%
-,»...fiu

:«n

RWAW‘V‘ . 3.11m:
Infirmatut. I’MMdcflnb.‘
‘ i

:5 .1

‘ it...
, ii}...

Uhflfifiwmau
fit!

Y- “. év ll
17%;. ..l i
I .

.. . {1,
U. ’13 V 51.11
Lr5.‘n'.uul|n. , . 5'-
i V . . . . . I. ,
.. its! .3...
.1133.

 

 

LIBRARY
Mict“ _' ‘n State
University

 

 

 

This is to certify that the
dissertation entitled

TRANSFER OF LISTERIA MONOCYTOGENES DURING
SLICING OF READY-TO-EAT DELICATESSEN MEATS

Doctoral

 

presented by

KEITH VORST

has been accepted towards fulfillment
of the requirements for the

Food Science and Human
Nutrition

degree in

 

:1 m J: flm

Major Professor’s Signature

«5 /3 /05

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

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

if

im.07mn

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 cycficiomomeu-ms

 

TRANSFER OF LIS TERI/1 MONOC YT OGENES DURING SLICING OF
READY-TO-EAT DELICATESSEN MEATS

By

Keith Vorst

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Food Science and Human Nutrition

2005

ABSTRACT

TRANSFER OF LISTERIA MONOCYTOGENES DURING SLICING OF
READY-TO-EAT DELICATESSEN MEATS

By

KEITH VORST

In response to continued concerns regarding Listeria cross-contamination of
ready-to-eat meat and poultry products in both retail and home kitchens, a series of
studies was conducted to: (1) optimize the quantitative recovery of L. monocytogenes
from stainless steel surfaces, (2) determine direct and sequential transfer rates for L.
monocytogenes from artificially contaminated ready-to-eat luncheon meats to a
delicatessen slicer and vice versa, (3) determine the effects of cutting force, stainless steel
grade, sharpness, and product composition on transfer of L. monocytogenes from
artificially contaminated ready-to-eat luncheon meats to knives and vice versa, and (4)
develop a mathematical model based on the transfer coefficients obtained from the
previous three objectives that will predict the numbers of L. monocytogenes cells
transferred during slicing of delicatessen meats.

Initially, four sampling devices: (1) sterile environmental sponge (ES), (2) sterile
cotton-tipped swab (CS), (3) sterile calcium alginate fiber-tipped swab (CAS), and (4) 1-
ply composite tissue (CT), were evaluated for quantitative recovery of L. monocytogenes
from food-grade stainless steel. Recovery was 2.70, 1.34, and 0.62 log greater using CT
compared to ES, CS, and CAS, respectively. The CT device, which is inexpensive and

easy to use, represents a major improvement over other methods in quantifying L.

monocytogenes.

Thereafter, a commercial delicatessen slicer blade and simulated kitchen knife
blades were used as vectors for sequential transfer of L. monocytogenes from (a) an
inoculated blade (~108, 105, 103 CFU/blade) to 30 slices of uninoculated delicatessen
turkey, bologna, and salami, (b) inoculated product (~108 cmz) to the blade and (c)
inoculated product (108, 105, 103 CFU/cmz) to 30 slices of uninoculated product via the
blade with cutting force and product composition also assessed for their impact on
Listeria transfer. Using slicer blades inoculated at 108 CPU/blade, Listeria populations
decreased logarithmically to 102 CFU/slice after 30 slices. Findings for inoculated slicer
blades and products (105 CPU/blade or cmz) were similar with Listeria counts of 102
CFU/slice after 5 slices and enriched samples generally negative after 27 slices. Using
103 CPU/slicer blade, the first 5 slices typically contained ~10l CPU/slice by direct
plating with enrichments negative after 15 slices. Knife blades containing 105 and 103
CPU/blade typically yielded direct counts out to only 20 and 5 slices, respectively, with
“tailing” observed thereafter. Variables that enhanced Listeria transfer during slicing and
cutting included higher fat and lower moisture content, application force, blade surface
roughness, and stainless steel grade with greater transfer using 304 as opposed to 316.

These finding were then used to develop four fitted predictive models in the form
[CFU (X) = kax ] along with a program written in GWBasic. These models can be used
if any two of the following three values are known: (a) initial inoculum level, (b) total
bacterial transfer, (c) fraction of bacteria remaining on blade after consecutive slicing,
solving for each model parameter CFU(X), k, or a. Based on our models, the greatest

number of Listeria (>90%) will be found in the first 15 slices.

To my wife for never ever letting me give-up

To my son and future son or daughter may you build on my successes, learn from my

failures, and rise above my weaknesses

To my parents for pushing enough but not too much

iv

ACNOWLEDGEMENTS

This dissertation was not solely written by me but with the help of many
influential people. Foremost among them is my wife who never let me give-up when all
hope seemed lost. After receiving my Masters degree she constantly encouraged me to
further develop my professional skills and use these to make my contribution to science
and society. We share this degree on many levels and none more important than my love
for her and my son.

I would like to express my sincere appreciation to Dr. Elliot T. Ryser. Dr. Ryser
was not only my mentor but my best friend and saw my talents when many had given up.
He worked to make sure I gave everything even when I thought it was hopeless. Without
Dr. Ryser this would not have been possible and we will remain friends for life.

It is no less important to express my gratitude for Dr. Alden Booren and Dr.
Bruce Harte for not only completing a successful Masters but seeing my through a
Doctorate. In many ways Dr. Booren was a father figure that made sure I was on the
right track and always helped me out in a pinch. With his guidance I found a successful
project and I will never forget his help and words of wisdom. I owe a great deal of
gratitude to Dr. Harte for keeping my thoughts positive and always seeing the application
of my work. Dr. Harte never second guessed if I would finish my degree but only with
his help and support I am able to complete this project with a smile.

I would like to thank Dr. Burgess for opening my mind to critical thinking. I will

be forever grateful to him for coming through at the last minute with modeling and

always making time for me no matter what the circumstance. I only hope Dr. Burgess
will touch more lives like he has mine with respect to understanding mathematical
relationships and tools that all doctoral students should have.

I would like to thank Dr. Ewen Todd for his continued support and willingness to
serve as a mentor. Dr. Todd provided great insight into my scientific thought process and
always made time for my committee meetings and manuscript reviews.

I would like to thank Dr. John Linz for serving as my committee chair and
providing his insight on the research project. Over the years I have enjoyed my
conversations with Dr. Linz in the parking lot and at many seminars. I also felt his
insight to be genuine and in the best interest of the student.

I would like to acknowledge all my past and present colleages in my laboratory. I
appreciate their support through the good times and bad. My sincere appreciation goes to

my friends and family who helped me achieve my goals.

vi

TABLE OF CONTENTS

ABBREVIATIONS ............................................................................. xviii

CHAPTER 1

LITERATURE REVIEW .............................................................................. 5

1.1

1.2

1.3

1.4

1.5

1.6

1.7

Listeria monocytogenes ........................................................... 5

1.1.1 Characteristics of Listeria ....................................... 5
1.1.2 Manifestations of listeriosis .................................... 6
1.1.3 Susceptible populations ......................................... 6
Listeriosis outbreaks ............................................................... 9
Incidence in ready-to-eat meats ................................................. 1 1
Listeria recalls ..................................................................... 14
USDA and FDA guidelines ...................................................... l6
Listeria post-process contamination ............................................ 21
1.6.1 Food processing environments ................................ 21
1.6.2 Sampling and recovery ......................................... 23
Listeria attachment and transfer ................................................. 24

1.7.1 Stainless steel .................................................... 25

vii

1.7.2 Attachment and release of Listeria from stainless steel. . .28

1.7.3 Other food contact surfaces ................................... 34
1.8 Risk assessment .................................................................. 35
1.8.1 FDA, USDA, and F AO/WHO ................................ 36
1.9 Predictive modeling of microbial growth and transfer ...................... 43
1.9.1 History and development of microbial modeling .......... 44

1.9.2 Techniques for predictive modeling of microbial growth

and transfer ...................................................... 44

1.9.3 Bacterial growth and thermal inactivation models ......... 45

CHAPTER 2. Improved quantitative recovery of Listeria monocytogenes from stainless

steel surfaces using a l-ply composite tissue

[Published in J. Food Prat. 67:2212-2217] ................................... 51
2.1 abstract ................................................................................................. .52
2.2 Introduction .......................................................................................... .53

2.3 Materials and MethodsS6

2.3.1 Preparation of strains ........................................... 56
2.3.2 Stainless steel .................................................... 56
2.3.3 Sampling devices ................................................ 57
2.3.4 CT and CAS ..................................................... 57
2.3.5 ES and CT ........................................................ 58
2.3.6 Potential inhibition in PBS .................................... 61
2.3.7 Modified disc diffusion assay ................................. 61

viii

2.4

2.5

2.6

CHAPTER 3.

3.1

3.2

3.3

2.3.8 Evaluation of recovery methos using SEM .................. 62
2.3.9 Statistical analysis ............................................... 62
Results 62
Discussion ............................................................................................ .68

Summary ............................................................................ 71

Transfer of Listeria monocytogenes during mechanical slicing of

turkey breast, bologna, and salami .............................................. 72
Abstract ............................................................................. 73
Introduction ........................................................................... 74
Materials and Methods ............................................................ 76
3.3.1 L. monocytogenes strains ................................................ 76
3.3.2 Delicatessen meats ....................................................... 77
3.3.3 Delicatessen slicer ........................................................ 78
3.3.4 Identification of delicatessen slicer product contact areas .......... 78
3.3.5 Surface profiling of delicatessen slicer blade ........................ 80
3.3.6 Evaluation of slicer blade using SEM ................................. 81

3.3.7 Impact of force on L. monocytogenes transfer from turkey
to a delicatessen slicer ................................................... 81
3.3.8 Slicer blade inoculation .................................................. 81
3.3.9 Transfer of L. monocytogenes from an inoculated delicatessen
slicer blade to uninoculated product ................................... 82

3.3.10 Transfer of L. monocytogenes form inoculated product via the

ix

3.4

3.5

3.6

CHAPTER 4.

4.1

4.2

4.3

slicer to uninoculated product .......................................... 83

3.3.11 Quantification of injured Listeria on slicer blades ................... 84
3.3.12 Cleaning and decontaminating the slicer .............................. 84
3.3.13 Statistical analysis ........................................................ 85
Results ............................................................................... 85
3.4.1 Proximate analysis ....................................................... 85

3.4.2 Impact of force on L. monocytogenes transfer from turkey to a
delicatessen slicer ........................................................ 86

3.4.3 Transfer of L. monocytogenes from an inoculated delicatessen
slicer blade to uninoculated product ................................... 88

3.4.4 Sequential transfer of L. monocytogenes from inoculated

product to a delicatessen slicer ......................................... 93
3.4.5 Slicer blade surface profiling ........................................... 95
3.4.6 Quantification of injured Listeria on slicer blades .................. 96

D15cussron96

Summary .................................................................................................. 101

Transfer of Listeria monocytogenes during slicing of turkey breast,

bologna, and salami using kitchen knives .................................... 102
Introduction104
Materials and Methods .......................................................... 106

4.3.1 L. monocytogenes strains .............................................. 106

4.4

4.3.2

4.3.3

4.3.4

4.3.5

4.3.6

4.3.7

4.3.8

4.3.9

Delicatessen meats ...................................................... 107
Knife blades ............................................................. 108
Quantification of cutting force and speed ........................... 109
Surface profiling of knife blade ....................................... 110
Knife blade inoculation ................................................ 111

Transfer of L. monocytogenes from inoculated grade 304

stainless steel knife blades to uninoculated product ................ 1 11
Transfer of L. monocytogenes from inoculated grade 304 and

316 stainless steel knife blades to uninoculated product .......... 1 12
Transfer of L. monocytogenes from inoculated product to

grade 304 stainless steel knife blades and then to uninoculated

product ................................................................... 113
4.3.10 Clearing/decontaminating knife blades .............................. 1 13
4.3.11 Quantification of injured Listeria on knife blades .................. 1 14
4.3.12 Statistical analysis ...................................................... 114
Results ................................................................................. 115
4.4.1 Knife blade surface profiling .......................................... 115
4.4.2 Proximate analysis ...................................................... 115
4.4.3 Effect of stainless steel grade, product, and sharpness on

4.4.4

4.4.5

transfer ................................................................... 115
Transfer of L. monocytogenes from inoculated grade 304 and
316 stainless steel knife blades to uninoculated product .......... 1 16

Sequential transfer of L. monocytogenes from inoculated

xi

4.5

4.6

CHAPTER 5.

5.1

5.2

5.3

5.4

product to a knife blade and then to uninoculated product

using grade 304 stainless steel knife blades ......................... 123
4.4.6 Quantification of injured Listeria on 304 and 316 knife blades..125
Discussron125

Summary .................................................................................................. 128

Transfer coefficients and predictive models for Listeria monocytogenes

during slicing of ready-to-eat turkey, bologna, and salami ..................... 129
Abstract ................................................................................................... 130
Introduction ......................................................................................... 13 1
Materials and Methods ............................................................................. 133

5.3.1 Transfer coefficients for L. monocytogenes during slicing of
turkey, bologna, and salam1133
5.3.2 Predictive modeling of L. monocytogenes transfer during
slicing turkey, bologna, and salam1134

5.3.3 Predicting CFU’s on meat as a function of slice number (X). 136

5.3.4 Fitting the equation to data (finding “k” “a”) ....................... 136
5.3.5 Interpretation of fit results ............................................. 137
Results................. ............................................................................... 138

5.4.1 Transfer coefficients of L. monocytogenes during slicing
ofturkey, bologna, and salam1138
5.4.2 Predictive model for L. monocytogenes transfer during slicing

of turkey, bologna, and salami using a mechanical slicer ............ 147

xii

5.5 Discussion ............................................................................................... 1 52

5.6 Smnmary ................................................................................................. 155
CONCLUSIONS AND FUTURE REASEARCH ............................................ 157
APPENDIX I Preliminary modeling of results by Dr. Robert McMasters. . 1 62
APPENDIX 11 Screen shot of GWBasic program output ............................ 168
APPENDIX III Lag Time (LT) and Generation Time (GT) of

L. monocytogenes on turkey, bologna, and salami at 4° C ........... 170

APPENDIX IV Correspondence with Dr. Richard Whiting (FDA/CFSAN) ........ 172
APPENDIX V Example calculations and modeling of turkey delicatessen

slicer data ..................................................................................... 191

BIBLIOGRAPHY ............................................................................ l 96

xiii

Table 1.1

Table 1.2

Table 1.3

Table 1.4

Table 1.5

Table 1.6
Table 1.7
Table 1.8
Table 1.9
Table 1.10

Table 2.1

Table 3.1

Table 4.1

Table 4.2

LIST OF TABLES
Predicted relative risk rankings for listeriosis among food categories
for three age-based subpopulations and the United States total
population using median estimates of relative predicted risks for
listeriosis on a per annum basis .................................................... 7

Listeriosis outbreaks in the United States 1979-2003 ........................ 11

Prevalence (%) of L. monocytogenes in RTE meat and products
CY 1990-2000 ...................................................................... 13

Generation (GT) and Lag Times (LT) of L. monocytogenes in meats.......18

Estimated storage temperature and duration between manufacture and

retail for predicted median growth .............................................. 19
Physical properties of stainless steel .......................................... ' . . .27
Chemical properties of stainless steel ............................................ 30
Finish grades of stainless steel .................................................... 31
Typical stainless grades and applications in the food industry ............... 33
Variables affecting dose for risk of isteriosis ................................... 37

Least squares means for effect of method Pr > M for Ho: LSMean (i) =
LSMean (j) with log count as dependent variable .............................. 64

Number of samples yielding Listeria by direct count and/or

enrichment (N=3) for delicatessen slicer-product (DS-P) and
product-delicatessen slicer-product (P-DS-P) transfer for

turkey (T), bologna (B), and salami (S) .......................................... 92

Average slicing force (lbs) for turkey, salami, and bologna using
medium sharp (MS) and sharp (S) knife blades manufactured from
304 and 316 grade stainless steel ................................................ 116

Number of direct counts and positive enrichments for blade-product

(BP) and product-blade-product (PBP) transfer for turkey (T), and
bologna (B), and salami (S) ...................................................... 119

xiv

Figure 1.1

Figure 1.2

Figure 1.3
Figure 1.4
Figure 1.5

Figure 1.6

Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

LIST OF FIGURES

Class I recalls of delicatessen meat products 1994-2004 .................. 15
USDA/F SIS recall of RTE delicatessen meats identified as

Sliced or unsliced from January 1994-April 2005 .......................... 39
Refi'igerated RTE luncheon meat sales 1999-2004 ........................ 40
Sliced RTE luncheon meats 1999-2004 ...................................... 41
Projected RTE luncheon meat sales 2004-2009 ............................. 42
Mathematical framework for relationship of raw product,

environment and finished products ........................................... 47
Folding pattern of CT ........................................................... 58
CT before homogenization ..................................................... 60
CT after homogenization ....................................................... 60 1
Recovery of L. monocytogenes from stainless steel ........................ 63

Scanning electron micrograph of Listeria attached to stainless steel

plates after using ES device ..................................................... 66
Scanning electron micrograph of Listeria attached to stainless steel

plates after using CS device ..................................................... 66
Scanning electron micrograph of Listeria attached to stainless steel

plates after using CAS device .................................................. 67

Scanning electron micrograph of Listeria attached

to stainless steel plates after using CT device ................................ 67

Contact areas of gravity feed delicatessen slicer ............................. 79

Number of Listeria monocytogenes recovered at an application force
of 0 and 10 lbs ................................................................... 87

Transfer of L. monocytogenes from inoculated slicer blade
(108 CPU/blade) to uninoculated turkey, salami and bologna ............. 88

Transfer of L. monocytogenes from inoculated slicer blade
(105 CPU/blade) to uninoculated turkey, salami and bologna ............. 89

XV

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Figure 4.9

Figure 4.10

Figure 5.1

Transfer of L. monocytogenes from inoculated slicer blade
(103 CFU/blade) to uninoculated turkey, salami and bologna ............. 90

Transfer of L. monocytogenes from inoculated turkey
(108 CFU/cmz) to slicer blade to uninoculated turkey ...................... 93

Transfer of L. monocytogenes from inoculated turkey (IT) and
Inoculated salami (IS) (105 CFU/cmz) to uninoculated turkey (UT)
and uninoculated salami (US) during slicing ................................. 94

Scanning electron micrograph of new (A) and used (B) slicer blades
after 1 year of use ............................................................... 96

Instron 5565 electromechanical compression analyzer .................... 108

Surface scoring of used 316 (A) and 304 (B) electropolished
stainless steel knife blades ..................................................... 109

Stylus locations on blade for surface profile measurements .............. 110

Transfer of L. monocytogenes from an inoculated knife blade
(108 CPU/blade) to uninoculated turkey, salami and bologna ............ 117

Transfer of L. monocytogenes from an inoculated knife blade
(105 CPU/blade) to uninoculated turkey, salami and bologna ............ 118

Transfer of L. monocytogenes from an inoculated knife blade
(103 CPU/blade) to uninoculated turkey, salami and bologna ............ 120

Transfer of L. monocytogenes from inoculated knife blades
(grade 304, grade 316, 103 CPU/blade) to uninoculated turkey ......... 122

Transfer of L. monocytogenes from inoculated knife blades
(grade 304, grade 316, 103 CPU/blade) to uninoculated salami .......... 122

Transfer of L. monocytogenes from inoculated knife blades
(grade 304, grade 316, 103 CPU/blade) to uninoculated bologna ....... 123

Transfer of L. monocytogenes from inoculated turkey (IT)

(10‘ CFU/cmz) and inoculated salami (IS) (105 CFU/cmz) to

uninoculated turkey (UT) and uninoculated salami (US) during

slicing ............................................................................ 124

Cumulative L. monocytogenes transfer (%) from an inoculated slicer
blade (108 CPU/blade) to turkey, salami and bologna ..................... 139

xvi

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Figure 5.11

Figure 5.12

Figure 5.13

Cumulative L. monocytogenes transfer (%) from an inoculated slicer
blade (105 CPU/blade) to turkey, salami and bologna ..................... 14o

Cumulative L. monocytogenes transfer (%) from an inoculated slicer
blade (103 CPU/blade) to turkey, salami and bologna ...................... 141

Cumulative L. monocytogenes transfer (%) from inoculated
turkey (IT) and salami (IS)(10S CFU/cmz) to uninoculated turkey
(UT) and salami (US) during mechanical slicing ............................ 142

Cumulative L. monocytogenes transfer (%) from an inoculated)
knife blade (108 CFU/blade) to turkey, salami and bologna ............... 143

Cumulative L. monocytogenes transfer (%) from an inoculated knife
blade (105 CPU/blade) to turkey breast, salami and bologna ............. 144

Cumulative L. monocytogenes transfer (%) from an inoculated knife
blade (103 CPU/blade) to turkey breast, salami and bologna .............. 145

Cumulative L. monocytogenes transfer (%) from inoculated turkey
(IT) and salami (IS)(105 CFU/cmz) to uninoculated turkey (UT)
and salami using an uninoculated knife blade ............................... 146

Example: GWBasic output for salami sliced using an inoculated knife
blade (108 CPU/blade) .......................................................... 147

Plotted output using GWBasic for assessing L. monocytogenes
transfer from an inoculated slicer blade (10 CPU/blade) to salami ...... 148

Plotted output using GWBasic for predicting L. monocytogenes transfer
from an inoculated slicer blade (10 CFU/blade) to turkey ................ 149

Plotted output using GWBasic for predicting L. monocytogenes
transfer from aninoculated slicer blade (103 CPU/blade) to turkey
and bologna ....................................................................... 150

Plotted output using GWBasic for predicting L. monocytogenes
transfer from an inoculated slicer blade (103 CPU/blade) to salami ...... 151

xvii

AISI

AOAC

BI-IA
BHT
CAS
CDC
CFSAN
CFU
CS
CT
CY
ns
ES

FAO

F SIS
GLM
GT

ICMSF

ABBREVIATIONS

Atomic Force Microscopy

American Iron and Steel Institute
Association Official Analytical Chemists
Blade

Butylated Hydroxyanisole

Butylated Hydroxytoluene

Calcium Alginate Swab

Center for Disease Control and Prevention
Center for Food Safety and Applied Nutrition
Colony Forming Units

Cotton Swab

Composite tissue

Calendar Year

Delicatessen

Environmental Sponge

Food and Agriculture Association

Food and Drug Administration

Food Safety Inspection Service

General Linear Model

Growth Time

International Commission on Microbiological
Specifications for Foods

xviii

LT
MOX
MS
mTPA

mTPAN

PBS

RTE

SAS
SEM
TSA-YE
TSB
TSB-YE
USDA

WHO

Lag Time

Modified Oxford

Medium Sharp

Modified Tryptose Phosphate Agar

Modified Tryptose Phosphate Agar Sodium Chloride
(NaCl)

Product

Phosphate Buffered Saline

Relative Humidity

Ready-to eat

Sharp

Statistical Analysis Systems
Scanning Electron Microscopy
Trypticase Soy Agar-Yeast Extract
Trypticase Soy Broth

Trypticase Soy Broth-Yeast Extract
United States Department of Agriculture

World Health Organization

xix

LIST OF APPENDICES
APPENDIX I Preliminary modeling results by Dr. Robert McMasters ................. 162
APPENDIX 11 Screen shot of GWBasic program output .................................. 168
APPENDIX III Lag Time (LT) and Generation Time (GT) of

L. monocytogenes on turkey, bologna, and salami at 4° C ................... 170

APPENDIX IV Correspondence with Dr. Richard Whiting (FDA/CF SAN ) ................. 172
APPENDIX V Example calculations and modeling of turkey delicatessen

slicer data .............................................................................................. 191

XX

INTRODUCTION

Listeria monocytogenes is a serious bacterial foodborne pathogen that can
reside in food processing facilities for many years (Tompkin, 2001). Not
surprisingly, those strains that are most persistent in factory environments have been
shown to possess greater capability to adhere to food contact surfaces (Lunden et al.,
2000, 2002; Norwood and Gilmour, 1999) with attachment of L. monocytogenes to
stainless steel occurring in as little as 20 min (Mafu etal., 1991). Such persistence of
L. monocytogenes in food processing environments allows this pathogen to
contaminate finished product and previously uncontaminated facilities via processing
equipment and other food contact surfaces.

Processing equipment increases the risk of widespread dissemination of
foodborne pathogens with bacterial transfer, reversible/irreversible attachment and
biofilm formation influenced by stainless steel surface finish and composition of the
food contact surface. In a comparative study of different food contact surfaces
(Beresford et al., 2001), coupons of grade 304 (2B finish and 2B sand-blasted finish),
316 (electropolished) and 430 stainless steel (2BA finish) were immersed in a broth
culture of L. monocytogenes for 2 h and then removed. After gentle rinsing, 25% of
the L. monocytogenes cells were released from both types of grade 304 stainless with
only 7 and 5% of the population shed from grades 316 and 430, respectively, thus
suggesting involvement of stainless steel grade and surface finish in bacterial
attachment. Arnold and Bailey (2000) assessed bacterial attachment rates for four
different surface finishes of grade 304 stainless steel. When exposed to a mixed

bacterial culture obtained from a poultry carcass rinse, bacterial attachment was at

least 1 log lower on electropolished stainless steel compared to the other three
surfaces. These findings clearly have important ramifications in the manufacture of
stainless steel knife blades, delicatessen slicer blades and other food contact surfaces
found on food processing equipment as well as in commercial and retail food
processing environments. However, the transfer rates for L. monocytogenes to and
from contaminated ready-to-eat (RTE) meat products to stainless steels of different
compositions and surface finishes remain poorly understood.

Transfer of Listeria from a contaminated product to a slicer blade or knife and
then to a previously uncontaminated product is likely a major route of dissemination
in retail delicatessens. However, adherence and subsequent release of any
microorganisms from a blade during slicing are likely impacted by a multitude of
factors that relate to the particular bacterial strain as well as the type of equipment
(e.g., slicing machine, knife), type and condition of the blade (e.g., stainless steel
grade, blade sharpness, extent of wear and corrosion) and cutting force. The
aforementioned factors and the particular product being sliced will likely result in
different transfer rates Our work has shown that L. monocytogenes can transfer from
artificially contaminated delicatessen slicer and knife blades to delicatessen meats and
vice versa during slicing with the extent of transfer being product dependent.
Additional findings suggest less prolonged transfer using grade 316 as Opposed to
grade 304 stainless steel knife blades with surface roughness and stainless steel grade
impacting the rate of Listeria transfer.

Based on informational gaps identified in the current Listeria risk assessments

(FAQ/WHO 2004; FDA 2003; F818, 2003) specific information is needed concerning

the extent of Listeria transfer (a) from contaminated foods to soiled and unsoiled
surfaces and (b) from contaminated surfaces (soiled and unsoiled) to foods. This
transfer potential needs to be quantified and expressed in terms of transfer
coefficients that can be incorporated into the various L. monocytogenes risk
assessments to more accurately define the risks associated with consumption of
ready-to-eat (RTE) foods. Based on the F SIS risk assessment for Listeria in RTE
meat and poultry products (F818, 2003), of the approximately 500 listeriosis fatalities
each year, an estimated 242 deaths are related to consumption of Listeria-
contaminated delicatessen meats. These findings suggest that minimizing
contamination at delicatessens will clearly have a major impact on reducing the
incidence of listeriosis and in meeting and/or exceeding the goals identified in
Healthy People 2010 (2004).

The research being reported in this dissertation was conducted in direct
response to the identification of the Listeria transfer rate as an informational gap in
the 2003 FDA Listeria Risk Assessment (FDA/FSIS/CDC, 2003). The approach was
to first calculate a series of transfer coefficients for L. monocytogenes during slicing
of various luncheon meats with a delicatessen slicer or knife blade and then develop a
predictive mathematical model for Listeria transfer that can be used to refine the
current Listeria risk assessments. The general hypothesis for this research is that the
rate of Listeria transferred during slicing of delicatessen products is influenced by the

type of slicer or knife blade and product composition.

The specific objectives of this three-year study were as follows:

m1:

Objective 2:

W

W

Optimize the quantitative recovery of L. monocytogenes from stainless

steel surfaces.

Determine the direct and sequential transfer rates for L.
monocytogenes from artificially contaminated ready-to-eat luncheon

meats to a delicatessen slicer and vice versa.

Determine the effects of cutting force, stainless steel grade, sharpness,
blade wear and product composition on transfer of L. monocytogenes
from artificially contaminated ready-to-eat luncheon meats to knives

and vice versa.

Develop a mathematical model based on the transfer coefficients
obtained in Objectives 1, 2, and 3 that will predict the numbers of L.
monocytogenes cells transferred in delicatessens when either the slicer

blade/knife or product is contaminated.

CHAPTER 1

LITERATURE REVIEW

1.1. LISTERIA MONOCYTOGENES

Listeria monocytogenes is a ubiquitous foodborne pathogen found in many
raw foods and processing environments. The ability of this organism to attach to
many different substrates and survive in harsh environments has prompted great
public concern and resulted in strict regulatory policies including the currently
enforced ‘zero tolerance” policy for L. monocytogenes in cooked and/or otherwise

processed ready-to-eat (RTE) foods.

1.1.1. Characteristics of Listeria

The genus Listeria comprises a group of gram positive, non-spore forming,
short rod-shaped bacteria and contains the following six species: L. monocytogenes,
L. innocua, L invanovii, L. seeligeri, L. welshimeri, and L. grayi. Listeria
monocytogenes — the primary human pathogen of the aforementioned species, is of
particular concern as a foodborne pathogen due its psychrotrophic characteristics and
ability to grow in refrigerated foods (Rocourt, 1999). Other characteristics of
importance to food manufacturers include the organism’s resistance to acid, salt and
low moisture environments. Thus, while typically found in soil and water as well as
wild and domesticated animals, it is not surprising that L. monocytogenes is also

common in many food-processing environments.

1.1.2. Manifestations of listeriosis

Listeriosis, the disease caused by infection with L. monocytogenes, is
generally confined to high-risk groups including immunocompromised adults,
pregnant women, and neonates. Such infections are relatively rare with an estimated
2500 cases occurring annually in United States, 500 of which are typically fatal
(Mead et al., 1999).

Listeria monocytogenes primarily infects humans via contaminated food and
invades the intestine resulting in multiplication and systemic spread via the
circulatory system. Manifestations of listeriosis include flu-like symptoms,
meningitis, spontaneous abortion, fetal death, neonatal septicemia and less commonly
gastroenteritis (Slutsker and Schuchat, 1999). The incubation time for listeriosis
typically ranges from 14 to as long as 70 days greatly complicating the investigation

of potential foodborne outbreaks.

1.1.3. Susceptible populations

Based on relative risk of population subgroups for listeriosis, the Food and
Drug Administration / Center for Food Safety and Applied Nutrition (FDA/CFSAN)
has ranked food categories based on a per serving basis. Delicatessen meats ranked
first in the predicted relative risk rankings for listeriosis among food categories for
three US. age-based subpopulations (Table 1.1). These predicted risk rankings along
with the increased number of meals consumed outside the home and the growing
population of immunocompromised adults have resulted in heightened scrutiny of

meat processors and retailers alike.

Table 1.1. Predicted relative risk rankings for listeriosis among food categories for
three age-based subpopulations and the United States total population
using median estimates of relative predicted risks for listeriosis on a per

annum basis (F DA/F SIS/CDC, 2003)

Seafood

Seafood

SEAFOOD Fish
Ready-

PRODUCE
Soft

Unripened
>50%

Ripened
>50%

39-50%

Cheese,
moisture

 

Table 1.1. (Con’t)

 

 

 

 

 

 

 

 

 

 

 

 

 

Ice Cream and
Frozen Dairy 21 19 20 20h
Products
. . Cultured Milk
Dairy (con t) Products 22 22 22 22h
High Fat and
Other Dairy 3 3 3 3a
Products
Dry/Semi-Dry
Fermented 13 13 13 13
MEATS ausages
Deli Meats 1 1 1 1
Paté and Meat
preads 6 6 6 6b,c,d
COMBINATION .
FOODS Dell-type Salads 18 14 18 17f

 

aFood categories are grouped by type of food but are not in any particular order.

b A ranking of 1 indicates the food category with the greatest predicted relative risk of
causing listeriosis and a ranking of 23 indicates the lowest predicted relative risk of

causing listeriosis.

°'h Ranks with the same letter are not significantly different based on the Bonferroni

Multiple Comparison Test (alpha=0.05).

 

Due to the high fatality rate (~20%) and hardiness of this organism, a
regulatory policy of “zero tolerance” was established in the United States for all RTE
foods in 1985 following the now infamous outbreak in California in which
consumption of Jalisco-brand Mexican style-soft cheeses was linked to at least 300
cases of listeriosis including 85 fatalities (MMWR, 1985). Typical heat treatments
given to RTE and processed foods will eliminate L. monocytogenes (Frye et al.,
2002). However, many processed foods, including smoked salmon and turkey, ham,
and beef luncheon meats, remain prone to Listeria contamination and have been the
subject of numerous Class I recalls. These findings reaffirm the importance in
minimizing post-processing contamination of RTE foods. While regulatory agencies
and food processors struggle with the challenges of controlling Listeria contamination
during processing, many retail delicatessen environments afford ample opportunity
for cross-contamination and subsequent growth of Listeria in RTE products due to

increased handling and reduced regulatory control.

1.2. Listeriosis Outbreaks

Listeria monocytogenes first emerged as a foodborne pathogen in 1981 when
17 of 41 people died of listeriosis after consuming coleslaw that was marketed in the
Maritime Provinces of Canada (Schlech, 2000). Since 1979, a total of 13 listeriosis
outbreaks have been reported with nine from contaminated RTE meat products (CDC,
2004). The California outbreak in 1985 led to numerous stillbirths from Mexican-
Style cheese. More recent outbreaks (1998, 2001 , and 2002) involving various

processed meats have made L. monocytogenes a pathogen of heightened public

concern, most notably with the 1998 outbreak resulting in 21 fatalities from BallPark
brand turkey hotdogs manufactured by BilMar Foods (Zeeland, MI). Since May of
2000, four major outbreaks of foodborne listeriosis have been documented in the
United States; three were linked to consumption of delicatessen-sliced turkey breast
(Table 1.2). The first of these outbreaks was responsible for 29 cases of listeriosis,
including 4 deaths and 3 miscarriages/stillbirths, in 10 states and prompted the recall
of 16.9 million pounds of product (MMWR, 2000). In June of 2001 , another
listeriosis outbreak characterized by 16 cases of acute febrile gastroenteritis (no
fatalities) was identified in Los Angeles County, California (MMWR, 2001).
Precooked delicatessen-sliced turkey was identified as the vehicle of infection
with L. monocytogenes serotype 1/2a of the same molecular fingerprint recovered
from six of the victims and from some of the leftover turkey at levels of 1.6 x 109
CFU/g. The latest and largest of these three outbreaks involved 46 culture-confirmed
cases of listeriosis, including 7 deaths and 3 miscarriages/stillbirths in eight primarily
northeastern states and led to the recall of 27.4 million pounds of RTE turkey and
chicken products (MMWR, 2002). In all cases, processed RTE meats (e.g.,
delicatessen turkey and turkey frankfurters), were the most common vehicle of

infection.

10

Table 1.2. Listeriosis outbreaks in the United States 1979-2003 (CDC, 2004)

 

 

 

 

 

 

 

 

 

 

 

 

 

Year State Serotype Vehicle
2003 TX 4b Mexican-Style Cheese
2002 9 * 4b Deli turkey meat
2001 CA 1/2a Deli Turkey meat
2000 10 * 4b Deli type turkey, chicken meat
2000 . NC 4b Mexican-Style Cheese
1999 3 * 1/2a Pate
1999 4 * 4b Mexican-Style Cheese
1998 22 * 4b Hot Dogs
1994 IL 1/2b Chocolate Milk
1989 CT 4b Shrimp
1985 CA 4b Mexican-Style Cheese
1983 MA 4b Milk
1979 MA 4b Produce

 

 

 

 

* Multistate outbreak

1.3. Incidence of Listeria in RTE Meats

Beginning in 1987, the USDA/F SIS setup a Listeria monitoring program for
meat products (U SDA/F SIS, 2003). Initial sampling included cooked beef products,
but was expanded in 1993 to include meat/poultry products as well as meat/poultry
spreads. Data from 1990-2000 (FSIS, 2005) shows the prevalence of L.
monocytogenes to be highest in ham and luncheon meats compared to 7 other
categories of RTE products (Table 1.3).

In the most recent comprehensive survey of sliced luncheon meats, 31,705
samples of ready-to-eat meat and cheese products were tested for L. monocytogenes
over 14-23 months in retail markets across the United States. L. monocytogenes was
identified in 0.4% of prepackaged vs. 2.7% of delicatessen-sliced luncheon meats

with deli meats sliced on demand accounting for approximately 75% of all deli meat

11

sales (Gombas et al., 2003). Thus, consumer exposure to Listeria is approximately 7
times greater from luncheon meats sliced on-site at delicatessens as compared to pre-

packaged commercially sliced products.

12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cod $.m as 3; and 2: SN Dav mod 3:32:80

cod 9o.— wod em; mud Gd cm; mod «Nd coon
<\Z oo.m m _ ._ 3e.— ood $6 on; and Sum 33
<\Z SN :.m NNN cm; a: a: M26 2N M32
32 cad 2mm 36 cod mo; Sum owe mod 32
$2 $2 5 _.m 5 fl .m cod mad end 36 mm.m coo—
SZ 32 $6 mmd cod ‘2.— ooé cod mod 32
sz 32 Sam RN «Na 3.. :34 $46 cod vam—
<\Z 32 SN 5; cod 2N and mod worm .82
32 $2 mm.m SN cod Ned 86 as on...” moa—
sz 32 h _ .m Nod cod cod ems «Wm No4e 3o—
32 332 M36 and cod mm.m _~.v $5 wmd o3—
a mouse—Em $3.:— floseeum Natural rowan—Em woman—3m 33: noon 23>

6858.5..— Ecaoam 5.5—sen— eoo—eeO 63—on 52—254— .3950

Eta—am woe—eoO owner— :aEm Q Ea: Janey—
cooam 633.5

83.32 >0 .3369:— bzzoa v5 :88 EM 5 autumoekoozofi Q .«o Ao\ev coco—«>05 .Q— nigh.

l3

1.4. Listeria recalls

Post-process contamination of cooked/RTE delicatessen products with L.
monocytogenes has made this pathogen the leading cause of Class I microbiologically
related recalls (Levine et al., 2001). Since 1994, Listeria contamination of
delicatessen meats has resulted in 66 Class I recalls (Figure 1.1). Of these 66 recalls,
ham was most frequently contaminated (44 recalls) followed by beef (24 recalls),
turkey (14 recalls), and chicken (7 recalls). Although responsible for the greatest
number of recalls, ham has not been implicated in large multi-state outbreaks unlike
turkey and chicken (Figure 1.1). Within the last decade more than 50 million pounds
of hot RTE hot dogs, chicken, and turkey luncheon meats have been recalled for
Listeria contamination. These recalled RTE meat products were linked to more than
130 cases of listeriosis and 28 fatalities during two separate outbreaks in 1998 and
2001 (MMWR, 2000).

Products recalled during these outbreaks were suspected of becoming
contaminated during packaging. While ham products have seen the highest total
number of recalls (44) since 1994, refrigerated growth studies by Glass and Doyle
(1989) demonstrated a greater growth of Listeria in processed poultry products as
compared to ham, bologna, and bratwurst. When processed meats were inoculated
with a five- strain cocktail of L. monocytogenes (<102 CF U/g) and stored at
refrigeration temperatures, the pathogen grew to 103-104 CFU/g on ham after 6 weeks
and tolO3-105 CFU/g on turkey and chicken after 4 weeks. Consuming low levels of
L. monocytogenes in food is not uncommon and is not considered a significant risk to

most people (Chen et al., 2003).

14

uno> .....
QOON noon Nch SON cocN mam“. ham r mam r cmmv

.E

\

\\\\\\\\\\\

 

 

350 m

eoom 02:00 I
3cm D .
:eono I
>995... I fi

Ea: I

 

 

 

 

Amoom .mHmSQmDv vcoméoe 30:on “was commoamomow mo £32 H awe—O

, .v

c
N

(D
sueoel jo JeqwnN

2.

Nr

.3 2.5:.

15

However, the aforementioned studies illustrate the ability of L.
monocytogenes to reach levels potentially hazardous levels (>103 CF U/ g) in certain
delicatessen products such as chicken and turkey luncheon meat. Delicatessen-sliced
products continue to be a concern due to temperature abuse and varying degrees of
regulatory scrutiny at the retail level. Previous studies have not addressed the
differences in growth, transfer, and distribution of foodborne pathogens in retail
delicatessens that result from different storage, handling, and food preparation
practices.

In response to previously mentioned outbreaks that were traced to
consumption of delicatessen-sliced turkey, quantitative transfer to and from
commercial slicing machines, knives, and cutting boards in delicatessens was
identified as both a major public health concern and a key informational gap in the

2003 FDA Listeria Risk Assessment (FDA/FSIS/CDC, 2003).

1.5. USDA and FDA guidelines

The public health significance surrounding L. monocytogenes has led to a
regulatory policy of "zero tolerance" in the United States for this organism in cooked
and/or otherwise processed RTE foods. The three previously discussed outbreaks
involving delicatessen turkey have prompted the development of three USDA-
mandated alternatives for controlling Listeria in delicatessen meats - (a) post-package
pasteurization, (b) product reformulation to prevent Listeria growth and/or (c)
increased product and environmental testing (FSIS, 2003b). While food processing

environments remain major sources of contamination, the extent to which Listeria is

16

transferred from food contact surfaces and utensils to RTE products at retail
delicatessens remains largely unknown. In one of two studies reported, Hudson and
Mott (1993) collected various environmental swab samples from a supermarket
delicatessen and isolated L. monocytogenes from a knife and slicing machine with the
pathogen also found at most sites near a display case of processed meats. In the
remaining study, Humphrey (1990) evaluated retail delicatessen meat slicers in the
UK and found L. monocytogenes on 10 of 32 slicer blades, thus suggesting ample
opportunity for Listeria transfer. These studies support the need for increased
scrutiny at the retail level.

Many delicatessen meats have an estimated 10-30 day shelf-life at l-5° C with
growth estimates showing up to a 2 log increase during this period (Glass and Doyle,
1989) (Table 1.4). RTE products are exposed to wide fluctuations in temperature
from the point of manufacture to the time of consumption allowing Listeria ample

time to reach potentially infectious levels.

17

Table 1.4 Generation (GT) and Lag Times (LT) of L. monocytogenes in meats

 

 

Temperature
Food (C) at (h) LT (11) REF
Roast Beef -1.5° 100 173.7 1'
3. 26.7 59 1
3b 80.9 477.1 1
Corned Beef 0 110 1
Cooked Meat 5 44-61 1
Ham 5 33.2 1
10 13.4
15 6.1
Cooked beef 5 18.6-22.6 80.6-83.4 1
10 8.5-9 22.6-30.4 1
Sliced Turkey 4.4 15.6-37.8 2‘
Pate 7 19.7 48 1
23.53 1.6 24 1
Pate 10‘il 9.12 27.6 1

 

1" — J. Farber and P. Peterkin, 1999

2‘ — Glass and Doyle, 1989

The wide temperature range in home refrigerators is more likely to promote
the grth of L. monocytogenes in RTE foods to infectious levels than are the more
tightly temperature controlled commercial refrigeration units with the former (Pinner
et al., 1992) conditions more likely to adversely affect “at risk” consumers. Given the
relatively short generation time (<16 h) for Listeria in some products at refrigerated
temperature and the extended storage and distribution times for delicatessen meats
(>10 days), the risk of contracting listeriosis from processed meats has become a

major concern for pregnant women and the elderly (Table 1.5).

18

Table 1.5. Estimated storage temperature and duration between manufacture and
retail for predicted median growth (FDA/FSIS/CDC, 2003)

SEAFOOD

PRODUCE

 

Table 1.5. (Con’t)

Cultured Milk
Products
Dairy (Con't) High Fat and
Other Dairy
Products
Frankfurters
Dry/ Semi-dry
Fermented

Deli Meats 1O 30
Pete and Meat 1 7

COMBINATION Deli-type
FOODS Salads

NAb NAb NAb

 

‘ Rectangular distributions were used for both the temperature range and storage
times.

l’NA - Not applicable because none of the samples were collected at manufacture so
growth between manufacture and retail was not calculated for these food categories.
° Median growth (log cfu) is calculated by multiplying the storage times and the
exponential growth rates

20

1.6. Listeria post-process contamination

The transfer of foodborne pathogens from contaminated food to previously
uncontaminated food via slicing machines was recognized over 40 years ago. During
the 1960’s, an outbreak in Aberdeen, Scotland led to 469 cases of typhoid fever from
contaminated corned beef that was sliced at a retail delicatessen, thus providing
evidence for delicatessen slicing machines as vectors for contamination of previously
uncontaminated product (Howie, 1968). In a far more recent study, Lin et al. (2004a)
showed increased Listeria transfer from contaminated slicer parts to turkey, salami,
and bologna during slicing. In their study, a commercial delicatessen slicer blade was
inoculated with a five-strain cocktail of L. monocytogenes (102 CPU) and then used to
slice turkey, bologna, and salami. Packages containing five slices of each product
were vacuum-sealed and assessed for Listeria growth afier l, 30, 60, and 90 days of
storage at 4°C. While Listeria populations increased in roast turkey breast, numbers
gradually declined in salami and bologna and fell below detectable limits after 60 and

90 days of storage.

1.6.1. Food processing environments

Listeria monocytogenes can reside in food processing facilities for many years
(Tompkin, 2001) with those strains that are most persistent in factory environments
possessing greater capability to adhere to food contact surfaces (Kim and Frank,
1994; Lunden et al., 2000, 2002; Norwood and Gilmour, 1999; Tiwari and Alenrach,
1990). Attachment of L. monocytogenes to stainless steel surfaces can occur in as

little as 20 min allowing ample time for long-term transfer (Mafu et al., 1991). While

21

many bacteria are capable of producing biofilms, true biofilm-forming strains of L.
monocytogenes are relatively rare (Kalmokoff et al., 2001) with this pathogen more
commonly seen in biofilms containing a mixed microflora (Bremer et al., 2001).
Adherence and subsequent transfer of Listeria is impacted by various environmental
conditions including temperature, relative humidity and substrate composition. The
persistence and spread of Listeria in food-processing environments (Lunden et al.,
2002; Tompkin, 2001; Vogel et al., 2001), and ability to grow to populations of 104 to
106 CFU/ g on many refrigerated RTE foods such as smoked salmon and luncheon
meats (Miettinen et al., 2001, Pinner et al., 1992) has made this pathogen a great
concern to the food industry.

Many studies have demonstrated the ability of L. monocytogenes to attach to
various materials and transfer to subsequent food products (Ak et al., 1994; Arnold
and Bailey, 2000; Beresford et al., 2001; Midelet and Carpentier, 2002, Schaffner et
al., 2004). Results of these studies suggest increased scrutiny of food contact surfaces
and development of new and innovative materials to reduce bacterial attachment and
subsequent transfer.

The aforementioned persistence of L. monocytogenes in food processing
plants allows this pathogen to enter previously uncontaminated facilities via
processing equipment and other food contact surfaces. In one study, Lunden et al.
(2002) demonstrated plant-to-plant transfer of L. monocytogenes via a dicing machine
with the same strain of Listeria identified at three different facilities. Studies have
also shown the attachment and transfer of bacteria during food preparation and

handling in food service environments. A study by Ak et al., (1994) demonstrated that

22

the absorption and bactericidal activity of wooden cutting boards represented a
significant advantage over plastic cutting boards with subsequent transfer more likely
from plastic boards to other food products. In other work by Schaffner et al. (2004),
plastic cutting boards were identified as a source of contamination when slicing
various raw meats and vegetables with colifonn counts ranging from 2-4 log CFU/g.
These studies clearly demonstrate the impact of material composition, surface
conditioning, and surface structure on attachment of bacteria and dissemination to

foods.

1.6.2. Sampling and Recovery

Current methods relying on cellulose sponges and cotton swaps for
environmental sampling of food processing facilities continue to be plagued by poor
repeatability and efficacy. The medical and pharmaceutical industries have evaluated
newer direct plating methods utilizing technologies such as adhesive sheets
(Yamaguchi et al., 2003) and pads (Tominaga et al., 2001). These devices are
reportedly superior to traditional swabbing in terms of both recovery and
reproducibility. However, the food industry is faced with several major obstacles that
make these sampling devices far less advantageous. Presence of food particulates,
fat, and oil on food processing equipment and other food contact surfaces makes
recovery with adhesive sheets and pads very inefficient. In addition to poor
performance. on heavily soiled food contact surfaces, these devices are costly when

compared to traditional swabs and environmental sponges.

23

Thus far, recovery strategies that have been developed to assess presence or
absence of Listeria in food processing environments lack the needed sensitivity to
quantify Listeria on solid surfaces (Vogel et al., 2001). The efficiency and reliability
of traditional environmental sampling devices such as the sponge and cotton swab
have been debated since their introduction in the 1970's (Ware et al., 1999). Studies
on meat surfaces have also shown that traditional destructive surface sampling
methods such as excision yield higher bacterial recovery (_>_ 50%) than non-
destructive swab and sponge methods (Tompkin, 2001). Direct agar contact plating,
including the use of Rodac® plates, has been plagued by limited sampling area, cost,
and difficulty with food particulates. While Rodac® plates have been successful for
sampling aerosols (Crozier-Dodson and Fung, 2002), their relatively small size and
their inability to withstand even a modest amount of mechanical energy during
sampling make them inadequate for many food contact surfaces (Moore and Griffith

20021

1.7. Listeria attachment and transfer

Given the hardiness and wide occurrence of Listeria in the environment, L.
monocytogenes has been successfully recovered from many raw and processed foods.
Attachment and subsequent transfer of Listeria to various materials can occur in a
very short time. Bacterial attachment to stainless steel has been widely studied (Akier
et al., 1990; Arnold and Bailey, 2000; Arnold and Silvers, 2000; Norwood and
Gilmour, 1999; Vantanyoopaisarn et al., 2000). In addition to stainless steel, limited

work using various polymers and fabrics (Ak et al., 1994; Beresford et al.,

24

2001;Montvi11e et al., 2001; Satter et al., 2001) also suggests that Listeria can be

transferred during common food handling tasks.

1.7.1. Stainless steel

Based on compositional differences, stainless steel can be classified as ferritic,
martinistic, austenitic, or precipitation hardened. Austenitic stainless steel is most
commonly used for food processing equipment, whereas ferritic and martinistic
stainless steels have seen limited use in the food industry due to their high cost and
inferior physical/chemical properties. High tensile strength, yield stress, and hardness
are desirable in food-grade stainless steel (Table 1.6.). The high chromium (Cr) and
nickel (Ni) (16 to 25 wt.% and 7 to 20 wt.%, respectively) content of austenitic
stainless steel results in excellent formability at room temperature with relatively
good resistance to oxidation (Table 1.7). American Iron and Steel Institute (AISI)
stainless grades 304, and 316 along with their low carbon counter parts 304L and
316L are the most commonly used alloys in the food industry (Smith, 1993) with the
400 series also used for knives. The standard mill surface finishes of stainless steel
are designated by AISI and range from basic hot rolled (No. 0 finish) - a very rough
finish that is not fully corrosion resistance, to cold rolled (No. 2, including 23, 2D,
and ZBA), mechanically produced polished finishes (No. 3 — 8) and electropolished
finishes (Table 1.8).

Corrosion of stainless steel after repeated cleaning and sanitizing likely
enhances bacterial attachment and transfer (Barkley, 1979; Bohner and Bradley,

1991). Mechanical and/or physical abuse during cleaning can produce additional

25

attachment sites as a result of surface marring and scratching. Bacterial attachment
and transfer can reportedly be decreased by ‘passivation’ (i.e., treatment with a mild
oxidant to remove surface iron and iron compounds) and by choosing 316 rather than
304 grade stainless with the latter being more corrosion resistant due to the addition
of 2.5 wt.% Mo (Arnold and Bailey, 2000). In limited work by Percival (1999), the
Mo concentration used in grade 316 stainless steel decreased bacteria viability and
reduced biofilm formation. In addition to corrosion resistance and potential biocidal
activity of Mo, grade 316 stainless steel has a smoother finish after manufacture and
polishing which decreases the number of bacterial attachment sites. (Leclercq-Pelat
and Lalande, 1994). Various alloys and polishes including electropolished, ZBA, and
4 finishes are available for food industry applications. However, other alloys and

polishes are often too costly for such use (Table 1.9).

26

Table 1.6. Physical properties of stainless steel (AISI, 2005).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Type TensrilienWsi) Yielgigksi) Elongation (Bl-13:23:22 x (RI-5:132:88)
max
300 Series Austlnetic
304 75 30 40% in 2" 183 88
304L 70 30 40% in 2" 183 88
316 75 30 40% in 2" 217 95
316L 7O 25 35% in 2" 217 95
400 Series Martinistic
410 65 30 20% in 2" 217 95
400 Series Ferritic
430 65 30 22% in 2" 183 88

 

 

 

 

 

 

 

 

27

1.7.2. Attachment and release of Listeria from stainless steel

Food interactions with various grades of stainless steel used in the food
industry can result in oxidation, pitting, and scoring over time due to both food
acidity and cleaning regimens. While many studies have addressed surface wear and
oxidation during processing (Arnold and Bailey, 2000; Bohner and Bradley, 1991;
Beresford et al., 2001; Bremer et al., 2001), only limited work has been done with
respect to retail food contact surfaces such as delicatessen slicers, kitchen knives, and
countertops. Studies have shown the influence of stainless steel grade and structure
on wear and attachment of bacteria (Akier et al, 1990; Arnold and Bailey, 2000;
Arnold and Silvers, 2000; Bohner and Bradley, 1991). Each of these studies
indicated that surface finish, grade, and conditioning had a major impact on bacterial
attachment and transfer. These studies also emphasized the need for knowledge of
stainless grades with the physical and chemical attributes least conducive to bacterial
attachment during mechanical and oxidative abuse commonly seen in processing and
retail food establishments.

Processing equipment increases the risk of widespread dissemination of
foodborne pathogens. Bacterial attachment to surfaces is influenced by surface
profile and composition. Previous studies have examined the impact of different
surface material compositions and surface finishes on bacterial attachment and
biofilrn formation (Arnold and Silvers, 2000, Briandet et al., 1999, Jeong and Frank,
1994; Norwood and Gilmour, 1999). According to Briandet et al., (1999) greater
bacterial attachment was seen for stainless steel than for rubber when semi-

quantitative absorbance values were obtained from recovered bacterial suspensions.

28

When surface morphologies of different stainless steels of various finishes were
examined, differences in bacterial attachment were observed. Relative differences in
stainless steel surface morphology based on the type of surface finish: 2B finish,
sandblasted, sanded, or electropolished, were also seen using scanning electron
microscopy (SEM) and atomic force microscopy (AFM) (Akier et al., 1990; Arnold
and Bailey, 2000). Significant differences in attachment to grade 304 stainless steel
that was inoculated with a bacterial chicken rinse suspension (2 x 106 CFU/ml) and
incubated for 18 h at 37° C were observed using various surface finishes.
Electropolished stainless steel had significantly fewer attached bacterial cells (102
cells) compared to other surface finishes (103 cells). No clumps were observed on
electropolished surfaces, whereas more than 12 clumps were seen for the other
surface finishes. Evidence from SEM also suggests that bacterial attachment is more
prevalent at grain boundaries in stainless steel with corrosion and pitting at these sites
further enhancing bacterial attachment (Arnold and Bailey, 2000). Hence, the
reduced grain boundary corrosion of low carbon alloys (304L and 316L) would be

expected to reduce the number of attachment sites and decrease bacterial transfer.

29

Table 1.7. Chemical properties of stainless steel (AISI, 2005)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Type C Mn P S Si Cr Ni Mb
300 Series Austinetic

304 0.08 2 0.045 0.03 1 1800/2000 800/1050 -

304L 0.03 2 0.045 0.03 1 1800/2000 800/1200 -

316 0.08 2 0.045 0.03 1 1600/1800 1000/1400 2.00/300

316L 0.03 2 0.045 0.03 1 1600/1800 1000/1400 2.00/300
400 Series Martinistic

410 0.15 1 0.04 0.03 - 1 115/13.5 -
400 Series Ferritic

430 0.15 1 0.04 0.03 - 1 1600/1800 -

 

30

 

Table 1.8. Finish grades of stainless steel (AISI, 2005).

 

Finish Description Application

 

Broad definition of manufactured steel finish

Standard to be used for further processing. Hot rolled Further processing needed
Mill Finish resulting in scaling which is subsequently for food applications
removed by nitric acid

 

Referred to as Hot Rolled Annealed (HRA)
No 0 resulting in sealed black finish. Does not Tool and dye applications.
' develop fully corrosion resistant film on Not for food applications.

stainless steel.

 

Hot rolled annealed, pickled and passivated. Further roce i d d
No. 1 Dull slightly rough finish. Starting finish for f g 15,5 “f “cc 6
cold rolled steel with bright finishes. 0° aPP lca Ions

 

Industry equipment
A No. I finish after being cold rolled, applications. Not used for
No. 2D annealed and passivated. Slightly dull finish food contact surfaces
improved corrosion resistance. where bright finish is

needed

 

Given a skin pass between cold rolling
No ZB operations between polishing rolls. Brighter Sheet metal applications
' than 2D and precursor to further finishing and industry equipment.

polishes

 

Commonly referred to as Bright Annealed Sheet metal, construction,

 

N 2B A finish (BA). Cold rolled using high polished and equipment
0' rolls for bright finish. Mirror finish similar to applications where bright
7 or 8. surface is needed.
Ground unidirectional finish using 80-100 Smfii finish) giggle r
No. 3 grit abrasive sanding. Starting finish for p g. g y

used in food or industry

further POhShmg applications.

 

 

Food service applications
such as countertops used
in restaurants and
delicatessens.

Ground unidirectional finish using 150 grit
No.4 abrasive sanding. Good general-purpose
finish subject to rough handling.

 

 

 

 

31

 

 

 

 

Table 1.8. (Con’t)
Finish produced using rotating cloth mops . .
(Tarnpico, fibre, muslin, or linen). Non- Sp ecralty decoration or
No. 6 . . . . construction. Not a typlcal
directional texture wrth varying f d l' t' f . h
reflectiveness. Referred to as satin blend. 00 app lca lon ““5 '
Buffed finish with high degree of Used where bright highly
No 7 reflectiveness. Produced using successively polished surface is needed
' finer and finer buffing compounds and (delicatessen displays,
abrasives. Minimal fine scratches remaining. equipment fascias).
Produced similarly to No. 7 finish with even US.“ where bright highly
. . . pollshed surface is needed
No. 8 higher degree of buffing. Flnal surface rs . .
. . . (medlcal industry). Costly
blemish free and true mlrror-llke appearance. . .
for food applicatlons.
Electrochemical process where phosphoric Commonly used in food
and sulphuric acids are used in conjunction industry applications for
Electro lishlwith high current density to clean and smooth food contact surfaces
p0 the surface. Raises the proportion of where high corrosion
chromium at the surface. Very bright mirror- resistance and mirror like
like finish. finish is needed

 

 

 

 

32

 

Table 1.9. Typical stainless grades and applications in the food industry (AISI,

2005).

 

Grade

Cost

Application

 

304

Most widely used of all stainless steel grades. Used for numerous food

Moderate equipment applications and contact surfaces. Moderate

corrosion/oxidation resistance and good weldability and physical
characteristics. Considered a versatile and reasonably priced stainless
grade for many applications.

 

304 L

A low carbon derivative of 304 as noted by the "L". Enhanced

"0‘19th corrosion/oxidation resistance with excellent weldability due to

lowered carbon content. Slightly lower physical strength at high
temperatures. Used where corrosion and acid resistance is necessary
land high temperature strength not a factor (i.e. food processing).

 

316

Moderate

Improved version of 304 with added molybdenum and slightly higher
'ckel content. Improved corrosion resistance over 304 with increased
hysical strength at low temperatures. Lower rate of general corrosion
with increased low temperature strength when compared to 304. Many
food application uses in low temperature corrosive environments such
as brines.

 

 

316 L

High

A low-carbon derivative of 316 as noted by the letter "L". Excellent
weldability and corrosion resistance at high temperatures. Limited use
lin food applications.

 

410

Low

lLowest alloy content of all basic stainless grades (304, 304L, 316,
316L, 430). A Low cost and general-purpose stainless steel grade.
Typical uses secondary handling of food products (packaging, transfer,
conveyance).

 

 

430

Low

 

A low carbon plain chromium stainless steel. Good
corrosion/oxidation resistance in mild environments. Brittle at low
temperatures. Limited use in food processing as a low cost alternative

 

ot used.

E0 mild food processing environments where strong organic acids are

 

33

 

1.7.3 Other food contact surfaces

Many studies have demonstrated the ability of L. monocytogenes to attach to
various materials and transfer to subsequent food products (Ak et al., 1994; Arnold
and Bailey, 2000; Beresford et al., 2001; Midelet and Carpentier, 2002, Schaffner et
al., 2004). Results from these studies suggest a need for increased scrutiny of food
contact surfaces and develop of new and innovative materials to reduce bacterial
attachment and subsequent transfer. In another study, the microflora on plastic and
wooden cutting boards was assessed by either soaking S-cm square blocks in nutrient
broth or by direct plating on nutrient agar (Ak et al., 1994). Escherichia coli, Listeria
innocua, L. monocytogenes, and Salmonella Typhimurium were readily recovered
from the plastic boards up to 12 h after inoculation. Recovery was less from wooden
boards with all pathogen populations decreasing 98% after 12 h. The absorption and
bactericidal activity of wooden cutting boards represented a significant improvement
over plastic cutting boards. This study clearly demonstrates the impact of material
composition, surface conditioning, and surface structure on bacterial attachment.

Work done by Satter et al. (2001) assessed transfer of Staphylococcus aureus
from fabrics (100% cotton and 50% cotton/polyester) to hands and other fabrics such
as bed linens and garments. Pre-moistened poly cotton had the highest transfer rate
(>20%) with the amount of friction significantly impacting transfer. Beresford et al.,
(2001) also evaluated various polymers and rubber for attachment and subsequent
release of Listeria after 2 h incubation. Lexan and polypropylene were found to shed

33% and 27% of the Listeria population, respectively, demonstrating the impact of

34

material composition, surface conditioning, and surface structure on bacterial

attachment.

1.8. Risk Assessment

Risk assessment is one of the many tools used by both national and
international governing organizations to identify chemical, biological, or physical
hazards to humans, plants, animals and environments alike. A food-safety risk
assessment is compiled to provide a framework for evaluating scientific data to
identify the potential for risk of illness and death to a population from exposure to a
foodborne pathogen. Several risk assessments developed over the last few years may
useful in instituting regulatory policies for control of foodborne pathogens with a
possible move away from the “zero tolerance” policy still being enforced in the

United States for L. monocytogenes in cooked and/or otherwise processed RTE foods.

35

1.8.1. FDA/CFSAN, USDA, and FAO/WHO

Since 2001, FDA/CFSAN and USDA have developed comprehensive risk
assessments for consumer exposure to L. monocytogenes through selected categories
of RTE foods with the latest revision made available in 2003. The aforementioned
risk assessments have ranked delicatessen meats as the leading product for the
establishment of listeriosis in pregnant women, neonates, and immunocompromised
adults.

Internationally, the Food and Agriculture Organization (FAO) of the United
Nations and World Health Organization (WHO) has compiled a risk assessment of L.
monocytogenes in RTE foods (FAQ/WHO, 2004). Various factors including raw
ingredients, processing, distribution, and consumption were included as part of the

risk analysis for listeriosis (Table 1.10)

36

Table 1.10. Variables affecting dose for risk of listeriosis (FAQ/WHO, 2004)

 

Point in Food

Variables Affecting Dose

 

Concentration in

Prevalence of

 

 

 

 

 

 

 

Continuum Consumption Contaminated unite Contaminated Units
Frequency and heating; mixing with breakdown to
amount other components smaller units/serving
consumed (e.g.; vinegar in portions

Consumption affected by: salads); breakdown to

season, wealth, smaller units

ages, sex,

culture/region
time, temperature, cross-contamination

Homellfood product composition with other foods
Serwce

time, temperature, packaging and cross
product composition, contamination, '

Retail Sale breakdown to smaller Hportioning,

units breakdown to
smaller units
time, temperature,
"“3”" and product composition
Storage
Volumetric changes: cross-contamination,
mixing with other mixing with other
ingredients changes bulk ingredients,
due to dilution or splitting into smaller
concentration units for retail/food

Processing (evaporation, removal service

of whey) Growth

inactivation changes

brining, heating steps,

holding times and

temperatures,

Environmental sources Season, harvest

affecting concentration area fodder and
Raw Ingredients in ingredients feeding regimes,

 

 

 

irrigation water, etc.

 

37

 

Based on current estimates, consumers purchase 24.4 and 75.6% of their
luncheon meats prepackaged and delicatessen-sliced, respectively (U SDA/F SIS,
2003). USDA/CFSAN data for Class I recalls of Listeria-contaminated sliced and
unsliced delicatessen meats clearly validates previous estimates for increased sales of
delicatessen-sliced as opposed to prepackaged product. Since 1994, a total of
81,623,410 pounds of RTE meat products have been recalled (USDA/F SIS, 2005).
From this total, 242,000 pounds were identified as manufacture-sliced or unsliced
(e.g. to be sliced at retail delicatessens) luncheon meats. From this total,
approximately 186,000 pounds or 77% was destined for retail slicing at delicatessens,
which further supports the current consumer preference for delicatessen-sliced meats
(Figure 1.2).

Revenue from the sale of delicatessen RTE luncheon meats has grown 4.6%
over the last 5 years topping more than 16 billion dollars (Uetz, 2005) (Figure 1.3).
In 2004, consumer spending on delicatessen-sliced lunch meats topped 3.1 billion
dollars compared to 255 million for unsliced products (Uetz, 2005) (Figuresl.4-1.5).
Sales projections for RTE luncheon meats are expected to exceed 22 billion dollars

by the year 2009 (Uetz, 2005).

38

 

 

..oocc~

.-°°cc¢

-Vococo

-cocom

peueoes spunod

..ocoooF

 

..ccoch

 

h.2593.

A88 £53.88
38 £3.32 Eng see @8225 a 8% a commence see commonsense Ba .8 have 2335 .S as»;

39

Lae>
voom meow NOON room ooom mam?

 

. . i ..
.....tt i... o
i .. I. 1\
oft . u i.
.v L o A
i i
. l . n
:1. . . .

coma

 

       
 

 

    

_ 8a.;

‘l
,

 

com. 3

 

a '0
p a .nr I
”1?. . u I
i... a. . '0
L L. . ..,. o
.t.
._
.. u
J. . -
v .. . coo Na
. -
c. .,J
..
. , .,. ,
. . . r
. ,

oomda

 

 

Good»

 
 

 

i oomdw

Goon .502 8?. use a855,. m2 enscowaé .3 9.5»:—

4o

Figure 1.4. Sliced RTE luncheon meat sales (Uetz, 2005)

   

,..
‘ .41..
x. ,,_. .1.

b*“!.‘

.'
at

 
 

;

 

 

.

lift-Cr ,1: H. t;

 

 

$3.200 —

$3.100 ~

$3,000 ~

41

o o
o o
“5’. ‘0.

N
89‘ as

20

2003

2002

2001

2000

1 999

Year

OOON OOON NOON OOON mOON VOON

| --.. O l.... Ill 11. I- Fl- -llilllill ll! L I .Iilll ll.1.i-li-.ki|ill- . .i I . ti - .1»..- Iiii.l|..l..

r. 9»
t»
a;
9.»

ON»

suollllS

. 9N»

NNw

 

a. 3.»

 

a «we

38m .503 was as: SBQE m: Becca .3 2:3..—

42

Given the previously reported market trends, post-process contamination of
RTE luncheon meats will be a serious concern for many years to come. Work done
by Uyttendaele et al. (1999) in which 4.9% of cooked meat products sampled at retail
markets tested positive for L. monocytogenes emphasizes the risk of listeriosis to the
consuming public.

With an estimated 320 fatal listeriosis cases each year from RTE meat
product, 242 deaths of which are associated with delicatessen-sliced luncheon meats,
minimizing contamination at delicatessens will clearly have a major impact on
meeting the goals of Healthy People 2010 (2004), the Presidential directive for
reducing the number of listeriosis cases (and other foodborne illnesses) by 50%. The
current estimate of 2500 cases annually in the United States has seen a significant
reduction over the past 5 years. Laboratory confirmed listeriosis cases have declined
from 0.47 cases per 100,000 in 2000 to 0.26 cases per 100,000 in 2005, approaching

the target of 0.25 cases per 100,000 to be achieved by 2010.

1.9. Predictive modeling of microbial growth and transfer

Within the last decade risk assessments have promoted the development of
more dynamic models that include changes in bacterial survival during distribution
and subsequent storage. Reviews done by Buchanan et al., (1997) and Davies (1993)
describe the many approaches to experimental design, techniques and approaches to
modeling growth, transfer, and dissemination of pathogens in the microbiology,

mathematical, engineering and regulatory disciplines.

43

1.9.1. History and development of microbial modeling

The early years of predictive microbiology centered on growth studies and
thermal inactivation at high temperatures, both of which were defined by log-linear
relationships (Beck and Arnold, 1977; Bernaerts et al., 2004; Schaffner et al., 1998;
Zhoa and Schaffner, 2001) which were developed for food safety and quality
assurance programs. Recent cross-disciplinary collaborations have prompted the
development of a broad range of dynamic models using microbial, mathematical, and
environmental parameters to predict population outcomes at various growth and

transfer conditions. (Bernaerts et al., 2004).

1.9.2. Techniques for predictive modeling of microbial growth and transfer

Many growth models, beginning with the USDA Pathogen Modeling Program
Version 2.1 (Buchanan and Phillips, 1990) have been developed to predict the grth
of Listeria and other foodborne pathogens in foods based on pH, storage temperature,
and levels of salt and sodium nitrite (Houtsma eta1., Conner et al., 1986; Le Marc ct
al., 2002; and Tamplin, 2002). The most recent USDA model developed by Tamplin
comes with a pre-programmed graphical user interface and generates graphs and
tabular output for various growth parameters (Tamplin, 2002). This model is freely
available to both the public and private sector for estimation of contamination and
risk of exposure under various environmental conditions.

For empirical modeling it is desirable to solve for the minimum number of
parameters necessary to adequately fit the data. Using a minimum number of

parameters increases the degree of stability in the parameter estimation procedure and

44

the greatest degree of confidence in the calculated parameters. The method of least
squares is chosen because it is a simple method and the results are the same as those
obtained by maximum likelihood and Gauss-Markov, assuming that the following
statistical assumptions are valid, as given by (Beck and Arnold, 1977).
1. The measurement errors are additive in nature to the true (but unknown)
bacterial count.
2. The measurement errors, considered over the duration of the experiment,
have mean value of zero.
3. The measurement errors have a constant variance over the duration of the
experiment.
4. The magnitude of each measurement error is unrelated to its predecessor
or successor. In simple terms, the errors are not related.
5. The measurement errors, considered over the duration of the experiment,

fall in a normal, or Gaussian, distribution pattern.

Although in most experiments it is difficult to obtain detailed information
about experimental errors, the assumptions listed above are not unreasonable for most

types of measurements.

1.9.3. Bacterial growth and thermal inactivation models
Mathematical models were first developed to predict bacterial growth and
thermal inactivation of foodborne pathogens in various substrates with little regard to

the bacterial contamination and transfer that occurs up to the time of consumption

45

(Bernaerts et al., 2004). Schaffner (2004) recently discussed the framework for
developing models that can predict the extent to which pathogens can be transferred
from the food processing environment to the final product. Schaffner defined the
relationship between the raw product, environment, and finished product as shown in

Figure 1.6.

46

Figure 1.6.

Mathematical framework for relationship of raw product, environment

and finished products

Raw product CFU x Cross-Contamination Rate = Environmental

CFU 1

Environmental CFU x Persistence Rate = Environmental Reservoir

CFU 1

Environmental Reservoir CFU x Cross-Contamination Rate = Product

1

Product Contact Surface CFU x Persistence Rate = Product Contact

1

Product Contact Reservoir CFU x Cross-Contamination Rate =

Contact Surface CFU

Surface Reservoir CFU

Finished Product CFU

47

This framework helps clarify the experimental design and mathematical
manipulations for predictive modeling of cross-contamination and subsequent transfer
of foodborne pathogens. This model also illustrates the additive effect of each

parameter to the framework of the model where each fraction of transfer “ f, ” is an
additive function of the previous fraction or f, = fa * f, where

“ fa ” = raw product and

“ f, ” = cross contamination rate

This simplistic approach to predicted distribution of pathogens from raw to
finished product is a precursor to more dynamic models including the use of
empirical data with parameter estimation and mathematical manipulation (Bernaerts
et al., 2004).

In most models, empirical data is fitted to mathematical equations using
parameter estimation techniques. Hybrid models that use a combination of empirical
data fitting techniques and mathematical manipulations or mechanistic mathematical
translations have been described by Bemaerts et a1, 2004. These models which can
be used to develop manufacturing and mathematical parameters build on parameters
described by empirical data while applying mathematical manipulation to each of the
parameters as a function of total transfer.

Predictive modeling of microbial pathogens during food production and
storage has been approached using previously published models and methods
(Bernaerts et al., 2004; Buchanan and Philips, 1990; Houtsma et al., 1996). Each of
these approaches has many advantages and disadvantages and has been the subject of

much debate. While predictive models based on mathematical translation of

48

biological functions can be rapid and less costly than empirical models that require
little or no laboratory experimentation, they can be greatly influenced by
environmental factors not realized in a laboratory setting. In contrast to mathematical
translation, curve fitting models that predict population outcomes based on previously
obtained experimental data offer an arguably more accurate interpretation of
predicted environmental populations. However, these models are costly and may not
account for underlying biological parameters and in some cases may be dependent on
specific environmental or laboratory conditions. In limited work by Schaffner et al.
(2004), modeling of bacterial transfer to and from food contact surfaces was done
using a Monte Carlo simulation for plastic cutting boards used in a food service
kitchen for raw meats and vegetables over a 2-week period. Results from this study
and subsequent simulations predicted a contamination level greater than 20
CFU/4cm2 after 15 min and greater than 40 CFU/4cm2 after 45 min. While this study
provides some insight into modeling total bacterial transfer during slicing of deli
meats, other important parameters including surface scoring, surface roughness,
cutting force, and physiological differences in bacterial attachment were not
addressed.

In conclusion, empirical data obtained from three years of laboratory work
was used to mathematically model the relationship between cutting and slicing of
RTE delicatessen meats and transfer of L. monocytogenes. This research was
performed using a commercially available slicing machine, specially fabricated
stainless steel kitchen knives and retail delicatessen meats. The previously reported

incidence of Listeria contamination in RTE foods is considered a main factor

49

impacting further contamination of Listeria-free foods during subsequent handling at
retail. This research was conducted after Listeria transfer rates were identified as a
key informational gap in the Draft Listeria Risk Assessment that was published by the

federal government in September 2003 (FDA/FSIS/CDC, 2003).

50

CHAPTER 2

IMPROVED QUANTITATVE RECOVERY OF LIS T ERIA MONOCYTOGENES

FROM STAINLESS STEEL SURFACES USING A l-PLY COMPOSITE TISSUE

Vorst, K.L, Todd, E.C.D., Ryser, E.T.

Journal of Food Protection. 2004. 67 :2212-2217

51

2.1. ABSTRACT

Four sampling devices a sterile environmental sponge (ES), a sterile cotton-tipped
swab (CS), a sterile calcium alginate fiber-tipped swab (CAS), and a l-ply composite
tissue (CT), were evaluated for quantitative recovery of Listeria monocytogenes from a
food-grade stainless steel surface. Sterile 304 grade stainless steel plates (6 x 6 cm) were
inoculated with approximately ~ 106 CFU/cm2 L. monocytogenes strain Scott A and dried
for 1 h. The ES and CT sampling devices were rehydrated in phosphate buffer solution
(PBS). After plate swabbing, ES and CT were placed in 40 ml of PBS, stomached for 1
min and hand-massaged for 30 seconds. Each CS and CAS device was rehydrated in
0.1% peptone before swabbing. Afier swabbing, CS and CAS were vortexed in 0.1%
peptone for 1 min. Samples were spiral-plated on Modified Oxford Agar (MOX) with
MOX Rodac Contact” plates used to recover any remaining cells from the stainless steel
surface. Potential inhibition from CT was examined in both PBS and in a modified disc
diffusion assay. Recovery was 2.70, 1.34, and 0.62 log greater using CT compared to ES,
CS, and CAS, respectively, with these differences statistically significant (P<0.001) for
ES and CT and for CAS, CS and CT (P<0.05). Rodac® plates were typically overgrown
following ES, positive after CS and CAS, and negative afier CT sampling. CT was non-
inhibitory in both PBS and the modified disc diffusion assay. Using scanning electron
micrscopy, Listeria cells were observed on stainless steel plates sampled with each
sampling device except CT. The CT device, which is inexpensive and easy to use,
represents a major improvement over other methods in quantifying L. monocytogenes on

stainless steel surfaces and is likely applicable to enrichment of environmental samples.

52

2.2. INTRODUCTION

The public health significance surrounding Listeria monocytogenes has led to a
regulatory policy of "zero tolerance" in the United States for this organism in certain
ready-to-eat (RTE) foods. Listeria monocytogenes is a foodborne pathogen of major
concern due to its high fatality rate (20-30%) (Ryser and Marth, 1999) its persistence
(Tompkin, 2001) and spread in food-processing environments (Lunden et al., 2002;
Vogel et al., 2001), and ability to grow to populations of 104 to 106 CFU/g on many
refrigerated RTE foods such as smoked salmon and luncheon meats (Gombas et al.,
2003; Ryser and Marth, 1999). Post-process contamination of cooked/RTE delicatessen
products with L. monocytogenes has resulted in at least two major outbreaks (MMWR,
2000) and over 80 recalls involving more than 130 million pounds of product, making
this pathogen the leading cause of Class I microbiologically related recalls (Levine et al.,
2001). However, consumption of low levels of L. monocytogenes in food is not
uncommon and is not considered a significant risk to most people (Chen et al., 2000).

Environmental sampling within food processing facilities has been plagued by
poor repeatability and efficacy when compared to newer methods used in the medical and
pharmaceutical industries (Richard and Piton, 1986; Tominaga et al., 2001; Yamaguchi et
al., 2003). Several of these newer direct plating methods that utilize adhesive sheets
(Yamiguchi et al., 2003) and pads (Tominaga et al., 2001) are reportedly superior to
traditional swabbing in terms of both recovery and reproducibility for bacteria on medical
devices and in pharmaceutical products. However, recovery of pathogens from
equipment used in foodservice and processing environments poses a major hurdle for

these newer methods because of the presence of numerous food particulates.

53

Existing strategies that have been developed to assess presence or absence of
Listeria in food processing environments lack the needed sensitivity to quantify Listeria
on solid surfaces (Gombas et al., 2003; Midelet and Carpentier, 2002). The efficiency
and reliability of traditional environmental sampling devices such as the sponge and
cotton swab have been debated since their introduction in the 1970's (Ware et al., 1999).
Studies on meat surfaces also have shown that traditional destructive surface sampling
methods such as excision yield higher bacterial recovery (2 50%) than non-destructive
swab and sponge methods (Gill et al., 2001).

Quantitative transfer to and from slicing machines, knives, and cutting boards in
delicatessens was identified as a key informational gap in the 2003 Draft FSIS risk
assessment for Listeria in RTE meat and poultry products (FDA/F SIS, 2003). In limited
work, Humphrey (1990) evaluated retail delicatessen meat slicers in the UK and found L.
monocytogenes on 10 of 32 slicer blades, thus suggesting ample opportunity for Listeria
transfer.

In other work, the microfiora on plastic and wooden cutting boards was assessed
by either soaking 5-cm square blocks in nutrient broth or by direct plating on nutrient
agar (Ak et al., 1994). Escherichia coli, Listeria innocua, L. monocytogenes, and
Salmonella Typhimurium (106 CFU) were readily recovered from plastic boards up to 12
h after inoculation. Recovery was less from wood blocks with bacterial populations
decreasing 98% after 12 h. This study indicates the potential for pathogen transfer from
soiled surfaces after extended holding times. Transfer of Staphylococcus aureus from
fabrics (100% cotton and 50% cotton/polyester) to hands and other fabrics such as bed

linens and garments also has been assessed (Satter et al., 2001). The highest transfer rate

54

(>20%) was seen with moist poly cotton, with friction from moist or re-moistened fabrics
significantly impacting the transfer rate.

Foodborne pathogens are also easily transferred in domestic kitchens during
common food handling practices (Chen et al., 2000; Satter et al., 2001). One study
showed that 40 and 60% of samples from knife handles, chopping boards, wash clothes
tested positive for Salmonella and Camplyobactor, respectively, after contacting
contaminated chicken during normal kitchen usage. A correlation between frequency and
level of exposure and dose-response was identified as a key element in predicting the
relative risk of foodborne pathogens (Tamplin, 2002).

Several studies have used scanning electron microscopy (SEM) to evaluate
biofilm formation and bacterial attachment to 304 grade stainless steel of different
surface finishes (Arnold and Bailey, 2000; Arnold and Silvers, 2000; Kalmokoff et al.,
2001; Mafu et al., 1991). When Arnold and Bailey (2000) used a mixed culture,
attachment to stainless steel was 1 log greater on 304 stainless with a ZB or rough finish
when compared to 304 electropolished stainless with a mirror-like finish. Thus, SEM can
serve as another means to assess recovery of bacteria from stainless steel surfaces.

Given the importance of risk assessments in determining the most vulnerable
steps for contamination and growth in a food processing operation, quantification of
microbial contaminants on food contact surfaces has become an integral component in
establishing the degree of risk to the public. Hence, the objective of this study was to
compare the l-ply composite tissue (CT) to the environmental sponge (ES), cotton-tipped
swab (CS) and calcium alginate swab (CAS) for quantitative recovery of L.

monocytogenes from stainless steel surfaces.

55

2.3. MATERIALS AND METHODS

2.3.1. Preparation of strains

Listeria monocytogenes strain Scott A (GT 3864) was obtained from Dr. Joseph
Madden (Neogen Corp., Lansing, MI) and maintained at -80°C in trypticase soy broth
(TSB) (Difco/Becton Dickinson, Sparks, MD) containing 10% (v/v) glycerol. TSB
containing 0.6% yeast extract (TSB-YE) (Difco) was inoculated from the frozen stock
culture and incubated for 22-24 hours at 37°C. After a second transfer in TSB-YE, the
culture was pelleted by centrifugation at 9700 x g / 10 min / 4°C (Sorvall Super T21;
Sorvall Products, L.P. Newton, CT) and resuspended in 9 ml of 0.1% peptone (Difco).
Cell concentration was determined by spiral plating (Autoplate® 4000 Spiral Plater;
Spiral Biotech Inc., Norwood, MA) on trypticase soy agar containing 0.6% yeast extract

(TSA-YE) followed by 48 h of incubation.

2.3.2. Stainless steel preparation and inoculation

Unpolished scratch-free grade 304 sanitary stainless steel plates measuring 6 cm x
6 cm x 0.145 cm were obtained from ProAxis, Inc. (Lafayette, IN). Plates were
autoclaved at 121°C for 15 min and then inoculated. Afier use, the plates were treated
with mineral oil to prevent surface oxidation. To remove mineral oil before the next use,
the plates were rinsed in sterile deionized water, flamed with 95% ethyl alcohol, and then

autoclaved for inoculation.

56

2.3.3. Sampling devices

Four Listeria recovery devices were assessed - a sterile environmental sponge
(ES) (Nasco Speci-Sponges®; NASCO, Fort Atkinson, WI), a sterile cotton-tipped swab
(CS) (Pur-Wraps® Cotton Tipped Applicator; Harwood Products Co. LLC, Guilford,
MA), a sterile calcium alginate fiber-tipped swab (CAS) (Fisherbrand® Sterile Swabs;
Curtin Matheson Scientific, Houston, TX), and a l-ply white 11.4 x 21.5 cm tissue (CT)
with a basis weight of 11.2 lbs and 3 point thickness (Kim-wipe® Ex-L 1- ply white
tissue; Kimberley-Clarke Corp., Roswell, GA). The culture suspension (100,111) was
spotted on stainless steel plates, uniformly spread with a sterile inoculating needle to
obtain an inoculum level of ~106 L. monocytogenes CFU/cm2 and then allowed to dry for

1 h at ~23°C in a laminar flow hood.

2.3.4. CT and CAS

All CS and CAS devices were rehydrated in PBS with CS and CAS absorbing 0.1
and 0.2 ml of PBS, respectively. Thereafter, the stainless steel plates were swabbed 10
times vertically and horizontally while rotating the swab between movements at a 30°
angle as stated in the Compendium of Methods for the Microbiological Examination of
Foods (Sveum et al., .1992). Thereafter, swab applicators were vortexed (Genie 2;
Scientific Industries Inc., Bohemia, NY) in 10 ml of PBS for 60 sec to release Listeria.
Samples (50 ul) were then spiral-plated on duplicate plates of Modified Oxford Agar
(MOX) which were incubated 48 h at 35°C to determine numbers of L. monocytogenes

recovered.

57

2.3.5. ES and CT
The CT was folded twice from the side and top edges so as to measure 5.5 cm x

5.5 cm, producing a clean interior and exterior surface which eliminated contact between

gloves and the area sampled (Figure 2.1).

Figure 2.1 Folding patter of CT

 

lst fold

   

2nd Fold. l

 

Opposite of C

58

Using disposable gloves, the ES and CT devices were rehydrated with 10 mL of
PBS in a 24 oz Whirl-Part“ bag (NASCO; Fort Atkinson, WI) with ES and cr absorbing
9.8 mL and 0.7 ml of PBS, respectively. Afier squeezing the opened CT inside the sterile
Whirl-Pak" bag to remove excess diluent, the stainless steel plate was swabbed 10 times
vertically and horizontally with ES and with CT using the folded exterior surface for CT.

After sampling, each ES and CT was returned to the same Whirl-Pak" bag with the CT

partially unfolded (Figure 2.2).

59

Figure 2.2. CT before (A) homogenization

 

60

PBS (40 mL) was added after which the device was homogenized in a Stomacher
400 (Seward, London, UK.) for 60 sec and then hand-massaged for 30 sec. After
homogenization, the or was unfolded inside the Whirl-Pak® bag (Figure 2.3) and a so
uL aliquot was spiral-plated in duplicate on MOX as previously described for
enumeration of L.’ monocytogenes.

Each ES, CS, and CT recovery test was replicated 5 times using 12 plates (n = 60)
with CAS being replicated 3 times (11 = 36). After sampling the stainless steel plates with
ES, CS, CAS, and CT, Rodac® plates containing MOX were used to quantify any
remaining Listeria with these plates incubated 48 h at 35°C. Rodac“) contact plating was

replicated 3 times (n=36).

2.3.6. Potential inhibition in PBS
Sterile Whirl-Pelt“) bags containing 50 ml of PBS at 21-23°C were inoculated to
contain 108 L. monocytogenes CFU/ml. After homogenizing in a Stomacher for 60 sec, a
0.1 ml aliquot of the homogenate was spiral-plated on TSAYE to determine the initial
population. One CT device was then added to PBS with PBS similarly examined for
numbers of Listeria after 5, 15, and 30 min of exposure. Inhibition studies for each time

interval were conducted in triplicate using 10 sample bags (n=30).

2.3.7. Modified disc diffusion assay
TSAYE (20 ml) was inoculated with 100ul of an overnight L. monocytogenes
culture (108 CFU/m1) and poured into standard 100-mm diameter Petri dishes. A 25-mm

diameter disc was aseptically cut from the CT and placed in direct contact with the agar

61

surface after solidification. Plates were visually examined for inhibition zones after 24 h

of incubation at 35°C. This assay was replicated twice using 10 inoculated Petri dishes

with CT (n=20).

2.3.8. Evaluation of recovery methods using SEM

A field emission scanning electron microscope (CamScan 44F E; CamScan USA
Inc., Cranberry Twp., PA) was used to visually evaluate recovery of Listeria from
duplicate stainless steel plates after sampling with the four different devices. After
sampling, each stainless steel plate was treated with 4% formaldehyde (5 min) to fix any
remaining cells to the plate and then dehydrated using a series of 25%, 50%, 75% and
95% ethanol for 20 min. After the final dehydration in 95% ethanol, plates sampled
using each of the four sampling devices were air-dried for 60 min., placed in the SEM

chamber and scanned from end to end.

2.3.9. Statistical analysis
Data were analyzed using a general linear model (GLM) with a general
randomized complete block design used to compare ES, CS, CAS, and CT devices for

quantitative recovery of L. monocytogenes (SAS, 1996).
2.4. RESULTS

CT yielded the best recovery with populations 1.11 to 2.70 log/cm2 higher when

compared to the other methods (Figure 2.4).

62

Figure 2.4. Recovery of L monocytogenes from stainless steel.

Log CFUI cmz

 

ES CS CAS CT Initial

Means with different letters are significantly different (P < 0.05).

63

Least squares means by average log count were compared for effect of the method
and subjected to the least significant difference test. Differences between CT and the

other devices were statistically significant (P<0.05) (Table 2.1).

Table 2.1. Least squares means for effect of method Pr > It] for Ho: LSMean (i) =

LSMean (j) with log count as the dependent variable.

Dependent Variable: Log count

 

 

Method (i/j) ES CS CAS CT
ES <0.001 <0.001 <0.001
CS <0.001 >0.05 <0.05
CAS <0.001 >0.05 <0.05
CT <0.001 <0.05 <0.05

 

ES, which is recommended in most environmental testing protocols, was least
effective in quantitatively recovering Listeria from stainless steel. Differences between
these devices were statistically significant (P<0.001) with these responses closely
following the observed sampling errors for ES.

Rodac® plates were used to confirm the presence of L. monocytogenes on
stainless steel surfaces after swabbing. Confluent grth of Listeria was seen for
stainless steel plates previously sampled with ES with these Rodac“) plates being
overgrown. Following CS and CAS, Rodac® counts averaged 2.0 log L. monocytogenes
CFU/cmz, whereas no Listeria were detected after CT.

Subsequent CT testing showed no inhibition of L monocytogenes in PBS with

counts of 8.36 and 8.20 log L. monocytogenes CFU/ml after 30 min of exposure. The

64

modified disc-diffusion assay was also negative with no inhibition zone evident after 24 h
of incubation.
As seen by SEM, Listeria cells were present on the stainless steel plates after

recovery using each of the devices except CT (Figures 2.5-2.8).

65

Figure 2.5. Scanning electron micrograph of Listeria attached to stainless steel plates
after recovery using ES device.

 

3m

Figure 2.6. Scanning electron micrograph of Listeria attached to stainless steel plates
after recovery using CS device.

 

66

Figure 2.7. Scanning electron micrograph of Listeria attached to stainless steel plates
after recovery using CAS device.

 

lum—

Figure 2.8. Scanning electron micrograph of Listeria attached to stainless steel plates
afier recovery using CT device.

 

67

Scanning was performed uniformly after inoculation and recovery from stainless
steel plates using each device. Additional scans were performed after CT to confirm

absence of Listeria cells on the stainless steel plates.

2.5. Discussion

Quantifying pathogens on solid surfaces provides valuable risk assessment data
for modeling consumer exposure from cross-contamination in food manufacturing and
foodservice environments (International Commission on Microbiological Specifications
For Foods (ICMSF) Working Group on Microbial Risk Assessment, 1998). Approved
for environmental sampling, ES, CS and CAS are primarily used for bacterial enrichment
rather than quantification. Numerous studies have shown that these sampling devices are
often awkward to use and inefficient (Pinner et.al. 1992; Richard and Piton, 1986; Salo
and Laine, 2000; Scott et al., 1984; Ware et al., 1999;Yamaguchi et al., 2003).
Furthermore, CS and CAS as well as Rodacc’ plates are impractical for quantitatively
sampling large heavily soiled areas and typically yield results that are difficult to interpret
(Miettinen, 2001).

Listeria monocytogenes can reportedly attach to stainless steel, glass,
polypropylene, and Buna-N rubber in as little as 20 min at 4 and 20°C (Akier et al., 1990;
Beresford et al.,2001; Briandet et al., 1999; Djordjevic et al., 2002; Herald and Zottola,
1988). Characterization of stainless steel has played an important role in understanding
how bacteria attach to these surfaces. Studies using contact angle measurements have
shown little difference between flat and penicylinders in terms of the total surface energy

for stainless steel (Herald and Zottola, 1988; Hood and Zottola, 1997). The effect of

68

various cleaners on removal of L. monocytogenes from food contact surfaces also has
been assessed with the pathogen rapidly attaching to stainless steel and becoming
resistant to chemical sanitizers (Krysinski et al., 1992). Hence, proper cleaning and
sanitizing is necessary to minimize biofilm formation and subsequent transfer of Listeria
in delicatessens and food service environments.

The ES device is well suited for qualitative sampling of large heavily soiled areas
and offers numerous advantages over CS and CAS. However, the porous cellulose
matrix of ES is well known for entrapment of bacteria (Gill et al., 2001; Moore and
Griffith, 2002; Salo and Laine, 2000; Ware et al., 1999), which greatly hinders any type
of quantitative analysis.

Based on size limitations, CS and CAS are typically used to assess bacterial
contamination in cracks, narrow tubing, crevices, joints, and any other difficult-to-sample
areas on equipment. While superior to ES, the small size of CS and CAS makes these
devices poorly suited for sampling large flat areas such as floors and countertops. While
CAS afforded better release of Listeria when compared to CS (P<0.05), the size of the
device was limiting.

Using fiber-tipped swabs, Moore and Griffith (2002) reported that the amount of
mechanical energy was more important than the type of wetting solution [1/4 strength
Ringer’s solution, 2-N-Morpholino-ethanesulfonic acid, Tris buffer, 3% Tween,
Spraycult® (a disintegrating reagent)] for recovering Salmonella from stainless steel
surfaces. Although Salmonella recovery increased 16% using ‘/4 strength Ringer’s
solution compared to the other wetting solutions, the rate of Salmonella release from

swabs pre-moistened with 1/4 Ringer’s solution was lower (85.2%) compared to dry

69

swabs (88.2%). The amount of mechanical energy and the type of swab both played
pivotal roles with the coarse foam swab having greater recovery (70.4%) when compared
to cotton (69.6%), dacron (38.2%), or alginate (55.8%). These findings closely follow
our CT data, suggesting that the coarse CT composition enhances scouring of the surface
to remove attached cells at a lower application force. Increased recovery using CT may
also be due to the inherent antistatic coating, which would aide in the release of bacteria
by reducing the electrostatic discharge commonly seen in fiber-tipped swabs.

When CT was used similarly to the other devices, Rodac" plates were typically
negative after CT and positive after ES and CS. Following ES, CS, and CAS sampling,
Listeria cells were readily detected using SEM (Figures 2.5-2.8). However, no Listeria
cells were seen on stainless steel plates in repeated scans after CT sampling (Figure 2.8).-

Differences seen between initial inoculum and final recovery can be attributed to
entrapment of Listeria cells within the CT device. The problem of entrapment within
sampling devices has been of great concern when evaluating bacterial recovery. CT does
not have the large porous structure of ES, thus allowing for greater release of Listeria.
However, CT still likely entraps some cells, as suggested by our data.

CT was advantageous over the other three sampling devices in terms of
repeatability and recovery. On a per test basis, the cost of the CT device was also
advantageous at $0.02 per device compared to $0.73, 0.19, and 0.10 for each ES, CAS,
and CS, respectively (Fisher Scientific; Pittsburgh, PA). Like other currently used
environmental sampling devices, CT is amenable for enrichment of environmental
samples. The >100-fold increase in recovery of L. monocytogenes from stainless steel,

combined with ease of use and low cost, makes CT an ideal sampling device for

70

quantitative (and potentially qualitative) assessment of contamination on hard-to-clean

surfaces such as delicatessen slicing blades.

2.6 Summary

Overall findings in the study demonstrate the improved efficacy of CT for Listeria
recovery from stainless steel when compared to traditional ES, CS, and CAS devices. The
CT device was chosen to quantify transfer of L. monocytogenes from RTE meats to knife
and slicer blades and vice versa in Chapters 3-5. The enhanced ability of CT to detect low
levels of bacterial contamination on food contact surfaces and processing equipment will
aide in the development of more accurate risk assessments that directly address post-
processing contamination in food processing and foodservice establishments. Given the
inadequacy of current sampling devices to recover bacteria from large and heavily soiled
areas, the CT device represents a significant improvement in quantification of bacteria on

food contact surfaces.

71

CHAPTER 3

TRANSFER OF LIS T ERIA MONOC Y T OGENES DURING MECHANICAL

SLICING OF TURKEY BREAST, BOLOGNA, AND SALAMI

Vorst K.L., Todd, E.C.D., Ryser, E.T.

72

3.1. ABSTRACT

A commercial delicatessen slicer was used as the vector for sequential
quantitative transfer of Listeria monocytogenes from (a) an inoculated slicer blade (~108,
105, 103 CFU/blade) to 30 slices of uninoculated delicatessen turkey, bologna, and
salami, (b) inoculated product (~108cm2) to the slicer and (c) inoculated product (108,
10‘, 103 CFU/cmz) to 30 slices of uninoculated product via the slicer blade with cutting
force and product composition also assessed for their impact on Listeria transfer. Five
product contact areas on the slicer identified using product bathed in Glow GermTM were
also sampled after slicing inoculated product using a l-ply composite tissue technique.
After slicing with inoculated blades, each slice was surface- or pour-plated using
Modified Oxford Agar and/or enriched in University of Vermont Medium. Greater
transfer (P<0.05) was seen from inoculated turkey (108 CFU/cmz) to the five slicer
contact areas using a cutting force of 10 as opposed to 0 lbs. Using slicer blades
inoculated at 108 CFU/blade Listeria populations decreased logarithmically to 102
CFU/slice after 30 slices. Findings for inoculated slicer blade and product (105
CFU/blade or cmz) were similar with Listeria counts of 102 CFU/slice after 5 slices and
enriched samples generally negative after 27 slices. Using 103 CFU/blade, the first 5
slices typically contained ~10‘ CFU/slice by direct plating with enrichments negative
after 15 slices. The higher fat and lower moisture content of salami compared to turkey
and bologna produced a fat layer on the blade that prolonged Listeria transfer. When
cross-contaminated, delicatessen-sliced meats allowing growth of Listeria in home

refrigerators may pose an increased public health risk for certain consumers.

73

3.2. INTRODUCTION

Listeria monocytogenes has long been viewed as a serious post-processing
contaminant with this pathogen residing in some food processing facilities for many years
(Beresford et al., 2001; Tominaga ct al., 2001). Endemic strains that persist in food
manufacturing environments possess greater ability to adhere to food contact surfaces
(Beresford et al., 2001; Chmielewski and Frank, 2003; Lunden et al., 2000) with some
strains attaching to stainless steel in as little as 20 minutes (Mafu, 1991). In one study,
Lunden et al. (2002) demonstrated plant-to-plant transfer of L. monocytogenes via a
dicing machine with the same Listeria strain identified at three different facilities. Thus,
processing equipment and other food contact surfaces can serve as vectors for the spread
of Listeria during food manufacture.

Transfer of pathogens through slicing machines was recognized over 40 years
ago. An outbreak in Aberdeen, Scotland led to 469 cases of typhoid fever (International
Commission on Microbiological Specifications For Foods (ICMSF) Working Group on
Microbial Risk Assessment, 1998). The contaminated corned beef was delicatessen
sliced resulting in transfer to other meat products via the contaminated slicer over several
days. In 1990, Humphrey (1990) recovered L. monocytogenes from 10 of 32 retail
delicatessen slicers surveyed in the United Kingdom. Three years later, Hudson and Mott
(1993) reportedly isolated L. monocytogenes from a delicatessen knife and slicing
machines in Amsterdam supermarkets with the pathogen also found at most sites near a

display case of processed meats.

74

Listeria monocytogenes is now a well-recognized contaminant of delicatessen
products with Sauders et al. (2004) having identified this pathogen in smoked salmon,
deli meats and cheeses, hot dogs, and seafood from 20 of 47 retail food establishments
surveyed in New York State. In a large-scale survey by Gombas et al. (2003), the
incidence of L. monocytogenes was approximately seven times greater in delicatessen-
sliced (0.4%) as opposed to manufacture-sliced luncheon meats (2.7%) with difficult-to-
clean delicatessen slicers and other food contact surfaces presumably being responsible
for the higher contamination rate. These findings, along with a report indicating that 75%
of consumers purchase delicatessen-sliced rather than pre-packaged luncheon meats
(Gombas, 2003), suggest substantial consumer exposure to Listeria.

Four major listeriosis outbreaks have been documented in the United States since
May of 2000, three of which were traced to consumption of delicatessen-sliced turkey
breast, (MMWR 2000, MMWR 2002). These three outbreaks were responsible for a
combined total of 91 listeriosis cases, including 11 deaths and 6 miscarriages, in 22 states
and the recall of 44.3 million pounds of product. These outbreaks prompted the
development of three USDA-mandated alternatives for controlling Listeria in
delicatessen meats — (a) post-package pasteurization, (b) product reformulation to prevent
Listeria growth and/or (c) increased product and environmental testing (FSIS, 2003) and
also raised concerns regarding current food handling practices at the retail level as
specified in the Food Code (FDA/CF SAN, 2001). Based on a FSIS risk assessment for
Listeria in ready-to-eat meat and poultry products (FSIS, 2003 ), 242 of the estimated 500

listeriosis fatalities each year are thought to be traceable to delicatessen meats. Thus,

75

minimizing contamination at delicatessens will clearly have a major impact on reducing
the incidence of listeriosis and in meeting the goals of Healthy People 2010 (2004)

Given that the numbers of Listeria transferred between commercial slicing
machines and delicatessen meats was cited as both a major public health concern and a
key data gap in several Listeria risk assessments (FSIS, 2003; USDA/FSIS, 2003), the
objectives of this study were to assess: (a) impact of cutting force on transfer of L.
monocytogenes from contaminated RTE luncheon meats to a delicatessen slicer, (b)
transfer of L. monocytogenes from an inoculated delicatessen slicer blade to uninoculated
roast turkey, salami and bologna, (c) transfer of L. monocytogenes from inoculated
product to a delicatessen slicer and then to uninoculated product, and (d) slicer blade

wear over a 2-year duration.

3.3. MATERIALS AND METHODS

3.3.1. Listeria monocytogenes strains

The following six strains of Listeria monocytogenes (obtained from Dr. Catherine
W. Donnelly, University of Vermont, Burlington, Vermont): CWD 205 (source
unknown), CWD 578 (dairy plant), CWD 701 (Azore cheese), CWD 730 (dairy plant),
CWD 845 (dairy plant), and CWD 1002 (pork sausage) were chosen from a set of more
than 190 strains based on their ability to form weak (CWD 205, CWD 578), medium
(CWD 701, CWD 1002) or strong (CWD 730, CWD 845) biofilms in a microtiter plate
assay (Keskinen et al., 2003). All strains were maintained at -80°C in trypticase soy

broth (TSB) (DifcofBecton Dickinson, Sparks, MD) containing 10% (v/v) glycerol. TSB

76

containing 0.6% yeast extract (TSB-YE) (Difco) was inoculated from the frozen stock
cultures and incubated for at 37°C for 24 h. After a second transfer in TSB-YE, each
culture was pelleted by centrifugation at 9700 x g / 10 min / 4°C (Sorvall Super T21;
Sorvall Products, L.P. Newton, CT), resuspended in 9 ml of 0.1% peptone (Difco) and
combined in equal volumes to produce one 6-strain cocktail containing approximately
108 CFU/ml. Cell concentration was determined by optical density at 600 nm and spiral
plating (Autoplate® 4000 Spiral Plater; Spiral Biotech Inc., Norwood, MA) an
appropriate dilution on Modified Oxford (MOX) agar followed by 48 h of incubation at

35°C.

3.3.2. Delicatessen meats

One retail brand each of restructured roast turkey breast, Genoa hard salami and
bologna (5.5 to 6.5 lbs each) was purchased in chub-form from a local retailer (Gordon
Food Service, Lansing, MI), held at 4°C and used within 20 d. Based on the package
label, product compositions were as follows: turkey breast (composed of each turkey
breast, turkey broth, < 2% each of salt, dextrose, and sodium phosphates); salami
(composed of pork, beef, salt, < 2% each of dextrose, water, natural spices, sodium
ascorbate, lactic acid starter culture, garlic powder, sodium nitrite, BHA, BHT, and citric
acid); and bologna (composed of beef, pork, water, salt, and < 2% each of dextrose,
potassium lactate, sodium diacetate, sodium erythorbate, sodium nitrite, and oleoresin of

paprika).

77

Fat, moisture, and crude protein contents were determined in triplicate for two lots of
each product according to the Association of Official Analytical Chemists (AOAC)

methods 991.36, 950.46, and 992.15, respectively (AOAC, 2003).

3.3.3. Delicatessen slicer

A commercial gravity fed delicatessen slicer (Model 220F, Omcan
Manufacturing; Niagara, Falls, NY) manufactured with an electropolished 304 stainless
steel blade and other non-electropolished components was used for slicing. In order
better quantify numbers of Listeria recovered from the various slicer components, the
slicer blade was milled from a diameter of 22 cm to 15.5 cm while maintaining the
original surface profile, which had a beveled cutting edge 2.5 cm wide. The guard and

back plate were scaled down to conform to the milled blade.

3.3.4. Identification of delicatessen slicer product contact areas

A chub of turkey breast was bathed in Glow-GermTM powder (Glo-GermTM;
Moab, UT) and immediately sliced (5 slices) using the delicatessen slicer. The entire
slicer was then viewed under UV light (260 nm) to identify the most likely parts to be
contaminated. From this, the following product contact surfaces and areas for later
sampling: table (T) - 160 cm2, back plate (BP) - 192 cmz, guard (G) - 161 cm2, blade (B)

- 181 cm’, and collection area (C) - 176 cm2 (Figure 3.1).

78

Figure 3.1. Contact areas of gravity feed delicatessen slicer

 

(T) = table, (BP) = back plate, (B) = blade, (G) = guard, (C) = collection area

79

3.3.5. Surface profiling of delicatessen slicer blade

Blade roughness values and overall surface profiles were obtained at the
University of Illinois - Center for Microanalysis of Materials (Urbana, IL) using a Sloan
Dektak3 ST stylus surface profilometcr (Veeco Instruments Inc., Woodbury, NY).
Surface profilometer measurements were taken along three radial lO-mm lines marked at
approximately 120-degree intervals on the front and backside of new and used blades
after 1 and 2 years of use. Surface roughness data points were collected by recording the
height of the stylus 40 times per second while traveling along the lO-mm line.
Measurements were made along these lines with the stylus movement, ending
approximately 0.5 mm from the blade edge. The data was then short-pass filtered to
remove the effects of blade surface curvature during the milling process from the
manufacturer and provide a base line for pitting and scoring from slicing and cleaning

regimens.

3.3.6. Evaluation of slicer blade wear using SEM

A field emission scanning electron microscope (CamScan 44FE; CamScan USA
Inc., Cranberry Twp., PA) was used to visually assess new and used stainless steel slicer
blades for pitting and oxidation. Three 4 x 4 cm pieces were cut from new and 2-ycar old
slicer blades using a computerized numeric control laser cutter (ProAxis Inc.; West
Lafayette, IN). Each slicer blade piece was cleaned with 95% ethanol, placed in the SEM

chamber and scanned from end to end.

80

3.3.7. Impact of force on L. monocytogenes transfer from turkey to a delicatessen
slicer.

A replicated study (11 = 3) involving inoculated roast turkey breast (~10s
CFU/cmz) was conducted using the 6-strain cocktail. Each turkey chub (22 cm in length
x 8 cm in diameter) was surface inoculated with the 6-strain cocktail (100 pl) lengthwise
along a 1-cm wide strip and held for l h at 4°C to allow the inoculum to absorb into the
product. Forces of 0 and 10 j; 2 lbs were applied to the product against the back plate
while slicing and were continuously monitored using a ChatillionQD force gauge (Amtek,
Largo, FL) equipped with 10 x 10 cm product contact platform. After each slice, the five
previously identified contact areas were swabbed using the l-ply composite tissue (CT)
recovery method developed by Vorst et al. (2004) and added to stomacher bag-s
containing 50 ml of phosphate buffered saline (PBS). Samples were homogenized in a
Stomacher 400 (Seward; Norfolk, England) for l min and spiral plated on MOX followed
by 48 h of incubation at 35°C. For each replication, the numbers of Listeria transferred

as impacted by force were analyzed using multiple linear regression and analysis of

variance (SAS, 1996).

3.3.8. Slicer blade inoculation

A turkey slurry was prepared for inoculating the delicatessen slicer blade by
diluting 25 g of turkey breast 1:10 in sterile deionized water and homogenizing in a
model DIFP2 blender (General Electric; Bridgeport, CT) at high speed for l min.
Thereafter, the slurry was filtered through five layers of cheesecloth, heated in an 80°C

water bath for 20 min, cooled and stored in 50 ml aliquots at -20°C. For use, 50 ml of the

81

turkey slurry was thawed overnight at 4°C, poured into a sterile 15-cm diameter glass
bowl to a depth of 1.5 cm and inoculated with 0.35 i .05 ml of the 6-strain cocktail so as
to contain 109, 106, or 104 CF mm. The product-blade contact area of the alcohol-flamed
and cooled slicer blade, as previously identified by Glo-Germm (161 cmz), was
inoculated by rotating the blade through 5 revolutions in the bowl so as to contain 108,
105, or 103 CFU/blade and then dried for l h in a laminar flow cabinet. Although
unrealistically high, these inoculations level were deemed necessary to quantify Listeria

transfer during sequential slicing.

3.3.9. Transfer of L. monocytogenes from an inoculated delicatessen slicer blade to

uninoculated product

The previously inoculated slicer blade was used to obtain thirty 2 to 3 mm-thick
slices of turkey, salami or bologna weighing approximately 25 g each with each
experiment replicated 3 times. For slicer blades containing 108 CFU/blade, all 30 slices
were diluted 1:5 (w/v) in PBS, homogenized in a Stomacher for 2 minutes and spiral
plated (50 pl) on MOX. For inoculum levels of 105 and 103 CFU/blade, all slices were
diluted 1:5 in UVM and homogenized in a Stomacher for 1 minute. Duplicate 5 ml
aliquots of the homogenized sample were pour-plated in 25 ml of MOX using ISO-mm
diameter disposable Petri dishes (Fisher Scientific; Chicago, IL) and incubated at 35°C
for 48 h with populations determined as the number of listeriae per slice. When Listeria
was not detected by direct plating, MOX plates from the previously enriched samples

were examined for presence/absence of Listeria after 48 h of incubation at 35°C.

82

3.3.10. Transfer of L. monocytogenes from inoculated product via the slicer to

uninoculated product

Transfer of L. monocytogenes from inoculated turkey, salami, and bologna to the
delicatessen slicer and then to uninoculated product was replicated 3 times for each of the
three products. The turkey, salami, and bologna chubs were surface-inoculated with the
aforementioned 6-strain L. monocytogenes cocktail to obtain approximately 108 and 105
CFU/cm2 as determined from spiral plating. These inoculation levels were again
necessary to quantify numbers of Listeria in consecutive slices. After 1 h at 4°C to allow
the inoculum to absorb, three to five slices were generated from each chub to artificially
contaminate the blade with this same blade then immediately used to obtain 30 slices of
uninoculated product of the same or different type.

When product containing 108 CFU/cm2 was cut to contaminate the blade and
followed by uninoculated product, the first 20 slices were diluted 1:5 in PBS and spiral-
plated on MOX. The 10 remaining slices were diluted 1:5 in UVM, incubated 48 h at
35°C and streaked to MOX. For products containing 105 CFU/cmz, L. monocytogenes
was recovered by homogenizing a 1:5 dilution in UVM and then pour plating duplicate 5
ml aliquots in 25 ml of MOX using ISO-mm dia. Petri plates. After 48 h of incubation at
35°C, all Listeria-like colonies on the MOX plates were counted to determine the number
of listeriae per slice. When Listeria was not detected by direct plating, the MOX plates
streaked after enrichment were examined for presence/absence of Listeria after 48 h of

incubation at 35°C.

83

3.3.11. Quantification of injured Listeria on slicer blades

Five 4 x 4 cm pieces cut from the cutting edge of a new stainless steel slicer were
inoculated by spreading 100 pl of the aforementioned six-strain cocktail on the surface
and then drying in a laminar flow cabinet for 1 h. The five slicer blade pieces were
sampled using the previously described CT method with 1 m1 of PBS added to CT before
swabbing. After adding the CT to 9 ml of PBS and homogenizing in a Stomacher for 1
min, aliquots (50 pl) were spiral-plated in duplicate on tryptose phosphate agar (DIFCO)
containing ferric ammonium citrate (0.5 g/l) and esculin (1 g/l) (mTPA) for recovery of
healthy and injured cells, and on mTPA with sodium chloride (40 g/l) (NaCl) (mTPAN)
and MOX for recovery of healthy cells as previously described (Matthew and Ryser,
2002). All plates were counted after 48 h at 35° C. Percent injury was determined by the

following equation:

% injury = [(non-selective count - selective count)/ non-selective count]* 100.

3.3.12. Cleaning and decontaminating the slicer

After use and complete disassembly, the slicer table, guard, and blade were wiped
with a l-ply composite tissue and soaked for 30 min in a pan containing an activated 32%
alkaline glutaraldehyde solution (CIDEX®; Advanced Sterilization Products, Irvine, CA).
Non-removable components of the slicer (back plate and collection area) were disinfected
with the same 32% alkaline glutaraldehyde solution and allowed to air dry for 30 min.
Thereafter, a l-ply composite tissue was soaked in 70% ethanol (v/v) was used to clean

all removable and non-removable parts of the slicer after which all components were

84

rinsed with deionized water and dried. F ollow-up sampling using the CT method showed
that the slicer was free of Listeria. To prevent surface oxidation during storage, the slicer
blade was coated with a thin layer of mineral oil, which was removed by flaming with
o

95% ethanol, rinsing with sterile deionized water and drying with KimWipe

immediately before use.

3.3.13. Statistical analysis

All Listeria transfer experiments were replicated three times. Impact of cutting
force on transfer of Listeria to the five slicer contact areas and direct/sequential transfer
from the inoculated slicer blade to uninoculated product and from inoculated product to
uninoculated product via the slicer blade were analyzed using a general linear model and
analysis of variance (ANOVA) for least significant differences in mean recovery (Scott et
al., 1984). Mean differences in surface topography were analyzed using a general liner

model at each time point (n=3) (Scott et al., 1984).

3.4. RESULTS

3.4.1. Proximate analysis

Based on analyses of duplicate lots, roast turkey breast, bologna and salami
contained an average of 78, 60 and 43 % moisture, <1, 27 and 36% fat, and 19, 10 and

17% protein, respectively.

85

3.4.2. Impact of force on L. monocytogenes transfer from turkey to a delicatessen
slicer

A force of 4.5 kg applied against the product while slicing yielded significantly
greater Listeria transfer than 0 kg. (P<0.05). Less transfer was seen to the table than
other slicer contact areas at 0 kg (P<0.05) with transfer of Listeria to the back plate,
guard, and blade not significantly different (P>0.05) (Figure 3.2). However, significantly
greater (P<0.05) numbers of listeriae were recovered from the collection area using an
application force of 4.5 kg. No significant difference (P<0.05) was seen between the

table, back plate, guard and blade at a force of 4.5 kg, suggesting uniform contamination

(Figure 3.2).

86

Figure 3.2. Number of Listeria monocytogenes recovered at an application force of
0 and 10 lbs.

b,‘

DOIbs
l10|bs

Log CFU

 

 

Table Back Plate Guard Blade Collection Area

Contact Area

Means with different superscripts are significantly different for total transfer (P <0.05
Means with (*) are significantly different within each contact area between force treatments (P <0.05)

87

3.4.3. Transfer of L. monocytogenes from an inoculated delicatessen slicer blade

to uninoculated product
Listeria monocytogenes transfer from an inoculated slicer blade containing 108
CFU/blade to uninoculated roast turkey and bologna was generally logarithmic (R2

>0.92) and linear for salami (R2 = 0.93) with no significant differences (P<0.05) in

average recovery seen between the three products (Figure 3 .3).

Figure 3.3. Transfer of L. monocytogenes from inoculated slicer blade 108 (CFU/blade)
to uninoculated turkey, salami and bologna

 

 

eTurkey
8'0] ASalami
7.0 IBologna
'A
BOI 'fi“.‘a‘
I I I I
I...-A .

84o ’.°o ll ‘.II'- .d

' I 9 AA
g ...O:. .9.
43.0“ ..

2.0a

1.0a

0.0a1 . . e - a --._.~_---_--_-.___.._--.._. .__-.-_._

13 5 7 911131517192123252729
Slicenumber

88

All three products yielded direct counts for each of the 30 slices with a 2-log

reduction seen after the first 20 slices. Except for salami, similar results were obtained at

an inoculation level of 10s CFU/blade (Figure 3.4).

Figure 3.4. Transfer of L. monocytogenes from an inoculated slicer blade 105
(CFU/blade) to uninoculated turkey, salami and bologna. Open symbols

not quantifiable by direct plating.

3 5 9 Turkey
' l I Bologna
A Salami

 

Log CFU
I
I
I
I

0.5 ~

 

 

 

0.0 ... W Tees.-—
13 5 7 91113151719212325272

Slice number

89

While a linear and logarithmic decrease in numbers of Listeria transferred was not

seen for salami, enrichment results were positive out to 30 slices at an inoculation level of

lo3 CFU/blade (Figure 3.5).

Figure 3.5. Transfer of L. monocytogenes from an inoculated slicer blade 103
(CFU/blade) to uninoculated turkey, salami and bologna. Open symbols
not quantifiable by direct plating.

 

2.5 a. 0 Turkey
1 A Salami
i ‘ ‘ I Bologna
2.0 ’
‘A A‘ A‘ “
I‘A MA
3 1.5 '1 . ‘ ‘
u. . I
o * I I . A
a e
3 1.0 4. .I ‘
I c
I
0.5 i
.l
l
0.0 WI} T r . r ~ ~—

13 5 7 911131517192123252729
Slicenumber

90

 

 

Salami was significantly different (P <0.05) from turkey and bologna at 105

CFU/blade. Unlike turkey and bologna, a decrease in Listeria transfer was not evident
during slicing of salami. Listeria populations transferred to turkey and bologna were not
significantly different (P >0.05) at an inoculation level of lo3 CFU/blade. Similarly to
105 CFU/blade, salami was significantly different (P <0.05) from both turkey and
bologna. Using 105 CFU/blade, enrichments for turkey and bologna were typically
positive out to 22 slices with all 30 slices of salami positive by direct plating. At 103

CFU/blade, enrichments were typically positive out to 23 slices for turkey and salami and

20 slices for bologna (Table 3.1).

91

Table 3.1. Number of samples yielding Listeria by direct count and/or enrichment
(N=3) for delicatessen slicer-product (DS-P) and product-delicatessen
slicer-product (P-DS-P) transfer for turkey (T), bologna (B), and salami

 

 

 

(S).
10’CFUIbIade 10‘CFUIblade 10“c1=u1cntt2
(DS-P) (DS-P) (P-DS-P)
Slice 7 a s 7 a s 3.03.3 3.03.7 r-oss
1 313' 313 313 313 313 31N7b 313 313 313
2 313 313 313 313 313 3/NT 313 113 113
3 313 213 313 313 313 3INT 313 313 313
4 013 113 313 313 313 3/NT 313 113 313
5 213 213 313 313 313 31N7 313 113 313
3 113 113 313 213 313 3/NT 313 113 313
7 013 113 313 113 313 3/NT 313 112 313
3 213 213 313 113 313 3/NT 213 212 313
9 013 213 313 213 313 3INT 313 010 213
10 113 013 313 113 313 3/NT 013 011 213
11 013 012 313 013 213 3/NT 313 010 213
12 013 012 313 013 213 3INT 213 010 113
13 013 012 313 013 213 3/NT 013 010 113
14 013 012 313 013 013 3/NT 013 010 012
13 013 012 313 013 213 3INT 013 010 012
13 N713 N711 213 013 013 3/NT N713 N7 N710
17 N713 N710 313 013 013 3INT NTI3 NT NTI1
13 N712 NTI1 213 013 012 3/NT NT/3 NT N710
1s NTI2 N712 212 013 012 3/NT NTI3 NT N710
20 N713 N711 V2 011 012 3/NT NTI3 NT N710
21 N713 NT/O NTI3 N711 N712 3/NT NTI3 NT NT
22 NTI3 NT/0 um um (um 31N7 NTI3 NT NT
23 N112 N710 N712 NT/0 N712 31N7 N713 N7 NT
24 N710 NTIO N710 N711 N711 3INT NTI3 NT NT
25 N712 NT/O N710 NT/0 N711 3/NT NT/3 NT NT
23 NT/0 NTIO N710 N710 NT/1 3/NT N712 NT NT
27 N710 NTIO N710 N710 NT/1 3/NT N712 NT NT
23 N711 NT/O NTIO N710 N710 3INT N712 NT NT
29 NT/O NT/O N710 NTIO N711 3INT NT/2 NT NT
30 NT/0 NT/O N710 N710 NT/0 3/NT NTI2 NT NT

92

' Number of samples positive by direct plating / number of samples positive by enrichment,
" NT— Not Tested

 

3.4.4. Sequential transfer of L. monocytogenes from inoculated product to a
delicatessen slicer and then to uninoculated product
Numbers of Listeria transferred from surface-inoculated turkey containing 105
CFU/cm2 to uninoculated turkey during slicing were not quantifiable by direct plating.
At the higher inoculum level (108 CFU/cmz), L. monocytogenes populations decreased ~2

logs after 15 slices with enriched samples positive out to 30 slices (Figure 3.6).

Figure 3.6. Transfer of L. monocytogenes from inoculated turkey 108 (CFU/cmz) via
the slicer blade to uninoculated turkey.

6.0

00
5.0 99...
.e

 

 

0.0 7 . . . . . . ..___-,-_-.-__-_,__
13 5 7 911131517192123252729

Slice number

93

Using product inoculated to contain 105 CFU/cmz, Listeria was quantifiable when
inoculated turkey was sliced before uninoculated salami, when inoculated salami was
sliced before uninoculated turkey, and when inoculated salami was sliced before
uninoculated salami. After slicing inoculated turkey, the first 14 slices of uninoculated
salami yielded Listeria by direct plating. After slicing inoculated salami, Listeria was
also quantifiable in uninoculated turkey with a high degree of variability seen between
slices. Slicing inoculated salami followed by uninoculated turkey yielded the greatest

variability in direct counts for the first ten slices (Figure 3.7).

Figure 3.7. Transfer of L. monocytogenes from inoculated turkey (IT) and
inoculated salami (IS) 105(CFU/cm2) to uninoculated turkey (UT) and
uninoculated salami (US) during slicing. Open symbols not quantifiable

 

 

 

by direct plating.
2.5 a
9 O
2.0 a I
0 e
l ' .
a 15 . ‘ . ‘ . GIT-US
3. l ‘ e : . e c I llS-UT
° I ' Ans-us
-I 1.01| A A
A
0.5 «
0.0 . , U—B—D—-—D~--l}——-—~~—a
1 3 5 7 9 11 13 15

Slice number

94

A product inoculation level of 105 CFU/cm2 yielded positive enrichment results
out to 30, 8 and 15 slices when inoculated salami was followed by uninoculated salami or
turkey or when inoculated turkey was followed by uninoculated salami, respectively
(Table 3.1). At an inoculation level of 105 CFU/cmz, a comparison of means showed
significant differences (P <0.05) for average recovery between inoculated turkey
followed by uninoculated salami when compared to inoculated salami followed by

uninoculated turkey and inoculated salami followed by uninoculated salami.

3.4.5. Slicer blade surface profiling

A significant difference (P<0.001) in surface topography was seen for both the
front and back surfaces of the grade 304 stainless steel electropolished slicer blades over
time. Initial average roughness values for the front and back sides increased from 653 and
752 pm to 935 and 836 pm after year 1 and to 3251 and 5045 pm after year 2. Electron
micrographs taken with the SEM showed substantial wear and pitting on used as

compared to new slicer blade chips (Figure 3.8).

95

Figure 3.8. SEM micrographs of new (A) and used (B) slicer blades after 1 year of use

200 pm 200 pm

:«‘su5'
r

      

3.4.6. Quantification of injured Listeria on slicer blades

The non-selective medium (mTPA) afforded greater recovery of healthy and
injured Listeria cells from stainless steel slicer blade pieces blade compared to selective
media. Using mTPAN and MOX, 61% and 73% of the Listeria inoculum was injured,

respectively, after 1 h of drying in a laminar flow cabinet.

3.5. DISCUSSION

The three products for slicing were chosen based on differences in fat and
moisture content with turkey having the lowest fat (<1%) and highest moisture (78%) and
salami having the highest fat (36%) and lowest moisture (43%). These variations in
product composition resulted in visual differences in soiling after slicing. The higher fat

and lower moisture content of salami compared to turkey and bologna produced a

96

pronounced fat layer on the blade, which provided an excellent menstrum for Listeria
dispersion as well as protection from the normal frictional forces during slicing. Unlike
salami, the higher moisture and lower fat content of turkey had a washing affect on the
slicer blade with fewer visible meat particles remaining on the blade alter consecutive
slicing. Lin et al. (2004a) also reported that a layer of fat developed on slicer blades and
conveyor belts after slicing salami but not after slicing bologna or turkey. Salami does
not support growth of L. monocytogenes and has not been implicated as a vehicle for
listeriosis. However, the fat layer that develops on these blades after slicing salami
provides an ideal mechanism for prolonged Listeria transfer and cross-contamination of
other Listeria-free products. The amount of mechanical energy applied to a stainless
steel food contact surface during microbiological sampling also has a significant impact
on bacterial recovery. Findings by Moore and Griffith (2002) demonstrated greater
bacterial recovery from stainless steel with increased mechanical energy. Our findings
also suggest a direct relationship between mechanical energy, application force and
Listeria transfer. Significantly greater (P<0.05) transfer of L. monocytogenes was seen
from inoculated turkey breast to the slicer using a cutting force of 10 lbs as opposed to 0
lbs. Applying a 10 lb force to turkey breast during slicing led to more uniform
contamination of the slicer with the amount of exudate increasing as more force was
applied to the product during slicing.

Further differences in the numbers of Listeria transferred can be attributed to the
fat and moisture content of these products. The presence of a fat layer on a conveyor belt
during slicing of salami yielded the greatest numbers of Listeria on the blade, slicer

housing and conveyor belt. Further work done by Lin et al. (2004a) showed increased

97

Listeria transfer from contaminated slicer parts to turkey, salami, and bologna during
slicing. In their study, a commercial delicatessen slicer blade was inoculated with a five-
strain cocktail of L. monocytogenes (102 CFU) and then used to slice turkey, bologna, and
salami. Packages containing five slices of each product were vacuum-sealed and
assessed for Listeria growth after 1, 30, 60, and 90 days of storage at 4°C. While Listeria
populations increased in roast turkey breast, numbers gradually declined in salami and
bologna and fell below detectable limits after 60 and 90 days of storage. In most cases,
our results demonstrated Listeria transfer out to 30 slices. Thus, the potential exists for
growth of Listeria to potentially hazardous levels in roast turkey breast during
refiigeratcd storage as demonstrated by Lin et al. (2004b).

Previous research has helped identify surfaces such as stainless steel and
polyethylene that more conducive to bacterial transfer than rubber (Arnold and Silvers,
2000; Beresford, 2001; Midelet and Carpentier, 2002). Stainless steel is prone to
scratching, pitting and corrosion with chlorine- and acid-based sanitizers hastening this
process (Barkley, 1979; Bohner and Bradley, 1991; Stone and Zottola, 1985). When
present on slicer blades, such pits provide preferential sites for bacterial attachment and
biofihn formation (Arnold and Bailey, 2000; Chmielewski and Frank, 2003), which
impact the bacterial transfer rate during slicing. Recent studies have also demonstrated
areas of a table-top bowl chopper most susceptible to contamination during processing of
beef contaminated with Escherichia coli 0157:H7 (Flores, 2004). Areas of the bowl
chopper most likely to be contaminated were the top of the comb/knife guard and the
knife. Cleanability of stainless steel is a problem in the dairy industry. Similar studies

have evaluated cracks in stainless steel surfaces after cleaning with solvent-based

98

cleaners. Hairline cracks were found in 6 of 9 stainless steel milk holding tanks and all
13 cheese vats in one dairy processing facility (Barkley, 1979), suggesting environmental
niches for bacterial pathogens.

When SEM and atomic force microscopy (AF M) were used to assess stainless
steel surfaces of different finishes Arnold and Bailey (2000) saw relative differences in
surface morphology for different surface finishes with fewer bacterial cells attaching to
electropolished stainless steel (102 cells/SEM area) compared to 2B finished, sandblasted,
sanded, and electropolished stainless steel (103 cells/SEM area). In addition, no bacterial
cell clumps were observed on electropolished surfaces, whereas greater than 12 clumps
were seen on all other surfaces.

In our study, significant changes in slicer blade surface topography were observed
during two years of continuous use with repeated, mildly abrasive cleaning and sanitizing
producing pits or areas of high oxidation from the disintegration of the electropolished
finish. These topographical changes likely impacted Listeria transfer with the rougher
blades allowing for increased attachment. Electron micrographs of new and used slicer
blades (Figure 3.8) illustrate differences in surface finish topography as a result of use
and cleaning cycles. Numerous pits and scores on worn slicer blades can serve as
harborage sites for bacterial attachment and thus lead to more extended transfer during
slicing.

Transfer of the low-level inoculum (103 CFU/blade) was difficult to quantify with
a tailing effect observed followed by sporadic recovery after 5 slices for all three
products. Enrichment data provided some insight on likelihood of transfer after extended

slicing (>10 slices). Differences seen between initial inoculum levels and total numbers

99

of Listeria cells recovered can be accounted for in part by injury. Using selective and
non-selective media, an average of 67% of the Listeria inoculum became injured on our
stainless steel blades after 1 h of drying with this injury accounting for differences in
recovery using selective media for direct plating and enrichment. Salami continued to be
an outlier when compared to the other two products with 22% of the initial inoculum (103
CFU/blade) recovered by direct plating during sequential slicing. Listeriae not accounted
for by direct plating and injury results were likely transferred to other unsampled surfaces
of the delicatessen slicer or lost as aerosols.

Based on guidelines established in the 2001 Food Code (FDA/CFSAN, 2001),
equipment used for food preparation with food contact services must be cleaned every 24
h if held at < 5° C or every 10 h when held at 10-12.8° C with cleanliness being defined
as “clean to sight and touch”. These recommendations clearly allow ample time for
bacterial attachment, growth, and subsequent transfer to previously uncontaminated
products between cleanings. Food preparation equipment such as mechanical slicers
have numerous components including the slicer blade and guard that are difficult to clean
with soiling not always visually apparent. Findings presented in this study suggest ample
opportunity for transfer of Listeria in delicatessens via mechanical slicers with the
highest risk of consumer exposure coming from the first 10 slices. Thereafter, sporadic
transfer was seen out to 28-30 slices for all inoculation levels and transfer scenarios
(inoculated slicer to uninoculated product and inoculated product to uninoculated product
via the slicer). Based on one recent report (Draughon, 2005), delicatessen sliced

luncheon meats were more frequently contaminated with Listeria monocytogenes when

100

sliced in succession thus suggesting repeated cross contamination from delicatessen

slicer.

3.6 SUMMARY

Overall findings presented in this study suggest that the greatest risk of exposure to
Listeria during slicing of delicatessen meats occurs within the first 10 slices. Given that
an estimated 75% of all luncheon meats sold are being sliced at delicatessens, ample
opportunity exists for the contamination of delicatessen-sliced meats. Depending on
product formulation, certain delicatessen meats that permit growth of Listeria may pose a
public health risk to consumers when stored in home refrigerators for long periods of
time. These findings identified improved equipment design, stainless steel grade, and
finish as future research areas of importance for food manufacturers and retail

establishmets.

101

CHAPTER 4

TRANSFER OF LIS T ERIA MONOC Y T OGENES DURING SLICING OF

TURKEY BREAST, BOLOGNA, AND SALAMI USING KITCHEN KNIVES I

Vorst, K.L., Todd, E.C.D., Ryser, E.T.

102

4.1.

ABSTRACT

In response to continued concerns regarding Listeria cross-contamination of
ready-to-eat meat and poultry products in both retail and home kitchens, a series of
specially prepared grade 304 and 316 stainless steel knife blades were inoculated with
a 6-strain L. monocytogenes cocktail comprised of two weak, two medium, and two
strong biofilm forming strains so as to contain ~108, 105, 103 CFU/blade. Thereafter,
whole chubs of delicatessen turkey breast, bologna, and salami (3 replicates) were
sliced to entirety (30 slices) at a cutting speed of 20 mm/min using an Instron 5565
electromechanical compression analyzer. Slices were diluted 1:5, homogenized and
then surface- or pour-plated using Modified Oxford Agar and enriched in University
of Vermont Medium, Listeria transfer from knife blades inoculated at 108 CF U/blade
was logarithmic with a 2-log decrease after 12 slices and direct counts obtained
thereafter out to 30 slices. However, blades containing 105 and 103 CFU/blade
typically yielded direct counts out to only 20 and 5 slices, respectively. Normalizing
data on a log scale for the first 10 slices resulted in significantly greater Listeria
transfer and “tailing” from grade 304 as opposed to grade 316 stainless (P<0.05) for
all three products. After one year of use, knife blade roughness values as determined
by surface profilometry were significantly greater (P<0.001) for grades 304 than 316
stainless. Force and knife sharpness were not significantly different (P>0.05) within
stainless steel grade (P<0.05) for each product. However, significant differences in
force were seen between salami and turkey (P<0.05) for grades 304 and 316 stainless
steel respectively. Compositional differences of deli meats and knife blades, knife

blade wear and scoring will also likely increase Listeria transfer during slicing.

103

4.2. INTRODUCTION

Cross-contamination of cooked and ready-to-eat (RTE) foods with Listeria
monocytogenes has been identified as a serious public health concern with delicatessen
meats ranking fourth for predicted relative risk to the North American population (FDA,
2003). Four major listeriosis outbreaks have been documented in the United States since
May of 2000, three of which were traced to consumption of delicatessen-sliced turkey
breast, (MMWR, 2000; MMWR 2002). These three outbreaks were responsible for a
combined total of 91 listeriosis cases, including 11 deaths and 6 miscarriages, in 22 states
and the recall of 44.3 million pounds of product. These outbreaks prompted the
development of three USDA-mandated alternatives for controlling Listeria in
delicatessen meats — (a) post-package pasteurization, (b) product reformulation to prevent
Listeria growth and/or (c) increased product and environmental testing (FSIS, 2003).
While processing environments are still major sources of contamination, very little
research has been done with respect to Listeria cross-contamination from food contact
surfaces and utensils in delicatessens (Chen et al., 2000; Lunden et al., 2002, Uttendaele,
1999). Studies have shown a high incidence of microbial contamination associated with
retail delicatessens. In one UK survey, 10 of 32 slicer blades yielded L. monocytogenes
(Humphrey, 1990) with Uyttcndaele et a1. (2001) reporting that 4.9% of cooked meat
products sampled at retail markets tested positive for Listeria monocytogenes. This is
higher than for commercially processed meat [data such as Gombas et al., 2003] and
indicates that there is a likelihood of cross contamination in the retail operations.

Routes of cross—contamination in kitchen environments have been well

documented with studies assessing bacterial survival on common kitchen items including

104

cutting boards (Ak et al., 1994; Akier et al., 1990; Sattar et al., 2001), sponges (Michaels
et al., 2002), oven mitts (Michaels et al., 2002), pot holders (Michaels et al., 2002) and
cloth towels (Satter et al., 2001) as well as food contact surfaces comprised of stainless
steel (Arnold and Bailey, 2000; Arnold and Silvers, 2000; Herald and Zottola, 1988;
Norwood and Gilmour, 1999) or other materials (Beresford et al., 2001). Using plastic
and wood cutting boards, Ak et al. (1994) showed increased transfer of a bacterial
suspension including Listeria, Escherichia coli, and Salmonella species with scored or
scratched plastic boards compared to wooden boards. Montville et al. (2001) reported
bacterial transfer rates of 0.01 and 10% when food workers handled chicken meat with
and without vinyl gloves, respectively. Due to sampling difficulties and wide variations
in both design and use, knife blades have received inadequate attention in regards to
bacterial transfer. Listeria has been shown to attach to stainless steel in as little as 20
minutes (Mafu et al., 1991) with the extent of attachment dependent on both the grade of
stainless steel and the type of surface finish. Using scanning electron microscopy (SEM)
and atomic force microscopy, Arnold and Bailey (2000) measured biofilm formation and
surface morphology of grade 304 stainless steel of different surface finishes including
2B, sandblasted, sanded, and electropolished. When all four surfaces were exposed to a
mixed bacterial culture obtained from a poultry carcass rinse, bacterial attachment was at
least 1 log lower on electropolished stainless steel compared to the other three surfaces.
These findings have important ramifications in the manufacture of stainless steel knife
blades, delicatessen slicer blades and other food contact surfaces found on processing

equipment as well as in retail delis.

105

In limited work by King (1999) using an Instron electromechanical compression
analyzer, a light weight high speed knife sustained less damage during slicing of lamb rib
bones compared to traditional knife blades used for processing with fewer meat
particulates being generated. This work suggests that cross-contamination and
subsequent contamination of processing environment can be lessened by improving knife
blade designs with reduced aerosols and meat particulates as a result of the cutting and
sawing process.

The three primary objectives of the present study were to quantify transfer of L.
monocytogenes from (a) contaminated knife blades to turkey, salami, and bologna, (b)
contaminated roast turkey, salami and bologna to knife blades and (c) inoculated product
to a knife and then to uninoculated product. As a secondary objective, stainless steel
grade, surface roughness, knife sharpness, and cutting speed were also assessed for their

impact on Listeria transfer during slicing of deli meats. .

4.3. MATERIALS AND METHODS

4.3.1. Listeria monocytogenes strains

Six strains of Listeria monocytogenes (obtained from Dr. Catherine W. Donnelly
at the University of Vermont, Burlington, Vermont): CWD 205 (source unknown), CWD
578 (dairy plant environment), CWD 701 (Azore cheese), CWD 730 (dairy plant
environment), CWD 845 (dairy plant environment), and CWD 1002 (pork sausage) were
chosen from more than 190 strains based on their ability to form weak (CWD 205, CWD

578), medium (CWD 701, CWD 1002) or strong (CWD 730, CWD845) biofilms in a

106

microtiter plate assay (Keskinen et al., 2003). All strains were maintained at -80°C in
trypticase soy broth (TSB) (Difco/Becton Dickinson, Sparks, MD) containing 10% (v/v)
glycerol. TSB containing 0.6% yeast extract (TSB-YE) (Difco) was inoculated from the
frozen stock cultures and incubated at 37°C for 24 h. After a second transfer in TSB-YE,
each culture was pelleted by centrifugation at 9700 x g / 10 min / 4°C (Sorvall Super
T21; Sorvall Products, L.P. Newton, CT), resuspended in 9 ml of 0.1% peptone (Difco)
and combined in equal volumes to produce one 6-strain cocktail containing
approximately 108 CFU/ml. Cell concentration was verified by optical density at 600 nm
and spiral plating (Autoplate® 4000 Spiral Plater; Spiral Biotech Inc., Norwood, MA) an
appropriate dilution on Modified Oxford Agar (MOX) followed by 48 h of incubation at

35°C.

4.3.2. Delicatessen meats

One retail brand each of restructured roast turkey breast, Genoa hard salami and
bologna (5.5 to 6.5 lbs each) was purchased in chub-form from a local retailer (Gordon
Food Service, Lansing, MI), held at 4°C and used within 20 d. Based on the package
label, each product contained the following ingredients: turkey breast (turkey breast,
turkey broth, < 2% each of salt, dextrose, and sodium phosphates); salami (pork, beef,
salt, < 2% each of dextrose, water, natural spices, sodium ascorbate, lactic acid starter
culture, garlic powder, sodium nitrite, BHA, BHT, and citric acid); and bologna (beef,
pork, water, salt, and < 2% each of dextrose, potassium lactate, sodium diacetate, sodium
erythorbate, sodium nitrite, and oleoresin of paprika). Fat, moisture, and crude protein

content were determined in triplicate for two lots of each product according to the

107

Association Official of Analytical Chemists International (AOAC Int.) methods 991.36,

950.46, and 992.15, respectively (AOAC, 2003).

4.3.3. Knife blades
A series of sharp and medium sharp electropolished grade 304 and 316 stainless
steel knife blades measuring 12 cm x 5 cm (product contact area of 60 cm2 for each side

of the blade) with a thickness of 1.4mm were manufactured by ProAxis, Inc. (Lafayette,

IN) (Figure 4.1).

Figure 4.1. Instron 5565 electromechanical compression analyzer

~I

._‘
I
i
1
I
a
I
‘.
1

 

108

Each knife blade had a built-in 1 cm x 2 cm flange at each end so that the blade could
be screwed to the support bracket and then secured to an Instron electromechanical
compression analyzer. Sharp knives were machined to allow for sharp point by milling at
a 45° angle 10 mm from the end of the blade. Medium sharp blades were machined with

a blunt end (0.5 mm from tip) to simulate a broken knife blade.

4.3.4. Quantification of cutting force and speed

An Instron 5565 electromechanical compression analyzer (Instron; Canton, MA)
was used to quantify force at a cutting speed of 20 mm/min. A custom-made knife
support bracket to which all knife blades were attached was seemed to the upper load cell

(1124 lb) for complete cutting of all deli meats (Figure 4.2).

Figure 4.2. Surface scoring of used grade 316 (A) and 304 (B) electropolished
stainless steel knife blades after 6 months of use (approximately 500 slices).

 

109

4.3.5. Surface profiling of knife blades.

Knife blade roughness values and overall surface profiles were obtained from the
University of Illinois Center for Microanalysis of Materials (Urbana, IL) using a Sloan
Dektak3 ST stylus surface profilometer (Veeco Instruments Inc., Woodbury, NY).
Initially and after one year of use, surface profilometer measurements were taken along

six defined 10-mm lines on the front and backside of the blade (Figure 4.3).

Figure 4.3. Stylus locations on blade for surface profile measurements

 

Scan Lines

«W*‘Mi .a_. L
.‘ ___ I .n

.‘

‘-.__fi 4 ._n a ‘
.‘.__.~___an LW-L_L-W._ -_
‘..___I-__._._n-_,-_ _

Surface roughness values were obtained by recording the stylus height 40 times per
second as the stylus traveled towards the edge of the blade along each lO-mm line at
predetermined intervals. Measurements were stopped when the stylus came within 0.5

mm of the blade edge.

110

4.3.6. Knife blade inoculation

A turkey slurry was prepared for inoculating the knife blade by diluting 25 g of
turkey breast 1:10 in sterile deionized water followed by homogenization in a model
DIF P2 blender (General Electric; Bridgeport, CT) at high speed for 1 min. The slurry
was then filtered through five layers of cheesecloth, heated in an 80°C water bath for 20
min, and stored in 50 ml aliquots at -20°C. For use, 50 ml of the turkey slurry was
thawed overnight at 4°C after which 1 ml of the 6-strain cocktail was inoculated into 9 ml
of turkey slurry. One face of the 75% (v/v) ethanol flame sterililzed knife blade. was then
inoculated with 100 pl of this turkey slurry. Alter uniformly spreading the inoculum on
the 60 cm2 surface of the blade with a inoculating needle so as to contain 108, 105, and
103 CFU/blade, the blades were dried for 1 h in a laminar flow cabinet at 23°C and 30 ‘-
40% relative humidity recorded with a hygrometer (Fisher Scientific; Hampton, NH) and

then immediately used to obtain 30 slices of turkey, salami or bologna.

4.3.7. Transfer of L. monocytogenes from inoculated grade 304 stainless steelknife

blades to uninoculated product

In a replicated study (n=3), 30 slices each of previously uninoculated turkey
breast, salami, and bologna were obtained using knife blades inoculated to contain 108,
105, and 103 L. monocytogenes CFU/blade. These unrealistically high inoculation levels
were necessary to quantify numbers of Listeria in consecutive slices for subsequent
modeling of Listeria transfer. For knife blades containing 108 CFU/blade, all slices were
diluted 1:5 (w/v) in PBS, homogenized in a Stomacher 4000 (Seward; Norfolk, England)

for 1 min and spiral-plated (50pl) on MOX. For blade inoculum levels of 105 and 103

111

CF U/blade, all slices were diluted 1:5 (w/v) in University of Vermont Medium (UVM)
(Difco-Becton Dickenson; Detroit, MI) and homogenized in a Stomacher for 1 min.
Duplicate 5 ml aliquots of the homogenized sample were pour-plated in 25 ml of MOX
using ISO-mm diameter disposable Petri dishes (Fisher Scientific; Chicago, IL) and
incubated at 35°C for 48 h with populations determined as the number of listeriae per
slice. When Listeria was not detected by direct plating, MOX plates streaked from the
previously enriched sample at 30°C were examined for presence/absences of Listeria

afier 48 h at 35°C.

4.3.8. Transfer of L. monocytogenes from inoculated grade 304 and 316 stainless
steel knife blades to uninoculated product.

Listeria transfer studies from inoculated grade 304 and 316 stainless steel knife
blades to turkey, salami and bologna were replicated three times for each of the three
products. Sterile grade 304 and 316 stainless steel knife blades were inoculated to
contain 103 CFU/blade as previously described and used to obtain 20 slices of each
product. All slices were diluted 1:5 (w/v) in UVM and homogenized in a Stomacher for
1 min. Duplicate 5 ml aliquots of the homogenized sample were pour-plated in 25 ml of
MOX using ISO-mm diameter disposable Petri dishes (Fisher Scientific; Chicago, IL) and
incubated at 35°C for 48 h with populations determined as the number of listeriae per
slice. When Listeria was not detected by direct plating, MOX plates streaked from the

previously enriched samples were examined for presence/absences of Listeria after 48 h

at 35°C.

112

4.3.9. Transfer of L. monocytogenes from inoculated product to grade 304 stainless

steel knife blades and then to uninoculated product

Replicated studies (n=3) assessing transfer of L. monocytogenes from inoculated
turkey, salami, and bologna to knife blades and then to uninoculated product were
conducted using each of the three products. The turkey, salami, and bologna chubs were
surface-inoculated with the aforementioned 6-strain cocktail along a l x 1 cm strip to
obtain approximately 105 CFU/cm2 as determined from subsequent spiral plating of
diluted samples on MOX. This high inoculation level was again necessary for
quantification of Listeria in consecutive slices. After a l-h hold at 4°C, each product was
sliced 3-5 times to contaminate the grade 304 stainless steel knife blade with this same
blade then immediately used to obtain 30 slices of uninoculated product of the same or
different type.

Listeria was recovered by homogenizing a 1:5 dilution in UVM followed by pour
plating duplicate 5 ml aliquots in 25 ml of MOX using ISO-mm dia. Petri plates. After
48 h of incubation at 35°C, all MOX plates were counted to determine numbers of
listeriae per slice. When Listeria was not detected by direct plating, MOX plates streaked

after enrichment were examined for presence/absence of Listeria after 48 h at 35°C.

4.3.10. Cleaning/decontaminating knife blades

The knife blades were removed from the support bracket after use, soaked in 75%
ethanol (v/v), wiped with a l-ply composite tissue (CT) and rinsed with deionized water.
Adequacy of this cleaning/sanitizing regimen was confirmed using the CT developed by

Vorst et al. (2004). Before use, the knife blades were rinsed with sterile deionized water

113

and dried with a l-ply composite tissue. To prevent surface oxidation during storage, the
knife blades were coated with a thin layer of mineral oil, which was removed before use

by flaming with 95% ethanol and a final rinse in sterile deionized water.

4.3.11. Quantification of injured Listeria on knife blades

Five unused grade 304 and 316 knife blades were inoculated at 105 CFU/blade as
previously described. All knife blades were sampled using the CT method of Vorst et al.
(2004) with 1 ml of PBS added to CT before swabbing. After adding the CT to 9 ml of
PBS and homogenizing in a Stomacher for 1 min, aliquots (50 pl) were spiral-plated in
duplicate on tryptose phosphate agar (DIFCO) containing ferric ammonium citrate (0.5
g/l) and esculin (1 g/l) (mTPA) for recovery of healthy and injured cells, and on mTPA
with sodium chloride (40 g/l) (mTPAN) and MOX for recovery of healthy cells as
previously described (Mathew and Ryser, 2002). All plates were counted after 48 h at

35° C. Percent injury was determined by the following equation:

% injury = [(non-selective count - selective count)/ non-selective count]* 100

4.3.12. Statistical Analysis

All Listeria transfer experiments were replicated three times. Listeria transfer
to/from knife blades and direct/sequential transfer from inoculated blades to product and
inoculated product to uninoculated product via the knife blade were analyzed using a
general linear model and analysis of variance (ANOVA) for least significant differences

in mean recovery (SAS, 1996). Mean differences in surface topography for 304 and 316

114

grade stainless were replicated five times and analyzed using a general liner model at
each of 5000 points for 6 defined areas across the front and back of the knife blade

(Figure 4.3) (SAS, 1996).

4.4. RESULTS

4.4.1. Knife blade surface profiling

Initial roughness values for new knife blades prepared from grades 304 and 316
stainless steel were 105 and 70 pm, respectively. After one year of use, a significant
difference (P<0.0001) in surface topography was observed for both the front and back of
grade 316 blades with average roughness values of 2083 and 3079 pm, respectively,
whereas no such difference (P>0.05) was seen for grade 304 blades. The total mean
roughness values of 7409 and 2581 pm (front and back) for 304 and 316 blades

respectively were significantly different (P<0.001).

4.4.2. Proximate Analysis
Based on proximate analyses, the turkey breast, salami, and bologna sliced in this
study contained 78, 43 and 60% moisture, <1, 36, and 27% fat, and 19, 17, and 10

protein, respectively.
4.4.3. Effect of stainless steel grade, product, and sharpness on transfer

The average force needed to cut salami was the highest for grades 304 and 316

stainless steel blades at 50 :1: 7 and 48 i 5 lbs., respectively for sharp blades. Significant

115

differences in cutting force were seen between salami, bologna and turkey (P<0.05) for

grades 304 and 316 stainless steel (Table 4.1).

Table 4.1. Average slicing force (lbs) for turkey, salami, and bologna using medium
sharp (MS) and sharp (S) knife blades manufactured from 304 and 316 grade

 

 

 

stainless steel
Grade 304 Grade 316
Product MS S MS S
Turkey 3434: 213:31, 30423a 22a3b
Bologna l 1i3° 83:1c 8&1c 8:1c
Salami 57a5d 50:7d 59sec“ 483:5“

 

Means with different superscripts significantly different (P<0.05)

Average cutting force for turkey and salami was 22 :L- 3 and 48 :1: 5 lbs for grade
316 and 21 :1: 3 and 50 :1: 7 lbs. for grade 304 stainless steel blades, respectively. Bologna
had the lowest average cutting force at 8 :1: 1 lbs. Preliminary data showed no significant
differences (P>0.05) in transfer using inoculated sharp (S) and medium sharp (MS) 304
grade stainless knife blades (108 CFU/blade) with transfer at 5.02 and 5.07 CFU/slice for

MS and S respectively.

4.4.4. Transfer of L. monocytogenes from inoculated grade 304 and 316 stainless
steel knife blades to uninoculated product
Transfer of Listeria from inoculated knife blades (108 CFU/blade) was generally
logarithmic for all three products with populations decreasing 2 logs on the blade after

the first 8-12 slices (Figure 4.4).

116

Figure 4.4. Transfer of L. monocytogenes from an inoculated knife blade
(108 CFU/blade) to uninoculated turkey, salami and bologna

 

8.0 a! I Bologna
7.0 i A Salami
50,!00 . eTme
50 Ix..==' I
. A I

g T ‘ ‘ :AI‘.

u401 0:. A I e. I

o 8‘ IIAII .

O O. A
2.0 . ‘ J
o

1.0 1 ‘ . O
0.0 l . . 1 - e . . ———

 

13 5 7 911131517192123252729
Slice number

The total number of Listeria CFUs transferred was not significantly different
(P<0.05) between products. At an inoculation level of 10s CFU/blade, Listeria was

quantified in slices 13 to 20 by direct plating for all three products (Figure 4.5).

117

Figure 4.5. Transfer of L. monocytogenes from an inoculated knife blade
(105 CFU/blade) to uninoculated turkey, salami and bologna. Open
symbols not quantifiable by direct plating.

I Bologna
. I g ' I A Salami
0 Turkey

Log CFU

N
01

e

>

e

I
p

 

 

 

0,0 1 r f 7 Tfl—‘l—O—OHr r T i 'T‘
13 5 7 91113151719212325272

Sllce number

Enrichment results were typically negative for turkey and bologna after 26 slices and

positive for salami out to 30 slices (Table 4.2).

118

Table 4.2. Number of direct counts and positive enrichments for blade
-product (B-P) and product-blade-product (P-B-P) transfer for turkey (T),

 

 

 

 

bologna (B), and salami (S)
10°CFUIBlade 10‘CFUIBlade 1o"c1=u1cnn2
(3") (BP) (PBP)
Slice 7 B s 7 e s 737“ 337° 73s“
1 313' 313 213 313 313 313 313 313 213
2 313 313 313 313 313 313 313 313 213
3 313 3/3 313 313 313 313 313 313 213
4 313 2/3 313 3/3 313 313 313 313 113
5 313 113 313 313 313 313 313 213 113
6 213 1/3 0/3 3/3 3/3 3I3 3I3 .1/3 1/3
7 113 1/3 213 313 313 313 313 213 113
a 013 1/3 213 3/3 3/3 3/3 3/3 313 113
9 013 2/3 113 3/3 313 313 313 313 113
10 013 213 313 313 313 313 313 113 313
1 1 011 113 313 313 313 313 313 013 113
12 012 012 213 313 213 313 313 013 012
1 3 011 113 011 313 213 313 313 013 112
14 011 013 111 213 013 313 313 013 011
15 011 013 012 213 113 313 313 113 212
13 011 012 012 112 113 313 313 113 013
17 011 012 011 013 113 313 313 013 011
13 010 012 010 112 013 313 313 111 011
19 010 013 010 213 113 313 213 011 012
20 010 011 010 113 013 313 313 013 012
21 N7‘ N7 NT 112 012 N713 313 012 011
22 NT NT N7 112 011 N713 213 012 011
23 NT NT NT 012 011 NTI3 213 011 010
24 N7 NT NT 012 012 NTI3 313 012 012
25 NT N7 NT N712 NTI1 NTI3 313 011 011
23 NT NT NT N712 N711 NTI3 313 011 010
27 N7 NT NT NT/O N710 N713 111 010 012
23 N7 N7 N7 NT/0 N710 NTI3 113 011 012
29 N7 N7 N7 N710 N710 NTI3 111 010 010
30 NT N7 NT NT/O NT/1 N713 111 - 010 010

" Direct counts separated by slash following enrichment results for 3
replicates

b T-B-T — inoculated turkey to blade to turkey

° S-B-T - inoculated salami to blade to turkey

d T-B-S — inoculated turkey to blade to salami

‘ NT-Not Tested

119

Low-level inoculation (103 CF U/blade) identified a weak logarithmic association for

bologna and no association for turkey and salami (Figure 4.6).

Figure 4.6. Transfer of L. monocytogenes from an inoculated knife blade
(103 CFU/blade) to uninoculated turkey, salami and bologna. Open
symbols not quantifiable by direct plating.

 

 

 

2.5 ~ I Bologna
A A Salami
20)....‘ eTurkey
' e
I
‘ I
3 1 5 T A I ' . l A A
as ‘ ‘ I ‘
ET 1 0 . I I
‘ A
I
0.5 1 I
0.0 r WOO—o:— , . . --__-
0 5 10 15 20 25 30

Sllce number

120

Differences in transfer between stainless steel grade (304 vs. 316) were compared
for each turkey, bologna, and salami product. Using 103 CFU/blade, stainless steel type
did not significantly impact (P>0.05) numbers of listeriae transferred during the first 20
slices. Listeria was quantifiable in the first 5 slices by direct plating, regardless of
stainless steel grade or product type. Direct counts were obtained out to 5 slices for all
products and both grades of stainless. Normalizing data on. a log scale for the first 10
slices resulted in significantly greater (P<0.05) transfer for grade 316 as opposed to grade
304 stainless steel. For slices 10 through 15, more direct counts were obtained from

grade 304 than grade 316 stainless steel (Figures 4.7-4.9).

121

Figure 4.7. Transfer of L. monocytogenes from inoculated knife blades (grade 304,
grade 316, 103 CFU/blade) to uninoculated turkey

ClT316
IT304

Log CFU

 

 

 

 

 

T" , 7.,, v

12 34 56789101112131415
Slice

Figure 4.8. Transfer of L. monocytogenes from inoculated knife blades (grade 304,
grade 316, 103 CFU/blade) to uninoculated salami

2.5 . y as 313

IS 304
2.0

 

 

12 34 5 67 8 9101112131415
Slice

122

Figure 4. 9. Transfer of L. monocytogenes from inoculated knife blades (grade 304,
grade 316, 103 CFU/blade) to uninoculated bologna

3.0 - :1 B 313
2.5 -‘ I B 304
2.0 1

Log CFU

 

ii 11111111 L

2 34567 89101112131415
Slice

For slices 10-15, total numbers of Listeria transferred were greater using grade
304 as opposed to grade 316 stainless steel blades, resulting in significant differences

(P<0.05) for mean recovery as a function of total CFUs transferred.

4.4.5. Sequential transfer of L. monocytogenes from inoculated product to a knife
blade and then to uninoculated product using grade 304 stainless steel knife
blades
Numbers of Listeria transferred from surface-inoculated turkey and to

uninoculated salami (10s CFU/cmz), inoculated salami to uninoculated turkey (105

CFU/cmz) and inoculated turkey (105 CFU/cmz) to uninoculated turkey were quantifiable

out to ll'h,l6"‘, and 30th slices using direct plating (Figure 4.8).

123

Figure 4.10. Transfer of L. monocytogeness from inoculated turkey (IT) (105 CF U/cm )
and inoculated salami (IS) (105 CFU/cm2 ) to uninoculated turkey (UT) and
uninoculated salami (US) during slicing. Open symbols not quantifiable by

 

direct plating. .
3-0 _ . I lS-UT
' A rr-us
2.5
l . e lT-UT
2.0 . ' ' e . ' e
3
LL 0 . O O .
:151'03'0‘ AI . 0' 9.. .e
3 ‘ 9 A e I
J ‘ I ' O o
1.0 ‘ A . ‘ .
A ' e
I
0.5 i ‘ A 4 I
l
0.0 i . r r .W s . e—T—m.

 

 

0 2 4 6 81012141618202224262830
Slicenumber

Mean recovery was significantly greater (P<0.05) for inoculated turkey sliced
before uninoculated turkey when compared to inoculated salami sliced before
uninoculated turkey and inoculated turkey sliced before uninoculated salami. A 2-log
reduction was seen within the first 15 slices for all transfer scenarios with transfer out to
more slices with inoculated turkey sliced before uninoculated turkey. Enrichment results
were typically negative after 28 slices for inoculated turkey sliced before uninoculated
salami and inoculated salami sliced before uninoculated turkey. Inoculated turkey sliced

before uninoculated turkey resulted in positive enrichments to 30 slices.

124

4.4.6. Quantification of injured Listeria on 304 and 316 knife blades

The non-selective medium (mTPA) afforded greater recovery of healthy and
injured Listeria cells from stainless steel blades compared to selective media. When
mTPAN and MOX was used for recovery of L. monocytogenes after 1 h of drying in a
laminar flow cabinet on grade 304 and 316 stainless steel, 46% and 72% of the Listeria

CFU were injured, respectively,.

4.5. DISCUSSION

Products chosen for this study were based on fat and moisture content with turkey
having the lowest fat (<1%) and highest moisture (78%). Salami had the highest fat
(36%) and lowest moisture (43%) content. The knife blades used in this study were
manufactured to specific specifications to minimize variability seen in commercially
available knife blades. Use of the Instron eliminated operator variations in cutting speed,
force, and cutting action (sawing versus chopping). Cutting force and knife sharpness
were not significantly different within stainless steel grades and product (P<0.05). Force
required to slice the product was significantly higher (P<0.05) for salami when compared
to turkey and bologna.

Compositional differences in products were not observed to have an impact on
transfer with the exception of inoculated turkey subsequently sliced with uninoculated
turkey. While Lin et al. (2004a) reported the development of a visible fat layer during
slicing of salami with a delicatessen slicer, this same fat layer was not pronounced after
slicing the same product with a knife blade. Hence, the chopping action did not support

formation of a fat layer when compared to the hi gh-speed rotation and centrifugal force

125

of a delicatessen slicer. In our study, Listeria transfer was similar when turkey, salami,
and bologna were sliced with an inoculated knife blade. However, when inoculated
turkey was sliced and followed with uninoculated turkey, greater transfer was seen to
more slices compared to when inoculated salami or turkey was followed by uninoculated
salami.

Stainless steels of different grades are used in all segments of the food industry
due to their superior mechanical properties (good ductility, toughness, strength, and
workability) and corrosion resistance. In food processing environments, AISI (American
Iron and Steel Institute) grades 304, 304L, 316, and 316L stainless steel are the most
common alloys (Smith, 1993; Stone and Zottola, 1985; Suzuki, 2000) with the 400 series
also commonly used for knives. Stainless steel knives are universally used in food 1
preparation with a seemingly endless variety of industrial and kitchen knives being
marketed. In addition to obvious differences based on intended use, knife blades also
differ in surface finish/polish type, stainless steel grade (e.g., carbon, nickel and
molybdenum content), cutting edge styles (serrated and straight) and overall surface area.
Knife blade sharpness was not shown to have an impact on Listeria transfer but did result
in a lower cutting force for sharp blades. Although likely having a limited impact on
transfer, the variability between within products sliced was too high to identify any
discemable differences. Furthermore, large changes in surface topography most likely
masked any subtle difference seen in knife force or sharpness. Current methods for
quantifying Listeria on heavily soiled food contact surfaces are too imprecise to allow the
use of more realistic contamination levels (101 - 102 CFU/cmz) (Gombas, 2003; Moore

and Griffith, 2002; Satter, 2001). Once improved recovery methods have been

126

developed, further research needs to be conducted at very low-level inoculation to fully
understand the impact of knife sharpness on bacterial transfer during slicing of RTE
foods.

Stainless steel grade did not significantly impact (P>0.05) the total number of
Listeria cells transferred during slicing. However, differences in stainless steel grade
were significantly different by slice (P<0.05) for the first ten slices. Most notably, a far
more pronounced tailing affect was seen when grade 304 rather than 316 knife blades
were used for slicing. In contrast, greater transfer of Listeria to fewer slices was evident
using grade 316 stainless steel knife blades which may be related to the smoother finish,
greater durability and easier Cleanability of grade 316 as opposed to 304 stainless steel
(Arnold and Bailey 2000, Leclercq-Pelat and Lalande, 1994). Our surface topography
results support these findings with overall surface roughness values being significantly
lower and less variable for 316 when compared to 304 electropolished stainless steel.
Surface scoring was also less pronounced on grade 316 stainless steel blades with
obvious score marks on the same areas of 304 grade steel blades after 6 months of
repeated use and cleaning. According to Percival, the molybdenum concentration used in
grade 316 stainless steel may be factor in terms of decreasing viability of bacteria and
reducing biofilm formation (Percival, 1999). While our findings cannot confirm or deny
these biocidal claims of molybdenum, use of grade 316 stainless steel resulted in greater
injury of Listeria (72%) compared to grade 304 grade stainless (46%).

Based on guidelines established in the 2001 Food Code (FDA/CFSAN, 2001), all
food contact surfaces on equipment must be cleaned every 24 h if held at < 5° C or every

10 h when held at 10-12.8° C with cleanliness being defined as “clean to sight and

127

touch”. These recommendations clearly allow sufficient time for bacterial attachment,
growth, and subsequent transfer to previously uncontaminated products between
cleanings. Food preparation equipment such as knives will score over time as

demonstrated in this study and have the potential to harbor pathogens even after cleaning.

4.6. SUMMARY

Based on our findings, ample opportunity exists for transfer of Listeria using
kitchen knives in both commercial and home settings with the highest risk of consumer
exposure coming from the first 5-15 slices, depending on the grade of stainless steel used.
While the numbers of listeriae transferred in such settings would be admittedly very low,
even these few cells may pose a public health risk to consumers if the product
formulation permits growth of Listeria in home refrigerators during extended storage.
Further recommendations to equipment manufacturers and food processors would include
use of electropolished knife blades and grade 316 stainless steel for reduced surface

scoring and transfer of bacteria to fewer slices.

128

CHAPTER 5

TRANSFER COEFFICIENTS AND PREDICTIVE MODELS FOR LIS T ERIA
MONOC Y TOGENES DURING SLICING OF READY-TO-EAT, TURKEY

BOLOGNA, AND SALAMI

Vorst, K.L., Burgess, G.B., Todd, E.C.D., Ryser, E.T.

129

5.1. ABSTRACT

Previous data was used to develop a series of Listeria transfer coefficients during
slicing of deli meats with a mechanical delicatessen slicer and kitchen knives prepared
from grades 304 and 316 stainless steel. Transfer coefficients were calculated for two
different Listeria transfer scenarios — (a) inoculated blade to uninoculated product and (b)
inoculated product to uninoculated product via an uninoculated blade. A mathematical
model was then developed to predict the least favorable conditions for Listeria transfer
during slicing of ready-to-eat turkey breast, salami, and bologna. The model and
subsequent program is based on the following three assumptions: 1) the expected number
of Listeria cells transferred during slicing is the fraction “f1” that describes the number of
Listeria cells on the blade just before each sequential slice, 2) the expected number of
Listeria cells transferred to the surrounding areas is the fraction “f2”, and 3) the number
of Listeria cells on the blade available for transfer before any slicing begins is No. Based
on these assumptions, a predictive model was developed that may prove to be beneficial
in refining the current risk assessments for RTE meat and poultry products purchased at

retail delicatessens.

130

5.2. INTRODUCTION

Since May of 2000, four major outbreaks of foodborne listeriosis have been
documented in the United States; three were linked to consumption of delicatessen-sliced
turkey breast. These later three outbreaks were responsible for 91 cases of listeriosis,
including 11 fatalities and 6 miscaniages, and resulted in the recall of at least 44.3
million pounds of product. In all three outbreaks, post-processing contamination of the
product was to blame.

Listeria monocytogenes can reside in food processing facilities for many years
(Tompkin, 2001) those strains that are most persistent in factory environments possessing
greater capability to adhere to food contact surfaces (Lundun et al., 2000, 2002; Norwood
and Gilmour, 1999). Attachment of Listeria monocytogenes to stainless steel surfaces in
as little as 20 min has been reported after which this pathogen can be transferred to
previously uncontaminated food and food contact surfaces during food preparation and
subsequent handling (Mafu et al., 1990). In one study, Lunden et al. (2002) demonstrated
plant-to-plant transfer of L. monocytogenes via a dicing machine with the same strain of
Listeria identified at three different facilities. At the retail level, Hudson and Mott (1993)
collected various environmental swab samples from a supermarket delicatessen and
isolated L. monocytogenes from a knife and slicing machine with the pathogen also found
at most sites near a display case of processed meats.

In response to the aforementioned outbreaks, quantitative transfer to and from
commercial slicing machines, knives, and cutting boards in delicatessens was identified
as both a major public health concern and a key informational gap in the 2003 FDA

Listeria Risk Assessment (FDA/FSIS/CDC, 2003). In one study of 32 retail meat slicers,

131

10 machines were positive for L. monocytogenes (Humphrey, 1990). This information,
combined with a recent report by Gombas et al. (2003) indicating L. monocytogenes
populations as high as 104 CFU/g on delicatessen luncheon meats, suggests ample
opportunity for transfer of L. monocytogenes via slicing machines to previous
uncontaminated delicatessen meats sold at retail with such contamination putting
susceptible individuals at greatest risk of infection.

Many growth models, beginning with the US Department of Agriculture (USDA)
Pathogen Modeling Program (Buchanan and Phillips, 1990) have been developed to
predict the growth of Listeria and other foodborne pathogens in foods based on pH,
storage temperature, and levels of salt and sodium nitrite (Houtsma et al., 1996; 1986; Le
Marc et al., 2002; and Tamplin, 2002). The most recent USDA model developed by
Tamplin (2002) comes with a pre-programmed graphical user interface and generates
graphs and tabular output for various growth parameters. In most of these models,
empirical data is fitted to mathematical equations using parameter estimation techniques.
The method proposed herein uses a combination of emipirical data fitting techniques and
mathematical manipulations as described by Bemaerts et a1, (2004). Techniques
proposed in this study build on parameters described by empirical data while applying
mathematical manipulation to each of the parameters as a firnction of total transfer.

The objective of this study was to develop a predictive model describing the
relationship between numbers of Listeria cells transferred based on previous data
obtained from the slicing of turkey, salami, and bologna with inoculated delicatessen
slicer and kitchen knife blades. a) the expected number of Listeria cells transferred

during slicing is the fraction “f1” that describes the number of Listeria cells on the blade

132

just before each sequential slice, b) the expected number of Listeria cells transferred to
the surrounding areas is the fraction “f2” that describes how many cells are on the blade
just before slicing, and c) the number of Listeria cells on the blade available for transfer
before any slicing begins is No. Based on these assumptions, a model was generated that
describes transfer of L. monocytogenes during slicing of RTE deli meats. The transfer
coefficients and mathematical model developed can be used to refine the current risk

assessments for RTE meat and poultry products.

5.3. MATERIALS AND METHODS

5.3.1. Transfer coefficients for Listeria monocyctogenes during slicing of turkey,
bologna and salami.

Previous data obtained from the slicing of roast turkey breast, salami and bologna
with a mechanical delicatessen slicer (V orst et al., 20053) and kitchen knives (Vorst et
al., 2005b) prepared from grades 304 and 316 stainless steel. Listeria transfer
coefficients were developed for two different scenarios — (a) inoculated blade to
uninoculated product and (b) inoculated product to uninoculated product via an
uninoculated blade. The following equation was used for calculating transfer coefficients
where: “N,” is the original inoculum on the blade, “ft” is the fraction of bacteria left on
the blade just before each sequential slice and “f2” is the expected number of bacteria
transferred to surrounding areas. The equation for the transfer coefficient presented as a

cumulative percentage of total transfer is as follows:

133

(~01
%transfer= 100*L13-f—2— (1a)

0

This can be arranged as

100*f,

% transfer = ——
ft + f2

(1b)

The cumulative transfer coefficient was found by summing sequential transfer

coefficients for each slice.

5.3.2. Predictive modeling of L. monocytogenes transfer during slicing of turkey
breast, bologna, and salami.

A model based on the following three assumptions was developed to predict the
previously calculated transfer coefficients: a) the expected number of colony forming
units (CFU) transferred during slicing is the fraction “ft” that describes what is left on the
blade just before each sequential slice b) the expected amount of transfer to the,
surrounding areas is a different fraction “f2” that describes what is left on the blade just
before slicing, and c) the number of CFU’s on the blade available for transfer before any
slicing begins is No.

The following consequences of these assumptions are as follows:

1St Slice
CFU on Meat = le0 (2a)
CFU to Surroundings = sz0 (2b)

134

CFU left on Blade = No —f,N0 —f2No

= (1 -fr -f.)No (2c)

2"d Slice
CFU on Meat = f,(1—f1 — f,)1vO (3a)
CFU to Surroundings = f,(l — f, - f,)1v, (3b)

CFU left on Blade= (l—fl —f2)N0 -f,(l-fl -—f2)N0 -f2(1—fl -f2)N0

= (l-fi-fszo (30)
3rd Slice
CFU on Meat = f,(1— fl —f2)2No (4a)
CFU to Surroundings = f,(1—fl — f2)"'N0 (4b) 1

CFU left on Blade= (l—fl -f,)21110 -f,(1—fl —f,)21v0 -f,(1-fl —f,)2No

=(1-f.-f.)’N. (4c)

xth Slice
CFU on Meat = f,(1 - f, -f,)’("1v0 (5a)
CFU to Surroundings = f,(1— f, - f,)""1v0 (5b)

CFUlefion Blade=(l—fl -f2)X-'N0 ‘ftU-ft "f2)X-INo -f2(1-fr ‘f2)X—|No

=(1_fl-f2)XNo (50)

135

5.3.3. Predicting CFU’s on meat as a function of slice number (X).

The model predicts that the number of CFUs transferred to slice X is:

CFU (X) = 1.0—f. —f.)"" N. 16a)

This can be arranged as:

crum=fi§¥f7 (1—11—121‘ (6b)

This can be rewritten as:

CFU (X) = kax (6c)
Where “k” and “a” are constants related to the model parameters.
5.3.4. Fitting the equation to data (finding “k” “a”). Taking the natural log of the

predictive equation gives the general equation for a straight line.

y=b+mx (7a)

136

This equation can then be fitted to the data to slope (m) and intercept (b)

Where:
y = ln(CF U )
b = ln(k)
m = ln(a)
It then follows that:
a:l—f;—f'2=em (7b)
k=ero(1-Ji-fs)=e" 17c)

5.3.5. Interpretation of fit results.
The previously identified equation has three unknowns that can be determined if
the original inoculum level is known and both the slope and y-intercept are derived. The

parameter “a” is the fraction of CF Us remaining on the blade after any slice. The slope

“m” from the fit will always be negative, so (1 - fl — f2) <1. The number of CFUS
transferred from the blade to the first slice is leo. From the relationships between the
fit parameters “m” and “b” and the model parameters “ fl ”, “ f2 ” and “ N0 ” it follows

that:
Fraction remaining on blade = e’"

CFUS transferred to lSt slice = e”’*”

137

Given the inoculation level or original number of CFU’S on the blade, “ fl ” “ f2 ” and

9

“ N0 ” can be found as follows using these previous equations:

 

ft = (83)

5.4. RESULTS

5.4.1. Transfer coefficients for Listeria monocyctogenes during slicing of turkey

breast, bologna and salami.

Listeria transfer coefficients for each mechanical slicer and knife blade scenario
are presented in Figures 5.1-5.8. Except for salami sliced on a mechanical slicer at 105
and 103 CFU blade, 3 similar trend was seen for all other mechanical slicer and knife
blade scenarios with 99% of the original Listeria population being transferred in the first
10-15 slices. When salami was sliced using a mechanical slicer blade containing 105
CFU/blade, Listeria continually transferred out to 30 slices without a plateau and out to
20 slices before reaching a plateau or becoming stationary for a slicer blade inoculated at

103 CFU/blade.

138

Figure 5.1. Cumulative L. monocytogenes transfer (%) from an inoculated slicer blade
(108 CFU/blade) to turkey, salami and bologna.

5 1 oTurkey
4-57 Dugguuuuuuuunuuuuuu Usa'am'
DE]

. ' A Bologna
4 l D U

383333333333330333330
32228

0
1.51 A
1:]

Cumulative Transfer (%)
N
0|

 

0 5 10 15 20 25 30
Slice number

139

Figure 5.2. Cumulative L. monocytogenes transfer (%) from an inoculated slicer blade
(105 CFU/blade) to turkey, salami and bologna.

 

5 ., oTurkey
l [313 DSalami
% D
A 5 . 0 Bot n
$ 1 DD A m
f i DD
34 D
2 1
E l ooooooooanooooooooooo
g 31' 0000 Dan
fr 1 DD
321 0 DUDE
E = 1:1
8 I 0 DD
1' a
<>1:1
D
0 EAAAAAAAAAAAAAAAAAAAAAAAAAAAan
0 5 10 15 20 25 30

Sllce number

140

Figure 5.3. Cumulative L. monocytogenes transfer (%) from an inoculated slicer blade
(103 CFU/blade) to turkey, salami and bologna.

 

25f
( 00000000000
1 0
. 0
$5201 DUB oTurkey
v I
5 I DO DSalami
a ' D
:151 0 ABologn
E l '3
3 l D
3101 a”
:1
C]
S 3 an
0 5n D
:1 099308666666666666666666666
0,5183 , Y ____F___._-
0 5 10 15 20 25 30

Slice number

141

Figure 5.4. Cumulative L. monocytogenes transfer (%) from inoculated turkey (IT) and
salami (1S)(10S CFU/cmz) to uninoculated turkey (UT) and salami (US)
during mechanical slicing.

 

 

0.7.
“0.61 DDDDUDDDDD
ii 1 DD OIS-UT
basil D
e-; BIT-US
2 0”
50-41 CID AIS-UT
1- .
£0.33 El AAAAAAAAAAAAAA
g. 121 AA
30.23 A
3
U [j A 00000000000000
()4. 0°,
0
03 ,

O
o -12.

1

l

5 10 15 20
Sllce number

142

Figure 5.5. Cumulative L. monocytogenes transfer (%) from an inoculated) knife blade
(108 CFU/blade) to turkey, salami and bologna.

gunuunuuuunuuuuuuuunuuu

 

9‘:
J
8*?
A 9 DD
3574. E3000 oTurkey
h I
agar} uSalami
: i
254 ABologna
" i
0
>4] AAAAAAAAAAAAAAAAAAAAAA
g 1 AAAéoooooooooooooooooooooo
.231 AAo°°
=2J A 0
O ‘ <>
1 A0
la
07' . ' T T Y 1
O 5 10 15 20 25 30

Slice number

143

Figure 5.6. Cumulative L. monocytogenes transfer (%) from an inoculated knife blade
(105 CF U/blade) to turkey breast, salami and bologna

Cumulative Transfer (%)

501

45*

40

35a
30a

25

201

15
10

5-

0

._.

4

J
J

 

O

AAAAAAAAAAAAAA
AA

000000000
00

Dugnaununnnuuuun

 

T r

5 10 15 20
Sllce number

144

0 Turkey
E1 Salami

A Bologn

Figure 5.7. Cumulative L. monocytogenes transfer (%) from an inoculated knife
blade (103 CFU/blade) to turkey breast, salami and bologna

Cumulative Transfer (%)

 

 

405 0Turkey
351 OOOOOOOOOOUSaIami
aBologna
304 AAAAAA
A
25~ AAA
DDDDU
20~ ° C]
A Dunn
154 A
A0
El
10- U
5- o
o [9,, ,, ,, - a, ,L
0123 6789101112131415

Sllce number

145

Figure 5.8. Cumulative L. monocytogenes transfer (%) from inoculated turkey (IT)
and salami (lsxloS CFU/cmz) to uninoculated turkey (1m and salami
using an uninoculated knife blade.

 

 

21 GIT-UT
1.8‘! DlT-US
£16.,i 000000 AIS-UT
a. 0
1.4
'3 1 00°
51.2% o <>°°
*- 1J 0°
g :
2:0.8~
£06
:1 ,4
E 000
30.4—
002- AAAAAAAAAAAAAA
.O[ Agggguuunnuggnuuunn
0 5 10 15 20

Slice number

146

5.4.2. Predictive model for L. monocytogenes transfer during slicing of turkey,
bologna, and salami using a mechanical slicer.

Using the previously identified predictive model, a program was developed using
GWBasic for a series of correlation coefficients of predicted versus observed values for
transfer using slicer and knife blade scenarios with example output illustrated in Figures
11-14. High-level inoculation of products (108 CFU) resulted in the lowest variance or

fractional differences (R2>0.9O) for observed vs. predicted values for all models tested

(Figure 5.9).

Figure 5.9. Example: GWBasic output for salami sliced using an inoculated knife
blade (108 CFU/blade)

Fraction left on blade during each slice= .8150896

CFUs transferred to lst slice= 2173269

Above results are independent of No

If initial CFUs on the blade= 1E+O8, then .

fraction transferred to the product during each slice = 2.173269E-02
fraction transferred to surroundings during each slice = .1631777
fitted equations (all equivalent) are:

1) 1n CFUs = -.2044573 *5 + 14.7962

2) CFUS = 2666295 "' .8150896 "5

3) CFUs = 2666295 *e"(-.2044573 *s)

4) CFUs == 2666295 *10"(-8.879466E-02 *s)

correlation coefficient for fit is R2 = .964417

147

Figure 5.10.

Plotted output using GWBasic for assessing L. monocytogenes transfer
fi'om an inoculated slicer blade (108 CF U/blade) to salami

—o—— Observed
------- Predicted

   

R2 = 0.92

 

5 10 15 20 25 30
Slice number

Identifying a predictive model for lower level inoculations (10S and 103

CF U/blade) was much more difficult with different models being selected based on the

product sliced.

Turkey and bologna were similar (R2 > 0.90) at 105 CFU/blade (Figure

5.11). Salami was not modeled at 105 CFU/blade for product sliced on the delicatessen

slicer due to continued transfer out to 30 slices in the absence of any trend or regression

(Figure 5.2).

148

Figure 5.11. Plotted output using GWBasic for predicting L. monocytogenes transfer

  

 

from an inoculated slicer blade (10 CF U/blade) to turkey.
3.5
] ——o—Observed
3.0 l ------ Predicted
l
2.5 ii
a 2.0 1
0 1
§ 1.5 1'
1.0 J
l
0.5 J
0.0 l . . - - .
0 2 4 6 8 10 12

Sllce number

Low-level inoculation (103 CFU/blade) resulted in greater variance in all models
tested. The model identified for turkey and bologna resulted in large differences between
observed and predicted values (0.52 s R2 S 0.65) when compared to higher inoculation
levels (Figure 5.12). Salami showed a weak correlation coefficient (R2 <0.40) for all

models tested (Figure 5.13) with residuals greater than 50% of the measured values.

149

 

Figure 5.12. Plotted output using GWBasic for redicting L. monocytogenes transfer

from aninoculated slicer blade (10? CFU/blade) to turkey and bologna

  
 
  

—e— Observed Turkey
....... Predicted Turkey
+ Observed Bologna
_____ Predicted Bologna

Log CFU

 

Slice number

150

Figure 5.13. Plotted output using GWBasic for predicting L. monocytogenes transfer
from an inoculated slicer blade (10 CFU/blade) to salami

—0— Observed
------ Predicted

  

A
0'!
.l__2 _.... -L__.a.__m_ ._l____._-._

 

0.0 , ‘ —.v———. , ._-- . . , __ _-.-_
0 2 4 6 8 1O 12 14 16 18 20 22
Slice number

 

 

 

151

 

 

5.5. DISCUSSION

Predictive modeling of microbial pathogens during food production and storage
has been approached using various mathematical models and methods. Common
approaches have included such empirical modeling or mechanistic mathematical
translation of various factors including attachment properties and metabolic functions
(Bernaerts et al., 2004). Each of these approaches has many advantages and
disadvantages. Predictive models based on mathematical translation of biological
functions can be rapid and less costly than empirical models requiring little or no
laboratory experimentation. While derived mathematical models use biological
mechanisms to predict population outcome, environmental scenarios may strongly
influence an outcome that is not readily identified or foreseen with mathematical
derivations and manipulation. Curve fitting models predict population outcomes based on
previously obtained experimental data offering arguably more accurate interpretation of
predicted environmental populations. However, many curve fitting models based on
empirical data do not account for underlying biological factors that influence the fitting
of results and in some cases may be dependent on specific environmental or laboratory
conditions. Limited work has been done modeling bacterial transfer to and from food
contact surfaces. A study by Schaffner et al. (2004) used a Monte Carlo simulation for
predicting bacterial populations on cutting boards over time. Results from this study and
subsequent simulation predicted a contamination level greater than 20 CF U/4cm2 after 15
min and greater than 40 CFU/4cm2 after 45 min. While this study provides some insight

into modeling total bacterial transfer during slicing of deli meats, other important

152

parameters including surface scoring, surface roughness, cutting force, and physiological

differences in bacterial attachment were not addressed.

Models generated for this research were based on empirical data generated in a

laboratory during mechanical and knife slicing of turkey, salami, and bologna with the

following 12 product, environmental and blade variables affecting the outcome of these

models:

8.

9.

. Product fat content — high fat (salami) vs. low fat (turkey)

Product moisture content — high moisture (turkey) vs. low moisture

(salami)

Product composition — homogeneous (bologna) vs. heterogeneous
(salami)

Product temperature — frozen (< 0°C) vs. refrigerated (<0-7°C) or
abusive (7-23° C)

Environment - low (<50%) versus high (>50%) relative humidity

Blade stainless steel grade - 304 vs.3 16
Blade sharpness — sharp vs. dull or broken
Blade thickness - thin vs. medium or thick.

Blade cutting speed/force — slow vs. fast

10. Blade age — changes in surface roughness, /wear, and scoring, and pitting

over time.

11. Blade surface finish - ZB vs. electropolished.

12. Blade/knife edge — serrated vs smooth

153

Assumptions made for fitted equations in this model were: a) the expected
number of Listeria transferred (CF U) during slicing is the fraction “f1” that describes how
many Listeria are on the blade just before each sequential slice b) the expected number of
Listeria transferred to the surrounding areas is a different fraction “f2” that describes how
many Listeria are on the blade just before slicing, and c) No . the number of Listeria cells
on the blade that are available for transfer before any slicing begins. These fractional
assumptions “f1” and “f2” are expected to be constant because of the degree of adhesion
between Listeria and the blade/meat surface. A surface exhibiting a low-level of

adhesion would allow the fraction transferred to the meat to approach 1 and the fraction

 

transferred to the surface to approach 0. This would allow the fraction left on the blade
“ft” to approach 0 while the fraction left on the surface “f2” would approach 1. The
reverse is true for a surface exhibiting a high-level of adhesion where the fraction left on
the surface “f1” approaches 1 and the fraction on the meat “f2” approaches 0. When using
a mechanical slicer, it is assumed that any Listeria on the blade are uniformly and
randomly distributed due to the exceedingly high rotation rate of 8 rotations per second
during slicing. Given the high number of revolutions between slicing (60-80), Listeria is
uniformly transferred from the blade to the product or from the product to the blade. This
assumption was verified when turkey, salami, and bologna were sliced using knife
blades. Listeria transfer coefficients were dependent on both the type of product being
sliced and the method of slicing. Using the delicatessen slicer, transfer coefficients were
higher for high fat, low moisture salami when compared to low fat and high moisture
turkey. When turkey and salami were sliced using the knife blades, turkey yielded a

higher transfer coefficient for blades inoculated at 105 and 103 CF U/blade. Lin et al.

154

 

(2004a) reported similar findings using a mechanical slicer with the presence of a fat
layer extending transfer out to more slices. These findings indicate that different Listeria

transfer scenarios can be expected based on the previously identified 12 parameters.

The four fitted models described herein of the form [CFU (X) = kax ] along with
the program written in GWBasic can be used if any two of the following three values are
known: a) initial inoculum level, b) total bacterial transfer, c) fraction of bacteria
remaining on blade after consecutive slicing, solving for each model parameter CFU(X),
k, or a. Further extrapolations can be done using previously reported data for turkey,
salami, and bologna to estimate each of these model parameters using predicted fitted
equations having correlation coefficients greater than 85%. While many low inoculum
levels exhibited weak correlation coefficients, the last 10 slices and the first 5-10 slices of
high-level (108 CFU/blade or cm2) and low-level inoculations (103 CFU/blade or cmz)

can be used to estimate the number of Listeria transferred during slicing.

5.6. SUMMARY

Based on our model, the greatest number of Listeria (>90%) will be found in the
first 15 slices of delicatessen meats afier mechanical or knife slicing. The model
presented in this research has been simplified using limited factors based on fractions
transferred to the blade, meat and surroundings. This model can be used as a starting
point to identify additional parameters that impact bacterial transfer. Depending on any
one or combination of the aforementioned 12 parameters affecting transfer, certain
delicatessen meats that permit growth of Listeria may pose a public health risk to certain

consumers if the product is subjected to extended refrigerated storage. The mathematical

155

construction of the model presented in this study provides a framework for designing

future models with different parameters, environments, and processes.

156

CONCLUSIONS AND FUTURE RECOMMENDATIONS

Post-process contamination of ready-to-eat meats from Listeria monocytogenes
continues to be a health risk to the public and a safety and financial risk to the processor
and retailer alike. Results from this research demonstrate the ability of L. monocytogenes
to transfer from contaminated product to uncontaminated products using a slicer or knife
blade as a vector. Quantitative recovery of Listeria from solid surfaces was identified as
a hurdle for both accurate and precise assessmen t of transfer from contaminated products
to food contact surfaces.

The first objective of this research, optimizing quantitative recovery of L.
monocytogenes from stainless steel surfaces, was improved (> 0.5 log cfu) using a
composite tissue device, which was inexpensive and easy to use when compared to
traditional sponge, and swabbing devices. In addition to improved efficacy, larger
heavily soiled areas could be sampled without using multiple devices. While this device
represents a significant improvement to traditional sampling devices, future research
needs to be conducted to further develop new and innovative sampling devices for
improved accuracy and precision when sampling solid surfaces.

The second objective of this research was to determine direct and sequential
transfer rates for L. monocytogenes fi'om artificially contaminated ready-to-eat luncheon
meats to a delicatessen slicer and vice versa. Our findings provided valuable insight into
distribution of L. monocytogenes during slicing of contaminated ready-to-eat meats.
Results fiom this study demonstrated greater transfer (P<0.05) from inoculated turkey
(108 CFU/cmz) to the five slicer contact areas using a cutting force of 10 as opposed to 0

lbs. When slicer blades were inoculated at 108 CFU/blade Listeria populations decreased

157

 

 

logarithmically to 102 CFU/slice after 30 slices. Findings for inoculated slicer blade and
products (105 CFU/blade or cmz) were similar with Listeria counts of 102 CFU/slice after
5 slices and enriched samples generally negative afier 27 slices. Using 103 CFU/blade,
the first 5 slices typically contained ~10l CFU/slice by direct plating with enrichments
negative after 15 slices. Product composition had a major impact on transfer during
slicing. Higher fat and lower moisture of salami compared to turkey and bologna
resulted in prolonged Listeria transfer with a fat layer developing on the blade. Our
finding also suggest a major impact of product composition on transfer with blade surface
roughness changing significantly (>100%) during one year of use. Numerous areas of
future research were identified during this study. Material composition of the slicer,
surface finish, effects of cleaning regimens on slicer oxidation and pitting, as well as
environmental temperature and relative humidity are limited examples of interactions that
are poorly understand and in most cases not documented.

Similar to objective two, the third objective, determining the effects of cutting
force, stainless steel grade, sharpness, and product composition on transfer of L.
monocytogenes from artificially contaminated ready-to-cat luncheon meats to knives and
vice versa, identified product composition as having a significant impact on transfer.
Listeria transfer from knife blades inoculated at 108 CFU/blade was logarithmic with a 2-
log decrease after 12 slices and direct counts obtained thereafter out to 30 slices.
However at lower inoculation levels of 105 and 103 CFU/blade, direct counts were
typically only observed out to 20 and 5 slices, respectively. Stainless steel grade had a
significant impact on transfer with greater “tailing” or more prolonged transfer from

grade 304 as opposed to grade 316 stainless (P<0.05) for all three products. After one

158

 

 

year of use, knife blade roughness values were significantly greater (P<0.001) for grades
304 than 316 stainless. Surprisingly, force and knife sharpness were not significantly
different (P>0.05) within stainless steel grade (P<0.05) for each product. However,
significant differences in force were seen between salami and turkey (P<0.05) for grades
304 and 316 stainless steel. This study also identified numerous areas of future research
with material composition (chemical and physical properties), surface finish, effects of
cleaning regimens on blade oxidation and pitting, as well as environmental temperature
and relative humidity being limited examples of interactions that are poorly understand
and in most cases not documented.

In the final objective, development of one or more mathematical models based on
the transfer coefficients obtained from the previous three objectives that will predict the

numbers of L. monocytogenes cells transferred during slicing of delicatessen meats, four

variations of a model identified as [CFU (X) = kax ] along with a program in GWBasic.
The model variations and subsequent program were based on the following three
assumptions: 1) the expected number of Listeria cells transferred during slicing is the
fraction “ft” that describes the number of Listeria cells on the blade just before each
sequential slice, 2) the expected number of Listeria cells transferred to the surrounding
areas is the fraction “f2” after each slice, and 3) the number of Listeria cells on the blade
available for transfer before any slicing begins is No. The aforementioned model and
variations thereof can be used if any two of the following three values are known: (a)
initial inoculum level, (b) total bacterial transfer, (c) fraction of bacteria remaining on
blade after consecutive slicing, solving for each model parameter CF U(X), k, or a. The

overall finding using transfer coefficients in the models resulted in the greatest number of

159

 

Listeria (>90%) being found in the first 15 slices. As a result of this study, numerous
areas of future research were identified emphasizing the future research needs presented
in objectives two through four. In addition to better understanding of material and
product composition, experimental design using a hybrid of mathematical manipulation
and empirical data gathering is needed to evaluate parameter effects. Traditional
approaches using empirical data gathering to find parameter effects are very costly and
may result in the neglect of underlying factors hidden with the parameters being
evaluated. Modifications to each of these approaches (mathematical manipulation and
empirical data gathering) will minimize cost and time spent finding each of these
parameter effects. In addition to model identification and development, future research
needs to be done validating such models using independent researchers and laboratories. A

Many recommendations to food manufacturers, retailers, and researchers can be
taken fiom these studies. Transfer of bacterial pathogens was found to be sequential with
scoring and surface finish changing significantly with usage for both slicer and knife
blades. As a result of these studies recommendations to slicer manufactures would
include better polishing of food contact areas using electropolishing or high grade
buff'mg. Current slicer configurations use a low-grade stainless (304) and polish (No. 4)
on all areas except the slicer blade which is typically electropolished. A higher grade
stainless such as 316 would result in greater resistance to oxidation and corrosion and
subsequently decrease scoring and areas of attachment for bacteria. Improving surface
finish by electropolishing to obtain a mirror-like buffed finish on all food contact areas
would decrease food particulate and bacterial attachment as well as enhancing visual

soiling. Slicer design has also been identified as an area for improvement by eliminating

160

unnecessary machining and creating smoother transitions between contact areas to ease in
cleaning.

Additional recommendations resulting from this study can be applied to food
utensil manufacturers. Food utensils have many variations in design, composition and
surface finish. Based on our findings, 316 grade stainless was superior to grade 304 and
is recommended for improved initial smoothness and surface wear after continued use.
Future research needs include the compositional effects of grade 316 stainless on cell
recovery. Addition of molybdenum and other compounds may result in new and
innovative alloys with bactericidal properties, thus minimizing biofilm formation and
subsequent transfer.

Overall, this work represents a major contribution to the area of bacterial transfer
in processing and retail environments and has generated many new research avenues for
further investigation of post-process contamination at the retail level. Future research
building on this study will contribute greatly to a better understanding of the distribution
and dissemination of bacterial pathogens in the food supply along with major

improvements in the current risk assessments.

161

 

APPENDIX I

 

162

Preliminary Analysis of Dr. Robert McMaster

11 October 2003

Background:

This is a preliminary report based on the two sets of data I've received from Keith
Vorst on slicing the turkey breast. There were 21 data points taken on 18 June 2003 and
29 points measured on 28 August 2003. Keith forwarded the information to me on 15

September 2003.

Analysis:

I have analyzed the data using seven different functions, allowing them to
compete on the basis of closeness of fit. Each function contains either 2, 3 or 4
parameters, which are optimized in order to obtain the closest fit for the respective
functional form. Each of these parameters could be considered a type of "transfer

coefficient" as we had set forth in our original proposal.

Once the best fit for each function is found, the standard of deviation of the
residuals is then compared for each function in order to determine which function
provides the best fit. The residuals are simply the differences between each measured

data point and the value of the same point as calculated by the model. The model with

163

 

the smallest standard deviation in the residuals is generally assumed to be the most
appropriate. However, this must be tempered by the number of parameters used in the

function.

Functions with more parameters may provide a better fit, not necessarily on the
basis of a more appropriate modeling of the physical behavior of the problem, but
because of the additional degrees of freedom offered by the higher order model. In order

to determine the most appropriate model, then, statistical methods can be employed to

 

discern whether the additional parameters in a higher order model provide improvement ‘
in the residuals of statistical significance. For example, using the “F” statistic, the “F”

. test can be employed in evaluating the reduction in the standard deviation of the residuals

in comparing a 2 parameter model to a 3 parameter model. The “F” statistic table is

entered with the number of parameters and the number of measurements and the result

gives the necessary reduction in the standard deviation of the residuals from the 3

parameter model to justify inclusion of the 3rd parameter beyond the 2 parameter model.

For cases where there is no physical reason to justify the inclusion of the 3rd parameter,

the “F” test should show that the inclusion of the additional parameter in the model is not

warranted.

 

I64

Results:

Of the two tests on hand to date, seven models were fitted to each as shown in
Tables 1 and 2 below. Table 1 summarizes the 21 data points taken on 18 June 2003 and
Table 2 summarizes the 29 points measured on 28 August 2003. The accompanying
figures show plots of the raw data with the best fit function superimposed. As can be
seen in Table I, the lowest standard deviation in the residuals is achieved by both Models

3 and 4. However, Model 3 has only 2 parameters (or "transfer coefficients") and Model

 

4 requires 3 parameters. Therefore, the "F" test is not required in this case, since the ‘

additional parameter in model 4 offered no improvement in the model performance.

Sliced Turkey 18 Jun 03

 

6.00

500 ~

4.00 a

3.00 ~

 

Log Count

2.00 n

1.00 -l

 

 

 

0.00 r r . .
0 5 10 15 20 25
Number of Sllce.

165

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table l
21 data points taken on 18 June 2003
Standard
Mod Equation A B C D Deviation
el of
Residuals
1 1701) = A + 39-“ 1.88534 3.070936 0.097862 0.192888
8
2 F07) = Ae-B" 4.61315 0.037861 0.216080
6
3 F(n)= A-Bln(n) 5.05141 0.891713 0.185079
5
4 F(n) = A - Bln(Cn) 5.02345 0.891714 0.969131 0.185079
' 5
5 FM) = A _ BnC 0.63240 4.596126 -0.29112 0.236264
3
6 170,) = A— 3,“.an 4.77643 0.217829 0.00472 0.200547
9
7 F01) = A-Bn+Cn2 +Dn3 5.00045 0.328542 0.017063 - 0.190184
1 0.00037 ~

 

In Table 2, we see that the same seven models are used, since the curve has the same

basic shape as the data analyzed in Table 1. In this case, Model 7 emerges as the best fit.

However, this model requires 4 parameters and the magnitude of the improvement in the

standard deviation of the residuals is very small. Moreover, the "F" test would not show

justification to include the additional parameter added in model 7 from model 6. More

significant still is the fact that Model 1 is superior to Model 6, even though it has one

fewer parameter. Therefore, Model 1 is clearly the most appropriate choice of the

functions which were fit to this set of data.

166

 

 

Sliced Turkey 28 Aug 03

 

700

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6,00 1 E

5 00 '1 E

5.... a

o i

E 3.00 ~ '

1 00 . i

0 00 r r r r r r i

O 5 1O 15 20 25 30 35
Number of Sllce:
Table 2
29 data points taken on 28 August 2003
Standard
Model Equation A B C D Deviation
of

Residuals
1 1701): A + Be‘c" 1.961016 4.219396 0.113016 0.253891
2 F(n) = Ae‘B" 5.316928 0.038814 0.373251
3 F(n) = A - B ln(n) 6.068183 1.197134 0.296056
4 F (n) = A — B ln(Cn) 6.059808 1.19713 0.993037 0.296056
5 F(n) = A _ BnC -58.4602 64.55685 -0.0192 0.298025
6 F(") = A .. Bn + 012 5.783087 0.296413 0.00606 0.263702
7 F(n) = A _ Bn+ of + Dn3 6.107309 0.414355 0.01547 -0.0002 0.242961

 

 

167

 

 

 

 

APPENDIX II

 

168

The example program using GWBasic is as follows:

3 L ‘ bWBuslc'- 139/BASIC. tXE

r

:11- l iljll

I:‘.'.I4\l

14"}

Using data obtained from a delicatessen slicer inoculated at 108 CFU/blade the

din iqu
1‘i11 :mlanlirl.

actual
‘1Ur- n (31:11
N Ilii: SIUI

l
(rl~“\'.l)- B ill-"61 I

program developed in GW basic is as follows:

"ii:
.17.

Uri blriilfl
rail 1..)

.li‘l! irn'u, ..

C’fii .: on
l1.‘1fi“t-‘:I1l‘iti

fI‘-‘1Li011 Eran." f (:1‘! tin

corruie‘t’c ion cut-ff it: iullt

h it EN I' Eli 1:1:
7

blarin‘ If: '03

— giutrn t—t—fu-"
I>N Ilil‘iN '.3I‘. )‘i

XI‘CNNS'TMHL" Hi .41 1

during r-.ll:l’r '.IILL‘ ’ "
t ' ' - ‘f.’
(.1 .iill.l.tI
.tllt-l.
nurll’ «luring each
r, lli‘l‘flllltli int, :1

'~llU.tIl::‘ 1 3' ”yr:

f U I"

i’ it is li-‘-

169

.tallj.

42f”,

v.‘
(luring 41.11711

.9411820

pro-ll it L xii

,m-Hl it 11L"...

liIutlt:

set: actual us predicted data (10 at it Lime)

(18.11

L‘

L 11‘“: 3"

 

 

 

 

APPENDIX III

170

Lag (LT) and Generation Time (GT) of Listeria monocytogenes on turkey, bologna

and salami at 4° C.

 

 

 

 

Temperature
Food GT (h) LT (h)
(C)
Roast Turkey 4 29 24
Bologna 4 MG N/G
Salami 4 N/G N/G

 

+ Turkey
-l— Bologna
+ Salami

3.0‘1 \\

.N
on

W

Log CFU/g
3 a E“.

 

 

0.0 I j T '—T— f ‘T—_——".'-'—‘_"_‘--T'—_‘—" "“ “_“-"T’—‘"T

o 5 10152025303540455055606570758085
l-lour3(h)

171

 

 

 

APPENDIX IV

 

172

From: Whiting, Richard C [mailto:Richard.Whiting@cfsan.fda.gov]
Sent: Thursday, January 13, 2005 4:39 PM

To: 'Keith Vorst'

Cc: Ryser, Elliot; Ewen Todd; 'carl.custer@fsis.usda.gov'
Subject: RE: Data for Deli Slicer and knife transfer

Keith,

I've looked at the data and propose you consider an exponential or first-order
decay model. This has a constant amount of material removed with each
subsequent slice. Log Ns = log No - ks Ns is the counts on a slice, No is the
original amount on the knife/blade, s is the slice number and k is the parameter
value. On the attached spreadsheet of yours the k appears to be about 0.1, this
means that 90% of the material remains on the knife and 10% goes to the slice.
The next slice that passes the knife removes another 10% but from a slightly
smaller amount remaining on the knife. The spreadsheet has my electronic
doodling and graphing of your data.

The difference between the inoculum and first slice appears to be about 2. 08 log
10 or about 0.8%.

However, I see this as a mass transfer not really a microbial transfer. It would be
useful to know the mass transferred from the original contaminated product to the
knife/blade. Then 10% of that product would be transferred with the first slice.

The three products appear to be similar, it would be interesting to see how other
products (including cheeses) would transfer. This might be related to the
product's friability and adhesion or it might be mostly adhering water. Your
pressure, surface texture would also be factors here.

I could see one model for the amount of mass (with bacteria) transferred to a
blade and the second model above for the removal from the blade to subsequent
clean product with each slice.

Let me know what this sounds like to you.

Richard Whiting

FDA, CFSAN HFS-302
5100 Paint Branch Parkway
College Park, MD 20740-3835

phone 301-436-1925

fax 301-436-2632
e-mail rwhiting@cfsan.fda.gov

173

Model Generated by Dr. Whiting FDA/CFSAN

 

 

 

 

 

 

N5 = No -
exp(l<s)
No = amount on first slice
3 '3
slice
No k
7 0.1
slice Ns l
1 6.90
2 6.80
3 6.70
4 6.60
5 6.50
6 6.40 a
7 6.30 g
8 6.20 .2
9 6.10
10 6.00
11 5.90
12 5.80 E
13 5.70 l
14 5.60 l
15 5.50 we. 3
L J
No k

10"8 Bologna 6.843802 0.137483 5.853894 0.073729 5.374298 0.031559
Salami 6.246282 0.097275 6.363965 0.085096 6.50417 0.086507
Turkey 6.059119 0.110711 5.846207 0.105168 5.395735 0.075915

1045 Bologna
Salami
Turkey

with exception of 1045
salami:
If appears about 1% of pathogens were transferred to blade and were removed with first slice

Subsequent slices have a k = 0.10. this means that 90% remains and 10% removed each slice.
If back calculate to slice zero. than original concentration was 2.08 + 0.1 + 2.19

Nprod-
This is a transfer of mass, bacteria are along for the ride. No
ave 0.08 1.98
6.02 0.09 1.63
6.37 0.10 2.23
5.77
0.11 2.56
2.44 0.01 2.10
2.90 0.11 1.99
3.01 0.10 2.08
ave =5

174

1.
“3' h.

 

Slicer Blade to Product (108 CFU/Blade)

Product
REP
Slice #
Count
Level

Rep

J—b-L-L—LA—L—b—k—S—t«L—L—L—L-A—fi—A—L—L—l—l—Ae—L-Au-l—l—A-Au-L

NNNNNNNN

Bolgona = 1, Salami = 2. Turkey =3

CFU
10"8

(DNQUIth—l

Count

7.785406
8.35E+06
2555-06
1.71 E+06
8.54E+05
1.01 E+06
4.42E+05
2.94E+05
3.49E+05
4.48E+05
3.98E+05
3.36E+05
5.97E+05
9.25E+04
5.70E+04
8.86E+03
1.22E+04
8.20E+03
9.27E+03
1.11E+04
68551-03
2.12E+04
3.84E+03
2.27E+03
1.61 E+03
2.69E+03
5.1GE+03
0.00E+00
8.76E+02
0.00E+00

2.895+06
5.74E+05
3.31 E+05
3.23E+05
2.24E+05
2.34505
1.76505
8.76E+04

Product

AAd—fi—fi—L—L—L—l-fi-fi—L—L—l—l-fi—L—LAAA—l—L—L-A—L—L—l—L-L

AAA-AAAa—k—A

log count
6.89
6.92
6.41
6.23
5.93
6.00
5.65
5.47
5.54
5.65
5.60
5.53
5.78
4.97
4.76
3.95
4.09
3.91
3.97
4.05
3.84
4.33
3.58
3.36
3.21
3.43
3.71

2.94

6.46
5.76
5.52
5.51
5.35
5.37
5.25
4.94

175

bologna rep 1

No

R

6.843802 0.137483092

model
6.71
6.57
6.43
6.29
6.16
6.02
5.88
5.74
5.61
5.47
5.33
5.19
5.06
4.92
4.78
4.64
4.51
4.37
4.23
4.09
3.96
3.82
3.68
3.54
3.41
3.27
3.13
2.99
2.86
2.72

5.78
5.71
5.63
5.56
5.49
5.41
5.34
5.26

diff sq

0.034
0.125
0.001
0.004
0.051
0.000
0.056
0.076
0.004
0.033
0.072
0.110
0.518
0.002
0.001
0.485
0.177
0.207
0.070
0.002
0.015
0.258
0.010
0.035
0.040
0.026
0.337

0.007

0.464
0.003
0.013
0.003
0.018
0.002
0.008
0.103

bologna
rep 2

No
5.853894
k
0.073729

SUlTl
2.756232

NNNNNNNNNNNNNNNNNNNNNN

wwwwwwmwwwwwmwwwwwwwwwwwwww

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

.5
OCDONCDOI-wa—l

NNNNNNNN—L—sa—saagg.‘
“mthN-lbOCDmNOOI-th-A

2.02E+05
9.63E+04
8.34E+04
1.1 1 E+05
1.3OE+05
9.28E+04
6.22E+04
2.36E+04
2.48E+04
3.71 E+04
2.23E+04
1.90E+04
1.82E+04
1.59E+04
8.7QE+03
1.71 E+04
1 .03E+04
1 .02 E+04
1 4854-04
3.95E+03
5.90E+03
7.57E+03

3.17E+06
86254-05
2.23E+05
1.60E+05
1095-1-05
8.29E+04
7.54E+04
1 .39E+05
7.23E+04
1 .86E+05
1.33E+05
9.88E+04
8.91 E+04
7.35E+04
1.01 E+05
6.32E+04
6.3OE+04
8.33E+04
1.02E+05
2.01 E+04
2.27E+04
6.77E+04
4.96E+04
2.81 E+04
47654-04
4.92E+04
5.24E+04

A—b—l—l—A—A-A—t-L—lé-S-fi—L—L—L—l-L—Ld—L—L—L—L-A—l-L

5.31
4.98
4.92
5.05
5.11
4.97
4.79
4.37
4.40
4.57
4.35
4.28
4.26
4.20
3.94
4.23
4.01
4.01
4.17
3.60
3.77
3.88

6.50
5.94
5.35
5.20
5.04
4.92
4.88
5.14
4.86
5.27
5.12
4.99
4.95
4.87
5.00
4.80
4.80
4.92
5.01
4.30
4.36
4.83
4.70
4.45
4.68
4.69
4.72

176

5.19
5.12
5.04
4.97
4.90
4.82
4.75
4.67
4.60
4.53
4.45
4.38
4.31
4.23
4.16
4.08
4.01
3.94
3.86
3.79
3.72
3.64

5.34
5.31
5.28
5.25
5.22
5.18
5.15
5.12
5.09
5.06
5.03
5.00
4.96
4.93
4.90
4.87
4.84
4.81
4.77
4.74
4.71
4.68
4.65
4.62
4.59
4.55
4.52

0.013
0.018
0.015
0.006
0.047
0.021
0.002
0.090
0.042
0.002
0.011
0.010
0.002
0.001
0.046
0.022
0.000
0.005
0.094
0.037
0.003
0.056

1.343
0.390
0.005
0.002
0.032
0.071
0.076
0.000
0.053
0.044
0.009
0.000
0.000
0.004
0.01 1
0.005
0.001
0.013
0.054
0.193
0.127
0.023
0.002
0.028
0.009
0.019
0.039

SUN
1 .158075

00000)

AAAA—l—t—LJ—fi-A—L—L—b—bu—L-Q—l—L-J—IAJAAA-h-AAAA

NNNNNNNNNNNNNNN

28
29

(Dmflmm-hOJN-b

.5
OfOCDNCDUItwa-b

4.3.5.3.;
(fléwNfi

4.38E+04
1.77E+04
3.44E+04

6.59E+05
1.70E+06
1.58E+06
1.1SE+06
8.63E+05
5.56E+05
8.57E+05
5.73E+05
3.19E+05
5.1 1 E+04
7.45E+04
8.64E+04
6.40E+04
7.01 E+04
1.69E+04
3.91E+04
2.40E+04
20354-04
6.70E+04
3.80E+04
1.8OE+03
2.16E+04
1.54E+04
1.81 E+04
1.91 E+04
1.04E+04
4.35E+03
1.18E+03
2.07E+03
3.38E+03

2.88E+05
3.93E+06
5.10E+06
1.50E+06
1.09E+06
2.49E+06
4.22E+05
5.33805
3.81 E+05
4.33E+05
1 .64E+05
9.29E+04
1 .02 E+05
8.41 E+04
8.86E+04

—L

NNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

NNNNNNNNNNNNNNN

4.64
4.25
4.54

5.82
6.23
6.20
6.06
5.94
5.75
5.93
5.76
5.50
4.71
4.87
4.94
4.81
4.85
4.23
4.59
4.38
4.31
4.83
4.58
3.26
4.33
4.19
4.26
4.28
4.02
3.64
3.07
3.32
3.53

5.46
6.59
6.71
6.18
6.04
6.40
5.63
5.73
5.58
5.64
5.21
4.97
5.01
4.92
4.95

177

4.49
4.46
4.43

6.15
6.05
5.95
5.86
5.76
5.66
5.57
5.47
5.37
5.27
5.18
5.08
4.98
4.88
4.79
4.69
4.59
4.50
4.40
4.30
4.20
4.11
4.01
3.91
3.81
3.72
3.62
3.52
3.43
3.33

6.28
6.19
6.11
6.02
5.94
5.85
5.77
5.68
5.60
5.51
5.43
5.34
5.26
5.17
5.09

0.023
0.044
0.012

0.11
0.03
0.06
0.04
0.03
0.01
0.14
0.08
0.02
0.32
0.09
0.02
0.03
0.00
0.31
0.01
0.05
0.04
0.18
0.08
0.90
0.05
0.03
0.12
0.22
0.09
0.00
0.20
0.01
0.04

0.67
0.16
0.36
0.02
0.01
0.29
0.02
0.00
0.00
0.02
0.05
0.14
0.06
0.06
0.02

sum

1 .289239

sum

3.31

NNNNNNNNNNNNNNN

wwwmwwwwwwwwwwwwwwwwwwwwwwwwww

16
17
18
19
20
21
22
23
24
25
26
27
28
29

CDQNGU'l-bWN—I

8.23E+04
8.36E+04
7.12E+04
4.12E+04
4.05E+04
3.91E+04
3.13E+04
3.69E+04
3.30E+04
1.35E+04
8.48E+03
8.95803
1.10E+04
1.64E+04
9.83E+03

1 .70E+05
4.55E+06
3.34E+06
1 .95E+06
2.42E+06
1.18E+06
7.48E+05
8.82E+05
1 7861-06
5.25E+05
3545-1-05
3.84E+05
2.47E+05
1 .31 E+05
1 .08E+05
7.47E+04
6.56904
7.81 E+04
5.15E+04
6.095404
4.95E+04
3.64E+04
2.74E+04
3.55E+04
2.99E+04
3.23E+04
1 .01 E+04
1 .08E+04
9.07E+03
9.37E+03

9.35E+06
4.01E+06
1.28E+06

NNNNNNNNNNNNNNN

MNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

000000

4.92
4.92
4.85
4.61
4.61
4.59
4.50
4.57
4.52
4.13
3.93
3.95
4.04
4.21
3.99

5.23
6.66
6.52
6.29
6.38
6.07
5.87
5.95
6.25
5.72
5.55
5.58
5.39
5.12
5.03
4.87
4.82
4.89
4.71
4.78
4.69
4.56
4.44
4.55
4.48
4.51
4.00
4.03
3.96
3.97
6.97

6.60
6.11

178

5.00
4.92
4.83
4.75
4.66
4.58
4.49
4.41
4.32
4.24
4.15
4.07
3.98
3.90
3.81

6.42
6.33
6.24
6.16
6.07
5.99
5.90
5.81
5.73
5.64
5.55
5.47
5.38
5.29
5.21
5.12
5.03
4.95
4.86
4.77
4.69
4.60
4.51
4.43
4.34
4.26
4.17
4.08
4.00
3.91

5.95
5.84
5.73

0.01
0.00
0.00
0.02
0.00
0.00
0.00
0.03
0.04
0.01
0.05
0.01
0.00
0.10
0.03

1.41
0.11
0.08
0.02
0.10
0.01
0.00
0.02
0.28
0.01
0.00
0.01
0.00
0.03
0.03
0.06
0.05
0.00
0.02
0.00
0.00
0.00
0.01
0.01
0.02
0.06
0.03
0.00
0.00
0.00

1.05
0.59
0.14

sum

sum

2.19

7.87

AAA—8AA—LA—b—L—A—A—l—l—L-A—l—L-A—L—b—l—l—h—L-L—l

NNNNNNNNNNNNNNNNNNNNNN

LOST

4.32E+05
2.12E+05
2.35E+05
2.19805
1.1 1 E+05
9.83E+04
1 .25E+04
8.07E+03
2.09804
1.22E+04
1.24E+04
5.25E+03
5.46E+03
2.05E+03
4.64E+03
1.22E+03
3.03E+03
7.1 1 E403
5.49E+03
7.37E+03
5.47E+03
4.35E+03
2.61 E+03
1.67E+03
1.27E+03
1.58E+03

1 .52E+06
1 .30E+06
5.63E+05
1 3954-05
2.41 E+05
1 .75E+05
6.55E+04
1.09E+05
9.09E+04
5.85E+04
3.75E+04
9.03E+03
5.71E+04
2.56E+04
8.85E+03
9.32E+03
4.64E+03
6.06E+03
2.55E+03
2.32E+03
3.17E+03
2.985+03

onwwwwwwwwmwwwwwwwwwwwwwwwww

wwwwwwwwwwwwuwwwmwwwww

5.64
5.33
5.37
5.34
5.05
4.99
4.10
3.91
4.32
4.09
4.09
3.72
3.74
3.31
3.67
3.09
3.48
3.85
3.74
3.87
3.74
3.64
3.42
3.22
3.10
3.20

6.18
6.11
5.75
5.14
5.38
5.24
4.82
5.04
4.96
4.77
4.57
3.96
4.76
4.41
3.95
3.97
3.67
3.78
3.41
3.37
3.50
3.47

179

5.62
5.51
5.39
5.28
5.17
5.06
4.95
4.84
4.73
4.62
4.51
4.40
4.29
4.18
4.07
3.96
3.84
3.73
3.62
3.51
3.40
3.29
3.18
3.07
2.96
2.85
2.74

5.74
5.64
5.53
5.43
5.32
5.22
5.11
5.00
4.90
4.79
4.69
4.58
4.48
4.37
4.27
4.16
4.06
3.95
3.85
3.74
3.64
3.53

0.02
0.00
0.01
0.03
0.00
0.00
0.55
0.68
0.09
0.18
0.09
0.32
0.19
0.57
0.08
0.58
0.06
0.05
0.05
0.22
0.20
0.21
0.12
0.07
0.07
0.21

0.19
0.23
0.05
0.08
0.00
0.00
0.09
0.00
0.00
0.00
0.01
0.40
0.08
0.00
0.10
0.04
0.15
0.03
0.19
0.14
0.02
0.00

sum

6.43

 

NNNNNNNN

wwwwcommonwwwwwwwmwwwwwwwmwwwwww

23
24
25
26
27
28
29

.5
ODONCDU’IthN-b

NMNNNNNBNN-L-LAAAJA—ng
(Dmflmmvhb) AO‘OQ‘IO’U’l-wa-I

30

5.965+03
1 .07E+04
3.00E+03
5.48E+03
1.33E+03
2.77E+03
3.22E+02
1 545-02

‘ 5.37E+06

7.00E+05
4.45E+05
2.39E+05
1 .06E+05
6.38E+04
3.29E+04
7.44E+04
4.50E+04
1 .9854-04
1 .39E+04
1 .36E+04
56454-03
4.28E+03
2.155+03
4.29E+03
8.27E+03
5.17E+03
3.60E+03
2.89E+03
2.19E+03
7.47E+03
1.74E+04
1 .16E+04
1.91 E+04
7.80E+03
6.06E+03
4.25E+03
2.03E+03

LOST

wwwwwwww

wwwmwwwuwwuwwwmwwwwwwwwwwwmwww

3.78
4.03
3.48
3.74
3.12
3.44
2.51
2.19

6.73
5.85
5.65
5.38
5.03
4.80
4.52
4.87
4.65
4.30
4.14
4.13
3.75
3.63
3.33
3.63
3.92
3.71
3.56
3.46
3.34
3.87
4.24
4.06
4.28
3.89
3.78
3.63
3.31

180

3.43
3.32
3.22
3.11
3.01
2.90
2.80
2.69

5.32
5.24
5.17
5.09
5.02
4.94
4.86
4.79
4.71
4.64
4.56
4.48
4.41
4.33
4.26
4.18
4.11
4.03
3.95
3.88
3.80
3.73
3.65
3.57
3.50
3.42
3.35
3.27
3.19
3.12

0.12
0.50
0.07
0.39
0.01
0.29
0.08
0.25

1.99
0.36
0.23
0.08
0.00
0.02
0.12
0.01
0.00
0.12
0.17
0.12
0.43
0.49
0.85
0.30
0.04
0.10
0.16
0.17
0.21
0.02
0.35
0.24
0.61
0.22
0.19
0.13
0.01

sum

sum

3.54

7.76

bologna rep 3
No k
5.374298 0.031559

 

 

W300?

 

' 7.00 "31:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l .
l l—fi“ 1
l l E—A—Repi‘;
\ . I . : 1112 I
II . ' -‘
\ . . l I Rep32l
It2 I ' ?’
. \,\ - ‘L m3
0 A.
0 i
e 9 i
l e
l 3.00
l
I
l
l
12.00 V . . . .. , .
l 0 5 10 15 20 25 30 35

 

181

Salami rep 1 rep 2 rep 3
No k No k No k
6.246282 0.097275 6.363965 0.085096 6.50417 0.08651

 

 

 

l
Tilers—55;?
i_fit1 ii
1+rep2 l
‘—a—fit2
i o rep3

l,--, fit =1_-

 

 

 

 

 

182

turkey rep 2 turkey rep 3
k N0 k N0 k
turkey
rep 1 0.110711 5.846207 0.105168 5.395735 0.075915
N0
6.059119

 

 

8.00 f 7 7 -. .. .7, a ,

 

 

i»‘arep21|

.—fit2 ;
Hit—repSli
l—fits j

 

i
l
l

 

 

 

2.00 Leaf”- . . Mnefi .
o 5 10 15 20 25 30 35 ,

 

183

Slicer blade to product (105 CFU/blade)

Product Bolgona = 1, Salami = 2. Turkey =3
REP 1-3

Slice # 1-15

Count CFU

Level 1045

* zero indicates negative direct plate but positive enrichement

'09

Rep Slice Count Product count
1 1 4.0E+02 1 2.60
1 2 1.QE+02 1 2.27
1 3 2.2E+01 1 1.34
1 4 2.6E+01 1 1.41
1 5 3.3E+01 1 1.52
1 6 1.3E+01 1 1.11
1 7 1.3E+01 1 1.11
1 8 3.7E+01 1 1.57
1 9 2.2E+01 1 1.35
1 10 1.6E+01 1 1.21
1 11 0.0E+00 1 .
1 12 0.0E+00 1
1 13 0.0E+00 1
1 14 0.0E+00 1
1 15 0.0E+00 1
2 1 8.0E+02 1 2.90
2 2 3.7E+02 1 2.57
2 3 2.3E+02 1 2.37
2 4 1.5E+02 1 2.18
2 5 1.7E+02 1 2.22
2 6 2.2E+02 1 2.33
2 7 1.1E+02 1 2.05
2 8 1.1E+02 1 2.05
2 9 5.3E+01 1 1.72
2 10 2.3E+01 1 1.35
2 1 1 1.1E+01 1 1.04
2 12 3.3E+01 1 1.52
2 13 1.3E+01 1 1.11
2 14 0.0E+00 1
2 15 1.ZE+01 1 1.09
3 1 2.4E+02 1 2.37
3 2 2.3E+02 1 2.36
3 3 8.6E+01 1 1.94
3 4 4.6E+01 1 1.66

184

wwwmuwwwwww

NNNNNNNNNNNNNNN

000003030300

1.2E+01
9.1 E+01
5.0E+01
3.5E+01
1.1E+01
2.3E+01
4.58-01
2.2E+01
2.2E+01
0.0E+00
1.9E+01

3.4E+03
3.4E+03
2.4E+03
2.7E+03
1 .7E+03
1 .7E+03
8.4E+02
0.0E+00
0.0E+00
0.0E+00
3.3E+02
1 .OE+03
6.4E+02
1.5E+03
2.3E+03

3.7E+02
7.8E+02
3.5E+02
2.4E+02
4.8E+02
1 .4E+02
4.1 E+03
3.1 E+02
3.ZE+02
7.SE+02
4.85-0'02
8.4E+02
2.0E+03
2.0E+02
4.1 E+02

5.9E+02
2.7E+02
1 .4E+03
7.7E+01
9.BE+01
0.0E+00

A—t-A—A—h—l—l—l-A—l-h

NNNNNNNNNNNNNNN

NNNNNNNNNNNNMNN

NNNNNN

1.06
1.96
1.70
1.55
1.04
1.36
1.65
1.34
1.35

1.28

3.53
3.53
3.38
3.43
3.22
3.23
2.92

2.52
3.02
2.81
3.17
3.37

2.57
2.89
2.55
2.38
2.68
2.13
3.62
2.50
2.51
2.87
2.68
2.93
3.30
2.31
2.61

2.77
2.43
3.14
1.89
1.99

185

wwwwwwmww

NNNNNNNNNNNNNNN

wwwwwwww

10
11
12
13
14

CDQNGO$O>NA

(DVCDUI-th—h

7.9E+02
7.5E+02
1.4E+03
7.0E+02
3.8E+02
7.3E+02
6.0E+00
4.4E+02
6.4E+02

1.3E+03
2.0E+03
1.1E+03
3.4E+02
7.4E+01
0.0E+00
1.1E+02
0.0E+00
1.8E+02
1.4E+02
0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00

1 .1E+03
6.5E+02
2.4E+02
2.9E+02
3.0E+02
3.6E+02
0.0E+00
0.054-00
1.8E+02
0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00

1.9E+03
4.9E+02
2.3E+02
2.6E+02
2.8E+02
1.1E+02
0.0E+00
1.2E+02

NNNNNNNNN

wwwwwwwwwwwwwww

wwwwwwwwwwwwwww

00030303000000)

2.90
2.87
3.15
2.84
2.58
2.86
0.78
2.64
2.81

3.12
3.30
3.05
2.53
1.87

2.06

2.25
2.15

3.06
2.81
2.39
2.47
2.48
2.56

2.25

3.27
2.69
2.36
2.41
2.45
2.03

2.09

186

wwuwwww

10
11
12
13
14
15

0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00
0.0E+00

0000000303000)

187

bologna

Slice

(DQVO'DUIJXWN—l

Rep 1
2.598572
2.273418
1.342423
1.407561
1.518514
1.108565
1.112605
1.567732
1.347135
1.21272

Rep 2
2.900695
2.573336
2.366983
2.180928
2.219899
2.334212
2.052309
2.047664
1.724276
1 .353339
1 .04454
1 .516932
1 .1 13943

1.09482

Rep 3

2.374382
2.360972
1.936514
1.658965
1.061452
1.958468
1.702431
1 .546049
1.038223
1.362482
1 .651278
1.339253
1 .351796

1.277838

Ave

2.62
2.40
1.88
1.75
1.60
1.80
1.62
1.72
1.37
1.31
1.35
1.43
1.23

1.19

No

2.43755

fit

2.33
2.22
2.11
1.99
1.88
1.77
1.66
1.55
1.44
1.33
1.22
1.11
1.00
0.89
0.78

0.110726

(ave-

fit)"2
0.089
0.035
0.050
0.060
0.081
0.001
0.002
0.028
0.005
0.000
0.016
0.102
0.055
0.787
0.168
1.479

 

log cfu

Bologna (3 reps)

 

 

 

 

 

188

 

 

salam

slice

‘DONQUI#QN—K

1
3.528039
3.534517

3.37541
3.425899
3.220788
3.227914
2.922195

2.520237

3.02111
2.808929
3.171538
3.369162

2
2.572058
2.8927
2.547701
2.376526
2.678518
2.134993
3.617631
2.495822
2.507316
2.872785
2.679064
2.926748
3.302902
2.306245
2.60892

3 Ave

2.771914
2.425906
3.135324
1.886626
1.990836

2.895644
2.874366
3.148997
2.844104
2.576756
2.863085
0.778151
2.643528
2.808621

salami (3 reps)

2.96
2.95
3.02
2.56
2.63
2.68
3.15
2.69
2.83
2.86
2.59
2.94
2.30
2.71
2.93

No

2.895622

fit

sum

2.88
2.87
2.85
2.84
2.83
2.81
2.80
2.79
2.77
2.76
2.74
2.73
2.72
2.70
2.69

 

 

 

 

0.013779

(ave-

fit)"2
0.006
0.007
0.027
0.077
0.039
0.017
0.120
0.010
0.003
0.010
0.023
0.043
0.176
0.000
0.058
0.616

 

 

 

 

 

 

 

I89

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 12

No R
3.010 0.106
turkey
(ave-
slice 1 2 3 Ave fit fit)"2
1 3.11883 3.058046 3.272306 3.15 2.90 0.060
2 3.296054 2.80956 2.686636 2.93 2.80 0.018
3 3.050622 2.38739 2.361728 2.60 2.69 0.008
4 2.525796 2.468347 2.41162 2.47 2.58 0.014
5 1 .870989 2.480007 2.450249 2.27 2.48 0.045
6 2.559907 2.033424 2.30 2.37 0.006
7 2.055799 2.06 2.27 0.044
8 2.08636 2.09 2.16 0.005
9 2.253411 2.247973 2.25 2.05 0.039
10 2.148726 2.15 1.95 0.041
1 1
12 sum 0.280
] turkey '
l i
I 3.5 l
l 3 L
l
i 2.5 .
! 2 '—0— rep 1 l i
l u -I— rep 2 t
r 3 , Ii
. ”a l rep 3 {1
i
l 1 u
i 0.5 — W
i
l
l

190

APPENDIX V

191

Example calculations and modeling of turkey sliced on a delicatessen slicer

(inoculated blade to uninoculated product)

GW Basic Output: turkey (108 CFU/blade)
fraction left on blade during each slice= .7995 544
cfu's transferred to lst slice= 5680164

above results are independent of initial cfu's on blade

if initial cfu's on blade= 1E+08 ,then
fraction transferred to meat during each slice= 5.680164E-03

fraction transferred to surroundings during each slice= .1947655

fitted equations (all equivalent) are:

1) ln cfu(s)= -.2237008 *s + 13.47361
2) cfu(s)= 7104162 * .7995544 "5

3) cfu(s)= 7104162 *e"(-.2237008 *s)
4) cfu(s)= 7104162 *lO"(-.097152 *s)

correlation coefficient for fit is R= .8978895

192

Example calculations of output for turkey (108 CFU/blade)

100*f,
fl+f2

% transfer =

f1=5.680164E-3 = 0.005680164

f2 = 0.1947655

100 * 0.005680164
0.005680164 + 0.1947655

% transfer =

 

% transfer = 2.83

193

GW Basic Output: turkey (10s CFU/blade)

fraction lefi on blade during each slice= .672946
cfu's transferred to lst slice= 1144.353

above results are independent of initial cfu's on blade

if initial cfu's on blade= 100000 ,then
fraction transferred to meat during each slice= 1.144353E-02

fraction transferred to surroundings during each slice= .3156105

fitted equations (all equivalent) are:

1) In cfu(s)= -.3960902 *5 + 7.438685
2) cfu(s)= 1700.512 "‘ .672946 "8

3) cfu(s)= 1700.512 *e"(-.3960902 *s)
4) cfu(s)= 1700.512 *10"(-.1720198 *s)

correlation coefficient for fit is R= .9150966

194

Example calculations of output for turkey (10s CFU/blade)

100*f,
fr+f2

% transfer =

f1= 1.14435E-2 = 0.0114435

f2=0.315610

100*0.0114435
0.0114435+0.315610

 

% transfer =

% transfer = 3.49

195

BIBILIOGRAPHY

196

Ak, N.O., D.O. Cliver, and CW. Kasper, 1994. Cutting boards of plastic and wood
contaminated experimentally with bacteria. J. Food Prot. 57:16-22.

Akier, A.M., R. Denis, G. Jacques, and M. Pierre. 1990. Attachment of Listeria
monocytogenes to stainless steel, glass, polypropylene, and rubber surfaces after
short contact times. J. Food Prot. 53:742-746.

American Iron and Steel Institute (AISI). 2005. http://www.steel.org/mt/. Accessed,
February, 22, 2005.

Arnold, J.W., and G.W. Bailey. 2000. Surface finishes on stainless steel reduce bacterial
attachment and early biofilm formation: scanning electron and atomic force
microscopy study. Poultry Sci. 79:1839-1845.

Arnold, J .W., and S. Silvers. 2000. Comparison of poultry processing equipment surfaces
for susceptibility to bacterial attachment and biofilm formation. Poultry Sci
79:1215-1221.

Association of Official Analytical Chemists. 2003. Meat and Meat Products, p. 2-23. In
AOAC Official Methods of Analysis; 17th ed. Gaithersburg, MD.

Barkley, MB. 1979. Detecting cracks in stainless steel dairy equipment. Aust J. Dairy
Tech. 34:75-77.

Beck, J ., and K. Arnold, 1977. Parameter Estimation. Wiley, New York.

Beresford, M.R., P.W. Andrew, and G. Shama. 2001. Listeria monocytogenes adheres to
many materials found in food-processing environments. J. Appl. Microbiol.
90: 1000- l 005.

Bemaerts, K., E. Dens, K. Vereecken, A.H. Geeraerd, A.R. Standaert, F. Devlieghere, J.
Debevere and J .F . Van Impe. 2004. Concept and tools for predictive modeling of
microbial dynamics. J. Food Prot. 67:2041-2052

Bohner, HF. and R.L. Bradley. 1991. Corrosivity of chlorine dioxide used as a sanitizer
in ultrafiltration systems. J. Dairy Sci. 74:3348-3352.

Bremer, P.J., I Monk, and CM. Osborne. 2001. Survival of Listeria monocytogenes
attached to stainless steel surfaces in the presence or absence of Flavobacterium

spp. J. Food Prot. 64:1369-1376.
Briandet, R., V. Leriche, B. Carpentier, and M. N. Bellon-Fontaine. 1999. Effects of

growth procedure on the surface hydrophobicity of Listeria monocytogenes cells
and their adhesion to stainless steel. J. Food Prot. 62:994-998.

197

Buchanan, R., and J. Phillips. 1990. Response surface model for predicting the effects of
temperature, pH, sodium chloride content, sodium nitrate concentration and
atmosphere on the growth of Listeria monocytogenes. J. Food Prot. 5 3:370-3 76.

Buchanan, R.L., R.C. Whiting, and WC. Damert. 1997. When is simple good enough: a
comparison between Gompertz, Baranyi and three phase linear models for fitting
bacterial grth curves. Food Microbiol. 14:313-326.

CDC. 2000. Multistate outbreak of listeriosis. Morb. Mortal. Wkly. Rep. 49:1129-1130.

Chmielewski, R-A-N, and J. Frank. 2003. Biofilm formation and control in food
processing facilities. Comp. Rev. Food Sci. and Food Safety. 2:22-32.

Chen, Y., K.M. Jackson, F.P. Chea, and D.W. Schaffner. 2000. Quantification and
variability analysis of bacterial cross-contamination rates in common food service
tasks. J. Food Prot. 64:72-80.

Chen, Y., H. Ross, V.N. Scott, and DE. Gombas. 2003. Listeria monocytogenes: Low
levels equal low risk. J. Food Prot. 66:570-577.

Crozier-Dodson, B.A. and D.Y.E. Fung. 2002. Comparison of recovery of airborne
microrganisms in a dairy cattle facility using selective agar and thin agar layer
resuscitation media. J. Food Prot. 65: 1488-1492.

Davies, K.W. 1993. Design of experiments for predictive microbial modeling. J. Ind.
Microbiol. 12:296-300.

Djordjevic, D., M. Wiedmann, and LA. McLandsborough. 2002. Microtiter plate assay
for assessment of Listeria monocytogenes biofilm formation. Appl. Environ.
Microbiol. 68:2950-2958.

Draughon, AF. 2005. Personal communication.
Food and Agriculture Organization of the United Nations / World Health Organization
FAQ/WHO). 2004. Microbiological Risk Assessment. No. 4. Interpretive

summary. http://www.who.int/foodsafety/publications/micro/mra_listeria/en/

Food and Drug Administration Center for Food Safety and Applied Nutrition. 2001 Food
Code. httnzllwwwcfsgnfda.gov/~dms/fc01-4.html Accessed February 22, 2005.

Flores, RA. 2004. Distribution of Escherichia coli 0157:H7 in beef processed in a table-
top bowl cutter. J. Food Prot. 67:246-251.

198

Food Safety Inspection Service. 2003. Draft F SIS risk assessment for Listeria in ready-
to-eat meat and poultry products.
http://www.fsis.usdgov/OPHS/lmrisk/DrafthZ2603 .pdf. Accessed February
22, 2005.

Food Safety Inspection Service. 2003. F SIS strengthens regulations to reduce Listeria
monocytogenes in ready-to-eat meat and poultry products. Available from:
htg)://www.fsis.usdggov/OA/news/2003/lmfinal.htm. Accessed February 22,
2005.

Frye, D.M., R. Zweig, J. Sturgeon, M. Tormey, M. LeCavalier, 1. Lee, L. Lawani, and L.
Mascola. 22002. An outbreak of febrile gastroenteritis associated with
delicatessen meat contaminated with Listeria monocytogenes. Clin. Infect. Dis.
35:943-949.

Gill, 00., M. Badoni, and J .C. McGinnis. 2001. Microbiological sampling of meat cuts .
and manufacturing beef by excision or swabbing. J. Food Prot. 64:325-334.

Glass. K.A. and MP. Doyle. 1989. Fate of Listeria monocytogenes in processed meat
products during refrigerated storage. Appl. Environ. Micrbiol. 55: 1565-1569.

Gombas, D.E., Y. Chen, R.S. Clavero, and V.N. Scott. 2003. Survey of Listeria
monocytogenes in ready-to-eat foods. J. Food Prot. 66:559-569.

Herald, P.J., and B.A. Zottola. 1988. Effect of various agents on the attachment of
Pseudomonas fragi to stainless steel. J. Food Sci. 54:461-464.

Herald, P.J., and B.A. Zottola. 1988. Attachment of Listeria monocytogenes to stainless
steel surfaces at various temperatures and pH values. J. Food Sci. 53:1549-1562.

Healthy People 2010. 2004. United States Department of Health and Human Services.
httm/lwwwhealthypeople.gov/About/goals.htm. Accessed February 22, 2005.

Healthy People 2010. 2005. United States Department of Health and Human Services.
http://www.healthypeople.gov/data/PROGRVW/default.htm#20 l 0.

Hood, SK, and B.A. Zottola. 1997. Growth media and surface conditioning influence the
adherence of Pseudomonas fragi, Salmonella typhimurium, and Listeria
monocytogenes cells to stainless steel. J. Food Prot. 60: 1034-1037.

Houtsma, P., M. Kant-Muermans, F. Rombouts, and M. Zwietering. 1996. Model for the
combined effects of temperature, pH, and sodium lactate on growth rates of
Listeria innocua in broth and bologna-type sausages. Appl. Environ. Microbiol.
62:616-1622.

Howie, J.W., 1968. Typhoid in Aberdeen, 1964. J. Appl. Bact. 31:171-178.

199

Hudson, J.A., and SJ. Mott. 1993. Presence of Listeria monocytogenes, motile
aeromonads and Yersinia enterocolitica in environmental samples taken from a
supermarket delicatessen. Int. J. Food Microbiol. 18:333-337.

Humphrey, T.J. 1990. Listeria contamination of retail meat slicers. PHLS Microbiology
Digest 7:57.

International Commission on Microbiological Specifications For Foods (ICMSF)
Working Group on Microbial Risk Assessment. 1998. Potential application of risk
assessment techniques to microbiological issues related to international trade in
food and food products. J. Food Prot. 61 :1075-1086.

Jeong, D.K., and J .F. Frank. 1994. Growth of Listeria monocytogenes at 10° C in biofilms
with microorganisms isolated from meat and dairy processing environments. J.
Food Prot. 57:576-586.

Kalmokoff, M.L., J.W. Austin, X.-D. Wan, G. Sanders, S. Banerjee, and J.M. Farber.

2001. Adsorption, attachment and biofilm formation among isolates of Listeria .

monocytogenes using model condition. J. Appl. Microb. 91 :725-734.

Keskinen, L.A., E.C.D. Todd, and E.T. Ryser. 2004. Impact of biofilm-forming ability on
transfer of surface-dried Listeria monocytogenes cells from knife blades to
smoked turkey breast. Ann. Mtg. Int. Assoc. Food Prot., Phoenix, AZ.

Keskinen, L.A., E.C.D. Todd, and E.T. Ryser. 2003. Variation in biofilm formation by
Listeria monocytogenes strains at 4°C and 22°C. Ann. Mtg. Inst. Food. Tech.,
Chicago, IL.

Kim, K.Y., and J.F. Frank. 1994. Effect of growth nutrients on attachment of Listeria
monocytogenes to stainless steel. J. Food Prot. 57:720-724

King, M. J. 1999. Knife and impact cutting of lamb bone. Meat Sci. 52:29-38

Krysinski, E.P., L.J. Brown, and T.J. Marchisello. 1992. Effect of cleaners and sanitizers
on Listeria monocytogenes attached to product contact surfaces. J. Food Prot. 55:

246-25 1 .

Leclercq-Pelat, MN. and M. Lalande. 1994. Cleanability in relation to surface chemical
composition and surface finishing of some materials commonly used in food

industries. J. Food Eng. 23:501-517.
Le Marc, Y., V. Huchet, C. Bourgeois, J. Guyonnet, P. Mafart, and D. Thuault, 2002.

modeling the growth kinetics of Listeria as a function of temperature, pH, and
organic acid concentration, Intern. J. Food Microbiol. 73:219-23 7.

200

 

 

 

Levine, P., B. Rose, 3. Green, G. Ransom, and w. Hill. 2001. Pathogen testing of ready-
to-eat meat and poultry products collected at federally inspected establishments in
the United States, 1990 to 1999. J. Food Prot. 64:1188-1193.

Lin, C.M., K. Takeuchi, M.P. Doyle, C.B. Dohm, J.D. Meyer, and RA. Hall. 2004a.
Cross-contamination of deli meats by Listeria monocytogenes on a commercials
slicer. Ann. Mtg. Int. Assoc. Food Prot., Phoenix, AZ.

Lin, C.M., L. Zhang, M.P. Doyle, C.B. Dohm, J.D. Meyer, and RA. Hall. 2004b.
Comparison of polymerase chain reaction (PCR) and USDA culture procedure to
detect Listeria monocytogenes in deli meats. Ann. Mtg. Int. Assoc. Food Prot.,
Phoenix, AZ.

Lunden, J.M., T.J. Autio, and H.J. Korkeala. 2002. Transfer of persistent Listeria
monocytogenes contamination between food-processing plants associated with a
dicing machine. J. Food Prot. 65:1129-1133.

Lunden, J.M., M.K. Miettinen, T.J. Autio, and H.J. Korkeala. 2000. Persistent Listeria
monocytogenes strains show enhanced adherence to food contact surface after
short contact times. J. Food Prot. 63: 1204-1207

Mafu, A.A., D. Roy, J. Goulet, and L. Savoie, 1991. Characterization of physiochemical
forces involved in adhesion of Listeria monocytogenes to surfaces. 1991. Appl.
Environ. Microbiol. 57: 1969-1973.

Mathew F.P.and E.T. Ryser. 2002. Competition of thermally injured Listeria
monocytogenes with a mesophilic lactic acid starter culture in milk for various
heat treatments. J. Food Prot. 65:643-650.

Mead, P.S., L. Slutsker, V. Dietz, L.F. McCraig, S. Bresee, C, Shapiro, P.M. Griffin, and
RV. Tauxe. 1999. Food-related illness and death in the United States. Emerg.

Infect. Dis. 5:607-625.

Michaels, B., T. Ayers, and W. Birbari. 2002. Hygiene issues associated with food
service potholders and oven mitts. Food Service Tech. 2:81-86.

Midelet, G., and B. Carpentier. 2002. Transfer of microorganisms, including Listeria
monocytogenes, from various materials to beef. Appl. Environ. Microbiol.
68:4015-4024.

Miettinen, H., K. Aarinsalo, S. Salo, and A. Sjoberg. 2001. Evaluation of surface

contamination and the presence of Listeria monocytogenes in fish processing
factories. J. Food Prot. 64:635-639.

201

 

MMWR. 1985. Epidemilogic notes and reports listeriosis outbreak outbreak associated
with Mexican styles cheese in California. Morbid. Mortal.
Wk. Rep. 34:357-359.

MMWR. 2000. Multistate outbreak of listeriosis - United States, 2000. Morbid. Mortal.
Wk. Rep. 49:1129-1130.

MMWR. 2002. Public health dispatch: Outbreak of listeriosis — Northeastern United
States, 2002. Morbid. Mortal. Wk. Rep. 51:950-951.

Montville, R., Y. Chen, and D.W. Schaffner. 2001. Glove barriers to bacterial cross-
contamination between hands to food. J. Food Prot. 64:845-849.

Moore, G., and C. Griffith. 2002. Factors influencing recovery of microorganisms from
surfaces by use of traditional hygiene swabbing. Dairy Food Environ. Sanit.
22:410-421.

Murio, D.l993. The Mollification Method and the Numerical Solution of 111 posed
Problems. Wiley-Interscience, New York.

Norwood, DE, and A Gilrnour. 1999. Adherence of Listeria monocytogenes to
stainless steel coupons. J. Appl. Micrbiol. 86:576-582.

Percival, S. L. 1999. The effect of molybdenum on biofilm development. J. Ind.
Micrbol & Biotech 23:112-117

Pinner, R.W., A. Schuchat, B. Swaminathan, P.S. Hayes, K.A. Deaver, R.E. Weaver,
B.D. Plikaytis, M. Reeves, C.V. Broome, J.D. Wenger, and the Listeria study
group. 1992. Role of foods in sporadic listeriosis, II. Microbiologic and
epidemiologic investigations. JAMA. 267:2046-2050.

Richard, J. and C. Piton. 1986. Semi-log model for interpreting the results of swabbing
surfaces naturally contaminated. J. Appl. Bacteriol. 60:243-249.

Rocourt, J. 1999. The genus Listeria and Listeria monocytogenes: Phylogenetic Position,
Taxonomy, and Identification. In Listeria, Listeriosis and Food Safety. E.T. Ryser
and EH. Marth eds. Marcel Dekker, Inc. New York, NY. pp. 1-12.

Salo, S., and A. Laine. 2000.} Validation of the microbiological methods hygicult dipslide,
contact plate, and swabbing in surface hygiene control: A Nordic collaborative
study. J. AOAC Int. 83:1357-1364.

SAS Institute, Inc. 1996. SAS User's Guide: Basic Statistical Analysis, Version 6.12,
North Carolina: SAS Institute Inc.

202

$1..

Sattar, S.A., S. Springthorpe, S. Mani, M. Gallant, R.C. Nair, E. Scott, and J. Kain.2001.
Transfer of bacteria from fabrics to hands and other fabrics: Development and
application of a quantitative method using Staphylococcus aureus as a model. J.
Appl. Microb. 90:962-970.

Sauders, B.D., K. Mangione, C. Vincent, J. Schermerhorn, C. M. Farchione, N. B.
Dumas, D. Bopp, L. Komstein, E. D. Fortes, K. Windham, and M. Wiedmann.
2004. Distribution of Listeria monocyctogenes molecular subtypes among human
and food isolates from New York State shows persistance of human disease-
associated Listeria monocyctogenes strains in retail environments. J. Food Prot.
67:1417-1428.

Schaffner, D.W. 2004. Mathematical frameworks for modeling Listeria cross-
contarnination in food processing-plants. J. Food Sci. 69:R155-R199.

Schaffner, D.W. 1998. Predictive food microbiology Gedanken experiment: why do
microbial growth data require transformation. Food Microbiol. 15:185-189.

Schaffner, D.W., S. Sithole and R. Montville. 2004. Use of Microbial modeling and
Monte Carlo Simulation to determine microbial performance criteria on plastic
cutting boards in use in foodservice kitchens. Food Prot Trends. 24:14-19

Scott, 13., SF. Bloomfield, and CG. Barlow. 1984. A comparison of calcium alginate
swab techniques for quantitative assessment of bacteriological contamination of
environmental surfaces. J. Appl. Bacteriol. 56:317-320.

Schlech, W.F. 2000. Foodbome listeriosis. Clin. Infect. Dis. 31:770-775.

Slutsker L. and A. Shuchat.1999. Listeriosis in humans. In Listeria, Listeriosis and Food
Safety. E.T. Ryser and EH. Marth eds. Marcel Dekker, Inc. New York, NY. pp.
75-90.

Farber, J.M. and RI. Peterkin. 1999. Incidence and Behavior of Listeria monocytogenes
in Meat Products. In Listeria, Listeriosis and Food Safety. E.T. Ryser and EH.
Marth eds. Marcel Dekker, Inc. New York, NY. pp. 505-555.

Smith, W.F. 1993. Structure and properties of engineering alloys, 2nd ed., McGraw-Hill,
New York.

Stone, LS. and B.A. Zottola. 1985. Scanning electron microscopy study of stainless-steel
finishes used in food processing equipment. Food Technol. 39:] 12-114.

Suzuki, 0. 2000. Adsorption and fouling at stainless steel surfaces. Encyclopedia of
Surface and Colloid Science, Marcel Dekker Pub. New York, pp. 167-174.

203

Sveum, W.H., L.J. Moberg, R.A. Rude, and J.F. Frank. 1992. Microbiological
Monitoring of the Food Processing Environment. p. 51-72. In Compendium for
the Microbiological Examination of Foods, 3rd ed. APHA, Washington, DC.

Tamplin, M. 2002. Pathogen Modeling Program Version 6.1, Microbial Food Safety
Research Unit, US. Department of Agriculture,
http://www.gr§enc.grv/mfs/PATHOGEN.HTM. Accessed April 5, 2005.

Tiwari, NP, and Alenrath, S.G. 1990. Occurrence of Listeria species in food and
environmental samples in Alberta. J. Can. Inst. Food Sci. Technol. 23:109-113.

Tominaga, T., Y. Ichiba, Y., K. Miyakita, Y. Nabetani, H. Ueda, H. Suezawa, S.
Satomura, T. Saika, S. Senda, N. Yamaguchi, K. Tani, and M. Nasu. 2001.
Evaluation of microbial population in hospital using adhesive pad. J. Antbact.
Antifung. Agents 29:225-229.

Tompkin, RB. 2001. Control of Listeria monocytogenes in the food-processing
environment. J. Food Prot. 65:709-725.

USDA-FSIS. 2003. Draft FSIS risk assessment for Listeria in RTE meat and poultry
products. http://www.fsis.usdgggov/OPI-IS/lmrisk/DrafiLm22603.pdf Accessed

April 5, 2005.

Uyttcndaele, M., P. de. Troy and J. Debever. 1999. Incidence of Listeria monocytogenes
in different types of meat products on the Belgian retail market. Int. J. Food
Microbiol. 53:75-80.

Vantanyoopaisam, S., A. Nazil, C.E.R. Dodd, C.E.D. Rees and W.M., Waites, 2000.
Effect of flagella on initial attachment of Listeria to stainless steel. Appl. Environ.
Microbiol. 66:860-863.

Vogel, B.F., H.H. Huss, B. Ojenyi, P. Ahrens, and L. Gram. 2001. Elucidation of Listeria
monocytogenes contamination routes in cold-smoked processing plants detected
by DNA-based typing methods. Appl. Environ. Microbiol. 67:25 86-25 95.

Vorst, K., E.C.D. Todd, and E.T. Ryser. 2004. Improved quantitative recovery of
Listeria monocytogenes from solid surfaces. J. Food Prot. 67:2212-2217.

Ware, L.M., M.L. Kain, J .N. Sofos, K.E. Belk, and GO Smith. 1999. Comparison of
sponging and excising as sampling procedures for microbiological analysis of
fresh beef-carcass tissue. J. Food Prot. 62:1255-1259

Yamaguchi, N., A. Ishidoshiro, Y. Yoshida, T. Saika, S. Senda, and M. Nasu. 2003.

Development of an adhesive sheet for direct counting of bacteria on solid
surfaces. J. Microbiol. Meth. 53:405-410.

204

Zhoa, L., Y. Chen, and D.W. Schaffner. 2001. Comparison of logistic regression and
Linear regression in modeling percentage data. Appl. Environ. Microbiol.
67:2129-2135.

205

 

l11111111111I