method needed to be
veriﬁed. It had to be determined that the ﬁlling process did not signiﬁcantly inﬂuence
microbe penetration; i.e. that there was no signiﬁcant difference between the sterility
level of the agar and the sterility level of the agar-ﬁlled packages. (Reference

Appendix I, New Method Development).

13

Primary Research- Applying the New Methodology

Experiment Two, Primary Research - What effect does pressure differential have on
microbial penetration of a rigid, non-porous tray intended to contain a sterile medical
device?

During the normal course of handling, packages experience very small pressure
differentials across the sterile barrier that have the potential to act as a “driving force” for
microbial penetration. It has been suggested that for microbial penetration to occur, a
driving force must be present across the sterile barrier (Hackett, 2001). These driving
forces are pressure differences that result ﬁ'om changes in the environment surrounding
the package and are typically VERY small; for example, a calculation using Bernoulli’s
lawl suggests that a package being moved at a rate of 3 feet per second through air
experiences a pressure differential of 0.0000728 psi.

Microbial penetration is likely to occur even in a static environment, due in large part
to pure diffusion (Brownian motion). Additionally, events in distribution, like
descending from a mountain range or the ascent and descent of an airplane induce larger
pressure differentials. In an article presented at HealthPack in March of 2001, Earl
Hackett, a former Research Associate at Dupont, indicated the importance of the issue,
“the ﬂow of gas (and the particles it contains) into a package may be the most serious
threat to the sterility of a packaged device” (Hackett, 2001).

How does the magnitude of the pressure differential inﬂuence microbial penetration

into a package? It is tempting to say the larger the differential, the greater the microbial

 

2
I p = :v_ where p = pressure, y = density of air = 0.075 lb/ft3 at standard temperature and pressure
8

g=aecelerationduetogravity=32.2 ft/sec’v=windspeedinft/sec

l4

ingress. However, this neglects the possible effects of ﬂow ﬁltration theory, which have
been shown to apply to porous membranes (Scholla et al., 2000). This theory states that a
“moderate” ﬂow rate maximizes microbial penetration. On the low side, the low ﬂow
rates that accompany diffusion is theorized to be the predominant mechanism of
penetration. Under high ﬂow rates, impaction predominates, and the microbes are
removed because they are “smashed” into porous ﬁbers. According to this theory,
maximum penetration will occur when neither impaction nor diffusion is the predominant
mechanism.

Theoretically, the ﬂow of a gas through a cylindrical defect in a package under a
negative pressure differential can be calculated using the equation Q=AAP\I(RT/21rM)
where: R=the molar gas constant, T=absolute Temperature, M=gas molecular weight,
A=area of the oriﬁce, and AP=pressure differential. This ﬂow equation has been shown
to be valid for pressure differentials down to 1 psi (Hackett et al., 2000). Research
performed by Earl Hackett (Hackett et al., 2000) explored the effect of pressure
differentials on microbial penetration into porous structures and demonstrated that greater
permeability of the packaging material imparts a greater tolerance of package defects
because the more permeable barrier material will allow more of the exchanged air to
move through the material instead of forcing it through the defect. As a result of this
ﬁnding, this research examines ﬂow through a cylindrical oriﬁce in a package composed
entirely of non-porous material, a more severe and therefore, more conservative test of
microbial penetration in light of Hackett et al’s (2000) ﬁndings.

This research seeks to determine the need for a driving force other than that

introduced by gravity or molecular motion for microbial penetration into a non-porous

15

package. Factors that have been held constant include: the orientation of the defect
relative to the falling particles, the temperature and humidity of the test environment, the
aerosol delivering the microbes, the test organism concentration (0 CFU/mL and
106 CFU/mL), and the exposure time of the packages to the aerosol. Thermal laser
drilled holes in the test trays of 10 micron and 100 micron were compared to trays that
did not have holes.

“A major cause of pressure differential during transport is the ascent and descent of
transport aircraft” (Hackett, 2001). The descent phase results in a negative pressure
differential on the packaging system, with the outside air exerting pressure on the surface
of the package, and theoretically facilitating the penetration of outside contaminants into
any defects in the packaging system. This research focused on the potential ingress of
microorganisms as a result of a negative pressure differential. Test groups of sterile trays
were exposed to an induced pressure differential simulating a package exposure when
descending in an aircraft or ground shipment ﬁ'om 8,000 feet. Cargo and commercial
aircraft have a pressurized cargo hold, usually maintaining pressures of 8,000 feet. An
identical set of test groups were not exposed to an induced pressure differential,

therefore, gravity served as the driving force.

16

CHAPTER 2
LITERATURE REVIEW

Historical Perspective on Testing

Whole package microbial challenge testing was ﬁrst examined by Szczeblowski and
Rubinate in 1965. They submerged pouches carrying 560 micron holes generated with
hypodermic needles in a bacterial bath. Only half of the test pouches exhibited detectable
microbial penetration. They concluded that some other action was needed in addition to
submersion (Szczeblowski and Rubinate, 1968).

Based, in part, on Szczeblowski and Rubinate’s results, Maunder developed a
biotester that subjected pouches to a ﬂex action while submerged (Maunder et al, 1968).

D. T. Maunder, J. F. Folinazzo, and J. J. Killoran - Continental Can Company
and U.S. Anny Natick Laboratories, I968

Maunder’s et a1. (1968) simulated the handling abuse that a ﬂexible food pouch
might undergo when carried by a soldier. The handling was simulated by applying
mechanical pressure to two areas of the pouch, which induced ﬂexing. The
researchers believed that the ﬂexing would not only simulate handling but would
‘remove minute particles of product that might be blocking a leak.’ The test pouches
contained radiation-sterilized chicken-ala—king. This product was selected because of
its semi-ﬂuid consistency, which the researchers believed would contribute more to
pouch leakage than a solid food. The product also provided a good growth medium
for the test organism, Aerobacter aerogenes, a gas-producing biotest organism.
Organism concentration which the pouches were submerged in was reported to be in

‘excess of 20 x 106 cells/ml.’

l7

Defects were created in four ounce retort pouches using a sewing needle. The
holes were placed a half inch from the top seal midway between side seals. Defect
sizes ranged from 33 to 160 microns; the method of veriﬁcation was not included.

Custom equipment was constructed which would allow application of mechanical
pressure to the pouches, resulting in ﬂexing while the pouches were immersed in
water containing the test organisms at the aforementioned concentration. Biotesting
lasted 15 minutes; this allowed for approximately 30 complete ﬂexing cycles. The
pouches were then removed and incubated for 3 weeks at 30°C. Approximately
10 percent of the pouches with holes did not show contamination. A microscopic
examination indicated that, frequently, food product blocked the defects.

This team concluded that a biotest should be included in any package integrity
program.

M. J. M Michels and B .L. Sclmun - Unilever Research Duiven, I 978

The main objective of Michel’s and Schrarn’s (1978) research was to determine
the effect of handling procedures on contamination rates of pouches and not to
determine the minimum hole size which allows microbial penetration. However, they
did draw conclusions regarding hole size in relation to contamination rates. Their
work was performed with 130 x 170 mm multi-layered pouches composed of a
12 micron polyethylene terephthalate (outer layer), 9 micron aluminum (middle
layer), and 70 micron polyethylene (inner layer). The pouches were ﬁlled with
150 ml of yeast extract-peptone broth and sealed. The pouches were then heat

processed with steam with an F0 value between 10 and 12.

18

Package integrity was evaluated by the use of immersion testing in a medium
contaminated with Enterobacter aerogenes. The culture consisted of a concentration
of greater than 108 cells/ml, and testing was performed at 30°C with an exposure time
of 30 minutes. After pouches were exposed, they were incubated for a week or more,
examined for swelling, and pH was determined to identify any pouches that
experienced contamination. Each test group contained 10 pouches that were
purposely punctm'ed with a 100 micron needle. These pouches were intended to be
positive controls to ensure the efﬁcacy of the biological test. There was no
characterization of the resultant hole made by the needle.

The sampling procedure included three post processing handling steps, each with
2 replicates, a total of 9 tests, with 1,180 deliberately punctured pouches. One
hundred of these punctured pouches per handling procedure (300 pouches) served as
positive controls and were checked for leakage with the biotest.

The positive control group (300 pouches) exhibited 100 percent contamination.
Pouches from the worst case group (handling while wet) exhibited a 90 percent
contamination rate. The contamination results dropped signiﬁcantly as the handling
procedru'es improved; drying after manually handling wet resulted in 10 percent
contamination rates, and the group which was dried ﬁrst and then handled dry only
had a 1 percent contamination rate.

In this research, while the biotest was used only to test controls in the study, it
illustrates that under worst case conditions, a ‘large’ induced defect in the pouch
submerged in a high concentration of motile microbes may result in 100%

contamination.

l9

P. J. Anema and B. L. Schram - Unilever Research Duiven, 1979

Anema and Schram (1979) repeated the biotest preformed by Michels and Schram
(1978), but instead of using a retort package, Anema and Schram used semi-rigid
packages.

The packages were made by deep drawing a laminate material of approximately
100 micron aluminum with a polyolefm layer on the inside and a lacquer layer on the
outside. The containers were closed horizontally, and the seal width varied between
1.5 and 4 mm.

The results revealed a very small amount of defective seals (0.3 percent to
0.8 percent). While the extent of the defect that resulted in the contamination of the
tray was not reported, the study does indicate the viability of the biotest method for
detecting seal defects in semi-rigid packages.

Robert R. Reich - Etlwx Corporation, 1985

Reich’s (1985) evaluated the microbial barrier properties of intact medical device
packages exposed to a microbial aerosol test.

Ten devices from three lots of product were selected for the study. Each sealed
package had been ethylene oxide (EtO) sterilized at least 30 months prior to the
study. The devices were packaged in 0.020” clear polyvinyl chloride (PVC)
thermoformed trays heat sealed with a coated 107313 Tyvek® lid.

A tesla coil was used to place a known number and size of holes in the Tyvek®
lid. The voltage associated with the coil was used to control the size of the hole

created. The holes were examined via scanning electron microscopy (SEM) for

20

homogeneity of size and shape. The positive controls in the study contained one hole
of approximately 40 to 50 micron per 20 cm2 of surface area.

The test organism used was Bacillus subtilis ATCC 93 72. The spores were
grown, harvested, cleaned, and incubated at 35+/- 2°C. The ‘theoretical microbe
chamber concentration was 40,000 per cubic meter.’

A Plexiglass® chamber was used for the test. The aerosol was formed using a
nebulizer that produced particles of 0.3 to 2.0 mm, with some above 2 mm. A
circulating fan was used to ensure distribution of the aerosol.

The exposure duration was 30 minutes. Following aerosol exposure, the product
was exposed to 2 minutes of UV. The product was then aseptically transferred to
120 ml of Tryptic Soy Broth and incubated for 7 days at 35 +/- 2°C. During this
period, the packages were observed for evidence of growth of the indicator organism.

The positive control trays (with created holes) demonstrated growth of the
indicator organism. Under the conditions of the test, microbial penetration occurred
in holes of approximately 40 to 50 microns. The results of the study indicated that
the aerosol test method was a viable one and that the sterile medical device packaging
was adequate to maintain a sterile barrier for the contained product.

Ana M. Placencia, Gordon S. Oxborrow, and James T. Peeler-I 986

Placencia et al. (1986) illustrated the efficacy of an aerosol chamber method to
test packaging materials. The PlexiglassO chamber used measured 19.1 cm wide x
21.0 cm high x 31.8 cm long. The challenge organism used was Bacillus subtilis var.
niger in suspension at a concentration of 104 cells/ml. The organism was dispersed

through the chamber with a nebulizer and a fan. The microbial exposm‘e period was

21

15 minutes. Particle count was determined by drawing the aerosol through ﬁlter
paper for a given period of time and counting the number of spores.

To determine the sensitivity of the system, 50 mm polycarbonate membrane
ﬁlters with pore sizes of 0.8, 1, 2, 5 and 12 microns were used. There were 3 tests per
pore size using concentrations of the test organism from 101 to 105 cells/ml. These
data were also used to determine the probability of detection. The sensitivity
evaluation study also included the challenge of heat seal coated 1073B Tyvek® of an
area of 15.21 cm2 containing 3 categories of microholes (0, 1, and 5 per area).
Microhole diameter averaged 67 microns. Nine samples per category of hole quantity
and microbial concentration (101 to 105 cells/ml) were evaluated.

The testing to evaluate the biobarrier properties of packaging materials included

ﬁve different types of materials, each with an area of 15.21 cm2 (Table 2).

 

Table 2 - Materials Examined by Placencia et al (1986)

 

45 lb autoclavable paper

 

44 lb machine ﬁnished surgical grade paper

 

42 lb blue surgical kraft paper

 

Uncoated 1073B Tyvek ®

 

Coated 1073B Tyvek ®

 

 

 

Each material contained three categories of microholes (0, 1, and 5 per area). The
average microhole varied in diameter from 47 to 67 microns. There were 25 samples
per category of microholes. The materials were exposed to ‘8.55 spores per square

centimeter at an airﬂow rate of 2,831.7 ml/min for 15 minutes.’ The airﬂow rate is

22

based on a pressrue differential of 8.86 ian (4.35 psi). The ﬂow rate was based on
the assumption that packaging materials are exposed to altitudes up to 3000 meters
during ground transportation. A subset of the test materials and hole categories were
evaluated at different ﬂow rates; 1283 ml/min and 2831 ml/min.

The pore detectability study indicated that the method’s detectability increases
with the increase in microbial challenge concentration or the increase in pore size.
The method detected a 1 micron pore at a concentration of 103 cells/ml in the
polycarbonate membrane ﬁlters. The challenge detectability study with coated
Tyvek”) supported these results; the percent detection increased as the microbial
concentration increased. The method detected differences among the materials. A
signiﬁcant difference was observed among materials and number of holes and
between materials and number of holes. Uncoated TyvekQ was the most effective
biobarrier. The ﬂow rate evaluation indicated that as ﬂow rates increase so does
microbial penetration.

Robert R. Reich - Ethox Corporation, 1 986

Reich (1986) repeated his 1985 chamber method to assess the biobarrier
properties of packaging materials. This was analogous to the work conducted by
Placencia et al, 1986. Reich (1986) employed a challenge condition of 5 x 104
microbial concentration of Bacillus subtilis spores at a pressure differential of 2 ian
for an exposure time of 30 minutes. This was considered an extreme test of the

materials in question.

23

Reich concluded, as he did for the whole package challenge, in concurrence with
Placencia et al.’s (1986) ﬁndings, that the aerosol chamber method is a viable method
to evaluate the biobarrier properties of packaging materials.

P. Bankes and ME Stringer- The Conrpden Food Preservation Research
Association, Chipping Canrpdcn, UK, 1988

Bankes et al. (1088) developed a model system to investigate physical factors that
affect container leakage. Three types of organism were used as challenge organisms.
They were Pseudomonas sp, Bacillus sp and Staphyloccus aureus. All were prepared
into a ﬁnal cell concentration of 1 x 109 cells/ml.

An apparatus was constructed with glassware to simulate a container. A
membrane ﬁlter with pores of l, 5 and 12 microns for simulated transient leakage
channels and 5 microns for simulated permanent leakage channels was used in the
model. Twelve micron diameter ‘8’ shaped channels were created with glass ﬁber
bundles. A partial vacuum was applied to one of two pipettes containing liquid.
Clamping was used on the mechanism to simulate permanent vs transient leaks. The
vacuum range was 0 to 250 mmHg. This was accomplished by withdrawing air from
the system with a syringe.

Researchers concluded that a ‘positive correlation existed between microbial
penetration and pressure differential’; the number of cells transferring through the
5 micron diameter permanent channels was much less than when a vacuum was

applied.

24

James E. Gilchrist, Dhirendro B. Shah, Dove C. Rodlc, and Roger W. Dickerson,
Jr.-Divisions of Microbiology and Food Chemistry and Technology, Food and
Drug Administration, 1989

Gilchrist et al. (1989) compared trace gas testing, dye penetration testing, and
biotesting when ﬂexible pouches were subjected to each of the three tests. Biotesting
was the accepted means of integrity testing at the time of this research; the intention
of the team was to identify a faster, more practical alternative method to biotest
ﬂexible retort pouches containing food products.

The work was conducted with retort pouches that were 0.005” thick with
dimensions of 6.5” x 8.5” and 12” x 14.” Pouches were made of a trilaminate
material consisting of polyethylene, aluminum, and polypropylene; a completely
nonporous packaging system. The pouches were ﬁlled with Tryptic Soy Broth,
sealed and then sterilized in a still retort.

The biotesting was performed by submerging the pouches in a Tryptic Soy Broth
containing a 108 concentration of E. coli 0A811. The pouches remained in the broth
for 2 hours and were agitated every 15 minutes. Following the submersion period,
the pouches were rinsed with tap water and incubated at 37°C for 20 to 24 hours.
Pouches that were contaminated were considered leakers. ‘Leakers exhibited
swelling, and contained a large concentration of carbon dioxide and hydrogen in the
headspace, as well as turbid broth.’

Holes were introduced to the pouches by two methods. The ﬁrst method was
laser beam melting, producing holes from 17 to 81 microns in diameter. These holes
were described as being approximately round with no tears or ﬂaps. The second

method was puncturing the pouch with a 0.00 ” diameter stainless steel wire. The

25

resultant range of holes produced was 22 to 175 microns. These holes were described
as ‘aperatures with tears and flaps.’ In addition to microscopy, the hole sizes were
determined by the rate of helium ﬂow through the hole. The rate of helium ﬂow was
proportional to the square of the hole diameter at that pressure. Upon completion of
the biotesting procedure, the holes were remeasured with microscopy.

The results of the study indicated that the helium test method was more consistent
at identifying small leaks than the biotest or the dye test. The results of the biotest
indicated that the test bacteria passed through holes down to about 22 microns.

S. McEldowney and M. Fletcher-Department of Biology, University of Warwick,
Coventry and Department of Microbiology, Campden Food Preservation Research
Association, Chipping Campden, 1989

McEldowney (1989) investigated the effect of physical and microbiological
factors on container leakage through the development of a model system for container
leakage. The extent of interaction between numerous factors was determined.

Various organisms were used in the study including Acinetobacter calcoaceticus,
Staphylococcus sp., Pseudomonas sp., Bacillus sp, and a Coryneform bacillus. These
organisms were isolated from processed food cans and conveyor belts in a food
processing lab. Staphyloccus aureas was ‘isolated from a nose swab’. Two bacteria,
Enterobacter aerogenes and Enterobacter cloacae, were included in the study as
representative organisms of those used in biotests.

Pure bacterial suspensions were of concentrations of 3-4 x 108, 6-8 x 108, 1-5 x

109, or 2-5 x 10'0 cells/ml. Mixed suspensions were of a total cell concentration of 1-

3 x 109 cells/ml. Monolayer ﬁlms of organisms and bioﬁlrns were also prepared for

test purposes.

26

At a ﬂow rate of 500 microliter/min bacteria was leaked into a test chamber ﬁlled

with a phosphate buffer solution at varying pressures (0 mmHg for 60 minutes, 51,

102, 152, 203, 254 or 305 mmHg) for 60 seconds.

The following factors were then evaluated:

Bacterial Morphology-leakage of the test organisms was evaluated through a
120 micron straight channel.

Leakage Channel Size-leakage of the test organisms was evaluated through
straight leakage pathways with diameters of l, 3, 5, 8, 10 and 12 micron.
Leakage Channel Shape-leakage of the test organisms was evaluated through
15 micron diameter straight channels and 15 micron diameter convoluted
channels.

Content viscosity-the chamber was ﬁlled with phosphate buffer solution
(control) or 4 percent pectin solution. Leakage of test organism was evaluated
with 5 and 10 micron diameter straight leakage channels.

Bacterial Cell Concentration-the pure cell test concentrations were evaluated
through 15 micron diameter straight pathways.

Mixed bacterial populations-the mixed cell suspensions were evaluated
through 5 and 10 micron diameter straight leakage channels.

Monolayers and bioﬁlms-these were evaluated through straight leakage
channels of 10 micron diameter.

Vacuum and exposure duration- at a ﬂow rate of 500 microliter/min bacteria
were leaked into a test chamber ﬁlled with a phosphate buffer solution at 0

mmHg for 60 minutes, 51, 102, 152, 203, 254 or 305 mmHg for 60 seconds.

27

McEldowney (1989) observed that bacterial motility did not have a ‘consistent
effect on the rate of bacterial leakage.’ For example, motile strains such as
Pseudomonas sp did not penetrate at a faster rate than non-motile strains such as
Acinetobacter calcoaceticus without vacuum. The effect of progressive increases in
vacuum ﬁ'om 51 mmHg to 305 mmHg varied with bacterial species and morphology,
but was greater in the presence of partial vacuum. Fluid ﬂow through leakage
channels also increased in the presence of vacuum. Leakage varied with vacuum,
bacterial morphology, cell concentration, leakage channel size, and channel shape.
The number of leaked cells was not proportional to vacuum or channel size. The
effect of channel shape varied with bacterial species. Increased content viscosity
decreased leakage. There was a difference in leak rates against vacuum between the
pure bacterial suspensions and the mixed bacterial suspensions. The results indicated
that there are three major factors inﬂuencing leakage; ﬂuid availability, container
vacuum, and content viscosity.

While these are the dominant factors, they are impacted by the other factors
previously discussed. McEldowney’s research clearly illustrates the complexity of
understanding the container leakage process because of the wide range of

interdependent variables.

28

C. Chen, B. Harte, C. La], J. Pestka, and D. Henyon-Michigan State University,
School of Packaging and Department of Food Science and Human Nutrition,
Pacteco U.S.A, and Pure-Bah, Inc, 1991

Chen et a1 (1991) evaluated the method of aerosol biotesting for determining the
microbial integrity of aseptically ﬁlled juice packages as an alternative to microbial
immersion testing.

A challenge test was run ﬁrst using 30 ml sterile glass vials containing growth
medium to determine the minimum hole size that the organisms could penetrate in a
speciﬁc time period. Standard oriﬁce sizes of 5, 10, and 15 microns were placed in
the vial stoppers. Exposure times of 15, 30, 60, and 90 minutes were used for both
the aerosol and immersion methods. The vials were then cleaned with chlorine
solution and incubated. If turbidity in the growth medium developed after 48 hours, it
was considered a positive result.

The test packages were 1.89 L ﬂat-top cartons aseptically ﬁlled with apple juice.
The cartons were nitrogen ﬂushed resulting in a headspace composition of 97 percent
nitrogen and 3 percent oxygen with dimensions measuring 19.7 cm x 13.3 cm x 7.6
cm (HxWxD). The juice boxes were composed of a lamination of polyethylene
(PE/paper stock/Surlyn‘D/aluminum foil/PE). Standard pinhole oriﬁces were used to
create holes in the cartons. The oriﬁces were 0.00254 cm thick and 0.635 cm in
diameter.

The challenge organism was Lactobacillus cellobiosus (ATCC11739), an

organism that can survive and reproduce at pH values below 4.5. A concentration of

2.5 x 10‘5 cells/ml was used for all the tests.

29

Aerosolization occurred in a custom built Plexiglass" cabinet that contained 32
spray nozzles positioned throughout the cabinet. Packages were placed in the center
of the chamber and the spray pattern was designed to cover the package geometry.
Two exposure times were tested, 15 and 60 minutes. Following the microbe exposure
period, the packages were sprayed with a 2500 ppm chlorine solution for 30 minutes.
The packages were then removed and incubated at 35-3 7°C (ideal conditions for
growth of the challenge organism) for 2 weeks.

The immersion test employed the same microbe and microbial concentration. The
packages were submerged in the microbe suspension for 60 minutes. Aﬁer the 60-
minute period the packages were submerged in a 2500 ppm chlorine solution for 30
minutes, rinsed with tap water and incubated at 35-37°C for 2 weeks.

Detection of carbon dioxide, a product of the microbes in the headspace, indicated
a likely breach in integrity. Additionally, containers were evaluated for a change in
the pH level, and container swelling. Microbial growth in containers exhibiting these
attributes was veriﬁed after the incubation period by removing the juice and
performing standard colony counting.

“Five micron holes were detected aﬁer 30 minutes of spraying and 20 micron
holes were detected aﬁer 15 minutes of spraying.” The percentage of breached
containers increased with increasing hole size. Sixty-two and one-half percent of
containers tested positive for microbes when the 15 micron oriﬁce was subjected to
60 minutes of spraying. Hole size was determined to be statistically signiﬁcant,
which indicated the aerosol method to be capable of differentiating between different

oriﬁce diameters. Test duration was not signiﬁcant. Pinholes of 5 to 15 microns

30

were more easily detected with the aerosol method than the immersion method. The
aerosol methods showed better detectability than the immersion test.

B. Ahvenainen, T. Mattila-Sandholm, L. Axelson, and G. Wirtanen - Technical
Research Center of Finland, Swedish Packaging Research Institute, 1992

The objective of this research was to investigate ‘the effect of microhole size and
the food type on microbial penetration of aseptic plastic cups.’ A driver to conduct
the research was that, prior to 1992, very little research had been done on critical leak
size for plastic containers. The majority of the research had focused on cans and
retort pouches.

The test materials were high impact polystyrene cups with a volume of 115 cubic
centimeters. Microholes were made in the bottom of the cup using a needle punch
technique. These holes had diameters ranging from 5 to 100 microns, as measured by
microscopy. Hole length was reported to be 50 +/- 10 microns. Seventy-ﬁve cm3 of
3 different types of foodstuffs and soﬁ agar ﬁlled the cups. The hole sizes tested with
milk and agar ranged from 5 to 100 microns in diameter; for meat soup from 10 to
100 microns; and for sausage gravy from 25 to 100 microns.

The test cups were submerged in a container ﬁlled with Enterobacter aerogenes
(ATCC 13048) suspension at a concentration of 107 CFU/mL. The chosen test
organism is motile and is a gas producer. The submersion chamber, containing the
cups, was then incubated and shaken for 30 minutes at 30°C. The cups were
removed, wiped, and cleaned, and further incubated at 30°C for l to 6 weeks.
Samples were taken at 1, 3 and 6 weeks and examined for microbial growth.

Following incubation, the cups were carefully washed and examined microscopically

31

to see if the holes were blocked by the food contained in the cups. Bacterial growth
was conﬁrmed by the pour plate method.

Three hundred, sixty-eight cups were tested in total; 132 ﬁlled with soﬁ agar, 72
ﬁlled with milk, 102 ﬁlled with meat soup, and 62 with sausage gravy. The smallest
hole that the test bacteria could penetrate was 5 microns when the cup was ﬁlled with
milk. The smallest hole for soft agar was 10 microns, for meat soup 20 microns, and
sausage gravy 25 microns.

“The rate of penetration increased with lower viscosity product and increased
with hole size.” However, there were instances of 100 micron holes not being
penetrated for all products.

The results illustrate that while bacteria did not penetrate certain samples, it did

not guarantee package integrity. While a critical hole size for plastic containers
cannot be determined ﬁ'om this study, it does indicate that the probability of microbial
penetration is dependant on the product contained within and its viscosity.
Scott W: Keller, Joseph E. Marcy, Barbara A. Blakistone, George H. Lacy,
Cameron R. Hackney, and Walter H. Carter, Jr. - Virginia Polytechnic Institute
and State University, National Food Processors Association, and Medical College
of Virginia, 1995

Keller et al (1995) examined the effects of test organism motility, test organism
concentration, bioaerosol exposure time, hole diameter, and hole length with regard to
their inﬂuence on microbial ingress into a ﬂexible plastic pouch.

The retortable pouches used in the study were constructed of polyester (outer),
Saran“), and cast copolymer polypropylene (inner). Nickel microtubes were used to
create seal defects by sealing them in package heat seals. Four sizes of tubes were

used; 10 and 20 +/- 2 micron internal diameter each with 5 and 10 +/- 1 mm lengths.

32

The pouches were ﬁlled with 99 ml of Tryptic Soy broth and retorted. They were
allowed to thermally equilibrate to 25°C before testing.

Motile and mutated non-motile Pseudomonas ﬁagi TM 849 were used as test
organisms. Two source concentrations were evaluated. “The 102 cells/ml source
yielded a ﬁnal concentration of P. ﬁagi in the Plexiglass® chamber of l CFU/cm3 and
2 CFU/cm3 for exposure times of 15 and 30 minutes, respectively.” The 10° cells/ml
source yielded a ﬁnal concentration of P. ﬁagi in the Plexiglass® chamber of 87
(:FU/cm3 and 178 CFU/cm3 for exposure times of 15 and 30 minutes, respectively.
The inoculated aerosol was generated in the chamber by nebulizers connected to two
air cylinders, equipped with air-ﬂow meters.

Prior to aerosol exposure, the defect area was exposed by cutting the top seam of
the pouch. Four pouches, one of each defect size, were hung on a rack and placed in
the chamber. Eight experimental sets were run for a total of 128 pouches. Each of
the sets contained four combinations of the variables. Four replicates were run.

Six of the 128 pouches exhibited microbial contamination. All of the positive
units were exposed to the motile, higher microbial source concentration of
10° cells/ml. These two variables were found to be statistically signiﬁcant. The
variables of channel length and exposure time were not statistically signiﬁcant, but
‘the interaction between an exposure time of 15 minutes and a 10 micron hole was
statistically signiﬁcant.’

Sixty-six percent of the positive results were in the lower probability group, the
10 micron hole size. It was concluded that the ‘method of aerosolization can be used

to detect microholes previously considered statistically improbable to detect’.

33

Researchers concluded that an aerosol challenge was more representative of
conditions that the package would likely encounter in distribution, than that of
microbial immersion testing, and that it was a viable means of assessing package
integrity.

Carol L. Harper-National Center for Food Safety and Technology/FDA, I995

Harper’s (1995) efforts served to develop guidelines for a new, off-line, non-
destructive package integrity test system. Speciﬁcally, the research team needed to
determine the minimum channel diameter that would allow microbial penetration.

A motile organism, Enterobacter aerogenes, chosen because it exhibits robust
growth, was used at a concentration of l x 109 CFU/mL. Seal defects were created by
sealing tungsten wire, of known diameters, into the seals of a ﬂexible plastic. The
seal and the area around it was then cut out and placed into a ﬁlter holder. The edges
were sealed with silicone to avoid any leakage around the test material. Once the
material was sealed, the apparatus and the lower chamber holding 250 ml of medium
was autoclaved. While the material was still warm, the wire was removed, leaving
channels from 10 to 125 microns in the seal area. The diameters of the channels were
veriﬁed by microscopy. The apparatus was then incubated overnight to ensure
sterility. Solution inoculated with the challenge organism was poured into the top
reservoir of the apparatus. If a sample exhibited contamination, the ‘apparatus was
bubble tested to verify that there were no leaks in the ﬁlter ring seal’.

A growth curve was generated to determine the time needed for one colony to
multiply to a level that would be detectable by plating methods. This time was

determined to be 4.5 hours. After preliminary experimentation, Harper (1995) set the

34

ﬁnal duration for the experiment at 24 hours. The test organism was ‘capable of
penetrating channel leaks of a 10 micron diameter,’ channel length was not reported.
Barbara A. Blakistone, Scott W. Keller, Joseph E. Marcy, George H. Lacy,
Cameron R. Hackney, and Walter H. Carter, Jr-National Food Processors
Association, Department of Food Science and Technology, Virginia Polytechnic
Institute and State University, Medical College of Virginia, Virginia
Commonwealth University, 1995

The purpose of this research was to examine the critical defect threshold
dimensions by using artiﬁcially created channel leaks of 10 and 20 microns across
seals that were 5 and 10 mm in length in plastic pouches.

The pouch material was polyester-polyvinylidene chloride-cast copolymer
polypropylene. Microtubes were positioned in the seals to create the channels.
Another seal was placed above the microtubes to protect them from contamination
until use. The pouches were ﬁlled with 100 ml of Trypticase soy broth and then
retorted.

The test organism, Pseudomonas ﬁ'agi, was prepared at concentrations of 102 and
10° CFU/mL. Both motile and non-motile variation of this organism were evaluated.
The seal on the pouch was cut to expose the channels. The pouches were placed in a
sterile container and immersed in the test organism suspension. The container was
gently rocked every 5 minutes. The exposure duration was 15 and 30 minutes.
Following immersion testing, the media contained in the pouches was cultured to
determine if the test organism had penetrated.

The results indicated 56 of 128 pouches tested (44 percent) were positive.
Analysis showed that channel length had no inﬂuence on the positive results.

Microbe concentration level (102 and 10° CFU/mL) and the interaction with time

35

were found to be statistically signiﬁcant (P<0.05). Exposure time as a sole factor was
not statistically signiﬁcant nor was channel diameter. “Approaching statistical
signiﬁcance was the relationship of concentration and motility versus time.” The
increase of exposure time had a slightly negative effect on contamination with motile
organisms, while not statistically signiﬁcant. Motility appeared to be more of a
positive factor at the lower microbial concentrations than the higher concentrations.

Blakistone et al. (1995) concluded that biotesting can be “misleading if not
designed for the particular exposure conditions of the package.” While Blakistone et
al. (1995) did not draw any conclusions regarding the critical defect threshold, they
did emphasize the need for a “model for risk assessment of packages” under speciﬁc
conditions.

Joseph E. Marcy-Food Science and Technology, Virginia Polytechnic Institute and
State University, I 995

Marcy (1995) investigated the viability of an immersion biotest for ﬂexible retort
pouches.

The pouches used in this study were Meals, Ready to Eat (MRE) retortable
pouches. The laminate consisted of 48-gauge polyethylene terephthalate
(PET')/adhesive/0.00005 inch foil/adhesive/0.003 inch polypropylene (PP). The
pouches contained 8 ounces of product and were 4 3/4 x 8 ‘/2 inches in size.

Channel leakers were manufactured in the seals by placing a 20 micron Teﬂon®
ﬁber in the sea] area prior to scaling. After cooling, the ﬁber was removed from the
pouch. Another set of leakers was produced using platinum wires 100, 130, and 210
microns in diameter. The size of the Teﬂon® ﬁbers was veriﬁed with an optical

microscope and the size of the platinum wires was veriﬁed with a micrometer. The

36

20 micron hole size was veriﬁed microscopically before and after the biotest resulting
in a size of 20 i 1 micron. The larger holes were tested by using an electrolytic test
and correlating the diameter of a leaker to the intensity of the current ﬂow.

Five sets of retortable pouches were manufactured for biotesting; 150 MRE
pouches containing 20 micron channels and three sets of 50 MRE pouches with
channels of 100, 130, and 210 microns. The ﬁfth set did not contain any
manufactured leaks. All pouches were double sealed to protect the channels from
contamination until the biotest. Each pouch contained 80 ml of Tryptic Soy Broth.

The test organism was gas producing E. coli 9637 at a concentration of
4 x 108 cells/ml. The ﬂanges of the test pouches were wiped with the test organism
broth before immersion to ensure adequate wetting of the seal. The pouches
remained immersed in the suspension for 2 hours and agitated every 15 nrinutes.
Following the test, the pouches were rinsed with tap water and incubated at 37°C for
24 to 48 hours. A positive biotest was indicated by swelling of the pouch and
turbidity of the broth. A positive result was conﬁrmed by analyzing the gas content
of the headspace. To ensure the only leak present was the manufactured leak, the
units were dye tested.

The results indicated that 100 percent of the larger sized holes (100, 130, and
210 micron) exhibited contamination and 96 percent of the 20 micron hole pouches
exhibited contamination.

The researchers concluded that 20 micron leaks result in bacterial penetration.

They also concluded that channel leak length does not affect the likelihood of

37

contamination since there was no difference noted with 20 micron channels of 5 mm
and 127 microns in length.

Matthew Joseph Gibney IV- Department of Food Science and Technology, Virginia
Polytechnic Institute and State University, 2000

Employing test methods similar to Keller (1998) and Blakistone (1995), Gibney
compared air-ﬁlled package defects with ﬂuid-ﬁlled defects to determine the critical
leak size for air-ﬁlled defects. Gibney used a bioaerosol chamber to expose the
packages to the test organism. Nickel microtubes were employed to create channel
defects of 2, 5, 7, 10, 20, and 50 microns in diameter. Defect sizes and pressure
differentials were examined for signiﬁcance. Pressure differentials of 0, -6.9, -13.8,
and -34.5 kPa were employed. The test organism used was Pseuduornonas ﬁagi at a
concentration of 10° cells/ml.

Gibney (2000) referred to defects in packages that contained only air as ‘air-ﬁlled
microholcs’. He compared penetration into these holes with penetration into
microholes that were situated against liquid product. Gibney concluded that the
critical leak size of an air-ﬁlled microhole was approximately 7 micron for pressure
differentials ranging from 0 to -34.5 kPa. This differs from previous research with
liquid ﬁlled defects in which the critical defect coincided with the threshold leak size
or the ‘defect size which allows the initiation of leakage at a particular pressure’

(Keller, 1998).

38

ASTM F 1608 — 00 Standard Tat Method for Microbial Ranking of Porous
Packaging Materials (Exposure Chamber Method), 2000

This method is intended to compare the ability of porous packaging materials
intended for use in medical device packaging to ﬁlter airborne bacteria. Samples of
porous materials are subjected to an aerosol of Bacillus subtilis var niger spores at a
challenge concentration of 1 x 10° cells/ml in an exposure chamber. Spores that pass
through the porous material are collected and enumerated.

The materials are exposed to the aerosol for 15 minutes at a ﬂow rate of
2.8 L/min. The materials are removed and the ﬁlter paper cultured and the plates
enumerated.

The resultant data assessed the mile potential of materials to contribute to the
loss of sterility of the product contained in the package. It is not intended to predict
the performance of a material for a speciﬁc application.

Whole package microbial challenge testing was used extensively in the mid-19603
and 19703 during the development of the retort pouch (Lampi, 1981). As retort
pouches began replacing cans for the containment of food products, the assessment of
the bio-barrier changed to incorporate the stresses of package handling. Maunder
tested pouches being ﬂexed as they were exposed to immersion in a bath of microbes
in order to simulate the handling of Meal Ready-to-Eat pouches by soldiers in the
ﬁeld (Maunder et al, 1968).

At the same time that the food industry was testing the integrity of retort pouches,
medical devices were evolving. Prior to the 19703, medical devices were sterilized in
the hospital and re-used. This changed with the advent of the disposable device.

Disposable medical devices were (and are) packaged and sterilized by the device

39

manufacturer, who, in turn, become responsible for the product’s sterility from
manufacture to arrival at the hospital. In parallel with these device advancements, a
proliferation of new packaging developments also occurred. These developments
involved the introduction of ﬂexible materials for use in ﬂexible pouches, and
materials to package devices in thermoformed trays (Spitzley, 1994).

As the industry changed, the need for standardized packaging integrity testing
became increasingly apparent. Publications in the early 19803 (Reich, 1985;
Placencia et al, 1986) indicate that microbial challenge methods were the
recommended method for testing the integrity of device packages.

Validation protocols for medical device packaging typically involved a whole
package microbial challenge test. This was the case as recently as the 1990s. See
Figure 2 for a complete description of the typical procedures of the past. As
discussed, the whole package microbial challenge testing method drew much
criticism ﬁ'om the industry (Hansen et al, 1995; Spitzley 1994; Scholla, Sinclair et al,
1995).

When presented with the results of the work of the HIMA group, the FDA agreed
that physical test methods were ‘superior to microbial challenge tests for sterile
package integrity’ (Spitzley, 2002) and began recognizing both types of tests for the
evaluation of package integrity. The focus of the medical device packaging
community shifted to the enhancement, standardization, and validation of physical
test methods. Current tests (reference Basis for Developing a New Method Section
of the Introduction, Chapter 1) tend to be destructive and, therefore, costly. Only a

fraction of the packages manufactured can be analyzed when destructive techniques

40

are employed. Innovative integrity test methods were needed to allow for increased
sample sizes without increased destruction of packages.

Emerging techniques such as: trace gas sensing, pressure-decay, mass-ﬂow, and
acoustic micro-imaging have greatly improved sensitivities. Acoustic micro-imaging
has been shown to detect sub-micron gaps in test materials (Allen, 2002). The greatly
improved sensitivities of these new techniques concern some in the device industry
(Leventon, 2001; Spitzley, 2002). “As device packagers consider the alternatives [to
traditional integrity tests], how do they know how much sensitivity is enough and
how much is overkill?” (Leventon, 2001) This is an important question to the
industry. A test that passes holes that allow microbial penetration has the potential to
negatively impact patient health. However, tests that fail packages that do not present
health risks needlessly drive up costs when packages and products are destroyed
unnecessarily.

The intuitive response to the question of the appropriate sensitivity levels is that if
the hole is larger than the microbe, there is a potential for penetration. However, this
seemingly simple answer does not hold true. Various researchers have explored the
question, “What is the smallest defect that allows microbial penetration into sterile
packages?”; with widely varying results, as previously discussed (Maunder et al,
1968; Michels et al, 1978; Anema et al, 1979; Reich, 1985; Reich, 1986; Placencia et
al, 1986; Bankes et al, 1988; Gilchrist et al, 1989; McEldowney et al, 1989; Chen et
al, 1991; Ahvenainen et al, 1992; Keller et al, 1995; Harper, 1995; Blakistone et a1,

1995; Marcy et al, 1995; Gibney, 2000).

41

A natural consequence of the numerous factors available for investigation is that
the experimental objectives vary signiﬁcantly, as do the experimental designs that
researchers working in this area employ.

Varying Objectives

Maunder (1968), Marcy (1995), and Michels (1978) investigated the viability of
immersion biotests as an indicator of package integrity for retort pouches, and to examine
the impact of handling on microbial penetration (Maunder, 1968; Marcy, 1995; Michels,
1978). Anema also investigated the use of immersion biotesting, but used rigid trays
(Anema et al, 1979). Gilchrist compared the effectiveness of immersion biotesting with
helium and dye penetration testing on ﬂexible pouches (Gilchrist et a1, 1989).
Ahvenainen used immersion biotesting to evaluate the relationship of different foods
contained in aseptic cups and the microhole size on bacterial contamination (Ahvenainen
et al, 1992). Blakistone’s research was directed at determining the size of microhole,
which resulted in microbial contamination of plastic pouches when immersion tested
(Blakistone et al, 1995). Harper’s intent was to develop guidelines with respect to the
required sensitivity for non—destructive package integrity testing; this work employed
packaging materials, where many of the others involved an evaluation of the entire
package (Harper, 1995).

A team led by Chen evaluated the use of bioaerosol testing as an alternative to
immersion testing for evaluating aseptic juice packs (Chen et al, 1991). Reich (1985,
1986) evaluated the use of aerosol testing for medical device trays and in a separate study
evaluated the use of aerosol testing for packaging materials (Reich, 1985; Reich, 1986).

Placencia also used biotesting and investigated packaging materials as well as whole

42

package testing (Placencia et a1, 1986). Keller and Gibney evaluated numerous physical
and microbiological variables with regard to their effect on microbial penetration and
critical hole size using bioaerosol testing (Keller et al, 1995; Gibney, 2000).

And ﬁnally, Bankes and McEldowney studied various microbiological and physical
variables effecting microbial contamination by the development of model systems and
apparatus, which allowed them to study the variables and their interactions (Bankes et a1,
1985; McEldowney et al, 1989).

Because of the varying objectives of different research teams and the resultant
disparity in each of their experimental designs, it is not surprising that conclusions
regarding threshold vary greatly as well.

Varying Results

Many researchers report the threshold for microbial penetration is less than 10
microns. Howard et al, (1980) indicated that the penetration threshold was 0.2 micron.
Under certain circumstances, 0.2 micron rated ﬁlters allowed passage of certain bacteria
(Howard et al, 1980). Blakistone et al, (1996), Harper et al, (1995), Hurme et al (1997),
Keller et al, (1996) and Keller (1998) all indicate “less than 10 microns.” Chen et al
(1991) indicated it to be “10 and 5 microns.” (Chen et al, 1991) Gibney (2000) reported
7 microns (Gibney, 2000), and Lake et a1. (1985), reported “considerably larger” than 1
micron while both McEldowney (1990) and Placencia reported 1 micron “under certain
conditions” (McEldowney et al, 1990; Placencia et al, 1986).

Other researchers support the idea that the penetration threshold is greater than
10 microns. Lampi (1980) indicated that it was not likely that the critical leak size was

less than 10 microns. Gilchrist et a1 (1989), reported 22 microns, Reich (1985) indicated

43

a range of 40-50 microns and Axelson et a1 (1990), using an electrolytic test, determined
the critical hole size to be 80 microns.

As mentioned previously, variation in experimental objectives and subsequent
experimental design contribute signiﬁcantly to the reported results. Tables 3 and 4

summarize physical variables (Table 3) and microbiological variables (Table 4)

 

 

 

 

 

 

 

employed by the cited teams.
Table 3 - Physical Variables in Published Research
Researcher/ Whole Whole Material Product Defect Creation
Year Pkg, Package Description Contained Method/
Mt’ls or Description Veriﬁcation
Model
Maunder, Package Flexible Trilarninate Chicken- Needle/
1968 Pouch ala-king Microscopic
33 to 160 micron
Michels, Package Flexible Trilaminate Growth Needle/No
1978 Pouch Medium Veriﬁcation
Cited
100 micron
Anema, Package Rigid Tray/ Trilanrinate Growth Needle/No
1979 Flexible Medium Veriﬁcation
Pouch Cited
100 micron
Reich, Package Rigid Tray PVC/Tyvek Device Tesla coil/SEM
1985 40—50 micron
Reich, Material N/A Various N/A Tesla coil/SEM
1986 40-50 micron

 

 

 

 

 

 

 

 

 

Table 3 Leont’d)

 

Researcher/
Year

Whole
Pkg,
Mt’ls or
Model

Whole

Package
Description

Material
Description

Product
Contained

Defect Creation
Method/
Veriﬁed

 

Placencia,
1 986

Material

N/A

Various

N/A

Pores and
Microholes from
47 to 67 micron
diameters.
Microholes
present in
various
packaging
materials/
Microscopic

 

Bankes, l 988

Model

N/A

N/A

N/A

1, 5, 12 micron
pores to simulate
transient leaks
and 5 micron to
simulate
permanent leaks

 

Gilchrist,
1989

Package

Flexible
Pouch

Trilaminate

Growth
Medium

Laser Beam-17
to 81 micron in
diameter

Wire Puncture-
22 to 175 micron
in diameter-
Veriﬁed by
microscopy and
ﬂow rates

 

McEldowney
1 989

Model

N/A

N/A

N/A

Numerous

 

Chen, 1991

Package

Aseptic
Package

Trilaminate

Juice

Standard Pinhole
Oriﬁces

5 to 20
micron/No
Veriﬁcation
Cited

 

 

Ahvenainen,
1 992

 

Package

 

Aseptic
Cups

 

High
Impact
Poly-
styrene

 

Foodstuffs

 

Needle-5 to
100 micron/
Microscopy

 

45

 

 

Table 3 (cont’d)

 

Researcher/
Year

Whole
Pkg.
Mt’ls or
Model

Whole

Package
Description

Material
Description

Product
Contained

Defect Creation
Method/
Veriﬁed

 

Keller, 1995

Package

Flexible
Pouch

Plastic
Trilaminate

Growth
Medium

Microtubes-1 0
and 20 micron
diameter and 5
and 10 mm
length/No
Veriﬁcation
Cited

 

Harper, 1995

Material

N/A

Flexible
Plastic

N/A

Wires in seal-
channels 10-125
micron/Micro-
SCOPY

 

Blakistone,
1995

Package

Flexible
Pouch

Plastic

Growth
Medium

Microtubes-10
and 20 micron
diameter and 5
and 10 mm
length/No
Veriﬁcation
Cited

 

Marcy, 1995

Package

Flexible
Pouch

Trilaminate

Growth
Medium

Fiber and Wires
in Seals-20, 100,
130, and 210
micron/
Microscopy and
Electrolytic Test

 

Gibney, 2000

Package

Flexible
Pouch

Plastic
Trilaminate

Growth
Medium

Microtubes-2, 5,
7, 10, 20 and

50 micron/No
Veriﬁcation
Cited

 

 

ASTM F
1608-00

 

Materials

 

N/A

 

Porous

 

N/A

 

Pores

 

46

 

 

Table 4 - Microbiological Variables in Published Research

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Researcher Exposure Biotest Driving Challenge Organism
Duration Method Force Organism Concentration
Maunder, 15 Immersion Flexing Aerobacter 20 x 10°
1968 minutes aerogenes cells/ml
Michels, 1978 30 Immersion N/A Enterobacter 108 cells/ml
minutes Aerogenes
Anema, 1979 30 Immersion N/A Enterobacter 1.3 x 102 to
minutes Aerogenes 1.3 x 108
and varied cells/ml
Reich, 1985 30 Aerosol N/A Bacillus subtilis 4 x 105
minutes cells/m3
Reich, 1986 30 Aerosol N/A Bacillus subtilis 5 x 104
minutes cells/m3
Placencia, 15 Aerosol Flow Bacillus subtilis 10I to 105
1986 minutes Rate var niger cells/ml
Bankes, 1988 N/A N/A Vacuum Pseudomonas 1 x 10r
sp, Bacillus sp, cells/m1
Staphyloccus
aureus
Gilchrist, 2 hours Immersion Flexing Escherichia coli 1 x 10"
1989 0A811 cells/m1
McEldowney, Numerous N/A Vacuum Acinetobacter Multiple
1989 calcoaceticus
Staphylococcus
sp.,
Pseudomonas
sp., Bacillus sp.,
Coryneform
bacillus
Staphyloccus
aureas Ent.
aerogenes Ent.
Cloacae
Chen, 1991 15 and 60 Aerosol N/A Lactobacillus 2.5 x 10°
minutes and cellobiosus cells/ml
Immersion
Ahvenainen, 30 Immersion Agita- Enterobacter 1 x 107
1992 minutes tion Aerogenes cells/ml
Keller, 1995 15 and 30 Aerosol N/A Pseudomonas 1 x 102 and 1
minutes ﬁagi x 106 cells/ml

 

 

 

 

 

 

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4 (cont’dL
Researcher Exposure Biotest Driving Challenge Organism
Duration Method Force Organism Concentration
Harper, 1995 24 hours Immersion N/A Enterobacter l x 109
Aerogenes cells/ml
Blakistone, 15 and 30 Immersion N/A Pseudomonas 1 x 102 and 1
1995 minutes ﬁagi x 10° cells/n11
Marcy, 1995 2 hours Immersion Agita- Escherichia 4 x 10‘r
tion coli 9637 cells/n11
Gibney, 2000 15 and 30 Aerosol Vacuum Pseudomonas 1 x 10°
minutes ﬁ-agi cells/ml
ASTM F 15 Aerosol Flow Bacillus 1 x 10°
1608-00 minutes Rate subtilis var cells/ml
niger
Variables

These research teams, and those reviewing their work, place varying levels of
emphasis on the many factors that play a role in microbial ingress, largely based on the
objectives of their respective studies. A discussion of the importance of the variables,
based on conclusions drawn from their research follows.

Research has established that a driving force is required for microbial ingress
(Hackett, 2001; Scholla et al, 2000; Floros, 1994; McEldowney et al, 1990; Ahvenainen
et al, 1992; Placencia et al, 1986; Purohit, 1995; Harper, 1995; Maunder et al, 1968;
Tallentire and Sinclair, 2002). Results ﬁom a model system developed by Bankes and
Stringer supports the idea that the amount of leakage with a 5 micron hole varied with the
amount of vacuum present. Their results support ﬁndings by Put et a1 (1972) that a can
with a vacuum of 400 mmHg is more susceptible to leakage than one at atmospheric
pressure (Put et a1, 1972). In other words, a larger differential results in greater
penetration through the package. McEldowney (1990) also developed a model system to

study container leakage. She indicated that the effect of progressive increases in vacuum

48

from 51 mmHg to 305 mmHg varied with bacterial species and morphology
(McEldowney et al, 1990). Gibney’s (2000) research indicated that pressure did not play
a signiﬁcant role in microbial ingress, but that hydraulic diameter effects were signiﬁcant
(Gibney, 2000).

Keller et al (1995), reported factors critical to maintenance of the sterile barrier to be:
defect dimension, organism concentration, and exposure time (Keller et al, 1995).

F loros (1994) discussed the comparison of critical hole size for different types of
packaging systems. He indicated that a comparison cannot be made between cans and
aseptic packages because the headspace for aseptic packaging, which is at atmospheric
pressure, versus that of cans, is below atmospheric pressure (vacuum) impacts the
results. The driving force for bacterial penetration is greater for cans than for aseptically
packaged products. He hypothesized that this difference in driving force, would result in
a smaller penetration threshold for cans than aseptic packages (F loros, 1994).

Michels (1978) determined that wet handling of post-processed retort pouches
resulted in a signiﬁcant increase in microbial penetration compared with pouches dried
before handling (Michels et a1, 1978). This would seem to indicate that moisture may
serve to enhance the movement of contaminants though the sterile barrier.

Anema and Schram (1979) determined that microbial concentration had a signiﬁcant
effect on the percentage of packages contaminated (Table 5) with an immersion biotest,
for both ﬂexible and rigid packages. Higher concentrations of microbes resulted in
100 percent of package contamination. The results below are for ﬂexible pouches with

100 micron holes.

49

 

 

 

 

 

 

 

Table 5 - Percent of Flexible Pouches Found to have Contamination at Varying
ChallengLConcentrations (Anema and Schram—1979)
Concentration of Microbes cells/ml Percentage of Contamination
130 18
20,000 52
520,000 100
130,000,000 100
Cooling Water 4.2

 

 

 

 

Additionally, they evaluated the effect of exposure time on the rate of contamination
(Table 6). The results indicate that this is an important factor for consideration when

designing a microbial challenge test (Anema et al, 1979).

 

 

 

 

 

 

Table 6- Percent of Flexible Pouches Found to have Contamination at Varying
Exposure Times (Anema and Schram—1979)
Exposure Time Concentration of Microbes Percentage
(minutes) cells/ml Contaminated
1 1.5 x 104 15
5 4.0 x 10‘r 58
15 4.6 x 10“ 57
30 2.3 x 10T 45

 

 

 

 

 

Blakistone (1995) and Harper (1995) determined that the threshold defect size is
dependent on the type of microbial challenge test that is used to test packages. This team
concluded that immersion biotesting, which actually submerges a package in a bath of
microbes, represents extreme conditions, and is not representative of the production and
distribution environments for all package types. It was, however, indicated to be a good
representation of the processing faced by retort pouches, which do experience an

immersion procedure.

50

Blakistone (1995) and Harper (1995) also reported defect type as an important factor.
In addition to channel leaks, blisters, wrinkles, incomplete seals, abrasion and punctures
factor into the probability assessment of microbial ingress.

Additionally, Blakistone (1995) and Harper (1995) noted organism type as an
important factor impacting microbial ingress. Researchers cited attributes such as size
and shape, and motility as important determinants. As with Anema and Schram (1979),
Blakistone (1995) and Harper (1995) indicated the concentration of the organisms in the
test environment to be a critical factor. This was regardless of whether the organism was
a vegetative cell or a spore, and was also found to be true for bacteria, yeasts and molds
(Blakistone et al, 1995; Harper, 1995).

Marcy (1995) examined the defects themselves as having the potential to impact the
penetration threshold. He found that factors such as hole diameter and surface properties,
growth potential in the vicinity of the hole, and bacteria concentration were signiﬁcant
effects. He concluded that channel leak length does not affect the likelihood of
contamination in his research with biotest immersion and ﬂexible pouches
(Marcy, 1995).

Scholla (2000) hypothesized that the speed of the particles containing microbes to be
a signiﬁcant factor. Scholla indicated that microbes exist as discrete particles, bound
together and bound to other particles, typically present as an aerosol, and that complexity
exists with the function of the aerodynamic particle size, which has to do with the mass
and dimensions of a particle and its face velocity.

He employed ﬂow ﬁltration theory to explain the behavior of these complex systems.

Filtration theory indicates that particles that are moving as a result of nothing more than

51

Brownian motion will penetrate materials via diffusion. Particles traveling at great rates
of speed are likely to experience inertial impaction, be ﬁltered when they collide with
package materials at high rates of speed. In between these two extremes is interception,
the point where neither diffusion nor inertial impaction dominates; it is this mechanism
that is hypothesized to bring about the greatest microbial penetration. Added complexity
exists when all three (interception, inertial impaction, and diffusion) ﬁltration
mechanisms happen at the same time (Scholla, 2000).

Factors affecting diffusion capture include a decrease in particle velocity resulting in
higher retention times (pressure differential, ﬂow rate). Electrostatic charge affects
diffusion capture. This has signiﬁcant ramiﬁcations to the device industry, as radiation, a
type of sterilization technique for medical devices, is known to eliminate electrostatic
charge on objects (Scholla, 2000). Bacteria tend to posses a negative charge (Brooks,
2004), thus the net negative charge of the bacteria and the packaging materials would
likely result in the repulsion between the bacteria and the materials. The introduction of
the aerosol in the cabinet would ‘bleed’ off the negative charge thus reducing the
repelling forces.

Lake (1985) found the effect of post processing contamination a signiﬁcant factor.
This team stated that it is reasonable to assume that, theoretically, a 1 micron hole would
be the smallest hole that would permit the passage of a bacterial cell if one considers only
a particle size relationship to the opening. However, the situation is quite different when
considering the parameters for post-processing spoilage in canned foods in which a liquid
is necessary to transport the infecting organism into the can. The size relationship is an

oversimpliﬁcation of the situation because it does not account for physical properties

52

such as viscosity, temperature, surface tension, volatility, capillary pressure, etc., present
in a liquid system. This team considered the following factors to be important: Whether
the bacteria is in a chain, or clumps, bacterial encapsulation, the presence of ﬂagella or
other cell appendages, the presence of electrostatic forces and the physical characteristics
of the leak itself. Based on these factors, Lake (1985) states that the smallest critical hole
size for the post processing contamination of canned foods would be considerably larger
than 1 micron (Lake et al, 1985).

The objective of McEldowney’s (1989) research was to investigate the effect of
physical and microbiological factors on container leakage through the development of a
model system for container leakage. The extent of interaction between numerous factors
was determined. Physical factors such as container vacuum, content viscosity, leakage
channel size and shape were evaluated. Microbiological factors such as cell morphology
and motility, cell concentration, pure and mixed bacterial suspensions, bioﬁlms and
monolayer ﬁlms of organisms were also evaluated. The results indicated that there are
three major factors inﬂuencing leakage; ﬂuid availability, container vacuum, and content
viscosity. While these were considered the dominant factors of this research, they were
also impacted by the other factors previously discussed. McEldowney’s research clearly
illustrates the complexity of understanding the container leakage process
(McEldowney et al, 1989).

Put’s (1972) research focused on the mechanism of bacterial infection of canned
foods using a bacterial immersion testing. This team also examined the inﬂuence of can
construction and handling on microbial penetration. Several factors were determined to

contribute to the re-infection of canned foods. The factors include the construction of the

53

can which incorporates several measurable parameters, the size, shape, length, proﬁle of
the leak, and the capillary action within the passage, the concentration of the microbial
suspension, their motility, and reproductive rates, the time the can seams were exposed to
the contaminated liquid, the viscosity of the contaminated liquid, the atmospheric
pressure within the can after sterilization and cooling, handling, and sudden pressure
ﬂuctuations that may occur during sterilization and cooling.

Put’s team concluded that bacterial re-infection of the cans resulted ﬁ'om passive
transport, and occasionally active transport, of bacteria through a capillary passage. The
efﬁciency of this transport is inﬂuenced by the viscosity of the suspension, the ability of
the organisms to reproduce during the test and the ability of the passage to ﬁlter out
organisms (Put et al, 1972).

Ahvenainen (1972) recognized that while previous research regarding critical hole
size on cans and retort pouches cannot be applied to plastic containers, many variables
involved apply to both types of packages. The variables he cited include vacuum,
bacterial morphology, cell concentration, and leakage channel shape. The results of this
research indicate that the probability of microbial penetration is dependent on the type
and viscosity of the product contained within the plastic container (Ahvenainen et al,
1992).

Kamei et a1 (1991) examined microbial penetration for aseptically packaged milk in
paperboard cartons. The focus of their research was to compare the difference in
microbial penetration of cartons containing a high fat product, milk, and a low fat
product, coffee. They placed pinholes (0.2 to 0.3mm diameter) in the board with sewing

needles. They found that an increase in fat content in the contained product resulted in

54

higher rates of contamination through pinholes. They found no difference in samples
with only one induced pinhole and those with 10 induced pinholes (Kamei, 1991).

Placencia’s et al (1986) objective was to illustrate the efﬁcacy of an aerosol chamber
method to test packaging materials. They discussed the following factors as contributing
to the microbial penetration of a package; temperature changes, transportation, and
handling, storage conditions, altitude changes, changes in pressures during surgical
procedures, and microholes in the material, sterilization processes, and any conditions
that may result in ‘bellowing’ of the package (Placencia et al, 1986).

Bankes and Stringer (1988) developed a model system to evaluate the mechanisms of
food container leakage. Factors they concluded to affect the leakage of microorganisms
into the container included the presence of vacuum, the size and shape of the leakage
channel, and volume of liquid passing through the channel and the test organism used
(Bankes and Stringer, 1988).

Purohit et al (1995) stated, ‘To our preconceived notions of size as a restriction to
microbial ingress, we need to add the parameters of tortuous path, mean free path length,
pressure differentials, driving forces, random events and conditions of surface tension,
viscosity, hydraulic head pressure, capillary ﬂow and reversibility, and transient and
repairable leaks.’

Harper (1995) discusses the results of various microbial challenge tests cited in the
literature. She states that the results of these tests vary depending on test conditions such
as channel length, channel diameter, bacterial contaminate concentration, pressure

differential, application technique, and ﬂuid viscosity (Harper, 1995).

55

Maunder et a1 (1968) found that the biotest organism concentration used for can
testing, prior to 1968, of 5 x 10° cells/ml was inadequate for testing ﬂexible pouches.
This conclusion was based on the fact that 30% of 86 defective pouches failed to show
contamination at this level. Therefore, Maunder used a concentration of 20x10° cells/ml
for his work with ﬂexible pouches. This research team also noted that maximum
incubation time for the test product following the biotest is a function of the interaction of
the test organism, packaged product, and temperature (Maunder et al, 1968).

Tallentire and Sinclair (2002) discussed the microbial penetration effect of ﬂow rate
with porous materials. With very low ﬂow rates, the extent of spore capture is inversely
related to ﬂow rate, so the extent of penetration increases with increasing ﬂow rates
(diffusion predominates). The opposite is true of very high ﬂow rates. At higher ﬂow
rates, the particles leave the airﬂow and impact and adhere to the material ﬁbers,
resulting in less penetration at higher ﬂow rates (impaction predominates). They suggest
a moderate ﬂow rate results in maximum particle penetration, an important measure of
the materials biobarrier performance (T allentire and Sinclair, 2002).

Lampi (1990) concluded, based on physical principles, that microbial penetration
would be unlikely to occur in a hole less than 10 microns. The rationale was that in a
retort pouch approximately three pounds of pressure are required to overcome capillarity
in a 10 micron hole and 3 to 4 psi pressure differential is reasonable for the biotester
action. Lampi based his conclusions on a study that involved a 10 micron hole in a retort
pouch that was roughly 1/10 the thickness of the pouch material. The test organism

diameter was roughly 0.5 micron (Lampi, 1980).

56

Howard and Duberstein (1980) discovered that certain types of bacterial species
naturally occurring in water were able to penetrate 0.2 micron rated ﬁlters. They noted a
signiﬁcant retention of the organisms in the ﬁlter and that penetration increased over time
(Howard and Duberstein, 1980).

Conclusion

As a result of the variation cited in the preceding sections, and in some cases
contradiction, of results of cited research, it is clear that biotesting results can be
misleading. There is a need to develop a body of knowledge concerning critical hole size
for medical device packaging. The research detailed in this dissertation addresses the
question of the need for a driving force, other than that introduced by gravity or

molecular motion, for a non-porous tray intended for packaging medical devices.

57

CHAPTER3

MATERIALS AND METHODS

For reasons detailed in Chapter 1, Introduction, a new method to measure the
microbial penetration into packages was needed in order to obtain meaningful results for
the primary objective of this research, an examination of the effect of pressure differential
on microbial ingress. The new method consists of aseptically ﬁlling packages with sterile
agar and then incubating closed packages to see if any microbial growth takes place
(Appendix I; Figure 22). In order to utilize the new method, it was necessary to verify
that the process of ﬁlling the package did not signiﬁcantly inﬂuence microbe penetration.

Details of this preliminary experiment and results are presented in Appendix 1.

Materials

0 Escherichia coli K-12; ATCC Number 29181
This organism was examined via scanning electron microscopy (SEM) to measure
organism size and to observe morphology. E-Coli K-IZ is a gram-negative, motile,
straight-rod that ranges in size ﬁem 1.1-1.5um‘2.0-6.0um. (Brooks et al, 2004)
Figure 3 is a photo of the challenge organism used in Experiment One (New Method

Development; Appendix I) and Experiment Two (Primary Research).

58

 

Figure 3 - Escherichi coli K-12; A TCC 29181
SEM Photograph

0 Trays formed from blue tint, uncoated glycolized polyethylene terphthalate (PETG);
nominal preform material thickness; 25 mils. The trays were supplied by Perfecseal
from a Medtronic mold; Part Number 350215-001.

o Lidstock-LKF-002 Paper/PE/Foil/PE/HSC Die Cut supplied by Amcor Flexibles

Approximate
location of laser-
drilled holes in
PETG tray

 

 

Figure 4 - Test Tray

59

Corrugated Shippers and PE bags with ties for transport of trays to the sterilization
site

Nutrient Agar (Difco, a trademark of Becton, Dickinson and Company, San Jose, CA)
REF #212000, Lot #2043004

Nutrient Broth (Difco, a trademark of Becton, Dickinson and Company, San Jose,
CA) REF #233000, Lot #3188850

60 ml syringes with luer lock tip-Sterile; (Becton, Dickinson and Company, San Jose,
CA) REF #309653

l6—gauge needles, 1.5 inch length-Sterile (Becton, Dickinson and Company, San
Jose, CA) REF #305198

l8—gauge needles, 1.5 inch length-Sterile (Becton, Dickinson and Company, San Jose,
CA) REF #305196

Vented needles with 0.2 micron air ﬁlter-Sterile (B. Braun, Bethlehem, PA) REF

#40211

 

Figure 5 - B. Braun Vented Needle System

o Eppendorf tubes with screw cap and O-ring-Sterile (Becton, Dickinson and Company,
San Jose, CA)

0 1, 5, and 10 ml pipettes-sterile

0 Self Sealing Septum Material supplied by (Mocon, Minneapolis, MN)

0 70 percent Isopropyl Alcohol Swabs

o Non-Latex Exam Gloves

0 50 percent Glycerol Solution

0 Chlorine Solution (NaOCl)

0 Potassium Phosphate, Monobasic, Crystal (KHzPO4)

o Becton Dickinson EnterotubeTM II, Catalog #211832, for the identiﬁcation of gram
negative bacteria. (Becton, Dickinson and Company, San Jose, CA)

Services

0 Gamma Sterilization - The appropriate dose, in kiloGrays, was determined in
conjunction with the gamma sterilizer provider. The dose level and process records
were recorded. Sterigenics, the radiation sterilization provider, indicates that PETG
has a 1000 kiloGray tolerance level and the material remains very stable and
maintains clarity (“wnlzstcrigenics.c0m, 2005).

0 Laser Hole Drilling - Thermal laser drilled holes were placed in the same location on
each test tray; the formed PETG tray in the center of the bottom of the tray (Figure 8).
The laser holes were drilled by Lenox Laser, Glen Arm, MD. The holes drilled were
10 micron :1: 10% and 100 micron i 5% in diameter. The hole diameter and ﬂow rate
were provided by Lenox Laser for each test tray. (Reference Results Tables 15, 16,

17, and 18)

6|

Equipment

Aerosolization Cabinet

The test trays were sprayed with E. coli K—12 in a closed, airtight, Plexiglas® cabinet.
The internal dimensions of the cabinet are 28.5 in x 20 in x 36 in (LxWxH). The
Plexiglass® is 0.25 inch thick. The aerosol containing the microorganisms was
delivered through dual pressurized (40 psi) spray canisters mounted on the top of the
cabinet, the aerosol remained within the cabinet. The canisters were Central
Pneumatic Air Spray Guns Model #43760 with an external mix nozzle 0.082 inches
in diameter. The cabinet has an integrated racking system designed to hold four test
trays and syringes in a uniform orientation for testing. The rack surface is 12 inches
from the ﬂoor of the cabinet. The cabinet has two gloves integrated into the ﬁ'ont
wall (Reference Figure 7), enabling sample and syringe manipulation while the

cabinet is sealed.

62

Rack with

 

Figure 6 - Racking System Inside Aerosol Cabinet

  

Mounted Aerosol Canisters

Figure 7 - Mounted Canisters on Aerosol Cabinet

63

Quebec Colony Counter

A Darkﬁeld Quebec Colony Counter was used to provide even, glare-free
illumination of the colonies so they are bright and easily distinguished to allow for a
visual count (Reference Figure 17). The counter has an adjustable specimen holder
for centering round dishes with diameters up to 100 mm and square culture plates up
to 100 mm x 100 mm. A 1.5X auxiliary lens ﬁts over the standard lens increasing
magniﬁcation to 3.0X. The adjustable focusing rod allows the 1.5X standard lens to
be raised or lowered. The lens also rotates a full 360° for ready access to culture
specimens. A built-in tilt leg may be mounted in the front or rear of the instrument
allowing a convenient tilt angle, or it may be locked ﬂat to the instrument base. A

white-ruled Wolffheugel counting plate was used to facilitate the counting of colonies

Heat Sealer
SenCorp Tray Sealer, Series MDZ420 (SenCorp, Hyannis, MA), dual-shuttle heat

sealer with tray 9-up tooling - Tray die position was recorded.

Methods

Outlined below are procedures A though K. These procedures were developed

speciﬁcally for the research presented in this dissertation. Procedure A, ‘Microbial

Cabinet Analysis’ was run ﬁrst, followed by Procedure B, ‘Experiment Two Procedures.’

Procedures C through K support and are referenced in Procedures A and B.

All microbial and medium handling procedures were developed and conducted in

accordance with ‘Current Protocols in Molecular Biology’(2005).

Procedure A - Microbial Cabinet Analysis

The development of a microbiological growth curve was performed as described

here. It was intended to illustrate that:

The organisms in the aerosolized state were viable and to quantify viability.
The dosage level of microbes delivered during aerosolization.

The consistency of the aerosol within the cabinet.

The growth rate of E. coli.

No other contaminants are present within the system that can survive on the
medium in the trays.

An absence of contaminants after disinfecting the cabinet.

. To determine the dosage of microbes delivered during aerosolization and the

duration of aerosolization, the aerosol system was run with deionized water in an
empty, clean cabinet. The two aerosol canisters were each ﬁlled with 750 ml
deionized water as measured with a graduated cylinder. The set up was the same
as intended for the primary experiment.

The time that elapsed until the cabinet ﬂoor was covered with a ‘ﬁlm’ of liquid
was recorded as 15 seconds. A ‘ﬁlm’ is deﬁned as a complete, visible covering of
the surface with no pooling. The remaining volume in the reservoir was measured
as 725 ml for each canister. Therefore, the total vollune delivered was 50 ml.
This duration, 15 seconds, was used as the starting duration for the development
of a growth curve.

A total of 28 sterilized trays (8 trays per microbial concentration and 4 for

controls) sealed with non-porous lidstock were tested (per Materials Section).

65

Gamma sterilization was provided by Smith & Nephew Orthopedics and
performed at Sterigenics. The Run Number was SMN01 SNN 27050270; nm
date 11/22/05. The run was performed per Smith & Nephew process procedure
P000257-L. A minimum dose of 26.5 kGy and a maximum dose of 40.9 kGy was
recorded.

4. Using aseptic technique (steps 1-7 of Figure 22), a minimum of 25 ml of sterile
medium was added through the self-sealing septum material using a sterile
syringe and needle as described in Procedure H “Addition of Medium to Trays”.
Difco Nutrient Agar Manufacture Number 212000; Lot Number 2043004 was
used.

5. Once the medium solidiﬁed, the lidstock was removed ﬁem the tray (4 at a time)
and the trays were placed bottom-side down on the rack inside the cabinet,
Reference Figure 6. The trays were identiﬁed as ‘1’ through ‘4.’ The tray
orientation was consistent from run to run. Trays and dishes were positioned as
follows (Figure 8):

Tray 1 2 3 4

E] E] I] E]

 

 

0% Petri Dish

 

 

 

 

 

 

Cabinet Opening

Figure 8- Cabinet Test Conﬁguration-Top View

6. Also placed on the ﬂoor of the cabinet (uncovered) was a sterile petri dish
containing the same agar as the trays. The dish remained covered until use. The

removed lidstock and covers were handled and stored in such a manner as to

minimize contamination because they were to be used again as covers for the test
units during incubation. The cabinet was closed and the aerosolization process
begun.

Each group was run separately. The ﬁrst group was run with 102 CF U/mL
microbe concentration, the next 104, with the ﬁnal concentration of 10° (per
dilution chart in Procedure I, “Microbial Aerosolization Procedure”).

Since the groups were run in ascending order with regard to microbial
concentration, cleaning was not performed until the completion of all test groups.
Beginning with a microbe concentration of 102 (per dilution chart in Procedure I)
the aerosol process was run for the duration of 15 seconds as determined in Step 2
of this procedure. The trays and the petri dish remained in the cabinet for
30 minutes (including aerosolization time), then were moved to the environmental

chamber.

10. The original lidstock was placed over the opening of the tray and the cover was

11.

returned to the petri dish.

The trays and dishes were placed in an environmental chamber set at 37°C
+/- 2°C; 50% RH for 36 hours. Observable colony forming units were quantiﬁed
using a Quebec Colony Counter per Procedure J, “Enumeration of Colonies”.
Time intervals were originally set at 12 hours, 18 hours, 24 hours, 48 hours, and
72 hours because results from Experiment 1 (Appendix 1) indicated that evidence
of growth after 12 hours and the agar began drying after 72 hours. The actual
time intervals were 12 hours, 24 hours, and 36 hours. The 18-hour interval was

eliminated due to schedule constraints and the 48 and 72 hour eliminated due to

67

12.

13.

14.

colony saturation and agar drying. The agar drying earlier than noted in
Experiment 1 was likely due to the loose cover of the removed lidstock allowing
for increased air circulation.

Steps 8 through 10 of this procedure were repeated using an aerosol exposure
duration of 30 seconds. The remaining volume in the canisters for the doubled
time was 700 m1, resulting in a total volume delivered of 100 ml.

Steps 5 through 12 of this procedure were repeated for microbial concentrations
of 10‘ and 10‘.

The cell growth versus exposure time for each concentration was plotted. The
amount of growth on the petri dish was recorded and was used for comparison to
the dishes run in Experiment 2. The data that follows is from the 36-hour time
interval as it was the maximum growth interval before the agar began to
deteriorate. The results for the 10° concentration delivered for 30 seconds were
difficult to quantify because after the 12-hour point the CFUs became too
numerous to quantify with great accuracy. Therefore, the slope of this
concentration is not as meaningful as the other concentrations. The key data point
for this concentration was approximately 100 CFU/sqin and was noted at the

15 second aerosol duration.

68

Table 7
Colony Forming Unit Quantiﬁcation From
Growth Curve Study (Units CFU/sqin)

10"2 15 seconds Tray 1 Tray 2 Tray 3 Tray 4 A VG Plate

12 hour 3 8 l4 17 ll 4
24 how 8 27 21 27 21 12
36 hour 22 40 32 46 35 30

I094 15 seconds Tray 1 Tray 2 Tray 3 Tray 4 AVG Plate

12 hour 22 22 31 42 29 25
24 hour , 4s _ , 46 59 65 55 47
36 hour 72 75 73 34 76 68

1006 15 seconds Tray 1 Tray 2 Tray 3 Tray 4 AVG Plate

12 hour 58 62 72 52 61 54
24 how 72 71 86 77 77 72
36 hour 114 104 99 88 101 97

[0‘2 30 seconds Tray 1 Tray 2 Tray 3 Tray 4 AVG Plate

12 hour 22 38 34 21 29 24
24 hour 68 77 66 64 69 67
36 how 119 144 140 7 121 131 121

10"4 30 seconds Tray 1* ﬂ Tray 2 -, Tray 3 ,T ray 4 AVG Plate

12hour 4 66 ,f g 61 _ 52,, 57 59 54
24hour 10,8 85 g 93;_ 119 101 87
36hour » 171 153, g 157 ‘ 164 161 151

who 30 seconds Tray-I Tray 2 M Tray 3 Tray 4 AVG Plate

12 hour , w 89 g 105 g 107, 106 102 97
24 hour 250 250* 250 250 250 250
36 hour 250 250 250 250 250 250

Following Disinfection-30 second delivery of sterile buffer solutior

Tray] Tray2 Tray3 Tray4 AVG Plate

12 hour , o o 0 o 0 o
24 hour 0 0 o o 0 0
36 hour 0 o o o o o

69

 

 

10*2 36 Hour Growth

 

140

120 /

1

 

 

 

 

+102 Concentration

 

 

 

 

CFU/sqln
o B ‘5 8 8 8

 

 

 

15 30
Aerml Duration (seconds)

 

 

y=mx+b
y=6.33334x + (-60)
Duration resulting in 100 CFU/sqin is 100=6.33334x-60=25.2 seconds

Figure 9 - 102 Concentration 36 Hour Growth

 

10M 36 Hour Growth

CFU/sqln
easaaéééaé

 

[+1064 Concentration]

 

15 30
Aerosol Duration (seconds)

 

 

 

ymx+b
y=5.666667x + (-10)
Duration resulting in 100 CFU/sqin is 100=5.666667x-10=19.4 seconds

Figure 10 - 104 Concentration 36 Hour Growth

70

 

10*6 36 Hour Growth

—- ".1

 

 

 

 

 

 

 

 

31]]
250
2 2CD
g 150 / L+10’6 Concentratiohl
b 10]
50
0

 

 

 

15 30
Aerosol Duration peconda)

 

 

 

y=mx+b
y=10x + (-50)
Duration resulting in 100 CFU/sqin is 15 seconds

Figure 11 - 10° Concentration 36 Hour Growth

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Compilation of Growth Curves
31]]
250
-§. 211] +109 Concentration
g 150 ff. +1091 Concentration
ti ff—x/ 1% Concentration
1E1] V
50 e”
D r
15 30
Aerosol Duration (seconds)

 

 

 

Figure 12 — Compilation of Growth Curves for 3 Microbe Concentrations

71

15. The volume delivered for each microbe concentration that yields approximately

16.

100 CFU/in2 of media was determined, Reference Figures 9, 10, and 11. The time
to deliver that volume was determined. This became the aerosol duration time for
the remaining experiments. Per the above graphs, the times which yield
approximately 100 CFU/in2 of media, are:
25 seconds of aerosolization when using a concentration of I 02
I 9 seconds of aerosolization when using a concentration of I 04
15 seconds of aerosolization when using a concentration of 1 0°
For the duration time selected for the experiments (Step 14), the ‘Dose Delivered’
column in the following table was completed with the volume times the
concentration (cells/ml), and the ‘Viable Organisms’ column with the average
number of viable organisms on the tray per square inch (CFU/inz). This gives an

indication of the loss of microbes to the aerosol process.

 

Table 8-Microbial Dose Delivery

 

Canister Microbe Resultant Dose Delivered Viable
Volume Dilution Concentration Organisms
Volume

 

1500 ml 1.5 ml of 109 1x10° cells/ml 50 x 106 cells/ml lOl/sqin
(750 x 2) cells/ml

 

1500 ml 1.5 m1 of107 1 x104 cells/ml 63 x 10“ cells/ml 100/sqin
750 x 2) cells/ml

 

1500 ml 1.5 ml of105 1 x102 cells/ml 83 x10Tcells/ml 100/sqin
750 x 2) cells/ml

 

 

1500 m1 No addition 0 50 ml (sterile O/sqin
(750 x 2) buffer solution)

 

 

 

 

 

 

72

17. A ﬁnal group intended to illustrate the ‘absence of microbes’ was repeated after
the prescribed disinfection process using 2500ppm chlorine solution (NaOCl) to
wipe down the cabinet and autoclave the canisters and spray apparatus.

18. 0 CFU/mL microbe dilution (Sterile Butterﬁeld’s Buffer Solution) was
aerosolized for the 30 seconds on 4 trays and a petri dish placed as in Step 5 of
this procedure.

19. The trays were incubated and any growth counted as in Step 10 of this procedure.
Results are included in Data Table in Step 14 (Table 7).

20. The results were reviewed for veriﬁcation of functionality of the cabinet and
aerosol system, viability of the aerosolized E. coli, efﬁcacy of the disinfection
process, and adequacy of the prescribed time intervals for quantifying growth of
the E. coli K-12 to be used for this research.

Discussion-Procedure A
The generation of the microbiological growth curve for the newly constructed
aerosolization cabinet was performed successfully. The following was illustrated:

1. The organisms in the aerosolized state are viable and the viability was
quantiﬁable. Reference data in Step 14 (Table 7).

2. The dosage of microbes delivered during aerosolization was determined.
Reference data in Step 16 (Table 8).

3. The consistency of the aerosol within the cabinet. The data indicates an observed
consistent distribution of aerosol on the test trays. In Experiment Two the trays

were rotated for each run to minimize the impact of positional variation.

73

Figure 13 illustrates the consistent distribution of microbes on a petri dish from

the study. This is representative of growth noted on test trays.

 

Figure 13 - Petri Dish Exhibiting Microbial Growth

4. The growth rate of E. coli. Reference data in Step 14 (Table 7).

5. No other contaminants were present in the system that survived on the medium in
the trays. No particles or foreign colonies were noted during observation.

6. An absence of contaminants after disinfecting the cabinet. Following the
disinfection process no growth was noted on the test trays or the Petri dish.
Reference data in Step 14 (Table 7).

Procedure B - Emeriment 2 Procedures

1. A total of 400 PETG trays and 400 die-cut lids were obtained.

74

. The trays were packaged in double polyethylene bags. The bags were tied and the
trays were then placed in a corrugated case for transport to the laser hole drilling
provider, Lenox Laser in Glen Arm, Maryland.

. Lenox Laser drilled a 100 micron hole :t 5% in 200 PETG trays and a 10 micron
hole i 10% in 200 PETG trays. Each hole was placed in the center of the tray
bottom. The hole area was circled with black marker. Each tray was also labeled
with a pressure sensitive label carrying a tray number, hole size, and hole ﬂow
rate in standard cubic centimeters. Reference Figure 14 for an image of the label
and reference the Results Section, Chapter 5, for the raw data.

. Following the thermal laser drilling of holes in the tray, the trays were then
returned to Michigan State University.

. The packages were sealed on the SenCorp MD 2420 dual shuttle tray sealer.
Sealer die position was recorded. Sealing pmameters were set at the following:
Seal temperature — 325°F, Seal Duration — 2.5 seconds, and Seal Pressure at
70 psi. The trays and seals were visually inspected for defects in the tray and
lidstock. Seal adequacy was visually veriﬁed per ASTM F1886 ‘Visual Tests,
destructive and non-destructive.’ Any trays detected to have defects were
removed hour the study.

. The trays were again packaged in double polyethylene bags. The bags were tied
and the trays were then placed in a corrugated case for transport to the gamma
sterilization facility in Memphis, Tennessee. Gamma sterilization was provided
by Smith & Nephew Orthopaedics and performed at Sterigenics. The Run

Number was SMN01 SNN 27050270; rim date 11/22/05. The run was performed

75

per Smith & Nephew process procedure P000257-L. A minimum dose of
26.5 kGy and a maximum dose of 40.9 kGy was recorded.

Following the sterilization, the trays were returned to Michigan State University.
The trays remained in the cases until the addition of media.

Care was taken to avoid any contamination of test trays while handling. A
sampling of trays was examined visually as in Step 2 of this procedure to ensure
the integrity of the trays. Again, trays determined to have damage were

discarded.

10. Medium was added to the trays as described in Procedure H.

11.

12.

13.

The experiment was conducted in the test group order described in Chapter 4,
Experimental Design. For each group the trays were placed in an inverted
position in a racking system designed to secure the trays and syringes in a uniform
conﬁguration. Reference Figures 6 and 8 for a depiction of test tray
conﬁguration.

For the treatments that required the pressure differential, a self-sealing septum
was placed on the end of the tray. Prior to application to the tray, each septum
were immersed in 70 percent isopropyl alcohol and wafted though a ﬂame to
remove any excess alcohol. The trays that were not exposed to a pressure
differential did not require the addition of the septum material. All trays were
oriented identically.

A sterile l6-gauge needle was placed on a sterile 60 ml syringe. The septum was

pierced aseptically and the needle/syringe system was secured in the cabinet

76

racking system. The bracket was then placed over the end of the syringe
plungers, reference Figure 6.

14. An open petri dish, containing the same agar used in the trays, was placed on the
bottom of the cabinet. The placement of the petri dish was the same as that
depicted in Procedure A (Figure 8). The petri dish was sterile and remained
covered until placement into the cabinet. As mentioned, the petri dishes served as
an indicator of the viability of the organisms and the dose delivered when
compared to the plates produced in Procedure A.

15. Immediately following the sealing of the cabinet, Procedure 1 was initiated;
Microbial Aerosolization, at the appropriate test group concentration.

16. After the aerosol had run for 15 seconds, the prescribed duration, the bracket was
retracted, simultaneously withdrawing 62 ml of air volume from each of the test
trays for which a pressure differential was induced. The rate of retraction was
consistently one minute per plunger retraction to 62 ml mark (reference ‘Pressure
Differential’ section in Experimental Design section for a discussion concerning
rate). This was accomplished following the aerosolization process, without
opening the cabinet, using gloves integrated into the ﬁont cabinet wall (see
Figure 7).

17. Upon completion of the cycle (aerosolization duration and time in
cabinet = 30 minutes), the cabinet was opened and the trays removed from the
rack. The needles and syringes were removed from the trays prior to placing

them in the incubator.

77

l8. Trays were placed, lidstock down, in an environmental chamber controlled at
37°C +/- 2°C; 50% RH for 72 hours. Observable colony forming units were
quantiﬁed by using a Quebec Colony Counter per Procedure J. Inspection time
intervals were 12 hours, 18 hours, 24 hours, 48 hours, and 72 hours. The only
quantiﬁcation recorded was that of the last inspection time, 72 hours, as this was
the only time the lidstock was removed ﬁom the tray to obtain a full view of the
growth media. The pressure sensitive label placed on the tray bottom by the laser
provider identifying the hole size and ﬂow rate restricted ﬁrll view of the media.
Reference Figure 14.

19. To verify that CFUs observed were the test organism, E. coli K-12, the colonies
were sampled and cultured in a Becton Dickinson EnterotubeTM II System, a
multimedia tube intended to perform as an identiﬁcation system for speciﬁc
member organisms of the family of Enterobacteriaceae. Reference Procedure K.

Procedure C - Preparation of Sterile Media
An adequate number of Erlenmeyer Pyrex Glass Flasks for media containment

were obtained. Care was taken when selecting ﬂask volumes to not have media more

than 1 inch from any container surface for adequacy of sterilization.

1. Flasks were clean and ﬁee of any debris. Prior to use, they were rinsed
thoroughly with deionized water (DW).

2. An appropriate volume of DW was measured with a graduated cylinder and
transferred to the ﬂask.

3. Using a calibrated scale, the appropriate amount of nutrient agar powder was

measured and added to the ﬂask containing DW. Care was taken throughout the

78

operation to not introduce any contaminants. The ﬂask was gently swirled to mix.
Agar used for this experiment was Difco Nutrient Agar; REF #212000,
Lot#2043004. The dilution was 23g per 1000 ml of DW.

4. A minimum of 30 m1 of agar solution was transferred to a 100 ml test tube.

5. A foam stopper of adequate size was placed in the opening of the tube with the
liquid media. The stoppers ﬁt tightly. Aluminum foil was placed over the stopper
and top of tube.

6. A small piece of autoclave indicator tape was placed on the foil. Care was taken
to not put adhesive in direct contact with glass.

7. The ﬁlled test tubes were then ready to be autoclaved.

8. Prior to autoclaving, the water bath was brought to the correct temperature;
47°C +/- 2°C

9. The tubes were placed in a tube rack and placed directly on the bottom surface of
the autoclave chamber, centering as much as possible.

10. The chamber door was then closed.

11. Cycle time was set to 25 Sterilize (minutes); Dry (minutes) 0.

12. The cycle was started by pushing the ‘on’ button and ‘ﬂuid’ button.

13. The cycle begins when conditions reach 121°C and 15psi.

14. When complete, care was taken to stand away from the chamber opening as the
door was opened.

15. The door was then slowly opened to allow for steam release.

16. Using asbestos gloves, product was removed and transferred to the water bath.

79

17. The water bath level covers the ‘height’ of agar in the tubes, but care was taken to

ensure that it was below the level of the top of the test tubes. Too low of a level
has the potential to result in gelling of solution, too high a level can result in

surface condensation, a possible contaminate.

18. Agar was used within 6 hours after preparation. To ensure agar sterility, a ﬂask

of the each sterilized agar lot was incubated at 37°C while being gently agitated in
ﬂask shakers for 24 hours. The samples did not exhibit microbial growth

indicating adequacy of the sterilization process.

Procedure D - Preparation of Microbial Stock Solution

1.

Two 100ml lots of Difco Nutrient Broth; Product 233000, Lot #3188850, were
prepared in two, 250 ml Erlenmeyer ﬂasks.

The lots were autoclaved for 25 minutes at 121°C as described in Procedure C.
One ml of sterile nutrient broth was added to the lyophilized Escherichia coli
K-12 vial; ATCC 29181, using a sterile pipette.

The vial was agitated ensuring that reconstitution was complete.

0.5 ml of reconstituted solution was transferred, using a sterile pipette, to each
250 ml Erlenmeyer ﬂask containing 100 mls of sterile nutrient broth.

Solutions were incubated at 37°C for 24 hours. Flasks were gently agitated during
the incubation period.

200 Eppendorf Polypropylene tubes with O-Ring Screw Caps were autoclaved for
15 minutes at 121°C.

100ml of 50% Glycerol Solution was autoclaved. This served as a cryoprotective

agent for the microbial stock solution.

80

9.

100 Eppendorf Polypropylene tubes were prepared hour each 250ml Erlenmeyer
Flask. Each tube contained lml of microbial solution and 0.4mls of 50%
Glycerol. The transfer was performed with sterile pipettes using aseptic

technique.

10. One tube ﬁ'om each ‘lot’ was used to determine the starting microbial load of the

stock solution via the growth curve method. Starting Load for this E. coli stock

solution=lx109 CFU/mL.

11. The remaining tubes ﬁom each ‘lot’ were adequately labeled and placed in a

polystyrene sleeve and stored at —80°C until use.

Procedure E - Preparation of Microbe Dilutions

l.

E. Coli stock solution was removed from ~80°C freezer and placed in wet ice until

use.

When ready to use, the vial was manipulated by hand to melt. Once melted,

inverted to mix (approximately 10 times).

. Using a sterile lml pipette, lml was drawn from the Eppendorf tube. The ml was

placed in 100 mls of sterile Butterﬁeld’s Solution produced per Procedure G.

The dilution was agitated vigorously for a minimum of 30 seconds to ensure an
equal distribution of cells as well as to break up clumps of organisms into
individual cells. This gives a 10‘2 microbe dilution.

Once adequately mixed, lml of the 10'2 dilution was withdrawn in a sterile
pipette. The ml was added to a fresh 100mls of Butterﬁeld’s Solution.

Manipulation was repeated. This gives a 10'4 dilution.

81

6.

Further dilutions were made by transferring 1 ml aliquots into additional dilution

blanks. The process was repeated to obtain 106 and 10‘8 dilutions.

Procedure F - Preparation of Plates with Microbe Dilutions/Pour Method

1.

The dilutions prepared in Procedure B were agitated vigorously for a minimum of
30 seconds.

1 ml aliquots of a dilution were added to a sterile petri dish.

Sterile liquid nutrient agar was added to the plate enough to fully cover the
bottom of the dish.

The dish was gently swirled to mix the agar and the microbe dilution

. A cover was placed on the dish.

The plates were incubated at 32°C for a minimum of 24 hours.
Using a Quebec Colony Counter the colonies on the plate were counted.
To determine the Standard Plate Count (SPC) of the dilutions: SPC=diluted

plate count/dilution ratio

Procedure G - Preparation ofButterﬁeld’s Buﬂ'er Solution

1.

2.

Stock Solution contains:

Potassium Phosphate, Monobasic, Crystal (KH2P04) 34g J .T. Baker CAS 7778-
77-0 Distilled Water 500ml

Alter preparation of the stock solution, pH was adjusted to 7.2 with 1 N Sodium

Hydroxide (NaOH). The volume was brought to 1 liter with distilled water.

. The solution was autoclaved 15 minutes at 121°C and stored in reﬁigerator (4°C)

until use.

82

Procedure H - Addition of Medium to Trays (Reference Figure 22)

l.

2.

10.

ll.

12.

Sealed, sterile empty trays were obtained.

Test tubes containing 30 mls of medium (1 Test Tube per Test Tray) were
autoclaved. Cycle time was set to 25 Sterilize (nrinutes); Dry (minutes) 0.

The tubes of agar were equilibrated in water bath to 47°C +/- 2°C.

Each tray was uniquely coded.

With a 70 percent Isopropyl Alcohol swab, an approximately 2 sqcm area on the
center of the tray lidstock was swabbed using moderate pressure for
approximately 30 seconds.

A pre-sized (approximately 2 cm2) piece of septum material was picked up using
disinfected forceps.

The backing material was removed with forceps and discarded, exposing the
adhesive material.

The septum material was dipped in a 70 percent isopropyl alcohol bath.

The disinfected septum was waited in the ﬂame of a Bunsen burner to ﬂash off
any remaining alcohol.

Still using forceps, the septum material was placed, adhesive side down in the
center of the disinfected area of the tray.

The septum material was ﬁrmly pressed with the forceps to ensure adhesion.

The trays were placed in the jig designed to orient the trays lid-side down while
adding medium and providing a consistent point and depth of needle entry,

reference Figures 14 and 16.

83

 

Figure 14 - Jig Designed for Uniform Introduction
of Media to Test Trays

 

Approximate Laser Hole
Location

 

Figure 15 - Top View of Test Trays Loaded in Jig

 

Figure 16 - Bottom View of Trays Loaded in Jig

13. The test tube stopper was swabbed with a 70% isopropyl alcohol swab for
approximately 30 seconds.

14. Approximately 25 mls of equilibrated medium was drawn up aseptically in a
sterile syringe using a 16-gauge needle from the test tube. The tube remains
closed during this process and the needle pierces the tube stopper.

15. The 16-gauge needle was replaced with a sterile vented needle. The use of a
vented needle allows for out gassing as the medium is injected into the non porous
package-

16. With a 70 percent Isopropyl Alcohol swab, the septum material was swabbed
using moderate pressure for approximately 30 seconds.

17. The entire syringe volume was injected into the tray through the septum through
the ﬁxture in the jig, reference Figure 16.

18. The tray was agitated to ensure an even distribution of agar.

85

19. The medium was allowed to completely solidify undistm'bed.

20. If more than 3 minutes elapsed after a disinfecting procedure before the next
operation was performed, the disinfecting procedure was repeated.

21. This process was repeated for each test tray.

Procedure I - Microbial Aerosolization Procedure

1. Each aerosol canister was ﬁlled with 750 ml of sterile Butterﬁeld’s Buffer
Solution. The solution was then inoculated with 0.75 ml of the desired

concentration of E. coli per the table below:

 

Table 9-Microbe Dilution Volume/Resultant Concentration

 

Canister Volume Microbe Dilution Volume Resultant
Concentration

750ml x 2=1500ml 0.75 ml of 109 cells/ml per canister 1 x 106 cells/ml

 

 

750ml x 2=1500ml No addition 0

 

 

 

 

 

2. Both pumps are then turned on and the test trays sprayed for the duration
determined in Procedure A (15 seconds).

3. After the cycle, test trays designated for exposure to a negative pressure
differential, have syringes simultaneously drawn back via the cabinet racking
system. The trays and syringes remain in this position for 30 minutes (total cycle
time).

4. Following the complete cycle, the trays are removed ﬁ'om the chamber; syringes
and needles are removed and the trays are placed in the appropriate environmental

chamber. The hole created by the needle is covered upon needle withdrawal by

86

the self-sealing septum material placed on the tray before the needle puncture was

made.

. Following test completion, the cabinet is wiped with chlorine bleach and the

canisters rinsed with bleach and autoclaved.

Procedure I -Enumeration of Colonies

1.

The tray, PETG side facing out, was placed on the Quebec Colony Counter for
illumination.

To facilitate the count, a water-soluble marker was used to identify colonies
counted. If a marker was used care was taken to avoid the laser drilled hole. The
magnifying glass was used if needed.

The number of colony forming units (CFU) was recorded for each time interval
for each tray.

If the tray exhibited too much growth to accurately identify individual CFUs, this
was recorded as ‘too numerous to count’ or TNTC.

Any suspicious contamination observed was noted and removed ﬁ'om the tray
media after the ﬁnal time interval, and cultured on selective media for the
organism being tested to determine if it was E. coli. Microscopy was also used to

investigate the source of any unidentiﬁed contaminants.

87

 

Figure 17 - Tray Exhibiting Growth on Quebec Colony Counter

Procedure K - Idenaﬁcation of gram negative colonies via the EnterotubeTM I]

1.

All CFUs observed on test trays following incubation are prepared to be cultured

with this system for veriﬁcation as the test organism, E. coli K-12.

. Remove both caps from the tube ends.
. Place the tube inoculating wire into the well isolated colony, avoid touching agar.

. Inoculate the tube by inserting the wire through the twelve tube compartments and

then withdrawing using a turning motion.

. Stop wire withdrawal when the notch on the wire is aligned with the tube

opening.
Break the wire at the notch by bending.

7. Punch holes with the broken piece of the wire covering the air inlets of the last
eight compartments in order to support aerobic growth in these compartments.

8. Replace both caps to the tube.

9. Incubate at 37°C for 18 to 24 hours.

10. Remove ﬁom incubation and interpret the results visually per the results pad
provided with the system.

11. Following recording of the results, perform the indole test by adding via syringe
and needle two drops of Kovacs’ reagent through the plastic ﬁlm of the
stlindole compartment. A positive reaction is noted by development of a red
color in the added reagent on surface media within 10 seconds. This result is
added to the results pad for interpretation. Kovacs reagent consists of :

Amyl or isoamyl alcohol ............ 75 ml
p-Dimethylaminobenzaldehyde ...... 5 g
hydrochloric acid concentrated ....... 25 ml
12. The ID value arrived at on the provided results pad is located in the provided

interpretation guide. The corresponding organism is reported.

89

.' ’1

[l

 

Figure 18 — B-D Enterotubemll — Identification System for Enterobacteriaceae

CHAPTER 4
EXPERIMENTAL DESIGN
An experiment (Experiment 2-Primary Research) was designed to estimate the
effects of the three factors (called variables): microbial concentration, hole size and
pressure diﬂ’erential on the ingress of E. Coli K-12 into the test trays. Other factors
were held constant (called constants).
Variables
o Microbial Concentrations (2) 0 and 106CFU/mL
0 Hole Diameter (3)0, 10, 100 micron
0 Pressure Diﬁ‘erential (2) O and —3.78 psi
Constants
- Tray and Hole Orientation and Tray Geometry
o Tray Materials
0 Microbe; Motility
o Aerosolization Method
0 Microbial Exposure Time
0 Temperature
Response Variable
o Colony counts of E. coli K-12 in the test tray
Justiﬁcation of Design
It is understood that a variety of factors relating to the package have the potential to
inﬂuence microbial penetration (as previously outlined), but it is believed that keeping

the design of this experiment as simple as possible is the best approach, and will provide

91

information that will be useful to build ﬁrtm'e research that incorporates more of these
variables. Following are the details regarding the variables and the constants in this
experiment.

Variables Details

0 Microbial Concentrations (2) 0 and 106CFU/mL

The results ﬁom Experiment 1, detailed in Appendix 1, indicate the contamination
threshold lies between a concentration of 104 CFU/mL and 106 CFU/mL.

The theoretical exposure of the test trays is 2.5 x 106 cells per cubic meter. This is
based on the maximum microbe concentration used in this study. An average hospital
has an airborne microbe level of 125 cells per cubic meter (Schmidt et al., 1984).
Based on this estimate, approximately 20,000 times this load will be evaluated at the
106 being employed in this study. In a controlled medical device manufacturing area
the airborne microbial load would be expected to be signiﬁcantly less than that of a
hospital (Reich, 1985) or other end use environments.

A greater bio-burden than these packages would typically face was chosen for
several reasons. Anema et al (1980) indicated that when using a biotest to quantify
the amount of leakers, the concentration of organisms should be relatively high to
ensure that all leakers are infected. Put (1972) suggested that, based on her work with
canned foods, fewer leakers will be detected when microbial loads are lowered.
Testing labs currently conducting microbial aerosolization testing for various medical
device manufacturers indicate they routinely begin with a microbial concentration of
107 or 108 cells, resulting in product exposure of 104 and 105 cells respectively

(Mycoscience, Inc, Wilmington, CT and AppTec, Inc, St. Paul, W). Additionally,

92

the higher burden presents a more hostile challenge than packages face in normal
distribution environments, providing a ‘safety factor’ for published results.

Hole Diameter 0 (no hole), 10, 100 micron

0 (no hole) is a negative control.

Reference Figure 14 for a tray photo identifying approximate hole location.

Ten and 100 micron holes were drilled in 200 test trays of each hole size by
Lenox Laser (Glenn Arm, MD) using a thermal laser. Ten microns was selected as
the defect size to be tested, as it is representative of what previous research has
documented as critical (Blakistone et al, 1996; Harper, 1995; Hurme et al, 1997;
Keller et al, 1996 and Keller, 1998; Chen et al, 1991). It also represents a size that is
slightly larger than the upper end of the size range of a typical cell of E. coli, the
challenge organism. Theoretically, if the hole is 10 microns and the organism is less
than 5 microns it should be able to penetrate.

One hundred micron holes were also drilled in 200 test trays. One hundred
microns is substantially larger than what has been cited as a critical leak size. The
‘larger’ hole size was chosen to provide an inducement of a difference, with regard to
microbial penetration, between the two pressure differentials identiﬁed in the study
design. Previous research found pouches with 100 micron holes varied in the rate of
contamination ﬁ'om 18 percent to 100 percent as the concentration of the challenge
suspension increased (Anema et a., 1980). Therefore, at the high level of microbial
concentration (106 CFU/mL), it was likely that microbial contamination would occur

in these samples, providing a positive control for the study.

93

Table 3, “Physical Variables,” ﬁom Chapter 2, Literature Review, illustrates the
wide range of methods used by researchers to create and quantify holes in the
materials or packages for test. In the reviewed literature, pinholes were created by:
puncturing the package wall with needles or wires, melting a hole with a Tesla coil,
melting a hole with a thermal laser, or by inserting standard pinhole oriﬁces. Channel
leakers were created by inserting wires or microtubes in the seal area before sealing
the package and then pulling the item from the seal. Most researchers used
microscopy of varying types to characterize the size and shape of the hole. In many
cases they do not state the type of microscopy used.

Bix et al. (2004) states there are advantages and disadvantages for each different
method. The pinholes created mechanically leave defects in the package that are
more realistic than pinhole oriﬁces or microtubes. Mechanically created pinholes
create an oriﬁce for microbial penetration made of the package material; this is not
the case for pinhole oriﬁces and microtubes, which are typically metal or glass.
Gilchrist (1989) noted that punched holes in ﬂexible laminate materials resulted in the
creation of tears and ﬂaps as well as much variation (from 17-175 microns). The
ﬂaps could theoretically close during testing, resulting in the inability of microbes to
pass, regardless of the hole size, and thus potentially skew the results (Maunder et al,
1968; Gilchrist et al, 1989; Axelson et al, 1990). Additionally, following testing,
Gilchrist indicated that the holes had all decreased in size in varying amounts.
Axelson et al. (1990) and Maunder et al. (1968) also reported large variations in the

size of pinholes they created manually (ﬁ'om 5 to 100 micron and 33 to 160 micron,

respectively).

94

Extensive work was performed by researchers at Michigan State University on the
topic of the creation of defects in packages for research intended to determine
penetration threshold (Bix et al, 2004). The research team found a signiﬁcant effect
of measuring technique (Scanning Electron Microscopy, Confocal Microscopy and
Optical Microscopy) and that the entry side and exit side of a hole certiﬁed to be 50
microns were statistically signiﬁcantly different in size at a level of a=0.0001. It was
hypothesized that the most effective method for determining the hole size was
confocal microscopy. Their work was conducted with the same PETG tray that will
be used in this research.

For this study, each tray hole size was determined based on ﬂow rate. Flow
varies directly with the area of the hole or is a function of the hole diameter squared.
This measurement was performed by Lenox Laser, the company that drilled the holes.
The ﬂow rate is reported in standard cubic centimeters (sccrn) (Reference Results
Tables 15, 16, 17, and 18).

Pressure Differential (2) 0 psi and -3.78 psi

A test group at an elevation of 860 feet above sea level (Lansing, Michigan,
elevation) with a pressure of 14.1 psi, and a differential between outside the tray and
inside the tray of approximately 0' psi was evaluated. This group was not exposed to
an induced pressure differential, only Brownian motion and gravitational settling
were the driving forces.

The second test group was exposed to an induced pressure differential simulating
what a package would be exposed to when descending in an aircraft or ground

shipment ﬁ'om 8,000 ft to an altitude of 860 feet above sea level. Cargo and

95

commercial aircraft have a pressurized cargo hold, usually maintaining pressures of
8,000 ﬂ. The conditions the second test group was exposed to simulated package
exposure to an altitude of 8,000 it, at which the pressure is 10.32 psi. The pressure
inside the tray was 10.32 psi and the pressure outside the tray was 14.1 psi, resulting
in a pressure differential of -3.78 psi.

Presumably, the presence of a negative pressure differential will facilitate
penetration through defects as well as mimic a common mode of transportation for
these types of packages.

It is important to recognize that pressure differentials can also be created through
routine handling by the end user, such as squeezing or dropping, of a package.

To determine the necessary volume of air to be withdrawn ﬁom the tray in order
to induce the pressure differential of -3.78 psi, ﬁrst the tray volume was measured.
The tray volume was determined by measuring the maximum volume of water
contained in the tray with a graduated cylinder; Tray Volume (VT)=154 ml.

By withdrawing a volume of air (VS) from the sealed tray with a sterile syringe
and needle through a self-sealing septum material, the negative pressure differential
the tray experiences is 3.78 psi. This negative pressure differential results in the
higher pressure in the cabinet air pushing inward on the test trays at 3.78 psi
consistently on every square inch of surface area. This, theoretically, should result in
an inward ﬂow of particulates in the air.

The ﬁxed mass of air inside the tray is initially 14.1 psi and occupies a volume of

(VT). When air is withdrawn (VS) the same mass of air now occupies a larger

96

volume (VT + VS) assuming the tray behaves as a rigid system. According to the gas
law pV=nRT=constant so 14.1(VT)=10.32(VT+VS), giving VS=56 ml.

As stated previously, this volmne determination (V S=56 ml) is based on the
assumption that the test tray is a completely rigid system. In reality, this is not the
case. When exposed to a negative pressure differential the lidstock visibly deﬂects
inward and it is assumed, while not visible, that the plastic tray also compresses
slightly. To correct for this, a simple experiment was conducted, using water because
it is incompressible, unlike air. A sealed tray was ﬁlled with water using a syringe
and needle penetrating the lidstock. Air was allowed to escape and the syringe hole
was covered with a non-penetrated piece of septum material, effectively sealing the
hole made by the needle. The resultant tray was ﬁlled with only water (154 ml), no
air, and the lid was completely ﬂat. With a syringe and needle, water was removed
ﬁom the tray in 1 ml increments. This was repeated until the lid deﬂection (d) was
equivalent (visually detected) to the deﬂection seen with the trays containing air when
56 ml of air was removed. This deﬂection is approximately 0.125 inches. Whatever
water is in the syringe is the reduction of the internal volume of the tray. The

following results chart (Table 10) was generated:

 

 

 

 

 

 

 

 

 

 

 

Table 10 - Tray Lidstock Deﬂection
V (volume of water removed) (ml) (1 (deﬂection of lidstock) (inch)

0 0

1 0.0208

2 0.0417

3 0.0625

4 0.0833

5 0.1041

6 0.125

 

 

97

Based on this information the following correction was made to the formula to
account for the tray compressing inward and losing volume V=6 ml. According to
the gas law, 14.1(VT)=10.32(VT+VS-V).

VS=154 ml ((14.1psi/10.32psi)-l) + 6:62 ml

The withdrawal of the prescribed air volume (62 ml) was performed using a
syringe plunger withdrawal mechanism attached to the ﬁxed racking system designed
to support the 4 test trays per group in an inverted orientation (lidstock down);
reference Figure 6. The air volume was withdrawn ﬁom the trays simultaneously.
While it is recognized that the rate of descent ﬁ'om an altitude can vary widely rate
was not a factor in this research. The rate for inducing the pressure differential was
held constant from run to run. The racking system was designed to facilitate a
consistent withdrawal of air from the trays over a one-minute period. Following the
one-minute period of air withdrawal, it is also recognized that the rate of ‘relaxation’
of the trays will vary between the trays with a 10 micron hole and the trays with a 100
micron hole. The rate difference will be considered to be the same as that
experienced by actual packages with these size defects experiencing a descent ﬁom
8,000 feet. An analysis of the impact of the rate differences is outside the scope of

this research.

98

Constants Details

Tray and Hole Orientation and Tray Geometry
A hole was consistently placed in the bottom center of each PETG tray. The trays
were consistently placed, relative to the aerosol, in the aerosolizing cabinet, lid-side
down.
Tray Materials
PETG is a common material for thermoformed medical device trays. Researchers
chose to use a non-porous lidstock because, like the choice of an aggressive microbial
load, it represents a severe case. The limited porosity of the lid does not allow the
system to come to equilibrium with the ambient atmosphere as quickly as with a more
porous lid. As a result, equilibration of the system occurs, presumably, through the
pinhole alone.
Microbe; Motility
Escherichia coli K-12; ATCC Number 29181 was used throughout the
experiment. This microbe was selected for this research for 3 reasons:
1. It is safe and ubiquitous
2. It is routinely used in MSU laboratory work
3. The organism is smaller than Bacillus subtilis (an organism commonly
used in medical device microbial challenge studies)(Reich, 1985; Reich, 1986;
Placencia et al 1986; Bankes, 1988; McEldowney et a1 1989; ASTM F1608-00).
As a result, it is, presumably, a more severe “challenge” of the package system

than the device industry’s traditional organism, Bacillus subtilis.

E-Coli K-12 is a gram-negative, motile, straight-rod that ranges in size from
1.1-1.5um*2.0-6.0um (Brooks et al., 2004). In addition to being a more severe
challenge, it should serve to accurately represent Pseudomonas aeruginosa, a
curved rod that ranges in size ﬁ'om 0.5-1.0},un*1.5-5.0um that is frequently linked
to burn infections (Brooks et al, 2004).

The morphology of the E. coli used in this experiment was characterized via
scanning electron microscopy (SEM), reference Figure 3, and is morphologically
consistent with the published literature (Brooks et al, 2004).

Aerosolization Method
The spray pattern and a racking system were designed such that test samples

received full coverage of microbes during test cycles.

The American Conference of Governmental Industrial Hygienists deﬁne
bioaerosol as an airborne dispersion of particles containing whole or parts of
biological entities, such as bacteria, viruses, dust mites, fungal hyphae, or fungal

spores.

The behavior of bioaerosols is governed by the principles of gravitation,
electromagnetism, turbulence, and diffusion. Since bioaerosol particles are larger
than 1 micron, diffusion due to gravitational settling is the dominant mechanism.
Brownian motion, the irregular movement of microparticles, is not a major factor for
the movement of bioaerosols. Thermal gradients can also induce aerosol movements;
this also will be considered negligible as there is not a thermal gradient in the aerosol
chamber. It is likely that shortly after exiting the spray nozzle, the particles will be in

laminar ﬂow, carried along the airstream with the air molecules. The heavier

100

microparticles will break free and settle gravitationally (Penn State Department of

Physics, 2003).
Microbial Exposure Time

The aerosolization duration for this research was determined through
experimentation with the process intended to correlate the volume of solution
introduced into the cabinet, to the dosage delivered for a given starting concentration
of microbes. The total chamber exposure time of 30 minutes was selected for the
following reasons: Chamber exposure time of 30 minutes per package and a test
organism concentration of 10‘5 was chosen because Keller’s work (Keller et al, 1996)
showed signiﬁcant penetration at this concentration when working with Pseudomonas
fragi TM 849. Work performed by Anema and Schram (1980), found that increasing
exposure time for a given hole size and given microbe concentration from 1 minute to

30 minutes resulted in an increase of contamination ﬁom 15 percent to 58 percent.

ASTM 4169-05 Standard Practice for Performance Testing of Shipping
Containers and Systems Schedule I ‘Low Pressure (High Altitude) Hazard Testing’
prescribes a 1-hour duration. While there is no approved standard at this time for
whole package microbial testing, laboratories are exposing packages to an aerosol of
organisms for a few minutes then leaving the packages in the chamber with the settled
organisms for 1 hour. The descent of the aircraft typically ranges ﬁom 1,000 to
1,500 ﬁ/min. As such, it would take the plane approximately 8 minutes to complete
its descent. If a truck is descending ﬁ'om the mountains at 60 mph, depending on the
grade of the road could take signiﬁcantly longer than eight minutes. During a typical

aircraft landing, recording barometers have shown that the pressure in the cargo hold

101

will increase from 12.2 psi to 14.7 psi in approximately 30 minutes (Hackett et al,

2000).

When the test trays are removed ﬁ'om the chamber aﬁer 30 minutes and
transferred to the incubation chamber, there was no wiping of the tray or disinfection.
In particular, the disinfection step was a source of controversy regarding its possible
contribution to the generation of false positives and false negatives in the traditional
microbial challenge method (Reference Appendix I, New Method Development). So
the TOTAL exposure time for the trays was reported as the time from initial exposure
and included the entire incubation period.

The test conditions produced conditions far in excess of what might realistically
be found in product use. However, this employs a conservative approach intended to
create harsh conditions in order to determine the impact of pressure differential on
microbial ingress.

Temperature and Humidity (T, RH)

Testing: (25°C, 50% RH)

The testing was performed under controlled laboratory conditions (25°C .4:
2°C). Due to the signiﬁcant effect that temperature has on pressure differential, it
is important that the temperature not vary signiﬁcantly. While the temperature
was not monitored dming experimentation and laboratories do vary, no signiﬁcant
temperature excursions were noted. It is recognized that the %RH of the cabinet
air likely increased signiﬁcantly during aerosolization. Placencia et al (1986)
noted a 24 percent RH increase in the aerosol test cabinet used for their research

in the ﬁrst 1.5 minutes of aerosolization. The research team noted that the ‘%RH

102

returned to ambient conditions 5 minutes after the chamber was cleared’
(Placencia et al, 1986).
Incubation: (37°C, 50% RH)
The incubation conditions are optimum for E. coli growth and the
environmental chamber will minimize variation in conditions.
Test Groups
Four trays were tested per trial. Trial One consisted of a run with 4 trays with 100
micron holes at a microbial load of 0 with 2 trays each at 0 psi and —3.78psi. The second
run consisted of 4 trays with 100 micron holes at a microbial load of 106 with 2 trays each
at 0 psi and -3.78 psi. The tray positions with respect to pressure differential were
rotated to provide positional balance. Each run was repeated 3 times giving a total of
8 runs, 32 trays.
Trial Two consisted of a run with four combinations of hole size (no hole and
10 micron) and pressure differential (0 and ——3.78 psi) in the cabinet with a microbial load
of 0. The second run consisted of four combinations of hole size (no hole and 10 micron)
and pressm'e differential (0 and -3.78 psi) in the cabinet with a microbial load of 106. The
tray positions with respect to pressure differential were rotated to provide positional
balance. Each run was repeated 3 times giving a total of 8 runs, 32 trays. This gives
4 replicates of each of 8 treatments ignoring possible run effects. Total trays for Trial One
and Trial Two were 64 trays.
Statistical Design
There are three factors which were varied in the experiment. The factor microbial

load was tested at two levels 0 and 106, the factor hole size at no hole, 10, and

103

100 microns, and the factor pressure diﬂ’erential at two levels, 0 and -3.78 psi. The
aerosol cabinet allows runs in blocks of size 4. As part of the analysis, results were
examined to see if it was plausible to ignore block effects.

The experiment was broken into trials (stages) that proceeded in a sequential manner.
This approach allows for adaptation based on ﬁndings ﬁ'om earlier stages. As decribed
earlier, Trial One kept hole size ﬁxed at 100 microns. The intent was to develop evidence
that the microbes were viable and able as expected to easily penetrate the barriers through
a 100 micron hole. The factors concentration and pressure differential were varied
allowing for estimation of effects for these variables. Based on the results a decision was
made following Trial One to eliminate the 100 micron hole size ﬁ'om subsequent trials.

Trial Two implemented four replications of a complete 23 factorial design with
microbial load at levels 0 and 10’, hole size at levels no hole and 10 microns, and
pressure drﬁerential at levels 0 and -3.78 psi. A decision stemming ﬁ'om Trial Two was
to eliminate the 0 microbial load from subsequent experiments. In this case, the model
becomes a simple two by two factorial run in blocks of size 4 (cabinet capacity).

The response variables analyzed included whether the trays showed any colony
forming units (an attribute variable Yes or No) and the colony count at 72 hours.

The decision on the number of additional runs needed to detect meaningful
differences in the h'eahnents was made following the completion of Trials One and Two.
To increase the power of the results ﬁ'om the ﬁrst trials, two identical replicate trials
(Trials Three and Four) were conducted. These trials consisted of 36 trays each. Four of

the hays were exposed to no microbial load, a conh'ol group. The remaining 32 hays

104

contained a 10 micron hole and were exposed to a microbial concentration of
106 CFU/mL
0 Data Analysis
Frequency counts for the presence of colony forming units were compared across
conditions using the Fisher Exact Test to test the effects of pressure differential and
hole size on microbial penetration. Effects of hay position and hole size variation
ﬁ'om nominal were tested using counts of colony forming units with t-tests and F-

tests. All effects were considered at the 0.05 level of signiﬁcance.

105

CHAPTER 5
RESULTS
The trials and runs that make up the primary research (Experiment Two) are
described below using h'ee diagrams (Tables 11 through 14). Following these tables,
tables of colony counts (CFUs) following 72 hours of test hay incubation are given as the
response for the various factor level combinations that were tested. Run numbers and

hay positions within the runs are also given.

Table 11 - Trial One

Microbial Load! cells) Hole 8% micron) Pressure Differentialmsi) #TRAYS
Test Runs 1A, 18, 1C, ID

 

 

0 a 100 2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
16
Test Runs 2A, 23, 2c, 21)
106 , 100 2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
2(0) 2(-3.78) 4
16

106

Table 12 - Trial Two

Microbial Loadjcells) Hole SE micron) Pressure Differentialtmi) #TRAYS
Test Runs 3A, 3B, 3C, 3D

 

 

 

 

o<: 2(0) 2(-3.73) 4
o <: 2(0) 2(—3.78) 4
10 2(0) 2(-3.78) 4
< 2(0) 2(—3.7s) 4
16

Test Groups 4A, 43, 40, 41)
o 4 2(0) 2(-3.78) 4
10‘5 <: \ 2(0) 2(-3.73) 4
10 » 2(0)2(-3.78) 4
\ 2(0) 2(-3.78) 4
16

107

Table 13 - Trial Three (Replicate One)

Microbial Load! cells) Hole SE micron) Pressure Diﬁerenﬁaltpgi) #TRAYS
Test Run 5A

 

 

0 >10 > 2(0) 2(-3.78)

hl-ﬁ

Test Runs SB through 51

2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)
106 >10 2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)

 

WIAhhA-h-h-hb

N

Table 14 - Trial Four (Replicate Two)

Microbial Loadteells) Hole SQ micron) Pressure Diﬂ’erentialjpgi) #TRAYS
Test Run 6A

 

 

0 +10 > 2(0) 2(-3.78)

Jil-b

Test Runs 6B through 61

2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.7s)
106 4,10 2(0) 2(-3.73)
2(0) 2(-3.78)
2(0) 2(-3.78)
2(0) 2(-3.78)

 

wIA-h-h-hAA-h-h

N

108

Test

Run#

1A
1A
1A
1A
1B
13
1B
13
1C
1C
1C
1C
1D
1D
1D
1D

Test

Run#

2A
2A
2A
2A
ZB
2B
ZB
2B
2C
2C
2C
2C
2D
2D
2D
2D

Tray

Number/

Position

AWNﬂ-bWNﬂ-kWNﬂ-th—t

Tray

Number/

Position

#WN—hlJJN—‘hWN—‘hbéN—t

Table 15 — Trial One Results

Pressure
Differential

(Psi)

0
0
-3.78
-3.78
-3.78

-3.78
-3.78
-3.78

-3.78
-3.78

Pressure

Differential

(Psi)

0
0
~3.78
-3.78
-3.78

-3.78
-3.78
-3.78

-3.78
-3.78

Aerosolized
Microbe

Concentration
(cells/m1)

OOOOOOOOOOOOOOOO

Aerosolized

Microbe

Concenhation
(cells/ml)

10“
10“
10°
106
106
lo6
106
10“
10‘5
106
10"
106
10‘5
10‘5
10“
10‘5

109

Flow
Rate
(sccm)

175.547
163.328
172.794
165.632
171.290
168.761
180.278
177.090
153.391
160.792
169.735
179.604
175.718
170.977
176.004
165.458

Flow
Rate
(sccm)

Hole

Diameter
(micron)

100.154
96.605
99.365
97.284
98.932
98.199
101.494
100.593
93.620
95.852
98.510
101.304
100.202
98.841
100.284
97.233

Hole
Diameter
(micron)

155.083
161.31 1
164.869
166.778
166.536
160.455
170.197
168.076
166.809
157.798
173.677
183.373
169.156
181.738
176.079
172.251

94.135
96.034
97.060
97.620
97.549
95.779
98.616
97.999
97.629
94.983
99.619
102.362
98.3 13
101 .904
100.305
99.209

CFU
(Total
Count)

OOOOOOOOOOOOOOOO

CFU
(Total
Count)

68
121
193

180
76
148

92
136
0

Table 16 — Trial Two Results

Test Tray Pressure Aerosolized Flow Hole CFU

Run # Number/ Differential Microbe Rate Diameter (Total
Position (psi) Concentration (sccm) (micron) Count)

(cells/ml)

3A 1 0 0 0 0 0

3A 2 0 0 0 0 0

3A 3 -3.78 0 1.831 10.406 0

3A 4 -3.78 0 1.726 10.102 0

3B 1 -3.78 0 0 0 0

3B 2 0 0 0 0 0

3B 3 0 0 1.964 10.776 0

3B 4 -3.78 0 1.916 10.644 0

3c 1 -3.78 0 0 0 0

3c 2 -3.78 0 0 0 0

3c 3 0 0 1.697 10.018 0

3c 4 0 0 1.89 10.572 0

3D 1 0 0 0 0 0

3D 2 -3.78 0 0 0 0

3D 3 -3.78 0 1.710 10.057 0

3D 4 0 0 1.842 10.438 0

Test Tray Pressure Aerosolized Flow Hole CFU

Run # Number/ Differential Microbe Rate Diameter (Total
Position (psi) Concenhation (sccm) (micron) Count)

(cells/ml)

4A 1 0 106 0 0 0

4A 2 0 10‘5 0 0 0

4A 3 -3.78 106 1.673 9.946 3

4A 4 -3.78 10‘5 2.579 12.350 0

4B 1 -3.78 10‘5 0 0 0

4B 2 0 10‘5 0 0 0

413 3 0 106 1.430 9.197 0

413 4 -3.78 106 x 1.892 10.577 0

4c 1 3.78 106 0 0 0

4c 2 -3.78 106 0 0 0

4c 3 0 106 1.636 9.836 0

4c 4 0 106 2.549 12.277 0

4D 1 0 106 0 0 0

4D 2 -3.78 106 0 0 0

41) 3 -3.78 106 1.519 9.478 0

41) 4 0 106 1.959 10.762 0

110

Table 17 - Trial Three Results

Test Tray Pressure Aerosolized Flow Hole CFU
Run # Number/ Differential Microbe Rate Diameter (Total
Position (psi) Concenhation (sccm) (micron) Count)
(cells/ml)

5A 1 0 0 1.862 10.493 0
5A 2 -3.78 0 1.165 8.301 0
5A 3 0 0 1.935 10.696 0
5A 4 -3.78 0 1.439 9.226 0
53 1 0 106 1.636 9.837 0
53 2 0 106 1.820 10.376 0
53 3 -3.78 10“ 2.090 11.117 0
53 4 -3.78 106 1.716 10.075 0
5c 1 -3.78 106 1.787 10.279 2
5c 2 0 106 1.413 9.141 0
5c 3 0 10‘5 1.960 10.765 0
5c 4 -3.78 106 1.402 9.107 0
51) 1 -3.78 106 1.866 10.503 0
5D 2 -3.78 10‘5 1.762 10.207 0
51) 3 0 10‘5 2.277 1 1.604 0
51) 4 0 10" 1.589 9.694 0
53 l 0 106 1.786 10.277 0
53 2 -3.78 106 1.837 10.424 0
53 3 -3.78 106 1.396 9.085 0
53 4 0 10‘5 1.504 9.432 0
SF 1 0 10‘5 2.102 11.150 0
53 2 0 106 1.707 10.047 0
51: 3 -3.78 106 2.270 1 1.587 0
53 4 -3.78 . 106 1.664 9.921 2
SG 1 -3.78 10‘5 1.749 10.170 0
so 2 0 106 1.804 10.330 0
SG 3 0 106 1.617 9.779 0
so 4 -3.78 106 1.919 10.653 0
5H 1 -3.78 106 1.541 9.545 0
511 2 -3.78 10‘5 1.590 9.697 3
SH 3 0 106 2.082 1 1.096 0
511 4 0 106 1.424 9.177 0
51 1 0 106 1.824 10.387 0
51 2 -3.78 10“ 1.961 10.770 0
51 3 -3.78 10‘5 1.517 9.473 0
51 4 0 106 2.215 11.446 0

111

Test
Run #

6A
6A
6A
6A
6B
6B
6B
6B
6C
6C
6C
6C
6D
6D
6D
6D
6E
6E
6B
6E
6F
6F
6F
6F
6G
6G
6G
6G
6H
6H
6H
6H
61

61

61

61

Tray
Number/
Position

ALAN—thigh)—#wN—AMN—AMN—hlﬂN—AMN—bb’N—huh)!—

Pressure
Differential

(Psi)

0
-3.78
0
-3.78
0
0
-3.78
-3.78
-3.78
0
0
-3.78
-3.78
~3.78
0
0
0
-3.78
-3.78

-3.78
-3.78
-3.78

-3.78
-3.78
-3.78

-3.78
-3.78

Aerosolized

Microbe

Concenhation
(cells/ml)

0
0

0

0

10“
106
106
10"
106
106
106
10“
106
10‘
10“
106
10‘5
106
10“
10"
106
106
106
10“
10‘
106
106
10‘5
106
106
106
106
10"
106
10‘
10‘5

112

Table 18 - Trial Four Results

Flow
Rate
(sccm)

1.536
2.144
2.393
1.361
1 .610
1.271
1.344
1.679
1.822
1.370
1.628
1.463
1.784
1.874
1.846
1.803
1.454
1.466
1.415
2.701
1.601
1.376
1.462
2.005
1.445
1.545
1.831
1.793
1.460
1.925
1.454
1.136
1.726
1.739
1.368
2.004

Hole

Diameter
(micron)

10.493
8.301
10.696
9.226
9.837
10.376
1 1.1 17
10.075
10.279
9.141
10.765
9.107
10.503
10.207
1 1.604
9.694
10.277
10.424
9.085
9.432
1 1.150
10.047
1 1.5 87
9.921
10.170
10.330
9.779
10.653
9.545
9.697
1 1.096
9.177
10.387
10.770
9.473
1 1.446

CFU
(Total

0

O

.5.
v

oomoooooooooowooocooooocooowo—oooooo

Pehi dishes containing sterile media were placed on the ﬂoor of the cabinet for each
nm as an indicator of organism viability, microbe dosage delivered and an absence of
foreign organism in the system. Table 19 illustrates that the organisms were viable, the
dose delivered was consistent from run to run and that no foreign organisms were present

in the system.

113

Table 19 - Petri Dish Growth Quantiﬁcation Results

Trial Number/Run

1A
13
1C
1D
2A
ZB
2C
2D
3A
3B
3C
3D
4A
4B
4C
4D
5A
SB
5C
SD
5E
5F
SO
SB
51
6A
6B
6C
6D
6E
6F
6G
6H
61

Microbial Load
CFU/mL

None
None
None
None
106
106
106
106
None
None
None
None
106
106
106
106
None
106
10“
106
10‘5
106
10‘5
10‘5
10“
None
10“
106
106
106
106
106
10‘
106

114

Colonies (CFU) per sqin

COCO

102
121
113

138
106
111
98

 

Figure 19 - Petri Dish from a Trial Exhibiting Growth

Results Summary

All reported results are following 72 hours of incubation.
Trial One

All test hays (8/8) with 100 micron holes, exposed to a microbial concentration of

106, experiencing a pressure differential, exhibited growth, while only one (1/8)
100 micron test hay not experiencing a pressure differential exhibited growth. The
number of colonies observed in this test hay (2 CFU) was much less than those
experiencing a pressure differential (Average=127 CFU). No trays (0/16) exposed to
0 microbial concentration exhibited growth. Pehi dishes for each run with microbial

exposure were consistent with levels noted in other runs and in growth curve

development (Procedure A).

115

Trial Two

One test (1/4) with 10 micron holes, exposed to a microbial concenhation of 106,
experiencing a pressure differential, exhibited growth, while no (0/4) 10 micron test
hay not experiencing a pressure differential exhibited growth. No hays (0/16)
without a test hole exhibited growth under any conditions. No hays (0/16) exposed
to 0 microbial concenhation exhibited growth. Pehi dishes for each run with
microbial exposure were consistent with levels noted in other runs and in growth

curve development (Procedure A).

 

Figure 20 - Test Tray Exhibiting Microbial Growth; Trial Three

116

Trial Three

Three test trays (3/16) with 10 micron holes (Figure 20), exposed to a microbial
concentration of 106, experiencing a pressure differential exhibited growth
(average=2.3 CFU), while no (0/16) 10 micron test trays not experiencing a pressure
differential exhibited growth. Number of CFUs was very consistent among test
trays, as seen in the data presented. No hays (0/4) exposed to 0 microbial
concenhation exhibited growth. Petri dishes for each run with microbial exposure
were consistent with levels noted in other runs and in growth curve development
(Procedure A).
Trial Four

Four test trays (4/ 16) with 10 micron holes, exposed to a microbial concenhation
of 106, experiencing a pressure differential exhibited growth (average 2 CFU), while
no (0/16) 10 micron test hays not experiencing a pressure differential exhibited
growth. Number of CF Us was very consistent among test hays. No hays (0/4)
exposed to 0 microbial concenhation exhibited growth. Pehi dishes for each run with
microbial exposure were consistent with levels noted in other runs and in growth
curve development (Procedure A).
li‘nterotubeTM II Results

All identiﬁed CFUs ﬁom Trials Two, Three and Four were conﬁrmed as
Escherichia coli with no atypical tests; Identiﬁcation Code 24460 ﬁ'om the
EnterotubeTM H Interpretation Guide B-D Catalog Number 24330; Revised August
1993. Identiﬁcation was performed per Procedure K with the B-D EnterotubeTM II

system, reference Figure 21.

117

 

Figure 21 - EnterotubeTM II Results

118

Statistical Analyses

As indicated in Chapter 4 - Experimental Design, consistency among conditions for
all hials was of paramount importance and every effort was made to ensure this
consistency.

As indicated in Chapter 3 - Materials and Methods, the basis for the development of
the methods was to ensure run to run consistency. A custom jig was built to ensure
consistent hay handling during test preparation (Figures 14, 15, and 16). The procedures
were veriﬁed for adequacy and the primary research was run and results read by one
person. A primary requirement for the building of the custom aerosol cabinet was again
consistency. The racking system held the hays in a uniform position at a consistent level
in the chamber. The syringes were also held consistently (Figure 6) and the plunger
rehaction was performed simultaneously at a consistent rate (1 minute) with a bracket
which ﬁt over the plungers while the cabinet was sealed. Handling was facilitated during
this process with gloves integrated into the cabinet wall (Figure 7).

The results were examined and analyzed to look for effects due to run and to tray
position. These examinations and analyses showed no evidence of these effects. For
example, note that across all eight hays that showed positive counts with a nominal hole
size of 10 microns (microbial concentration of 106 CFU/mL and pressure differential of -
3.78 psi), the hay positions are balanced, speciﬁcally being Tray 1 (n=2), Tray 2 (n=2),
Tray 3 (n=3) and Tray 4 (n=1). In Trial One, Test Run 2, with a nominal hole size of
100 microns (microbial concenhation of 106 CFU/mL and pressure differential of -
3.78 psi), all eight hays were positive for microbial growth necessarily dividing equally

among the hay positions in the balanced design (Tables 11, 12, 13, and 14). The mean

119

counts for the hays in the four positions are not statistically different (p-value = 0.84
based on an F-test).

As indicated in the Results Tables 15, 16, 17, and 18 in this chapter, the measured
hole diameters varied about the nominal values. This leads to the question of whether
these differences relate to the CFU counts. To address this question, two analyses were
performed. For the eight positives for the 100 micron group (microbial concenhation of
106 CFU/mL and pressure differential of -3.78 psi), the correlation of measured diameter
with count is (correlation coefficient = -0.19, p-value = 0.65 based on a t-test). For the
eight positives in the 10 micron group (microbial concenhation of 106 CFU/mL and
pressure differential of -3.78 psi), the CFU counts were 1 (n=1), 2 (n=4) and 3 (n=3).
The mean hole diameters for the hays in these groups are not statistically different
(p-value=0.25 based on an F -test). These statistical tests, admittedly of little power
because of sample size, retain the hypothesis that the observed counts are not related to
the deviation from nominal hole diameter.

Trial One

Test Group 1 (100 micron. concentration of microbes=0)

Result: 0 positive in 8 hays with pressure differential of 0 psi

Result: 0 positive in 8 hays with pressure differential of -3.78 psi

T_est Group 2 (100 micron. concenhation of microbe§=1061

Result: 1 positive in 8 hays with pressure differential of 0 psi

Result: 8 positive in 8 hays with pressure differential of -3.78 psi

The large difference in magnitudes of the CFUs should be noted: (average

127 CFU per hay for pressure differential of -3.78 psi) versus (2 CFU for hay with

120

0 pressure differential). In total, very strong evidence that pressure differential has a
large effect when a 100 micron pinhole is present and microbial concentrations of 106
are tested. In addition, the effect of pressure differential at 100 micron and
concenhation=106 was assessed with a one-sided Fisher Exact Test. The comparison
of 8 out of 8 versus 1 out of 8 has P-value = 0.0007, demonshating a statistically
signiﬁcant effect.
Trial Two
Igst Group 3 (10 micron, concenhation of microbes=0)
Result: 0 positive in 8 hays with pressure differential of 0 psi
Result: 0 positive in 8 hays with pressure differential of -3.78 psi

Results support the conh'ol exhibited in the results of Test Group 1.

Test Grouﬁﬁ a_n_d 0 micron . concentration of microbes=10f)

Result: 0 positive in 8 trays (0 and 10 micron) with pressure differential of 0 psi

 

 

Result: 1 positive in 4 hays (10 micron) and 0 positive in 4 hays (0 micron) with
pressure differential of -3.78 psi.
For the 10 micron hays, the Fisher Exact assessment (one-sided) comparing 1
positive tray out of 4 hays at -3.78 psi versus 0 positive hays out of 4 hays at 0 psi
gives P-value = 0.5, not statistically signiﬁcant.

Trial Three -Replicate 1 (10 micron, concentration of nricrobes=106) Replicates
Test Group 4 acperinrent with sample sizes of 16 and 16

Result: 0 positive in 16 hays with pressure differential of 0 psi
Result: 3 positives in 16 hays with pressure differential of -3.78 psi
Fisher Exact assessment (one-sided) of P-value is 0.1129; not statistically

signiﬁcant in itself.

121

 

Trial Four -Replicate 2 (10 micron, concentration of microbes=l 0‘) Another
replicate of Test Group 4 experiment with sample sizes of 16 and 16

Result: 0 positive in 16 hays with pressure differential of 0 psi
Result: 4 positives in 16 hays with pressure differential of -3.78 psi

Fisher Exact assessment (one-sided) of P-value is 0.0506; not statistically
signiﬁcant in itself.

Conholling for possible experimental effects for the three separate experiments
(Test Group 4, Replicate l and Replicate 2), the assessment (one-sided) of P-value is
(0.5)(0.1129)(0.0506)=0.0029 for comparing pressure differential of -3.78 psi versus
0 psi at 10 micron and concenhation of microbes=106; statistically signiﬁcant. If
one ignores possible experimental effects for the three experiments and simply
aggregates the results, the Fisher Exact one-sided assessment of 8 out of 36 versus 0
out of 36 gives P-value = 0.0025.

The effect of hole size 100 microns versus 10 microns at pressure differential
-3.7 8 psi and concenhation=106 was assessed using data from Test Groups 2 and 4
and Replicates 1 and 2. The one-sided Fisher Exact comparison of 8 out of 8 versus 8
out of 36 gives P-value < 0.0001 , demonshating a statistically signiﬁcant effect. The
difference 1 out of 8 versus 0 out of 36 at pressure differential 0 psi is not statistically

signiﬁcant.

122

Conclusion

 

 

 

 

 

 

 

Table 20
Aggregated Results of Positive Responses for n = 136 Trays
Microbe Load
0 10‘
Pressure Diff A = -3.78psi A = 0181 A = -3.78 psi A = 0 psi
Hole Size
011 0outof4 Ooutof4 Ooutof4 0outof4
10 l! 0 out of 8 0 out of 8 8 out of 36‘“c 0 out of 368
100 p 0 out of 8 0 out of 8 8 out of 8'" 1 out of 8b

 

 

 

 

 

 

 

' Data ﬁom Test Group 4 and Reps 1 and 2. Statistically signiﬁcant difference.

b Results of Test Group 2. Statistically signiﬁcant difference.

° Data ﬁ'om Test Groups 2 and 4, Replicates 1 and 2. Statistically signiﬁcant difference.
The presence of a pressure differential has a statistically signiﬁcant effect on the

ingress of the test organism in a sterile, rigid medical device hay. The effect of hole size

(100 micron versus 10 micron) at a pressure differential of -3.78 psi is statistically

signiﬁcant.

123

CHAPTER 6

DISCUSSION

This section is a discussion of the results obtained hour the primary research
conducted and detailed in this dissertation. The goal of this ﬁrst section is to explain the
observed number of CFU observed in the test trays. It is assumed that the number of
CFU is the result of the induced pressure differential pulling a quantity of cabinet air
containing microbes into the test hay. This can be check as follows:

0 Two spray canisters deliver a total of 50 ml of liquid with a concentration of

106 CFU/mL in 15 seconds. So 50 x106 = 5 x107CFU enter the chamber.

0 Assuming the CFU spread out evenly inside the chamber (Volume = 28.5” x 20” x

36” = 20,520 in3), they are suspended in air at a concenhation of :

5x 107CFU = 2437 CFUperin3ofair
20,520 in3
0 Once the vacuum is applied (via the pull of the syringe), air (containing CFUs from
the chamber) enters the hole, carrying CFUs with it, and ﬁlls the hay. Air continues
to ﬂow in until the pressure inside the tray matches the pressure outside (equilibrium
is reached).
0 The volume of air that must enter the hay to equilibrate pressures:
Starts at pressure P. —- Ap where P, = 14.1 psi (ambient-Lansing, MI) and Ap =
3.78 psi
The number of moles of air in the hay at start (mum)
= mum = (B. — Ap) V; where V = Tray Volume (same as VT)

RT R = Universal Gas Constant
T = Temperature (absolute)

124

According to the gas law when pressures equilibrate,

nfinal = M
RT

Therefore, the number of moles of cabinet air that must enter is:

Nﬁnal - ninitial = M
RT

From the gas law, this is equivalent to a volume of cabinet air equal to incoming

air volume =

(ApY) gr: VAQ=154(3.78/14.1)=41.3 mL
(RT) P, P.

This is equivalent to 2.5 in3 of cabinet air containing
2.5 m3 x (2437 CFU/m3) = 6092 CFU.

In theory, this is hue regardless of hole size. The smaller the hole, the CFUs ﬂow

into the hay slower, the equilibrium time is longer, but the net CF Us ﬂowing in is the

same.

Why weren’t this many CFUs observed in this research? Some possible reasons:
1. The test time of 30 minutes was not long enough to reach equilibrium

2. There was injury to the micro-organisms

3. The mist containing the concenhation of microbes settled out of the air before

the equilibrium time

125

Exploring these theories:

1. The Bernoulli Equation (Potter and Foss, 1975):
V.2 + Pl V22 + P2 where g = gravity = 32.2 ft/sec2
—- — , — -— y=airdensity=l_lb @72°F
2g y 2g y 13.3 113
relates the ﬂow speed V to the pressure differential, as air ﬂows from State
1 to State 2.
State 1: Starts somewhere within air mass inside chamber:
P1 = pa and Vl = 0
State 2: Just after it passes through the hole and enters the hay, it sees
pressure
P2 = 'P. - Apandhasspeedof V2 =V.
Then the Bernoulli equation becomes:

ai+a=yj+P—A
28 Y 28 1

Solving for the air speed as if ﬂows through the hole gives

v = [2g/y-Ap1V’ where g = gravity = 32.2 11/sec2
y = air density = L11; @ 72°F
13.3 113
and Ap = 3.78 psi = 544.3 lb/ﬂ2 So it passes through the tray

hole with a speed of:

= [(2 x 32.2/1+13.3)(544.3)]'/’= 680 ﬁ/sec

126

The ﬂow rate (volume per unit time) through the hole is AV, where
A = hole area and V = speed (680 ﬁ/sec). For the 100 micron hole
(nominal), the radius is 1.64 x10'4 ft, so the hole area = 1:1:2 = 8.45 x10‘8
ﬁz. So the ﬂow rate initially is:
AV = 8.45 x 10‘8 (680 ft/sec) = 5.75 x 10’5 113/sec = 1.62 ml/sec.
As air enters the hay, the inside pressure rises, Ap drops, the speed
drops, and the ﬂow rate drops. The ﬂow rate eventually becomes zero, so
the average ﬂow rate is about:

1.62+0 = 0.81 ml/sec for the 100 micron hole
2

Since the hay with the 100 micron hole needs an air volume of
41.3 ml to flow in to equilibrate pressures, this should take only

41.3 ml = 51 seconds
0.81ml/sec

These calculations assume no air friction, which is not the case, so
this estimate represents a lower limit.
For the 10 micron hole (nominal), the only thing that changes is A,
the hole diameter, which is 100 times smaller, so it should take 100 x
51 = 5,100 sec = 1.42 hours. Therefore, insufﬁcient test time could
explain the 10 micron results, but does not entirely explain the 100
micron results.
2. Theoretically, the 100 micron hole should have allowed 6092 CFU into the
hay. There was enough time for this to occur, as discussed previously. It is

possible that microbe injury was a conhibuting factor for the reduced amounts

127

of CFUs actually hansferred. The work performed to verify the cabinet
functionality (Chapter 3, Procedure A) indicated signiﬁcant injury to the
organisms during the aerosolization process. The growth rate was 101
CF U/sqin for the dose delivered in this research. This yields a signiﬁcant,
approximate death rate of [(87,720-101)/87,720] x 100 = 99% (Reference
Section 3, below). Additional microbe injury could be occurring to the
organisms which survive the aerosolization process as they enter the hay hole.
Examining the hole in detail, the ratio of hole diameter to length is:

100micron = 4mi1 = 1/5
20 mil 20 mil

For the 10 micron hole, the ratio is 1/50.

The longer the hole compared to its diameter, the greater the opportunity for
impacts between the penetrating microbe and the sides of the hole, which
would likely result in devastating consequences for the micro-organism, given
the speed of 680 ft/sec.

. To make a liberal estimate regarding the rate of the mist dropping out of the
air in the chamber, the assumption is made that the mist settles and falls
vertically down, collecting only on horizontal surfaces inside the cabinet. The
concenhation of CFUs on these horizontal surfaces is 5 x 107 CFUs injected
in cabinet divided by the total horizontal area, which is 28.5” x 20”, so:

Concentration = 5 x107 CFU = 87,720 CFU/in2
28.5” x 20”

128

The 100 micron, hole occupies an area of 8.45 x 10:8 a2 = 1.22 x 10’5 inz, so
the number of CF U’s that land in the hole should be:

87,720 CFU/in2 x 1.22 x 10'5 = 1.06 or just 1 CFU
The 10 micron hole is 100 times less, or 0.01 CFU should land on the hole.
These estimates are low compared to the data collected. Considering the
death rate discussed in Part 2 which occurs during aerosolization, the estimate
for the 100 micron hole becomes 0.001 CFU and for the 10 micron hole
0.00001 CFU that should land on the holes. The data does in fact show that
the 10 micron holes had about 100 times fewer CFU than the 100 micron
holes, which supports the idea that settling accounts for the observed results,
but the CFU counts observed are far greater than these estimates calculated
above. Given such low probabilities, it is signiﬁcant that a 100 micron test
hay was penehated without the presence of a pressure differential. It appears
that the resultant CFU counts in the test trays are a combination of microbes
settling out of the mist onto the test hays and microbes in the mist being
drawn into the hays as a result of the pressure differential driving force.

In addition to the considerations surrounding the driving force and microbial
penetration, another factor that must be considered is surface tension. Surface tension is
the result of cohesive forces between liquid molecules. Molecules at the surface do not
have other like molecules on all sides of them and, consequently, they cohere more
shongly to those directly associated with them on the surface. The surface molecules

exist in a state of tension. In this state of tension, they resist deformation.

129

Keller’s (1998) research illushated that surface tension was a critical food product
parameter in determining the threshold leak size and pressure (Keller et a1, 1998).

The surface tension of water is 72 dynes/cm at 25°C. This means that it would take a
force of 72 dynes to break a surface of ﬁlm of water 1 cm long. Potassium Phosphate,
Monobasic, Crystal (KH2P04) diluted with water is commonly referred to as Butterﬁeld’s
Buffer Solution. This was the carrier solution for the aerosolized E. coli. Butterﬁeld’s
Solution has a surface tension of 65 dynes/cm at 25°C, slightly less than water, making it
a slightly better wetting agent than water at 25°C. The better the wetting agent, the more
likely the solution is to enter the pinholes rather than bridging them with surface tension
(llttp://ll\'perpllysics. plly-astl'.g311.cdu, 2005). Therefore, a threshold pressure must be
exceeded to allow the aerosol droplets of microbes contained in Butterﬁeld’s Solution to
ingress into the PETG hay.

In the Inh'oduction of this dissertation, the inﬂuence of pressure differential
magnitude on microbial ingress was brieﬂy discussed. The discussion above regarding
microbial injury due to the extremely high calculated ﬂow rates which was theorized as a
conhibuting factor to the low CFU counts does seem to support the concept of
‘impaction’. This theory states that a ‘moderate’ ﬂow rate maximizes microbial
penehation. On the low side, diffusion is theorized to be the predominant mechanism of
penehation. On the high side, impaction predominates, the microbes are removed
because they are ‘smashed’ into porous ﬁbers, or in this research, where a non-porous
system was examined, may experience catash'ophic injury while being pulled into the
hole. According to this theory, maximum penetration will occur when neither impaction

nor diffusion is the predominant mechanism. This research suggests the concept that

130

high ﬂow rates may negatively impact penehation in nonporous systems as well as
porous systems. However, because this research did not investigate an intermediary
pressure differential, and was not intended to speciﬁcally address this question, ﬁrrther
research is recommended to conﬁrm and quantify this phenomenon.

With regard to one of the goals of this research, to enhance patient safety, a brief
discussion of the concept of microbial infective dose and zero tolerance is salient. E-Coli
K-12, a gram-negative, motile, shaight-rod that ranges in size ﬁom 1.1-1.5 um‘2.0-
6.0 um (Brooks et al., 2004) was the selected test organism. As discussed earlier, one of
the reasons it was selected for this research was that it would serve as a model of a
microbe frequently linked to burn infections, Pseudomonas aeruginosa, a curved rod that
ranges in size ﬁ'om 0.5-1.0um*1.5-5.0um (Brooks et al., 2004). The penehation of
E. coli models penehation of relevant pathogens.

Infective dose is the number of a particular microorganism required to cause disease
in a susceptible host (Brooks et al., 2004). The concept of zero tolerance is commonly
applied to food born disease agents. For some disease agents, zero organisms are
acceptable in some foods. In this research the amount of reduction noted ﬁom the
starting test load (CFUs present in the test system) and the amount of resultant microbial
ingress into positive test hays was signiﬁcant. It is recognized that for sterile medical
devices the presence of any organisms is not acceptable. However, the real world
situation could exist that while microbial ingress may occur in a package, that the
quantity of pathogens is insufﬁcient as an infective dose if it is not a zero tolerance
pathogen. The concept of infective dose with regard to the aggregate of organisms

typically found in a hospital environment is an area warranting future research.

131

Recommendations For Future Research

There are numerous variables which conhibute to the determination of the minimal
hole size which may result in microbial ingress in a sterile medical device package.
These include:
0 Duration of time that the package is exposed to the microbes
0 Package Geometry
o Porosity of Package Components
0 Thickness of Package Components
0 Rigidity of the Package at Test
0 Method of Microbial Challenge (Immersion, Aerosol, Dynamic, Static)
0 Pressure Differential across the Sterile Barrier
0 Types, Sizes, States, and Concenhation of the Challenge Organisms
0 Types of Package Defects and the Methods of Generating the Defects

This research provides an answer regarding one of these variables, pressure
differential. The research indicates that pressure differential is a signiﬁcant factor
conhibuting to microbial ingress; reference Chapter 5, Results, Table 20. Additional
research exploring the other variables is recommended, speciﬁcally, exploring the effect
of gravitational settling and Brownian Motion as the driving forces on microbial
penehation, the effect of static on microbial penetration, examination of additional hole
sizes, additional pressure differentials and an exploration of the impaction phenomenon
discussed in Chapter 1, Inh'oduction, the impact on the effect of pressure differential
under varying rates of inducement, the effect of pressure differential on seal defects, and

the impact of temperature variation.

132

APPENDIX I

133

APPENDIX 1
NEW METHOD DEVELOPMENT

Introduction

For reasons detailed in Chapter 1, Introduction, an alternative means to measure the
microbial penetration into packages was needed in order to obtain meaningful results for
the primary objective of this research, an examination of the effect of pressure differential
on the microbial ingress. In order to employ this new method, aseptically ﬁlling
packages with sterile agar and then incubating closed packages to see if any microbial
growth takes place (Figure 22), it was necessary to verify that the process of ﬁlling the

package did not signiﬁcantly inﬂuence microbe penetration.

    

Step 1 Step 2 Step 3 Step 4 Step 5
e,
0135,
t“""“‘“‘L 1: Mai ..
Using a 70%
Trays are .
sealed and isobpropylbalcohol Wdt the Place septum Using the same
sterilized swa - 5W8 taln 86’? disinfected septum (adhesrve srde swdrbmg process
appgﬁgmuasﬁig . m the ﬂame of a down) in the described in Step
disinfected forceps Bunsen burner to location that was 2. svivab the self-
remove backing - flash of any disinfected dunng sejenl‘ng septttlhmt
material from self- remaning 3(30110l step 2 an 9 area - a
surrounds it

sealing septum to

 

 

Trays are challenged Incubate at optimal

ASGDﬁCdW ﬂ" With Agitate the tray 10 with microbes (talc, condtions for the
sterile QFOWYh ensure an even immersion or aerosol) chdlenge organisms

medium througl a dstribution of agar and inspect for growth

self-sealing septum on the tray's "

bottom periodically

Figure 22 — Proposed Whole Package Microbial Challenge Test through Aseptic
Introduction of Sterile Growth Medium

134

A possible criticism of the new technique is that inh‘oduction of sterile agar through
the self-sealing septum will result in unwanted contamination and growth and, therefore,
false positives. As a result, an experiment was conducted to examine the following
working hypothesis:

Working Hypothesis

0 Aseptic introduction of the appropriate agar growth medium into a sterile package
will not result in unwanted contamination of the package (false positives).

Materials and Methods

Three hundred, eighty uncoated glycolized polyethylene terephthalate (PETG) hays
(Figure 23) were sealed with lid stock composed of 1073B Tyveko with heat seal coating
PTH-017 using a SenCorp MD 2420 dual-shuttle heat sealer. Heat sealer settings were
411 K, 517 kPa, with a dwell time of 2.5 seconds, and the trays were thermoformed ﬁom
stock that had a nominal thickness of 0.0635 cm. Nominal hay dimensions were 16.6624
cm x 12.8524 cm x 2.2225 cm. Sealed hays were sterilized using ethylene oxide (EtO)
by Sterigenics, Willowbrook, m, (EtO cycle 154, prior to test, system run number,

SWB_20041007_5268, sterilization date,10/07/04).

Figure 23 - Uncoated PETG Tray Provided by Perfecseal Sealed with 1073B
Heat Seal Coated TyvekO Provided by Amcor Flexibles

135

Four test populations were created for the experiment (Table 21). The design of the
study was based on the fact that using the haditional procedure (Figure 2); a small
random sample was unlikely to contain enough instances of the rare event
(contamination) to make reliable inferences. However, a sample size large enough to
ensure a reasonable number of rare events to reliably observe was prohibitively
expensive. To resolve this situation, trays were subjected to more exheme conditions
(injection occurred through a solution containing varying concenhations of microbes-
Table 21) than will be present when employing the proposed methodology. In doing so,
the methodology became more robust. This allowed sample-to-sample differences in
contamination levels to be examined and distinctions to be made between the method and

the randomness of nature, as well as to decrease the sample size.

 

Table 21 - Description of Test Groups

 

Test

. Quantity of
Group D 'ption Concenhahons Tested

Trays
(CFU/mL) Challenged

 

Aseptic Technique with NO alcohol
swabbing through contaminated
ﬁelds with varying concenhations 4 5 3

1 ofmicrobes (steps 5 and 8 in Figure None, 102’ 10 ’ 10 ’ 10 150 (30 per
22 are eliminated but injection
occurs through a contaminated
ﬁeld)

concenhation)

 

Aseptic Technique WITH alcohol
swabbing through contaminated
ﬁelds with varying concenhations

2 of microbes (Figure 22 procedure None, 102, 104, 106, 108 150 (30 per
step 8 is eliminated but injection
occurs through a contaminated
ﬁeld)

concenhation)

 

 

Full aseptic technique through a

3 non-contammated ﬁeld None 3 0

(Figure 22 procedure step 8 is
eliminated)

 

 

 

 

 

136

 

Table 21 (cont’d)

 

 

 

Test . Quantity of
Group Description Concenhahons Tested Trays
(CFU/mL)

# Challenged ,
Inject through varying
contaminated ﬁelds with no self-
sealing septum or aseptic technique

4 (positive control) (steps 2-5 and 8 None, 102, 10‘, 106, 108 igtfcleigion)
are eliminated ﬁom Figure 22
injection occurs through a

 

contaminated ﬁeld)

 

 

 

 

137

 

Table 22 - Test Group Procedure Description

 

Test
Group

Methodology

 

1. With a 70% Isopropyl Alcohol swab, swab an approximate 6.5 cm2 area in
the center of the lidstock (top) of the hay using moderate pressure for
approximately 30 seconds.

2. Using disinfected forceps, pick up a pre-sized piece of septum material and
remove the backing material.

3. Dip the septum material in a 70% Isopropyl Alcohol Bath.

4. Watt the disinfected septum in the ﬂame of a Bunsen burner to ﬂash off
any remaining alcohol

5. Still using forceps, place the septum material, adhesive side down in the
center of the disinfected area of the hay.

6. Firmly press with the forceps to ensure adhesion of the septum material.

7. Using a sterile pipette, transfer 051111 of microbe dilution from the stock
solution to the center of the septum material.

8. Aseptically draw up approximately 25 mls of equilibrated medium in a
sterile syringe using a 16 gauge needle ham 3 test tube. A 16-gauge needle
was used for this procedure to facilitate the drawing up of the viscous agar.

9. Replace the l6-gauge needle with a sterile 18-gauge needle. Change to an
18-gauge nwdle in this step was for two reasons; ﬁrst to minimize any
potential contamination and to minimize the size of the enhy port into
sterile hay.

10. Inject the entire syringe volume into the hay through the septum.

11. Immediately agitate the hay ensuring dishibution of agar on the tray
bottom.

12. Allow the medium to completely solidify undisturbed.

13. This process is repeated for each concenhation lot of 30 trays.

14. Incubate hays for 72 hours in a conholled environmental chamber set at
37°C +/-2°C; 50% RH. Visually inspect for growth at a time point
between 24 and 72 hours.

15. E. coli will exhibit growth after 24 hours. After 72 hours the medium will
likelybe too dry to accurately measure growth.

 

 

 

This test group is identical to Test Group One; with the following exception.
There is an additional step after Step 7. Prior to the injection of the medium
through the septum, the same swabbing process employed in step 1 is
employed on the contaminated septum material. Once this is completed, the
medium is injected into the hay as described in Test Group One.

 

138

 

 

Table 22 (cont’d)

 

Test
Group Methodology

 

This group is procedurally identical to Test Group Two with the exception of
3 the steps associated with exposure to microbial contamination. This group was
not exposed to any microbial contamination.

 

 

This group procedurally is identical to Test Group One with the exception that
the self-sealing septum material was not used. The sterile medium was
4 injected through the contaminated ﬁeld directly through the tray lidding
material. No effort was made to cover the approximately 1mm hole left in the
lidstock ﬁom the needle. No aseptic technique employed.

 

 

 

Three test groups and a positive conh'ol were examined (T able 21). The ﬁrst two test
groups were grossly contaminated (with a solution of varying concenhations of
microbes) on the surface of the package lidstock to create a situation that is more hostile
than “normal testing” to provide a worst case environment. The microbial dilutions were
added to the injection ﬁeld prior to the placement of the self-sealing septum for Test
Groups 1 and 2 (Table 21). As discussed, injecting through these dilutions allowed for
fewer samples to be used.

The third test group represented the proposed test methodology without the microbial
challenge (Figure 22, less step 8). Trays were purposefully not contaminated and aseptic

ﬁlling was conducted after the injection site was swabbed with alcohol.

139

The fourth group consisted of positive controls. Trays were grossly contaminated, no
septum material was used and no aseptic technique alcohol swabbing was done prior to
the introduction of the sterile agar. The positive control was included to indicate:

1. The agar can support growth under the conditions of the study

2. The E. coli strain used was viable.

The organism used for Test Groups 1, 2, and 4 was Escherichia coli, ATCC
Number 29181. Media Preparation was performed per standard methods described in
Current Protocols in Molecular Biology (Volume 1; Section 1 ESCHERICHIA COLI;
Media Preparation and Bacteriological Tools; Solid Media Page 1.1.3). The Escherichia
coli was reconstituted and cultured as described in this same source (Section 1.2; Growth
in Liquid Media). The microbe dilutions were prepared as described in this same source
(Section 1.3; Growth on Solid Media; Basic Protocol One ‘Titering and Isolating
Bacterial Colonies by Serial Dilutions’). The Butterﬁeld’s stock solution used for the
dilutions and the application of the microbes to the test trays was prepared by adding
34 grams of Potassium Phosphate, Monobasic, Crystal (KH2PO4); J. T. Baker
CAS 7778-77-0; to 500 mls of distilled water. After preparation of the stock solution, the
pH was adjusted to 7.2 with 1N Sodium Hydroxide (NaOH). The volume was then
brought to 1 liter with distilled water. The solution was then autoclaved for 15 minutes at
121°C.

Sample hays were incubated at 37°C and 50 percent RH in a conh'olled
environmental chamber for no longer than 72 hours. Trays were inspected for growth
between 24 hours and 72 hours. Counting of the colony forming units (CFUs) was

performed visually using a Quebec Colony Counter. Figure 24 depicts a hay exhibiting

140

growth. Counting was performed in accordance with the plate counting rules and

guidelines described in ‘Modern Food Microbiology’ (4m Edition).

 

Figure 24 - Tray Exhibiting Growth as Seen Using the Quebec Colony Counter

141

Results

 

Table 23 - Test Group 1- Aseptic Technique with NO Alcohol Swabbing Through
Contaminated Fields with Varying Concenhahons of Microbes (Steps 5 and 8 in
Figure 22 are Eliminated but Injection Occurs Through a Contaminated Field)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tray CFU/mL Tray # CFU/mL
*** 0 10r 10‘ 111T 10tr 0 10’ 10‘ 10‘ 108
1 0 0 0 0 68 16 0 0 0 1 11
2 0 0 0 3 TNTC” 17 0 0 0 0 121
3 0 0 0 1 177 18 0 0 0 0 22
4 0 0 0 0 0 19 0 0 0 0 136
5 0 0 0 0 78 20 0 0 0 0 93
6 0 0 0 0 125 21 0 0 0 2 0
7 0 0 0 0 TNTC" 22 0 0 0 1 7
8 0 0 0 0 TNTC“ 23 0 0 0 0 30
9 0 0 0 0 TNTC“ 24 0 0 0 0 l
10 0 0 0 6 3 25 0 0 0 0 TNTC“
11 0 0 0 0 0 26 0 0 0 0 174
12 0 0 0 1 21 27 0 0 0 4 39
13 0 0 0 0 69 28 0 0 0 5 103
14 0 0 0 0 66 29 0 0 0 0 40
15 0 0 0 5 86 30 0 0 0 2 96

 

*TNTC = Too numerous to count

***Incubation Periods:

0
102
10‘
10‘
108

58 hours
44.5 hours
33 hours
33 hours
33 hours

142

 

 

Table 24 - Test Group 2 - Aseptic Technique WITH Alcohol Swabbing Through
Contaminated Fields with Varying Concentrations of Microbes (Figure 22
Procedure Step 8 is Eliminated But Injection Occurs Through a Contaminated Field)

 

 

 

 

Tray CFU/mL Tray CFU/mL

1‘: 0 10’ 10‘ 10‘ 10" # 0 102 10‘ 10‘ 10“
1 0 0 0 0 0 16 0 0 0 0 0

2 0 0 0 TNTC“ 0 17 0 0 0 0 TNTC“

 

3 0 0 0 TNTC“ TNTC“ 18 0 0 0 0 TNTC“

 

40000 0190001 0

 

5 O 0 0 TNTC“ TNTC“ 20 0 0 0 TNTC" 0

 

 

 

 

6 0 0 0 0 TNTC" 21 0 0 0 0 * "‘
7 0 0 0 ** TNTC“ 22 0 0 O 0 0
8 0 0 0 O TNTC" 23 0 0 0 0 0
9 0 0 0 0 TNTC“ 24 0 0 0 0 85

 

10 0 0 0 0 TNTC“ 25 0 0 0 4 TNTC*

 

ll 0 0 0 0 TNTC" 26 O 0 0 0 TNTC“

 

 

 

 

12 O O 0 0 TNTC" 27 0 0 0 0 0
l3 0 ** 0 0 0 28 0 0 0 0 TNTC“
14 0 0 0 0 0 29 O 0 0 0 **
15 0 0 0 0 TNTC“ 30 0 0 ** ** **

 

 

 

 

 

 

 

 

 

 

 

 

 

 

*TNTC=Too numerous to count
MSample Not Available-Insuﬂicieut Medium

***Incubation Periods:
0 40.5 hours
102 40.5 hours
10‘ 40.5 hours
10‘ 41 hours
108 41 hours

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 25 - Test Group 3 (CONTROL) - Full Aseptic Technique
Through a Non-Contaminated Field (Figure 22 Procedure Step 8 is
Eliminated)

Tray # 0 CF U/mL Tray # 0 CFU/mL
1 0 16 0
2 0 l7 0
3 0 18 O
4 0 19 0
5 0 20 0
6 0 21 0
7 0 22 0
8 0 23 0
9 0 24 0
10 O 25 0
11 0 26 0
12 0 27 0
l3 0 28 0
14 0 29 0
15 0 30 0

 

***Incubation Period: 38.5 hours

144

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 26 - Test Group 4 (POSITIVE CONTROL) -
Inject Through Varying Contaminated Fields with no Self-Sealing Septum or
Aseptic Technique (Positive Conh'ol) (Steps 2-5 and 8 are Elrm' inated from
Figure 22 Injection Occurs Through a Contaminated Field)
Tray # CFU/mL
m 0 102 10‘ 10‘ 108
1 0 0 3 26 TNTC“
2 0 0 0 165 186
3 0 0 0 TNTC“ TNTC"
4 0 0 0 3O TNTC“
5 0 0 0 29 122
6 0 0 0 48 TNTC“
7 0 0 0 24 TNTC“
8 0 0 0 10 TNTC“
9 0 0 0 32 TNTC"
10 0 0 0 123 TNTC“
*TNTC=Too Numerous to Count
***Incubation Periods:
0 27 hours
102 27 hours
10‘ 27 hours
10‘ 27 hours
108 27 hours

As established in Chapter 3, Materials and Methods, by inh'oducing the “gross
contaminated area,” conﬁdence in the methodology can be gained with fewer samples.
For each heahnent group and each concentration of 10", the proportion (rate) of hays
exhibiting microbial growth was calculated. Logistic regression was used to model the
relationship of rate to the concenhation 10", for each treahnent group. Fisher’s exact test
was used to assess the statistical signiﬁcance of group differences at speciﬁc

concenhations.

145

Discussion

The regression estimates smooth the raw data that are found in Tables 13, 14, and 16.
The tables illustrate that in Groups 1 and 2, no microbial growth was observed at
concenhations below 106 CFU/mL. In Group 4, no microbial growth was observed at
concentrations below 104 CFU/mL. This indicates that the contamination threshold lies
between a concenhation of 104 CFU/mL and 106 CFU/mL.

At a concenhation of 10‘5 CFU/mL, the rates of positive hays for the Groups 1 and 2
are 11/30 and 6/28. The one-sided, p-value of the Fisher-Exact Test of this difference is
p = 0.16 (not below the 0.05 standard for statistical signiﬁcance). Conclusion: No
statistically signiﬁcant effect from swabbing with alcohol.

At a concenhation of 108 CFU/mL, the rates of positive hays for the Groups 1 and 2
are 27/30 and 16/27. The one-sided p-value of the Fisher-Exact Test of this difference is
p = 0.008 (below the 0.05 standard for statistical signiﬁcance). Conclusion: There is a
statistically signiﬁcant effect ﬁ'om swabbing with alcohol.

In total, there is a statistically signiﬁcant eﬁ'ect from swabbing with alcohol, ie,
employing full aseptic technique (Steps 2-5 of Figiue 22) in the expected direction; that

is, reducing contamination.

146

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Logistic Regression Estimates
1.0
a Group 4 // /
it 0.8 -«~--——— A--- v——»
1— / Grainy
o
33 0.6 -________,L_ m- * ~ ~ /
.6 / /
3 0.2 ~ ~~~~~~ — -- ,
Group2
0.0
0 2 4 6 8
CFUIml-Mlcrobiologlcal Challenge
(0 I zero, 2 8 10*‘2, 4 8 10“4, etc)

 

 

 

Figure 25 - Logistic Regression Estimates for Groups 1, 2, and 4
(microbial concentrations for Group 3 did not vary)

Conclusion
Working Hypothesis:
0 Aseptic introduction of appropriate agar growth medium into a sterile package will

not result in unwanted contamination of the package (false positives).

The results indicate that there is no evidence to the conh'ary regarding Working
Hypothesis 1. Therefore, the method does appear to eliminate the shortcomings of the
microbial challenge test described in Chapter 1. This method shows promise as a viable

alternative to haditional whole-package microbial challenge testing.

147

BIBLIOGRAPHY

148

BIBLIOGRAPHY

(2000). "Acoustic Micro Imaging Speeds Development of Blood Tester." Pharmaceutical
and Medical Pack_agrr_r' g News.

‘3M customer centers. (1999). "3M Pehiﬁlm Aerobic Count Plates." Retrieved
November 18, 2002, ﬁom
http://solutions.3m.com/wps/portal/_s. 155/84505/_s.155/92845.

*Barrier Test Consortium, L. New Physical Particulate Barrier Test for Porous Medical
Packaging Materials Fact Sheet, ESPA.

*PTI Press Release. "New pti-525 ABUS non-desh'uctive inspection system detects
quality of heat seals before failure to improve process control and eliminate customer

complaints." Rehieved November 18, 2002, ﬁ'om
httpz/MWW.kineticcommunication.com/Clieant/PTl-rcl.lltml.

*Public Employees Occupational Safety and Health Program. (1997). "Indoor
Bioaerosols." from http://www.state.ni.us/healtlh/eoh/peoshweb/bioacro.htm.

Adams, T. (2001). "Seeing the bond: Acoustic imaging of seals in medical carhidges."
Sealing Technology: 7-10.

Adams, T. (2003). "No Need To Break The Seal." Tegtmg' Supplement: 2-3.

Ahl, T., N. Dalen, et al. (1995). ”Air contamination during hip and knee arthroplaties:
Hoizontal laminar ﬂow randomized vs. conventional ventilation." Acta Orthomﬂ'ca
Scandinavica 66(1): 17-20.

Ahvenainen, R., T. Mattila-Sandholm, et al. (1992). "The effect of microhole size and

foodstuff on the microbial integrity of aseptic plastic cups." Paclgagrg' g Technology
and Science 5: 101-107.

Allen, D. (2002). "Package Integrity Testing." Pharmaceutical & Medica_l_lﬂag°mg
News: 32-38.

Allen, D. (2003). "No small debate when it comes to defect size." Pharmaceutical and
Medical Pacl_t_agig_1g News 11(5): 10.

Allen, D. (2003). "Setting the Course for Packaging." Pharmaceutical and Medical
Pack_agm’ g News: 1-2.

Allen, D. (2003). "Testing Traditions." Pharmacequ & Medical Pacl_c_ag1_1_1° g News: 38-
40.

149

Amini, M. A. and D. R Morrow (1979). "Leakage and permeationztheory and practical
applications." Pack_age Dev. a_ngl 8178.: 20-27.

Anderson, G. (1989). Leak Testing. Noanesh'uctive evaﬂation and quality control: may
handbook. Arlington, VA, AOAC: 50-57.

Anema, P. J. and B. L. Schram (1980). "Prevention of post-process contamination of
semi-rigid and ﬂexible containers." Journal of Food Protection 43(6): 461-464.

ASTM D 4169-05 Standard practice for performance testing of shipping containers and
systems. In ASTM International (Ed.).

ASTM D 6653-01: Standard test method for determining the effects of high altitude on
packaging systems by vacuum method. In ASTM International (Ed).

ASTM D3078-02 Standard Test Method for the Determination of Leaks in Flexible
Packaging by Bubble Emission. American Society of Testing and Materials, 2005.

ASTM D4991-94(l999) Standard Test Method for Leakage Testing of Empty Rigid
Containers by Vacuum Method. American Society of Testing and Materials, 2005.

ASTM F2338-05 Standard Test Method for Non-Desh'uctive Detection of Leaks in
Packages by Vacuum Decay Method. American Society of Testing and Materials, 2005.

ASTM F2391-05 Standard Test Method for
American Society of Testing and Materials, 2005.

ASTM F2228-02 Standard Test Method for Non-Destructive Detection of Leaks in
Medical Packaging Which Incorporates Porous Barrier Material by C02 Tracer Gas
Method. American Society of Testing and Materials, 2005.

ASTM F1886-98(2004) Standard Test Method for Determining Integrity of Seals for
Medical Packaging. American Society of Testing and Materials, 2005.

ASTM F2096-04 Standard Test Method for Detecting Gross Leaks in Medical
Packaging by Internal Pressurization (Bubble Test). American Society of Testing and
Materials, 2005.

ASTM Fl929-98(2004) Standard Test Method for Detecting Seal Leaks in Porous
Medical Packaging by Dye Penehation. American Society of Testing and Materials,
2005.

ASTM F2095-01 Standard Test Method for Pressure Decay Leak Testing for Non-Porous

Flexible Packages With and Without Reshaining Plates. American Society of Testing
and Materials, 2005.

150

 

ASTM F1327-05 Standard Terminology Relating to Barrier Materials for Medical
Packaging. American Society of Testing and Materials, 2005.

Ausubel, F.,R. Brent, R. Kingston, D. Moore, J .G. Seidman, J. Smith, K. Sh'uhl. Current
Protocols in Molecular Biology; Volume 1 (2005); Sections 1.1, 1.2 and 1.3.
Massachusetts General Hospital and Harvard Medical School

Axelson, L., S. Calvin, et al. (1990). "Aseptic integrity and microhole determination of

packages by electolytic conductance measmements." Packggrg' g Technology and
Science 3: 141-162.

Bankes, P. and M. F. Shinger (1988). "The design and application of a model system to
investigate physical factors affecting container leakage." Int. J. Food M133; 6: 281-
286.

Barlow, W. (2004). Evaluation of seal integrity of ﬂexible food polyhays by dee
and non-desh'uctive techniques, University of Tennessee, Knoxville. Masters of Science:
114.

Bausum, H. T., S. A. Schaub, et al. (1982). "Comparison of coliophage and bacterial
aerosols at a wastewater spray irrigation site." Appl. Environ. Microbiol. 43(1):
28-38.

Beagley, K. (1998). "Packaging Testing: Is Standardization on the Horizon."
Pharmaceutical_ and Medical Packggm' g News Rehieved September 20, 2002, ﬁ'om
http://\4MW.deV-'iceli11k.com/pmpn/arcllive.

 

Berens, M. J. (2002). "Drug-resistant germs adapt, thrive beyond hospital walls."
Rehieved September 29, 2002, from
httpz/[pqilsbpqarchivercom/cllicagotribune/index.lltml?ts=l 1 18693603.

 

Berens, M. J. (2002, July 21, 2002). "Infection epidemic carves deadly path."
Investiggtion: Unhealthy hoﬂﬂg from

h11DI//\W’W.C111C£1g01rlblln€.C0m/n€WS/SDCCltllSkill-02072l027311121.SlOI"\j‘?llUll.

Bitkover, C., E. Marcusson, et al. (2000). "Spread of coagulas-negative staphylococci
during cardiac operations in a modern operating setting." Ann_als of ThoLacic Surggy
69(4): 1110-1115.

Bix, L., O. Kassarjian, R. Iwaszkiewicz, J. Severin (August 2004) Examining Defects of
Various Sizes in Device Packages. Business Brieﬁng: Medical Device
Manufacturing and Technology. 72-77.

Blakistone, B. and C. Harper (1995). New developments in sea] integrity testing. Plastic

pack_age integg'gy mm : Assuring seal gualigy. I. o. P. Professionals and F. P.
Institute. Washington, DC: 1-10.

151

Blakistone, B. A., S. W. Keller, et al. (1996). "Contamination of ﬂexible pouches
challenged by immersion biotesting." Journal of Food Protection 59(7): 764-767.

Bohn, W., D. Mickinsey, et al. (1996). "The effect of a portable HEPA-ﬁltered body
exhaust system on airborne microbial contamination in a conventional operating
room." Infection Conhol and Hospital Epidemiology 17(7): 419-422.

Bovallius, A., B. Bucht, et al. (1978). "Long-range air transmission of bacteria." A l.
Environ. Microbiol. 35(6): 1231-1232.

Boyce, J. and D. Pittet (2002). "Guideline for Hand Hygiene in Health-Care Settings."
Morbidity and Mortﬁalitv Weekly RM 51(16): 1-44.

Broad, N., P. Graham, et a1. (2002). "Guidelines for the Development and Validation of
Near-inﬂated Spectroscopic Methods in the Pharmaceutical Industry." Handbook of

Vibrational Smscopy.

Brooks, G.F., J.S. Butel, S.A. Morse (2004) Jawetz, Melnick, and Adelberg's Medical
Microbiology. 23rd Edition

Bryant, M. (1988). "Package failures: quality in design doesn't end with the ﬁnished
produc " Medicﬂevice and diggpostic indm 10(8): 30-33.

Burgess, G. (2003) Coursework PKG 805 Advanced Packaging Dynamics. Michigan
State University, School of Packaging.

 

Butler, L. D., J. J. Coupal, et al. (1978). "The detection of ampul leakers using short-lived
radionuclides." Journal of Pgrentergl Drug Association 32(1): 2-9.

CALCE. (2001). "Non-Desh'uctive Evaluation: Inﬁared Imaging." Retrieved November
18, 2002, from http://www.calce.umd.edu/general/Facilities/isishtm.

 

Center for Disease Conhol and Prevention (CDC). (2002). "Antimicrobial Resistance:
Vancomycin- Intermediate/ Resistant Staphylococcus aureus." Rehieved March 13,
2003, from http://mvw.cdc.gov/ncidod/hﬂaresist/Visahtm.

 

Center for Disease Control and Prevention (CDC). (2002). "Appendix: Antimicrobial
Spech'um and Characteristics of Hand-Hygeine Antiseptic Agents." Rehieved
February 12, 2003, ﬁ'om

http://www.cdc.gov/mmwr/prcview/mmwrlllml/rrSl 16212.11tm.

 

Center for Disease Conhol and Prevention (CDC). (2002). "Hand Hygeine Guidelines
Fact Sheet." Rehieved February 12, 2003, from
hllpI//W\VW.CC1C.gOV/Od/OC/mCCIla/DFCSSI'Gl/I‘SOZl025111111.

 

152

Center for Disease Control and Prevention (CDC). (2003). "Antimicrobial Resistance:
Methicillin Resistant Staphylococcus aureus Fact Sheet." Rehieved March 13, 2003,
from http://W\\~'w.cdc.gov/ncidod/hip/aresist/mrsat‘athtm.

 

Center for Drugs and Biologics, Center for Devices and Radiological Health, et al.
(1987). Guideline on general principles of process validation. Washington, DC.

Chen, C., B. Harte, et al. (1991). "Assessment of package integrity using a spray cabinet
technique." Joumpl of Food Protection 54(8): 643-647.

Clark, S., R. Rylander, et al. (1983). "Airborne bacteria, endotoxin and ﬁrngi in dust in
poulhy and swine conﬁnement buildings." American Ind. Hyg. Assoc. J. 44(7): 537-
541.

Clemens, A. (1993). "The Safe Medical Devices Act." National Medical-Legal Journal
4(3): 6-7.

Cutnell, J. (1995). Physics. New York, John Wiley and Sons Inc.

Davidson, P. M. and I. J. Pﬂug (1981). "Leakage potential of swelled cans of low-acid
foods collected ﬁ'om supermarkets." Journal of Food Protection 44(9): 692-695.

Demorest, R. (1988). "Non-desh'uctive leak detection of blister packs and other sterile
medical packages." oumal of Packagmg’ Technology 2(5): 182-184,1 86,188-190.

Dimmick, R. L., H. Wolochow, et al. (1979). "Evidence for more than one division of
bacteria within airborne particles." Appl. Environ. Microbiol. 38(4): 642-643.

DuPont Medical Packaging (2002). Technical reference guide for medical packaging. E.I.
du Pont de Nemours and Company, DuPont: 1-33.

Ehrman, J. "Drug-resistant bacterial: Old foes, new faces." HEALTHWise report
January, 1998. Rehieved October 1, hour
http://www.nih.gov/news/HealthWise/Jan98/storyS.htm#'l‘Ol’.

 

Eisenberg, D. and D. M. Crothers (1979). thsicafl chemi_strv with applications to the life
sciences. Menlo Park, CA, Benjamin/Cummings Publishing 00.: p.726.

Ezzedine, H., P. Baele, et al. (1991). "Bacteriologic quality of intraoperative
autohansfusion." Journal of Sggeg 109(3 pt 1): 259-264.

F eagin, J. A. (1979). "Bacteriology of the operatiing room with the use of helmist
aspiration systems." 114: 790-792.

F eliu-Baez, R. (2001). Pharmaceutical and Medigal Device Pacl_(_agir_1g Overview. Forum
for Innovation in Pharmaceutical and Packaging Technology, San Juan, PR.

153

 

Floros, J. (1994). "Critical leak size and package integrity." PTT, Pack_agipg Technologies
and Inspection News IV.

Food and Drug Adminishation (1997). Parts 808, 812 and 820 Medical Devices; Current
Good Manufactming Practice (CGMP); Final Rule. 21CFR.

Food and Drug Adminishation (2002). Part 58; Good Laboratory Practice For
Nonclinical Laboratory Studies. 21 CFR.

Food and Drug Administration (Revised- By the Modernization Act of 1997). Chapter H-
Deﬁnitions. Federal Food Md Cogmetic Act.

Food and Drug Adminishation (Revised 2002). 21CFR820.1: Quality System Regulation.
Washington, DC.

Franks, S. (2002a). "Strength and Integrity, Part Two: Basics of Seal-Strength Testing."
Pharmaceuticgl & Medical PacLagipg News: 52-53.

 

Franks, S. (2002b). "Shength and Integrity: The Basics of Medical Package Testing."
Pharmacemal & Medical Pack_agipg News: 28-34.

Franks, S. (2003). "Strength and Integrity, Part Three: Packaging Integrity Testing."
Pharmaceutical & Medical Pacl_c_agmg° News: 41-43.

Frazier, C., A. Ozguler, et a1. (2000). "High conh'ast ulhasound images of defects in food
package seals." IEEE Transactions on ulhasonicp, ferroelechics, and fr_’e_quency
control 47(3): 530-540.

Freiherr, G. (1998). "Two Packaging Standard Give Manufacturers the Ability to Pick
and Choose." Medicarl device 24nd diggpostic ind_us_h'y 20(1): 94.

Friberg, B., L. Burman, et al. (1998). "Zoned exponential, vertical and horizontal ulha-
clean laminar airﬂows. No differences in bacteriological efficiency." Acta
Orthomljca Scandinavica 69(2): 169-172.

Friberg, B., L. Burman, et al. (2001). "Surgical area contamination." Journg of HospiLal
Infection 47(2): 110-115.

Friberg, B., S. Friberg, et a1. (1996). "Zoned vertial ulhaclean operating room ventilation.

A novel concept making long side walls unnecessary." Acta Orthopgedjca
Scandinavica 67(6): 578-582.

Friberg, B., S. Friberg, et al. (1999). "Correlation between surface and air counts of

particles carrying aerobic bacteria in operating rooms with turbulent ventilation: an
experimental study." ournal of Hospital Infection 42(1): 61-68.

154

Friberg, B., S. F riberg, et al. (1996). "Inefﬁciency of upward displacement operating
theahe ventilation." Journal of Hospital Infection 33(4): 263-272.

Friberg, B., M. Lindgren, et al. (2002). "Mobile zoned/ exponential LAF screen: a new
concept in ulha clean air technology for additional operating room ventilation."
Jorunal of Hospital Infection 50: 286-292.

Gilchrist, J. E., U. S. Rhea, et al. (1985). "Helium Leak test for micron-sized holes in
canned foods." Journal of Food Protection 48(10): 856-860.

 

Gilchrist, J. E., D. B. Shah, et a1. (1989). "Leak detection in ﬂexible retort pouches."
Journal of Food Protection 52(6): 412-415.

Gilliland, D., Bix, L., Lockhart, H., & Fotis, N. (2004). Sterile medical device packaging:
Statistics and sample sizes in operational qualiﬁcations. Medical device and
diagnostic industry.

Global Harmonization Task Force (GHTF) (1999). Process Validation Guidance.

Gnanasekharan, V. and J. Floros (1995). A theoretical perspective on the minimum leak
size for package integrity evaluation. Plastic e inte ' testin : Assurin seal
quality. B. Blakistone and C. Harper. Hemdon, VA, Institute of Packaging
Professionals . Food Processors Institute: 55-65.

Gold, D. (1996). "Validation: Why, what, when how much." JPDA Journal of
Pharmaceutical Science app TeQLnolggy 50(1): 55-60.

Goode Jr., J. and A. Souter (1995). "Rancidity in packaged nuts: end-use technical
service." J. Plague Film ﬂl Sheeting 11(7): 235-247.

Guazzo, D. M. (1994). Package Integrity Testing. Parenteral Quality conhol: sterility.

'culate and e inte ' testin . New York, Mercel Dekker:
247-276.

Hackett, E. (1996). "Dye penehation effective for detecting package seal defects."
Pack_agipg Technology and Engipeering.

Hackett, E. (2001). Effect of pinholes on §terile barrier mpﬂ'es. HealthPack,
St. Petersburg, Florida.

Hackett, E., M. Scholla, et al. (2000). "How small is small? The effects of pinholes on
medical packages." MedicaLlevice 1nd dipgpostic indm: 52-66.

155

 

Hansen, J ., Jones, L., Anderson, H., Larsen, C., Miller, H., Scholla, M., Spitzley, J., &
Baldwin, A. (1995). In quest of sterile packaging: Part I Approaches to package
testing. Medical device and diagnostic industry, 1 7(8), 56-61

Harper, Carol L. (1995) A Microbial Challenge Procedure for Identiﬁcation of Defective
Flexible Plastic Pouch Seals. Plastic Package Integrity Testing-Assuring Seal
Quality. Institute of Packaging Professionals and Food Processors Instititute.

Howard G, Duberstein R. A case of penehation of 0.2 micron rated membrane ﬁlters by
bacterial. Journal of the Parenteral Drug Association. 1980;34(2):95-102.

Hurme, E.U., and G. Wirtanen, L. Axelson-Larsson, N.A.M. Pachero, and R.
Ahvenainen. 1997. Penetration of bacteria through microholes in semirigid aseptic
and retort packages. Journal of Food Protection. 60(5):520-524

ISO 11607-2003. International Organization for Standards.

Jones, L., Hansen, J., Anderson, H., Larsen, C., Miller, H., Scholla, M., Spitzley, J., &
Baldwin, A. (1995). In quest of sterile packaging: Part 2 Physical package integrity
test methods. Medical device and diagnostic industry, 1 7(9), 81-85.

Kamei, T., J. Sato, Y. Nakai, A. Natsume and K. Noda.(1991). Microbiological Quality
of Aseptic Packaging and the Effect of Pin-holes on Sterility of Aseptic Products.
Packaging Technology and Science. (4): 1 85-1 93.

Keller, S. 1998 Determination of the leak size critical to package sterility maintenance.
PhD Dissertation, Virginia Polytechnic Institute and State University, Blacksburg,
VA.

Keller S, Marcy J, Blakistone B, Lacy G, Hackney C, Carter W. Bioaerosol exposure
method for package integrity testing. Journal of Food Protection. 1996;89(7):768-
771 .

Kutner, M.,C. Nachtsheim, J. Neter, w. Li (2005) Applied Linear Statistical Models, 5th
Edition.

Lake, D.D., R.R. Graves, R.S. Lesniewski, and J.E. Anderson. 1985 Post-processing
spoilage of low-acid canned foods by mesophilic anaerobic sporeformers. Journal of
Food Protection. 48(3):221-226.

Lampi R. Retort pouch: The development of a basic packaging concept in today's high
technology era. Journal of Food Process Engineering. 1980;4:1-18.

Leventon, W. (2001). Package testing: A touchy subject. Medical Devicelink. Rehieved
December 19, 2002, hour the World Wide Web:
http://www.devicelink.com/pmpn/archive/O1/09/0021111111

 

156

 

Linz, J. (2000, September 19). The current state of food pathogens. Paper presented at the
PKG 992, School of Packaging, Michigan State University.

Littell, MA. 2004. Targeting Hospital-Based Infections. UMDNJ Magazine. 1(1).

Marcy, Joseph E. (1995) Integrity Testing and Biotest Procedures for Heat Sealed
Containers. Plastic Package Integrity Testing-Assuring Seal Quality. Institute of
Packaging Professionals and Food Processors Instititute.

Maimder, D.T., J .F. Folinazzo, J .J . Killoran. (1968) Bio-test Method for Determining
Integrity of Flexible Packages of Shelf-Stable Foods Food Technology. 22(5): 81-
84.

McEldowney, S., and M. Fletcher. 1990. The effect of physical and microbiological
factors on food container leakage. Journal of Applied Bacteriology. 69: 190-205.

Morton, D., N. Lordi, et al. (1989). "Quantitative and mechanistic measurements of
container/ closure integrity: bubble, liquid, and microbial leakage tests." J. Parenteral
Sci. Technol. 43: 104-108.

National Nosocomial Infection Suveillance (NNIS) (2001). "National Nosocomial
Infection Suveillance System Report, Data Summary from January 1992-June 2001 ,
Issued August 2001." NNIS Reoort 29(6): 404-421.

Nave, C. (2000). "Mean ﬁee path and reﬁnement of mean ﬁee path." from

http://\nvw.230nscl.phy-astrgsuedu/hbase/kinetic/menfrehtml.

 

Nolan, P. "Package integrity test methods." Rehieved 9/25/2002, 2002, ﬁom
http://wvwv.wmrc.com/businessbricling/pdf/medicaldevice200 l /technologv/D DLirdf.

Nolan, P. (2001). "Package integrity test methods." 2002, from
http://www.wmrc.com/businessbiefl ng/pdemedicaldeviceZOO l ,I’Tecllnology/DDL.pd f.

Packaging Technologies and Inspection LLC (pti). (2002). "Package Integrity Testing-
Leak testing." Rehieved November 11, 2002, ﬁom
http://wwxmrtipacktechcom/packageintegritycfm.

Parisi, A., H. Miller, et al. (1993). "Toward a new consensus on sterile device
packaging." Medich device and diggpostic indpm 15(1): 84-87.

Pereira, L., G. Lee, et al. (1990). "The effect of surgical handwashing routines on the
microbial counts of operating room nurses." American Journal of Infection Conhol
18(6): 354-364.

 

157

 

Pﬂug, I. J ., P. M. Davidson, et al. (1981). "Incidence of canned food spoilage at the retail
level." J. Food Prot. 44(9): 682-685.

Pich, J. (1972). "Theory of gravitational deposition of particles from laminar ﬂows in
channels." J. Aerosol Sci. 3: 351-361.

Pilchick, R. (1999). "Material for Preformed Rigid-Tray Packages." Medical Device
Technology 10(3): 20-24.

Placencia, A., M. Arin, et al. (1988). "Physical tests are not enough." Medical device app
diggpostic indﬂy 10(9): 72, 74-78.

Placencia, A., G. Oxborrow, et a1. (1986). "Package integrity methodology for testing the
biobarrier of porous packaging. part 1]: FDA exposure-chamber method." Medical
device grid dipgnostic indugg 8(4): 46-53.

Potter, Merle C. and Foss, John F. (1975). Fluid Mech_anics: 59-51.

Purohit, Kailash S. Biotesting of Plastic Containers. (1995). Plastic Package Integrity
Testing-Assuring Seal Quality. Institute of Packaging Professionals and Food
Processors Instititute.

Put, H., H. Witvoet, et al. (1972). "Mechanism of microbiological leaker spoilage of
canned foods: a review." J. Appl. Bacteriol. 35: 7-27.

Put, H., H. Witvoet, et al. (1980). "Mechanism of microbiological leaker spoilage of
canned foods: biophysical aspects." J. Food Prot. 43(6): 488-497.

Raum, K., A. Ozguler, et al. (1998). "Channel defect detection in food packages using
integrated backseatter ulhasound imaging." IEEE Transaction_s on ulha_sonicg
ferroelechics, and ﬂuency conhol 45(1): 30-40.

Reich, R R. (1985). "A method for evaluating the microbial barrier properties of intact
packages." Medica_Ldevice and diggpostic indghy 7(3): 80-88.

Reich, R. R. (1986). "A method for evaluating the microbial barrier properties of intact
packages." Medical Diggpostics Indm 7(3): 80, 82, 86-88.

Resonetics, Laser Machining Group. (1999) Excimer Lasers Drill Small Holes in Plastic
Packages and Metal Containers. http://www.resonetics.com

 

 

Ritter, M. (1999). "Operating room environment." Clinical Ortho 'cs and Related
Research 369: 103-109.

Rose, D. (1994). Micrommn pgnneation tegpg' . Conference on Plastic Package
Integrity Testing: Assuring Seal Quality, Oak Lawn, IL.

158

 

Rosen, Michael (2005). Medical Device Indushy Growing in Importance in Midwest.
Wisconsin Technology Network Rehieved August 10, 2005
http://wistechnology.com.

Rubin Press Release. (1999). "Hospital Stays Can Be Harmﬁrl for Your Health."
Retrieved October 1, 2002, from http://w1mv.cdc.gov/ncidod/ ElD/press _r_/rubin.htm.

 

Rubin, R, C. Harrington, et al. (1999). "The Economic Impact of Staphylococcus aureus
infection in New York City hospitals." Eme_rgipg Infectious Disease 5(1): 9-17.

Schierholz, J. and J. Beuth (2001). "Implant infections:a haven for opportunistic." 1h;
ioumal of hospital infection 49(2): 87-93.

Schmidt, E.A., D.L. Coleman, and GP. Mallison (1984) Improved System for Floor
Cleaning in Health Care Facility. Journal of Applied Environmental Microbiology.
47:942-946.

Schneider, P. (1980). "Microbiological evaluation of pmkage and packaging material
integrity." 2(5): 29-37.

Scholla, M. Elucidation of the Mechanism for Microbial Barrier Effectiveness, Dupont:
1-30.

Scholla, M. F DAMA & HIMA's New Focus on Standards, Dupont: 1-18.
Scholla, M. Global Medical Packaging Update, Tyvek Medical Packaging: 1-84.
Scholla, M., E. Hackett, et al. (1999). "The effects of radiation sterilization on the

microbial barrier properties of Tyvek." Rehieved November 20, 2002, ﬁ'om
http://Vi~w~wv.tyvek.com/na/medicalngk/english/technillfo/radsterlitml.

 

Scholla, M., E. Hackett, et al. (2000). "A potential method for the speciﬁcation of
microbial barrier properties." Medical device a_pd technology.

Sharp, J. (1994). "Comment on Validation, risk-beneﬁt analysis." PDA Journal of
Pharmaceutical Scignce & Technology 50(1): 3-4.

Shaw, J ., M. Bordner, et a1. (1996). "Efﬁcacy of Steri-Sheild ﬁltered exhaust helmet in
limiting bacterial counts in the operating room during total joint arthroplasty." Journal
of Arthropla_s_ty 11(4): 469-473.

Sheehan, M. J. and J. V. Giranda (1994). "Bioaerosol generation in a food processing

plant: a comparison of production and sanitation operation." Appl. Occup. Environ.
Hyg. 9(5): 346-352.

159

 

Singh, J ., P. Singh, et al. (2002). Pack_age Erformance for hazardous materials in high
altitude shipments. WorldPak, East Lansing, MI.

Spitzley, J. (1994). "The relevance of microbial challenge tesing to package sterility."
Medical device and diggpostic indus_trv_ 15(1): 142-148.

Spitzley, J. (1993). "How effective is microbial challenge testing for intact sterile
packaging." Medical device and diggpostic indm 15(8): 44-46.

Spitzley, J. (2002a). Meeting- IOPP Medical Device Packaging Manufacturers. Chicago,
IL.

 

Spitzley, J. (2002b). The state of sterile Mage integp'ﬂ testing in the medical device
M. WorldPak, East Lansing, MI.

Stauffer, T. (1990). "non—desh'uctive testing on ﬂexible packaging." J. Pack. Technol.:
27-29.

Sterile Barrier Integrity Task Group (2002). "Update ﬁ'om the Institue of Packaging
Professionals." 1-8.

Stersky, A., E. Todd, et al. (1980). "Food poisoning associated with port process leakage
(PPL) in canned foods." J. Food Prot. 43(6): 465-476.

Stover, C., X. Pham, et al. (2000). "Pseudomonas Genome Project: Pseudomonas
aeruginosa." Rehieved May 13, 2003, ﬁ'om
llttp://\nmv.pseudomonas.com/p_aerug.html.

Swain, E. (1998). "Machine Vision: Easier to Implement, Easier to Use." Pharmaceutical
& Medical Pacl_c_agm g News 6(7): 38,42.

Szczeblowski, J .W. and Rubinate, F.J. (1968) Integrity of Food Packages. Modem
Packaging. 38(10), 131-135.

Tallentire, A. and C. S. Sinclair (2002). "A Discriminating Method for Measuring the
Microbial Barrier Performance of Medical Packaging Papers." Medicgl device a_nd
diggpostic indmy.

Tammelin, A., P. Domecil, et al. (2000). "Dispersal of methicillin-resistant

staphylococcus epidermidis by staff in an operating suite for thoracic and
cardiovascular surgery: relation to skin carriage and clothing." 44(2): 119-126.

Tammelin, A., A. Hambraeus, et al. (2001). "Routes and sources of Staphylococcus
aureus hansmitted to the surgical wound during cardiothoracic surgery: possibility of
preventing wound contamination by use of special scrub suits." Infection Control Q91
Hogpital Epidemiology 22: 338-346.

160

 

 

Tammelin, A., A. Hambraeus, et al. (2001). "Source and route of methicillin-resistant
Staphylococcus epidermis transmitted to the surgical wound during cardio-thoracic
surgery. Possibility of preventing wound contamination by use of special scrub sui ."
47: 266-276.

Teltsch, B., H. Shuval, et al. (1980). "Die-away kinetics of aerosolized bacteria from
sprinkler application of wastewater." Appl. Environ. Microbiol. 39(6): 1191 -l 197.

Thompson, P. and M. Griﬂith (1983). "Identity of mesophilic anaerobic sporeformers
cultured from recycled cannery cooling water." J. Food Prot 46(5): 400-402.

US. Food and Drug Administration (1938). Federal Food Drug and Cosmetic Act
(F D&C).

US. Food and Drug Administration. Code of Federal Regulations. Title 21, Volume 8, :-
Part 814-Premarket Approval of Medical Devices 1976 Amendment. i

Van Steenberghe, D., K. Yoshida, et al. (1997). "Complete nose coverage to prevent
airborne contamination via nostrils is unnecessary." Clincial Oral Iplants Res 8(6): .2»
5 12-5 16.

 

Vuong, C. and M. Otto (2002). "Staphylococcus Epidermis Infections." Microbes and
Infection 4(4): 481-489.

 

West, R. (1996). "Surface Analysis of Packaging Materials Part II." Medig Device
Technology 7(4): 36-38.

Whitaker, S. (1992). Introduction to ﬂuid mechanics. Malabar, FL., Krieger Publishing
Company.

 

Wynd, C., D. Samstag, et al. (1994). "Bacterial carriage on the ﬁngernail of OR nurses."
AORN Journal 60(5): 796, 799-805.

Yeh, H. and A. Benatar (1997). "Technical evaluation: methods for sealing paper/foil
aseptic food packages." TAPPI Journal 80(6): 197-203

161