EPIDEMIOLOGY OF SHIGA TOXIN-PRODUCING E. COLI (STEC) IN THE
FINISHING PIGS AND HUMANS
By
Marion Tseng

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Comparative Medicine and Integrative Biology-Doctor of Philosophy
2014

ABSTRACT
EPIDEMIOLOGY OF THE SHIGA TOXIN-PRODUCING E. COLI (STEC) IN THE
FINISHING PIGS AND HUMANS
By
Marion Tseng
Shiga toxin-producing E. coli (STEC) infections are a significant public health concern.
The epidemiology of STEC in swine remains largely unknown, and the role of swine play in
STEC transmission to humans is not yet elucidated. The objectives of this dissertation are to
provide descriptive epidemiology of STEC shedding in the finishing pigs, characterizing swine
STEC using molecular methods, and to understand the epidemiology of STEC, specifically
STEC in non-O157 serotypes, in human STEC cases in Michigan.
A descriptive longitudinal study was conducted to investigate the fecal shedding of STEC
in the finishing pigs. Three cohorts of finishing swine (n=50/cohort; total 150 pigs) were
included in the longitudinal study. Individual fecal samples were collected every 2 weeks (8
collections/pig) from the beginning (pig age=10 weeks old) to the end (pig age =24 weeks old)
of the finishing period. STEC isolates were recovered in at least one sample from 65.3%
(98/150) of the pigs, and the frequency distribution of first-time STEC detection during the
finishing period resembled an outbreak curve. Nineteen O:H serotypes were identified among the
STEC isolates. Most STEC isolates (n=148) belonged to serotype O59:H21 and carried the stx2e
gene. One O49:H21 STEC isolate carried the stx2e and eae gene. High prevalence rates of STEC
during the finishing period were observed, and STEC isolates in various non-O157 serotypes
were recovered.
To investigate whether there were actual STEC outbreaks within the finishing pigs as
suggested by the descriptive epidemiologic study, a subset of swine O59:H21 STEC strains

(n=29) was analyzed by pulsed field gel electrophoresis (PFGE) to examine their genetic
relatedness. Moreover, the presence and absence of a large panel of virulence genes was
examined in a subset of swine STEC strains (n=155) recovered by a high-throughput real-time
PCR array system. Seventeen different combinations of virulence gene profiles and serotypes
were determined in the swine STEC strains. The majority of the swine O59:H21 STEC strains
(n=120) carried the same virulence gene profile. Genes encoding adhesins were identified, for
example, the iha gene (n=154). The PFGE results revealed that swine STEC strains from pigs
raised in the same finishing barn were closely related, supporting the observations that there
were STEC outbreaks within the finishing barn. This work is the critical first step to understand
the swine STEC epidemiology and potential pathogenic mechanisms of swine STEC in human
disease.
To better understand the demographic and clinical characteristics of STEC cases,
specifically non-O157 STEC, STEC cases reported to the Michigan Department of Community
Health (MDCH) from 2001 through 2012 were described. An increasing trend of non-O157
STEC cases was observed in this 12-year period, and the incidence rates were similar for O157
and non-O157 STEC cases in 2012. No demographic characteristics were significantly different
between O157 and non-O157 STEC cases. However, the odds of hospitalization were 2.36 times
higher in O157 STEC cases than in non-O157 STEC cases when adjusted for age and gender.
The information enhances our understanding in epidemiology of non-O157 STEC in Michigan,
and future research is warranted to understand these pathogens in order to improve prevention
and control efforts.

ACKNOWLEDGEMENTS

I would like to express my gratitude to all of the people who have contributed to this PhD
dissertation research. This dissertation represents the accumulated efforts by all the collaborators
and people who have made contributions. I am forever grateful for all of the assistance and
support throughout my graduate school career, and without these helps I will not be able to
complete this dissertation.
First and the most importantly, I would like to acknowledge my major advisor, Dr. Julie
Funk. It is an honor and it has been a true pleasure to work under Dr. Funk’s guidance, support,
and encouragement. Words cannot describe how grateful I am to have Dr. Funk as my PhD
advisor. Dr. Funk has not only enlightened me as to scientific research, but also inspires me to
become a better person.
I would like to thank my PhD guidance committee: Dr. Paul Bartlett, Dr. Shannon
Manning, and Dr. Melinda Wilkins. It has been very rewarding and pleasant to work with each
one of them, whether it is for scientific research or guidance for my future career goal.
I sincerely thank all the collaborators of my research projects. First of all, I would like to
thank Dr. Pina Fratamico at the USDA ARS ERRC (U.S. Department of Agriculture,
Agricultural Research Services, Eastern Regional Research Center) for all of the efforts
contributing to this research project. Dr. Fratamico has provided priceless inputs and support into
the microbiology component of this research, and it has been a pleasure to collaborate with her. I
also would like to thank other members at the USDA ARS ERRC, Lori Bagi and Danielle
Manzinger, for their efforts and assistance contributing to sample processing and microbiologic
experiments. I would like to acknowledge Dr. Patrick Fach and Dr. Sabine Delannoy at the

iv

ANSES (French Agency for Food, Environmental and Occupational Health and Safety) for their
efforts and collaboration for determining the virulence gene profiles for the bacterial strains.
I sincerely thank all the following people who have provided assistance in part of my
dissertation research: Dr. Chitrita Debroy at E. coli Reference Center, the Pennsylvania State
University for the STEC strains characterization, MSU DCPAH (Diagnostic Center for
Population and Animal Health) Bacteriology, specifically Niesa Kettler and Terrence MacManus
for their assistance in molecular genotyping methods for swine STEC, and the Michigan
Department of Community Health for their assistance in the data collection of human STEC
cases.
I also would like to thank the swine producers who have participated in this research
project. Many thanks to the undergraduate and veterinary students at MSU who helped with the
visits to the swine farms, sample collections, and assisting with the data management: Alexandra
Buckley, Allan Mergener, Tasha Likavek, and Joel Sparks.
I would like to acknowledge my former colleague Dr. Alda Pires for all the assistance
with swine farm visits, sample collections, data management, and statistical analysis. It has been
wonderful to have Dr. Pires as a colleague and friend throughout the graduate school. Many
thanks to Dr. Pires’ support and all the good memories on and off work.
I also would like to thank Dr. Vilma Yuzbasiyan-Gurkan for all the assistance and
support as a graduate program director through my graduate school. I also would like to
acknowledge my friends at MSU for their support, assistance and friendship. Finally, but not the
least, I am grateful for the support from my family and friends. Specifically, I truly appreciate
my parents, who always support and encourage me to pursue my professional career. Most
importantly, I am forever grateful for the support from my husband, who has been always there

v

for me, including in-house scientific discussion involving E. coli. I would like to thank for the
support from the U.S. Department of Agriculture, National Needs Fellowship in Epidemiology
and Infectious Disease, National Pork Board (grant number 12-069), the U.S. Department of
Agriculture, National Institute of Food and Agriculture, Agriculture and Food Research Initiative
(award number 2013-67005-21189), and Michigan State University the Graduate School
Dissertation Completion Fellowship.

vi

TABLE OF CONTENTS

LIST OF TABLES

ix

LIST OF FIGURES

x

KEY TO ABBREVIATIONS

xi

CHAPTER 1. Literature Review: Shiga toxin-producing Escherichia coli in swine
1
ABSTRACT
2
INTRODUCTION
3
FOOD-BORNE OUTBREAKS ASSOCIATED WITH PORK PRODUCTS
5
STEC PREVALENCE IN SWINE POPULATIONS
6
STEC IN SWINE: EDEMA DISEASE
13
HUMAN STRAINS OF STX2E-PRODUCING E. COLI
14
MOLECULAR EPIDEMIOLOGY OF SWINE STEC
17
THE EFFECT OF SWINE STEC STRAINS ON HUMAN INTESTINAL CELLS
21
STEC INFECTION IN SWINE: ANIMAL MODEL STUDIES
22
FUTURE DIRECTIONS
25
APPENDIX
27
REFERENCES
30
CHAPTER 2. Shiga toxin-producing E. coli (STEC) in swine: prevalence over the finishing
period and characteristics of the STEC isolates
43
ABSTRACT
44
INTRODUCTION
45
MATERIAL AND METHODS
47
Study design
47
Sample collection
48
STEC detection and isolation in swine fecal samples
49
Serotyping of the STEC isolates
50
Sample size calculation
51
Variables and statistical methods
51
RESULTS
53
Finishing pigs and sample collection
53
Distribution of STEC positive pigs over the finishing period
53
Characterization of the STEC isolates
54
Estimation of duration of STEC shedding in finishing swine
55
DISCUSSION
56
ACKNOWLEDGEMENTS
60
APPENDIX
61
REFERENCES
68

vii

CHAPTER 3. Molecular characterization of Shiga toxin-producing E. coli (STEC) from
finishing swine in a longitudinal study
ABSTRACT
INTRODUCTION
MATERIAL AND METHODS
Swine STEC strains
Selection of virulence gene targets
High-throughput real-time PCR microarray
Strain selection strategy for pulsed-field gel electrophoresis (PFGE)
PFGE
RESULTS
Virulence gene profiles of swine STEC strains
PFGE
DISCUSSION
ACKNOWLEDGEMENTS
APPENDIX
REFERENCES

73
74
75
78
78
78
79
79
80
81
81
82
84
88
89
104

CHAPTER 4. Shiga toxin-producing E. coli (STEC) cases in Michigan, U.S., 2001-2012
ABSTRACT
INTRODUCTION
MATERIAL AND METHODS
Data source
Study design
Data management and analysis
RESULTS
Descriptive epidemiology of STEC cases
The odds of hospitalization in STEC cases
DISCUSSION
ACKNOWLEDGEMENTS
APPENDIX
REFERENCES

112
113
114
117
117
118
119
112
122
122
126
131
132
149

viii

LIST OF TABLES

Table 1.1. Outbreaks associated with pork products

28

Table 2.1. Sample time periods, demographic information and numbers of fecal samples of
finishing pigs
62
Table 2.2. Distribution of STEC isolates by O:H serotype, Shiga toxin gene subtypes, and eae
63
Table 3.1. Virulence gene targets, their encoded proteins and locations

90

Table 3.2. Distribution of swine STEC strains by serotype and virulence gene profiles

98

Table 4.1. Demographic characteristics of STEC cases by STEC O serotypes, Michigan, 20012012
113
Table 4.2. Clinical outcomes of STEC cases by STEC O serotypes, Michigan, 2001-2012 135
Table 4.3. Univariate analysis of characteristics associated with hospitalization in STEC cases,
Michigan, 2001-2012
137
Table 4.4. Multivariate logistic regression model for characteristics associated with
hospitalization in STEC cases, Michigan, 2001-2012

ix

140

LIST OF FIGURES

Figure 2.1.a. Proportion of pigs from which STEC were isolated by pig age over the finishing
period
65
Figure 2.1.b. Frequency distribution of pigs at the age of first-time STEC isolation

66

Figure 2.2. Kaplan-Meier survival curves for duration of fecal STEC shedding in finishing swine
67
Figure 3.1. PFGE analysis of swine O59:H21 STEC strains. Key represents strain numbers.
Source represents the strain from a pig in which cohort, which time of the eight farm visits, and
the individual pig number
101
Figure 3.2. PFGE analysis of swine O untypeable:H19 STEC strains. Key represents strain
numbers. Source represents the strain from a pig in which cohort, which time of the eight farm
visits, and the individual pig number
102
Figure 3.3. PFGE analysis of swine O98 STEC strains. Key represents strain numbers. Source
represents the strain from a pig in which cohort, which time of the eight farm visits, and the
individual pig number
103
Figure. 4.1.a. Numbers of O157 STEC cases and non-O157 STEC cases, Michigan, 2001-2012
(representing total n=1294; excluding n=203 with no serotype information)
141
Figure 4.1.b. Proportionsa of O157 STEC cases and non-O157 STEC cases, Michigan, 20012012 (representing total n=1294; excluding n=203 with no serotype information)
142
Figure 4.2. Age-adjusted incidence rates of STEC cases per 100,000 population by year,
Michigan, 2001-2012 (representing total n=1294; excluding n=203 with no serotype
information)
143
Figure 4.3.a. Numbers of STEC cases by month of report, Michigan, 2001-2012 (representing
total n=1294; excluding n=203 with no serotype information)
144
Figure 4.3.b. Proportionsa of STEC cases by month of report, Michigan, 2001-2012
(representing total n=1294; excluding n=203 with no serotype information)

145

Figure 4.4.a. Numbers of O157 STEC cases and non-O157 STEC by age groups, Michigan,
2001-2012 (representing total n=1294; excluding n=203 with no serotype information) 146
Figure 4.4.b. Proportionsa of O157 STEC cases and non-O157 STEC by age groups, Michigan,
2001-2012 (representing total n=1294; excluding n=203 with no serotype information) 147
Figure 4.5. Proportion of hospitalized STEC cases by age group and serotype, Michigan, 20012012 (representing total n=1294; excluding n=203 with no serotype information)
148
x

KEY TO ABBREVIATIONS

EHEC Enterohemorrhagic E. coli
HC

Hemorrhagic Colitis

HUS

Hemolytic Uremic Syndrome

MLVA Multilocus Variable Number Tandem Repeat Analysis
PCR

Polymerase Chain Reaction

PFGE Pulsed-Field Gel Electrophoresis
STEC Shiga toxin-producing E. coli

xi

CHAPTER 1
Literature review: Shiga toxin-producing Escherichia coli in swine

Manuscript based on the contents of this chapter has been published in M. Tseng, P. M.
Fratamico, S. D. Manning, J. A. Funk . 2014. Shiga toxin-producing Escherichia coli in swine:
the public health perspective. Animal Health Research Reviews.
doi:10.1017/S1466252313000170. (Published on Animal Health Research Reviews, first view on
January 8th, 2014)

1

ABSTRACT

Shiga toxin-producing Escherichia coli (STEC) strains are food-borne pathogens that are
an important public health concern. STEC infection is associated with severe clinical diseases in
human beings, including hemorrhagic colitis (HC) and hemolytic uremic syndrome (HUS),
which can lead to kidney failure and death. Cattle are the most important STEC reservoir.
However, a number of STEC outbreaks and HUS cases have been attributed to pork products. In
swine, STEC strains are known to be associated with edema disease. Nevertheless, the
relationship between STEC of swine origin and human illness has yet to be determined. This
review critically summarizes epidemiologic and biological studies of swine STEC. Several
epidemiologic studies conducted in multiple regions of the world have demonstrated that
domestic swine can carry and shed STEC. Moreover, animal studies have demonstrated that
swine are susceptible to STEC O157:H7 infection and can shed the bacterium for two months. A
limited number of molecular epidemiologic studies, however, have provided conflicting evidence
regarding the relationship between swine STEC and human illness. The role that swine play in
STEC transmission to people and the contribution to human disease frequency requires further
evaluation.

2

INTRODUCTION

Infection with Shiga toxin-producing Escherichia coli (STEC) is a critical public health
concern. STEC infections are associated with outbreaks and sporadic cases of diarrhea and
severe clinical diseases in human beings, including hemorrhagic colitis (HC) and hemolytic
uremic syndrome (HUS) (1, 2). HUS is a life-threatening thrombotic disorder characterized by
acute renal failure, thrombocytopenia and microangiopathic hemolytic anemia (1, 3, 4). It is one
of the leading causes of acute renal failure in young children worldwide (4-6). Every year in the
U.S., more than 170,000 human illnesses are attributed to STEC (7), and food-borne STEC
illnesses represent an estimated economic burden of 280 million dollars (8).
STEC represent a subset of E. coli that produce a cytotoxin known as the Shiga toxin
(Stx), or verotoxin. Various STEC transmission routes have been identified, though STEC
infections are most frequently associated with consumption of contaminated food (meat, dairy
products, produce, and others) and water (9-11). Other sources of infection include animal
contact (12-14), person-to-person contact (15-17), and airborne transmission (18). Cattle have
been suggested to be the important animal reservoir of STEC, and they do not typically present
with any STEC-associated clinical symptoms (9, 19, 20), except in calves under experimental
conditions (21). Nevertheless, many other food products, including pork products, have been
confirmed as vehicles of STEC transmission (22-28). Feral swine, for example, were found to be
a risk factor for STEC O157:H7 contamination of the spinach implicated in a 2006 North
American outbreak (29). In pork-associated outbreaks and cases, it is unknown whether the
contamination of pork products occurs during swine processing or via cross-contamination from
other foodstuffs (25, 27, 28, 30). Although swine-associated outbreaks have been reported less

3

frequently than cattle-associated outbreaks, the likelihood that swine represent an important
source of STEC infections in human beings cannot be overlooked.
The role that swine play in STEC transmission to human beings needs to be further
evaluated to determine whether swine-derived STEC strains are an important public health
concern. The severe symptoms associated with STEC infections and the increasing frequency of
infections caused by a large variety of STEC serotypes highlight the fact that more research is
needed to better understand these pathogens to aid in developing prevention and control
strategies. In this literature review, previous studies focusing on STEC of swine origin will be
critically summarized.

4

FOOD-BORNE OUTBREAKS ASSOCIATED WITH PORK PRODUCTS

Though only in a few instances, pork has been reported as a potential vehicle involved in
outbreaks of STEC infections (22-28). Table 1.1 summarizes the food products involved in these
outbreaks and the serotype and virulence gene characteristics of the STEC strains isolated in
these outbreaks. Mixed meat products were involved in most of these reported outbreaks. Thus,
it was difficult to attribute the source of contamination to any of the animal species. Notably, the
most recent outbreaks were associated with salami containing only pork (27) and large cuts of
pork from a whole roasted pig (28). It is impossible to eliminate the possibility that crosscontamination from different foodstuffs took place during processing. However, pigs were
suggested to be the source of STEC O157:H7 infection in the most recent outbreak (28). Though
the responsible source of STEC was not conclusive in these outbreaks, pork products cannot be
excluded as a potential vehicle of STEC transmission.

5

STEC PREVALENCE IN SWINE POPULATIONS

Although outbreaks have been rarely reported, the incidence of human STEC outbreaks
associated with pork products has raised questions about the sources of STEC in the food chain,
from pork products, pigs at slaughter facilities, to the most upstream origin, on-farm swine. The
presence of STEC in pork or pork-associated products has been reported in some epidemiologic
studies (31-37). These data were sparse, and the numbers related to STEC prevalence in pork or
pork-associated products were not consistent. The discrepancies are potentially attributable to
multiple factors. For example, different isolation protocols and different types of pork products
were used in these studies. In short, these positive findings of STEC in pork and pork-associated
products indicate that STEC are present on retail pork products. The possibility that swine play a
role in the contamination of pork products cannot be excluded.
The prevalence of STEC in clinically healthy swine populations has been reported in
numerous studies in multiple regions of the world. In the U.S, STEC detection in the swine
population has been reported. A study was published in which colon samples were obtained
from pigs at a U.S. slaughter facility to detect the presence of STEC O157 (38). Six (1.9%) of
the 305 colon samples were positive for STEC O157:H7. Interestingly, no STEC O157:H7
isolate was recovered in the National Animal Health Monitoring System’s (NAHMS) Swine
1995 study (39) and the later NAHMS 2000 study (38). However, it is impossible to directly
compare the number of STEC O157:H7 positive samples in these studies due to different study
designs and sample collection methods. In short, these reports provided evidence that swine in
the U.S. carry STEC serotype O157:H7, although at a much lower rate compared to cattle.

6

Besides STEC O157:H7, the NAHMS 2000 study also recovered and characterized nonO157 STEC isolates (40, 41). At least one STEC isolate was recovered from 196 (28.5%) of 687
fecal samples. A total of 219 STEC isolates were recovered, and they were categorized into
various non-O157 O serogroups. Most importantly, some O serogroups had previously been
associated with human clinical cases, namely O9, O20, O91, O101, and O121. The results of
this study indicated that clinically healthy pigs from multiple states in the U.S. shed non-O157
STEC at some point during the late finishing period.
A recent study reported no significant effect of antimicrobials (chlorotetracycline and
bacitracin) in feed on fecal shedding of non-O157 STEC in finishing swine (42). Notably, STEC
belonging to important O serogroups, namely O26, O103 and O145, were isolated from 6.9, 2.4,
and 4.8% of all the swine rectal swab samples. However, this study targeted a selected group of
non-O157 STEC for detection in the samples (O26, O103, O111, O121, O145). In spite of that,
the high prevalence of the Shiga toxin gene (stx) in enriched samples (58% prior to slaughter)
and the detection of non-O157 STEC belonging to important O serogroups highlight the need to
further investigate STEC shed by clinically-healthy swine.
Besides the U.S., studies in other regions of the world reported wide ranging estimates of
STEC prevalence from 0% (43, 44) to 68.3% (45). The highest prevalence, 68.3%, was reported
from one study in Chile (45). Rectal swab samples were collected during the evisceration
process from swine and cattle at slaughter facilities. Surprisingly, the STEC prevalence in swine
samples was higher than that in cattle (28.7%). In addition to the strains identified as nontypable by O serogrouping, O157 was the most prevalent O serogroup among the swine STEC
isolates. Other important STEC serogroups, such as O26 and O111, have also been reported
(46). In one state in Brazil, the prevalence of STEC was 1.4% (one STEC-positive sample) from

7

74 fecal samples obtained aseptically from swine intestines at slaughter facilities (47). A later
study performed in another state in Brazil reported one (0.4%) STEC O103 isolate recovered
from 215 sponge-rubbed samples from swine carcasses at three slaughter facilities (48).
In a South African study, E. coli O157:H7 was more frequently found in swine fecal
samples and in pork (67.7%) than in cattle fecal samples and beef (27.7%) (36). In this study,
only a selected subset of E. coli O157:H7 isolates was analyzed for the presence of stx.
Therefore, it is difficult to determine the number of STEC O157:H7 positive samples in this
study. Nevertheless, this study was in agreement with the previous Chilean studies (45, 46)
showing that swine are able to shed STEC at a relatively high rate and are potentially important
STEC reservoirs in some places in the world.
Several European studies have reported STEC prevalence in swine populations, and most
of the earlier studies focused on STEC serogroup O157. A study conducted in the United
Kingdom reported the recovery of six (0.3%) STEC O157 isolates from over 2000 swine fecal
samples at slaughter facilities (49). Similarly, in Ireland, the prevalence of STEC O157:H7 was
0.6% (3 of 480) in fecal samples of pigs at slaughter facilities (50). In a study conducted in the
Netherlands, STEC O157 was isolated from 1 (0.7%) of 145 fecal samples from pigs at one
slaughter facility (51). A low proportion (0.7%, 1 of 150) of STEC O157-positive fecal samples
was reported from pigs at slaughter facilities in Northern Italy (52). In Norway, STEC O157:H7
was found in 2 (0.1%) samples of intestinal contents from 1976 pigs, and one of the two pigs was
housed on a farm that also reared cattle. However, no STEC O157:H7 isolate was recovered
from samples of cattle on the same farm (53). Interestingly, other studies also reported isolation
of STEC O157:H7 strains from pigs that were kept with cattle or other ruminants on the same
farm. For instance, STEC O157:H7 was recovered from 2 (0.08%) of 2446 fecal samples of

8

slaughtered pigs that were housed with ruminants on the same farm in a Swedish study (54). In
general, the prevalence of STEC O157:H7 in swine appeared to be low in these studies.
However, the presence of ruminants on the same farm that reared swine suggests that horizontal
transmission of STEC may occur between different animal species.
Several studies investigated the prevalence of both STEC O157 and non-O157 in
clinically healthy swine populations. A study in Belgium reported that non-O157 STEC strains
were isolated from 8 (7.9%) out of 101 swine colon fecal samples at a slaughter facility (55). In
Switzerland, the stx gene was detected in 138 (22%) of the 630 swine fecal samples at a
slaughter facility. Subsequently, forty-five non-O157 STEC isolates were recovered from 45
randomly-selected stx-gene-positive samples (56). It was difficult to estimate the true prevalence
of STEC in this study because not every stx-positive sample was analyzed for STEC isolates.
These reports suggest that swine may be potential reservoirs of non-O157 STEC. In these
studies, relatively low numbers of STEC isolates were recovered even though the numbers of
stx-positive samples were high. This fact reflects the challenges of non-O157 STEC isolation.
In contrast to the above positive findings, one study conducted in Northern Spain analyzed
pooled rectal fecal samples from a total of 510 pigs, and the test results were negative for both
O157 and non-O157 STEC (43). Negative isolation of STEC in 106 swine fecal samples was
also reported by a study conducted in Central Greece (44).
The presence of STEC in swine populations has also been documented in Asia. In Japan,
the prevalence of STEC O157:H7 in clinically healthy on-farm swine was 1.4%, and this number
was similar to the prevalence in Japanese cattle around the same time period (57). A later
Japanese national surveillance report stated that STEC isolates were recovered from 32 (14%) of
179 swine fecal samples. Some strains belonged to the serotypes frequently associated with

9

human diseases (58). In Korea, only reports of STEC O157:H7 in swine have been published.
One (0.3%) STEC O157:H7 isolate was recovered from 345 fecal samples from pigs at slaughter
facilities and on farms (33). Two STEC O157:H7 isolates were recovered from fecal samples of
pigs at slaughter facilities in another study (59). However, the prevalence was unknown because
the total number of samples from pigs was not provided. In a recent report from Korea, one
(0.3%) STEC O157 isolate was recovered from 291 swine fecal samples (60). In China, one
study reported STEC isolates recovered from 10 (2.1%) of 487 rectal swabs from pigs at a
slaughter facility in Hong Kong. The pigs were from a number of provinces in mainland China
(61). Another study reported the prevalence of STEC in one swine breeding farm in mainland
China. STEC O157 was detected in 8 (1.1%) of 720 fecal samples, while non-O157 STEC was
isolated from 33 (4.6%) samples (62). STEC strains were isolated from 10 (6%) of 169 swine
fecal samples obtained from a number of swine farms in one province in Vietnam (63).
In addition to domestic swine, feral swine and wild boars have been suggested to play a
role in STEC transmission. The most prominent evidence was obtained from a study that took
place in the U.S. in 2006. Feral swine were suggested to be involved in the contamination of
spinach, which was associated with a nationwide outbreak caused by STEC O157:H7. The
outbreak strain was isolated from feral swine feces, which were collected directly on the spinach
fields (29). Moreover, reports in other countries also indicated the presence of STEC in wild
boars. STEC O157:H7 was found in 1 (1.4%) of 68 wild boars in Sweden (64). One Spanish
study investigated the presence of STEC O157:H7 and non-O157 STEC in wild boars. STEC
O157:H7 was detected in 7 (3.3%) and non-O157 STEC was detected in 11 (5.2%) fecal samples
from 212 wild boars (65).

10

Human-animal contact is an important risk factor for STEC transmission, thus the
presence of STEC in swine in zoos and petting zoos has also been investigated. No STEC O157
was detected in 45 pig fecal samples collected at U.S. institutions accredited by the Association
of Zoos and Aquariums (14). In contrast, STEC O157:H7 was recovered in 13 (1.2%) of 1102
fecal samples from pigs in 19 county fairs in two states in the U.S (13). In the United Kingdom,
an epidemiologic study was conducted to detect the presence of STEC O157 in “open farms” in
which there is close contact with animals to attract the general public (66). STEC O157 was
isolated in thirty (17.9%) of 168 swine fecal samples. However, this number may not reflect the
true prevalence at one time point as it represents a collated number from samples collected from
22 different farms within a ten-year period. Moreover, it was not clear whether the pigs were
housed together or in close contact with other animal species. Though the prevalence of STEC
O157:H7 in these reports was low, pigs at facilities where direct contact with human beings may
occur are another potential risk factor of STEC transmission.
In general, these data indicated that STEC strains in swine are geographically widely
distributed given the fact that reports were from many regions in different continents. Although
the prevalence numbers are wide ranging, these data provide epidemiologic evidence that STEC
strains are prevalent in swine. Multiple factors may account for the discrepancies in the data
among these studies, such as differences in farm management systems, sample sources, sampling
methods, isolation protocols, and diagnostic tests used.
A number of reports have described detection of STEC O157 in swine fecal samples, and
various non-O157 STEC serotypes were also detected in swine (41, 55, 56). STEC belonging to
various serogroups of which some that have been associated with human illness were identified
in swine (41, 56), and, surprisingly, none of the serogroups were those commonly associated

11

with swine edema disease. In addition, most of the above epidemiologic studies were crosssectional studies, and they are unable to determine if the prevalence of STEC in swine changes
with time (67). More research needs to be done to understand STEC shed by clinically healthy
swine.

12

STEC IN SWINE: EDEMA DISEASE

Besides isolation from clinically healthy swine, STEC can cause edema disease, which
affects post-weaning pigs and young finishing pigs. Some STEC strains, in particular edema
disease-associated STEC, also play a role in fatal shock, that occurs in pre- and post-weaning
pigs (68). STEC strains belonging to certain serogroups, including O138, O139, O141, and
O147 are more frequently associated with edema disease (69). These STEC strains colonize in
the small intestine and typically produce the variant Shiga toxin 2e (Stx2e). The toxin enters the
bloodstream and binds to the specific receptor, globotetraosylceramide (Gb4), which is located
on epithelial and endothelial cells (70). The toxin then impairs blood vessels, leading to edema,
ataxia, and death (68). In general, the clinical presentations of swine edema disease are
somewhat different from those associated with human STEC diseases.
Besides Stx2e, STEC strains associated with edema disease have been found to possess
virulence factors different from those found in STEC strains isolated from human clinical cases.
For example, F18 fimbriae are absent in most human-derived STEC (71), but are essential for
adherence to swine epithelial cells (72), and therefore are likely to play a role in specificity and
adaptation to the host. There are risk factors, including dietary changes and the introduction of
pigs to new herds, suggested to be important for edema disease onset (68).

13

HUMAN STRAINS OF STX2E-PRODUCING E. COLI

Interestingly, in addition to their isolation from pigs, Stx2e-producing E. coli strains were
recovered from samples of wastewater treatment plants in France (73), and from meat (pork,
beef, wildlife) and milk in Germany (74). Notably, though rarely documented, Stx2e-producing
E. coli strains have been isolated from human patients with HUS (75) and uncomplicated
diarrhea (71, 76-79). Moreover, Stx2e-producing E. coli strains have been recovered from stool
samples of asymptomatic human beings (71, 78). The etiologic role of Stx2e-producing E. coli
strains in these human cases has not been determined. No particular source of infection has been
identified in these human cases associated with Stx2e-producing E. coli.
A study was conducted to analyze Stx2e-producing E. coli strains from asymptomatic
people, people with uncomplicated diarrhea, and diseased pigs to compare their virulence gene
profiles and adherence to human and swine intestinal epithelial cells (71). Virulence genes
commonly found in STEC strains associated with HC and HUS, such as the gene, eae, that
encodes for intimin, an important attachment protein, were not detected in both the human- and
swine-derived Stx2e-producing E. coli strains. This fact may suggest that unknown virulence
factors are involved in the pathogenicity of human Stx2e-producing E. coli strains. Additionally,
the human-derived Stx2e-producing E. coli strains adhered to human epithelial cells but not
swine epithelial cells. In contrast, swine-derived Stx2e-producing E. coli strains lysed human
epithelial cells and adhered to swine epithelial cells. This study only analyzed Stx2e-producing
E. coli strains from diseased pigs, and thus, Stx2e-producing E. coli strains from healthy pigs
need further examination.

14

Beutin and colleagues, however, have characterized Stx2e-producing E. coli from people
with uncomplicated diarrhea, people with no clinical symptoms, diseased pigs, and healthy pigs
at slaughter facilities to determine their serotypes, distribution of virulence genes, and Stx2e
production (80). They found that Stx2e-producing E. coli strains from different sources were
heterogeneous with regard to serotypes and virulence genes. In agreement with the results of
Sonntag et al. (71), virulence genes commonly detected in STEC strains associated with HC and
HUS, such as eae and ehxA, the latter of which encodes an enterohemolysin, were not detected
by PCR in both the human- and swine-derived Stx2e-producing strains analyzed in this study.
Additionally, the authors used an enzyme immunoassay to find that human-derived strains
produced significantly higher amounts of Stx2e than strains from diseased pigs. Interestingly,
the toxin-production level did not significantly differ between strains from human subjects with
uncomplicated diarrhea and those with no clinical symptoms. However, different environments
(in hosts or under experimental conditions) may contribute to the differences of toxin production.
Another likely explanation is that there are other unidentified virulence factors in the pathogenic
Stx2e-producing E. coli strains.
In all of these reports, the primary conclusion was that there was a lack of evidence to
suggest that Stx2e-producing E. coli strains are a critical public health concern. Nevertheless,
these results also suggest that unknown virulence factors may contribute to the pathogenesis of
Stx2e-producing E. coli strains in human hosts. Additionally, Stx2e-producing E. coli strains
have been mostly isolated from human patients with uncomplicated diarrhea. Patients with
uncomplicated diarrhea may not seek medical attention, which may contribute to the low
reporting frequencies of Stx2e-producing E. coli infections in human patients. Some commercial
serological tests do not detect the toxin Stx2e (81). The limitation of diagnostic tests may

15

contribute to missed detection of Stx2e-producing E. coli infections. On the other hand, the
picture remains unclear regarding the source of Stx2e-producing E. coli in human infections.
More research is warranted to address whether there is an association between human Stx2eproducing E. coli strains and strains from pork, pigs, or the associated swine environment.

16

MOLECULAR EPIDEMIOLOGY OF SWINE STEC

The presence of virulence genes other than Shiga toxin genes (stx) also plays a role in the
capability of STEC strains to cause disease (82). Therefore, examining the presence or absence
of the selected virulence genes (virulence gene profile) is essential in assessing the public health
risk of STEC strains (83, 84). A considerable challenge for understanding virulence gene
profiles of swine STEC is that the profile of virulence genes targeted varies in each report. The
most common virulence genes evaluated are stx1, stx2, eae, and ehxA. Different stx1, stx2, eae,
ehxA gene combinations have been detected in swine STEC strains (45, 85). In the NAHMS
2000 study in the U.S., although a majority of swine STEC strains carried the stx2e gene, 6% of
the strains carried other stx2 variants. The eae gene was not detected in any of the swine isolates
in this study (41), and this was in agreement with other studies (58, 61). In contrast, many
studies have identified the eae gene in STEC strains from clinically healthy swine (13, 33, 36,
45, 50-53, 55-57, 59, 85). Meanwhile, detection of the ehxA gene in swine STEC isolates was
reported in numerous studies (36, 41, 50, 51, 55, 56, 59, 65, 86). A limited number of studies
described testing for different subsets of STEC virulence genes in swine STEC strains (41, 50,
86, 87). For instance, some eae-negative swine STEC strains carried the saa gene (STEC
autoagglutinating adhesion), and other adhesins (86). The above studies suggest that swine
STEC strains carry various combinations of virulence genes. A subset of swine STEC strains
does carry important virulence genes in human pathogenesis, such as eae. More extensive
examination is needed because the presence or absence of many other STEC virulence genes has
not been examined in swine STEC strains.

17

Molecular genotyping methods, such as pulsed-field gel electrophoresis (PFGE), have
been frequently utilized to evaluate the pathogenic potential of STEC strains and provide
suggestions for future strategies in control and prevention of STEC transmission (88, 89). For
example, these methods are used to identify the source of STEC during food-borne outbreaks
and to determine the genetic relatedness of STEC strains (90-94). A limited number of studies
have incorporated a small number of swine STEC strains to examine their genetic relatedness
with strains from human subjects with diarrhea, HC, HUS, or without clinical symptoms (46, 53,
95, 96). Most of these studies focused on STEC O157:H7 strains, and only one study
investigated STEC O101 strains (95). The results from these studies were inconsistent. One
Chilean study indicated that the swine STEC O157 strains were categorized into the same cluster
with strains from human HUS cases in the same geographic area by analyzing the PFGE
patterns. However, no further epidemiologic relationship could be drawn between the swine
isolates and human cases (46). One study conducted in Norway reported similarities of PFGE
patterns among swine STEC O157:H7 strains and those isolated from human cases. However,
there was no other epidemiologic information provided about the human STEC cases (53). In
contrast, in an Irish study the investigators stated that swine STEC O157:H7 strains were not
genetically related to human strains at the PFGE pattern similarity criterion of 80% (96).
One study implemented repetitive element sequence-based PCR (rep-PCR) to examine
the genetic relatedness of STEC O101 strains from swine and a human case of diarrhea.
Interestingly, the result of the analysis indicated that swine STEC O101 strains showed a high
degree of relatedness with the human STEC O101 strain. Moreover, the stx2e gene sequences
from STEC O101 strains of two different origins shared a high degree of homology, and none of
the STEC O101 strains carried the eae or ehxA genes (95). A direct relationship cannot be

18

ascertained between the swine and human STEC since the strains were from two different
countries and isolated in different time periods. However, the molecular evidence highlights that
swine may be a potential STEC reservoir and a source of human infections.
A number of studies have examined the genetic relatedness of swine STEC strains to
strains isolated from other animal reservoirs. The PFGE patterns indicated that swine STEC
strains were not closely related to STEC strains of bovine origin (46, 96), ovine origin (96) and
from turkeys (51). A Korean study demonstrated that swine STEC O157:H7 strains had distinct
randomly amplified polymorphic DNA (RAPD) patterns compared with strains from cattle at a
similarity criterion of 63% (59). Another study showed that swine STEC O157:H7 strains
demonstrated some similarity in PFGE patterns to strains from cattle (53). A Polish study
reported that one swine STEC O157 strain clustered with STEC O157 strains from cattle at a
similarity criterion of 80% based on PFGE patterns. The swine STEC O157 strain was from a
weaned pig with diarrhea, and no other epidemiologic relationship was described between the
swine and cattle STEC strains (97).
In the epidemiologic investigation following a spinach-associated outbreak of STEC
O157:H7 in 2006 in the U.S., PFGE and multilocus variable number tandem repeat analysis
(MLVA) were used to analyze STEC isolates from different sources (29). Indistinguishable
PFGE patterns from STEC isolates from cattle, feral swine, surface water, and soil were
identified. MLVA also categorized the STEC isolates with indistinguishable PFGE patterns
together with the outbreak STEC strain as one cluster. This evidence led the authors to suggest
that feral swine may serve as a risk factor for STEC contamination of the spinach fields (29).
Inspired by this outbreak, a Spanish study used PFGE to examine the genetic relatedness of
STEC isolates from wild boar and a human patient with diarrhea (65). Surprisingly, a STEC

19

isolate from a wild boar shared an indistinguishable PFGE pattern with a STEC isolate from a
human patient with diarrhea in the same geographic area. Though there were no epidemiologic
data available to establish a relationship between the human case and wild boar, the molecular
genotyping results indicated that wild boars are a potential source of STEC contamination (65).
Results presented by these molecular epidemiologic studies are varied, which is
potentially due to multiple factors. For example, there is potential selection bias of the isolates
included in the genotypic analysis. Limited numbers of swine STEC strains and strains from
other origins were included in the above molecular genotypic studies. The analytical results may
not reflect the complete picture of genetic relatedness between STEC strains from swine and
other origins. Moreover, most of the studies focused on STEC O157:H7. There is a paucity of
knowledge on the genetic relatedness of non-O157 STEC from different origins. Conclusively,
the genetic relationship of swine STEC strains with human clinical strains is still poorly
understood. Filling these current knowledge gaps is relevant to not only better understand swine
STEC but also to assess the potential public health risk of swine STEC.

20

THE EFFECT OF SWINE STEC STRAINS ON HUMAN INTESTINAL CELLS

Kennedy et al. (98) infected human Caco-2 cells, which is a human colonic epithelial cell
line, with STEC O157:H7 strains isolated from human patients or from healthy pigs at slaughter
facilities. The human and swine STEC strains possessed the same virulence gene profiles,
including eae and other virulence genes, published in an earlier study (96). Notably, swine
STEC O157:H7 strains induced greater loss of monolayer cell integrity than human-source
STEC. Moreover, microarray experiments that examined RNA levels transcribed from different
STEC genes indicated that swine and human STEC strains had different gene expression
profiles. When exposed to swine STEC strains, expression levels of some genes involved in
cytoskeleton rearrangement were up-regulated in the human Caco-2 cells (98). However, more
research is needed to further examine the effect of infection with swine STEC strains on human
epithelial cells.

21

STEC INFECTION IN SWINE: ANIMAL MODEL STUDIES

Swine have been used as an animal model in studies examining STEC colonization (99),
virulence of different STEC strains (100, 101), and treatment of STEC infections (102).
Interestingly, some studies evaluated the pathogenicity and colonization of intimin-negative
STEC strains in swine models. One study found that intimin-negative non-O157 STEC strains
colonize and cause similar intestinal lesions and systemic diseases as intimin-positive STEC
O157:H7 strains in cesarean-derived colostrum-deprived (CDCD) neonatal pigs (103).
Moreover, a later study reported that there were no significant differences in bacterial counts and
the numbers of tissues infected between intimin-positive and intimin-negative STEC O157:H7
strains in 12-week-old adult pigs (104). Both intimin-positive and intimin-negative experimental
strains were recovered in the pigs 38 days after inoculation. Notably, the result of this study in
pigs is different from a previous study indicating that intimin aids the persistence of STEC
O157:H7 strains in adult cattle and sheep (105). This may suggest that other adhesins contribute
to the colonization of STEC in swine. The above evidence demonstrates that swine are
biologically capable of carrying STEC, including intimin-positive STEC strains.
To further understand if swine have the potential to become reservoirs of STEC,
researchers have pursued experimental challenge models. Booher and colleagues inoculated
three-month-old pigs with mixtures of STEC O157:H7 and other pathogenic E. coli strains to
observe the level and duration of fecal shedding (106). They reported that STEC O157:H7
strains were isolated from swine fecal samples up to two months after inoculation, and the
experimental strains were recovered in swine intestinal tissues. Shedding of experimental STEC
strains was dose-dependent, and strains inoculated at lower levels (107 CFU/strain/animal) were

22

not recovered in the feces two months after inoculation. The dose-dependent shedding is similar
to what was found in cattle under experimental conditions (107). Interestingly, one of the
experimental strains inoculated at the lower dose was recovered in swine cecum tissues at
necropsy two months after inoculation (106). This study suggested that swine could shed STEC
O157:H7 for at least two months.
Cornick and Helgerson conducted an experimental study to observe the transmission of
STEC O157:H7 among pigs (108). The three-month-old pigs were inoculated with a STEC
O157:H7 strain that originated from a human outbreak, via addition of the organisms added to
their feed. Moreover, they were interested in examining the effect of two different housing
conditions on STEC O157:H7 shedding level and duration. In this housing experiment, a STEC
strain of bovine origin was inoculated in conjunction with the STEC strain of human origin. One
group of pigs was housed individually on decks with feed provided from a trough, while the
other groups of pigs were housed two per pen with feed given on the cement floors.
Transmission of STEC O157:H7 was observed from infected donor pigs to naïve pigs, which
were housed in the same pen. Fecal shedding of STEC O157:H7 was observed in exposed naïve
pigs for at least two weeks after exposure. In addition, the level and duration of STEC O157:H7
shedding was not significantly affected by different housing conditions.

The above experimental study (108) supported a previous study (106), suggesting that
shedding of STEC is dose-dependent and that swine are biologically competent hosts for STEC
O157:H7. Additionally, results of this study (108) indicated that STEC O157:H7 infection may
be maintained in the swine population and that transmission readily occurs between animals. In
another subsequent study by the same research group, STEC O157:H7 was recovered not only

23

from pigs housed close to the infected donor in the adjacent pen (nose-to-nose contact), but also
from naïve pigs housed in pens apart from the infected donor (109). The isolation of the
experimental bacterial strain in the air samples provided further evidence to support that
contaminated aerosols are a means of indirect STEC O157:H7 transmission among pigs.
Antibiotic-free feed was provided in the above animal model studies. However, in most
conventional swine production systems, at least in the U.S., feed with antimicrobials is
commonly used. Thus, the effect of antimicrobials in feed on STEC O157:H7 shedding was also
examined (110). A number of frequently used antimicrobials for growth promotion in U.S.
swine production were selected for this study, namely bacitracin methylene disalicylate,
chlortetracycline, and tylosin. Three groups of young adult pigs were given feed with each of the
three selected antimicrobials, and another group of control pigs was fed with antimicrobial-free
feed. A STEC O157:H7 strain from a human outbreak in Washington State in the U.S. (111)
was orally inoculated at a dose of 107 CFU/strain/animal. One month after inoculation, the
numbers of pigs that shed STEC O157:H7 and levels of STEC O157:H7 shedding were
significantly lower in two groups given feed with tylosin and chlortetracycline compared to the
control group. However, no significant differences were found in levels and numbers of pigs
that shed STEC O157:H7 between the bacitracin and control groups. Interestingly, a more recent
study by Wells et al. (42) reported the numbers of pigs shedding STEC O26, O103, and O145
did not significantly differ between antimicrobial and control groups. More research is needed to
evaluate the effect of sub-therapeutic doses of antimicrobials in feed on STEC shedding.

24

FUTURE DIRECTIONS

Swine are not viewed as an important STEC reservoir given the rare known incidence of
cases of severe human illness associated with STEC of swine origin. Nevertheless, the
association between swine STEC and human illnesses needs to be further investigated. First of
all, pork products and feral swine have been associated with a number of STEC outbreaks. STEC
causes edema disease in swine. However, STEC can be also isolated from clinically healthy pigs
at rates occasionally higher than those for cattle. In addition, animal model studies suggested
that swine are biologically competent for colonization and for shedding of STEC for at least two
months after inoculation.
The increasing surveillance of non-O157 STEC infections has raised the awareness of
non-O157 STEC (20, 112-115). In the U. S., six non-O157 STEC serogroups, namely O26,
O103, O111, O121, O45, O145, were recently classified as adulterants, similar to O157, in raw,
non-intact beef product (116). Non-O157 serogroups, namely O26, O103, O145, O111, and
O91, were also identified as major public health concerns in Europe (117). Some of the nonO157 STEC strains were highly virulent and carried virulence gene profiles which were rarely
reported in previous studies. For example, the STEC O104:H4 strain involved in a large-scale
outbreak in 2011 in Germany was an enteroaggregative E. coli that carried genes typical of this
pathogenic E. coli group, but that also carried the stx2 gene (118). Most STEC strains detected
in clinically healthy swine were non-O157 STEC (41). However, there is still very limited
information about swine STEC.
Based on the increasing role of non-O157 STEC in human illnesses, the high prevalence
and limited epidemiologic investigation of STEC in swine, we recommend further research to

25

elucidate the unclear association between swine STEC and human illness. This will serve public
health through determining whether swine are an important reservoir for STEC infection in
people. If the evidence indicates swine are an important reservoir, efforts can be directed at
control of this pathogen in this species, and, if swine ultimately are determined to not be an
important reservoir, public health resources can be directed away from swine and towards further
understanding of the epidemiology of STEC, particularly non-O157 serotypes.

26

APPENDIX

27

Table 1.1. Outbreaks associated with pork products
Outbreak location

STEC

and year

serotype

Washington and

O157:H7

stx gene subtype

eae gene

food product

reference

not specified

not specified

salami (mixed

(22)

California, U.S.,

pork and beef)

1994
Australia, 1995

O111:H-

stx1, stx2

positive

mettwurst

(23, 24)

(mixture of raw
pork, beef and
lamb)
Ontario, Canada,

O157:H7

not specified

not specified

1998
British Columbia,

O157:H7

not specified

not specified

Canada, 1999

salami (mixed

(25)(Williams

pork and beef)

et al., 2000)

salami (mixed

(26)

pork and beef)

28

Table 1.1. (cont’d)
Outbreak location

STEC

and year

serotype

Italy, 2004

O157:non-

stx gene subtype

eae gene

food product

reference

stx1, stx2

positive

salami (pork only)

(27)

not specified

not specified

pork from a whole

(28)

motile
Ontario, Canada,

O157:H7

2011

roasted pig

29

REFERENCES

30

REFERENCES

1.

Karmali, M.A., et al., The association between idiopathic hemolytic uremic syndrome
and infection by verotoxin-producing Escherichia coli. J Infect Dis, 1985. 151(5): p. 77582.

2.

Nataro, J.P. and J.B. Kaper, Diarrheagenic Escherichia coli. Clin Microbiol Rev, 1998.
11(1): p. 142-201.

3.

Karmali, M.A., Infection by verocytotoxin-producing Escherichia coli. Clin Microbiol
Rev, 1989. 2(1): p. 15-38.

4.

Tarr, P.I., C.A. Gordon, and W.L. Chandler, Shiga-toxin-producing Escherichia coli and
haemolytic uraemic syndrome. The Lancet, 2005. 365(9464): p. 1073-1086.

5.

Williams, D.M., et al., Acute kidney failure: a pediatric experience over 20 years. Arch
Pediatr Adolesc Med, 2002. 156(9): p. 893-900.

6.

Karmali, M.A., V. Gannon, and J.M. Sargeant, Verocytotoxin-producing Escherichia coli
(VTEC). Vet Microbiol, 2010. 140(3-4): p. 360-70.

7.

Scallan, E., et al., Foodborne Illness Acquired in the United States—Major Pathogens.
Emerging Infectious Diseases, 2011. 17(1): p. 7-15.

8.

Hoffmann, S., M.B. Batz, and J.G. Morris, Annual cost of illness and quality-adjusted life
year losses in the United States due to 14 foodborne pathogens. Journal of Food
Protection, 2012. 75(7): p. 1292-1302.

9.

Rangel, J.M., et al., Epidemiology of Escherichia coli O157:H7 outbreaks, United States,
1982-2002. Emerg Infect Dis, 2005. 11(4): p. 603-9.

10.

Doyle, M.P., Archer J, Kaspar CW, Weiss R Human illness casued by E. coli O157:H7
from food and non-food sources 2006 02/23/2012 Feburary, 2013]; Available from:
http://fri.wisc.edu/docs/pdf/FRIBrief_EcoliO157H7humanillness.pdf.

31

11.

Kaspar, C., Doyle ME, Archer J. White paper on non-O157:H7 Shiga toxin-producing E.
coli from meat and non-meat sources 2010 02/20/2012 Feburary 2013]; Available from:
http://fri.wisc.edu/docs/pdf/FRI_Brief_NonO157STEC_4_10.pdf.

12.

Durso, L.M., et al., Shiga-toxigenic Escherichia coli O157:H7 infections among livestock
exhibitors and visitors at a Texas County Fair. Vector Borne and Zoonotic Diseases,
2005. 5(2): p. 193-201.

13.

Keen, J.E., et al., Shiga-toxigenic Escherichia coli O157 in agricultural fair livestock,
United States. Emerg Infect Dis, 2006. 12(5): p. 780-6.

14.

Keen, J.E., L.M. Durso, and T.P. Meehan, Isolation of Salmonella enterica and Shigatoxigenic Escherichia coli O157 from feces of animals in public contact areas of United
States zoological parks. Appl Environ Microbiol, 2007. 73(1): p. 362-5.

15.

Spika, J.S., et al., Hemolytic uremic syndrome and diarrhea associated with Escherichia
coli O157:H7 in a day care center. J Pediatr, 1986. 109(2): p. 287-91.

16.

Carter, A.O., et al., A severe outbreak of Escherichia coli O157:H7--associated
hemorrhagic colitis in a nursing home. N Engl J Med, 1987. 317(24): p. 1496-500.

17.

Rowe, P.C., et al., Diarrhoea in close contacts as a risk factor for childhood haemolytic
uraemic syndrome. The CPKDRC co-investigators. Epidemiol Infect, 1993. 110(1): p. 916.

18.

Varma, J.K., et al., An outbreak of Escherichia coli O157 infection following exposure to
a contaminated building. The Journal of Aermican Medical Association, 2003. 290(20):
p. 2709-12.

19.

Griffin, P.M. and R.V. Tauxe, The epidemiology of infections caused by Escherichia coli
O157:H7, other enterohemorrhagic E. coli, and the associated hemolytic uremic
syndrome. Epidemiol Rev, 1991. 13: p. 60-98.

20.

Gyles, C.L., Shiga toxin-producing Escherichia coli: an overview. J Anim Sci, 2007.
85(13 Suppl): p. E45-62.

21.

Dean-Nystrom, E.A., et al., Pathogenicity of Escherichia coli O157:H7 in the intestines
of neonatal calves. Infection and Immunity, 1997. 65(5): p. 1842-8.

32

22.

CDC, Escherichia coli O157:H7 outbreak linked to commercially distributed dry-cured
salami- Washington and California, 1994. Morbidity and Mortality Weekly Report,
1995. 44: p. 157-160.

23.

CDC, Community outbreak of hemolytic uremic syndrome attributable to Escherichia
coli O111:NM-South Australia, 1995. Morbidity and Mortality Weekly Report, 1995. 44:
p. 550-558.

24.

Paton, A.W., et al., Molecular microbiological investigation of an outbreak of hemolyticuremic syndrome caused by dry fermented sausage contaminated with Shiga-like toxinproducing Escherichia coli. J Clin Microbiol, 1996. 34(7): p. 1622-7.

25.

Williams, R.C., et al., Illness outbreak associated with Escherichia coli O157:H7 in
Genoa salami. E. coli O157:H7 Working Group. Canadian Medical Association Journal,
2000. 162(10): p. 1409-13.

26.

MacDonald, D.M., et al., Escherichia coli O157:H7 outbreak linked to salami, British
Columbia, Canada, 1999. Epidemiol Infect, 2004. 132(2): p. 283-9.

27.

Conedera, G., et al., A family outbreak of Escherichia coli O157 haemorrhagic colitis
caused by pork meat salami. Epidemiol Infect, 2007. 135(2): p. 311-4.

28.

Trotz-Williams, L.A., et al., Pork Implicated in a Shiga Toxin-producing Escherichia coli
O157:H7 Outbreak in Ontario, Canada. Canadian Journal of Public Health. Revue
Canadienne de Sante Publique, 2012. 103(5): p. e322-6.

29.

Jay, M.T., et al., Escherichia coli O157:H7 in feral swine near spinach fields and cattle,
central California coast. Emerg Infect Dis, 2007. 13(12): p. 1908-11.

30.

Karmali, M.A., et al., The serotype and biotype distribution of clinical isolates of
Campylobacter jejuni and Campylobacter coli over a three-year period. Journal of
Infectious Diseases, 1983. 147(2): p. 243-6.

31.

Samadpour, M., et al., Occurrence of Shiga-like toxin-producing Escherichia coli in
retail fresh seafood, beef, lamb, pork, and poultry from grocery stores in Seattle,
Washington. Appl Environ Microbiol, 1994. 60(3): p. 1038-40.

33

32.

Brooks, H.J., et al., Occurrence and virulence factors of non-O157 Shiga toxin-producing
Escherichia coli in retail meat in Dunedin, New Zealand. Lett Appl Microbiol, 2001.
32(2): p. 118-22.

33.

Jo, M.Y., et al., Prevalence and characteristics of Escherichia coli O157 from major food
animals in Korea. Int J Food Microbiol, 2004. 95(1): p. 41-9.

34.

Oteiza, J.M., et al., Isolation and characterization of Shiga toxin-producing Escherichia
coli from precooked sausages (morcillas). Food Microbiology, 2006. 23(3): p. 283-8.

35.

Xia, X., et al., Presence and characterization of Shiga toxin-producing Escherichia coli
and other potentially diarrheagenic E. coli strains in retail meats. Appl Environ
Microbiol, 2010. 76(6): p. 1709-17.

36.

Ateba, C.N. and M. Mbewe, Detection of Escherichia coli O157:H7 virulence genes in
isolates from beef, pork, water, human and animal species in the northwest province,
South Africa: public health implications. Res Microbiol, 2011. 162(3): p. 240-248.

37.

Li, M.C., F. Wang, and F. Li, Identification and molecular characterization of
antimicrobial-resistant Shiga toxin-producing Escherichia coli isolated from retail meat
products. Foodborne Pathogens and Disease, 2011. 8(4): p. 489-93.

38.

Feder, I., et al., Testing of swine feces obtained through the National Animal Health
Monitoring System's Swine 2000 study for the presence of Escherichia coli O157:H7. J
Food Prot, 2007. 70(6): p. 1489-92.

39.

Bush, E., U.S. swine herd appears free of Escherichia coli O157:H7. Food Safety Digest,
1997. 4.

40.

Fratamico, P.M., et al., Prevalence and characterization of Shiga toxin-producing
Escherichia coli in swine feces recovered in the National Animal Health Monitoring
System's Swine 2000 study. Appl Environ Microbiol, 2004. 70(12): p. 7173-8.

41.

Fratamico, P.M., et al., Characterization of Shiga toxin-producing Escherichia coli
strains isolated from swine feces. Foodborne Pathogen and Disease, 2008. 5(6): p. 82738.

34

42.

Wells, J.E., et al., Effects of antimicrobials fed as dietary growth promoters on fecal
shedding of Campylobacter, Salmonella, and Shiga-toxin producing Escherichia coli in
swine. Journal of Applied Microbiology, 2012. 114(2): p. 318-328.

43.

Oporto, B., et al., Escherichia coli O157:H7 and non-O157 Shiga toxin-producing E. coli
in healthy cattle, sheep and swine herds in Northern Spain. Zoonoses Public Health,
2008. 55(2): p. 73-81.

44.

Pinaka, O., et al., Shiga toxin-producing Escherichia coli in Central Greece: prevalence
and virulence genes of O157:H7 and non-O157 in animal feces, vegetables, and humans.
European Journal of Clinical Microbiology and Infectious Diseases, 2013.

45.

Borie, C., et al., Prevalence and characterization of enterohaemorrhagic Escherichia coli
isoated from healthy cattle and pigs slaughtered in Santiago, Chile. Archivos De
Medicina Veterinaria, 1997. 29(2): p. 205-212.

46.

Rios, M., et al., Clonal diversity of Chilean isolates of enterohemorrhagic Escherichia
coli from patients with hemolytic-uremic syndrome, asymptomatic subjects, animal
reservoirs, and food products. J Clin Microbiol, 1999. 37(3): p. 778-81.

47.

Martins, R.P., et al., Prevalence of enterotoxigenic and Shiga toxin-producing
Escherichia coli in pigs slaughtered in Mato Grosso, Brazil. J Infect Dev Ctries, 2011.
5(2): p. 123-7.

48.

Borges, C.A., et al., Shiga Toxigenic and Atypical Enteropathogenic Escherichia coli in
the Feces and Carcasses of Slaughtered Pigs. Foodborne Pathogens and Disease, 2012.
9(12): p. 1119-1125.

49.

Milnes, A.S., et al., Factors related to the carriage of Verocytotoxigenic E. coli,
Salmonella, thermophilic Campylobacter and Yersinia enterocolitica in cattle, sheep and
pigs at slaughter. Epidemiol Infect, 2009. 137(8): p. 1135-48.

50.

Lenahan, M., et al., The potential use of chilling to control the growth of
Enterobacteriaceae on porcine carcasses and the incidence of E. coli O157:H7 in pigs. J
Appl Microbiol, 2009. 106(5): p. 1512-20.

51.

Heuvelink, A.E., et al., Isolation and characterization of verocytotoxin-producing
Escherichia coli O157 from slaughter pigs and poultry. Int J Food Microbiol, 1999. 52(12): p. 67-75.
35

52.

Bonardi, S., et al., Detection of Salmonella spp., Yersinia enterocolitica and
verocytotoxin-producing Escherichia coli O157 in pigs at slaughter in Italy. Int J Food
Microbiol, 2003. 85(1-2): p. 101-10.

53.

Johnsen, G., et al., Escherichia coli O157:H7 in faeces from cattle, sheep and pigs in the
southwest part of Norway during 1998 and 1999. Int J Food Microbiol, 2001. 65(3): p.
193-200.

54.

Eriksson, E., et al., Verocytotoxin-producing Escherichia coli O157:H7 in the Swedish
pig population. Vet Rec, 2003. 152(23): p. 712-7.

55.

Botteldoorn, N., et al., Detection and characterization of verotoxigenic Escherichia coli
by a VTEC/EHEC multiplex PCR in porcine faeces and pig carcass swabs. Res
Microbiol, 2003. 154(2): p. 97-104.

56.

Kaufmann, M., et al., Escherichia coli O157 and non-O157 Shiga toxin-producing
Escherichia coli in fecal samples of finished pigs at slaughter in Switzerland. J Food Prot,
2006. 69(2): p. 260-6.

57.

Nakazawa, M. and M. Akiba, Swine as a potential reservoir of Shiga toxin-producing
Escherichia coli O157:H7 in Japan. Emerg Infect Dis, 1999. 5(6): p. 833-4.

58.

Kijima-Tanaka, M., et al., A national surveillance of Shiga toxin-producing Escherichia
coli in food-producing animals in Japan. Journal of Veterinary Medicine. B, Infectious
Diseases and Veterinary Public Health, 2005. 52(5): p. 230-7.

59.

Kim, J., et al., Isolation and identification of Escherichia coli O157:H7 using different
detection methods and molecular determination by multiplex PCR and RAPD. Journal of
Veterinary Science, 2005. 6(1): p. 7-19.

60.

Keum, H.O., Kim HK, Rho SM, Moon HJ, Park SJ and Park BK, Comparative study of
Clostridium perfringens, Salmonella spp. and E.coli focused on characteristics of E. coli
O157 isolated from pigs of HACCP- and non-HACCP-accredited swine farms in Korea.
Korean Journal of Veterinary Research, 2010. 50(2): p. 113-116.

61.

Leung, P.H., et al., The prevalence and characterization of verotoxin-producing
Escherichia coli isolated from cattle and pigs in an abattoir in Hong Kong. Epidemiol
Infect, 2001. 126(2): p. 173-9.

36

62.

Yan, Y., et al., Prevalence of Stx phages in environments of a pig farm and lysogenic
infection of the field E. coli O157 isolates with a recombinant converting phage. Curr
Microbiol, 2011. 62(2): p. 458-64.

63.

Kobayashi, H., et al., Prevalence of pathogenic Escherichia coli in a swine breeding
environment in Can Tho Province, Vietnam. JARQ, Japan Agricultural Research
Quarterly, 2003. 37(1): p. 59-63.

64.

Wahlstrom, H., et al., Survey of Campylobacter species, VTEC O157 and Salmonella
species in Swedish wildlife. Vet Rec, 2003. 153(3): p. 74-80.

65.

Sánchez, S., et al., Detection and characterisation of O157:H7 and non-O157 Shiga
toxin-producing Escherichia coli in wild boars. Veterinary Microbiology, 2010. 143(24): p. 420-423.

66.

Pritchard, G.C., et al., Verocytotoxigenic Escherichia coli O157 in animals on public
amenity premises in England and Wales, 1997 to 2007. Vet Rec, 2009. 164(18): p. 545-9.

67.

Dohoo, I., Martin W, Stryhn H, Cross-sectional studies, in Veterinary Epidemiologic
Research. 2010, VER Inc.: Charlottetown, Prince Edward Island, Canada. p. 158-162.

68.

Fairbrother, J.M., Gyles CL, Colibacillosis, in Diseases of Swine, L.K. Jeffrey
Zimmerman, Alejandro Ramirez, Kent Schwartz,Gregory Stevenson Editor. 2012, John
Wiley and sons, Inc.: West Sussex, UK. p. 723-747.

69.

Fairbrother, J.M., E. Nadeau, and C.L. Gyles, Escherichia coli in postweaning diarrhea
in pigs: an update on bacterial types, pathogenesis, and prevention strategies. Anim
Health Res Rev, 2005. 6(1): p. 17-39.

70.

DeGrandis, S., et al., Globotetraosylceramide is recognized by the pig edema disease
toxin. Journal of Biological Chemistry, 1989. 264(21): p. 12520-5.

71.

Sonntag, A.K., et al., Shiga toxin 2e-producing Escherichia coli isolates from humans
and pigs differ in their virulence profiles and interactions with intestinal epithelial cells.
Appl Environ Microbiol, 2005. 71(12): p. 8855-63.

72.

Bertschinger, H.U., Gyles CL Oedema disease of pigs, in Escherichia coli in doemstic
animals humans, C.L. Gyles, Editor. 1994, CAB International: Wallingofrd, UK. p. 193219.
37

73.

Vernozy-Rozand, C., et al., Serotyping, stx2 subtyping, and characterization of the locus
of enterocyte effacement island of Shiga toxin-producing Escherichia coli and E. coli
O157:H7 strains isolated from the environment in France. Appl Environ Microbiol,
2004. 70(4): p. 2556-9.

74.

Beutin, L., et al., Identification of human-pathogenic strains of Shiga toxin-producing
Escherichia coli from food by a combination of serotyping and molecular typing of Shiga
toxin genes. Appl Environ Microbiol, 2007. 73(15): p. 4769-75.

75.

Thomas, A., et al., Isolation of Vero cytotoxin-producing Escherichia coli serotypes
O9ab:H- and O101:H-carrying VT2 variant gene sequences from a patient with
haemolytic uraemic syndrome. Eur J Clin Microbiol Infect Dis, 1994. 13(12): p. 1074-6.

76.

Pierard, D., et al., Diarrhoea associated with Escherichia coli producing porcine oedema
disease verotoxin. Lancet, 1991. 338(8769): p. 762.

77.

Muniesa, M., et al., Characterization of a Shiga toxin 2e-converting bacteriophage from
an Escherichia coli strain of human origin. Infect Immun, 2000. 68(9): p. 4850-5.

78.

Friedrich, A.W., et al., Escherichia coli harboring Shiga toxin 2 gene variants: frequency
and association with clinical symptoms. J Infect Dis, 2002. 185(1): p. 74-84.

79.

Beutin, L., et al., Characterization of Shiga toxin-producing Escherichia coli strains
isolated from human patients in Germany over a 3-year period. J Clin Microbiol, 2004.
42(3): p. 1099-108.

80.

Beutin, L., et al., Evaluation of major types of Shiga toxin 2e-producing Escherichia coli
bacteria present in food, pigs, and the environment as potential pathogens for humans.
Applied and Environmental Microbiology, 2008. 74(15): p. 4806-4816.

81.

Feng, P.C., et al., Specificity of PCR and serological assays in the detection of
Escherichia coli Shiga toxin subtypes. Applied and Environmental Microbiology, 2011.
77(18): p. 6699-702.

82.

Bugarel, M., et al., Virulence gene profiling of enterohemorrhagic (EHEC) and
enteropathogenic (EPEC) Escherichia coli strains: a basis for molecular risk assessment
of typical and atypical EPEC strains. BMC Microbiol, 2011. 11: p. 142.

38

83.

Cobbold, R. and P. Desmarchelier, A longitudinal study of Shiga-toxigenic Escherichia
coli (STEC) prevalence in three Australian diary herds. Vet Microbiol, 2000. 71(1-2): p.
125-37.

84.

Slanec, T., et al., Molecular analysis of virulence profiles and Shiga toxin genes in foodborne Shiga toxin-producing Escherichia coli. Appl Environ Microbiol, 2009. 75(19): p.
6187-97.

85.

Feder, I., et al., Isolation of Escherichia coli O157:H7 from intact colon fecal samples of
swine. Emerg Infect Dis, 2003. 9(3): p. 380-3.

86.

Chapman, T.A., et al., Comparison of virulence gene profiles of Escherichia coli strains
isolated from healthy and diarrheic swine. Appl Environ Microbiol, 2006. 72(7): p. 478295.

87.

Zweifel, C., et al., Virulence profiles of Shiga toxin 2e-producing Escherichia coli
isolated from healthy pig at slaughter. Vet Microbiol, 2006. 117(2-4): p. 328-32.

88.

Cobbold, R. and P. Desmarchelier, Characterisation and clonal relationships of Shigatoxigenic Escherichia coli (STEC) isolated from Australian dairy cattle. Vet Microbiol,
2001. 79(4): p. 323-35.

89.

Lanier, W.A., et al., Pulsed-field gel electrophoresis subtypes of Shiga toxin-producing
Escherichia coli O157 isolated from ground beef and humans, United States, 2001-2006.
Foodborne Pathogens and Disease, 2009. 6(9): p. 1075-82.

90.

Barrett, T.J., et al., Laboratory investigation of a multistate food-borne outbreak of
Escherichia coli O157:H7 by using pulsed-field gel electrophoresis and phage typing. J
Clin Microbiol, 1994. 32(12): p. 3013-7.

91.

Samadpour, M., Molecular epidemiology of Escherichia coli O157:H7 by restriction
fragment length polymorphism using Shiga-like toxin genes. J Clin Microbiol, 1995.
33(8): p. 2150-4.

92.

Diekema, D.J., et al., A cluster of serious Escherichia coli infections in a neonatal
intensive-care unit. Infect Control Hosp Epidemiol, 1997. 18(11): p. 774-6.

39

93.

Izumiya, H., et al., Molecular typing of enterohemorrhagic Escherichia coli O157:H7
isolates in Japan by using pulsed-field gel electrophoresis. J Clin Microbiol, 1997. 35(7):
p. 1675-80.

94.

Singer, R.S., W.M. Sischo, and T.E. Carpenter, Exploration of biases that affect the
interpretation of restriction fragment patterns produced by pulsed-field gel
electrophoresis. J Clin Microbiol, 2004. 42(12): p. 5502-11.

95.

Franke, S., et al., Clonal relatedness of Shiga-like toxin-producing Escherichia coli O101
strains of human and porcine origin. J Clin Microbiol, 1995. 33(12): p. 3174-8.

96.

Lenahan, M., et al., Molecular characterization of IrishE. coliO157:H7 isolates of
human, bovine, ovine and porcine origin. Journal of Applied Microbiology, 2009. 107(4):
p. 1340-1349.

97.

Osek, J. and P. Gallien, Molecular analysis of Escherichia coli O157 strains isolated
from cattle and pigs by the use of PCR and pulsed-field gel electrophoresis methods.
Veterinarni Medicina, 2002. 47(6): p. 149-158.

98.

Kennedy, C.A., et al., Shiga toxin-producing Escherichia coli isolated from human and
pig origin induce different gene expression profiles in human Caco-2 epithelial cells.
Livestock Science, 2010. 133(1/3): p. 189-191.

99.

Jordan, D.M., et al., Long polar fimbriae contribute to colonization by Escherichia coli
O157:H7 in vivo. Infect Immun, 2004. 72(10): p. 6168-71.

100.

Baker, D.R., et al., Differences in Virulence among Escherichia coli O157:H7 Strains
Isolated from Humans during Disease Outbreaks and from Healthy Cattle. Applied and
Environmental Microbiology, 2007. 73(22): p. 7338-7346.

101.

Shringi, S., et al., Differential virulence of clinical and bovine-biased enterohemorrhagic
Escherichia coli O157:H7 genotypes in piglet and Dutch belted rabbit models. Infect
Immun, 2012. 80(1): p. 369-80.

102.

Sheoran, A.S., et al., Human antibody against Shiga toxin 2 administered to piglets after
the onset of diarrhea due to Escherichia coli O157:H7 prevents fatal systemic
complications. Infect Immun, 2005. 73(8): p. 4607-13.

40

103.

Dean-Nystrom, E.A., et al., Comparative pathogenicity of Escherichia coli O157 and
intimin-negative non-O157 Shiga toxin-producing E coli strains in neonatal pigs. Infect
Immun, 2003. 71(11): p. 6526-33.

104.

Jordan, D.M., S.L. Booher, and H.W. Moon, Escherichia coli O157:H7 does not require
intimin to persist in pigs. Infect Immun, 2005. 73(3): p. 1865-7.

105.

Cornick, N.A., S.L. Booher, and H.W. Moon, Intimin facilitates colonization by
Escherichia coli O157:H7 in adult ruminants. Infect Immun, 2002. 70(5): p. 2704-7.

106.

Booher, S.L., N.A. Cornick, and H.W. Moon, Persistence of Escherichia coli O157:H7 in
experimentally infected swine. Vet Microbiol, 2002. 89(1): p. 69-81.

107.

Cray, W.C., Jr. and H.W. Moon, Experimental infection of calves and adult cattle with
Escherichia coli O157:H7. Applied and Environmental Microbiology, 1995. 61(4): p.
1586-90.

108.

Cornick, N.A. and A.F. Helgerson, Transmission and infectious dose of Escherichia coli
O157:H7 in swine. Appl Environ Microbiol, 2004. 70(9): p. 5331-5.

109.

Cornick, N.A. and H. Vukhac, Indirect transmission of Escherichia coli O157:H7 occurs
readily among swine but not among sheep. Appl Environ Microbiol, 2008. 74(8): p.
2488-91.

110.

Cornick, N.A., Tylosin and chlorotetracycline decrease the duration of fecal shedding of
E. coli O157:H7 by swine. Vet Microbiol, 2010. 143(2-4): p. 417-9.

111.

Griffin, P.M., et al., Illnesses associated with Escherichia coli O157:H7 infections. A
broad clinical spectrum. Annals of Internal Medicine, 1988. 109(9): p. 705-12.

112.

Law, D., Virulence factors of Escherichia coli O157 and other Shiga toxin-producing E.
coli. J Appl Microbiol, 2000. 88(5): p. 729-45.

113.

Bettelheim, K.A., The non-O157 Shiga-toxigenic (verocytotoxigenic) Escherichia coli;
under-rated pathogens. Crit Rev Microbiol, 2007. 33(1): p. 67-87.

41

114.

Henderson, H., Direct and indirect zoonotic transmission of Shiga toxin-producing
Escherichia coli. Journal of the American Veterinary Medical Assocciation, 2008.
232(6): p. 848-59.

115.

CDC. National Shiga toxin-producing Escherichia coli (STEC) surveillance annual
summary, 2009. 2012 Sep, 10, 2012 January, 2013]; Available from:
http://www.cdc.gov/ncezid/dfwed/PDFs/national-stec-surv-summ-2009-508c.pdf.

116.

FSIS, Food Safety and Inspection Service. Shiga toxin-producing Escherichia coli in
certain raw beef products. Federal Register, 2011. 76: p. 58157-58165.

117.

EFSA, Technical specifications for the monitoring and reporting of verotoxigenic
Escherichia coli (VTEC) on animals and food (VTEC surveys on animals and food).
EFSA Journal, 2009. 7(11): p. 1366.

118.

Bielaszewska, M., et al., Characterisation of the Escherichia coli strain associated with
an outbreak of haemolytic uraemic syndrome in Germany, 2011: a microbiological study.
Lancet Infect Dis, 2011. 11(9): p. 671-6.

42

CHAPTER 2
Shiga toxin-producing E. coli (STEC) in swine: prevalence over the finishing period and
characteristics of the STEC isolates

Manuscript based on the data in this chapter has been has been accepted for publication in
Epidemiology and Infection on April 10th 2014. (M. Tseng, P. M. Fratamico, L. Bagi, D.
Manzinger, J. A. Funk. Shiga toxin-producing E. coli (STEC) in swine: prevalence over the
finishing period and characteristics of the STEC isolates)

43

ABSTRACT

This descriptive longitudinal study was conducted to investigate the fecal shedding of
Shiga toxin-producing E. coli (STEC) in finishing swine and to characterize the swine STEC
isolates that were recovered. Three cohorts of finishing swine (n=50/cohort; total 150 pigs) were
included in the longitudinal study. Individual fecal samples were collected every 2 weeks (8
collections/pig) from the beginning (pig age=10 weeks old) to the end (pig age =24 weeks old)
of the finishing period. STEC isolates were recovered in at least one sample from 65.3%
(98/150) of the pigs, and the frequency distribution of first-time STEC detection during the
finishing period resembled an outbreak curve. Nineteen O:H serotypes were identified among
the STEC isolates. Most STEC isolates (n=148) belonged to serotype O59:H21 and carried the
stx2e gene. One O49:H21 STEC isolate carried the stx2e and eae gene. High prevalence rates of
STEC during the finishing period were observed, and STEC isolates in various non-O157
serogroups were recovered. These data enhance understanding of swine STEC epidemiology,
and future research is needed to confirm whether or not swine STEC are of public health
concern.

44

INTRODUCTION

Shiga toxin-producing Escherichia coli (STEC) represent a subset of E. coli that produce
a cytotoxin known as Shiga toxin (Stx) encoded by the stx gene (1). Stx has two types (Stx1,
Stx2), and each Stx type has multiple variants (Stx1c, Stx1d; Stx2c, Stx2d, Stx2e, Stx2f, Stx2g)
(2). STEC infection is associated with outbreaks and sporadic cases of diarrhea and severe
clinical diseases in humans, including hemorrhagic colitis (HC) and the potentially fatal
hemolytic uremic syndrome (HUS) (2). Therefore, STEC are a critical public health concern,
and they cause more than 170000 cases of human illness yearly in the U.S. (3). The infections
are most frequently acquired by consuming contaminated food (meat, dairy products, produce,
and other foods) and water (2). While STEC serotype O157:H7 is viewed as the serotype
associated with most reported outbreaks and severe diseases, non-O157 STEC-associated
outbreaks have been increasingly documented (4). Notably, more than 50% of human STEC
infections have been attributed to non-O157 STEC (3), but our current knowledge of non-O157
STEC is very limited.
Although cattle are considered the primary STEC reservoirs for human infections, there
are food products from other animal species, including pork products, that have been implicated
as vehicles in STEC transmission (5-11). A recent STEC outbreak was associated with
consuming large cuts of pork from a whole roasted pig (11). Even though investigators were
unable to trace back the original sources of STEC in these investigations, there were possible
sources of contamination, including intrinsically infected swine or fecal contamination during
swine processing, or the STEC could come from non-swine sources, such as cross contamination
during preparation, processing or incorporation of contaminated ingredients from other animal

45

species (8-11). Further research is warranted to elucidate whether swine are a source of STEC
contamination. Swine, unlike cattle, may have clinical disease associated with STEC infection
(12). Edema disease, often occurring in post-weaning or young finishing age pigs, is caused by
STEC strains carrying the stx2e gene (12). Cross-sectional epidemiologic studies have been
performed to estimate the prevalence of STEC in clinically healthy swine in multiple regions of
the world. STEC prevalence in these studies ranged from 0% (13) to 68.3% (14). In the U.S.,
earlier studies focused on the detection of STEC O157:H7, and the prevalence estimates were
low, ranging from 0% (15) to 1.9% (16). In one more recent survey, STEC isolates were
recovered in 28.5% of the fecal samples collected, and all of the isolates were categorized as
non-O157 serotypes (17). The role that swine play in the epidemiology of STEC, specifically
non-O157 STEC, needs further investigation.
Previous studies clearly indicate that pigs can shed STEC, particularly non-O157 STEC
serotypes. Yet, they are limited as they employed cross-sectional study designs relying on point
estimates of STEC prevalence, and furthermore, only recently have efforts been made to identify
and characterize non-O157 serotypes. Little is known about fecal shedding of STEC in clinically
healthy swine over time. This longitudinal study was conducted to fill this gap by investigating
prevalence of STEC in swine over the finishing period, and to characterize the STEC isolates
recovered.

46

MATERIAL AND METHODS

Study design

This longitudinal study was conducted on two finishing sites (Site A and Site B) within
one all-in, all-out multi-site production system in the Midwestern U.S. This means that the pigs
originated from the same sow herd, but were reared separately from birth to marketing. These
two sites were selected based on convenience and the producers’ willingness to participate in the
study. Fecal samples were collected from three cohorts of finishing pigs (one cohort on Site A
and two cohorts on Site B) from 10 weeks of age until 24 weeks of age (approximately market
age). For cohort 1, the farm visits began in May 2011 and ended in August 2011. For cohort 2,
the farm visits began in July 2011 and ended in October 2011. For cohort 3, the farm visits
began in November 2011 and ended in February 2012.
Site A was a wean-to-finish facility, in which weaned pigs (pig age = 3 to 4 weeks old)
were placed into the barns and raised to market age. Site B was a finishing facility, in which
pigs, after being housed in a nursery facility (from 3 to 10 weeks of age), were moved into the
barns and raised until market age. After a batch of pigs of the same age was placed into the barn,
no new pigs were introduced into the barn. These two sites had similar building designs. Each
site had two separate buildings, and two separate barns were within each building. Thus, there
were four barns on each site. Each barn had 12 pens and was capable of housing a total of 1000
pigs (total site inventory of 4000 pigs). In each barn, there were eight “large” and four “small”
pens. Approximately 100 to 125 pigs were placed in each of eight large pens, and 50 pigs were

47

housed in 2 of the 4 small pens. The remaining small pens were used for housing sick pigs or
pigs deemed to be at high risk for disease. Pen dividers allowed pig to pig contact between pens.

Sample collection

A total of 150 individually identified finishing pigs (n=50/cohort; three cohorts) were
included in this study. For each cohort, the same proportional sampling scheme was followed.
When pigs were 10 weeks of age, 50 pigs in a single barn were randomly selected: six pigs per
pen in six large pens, five pigs per pen in two large pens, and two pigs per pen in 2 of the 4 small
pens (hospital and at risk pens were empty at the time of placement). The pigs were selected
based on random numbers generated by Microsoft Office Excel 2007 (Microsoft, U.S.).
From 10 weeks of age, fecal samples were collected from the selected pigs every 2 weeks for a
total of eight collections (16 weeks). A total number of 1200 fecal samples (50 pigs/cohort; 8
collections/cohort; 3 cohorts) were planned to be collected. At each collection period, health
conditions (e.g. diarrhea, lameness, clinically healthy) of the selected pigs were observed and
recorded by research personnel, and health records (e.g. use of medication, signs of clinical
symptoms, numbers of pigs that died, potential cause of death) documented by the producers
were also recorded.
Individual fecal samples were collected directly from the finishing pigs by gloved hands,
and new gloves were used for each pig. Fecal samples were placed into sterile VWR®
microbiology/urinalysis specimen containers (VWR International, U.S.). The specimen
containers were placed into a cooler under ambient temperature for transportation to the
laboratory at Michigan State University in East Lansing, MI. Fecal samples were stored at 2.7ºC

48

up to 48 h prior to shipping for culture depending on the availability of the shipping service. The
fecal samples where then shipped overnight on ice packs to the laboratory located at the Eastern
Regional Research Centre of the U.S. Department of Agriculture in Wyndmoor, PA. All fecal
samples were processed 24 to 72 h after collection at the farm.

STEC detection and isolation in swine fecal samples

The sample enrichment method was modified from Grant et al., 2009 (18). In summary,
a five-gram portion of each swine fecal sample was added to 95 mL of tryptic soy broth (TSB) at
pH 3.0 in a filter Stomacher bag. The bag was subjected to pummeling in a Stomacher for 30
sec, and then incubated at room temperature for 10 to 15 min. One hundred millilitres of TYTP
(TSB + 12.0 g/L yeast extract, 12.5 g/L Trizma base, and 1.0 g/L sodium pyruvate, with a final
pH of 8.7) were then added, and samples were incubated without rotation for 15 h at 41˚C.
DNA was then extracted using the PrepSEQ Rapid Spin Sample Preparation Kit (Life
Technologies Corporation, U.S.) according to the manufacturer’s instructions. A multiplex PCR
assay was performed using primers Stx 1/2-F and mod-Stx1/2-R and probes Stx1-P [FAM] and
Stx2-P [FAM], targeting stx1, stx2 and all variants except stx2f and Eae-F and mod-Eae-R primers
and EaeP [MAXN] probe, targeting the eae gene (19), which encodes for the outer membrane
protein intimin important for attachment to the intestinal epithelial cells. The PCR assay was
performed using 2 μl of template DNA and the TaqMan® Environmental Master Mix 2.0 (Life
Technologies Corporation, U.S.) as described by Wasilenko and co-workers (19). The PCR
assay was performed in an Applied Biosystems 7500 thermal cycler using a protocol consisting
of 95˚C for 10 min, and then 40 cycles of 95˚C for 15 sec and 60˚C for 60 sec.

49

Enrichment samples that were positive for the stx gene were then plated onto
CHROMagar STEC (DRG International Inc., U.S.). Although CHROMagar STEC, which
contains tellurite, prevents growth of tellurite-sensitve STEC strains, CHROMagar STEC has
been suggested to allow growth of 75 to 86.4% of STEC in various serotypes (20-22). From
each plate, three presumptive positive colonies were picked and confirmed as STEC using the
stx1/2-eae multiplex PCR assay described above. For the confirmed STEC isolates, stx gene
types (stx1, stx2) and stx2e variant were characterized by PCR assays described in earlier
publications (17, 23).

Serotyping of the STEC isolates

At least one confirmed STEC isolate was selected from every positive sample (154
STEC-isolate-positive samples/1040 collected samples) for O:H serotype characterization. For
O serogrouping, the confirmed STEC isolates were submitted to the E. coli Reference Centre at
the Pennsylvania State University in University Park, PA. Antisera were used against
serogroups O1 to O181, except for O31, O47, O72, O93, O94, and O122, which are not
designated. All STEC isolates with O serogroup information were submitted for H typing at the
French Agency for Food, Environmental and Occupational Health and Safety (Maisons-Alfort,
France). Selected flagellar antigen genes (H2, H7, H8, H11, H19, H21, H28, H16, H25, H4,
H32) were determined as part of a high-throughput real-time PCR system modified from
previously published methods (24). The STEC isolates which were not H typeable by the realtime PCR system were submitted to the E. coli Reference Centre (University Park, PA) for H
typing by PCR-Restriction Fragment Length Polymorphism of the flagellar antigen gene.

50

Sample size calculation

The sample size was estimated by using the population survey formula in the program
Epi Info version 7 (CDC, U.S.). An earlier study by Fratamico et al. (17) reported a 28.5%
individual animal prevalence of STEC in finishing swine in the U.S. We chose an estimated
prevalence of 30.0% with a 10.0% confidence limit and a population size of 1000, and the
confidence was set as 90.0%. This resulted in a target sample size of 50 for each cohort to
estimate prevalence at each sampling period.

Variables and statistical methods

All data were recorded and managed in Microsoft Office Excel 2007 (Microsoft, U.S.).
A pig was considered STEC positive when at least one confirmed STEC isolate was recovered
from the fecal sample. The outcome of interest was the proportion (prevalence rate) of pigs
positive for STEC at each farm visit. After data collection was completed, data validation was
performed by comparing the records of farm visits and STEC isolation results from the
laboratory to the database. Descriptive statistics of the proportion of pigs positive for STEC at
each farm visit was performed in Microsoft Office Excel 2007 (Microsoft, U.S.).
For each pig positive for STEC, the duration of STEC shedding was defined as the time
interval (days) between the first and last sampling date of positive STEC isolation, which was
before three consecutive STEC negative samples. An additional 14 days was counted for each
positive sampling date to account for sampling intervals. Right censored data included pigs that
died or were shipped to market before the observation of three consecutive STEC negative

51

samples. Survival analysis using the Kaplan-Meier method was performed to analyse the
duration of STEC shedding in finishing swine by STATA 13 (STATACorpLP, U.S.). The
equality of survivor curves by cohort, gender, and STEC serotype was examined by the PetoPeto-Prentice test. P-values <0.05 were considered significant.

52

RESULTS

Finishing pigs and sample collection

A total of 150 pigs (50 pigs in each cohort, three cohorts in total) were included in the
longitudinal sampling. Ten (6.7%) out of the 150 pigs died during the study period. Diarrhea
was observed in at least collection period in 12.0% (6/50) of the pigs in cohort 1, 12.0% (6/50) of
the pigs in cohort 2, and 10.0% (5/50) of the pigs in cohort 3, and samples were collected from
these pigs with diarrhea. There were seven instead of eight farm visits for cohort 1 due to
marketing of the pigs prior to the final collection, and eight farm visits occurred for cohort 2 and
for cohort 3. Eight farm visits were completed in 43.3% (65/150) of the pigs, and seven farm
visits were completed in 36.7% (55/150) of the pigs. At the end of sample collection, a total of
1040 fecal samples were collected. The main reasons why samples may not have been collected
at farm visits were that the pigs were too ill, the pigs were dead, or the pigs were shipped for
marketing before sample collection (n=14 in cohort 1, n=0 in cohort 2, n=9 in cohort 3). The
sample collection time and demographic information of the finishing pigs are summarized in
Table 2.1.

Distribution of STEC positive pigs over the finishing period

STEC isolates were recovered from at least one sample from 65.3% (98/150) of the pigs.
Specifically, STEC isolates were recovered in a least one sample from 62.0% (31/50) of the pigs
in cohort 1, 54.0% (27/50) of the pigs in cohort 2, and 80.0% (40/50) of the pigs in cohort 3.

53

Figure 2.1.a illustrates the proportions (prevalence rates) of pigs with STEC isolates by pig age
in weeks over the finishing period, and Figure 2.1.b displays numbers of pigs at the age of firsttime STEC detection. The highest prevalence rates of STEC positive pigs were detected when
pigs were 18 weeks old in cohort 1 (47.9%), 16 weeks old in cohort 2 (39.5%), and 14 weeks old
in cohort 3 (59.2%). The numbers of pigs at the age of first-time STEC detection peaked at the
same age as the proportions of STEC positive pigs in every cohort. At the end of finishing
period, the prevalence rates of STEC positive pigs were lower: 6.3% in cohort 1, 0% in cohort 2,
and 6.7% in cohort 3. STEC isolates were recovered in one fecal sample in 36.7% (55/150), two
samples in 20.7% (31/150), three samples in 7.3% (11/150), and four samples in 0.7% (1/150) of
the pigs. No STEC isolates were ever recovered from samples of the pigs that had diarrhea.

Characterization of the STEC isolates

Among the 1040 swine fecal samples, total 285 STEC isolates were recovered from
83.2% (154/185) stx-gene-positive samples. The presence of virulence genes (stx1, stx2, stx2e,
eae) were determined by PCR in all the 285 STEC isolates. Most of the STEC isolates (97.9%,
279/285) carried the stx2e gene. Four STEC isolates carried the stx1 gene. Two STEC isolates
carried the stx2 gene and were negative for stx2e. The eae gene was detected in only one of the
285 STEC isolates.
For O:H serotype characterization, at least one isolate was selected from every STEC
positive pig at each farm visit. A total of 200 STEC isolates were submitted for O:H serotype
determination. Among the 200 STEC isolates, ten different O serogroups were identified while
some isolates (n=29) were O serogroup non-typeable. Nine H types were identified in the 200

54

STEC isolates, and together, nineteen different O:H serotypes were identified. A majority
(73.6%, 148/201) of the STEC isolates were categorized as serotype O59:H21. The results of
O:H serotype and virulence gene (stx1, stx2, stx2e, eae) characterization are summarized in Table
2.2.

Estimation of duration of STEC shedding in finishing swine

The pigs (n=19) with samples positive for STEC isolates belonging to more than one
serotype were excluded from the survival analysis. Therefore, data from 79 pigs with samples
positive for STEC isolates for one serotype were included in the survival analysis. In these 79
pigs, data from 34.2% (27/79) of the pigs were right-censored. In the 27 right-censored pigs,
7.4% (2/27) of the pigs died during the study period, and 40.7% of them (11/27) were shipped to
market before the study ended. The event (three consecutive STEC negative samples) was not
observed in the remaining 14 of the 27 censored pigs before the study ended. In the 79 pigs
included in the survival analysis, the median of the duration of STEC fecal shedding was 28
days. Kaplan-Meier analysis suggested that the cumulative probability of finishing swine
shedding STEC for 28 days was 30.0%. The shortest shedding duration was 14 days at a
probability of 53.2%. The longest shedding duration was 56 days at a probability of 6.7%.
Figure 2.2. illustrates the Kaplan-Meier survival curve of duration of fecal STEC shedding in
finishing swine. No significant differences were observed among survival curves by cohorts (p=
0.20), gender (p=0.96), and STEC serotypes (p=0.84).

55

DISCUSSION

This study represents the first longitudinal study of STEC shedding in commercial swine.
Our results indicated that finishing swine shed STEC at relatively high prevalence rates during
the finishing period. Previous studies have reported a wide range of STEC prevalence in swine,
but most of them were cross-sectional study designs, challenging direct comparison to this study
(13-15, 17). In addition to study designs, comparison is limited because of the type of sample
collected, sample collection method, and STEC isolation protocols, which vary widely across
these previous reports. Although only one production company participated in this study, this
company sufficiently represents the majority of conventional swine production systems in the
U.S. (all-in, all-out multi-site production, etc.)
Variations in STEC prevalence rates over time were observed in this study, and this
highlights the importance of longitudinal sampling to determine the STEC shedding status on
farm. A shape similar to an outbreak curve was observed when plotting the numbers of pigs at
the age of first-time STEC detection, and the outbreak-like curves were observed from the results
of all three of the cohorts. This may suggest that the pigs were exposed to a single source of
STEC infection in the beginning of finishing period. The STEC prevalence rates were different
by pig age in these three cohorts. At the cohort level, no significant change of barn
environments and farm management protocols was observed during the study period in these
three cohorts. At the individual pig level, no significant factors (e.g. sex and health condition)
were observed in the STEC-positive pigs. Common sources of STEC infection may be from the
finishing barn environment, the finishing diet, or other factors. To further determine whether the
STEC strains were from a common source of infection, it will be essential for future studies to

56

examine the genetic relatedness of the swine STEC strains by molecular genotyping methods.
Moreover, expanding the scope of observations to more finishing sites will be crucial to
understand the potential cohort-level or individual risk factors associated with the dynamics of
STEC prevalence on swine farms.
The duration of detectable shedding in this study was similar to that described in
experimentally-challenged swine. Regardless of the inoculation dose, STEC O157:H7 strains
were recovered in fecal samples 14 to16 days after inoculation (25). However, when inoculated
with a higher dose at 1010 CFU/strain/animal, STEC O157:H7 strains were able to be recovered
in fecal samples 58 to 60 days after inoculation (25). Together, these data suggest that some pigs
in this study might be exposed to a higher dose of STEC during the finishing period, and this
resulted in a longer duration of STEC shedding. Moreover, re-infection of STEC from other pigs
in the same barn may occur. Re-infection of STEC from pigs in neighboring barns may also
occur because of potential pig to pig contact through the pen dividers in the barns in this study.
Additionally, the differences of shedding duration may be attributed to differential colonization
ability of different STEC strains and the host itself. One study reported that no difference in
duration of STEC shedding comparing pigs housed on crates and on cement floors (26).
However, what the potential factors are and how these factors contribute to the duration of STEC
shedding in pigs remain largely unknown. A better understanding of the factors associated with
the duration of shedding will be helpful for strategizing future plans for controlling STEC
shedding on farms.
The STEC isolates recovered in this study belonged to a number of different serotypes.
However, none belonged to O-serogroups O138, O139, O141, and O147, which are more
frequently associated with edema disease in swine (12). The serotypes were different from those

57

reported in the swine STEC isolates from the NAHMS 2000 study (17). Different STEC
isolation methods may contribute to the observations of different serotypes. For example,
CHROMagar STEC, which was used in this study, contains tellurite, and some STEC strains
were not able to grow on this medium (20-22). Nevertheless, CHROMagar STEC was suggested
to be useful for STEC isolates recovery from fecal samples, and it allowed the growth of 75% to
86.5% STEC strains in a wide variety of serotypes that were examined in these previous study
(20-22).
The majority of the STEC isolated in the current study were serotype O59:H21. This
particular serotype has not yet been reported in human cases but was reported in STEC isolates
from food (beef, pork, and others) (27). Some serotypes and O-serogroups of the swine STEC
isolated in this study have been reported in human cases. For instance, STEC O59:H19 (28),
O20 (29), O49, O98, and O119 (4) have been associated with human cases. Results of this study
show that swine carry a variety of different STEC serotypes, and this was observed with
sampling only three groups of pigs in one swine production company.
Most swine STEC isolates in this study carried the stx2e gene. Interestingly, Stx2eproducing E. coli strains have not only been recovered in pigs, but also in humans with HUS (30)
and uncomplicated diarrhea (31-34). Although no source of infection was ascertained in these
human cases, the association between Stx2e-producing E. coli and human illness requires further
examination. The eae gene was detected in only one STEC isolate (O49:H21, carrying the stx2e
gene) in this study. The presence of the eae gene in swine STEC was not consistent in previous
studies. Some studies reported that the eae gene was not detected in swine STEC (35), while
others have reported detection of eae in STEC O157:H7 strains recovered from swine colon fecal
samples and carcass swabs at slaughter houses (36). Although it is well-known that intimin is

58

essential in STEC attachment in humans (37), it has been suggested that intimin is not required
for STEC O157:H7 colonization in pigs (38). This may partially explain the fact that only one
out of 285 STEC isolates in this study was eae-positive. However, outbreaks and cases
associated with eae-negative STEC strains were reported (39). Many novel adhesins have been
proposed as possibly being important in eae-negative STEC, such as the STEC autoagglutinating
adhesin encoded by the saa gene (40). To better assess the potential risks posed by swine STEC,
there is a need to more extensively examine the virulence gene profiles in swine STEC isolates.
Swine have not been viewed as an important STEC reservoir. However, the findings of
this study provide insights into the epidemiology of fecal STEC shedding in finishing swine. It
was observed that swine can shed STEC at high prevalence rates during the finishing period.
The swine STEC isolates recovered in this study belonged to various non-O157 STEC serotypes,
and E. coli O157:H7 was not isolated. There is increasing awareness of the public health burden
associated with non-O157 STEC infections (3). In addition, the recent emergence of highly
virulent non-O157 STEC possessing unusual virulence gene combinations stress the need to
further understand these pathogens (39). From this study, relative high prevalence rates of STEC
isolation was observed during swine finishing period. Moreover, STEC isolates in various nonO157 serotypes have been recovered. Future study is warranted to confirm whether or not swine
are an important source of human STEC infections, specifically non-O157 STEC.

59

ACKNOWLEDGEMENTS

This work was supported by the National Pork Board (project number 12-069).
Sincere thanks to the producers who participated in this research, MSU students: Dr. Alda Pires,
Alexandra Buckley, Allan Mergener, Tasha Likavec, Joel Sparks for assistance with sample
collection and shipment. Dr. Chitrita DebRoy and the E. coli Reference Centre at The
Pennsylvania State University, Dr. Patrick Fach and Dr. Sabine Delannoy at the French Agency
for Food, Environmental and Occupational Health and Safety are acknowledged for assistance
with isolate characterization.

60

APPENDIX

61

Table 2.1. Sample time periods, demographic information and numbers of fecal samples of
finishing pigs
Cohort 1

Cohort 2

Cohort 3

Production site

A

B

B

Sample collection period

31 May 2011

18 July 2011

21 November

(visit 1) -

(visit 1) -

2011 (visit 1) -

22 August

22 October

18 February 2012

2011 (visit 7)

2011 (visit 8)

(visit 8)

Number of animals in each cohort

50

50

50

Number of animals died during

2

6

2

male

23

17

24

female

27

33

26

320

357

363

study period
Sex

Number of fecal samples collected

62

Table 2.2. Distribution of STEC isolates by O:H serotype, Shiga toxin gene subtypes, and eae

Number of isolates positive for
Total
number of

other stx2

Serotype

isolates

stx2e

variants

stx1 variants

eae

O15:H45

1

1

0

0

0

O15:H+e

1

0

1

0

0

O20:H21

1

1

0

0

0

O49:H21

1

1

0

0

1

O59:H19

3

3

0

0

0

O59:H21

148

148

0

0

0

O59:H21/H4b

3

3

0

0

0

O59:H19/H21c

2

2

0

0

0

O59/O54d:H21 2

2

0

0

0

O89:H19

1

1

0

0

0

O98:H12

4

0

0

4

0

O98:H19

1

1

0

0

0

O115:H19

1

1

0

0

0

O119:H21

1

1

0

0

0

O167:H21

1

1

0

0

0

ONTe:H19

19

19

0

0

0

ONTe:H4

7

7

0

0

0

63

Table 2.2. (cont’d)

Number of isolates positive for
Total
number of

other stx2

Serotype

isolates

stx2e

variants

stx1 variants

eae

ONTe:H21

2

2

0

0

0

ONTe:H27

1

0

1

0

0

Total

200

194

2

4

1

a

H+, flagellar gene fliC present, there were two bands for PCR.

b

These isolates were positive for H21 and H4 genes by part of the high-throughput real-time

PCR platform.
c

These isolates were positive for H19 and H21 genes by part of high-throughput real-time PCR

platform.
d

These isolates were positive with both O59 and O54 antisera.

e

ONT, O antigen non-typeable.

64

100
90
80

Proportion (%)

70
60
Cohort 1

50

Cohort 2
40

Cohort 3

30
20
10
0
10

12

14

16
18
Pig age by week

20

22

24

Figure 2.1.a. Proportion of pigs from which STEC were isolated by pig age over the finishing
period

65

25

Numbers of pigs

20

15
Cohort 1
Cohort 2

10

Cohort 3

5

0
10

12

14

16
18
Pig age by week

20

22

24

Figure 2.1.b. Frequency distribution of pigs at the age of first-time STEC isolation

66

Figure 2.2. Kaplan-Meier survival curves for duration of fecal STEC shedding in finishing
swine

67

REFERENCES

68

REFERENCES

1.

O'Brien, A.D. and G.D. LaVeck, Purification and characterization of a Shigella
dysenteriae 1-like toxin produced by Escherichia coli. Infection and Immunity, 1983.
40(2): p. 675-83.

2.

Gyles, C.L., Shiga toxin-producing Escherichia coli: an overview. J Anim Sci, 2007.
85(13 Suppl): p. E45-62.

3.

Scallan, E., et al., Foodborne Illness Acquired in the United States—Major Pathogens.
Emerging Infectious Diseases, 2011. 17(1): p. 7-15.

4.

CDC. National Shiga toxin-producing Escherichia coli (STEC) surveillance annual
summary, 2009. 2012 Sep, 10, 2012 January, 2013]; Available from:
http://www.cdc.gov/ncezid/dfwed/PDFs/national-stec-surv-summ-2009-508c.pdf.

5.

CDC, Escherichia coli O157:H7 outbreak linked to commercially distributed dry-cured
salami- Washington and California, 1994. Morbidity and Mortality Weekly Report,
1995. 44: p. 157-160.

6.

CDC, Community outbreak of hemolytic uremic syndrome attributable to Escherichia
coli O111:NM-South Australia, 1995. Morbidity and Mortality Weekly Report, 1995. 44:
p. 550-558.

7.

Paton, A.W., et al., Molecular microbiological investigation of an outbreak of hemolyticuremic syndrome caused by dry fermented sausage contaminated with Shiga-like toxinproducing Escherichia coli. J Clin Microbiol, 1996. 34(7): p. 1622-7.

8.

Williams, R.C., et al., Illness outbreak associated with Escherichia coli O157:H7 in
Genoa salami. E. coli O157:H7 Working Group. Canadian Medical Association Journal,
2000. 162(10): p. 1409-13.

9.

MacDonald, D.M., et al., Escherichia coli O157:H7 outbreak linked to salami, British
Columbia, Canada, 1999. Epidemiol Infect, 2004. 132(2): p. 283-9.

10.

Conedera, G., et al., A family outbreak of Escherichia coli O157 haemorrhagic colitis
caused by pork meat salami. Epidemiol Infect, 2007. 135(2): p. 311-4.
69

11.

Trotz-Williams, L.A., et al., Pork Implicated in a Shiga Toxin-producing Escherichia coli
O157:H7 Outbreak in Ontario, Canada. Canadian Journal of Public Health. Revue
Canadienne de Sante Publique, 2012. 103(5): p. e322-6.

12.

Fairbrother, J.M., Gyles CL, Colibacillosis, in Diseases of Swine, L.K. Jeffrey
Zimmerman, Alejandro Ramirez, Kent Schwartz,Gregory Stevenson Editor. 2012, John
Wiley and sons, Inc.: West Sussex, UK. p. 723-747.

13.

Pinaka, O., et al., Shiga toxin-producing Escherichia coli in Central Greece: prevalence
and virulence genes of O157:H7 and non-O157 in animal feces, vegetables, and humans.
European Journal of Clinical Microbiology and Infectious Diseases, 2013.

14.

Borie, C., et al., Prevalence and characterization of enterohaemorrhagic Escherichia coli
isoated from healthy cattle and pigs slaughtered in Santiago, Chile. Archivos De
Medicina Veterinaria, 1997. 29(2): p. 205-212.

15.

Bush, E., U.S. swine herd appears free of Escherichia coli O157:H7. Food Safety Digest,
1997. 4.

16.

Feder, I., et al., Isolation of Escherichia coli O157:H7 from intact colon fecal samples of
swine. Emerg Infect Dis, 2003. 9(3): p. 380-3.

17.

Fratamico, P.M., et al., Prevalence and characterization of Shiga toxin-producing
Escherichia coli in swine feces recovered in the National Animal Health Monitoring
System's Swine 2000 study. Appl Environ Microbiol, 2004. 70(12): p. 7173-8.

18.

Grant, M.A., M.A. Mogler, and D.L. Harris, Comparison of enrichment procedures for
shiga toxin-producing Escherichia coli in wastes from commercial swine farms. J Food
Prot, 2009. 72(9): p. 1982-6.

19.

Wasilenko, J.L., et al., Influence of primer sequences and DNA extraction method on
detection of non-O157 Shiga toxin-producing Escherichia coli in ground beef by realtime PCR targeting the eae, stx, and serogroup-specific genes. Journal of Food
Protection, 2012. 75(11): p. 1939-50.

20.

Hirvonen, J.J., A. Siitonen, and S.S. Kaukoranta, Usability and performance of
CHROMagar STEC medium in detection of Shiga toxin-producing Escherichia coli
strains. Journal of Clinical Microbiology, 2012. 50(11): p. 3586-90.

70

21.

Wylie, J.L., et al., Evaluation of a new chromogenic agar medium for detection of Shiga
toxin-producing Escherichia coli (STEC) and relative prevalences of O157 and nonO157 STEC in Manitoba, Canada. Journal of Clinical Microbiology, 2013. 51(2): p. 46671.

22.

Gouali, M., et al., Evaluation of CHROMagar STEC and STEC O104 chromogenic agar
media for detection of Shiga Toxin-producing Escherichia coli in stool specimens.
Journal of Clinical Microbiology, 2013. 51(3): p. 894-900.

23.

Blanco, M., et al., Genes coding for enterotoxins and verotoxins in porcine Escherichia
coli strains belonging to different O:K:H serotypes: relationship with toxic phenotypes.
Journal of Clinical Microbiology, 1997. 35(11): p. 2958-63.

24.

Bugarel, M., et al., Virulence gene profiling of enterohemorrhagic (EHEC) and
enteropathogenic (EPEC) Escherichia coli strains: a basis for molecular risk assessment
of typical and atypical EPEC strains. BMC Microbiol, 2011. 11: p. 142.

25.

Booher, S.L., N.A. Cornick, and H.W. Moon, Persistence of Escherichia coli O157:H7 in
experimentally infected swine. Vet Microbiol, 2002. 89(1): p. 69-81.

26.

Cornick, N.A. and A.F. Helgerson, Transmission and infectious dose of Escherichia coli
O157:H7 in swine. Appl Environ Microbiol, 2004. 70(9): p. 5331-5.

27.

Beutin, L., et al., Identification of human-pathogenic strains of Shiga toxin-producing
Escherichia coli from food by a combination of serotyping and molecular typing of Shiga
toxin genes. Appl Environ Microbiol, 2007. 73(15): p. 4769-75.

28.

Galli, L., et al., Virulence profile comparison between LEE-negative Shiga toxinproducing Escherichia coli (STEC) strains isolated from cattle and humans. Veterinary
Microbiology, 2010. 143(2-4): p. 307-13.

29.

Kappeli, U., et al., Human infections with non-O157 Shiga toxin-producing Escherichia
coli, Switzerland, 2000-2009. Emerg Infect Dis, 2011. 17(2): p. 180-5.

30.

Thomas, A., et al., Isolation of Vero cytotoxin-producing Escherichia coli serotypes
O9ab:H- and O101:H-carrying VT2 variant gene sequences from a patient with
haemolytic uraemic syndrome. Eur J Clin Microbiol Infect Dis, 1994. 13(12): p. 1074-6.

71

31.

Pierard, D., et al., Diarrhoea associated with Escherichia coli producing porcine oedema
disease verotoxin. Lancet, 1991. 338(8769): p. 762.

32.

Muniesa, M., et al., Characterization of a Shiga toxin 2e-converting bacteriophage from
an Escherichia coli strain of human origin. Infect Immun, 2000. 68(9): p. 4850-5.

33.

Friedrich, A.W., et al., Escherichia coli harboring Shiga toxin 2 gene variants: frequency
and association with clinical symptoms. J Infect Dis, 2002. 185(1): p. 74-84.

34.

Beutin, L., et al., Characterization of Shiga toxin-producing Escherichia coli strains
isolated from human patients in Germany over a 3-year period. J Clin Microbiol, 2004.
42(3): p. 1099-108.

35.

Fratamico, P.M., et al., Characterization of Shiga toxin-producing Escherichia coli
strains isolated from swine feces. Foodborne Pathogen and Disease, 2008. 5(6): p. 82738.

36.

Lenahan, M., et al., Molecular characterization of IrishE. coliO157:H7 isolates of
human, bovine, ovine and porcine origin. Journal of Applied Microbiology, 2009. 107(4):
p. 1340-1349.

37.

Frankel, G. and A.D. Phillips, Attaching effacing Escherichia coli and paradigms of Tirtriggered actin polymerization: getting off the pedestal. Cell Microbiol, 2008. 10(3): p.
549-56.

38.

Jordan, D.M., S.L. Booher, and H.W. Moon, Escherichia coli O157:H7 does not require
intimin to persist in pigs. Infect Immun, 2005. 73(3): p. 1865-7.

39.

Bielaszewska, M., et al., Characterisation of the Escherichia coli strain associated with
an outbreak of haemolytic uraemic syndrome in Germany, 2011: a microbiological study.
Lancet Infect Dis, 2011. 11(9): p. 671-6.

40.

Paton, A.W., et al., Characterization of Saa, a novel autoagglutinating adhesin produced
by locus of enterocyte effacement-negative Shiga-toxigenic Escherichia coli strains that
are virulent for humans. Infect Immun, 2001. 69(11): p. 6999-7009.

72

CHAPTER 3
Molecular Characterization of Shiga Toxin-Producing E. coli (STEC) from Finishing
Swine in a Longitudinal Study

This manuscript is in preparation for submission to the Journal of Applied and Environmental
Microbiology.

73

ABSTRACT

Shiga toxin-producing E. coli (STEC) infections are a critical public health concern
because of the severe clinical outcomes in humans, for example, potentially fatal hemolytic
uremic syndrome. Determining the presence or absence of virulence genes is essential in
assessing the potential pathogenecity of STEC strains. Currently, there is limited information
about the virulence genes carried by swine STEC strains. This study was conducted to examine
the presence and absence of a large panel of virulence genes in STEC strains recovered from
finishing swine in a longitudinal study. A subset of STEC strains was analyzed by pulsed field
gel electrophoresis (PFGE) to examine their genetic relatedness. Swine STEC strains (n=155)
were analyzed by a high-throughput real-time PCR array system, which included 69 virulence
gene targets. Seventeen different combinations of virulence gene profiles and serotypes were
determined in the swine STEC strains. The majority of the swine STEC strains (n=120) belonged
to serotype O59:H21, and carried the same virulence gene profile. The eae gene and two nle
variants were detected in one swine STEC strain (O49:H21). Other genes encoding adhesins
were identified, for example, the iha gene (n=154). The PFGE results revealed that swine STEC
strains from pigs raised in the same finishing barn were clonally related. This work is the critical
first step to understand the potential pathogenic mechanisms of swine STEC in human disease.

74

INTRODUCTION

Shiga toxin-producing E. coli (STEC) infections are a critical public health concern,
leading to 170,000 cases of human illness (1) and an economic burden of 280 million dollars (2)
annually in the United States. STEC represent a subset of E. coli that produce a cytotoxin known
as Shiga toxin, encoded by a phage carrying the stx gene (3, 4). The term enterohemorrhagic E.
coli (EHEC) was often used to define a subgroup of STEC strains, which are associated with
outbreaks and severe clinical cases (5). STEC infections are often acquired from consuming
contaminated food or water (6), and cattle are viewed as the most important animal reservoirs for
STEC (7). Nevertheless, many other food products, including pork products, have been
confirmed as vehicles of STEC transmission in a number of outbreaks (8-14).
Although the source of contamination of the pork products was uncertain in these
outbreaks (8-14) , the likelihood that on-farm pigs are the source of STEC contamination cannot
be overlooked. Unlike cattle which do not usually present clinical symptoms under STEC
infections, pigs, specifically post-weaning and young finishing pigs, suffer from edema disease
caused by STEC strains carrying the stx2e variant gene (15). Epidemiological studies conducted
in different regions of the world have reported wide ranging prevalences of STEC in swine
populations (16-18). However, the epidemiology and characteristics of STEC carried by on-farm
pigs and whether swine STEC strains contribute to human disease burden remain largely
unknown.
Human STEC infections are associated with a range of clinical presentations: diarrhea,
hemorrhagic colitis (HC), and the life-threatening hemolytic uremic syndrome (HUS) (5, 19).
The pathogenesis of STEC in human patients has been reviewed elsewhere (7, 20-24). Although

75

Shiga toxins are known to be crucial in STEC pathogenesis, as they inhibit host cell protein
synthesis (25) and are involved in host cell apoptosis (reviewed in (26)), virulence factors other
than Shiga toxins are associated with the capability of STEC strains to cause disease (27). Some
fimbriae are proposed to be involved in the initial attachment steps of STEC on host colonocytes,
such as the long polar fimbriae encoded by the lpf gene, which has many variants (21, 28).
Following the initial attachment, some STEC strains express intimin, encoded by eae gene,
which is located on the pathogenicity island locus of enterocyte effacement (LEE). Intimin
contributes to the formation of intimate attachment to host cells (29). For those STEC strains,
which cause HC and HUS in human patients but do not express intimin (LEE-negative STEC),
various potential adhesins have been identified, for example, the STEC autoagglutinating
adhesin encoded by the saa gene (30). A number of other virulence factors other than Stx have
been proposed to contribute to STEC pathogenesis, for instance, catalase peroxidase encoded by
katP gene (31). Moreover, some non-LEE encoded effector (nle) genes are associated with STEC
strains isolated from human patients with severe clinical diseases, namely HC or HUS (27, 32).
Knowing that various combinations of virulence factors contribute to STEC
pathogenesis, it is essential to determine the presence or absence of specific virulence genes to
better assess the potential pathogenicity of STEC strains (27, 32, 33). A few studies have
examined the virulence gene profiles of swine STEC strains (34-38). However, every study
selected different panels of virulence genes, and this fact makes it highly challenging to compare
results across different studies. The presence of many virulence genes, for example the nle genes
and their variants, has not been determined in swine STEC strains. In general, there is a
significant knowledge gap about virulence genes carried by swine STEC strains. To fill in the
current knowledge gap and better evaluate the potential pathogenicity of swine STEC strains,

76

this study was conducted to examine the presence of a large panel of virulence genes in STEC
strains recovered from finishing pigs in a longitudinal study. Moreover, the genetic relatedness
of these swine STEC strains was also examined to understand the transmission dynamics of
STEC in finishing swine.

77

MATERIALS AND METHODS

Swine STEC strains

A total of 155 STEC strains recovered from 97 finishing pigs in a longitudinal study were
included in this study. In the longitudinal study, individual fecal samples were collected from
three cohorts of finishing pigs (n=50/cohort, n=150 in total). Each cohort was raised in separated
finishing barns at two finishing sites (cohort 1 at site A, cohorts 2 and 3 at site B). The samples
were collected every two weeks during the finishing period in each cohort (8 farm visits/cohort).
A sample was considered STEC positive when an STEC isolate was recovered. The presence of
virulence genes (stx1, stx2, stx2e, eae) and O:H serotype were characterized in these STEC
isolates previously (M. Tseng, P. M. Fratamico, L. Bagi, D. Manzinger, and J. A. Funk, in press).
At least one STEC strain was selected from each positive sample for virulence gene
characterization. STEC strains in different O serogroups from the same sample were also
included in the study.

Selection of virulence gene targets

The general rationale to select virulence gene targets was based on their functions, roles
in pathogenesis, association with human illness and/or disease severity in human patients.
According to the previous characterization results, only one among the 155 swine STEC strains
carried the eae gene (M. Tseng, P. M. Fratamico, L. Bagi, D. Manzinger, and J. A. Funk, in
press). Additionally, all swine STEC strains recovered in the longitudinal study were in non-

78

O157 serogroups. Therefore, we expanded the list of virulence gene targets to include putative
genes encoding adhesins, toxins, fimbriae, and others in LEE-negative and non-O157 STEC
strains. We also included virulence genes associated with swine edema disease. The complete list
of the virulence genes (n=69) is summarized in Table 3.1.

High-throughput real-time PCR microarray

The BioMarkTM real-time PCR system (Fluidigm, USA) was used for high-throughput
microfluidic real-time PCR amplification using 96.96 dynamic arrays (Fluidigm). Amplifications
were performed using the EvaGreen DNA binding dye (Biotium Inc., Hayward, CA) with Gene
expression master mix in accordance with the recommendations of the manufacturer (Applied
Biosystems, Courtaboeuf, France). The thermal profile comprised 10 min at 95°C, followed by
35 cycles of 95°C for 15s and 60°C for 1 min, followed by a melting curve analysis.
Besides the 69 selected virulence genes, O-group associated genes O26, O157, O145,
O103, O111, O121, O45, O118, O128, O146, O91, O104, O113, and O55 were included in the
PCR microarray. Moreover, the microarray chip also contained gene targets for flagellar genes
H11, H7, H21, H2, H28, H8, H19, H16, H25, H4, and H32.

Strain selection strategy for pulsed-field gel electrophoresis (PFGE)

A subset of swine STEC strains (n=56) was selected for PFGE analysis to determine their
genetic relatedness. The selection criteria were based on serotype, virulence profiles and
epidemiologic information of the pigs. Within the predominant serotype O59:H21, we selected

79

strains recovered in the early, middle, and late stage of the finishing period. We also included
STEC strains in the same serotype and recovered from the same pig at different farm visits over
the finishing period. STEC strains in the same serotype but with different virulence gene profiles
were selected for PFGE analysis. In total, 29 O59:H21 STEC strains were selected, and two
O59:19 STEC strains were selected. Additionally 13 STEC strains in serotype O
untypeable:H19, four O98:H12 STEC strains, and one O98:H19 strain were also included in the
PFGE analysis.

PFGE

PFGE was conducted according to the standardized Centers of Disease Control and
Prevention (CDC) PulseNet protocol (39). In summary, STEC DNA was embedded in agarose
and digested with 50 U of XbaI for 2h at 37 ºC. A CHEF DR-III system (BioRad, Munich,
Germany) was used to separate the restriction fragments by electrophoresis at pulse times of
2.16-54.17 s in 0.5x Tris-borate-EDTA buffer with 50 μM thiourea at 14 ºC for 16.2h. The
H9812 Salmonella enterica serovar Braenderup strain (CDC, Atlanta, GA) was utilized as a
molecular size marker. BioNumerics software package 6.6 (Applied Maths, Ghent, Belgium)
was used to analyze the PFGE restriction-digested band patterns. The dendrogram was built by
analyzing Dice coefficients and the UPGMA method with 0.5% band position tolerance. The
genetic relatedness of the strains was assessed by the percentage similarity between the PFGE
patterns.

80

RESULTS

Virulence gene profiles of swine STEC strains

The frequency distribution of the 155 swine STEC strains by their serotype and virulence
gene profiles is summarized in Table 3.2. There were 12 virulence gene profiles within our swine
STEC strains. In combination with the serotypes, there were 17 different STEC strains by
combinations of serotypes and virulence gene profiles. Most strains (80.6%, 125/155) carried the
same virulence gene profile (virulence gene profile 1), which contained stx2e, iha, ecs1763, lpfAO113, estIa (STa), ehaA, paa, terE, and ureD. The second most prevalent virulence gene profile,
virulence gene profile 3, was found in 9.6% of the strains (15/155). This virulence gene profile 3
contained stx2e, iha, estIa (STa), paa, terE, and ureD. One strain, serotype O49:H21, carried the
eae and virulence gene profile 12, which included two nle variants: stx2e, eae, nleF, nleH1-2,
katP, iha, ecs1763, lpfA-O113, astA, estIa (STa), ecf1, ecf2, ecf3, ecf4, irp2, fyuA, ehaA, paa,
terE, and ureD.
Although one out of the 155 strains carried the eae gene, which encodes attachment
protein intimin, other attachment protein genes were present in the swine STEC strains. For
example, the iha gene, which encodes the iron regulated gene A homologue adhesin (40), was
detected in 99.4% (154/155) of the STEC strains. The lpfA –O113 gene which encodes long
polar fimibriae (41) was detected in 85.8% (133/155) of the strains. The fedA gene, which
encodes fimbrial adhesin F18 (42), was present in 0.6% (1/155) of the strains. The orfA and orfB
genes, which encode adhesins involved in diffuse adherence (AIDA) (43, 44) were present in

81

1.3% (2/155) of the strains. Moreover, the paa gene, which encodes porcine attaching and
effacing associated adhesion (45), was detected in 99.4% (154/155) of the strains.
A number of genes that encode toxins and hemolysins were present in the panel of swine
STEC strains. For instance, 151 of 155 the swine STEC strains carried the stx2e gene, and the
other 4 of the 155 strains carried the stx1 gene. The astA gene, which encodes enteroaggregative
E. coli heat stable toxin (46), was detected in 3.2% (5/155) of the strains. The estIa (STa) gene,
which encodes heat stable toxin (47), was detected in 94.8% (147/155) of the strains. Moreover,
the hlyA gene, which encodes the alpha hemolysin (48), was present in 0.6% (1/155) of the
strains. Interestingly, the ecs1763 gene, which was suggested to be associated with
enterohemorrhagic E. coli (EHEC) (49), was present in 85.2% (132/155) of the strains. The
following virulence genes were not detected in any of the 155 swine STEC strains: eae subtypes
alpha, beta, gamma, epsilon and theta, nleA, nleG5, ent/espL2, nle B, nle E, efa1/lifA, Z2096,
Z2098, Z2099, espM1, espM2, nleG6-2, espK, espN, espX7, espO1-1, espV, ecs1822, sfp, bfp,
lpfA-O26, lpfA-O157, cdt subtypes I and III, elt(LT), fasA, fimF, cnf2, ehxA, toxB, stcE, eibG,
epeA, espP, saa, subAB, and sab.

PFGE

Within the O59:H21 STEC strains, three major clusters (clusters A-C) were defined at a
cut-off value of 80% similarity (Figure 3.1). These three major clusters were related at a 71.1%
similarity. The strains isolated from pigs within the same cohort were clustered together, for
example, cluster A contained strains from pigs in cohort 3, and cluster B contained strains from
pigs in cohort 1. The only one exception was that strain number 297 from pig number 119 in

82

cohort 3 clustered with strains from cohort 2 (in cluster C). Strains from pigs in cohort 3 were
related to the strains from pigs in cohort 1 at a 75.1% similarity. Within each cluster,
indistinguishable PFGE patterns were observed among STEC strains recovered from samples in
the same pig over time during the finishing period. For example, strains 170 and 228 with
indistinguishable PFGE patterns were recovered from pig number 145 at the second and third
farm visits in cohort 3. The two O59:H19 strains, which carried virulence gene profile 3, were
not clustered with the O59:H21 strains (24.8% similarity, data not shown).
Within the O untypeable:H19 strains, 11 out of the 13 strains were clustered together at a
83.6% similarity. One of the O untypeable:H19 strain contained distinguishable PFGE patterns
from the other 11 strains (27.4% similarity). Moreover, strain 281 which carried a different
virulence gene profile was not clustered with the other 12 strains (Figure 3.2). Within the five
strains belonging to serogroup O98, the four O98:H12 strains had indistinguishable PFGE
patterns and were clustered together at a similarity over 90%. The O98:H19 strain (n=1) was not
clustered with the four O98:H12 strains (Figure 3.3).

83

DISCUSSION

The objective of this study was to use molecular methods to characterize swine STEC
strains to assess their potential pathogenicity and determine their genetic relatedness. This is the
first study to utilize a high throughput real-time PCR platform to examine the presence of a large
panel of virulence gene targets in swine STEC strains. Given the fact that these swine STEC
strains were recovered from samples of 97 healthy finishing pigs from three cohorts within 18
months, they are diverse with 16 different combinations of virulence gene profiles and serotypes.
The panel of virulence gene targets in this study included 69 targets, and our results were in
agreement with another study suggesting that increasing numbers of virulence genes in the panel
will increase the resolution of the virulence gene profiling (50). Various virulence gene profiles
in swine STEC strains have also been reported elsewhere (34-38, 51-53). However, it was
challenging to compare results of this study with these previous reports because every study
employed a different panel of virulence genes. Together, our results and the previous studies in
swine STEC strains have indicated that swine STEC consist of strains carrying diverse virulence
gene profiles.
Because the swine STEC strains used in this study belonged to non-O157 O serogroups,
and only one (O49:H21) of the 155 strains carried the eae gene, we selected many novel
virulence gene targets reported in non-O157 and LEE-negative STEC strains. Some attachment
protein genes which have been detected in human pathogenic STEC strains were present in the
swine STEC strains. For example, the iha gene and the lpfA –O113 gene were present in over
80% of the swine STEC strains. These two genes have been detected in LEE-negative STEC
strains which were associated with human clinical cases (54-56). Unlike the intimin-encoding

84

eae gene, other attachment protein genes and their roles in the pathogenesis of STEC in humans
have not been extensively studied. In addition to their potential ability to allow STEC to attach to
human cells, the high prevalence of these genes in swine STEC strains may also suggest that
they play a role in the colonization of STEC in swine population.
Interestingly, among all the swine STEC strains examined in this study, only one O
untypeable:H4 strain carried the fedA gene, which encodes fimbrial adhesin F18 and is
associated with swine edema disease (42). This finding may partially explain that all the pigs
included in the longitudinal study were clinically healthy. Moreover, the presence of the many
targets found in our panel of virulence genes has not been examined in swine STEC strains
elsewhere. For example, the ecs1763 gene, which was only found in the EHEC strains analyzed
in the previous study (49), was prevalent in a high proportion (85.2%) of the swine strains.
However, the function associated with this ecs1763 gene has not yet been characterized, and the
association between the presence of this gene and the clinical outcome in human cases requires
more research. In addition, the combination of espK with either espV, ureD, or Z2098 has been
suggested to be highly prevalent in EHEC strains, and can be utilized for identifying typical
EHEC strains which are both stx- and eae-positive (57). Although ureD was present in 99.4% of
our strains, the espK, espV, and Z2098 genes were absent in the panel of strains used in this
study.
One of the challenges in this study was that some of the primers in the real-time PCR
system used in this study may not detect genetic variation in some of the genes. For example, the
O49:H21 STEC which was positive for the eae, but was negative for the five eae variants
examined by this real-time PCR system. Therefore, misclassification of a strain as negative for
certain genes might occur when some strains carried genes with variants that the primers

85

included in this real-time PCR system were not able to target. Therefore, further validation of the
specificity or sensitivity of this PCR system is warranted. Another challenge in interpreting
virulence gene profiles is that there has not been a “gold standard” virulence gene profile
identified that accurately predicts virulence and pathogenicity of STEC strains in humans (58).
This challenge may result from multiple factors. For example, functions of many of the
associated virulence factors need further characterization, and much is unknown about the
interactions of STEC and human hosts. Notably, much of the current understanding about STEC
pathogenesis is based on STEC serotype O157:H7 and the LEE-positive STEC strains. To be
able to better interpret these virulence gene profiles and to elucidate their association with
pathogenicity, it is critical to understand the whole picture of STEC pathogenesis in humans and
colonization in both humans and their animal hosts. One of the many future directions is to use
cell culture assays to understand the interaction of these swine STEC strains with human
epithelial cell lines. For instance, it has been suggested that swine O157:H7 STEC strains affect
the integrity of human epithelial cells (59). Considering the uncertainties of some of these
putative virulence factors in causing human illness, it is difficult to determine the health risk of
many of these swine STEC strains. However, one may notice that some swine STEC strains had
only partial serotypes and most of those that were identified belonged to less common serotypes.
This is the first study to use PFGE to analyze STEC strains recovered from repeated
samples from pigs in a longitudinal study. Most of the previous studies used PFGE to determine
the genetic relatedness of STEC strains from swine to other species (53, 60-63). These results
were inconsistent, and most studies focused on STEC O157:H7 (53, 60, 62, 63). From the PFGE
results, the O59:H21 STEC strains, which were predominant in the current swine STEC
collection, were closely related to other strains from pigs within the same cohort at greater than

86

80% similarity. This may suggest that the same strain disseminated within each cohort,
supporting that there might be a point-source outbreak within each cohort which was raised on
three different barns at two finishing sites (M. Tseng, P. M. Fratamico, L. Bagi, D. Manzinger,
and J. A. Funk, in press). For example, the pigs were exposed to the same point source of
infections in the finishing site environment. This cohort-specific PFGE pattern was also observed
in a study, which was conducted in cattle (64). More research is needed to investigate potential
risk factors and the common source of infection associated with STEC strains shed by finishing
swine.
It was observed that swine STEC strains of the same serotype can carry different
virulence gene profiles. For example, there were two virulence gene profiles (1 and 2) within
O59:H21 strains. However, the strains with different virulence gene profiles were clonally
related by PFGE analysis. This was not unexpected because the two virulence gene profiles (1
and 2) were different by one gene estIa (STa), which is carried on a plasmid. Moreover, PFGE
can determine the genetic relatedness of the swine STEC strains, but cannot provide information
regarding the genetic components of the patterns (65). Future studies using sequence-based
molecular methods, for example, whole genome sequencing, can provide us more insights into
the genetic diversity of swine STEC. The results of this study help fill in the current knowledge
gap regarding swine STEC. Our results demonstrated the diverse virulence genes carried by
swine STEC and closely-related strains shed by pigs in the same cohort. However, whether these
swine STEC strains are potentially pathogenic to humans and the role swine play in the
transmission of STEC to humans require more research endeavors.

87

ACKNOWLEDGEMENTS

The work was supported by the National Pork Board (grant number 12-069) and U.S.
Department of Agriculture, National Institute of Food and Agriculture, Agriculture and Food
Research Initiative (award number 2013-67005-21189). The PCR microarray development was
partially financed by the French "joint ministerial program of R&D against CBRNE risks".
Sincerely thanks to Dr. Chitritia DebRoy, and the E. coli Reference Center at
Pennsylvania State University for assisting with characterization of the swine STEC strains,
Niesa Kettler, Terrence MacManus, and the Bacteriology Branch in Diagnostic Center for
Population and Animal Health (DCPAH) at Michigan State University for assisting with the
PFGE process.

88

APPENDIX

89

Table 3.1. Virulence gene targets, their encoded proteins and locations
virulence gene targets

proteins encoded

location

GenBank accession number

stx - all variants

Shiga toxin

chromosome

NC002695

stx2e

Shiga toxin 2e variant

chromosome

AB252836

eae

Intimin

chromosome, LEE (locus of

Z11541

enterocyte effacement)
eae-alpha
eae-beta
eae-gamma
eae-epsilon
eae-theta
iha

Iha, iron regulated gene A

chromosome, OI-43, OI-48

AF126104

homolog, adhesion
terE

tellurite resistant gene

chromosome, OI-43, OI-48

CP003301

ureD

urase transporter

chromosome, OI-43, OI-48

AE005174

espV

AvrA family effectors

chromosome, OI-44

AE005174

90

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

espK

non-LEE-encoded type III

chromosome, OI-50

AE005174

chromosome, OI-50

AE005174

chromosome, OI-50

AE005174

chromosome, OI -50

AE005174

chromosome, OI -57

AE005174

chromosome, OI -57

AE005174

secretion system effectors
espN

non-LEE-encoded type III
secretion system effectors

espX7

non-LEE-encoded type III
secretion system effectors

espO1-1

non-LEE-encoded type III
secretion system effectors

nleG5

non-LEE-encoded type III
secretion system effectors

nleG6-2

non-LEE-encoded type III
secretion system effectors

Z2096

putative gene marker for EHEC

chromosome, OI-57

AE005174

Z2098

putative gene marker for EHEC

chromosome, OI-57

AE005174

91

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

Z2099

putative gene marker for EHEC

chromosome, OI-57

AE005174

nleA

serine protease

chromosome, OI-71

AE005174

nleF

non-LEE-encoded type III

chromosome, OI-71

AE005174

secretion system effectors
nleH1-2

NleH

chromosome, OI-71

AE005174

espM1

non-LEE-encoded type III

chromosome, OI-71

AE005174

secretion system effectors
espM2

IpgM

chromosome, OI-108

AE005174

nleB

non-LEE-encoded type III

chromosome, OI-122

AE005174

chromosome, OI-122

AE005174

chromosome, OI-122

AE005174

secretion system effectors
nleE

non-LEE-encoded type III
secretion system effectors

efa1(lifA)

Efa1 EHEC factor for
adherence

92

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

ent/espL2

Enterotoxin

chromosome, OI-122

AE005174

pagC

phoP-activated gene C product

chromosome, OI-122

AE005174

(PagC)
lpfA-O157

long polar fimbriae

chromosome, OI-141, OI-154

AE005174

lpfA-O113

long polar fimbriae

chromosome, OI-154

AY057066

lpfA-O26

long polar fimbriae

chromosome

AB161111

ecs1822

hypothetical protein, putative

chromosome

BA000007

chromosome

BA000007

chromosome, Yersinia-like

CP000468

gene marker for EHEC
ecs1763

hypothetical protein, putative
gene marker for EHEC

irp2

iron responsible protein 2

high pathogenecity island
fyuA

pectin receptor

chromosome, Yersinia-like
high pathogenecity island

93

AFST01000010

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

ehaA

autotransporter, type V

chromosome

AE005174

secretion system
hlyA

alpha hemolysin

chromosome

AE014075

paa

porcine attaching and effacing

chromosome or plasmid

U82533

associated adhesion
cdtI

cytolethal distending toxin

chromosome or plasmid

AY365042

cdtIII

cytolethal distending toxin

chromosome or plasmid

AY365042

astA

enteroaggregative E. coli heat

chromosome or plasmid

L11241

stable enterotoxin
estIa (STa)

heat stable toxin

plasmid

M58746

elt (LT)

heat liable toxin

plasmid

KF733765

sfp

sorbitol-fermenting protein

plasmid

AF401292

plasmid

AB024946

fimbriae
bfp

bundle-forming pilus

94

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

fasA (F6, P987)

F6 fimbrial adhesion

plasmid

U50547

fedA (F18, F107)

F18 fimbrial adhesion

plasmid

M61713

fimF41a (F41)

F41 fimbrial adhesion

plasmid

X14354

cnf2

cytotoxic necrotizing factor

Vir plasmid

U01097

orfA

adhesin invovled in diffuse

plasmid

X65022

plasmid

X65022

plasmid O157

AF043470

plasmid O157

AF043470

plasmid O157

AF043470

adherence (AIDA)
orfB

adhesin invovled in diffuse
adherence (AIDA)

ecf1

enzyme that enhance bacterial
membrane structure

ecf2

enzyme that enhance bacterial
membrane structure

ecf3

enzyme that enhance bacterial
membrane structure

95

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

ecf4

enzyme that enhance bacterial

plasmid O157

AF043470

membrane structure
katP

catalase/peroxidase

plasmid O157

X89017

ehxA

EHEC hemolysin

plasmid O157

AF074613

toxB

potential adhesin for adherence, plasmid O157

AB011549

etpD

EHEC type II secretion system,

plasmid O157

CP001163

transporting protein across the
outer membrane
stcE

metalloprotease, mucinase

plasmid O157

AF074613

espP

serine protease autotransporter

plasmid O157

CP001163

eibG

immunoglobulin binding

plasmid O113

AB255744

plasmid O113

AY258503

protein
epeA

EpeA, serine protease

96

Table 3.1. (cont’d)
virulence gene targets

proteins encoded

location

GenBank accession number

saa

Saa, STEC agglutinating

plasmid O113

AF399919

adhesion
subAB

subtilase cytotoxin

plasmid O113

AF399919

sab

autotransporter

plasmid O113

AF399919

97

Table 3.2. Distribution of swine STEC strains by serotype and virulence gene profiles
Serotype

O59:H21

numbers of swine

numbers of strains

STEC strains

analyzed by PFGE

120

27

virulence gene profile

virulence gene
profile code

stx2e, iha, ecs1763, lpfA-O113, estIa (STa),

1

ehaA, paa, terE, ureD
2

2

stx2e, iha,ecs1763, lpfA-O113, ehaA, paa,

2

terE, ureD
3

3

stx2e, etpD, iha, ecs1763, lpfA-O113, estIa

11

(STa), ecf1, ecf2, ecf3, ecf4, ehaA, paa, terE,
ureD
O

13

12

stx2e, iha, estIa (STa), paa, terE, ureD

3

1

1

stx2e, iha, astA, estIa (STa), terE, ureD

4

2

2

stx2e, iha, estIa (STa), paa, terE, ureD

3

untypeable:H19

O59:H19

98

Table 3.2. (cont’d)
Serotype

O98:H12

numbers of

numbers of strains virulence gene profile

virulence gene

swine STEC

analyzed by

profile code

strains

PFGE

2

2

stx1, pagC, katP, iha, astA, ecf1, ecf2, ecf3,ecf4,

9

paa, terE, ureD
2

2

stx1, pagC, katP, iha, ecf1, ecf2, ecf3,ecf4, paa,

10

terE, ureD
O untypeable:H21 2

0

stx2e, iha, ecs1763, lpfA-O113, estIa (STa), ehaA,

1

paa, terE, ureD
O20:H21

1

0

stx2e, iha, ecs1763, lpfA-O113, estIa (STa), ehaA,

1

paa, terE, ureD
O49:H21

1

0

stx2e, eae, nleF, nleH1-2, katP, iha,ecs1763, lpfAO113, astA, estIa (STa), ecf1, ecf2, ecf3, ecf4, irp2,
fyuA, ehaA, paa, terE, ureD

99

12

Table 3.2. (cont’d)
Serotype

O89:H19

numbers of swine

numbers of strains

STEC strains

analyzed by PFGE

1

0

virulence gene profile

virulence gene
profile code

stx2e, iha, lpfA-O113, estIa (STa), ehaA,

7

paa, terE, ureD
O98:H19

1

1

stx2e, pagC, katP , iha, ecf1, ecf2, ecf3,

5

ecf4, paa, terE, ureD
O115:H19

1

0

stx2e, iha, ecs1763, lpfA-O113, astA, estIa

6

(STa),orfA, orfB, ehaA, paa, terE, ureD
O119:H21

1

0

stx2e, iha, ecs1763, lpfA-O113, estIa (STa),

1

ehaA, paa, terE, ureD
O167:H21

1

0

stx2e, iha, ecs1763, lpfA-O113, estIa (STa),

1

ehaA, paa, terE, ureD
O untypeable:H4

1

0

stx2e, fedA(F18), hlyA, orfA, orfB, paa,
terE

Total

155

56

100

8

A

B

C

Figure 3.1. PFGE analysis of swine O59:H21 STEC strains. Key represents strain numbers. Source represents the strain from a pig in
which cohort, which time of the eight farm visits, and the individual pig number.
101

Figure 3.2. PFGE analysis of swine O untypeable:H19 STEC strains. Key represents strain numbers. Source represents the strain
from a pig in which cohort, which time of the eight farm visits, and the individual pig number.

102

Figure 3.3. PFGE analysis of swine O98 STEC strains. Key represents strain numbers. Source represents the strain from a pig in
which cohort, which time of the eight farm visits, and the individual pig number

103

REFERENCES

104

REFERENCES

1.

Scallan, E., et al., Foodborne Illness Acquired in the United States—Major Pathogens.
Emerging Infectious Diseases, 2011. 17(1): p. 7-15.

2.

Hoffmann, S., M.B. Batz, and J.G. Morris, Annual cost of illness and quality-adjusted life
year losses in the United States due to 14 foodborne pathogens. Journal of Food
Protection, 2012. 75(7): p. 1292-1302.

3.

Unkmeir, A. and H. Schmidt, Structural analysis of phage-borne stx genes and their
flanking sequences in shiga toxin-producing Escherichia coli and Shigella dysenteriae
type 1 strains. Infection and Immunity, 2000. 68(9): p. 4856-64.

4.

O'Brien, A.D., et al., Environmental and human isolates of Vibrio cholerae and Vibrio
parahaemolyticus produce a Shigella dysenteriae 1 (Shiga)-like cytotoxin. Lancet, 1984.
1(8368): p. 77-8.

5.

Nataro, J.P. and J.B. Kaper, Diarrheagenic Escherichia coli. Clin Microbiol Rev, 1998.
11(1): p. 142-201.

6.

Rangel, J.M., et al., Epidemiology of Escherichia coli O157:H7 outbreaks, United States,
1982-2002. Emerg Infect Dis, 2005. 11(4): p. 603-9.

7.

Gyles, C.L., Shiga toxin-producing Escherichia coli: an overview. J Anim Sci, 2007.
85(13 Suppl): p. E45-62.

8.

Paton, A.W., et al., Molecular microbiological investigation of an outbreak of hemolyticuremic syndrome caused by dry fermented sausage contaminated with Shiga-like toxinproducing Escherichia coli. J Clin Microbiol, 1996. 34(7): p. 1622-7.

9.

CDC, Escherichia coli O157:H7 outbreak linked to commercially distributed dry-cured
salami- Washington and California, 1994. Morbidity and Mortality Weekly Report,
1995. 44: p. 157-160.

10.

CDC, Community outbreak of hemolytic uremic syndrome attributable to Escherichia
coli O111:NM-South Australia, 1995. Morbidity and Mortality Weekly Report, 1995. 44:
p. 550-558.
105

11.

Williams, R.C., et al., Illness outbreak associated with Escherichia coli O157:H7 in
Genoa salami. E. coli O157:H7 Working Group. Canadian Medical Association Journal,
2000. 162(10): p. 1409-13.

12.

MacDonald, D.M., et al., Escherichia coli O157:H7 outbreak linked to salami, British
Columbia, Canada, 1999. Epidemiol Infect, 2004. 132(2): p. 283-9.

13.

Conedera, G., et al., A family outbreak of Escherichia coli O157 haemorrhagic colitis
caused by pork meat salami. Epidemiol Infect, 2007. 135(2): p. 311-4.

14.

Trotz-Williams, L.A., et al., Pork Implicated in a Shiga Toxin-producing Escherichia coli
O157:H7 Outbreak in Ontario, Canada. Canadian Journal of Public Health. Revue
Canadienne de Sante Publique, 2012. 103(5): p. e322-6.

15.

Fairbrother, J.M., Gyles CL, Colibacillosis, in Diseases of Swine, L.K. Jeffrey
Zimmerman, Alejandro Ramirez, Kent Schwartz,Gregory Stevenson Editor. 2012, John
Wiley and sons, Inc.: West Sussex, UK. p. 723-747.

16.

Fratamico, P.M., et al., Prevalence and characterization of Shiga toxin-producing
Escherichia coli in swine feces recovered in the National Animal Health Monitoring
System's Swine 2000 study. Appl Environ Microbiol, 2004. 70(12): p. 7173-8.

17.

Oporto, B., et al., Escherichia coli O157:H7 and non-O157 Shiga toxin-producing E. coli
in healthy cattle, sheep and swine herds in Northern Spain. Zoonoses Public Health,
2008. 55(2): p. 73-81.

18.

Borie, C., et al., Prevalence and characterization of enterohaemorrhagic Escherichia coli
isoated from healthy cattle and pigs slaughtered in Santiago, Chile. Archivos De
Medicina Veterinaria, 1997. 29(2): p. 205-212.

19.

Karmali, M.A., et al., Escherichia coli cytotoxin, haemolytic-uraemic syndrome, and
haemorrhagic colitis. Lancet, 1983. 2(8362): p. 1299-1300.

20.

Kaper, J.B., J.P. Nataro, and H.L. Mobley, Pathogenic Escherichia coli. Nat Rev
Microbiol, 2004. 2(2): p. 123-40.

21.

Croxen, M.A. and B.B. Finlay, Molecular mechanisms of Escherichia coli pathogenicity.
Nat Rev Microbiol, 2010. 8(1): p. 26-38.

106

22.

Melton-Celsa, A., et al., Pathogenesis of Shiga-toxin producing escherichia coli. Curr
Top Microbiol Immunol, 2012. 357: p. 67-103.

23.

Petruzziello-Pellegrini, T.N., M. Moslemi-Naeini, and P.A. Marsden, New insights into
Shiga toxin-mediated endothelial dysfunction in hemolytic uremic syndrome. Virulence,
2013. 4(6): p. 556-63.

24.

Croxen, M.A., et al., Recent Advances in Understanding Enteric Pathogenic Escherichia
coli. Clinical Microbiology Reviews, 2013. 26(4): p. 822-80.

25.

Endo, Y., et al., Site of action of a Vero toxin (VT2) from Escherichia coli O157:H7 and
of Shiga toxin on eukaryotic ribosomes. RNA N-glycosidase activity of the toxins.
European Journal of Biochemistry, 1988. 171(1-2): p. 45-50.

26.

Johannes, L. and W. Romer, Shiga toxins--from cell biology to biomedical applications.
Nat Rev Microbiol, 2010. 8(2): p. 105-16.

27.

Bugarel, M., et al., Virulence gene profiling of enterohemorrhagic (EHEC) and
enteropathogenic (EPEC) Escherichia coli strains: a basis for molecular risk assessment
of typical and atypical EPEC strains. BMC Microbiol, 2011. 11: p. 142.

28.

Low, A.S., et al., Analysis of fimbrial gene clusters and their expression in
enterohaemorrhagic Escherichia coli O157:H7. Environ Microbiol, 2006. 8(6): p. 103347.

29.

Frankel, G. and A.D. Phillips, Attaching effacing Escherichia coli and paradigms of Tirtriggered actin polymerization: getting off the pedestal. Cell Microbiol, 2008. 10(3): p.
549-56.

30.

Paton, A.W., et al., Characterization of Saa, a novel autoagglutinating adhesin produced
by locus of enterocyte effacement-negative Shiga-toxigenic Escherichia coli strains that
are virulent for humans. Infect Immun, 2001. 69(11): p. 6999-7009.

31.

Brunder, W., H. Schmidt, and H. Karch, KatP, a novel catalase-peroxidase encoded by
the large plasmid of enterohaemorrhagic Escherichia coli O157:H7. Microbiology,
1996. 142 ( Pt 11): p. 3305-15.

107

32.

Coombes, B.K., et al., Molecular analysis as an aid to assess the public health risk of
non-O157 Shiga toxin-producing Escherichia coli strains. Appl Environ Microbiol, 2008.
74(7): p. 2153-60.

33.

Slanec, T., et al., Molecular analysis of virulence profiles and Shiga toxin genes in foodborne Shiga toxin-producing Escherichia coli. Appl Environ Microbiol, 2009. 75(19): p.
6187-97.

34.

Fratamico, P.M., et al., Characterization of Shiga toxin-producing Escherichia coli
strains isolated from swine feces. Foodborne Pathogen and Disease, 2008. 5(6): p. 82738.

35.

Meng, Q., et al., Characterization of Shiga toxin-producing Escherichia coli isolated
from healthy pigs in China. BMC Microbiol, 2014. 14: p. 5.

36.

Chapman, T.A., et al., Comparison of virulence gene profiles of Escherichia coli strains
isolated from healthy and diarrheic swine. Appl Environ Microbiol, 2006. 72(7): p. 478295.

37.

Zweifel, C., et al., Virulence profiles of Shiga toxin 2e-producing Escherichia coli
isolated from healthy pig at slaughter. Vet Microbiol, 2006. 117(2-4): p. 328-32.

38.

Sonntag, A.K., et al., Shiga toxin 2e-producing Escherichia coli isolates from humans
and pigs differ in their virulence profiles and interactions with intestinal epithelial cells.
Appl Environ Microbiol, 2005. 71(12): p. 8855-63.

39.

Ribot, E.M., et al., Standardization of pulsed-field gel electrophoresis protocols for the
subtyping of Escherichia coli O157:H7, Salmonella, and Shigella for PulseNet.
Foodborne Pathog Dis, 2006. 3(1): p. 59-67.

40.

Tarr, P.I., et al., Iha: a novel Escherichia coli O157:H7 adherence-conferring molecule
encoded on a recently acquired chromosomal island of conserved structure. Infect
Immun, 2000. 68(3): p. 1400-7.

41.

Doughty, S., et al., Identification of a novel fimbrial gene cluster related to long polar
fimbriae in locus of enterocyte effacement-negative strains of enterohemorrhagic
Escherichia coli. Infect Immun, 2002. 70(12): p. 6761-9.

108

42.

Frydendahl, K., Prevalence of serogroups and virulence genes in Escherichia coli
associated with postweaning diarrhoea and edema disease in pigs and a comparison of
diagnostic approaches. Vet Microbiol, 2002. 85(2): p. 169-82.

43.

Benz, I. and M.A. Schmidt, Isolation and serologic characterization of AIDA-I, the
adhesin mediating the diffuse adherence phenotype of the diarrhea-associated
Escherichia coli strain 2787 (O126:H27). Infection and Immunity, 1992. 60(1): p. 13-8.

44.

Benz, I. and M.A. Schmidt, Cloning and expression of an adhesin (AIDA-I) involved in
diffuse adherence of enteropathogenic Escherichia coli. Infection and Immunity, 1989.
57(5): p. 1506-11.

45.

An, H., et al., Distribution of a novel locus called Paa (porcine attaching and effacing
associated) among enteric Escherichia coli. Advances in Experimental Medicine and
Biology, 1999. 473: p. 179-84.

46.

Osek, J., Detection of the enteroaggregative Escherichia coli heat-stable enterotoxin 1
(EAST1) gene and its relationship with fimbrial and enterotoxin markers in E. coli
isolates from pigs with diarrhoea. Veterinary Microbiology, 2003. 91(1): p. 65-72.

47.

Blanco, M., et al., Genes coding for enterotoxins and verotoxins in porcine Escherichia
coli strains belonging to different O:K:H serotypes: relationship with toxic phenotypes.
Journal of Clinical Microbiology, 1997. 35(11): p. 2958-63.

48.

Welch, R.A., et al., The synthesis and function of the Escherichia coli hemolysin and
related RTX exotoxins. FEMS Microbiology Immunology, 1992. 5(1-3): p. 29-36.

49.

Abu-Ali, G.S., et al., Genomic diversity of pathogenic Escherichia coli of the EHEC 2
clonal complex. BMC Genomics, 2009. 10: p. 296.

50.

Brandt, S.M., et al., Molecular risk assessment and epidemiological typing of Shiga
toxin-producing Escherichia coli by using a novel PCR binary typing system. Appl
Environ Microbiol, 2011. 77(7): p. 2458-70.

51.

Kaufmann, M., et al., Escherichia coli O157 and non-O157 Shiga toxin-producing
Escherichia coli in fecal samples of finished pigs at slaughter in Switzerland. J Food Prot,
2006. 69(2): p. 260-6.

109

52.

Ateba, C.N. and M. Mbewe, Detection of Escherichia coli O157:H7 virulence genes in
isolates from beef, pork, water, human and animal species in the northwest province,
South Africa: public health implications. Res Microbiol, 2011. 162(3): p. 240-248.

53.

Lenahan, M., et al., Molecular characterization of Irish E. coli O157:H7 isolates of
human, bovine, ovine and porcine origin. Journal of Applied Microbiology, 2009. 107(4):
p. 1340-1349.

54.

Newton, H.J., et al., Shiga toxin-producing Escherichia coli strains negative for locus of
enterocyte effacement. Emerging Infectious Diseases, 2009. 15(3): p. 372-80.

55.

Hauser, E., et al., Phylogenetic and molecular analysis of food-borne shiga toxinproducing Escherichia coli. Applied and Environmental Microbiology, 2013. 79(8): p.
2731-40.

56.

Galli, L., et al., Virulence profile comparison between LEE-negative Shiga toxinproducing Escherichia coli (STEC) strains isolated from cattle and humans. Veterinary
Microbiology, 2010. 143(2-4): p. 307-13.

57.

Delannoy, S., L. Beutin, and P. Fach, Discrimination of enterohemorrhagic Escherichia
coli (EHEC) from non-EHEC strains based on detection of various combinations of type
III effector genes. Journal of Clinical Microbiology, 2013. 51(10): p. 3257-62.

58.

Anjum, M.F., et al., Use of virulence determinants and seropathotypes to distinguish
high- and low-risk Escherichia coli O157 and non-O157 isolates from Europe.
Epidemiology and Infection, 2013: p. 1-10.

59.

Kennedy, C.A., et al., Shiga toxin-producing Escherichia coli isolated from human and
pig origin induce different gene expression profiles in human Caco-2 epithelial cells.
Livestock Science, 2010. 133(1/3): p. 189-191.

60.

Johnsen, G., et al., Escherichia coli O157:H7 in faeces from cattle, sheep and pigs in the
southwest part of Norway during 1998 and 1999. Int J Food Microbiol, 2001. 65(3): p.
193-200.

61.

Franke, S., et al., Clonal relatedness of Shiga-like toxin-producing Escherichia coli O101
strains of human and porcine origin. J Clin Microbiol, 1995. 33(12): p. 3174-8.

110

62.

Osek, J. and P. Gallien, Molecular analysis of Escherichia coli O157 strains isolated
from cattle and pigs by the use of PCR and pulsed-field gel electrophoresis methods.
Veterinarni Medicina, 2002. 47(6): p. 149-158.

63.

Rios, M., et al., Clonal diversity of Chilean isolates of enterohemorrhagic Escherichia
coli from patients with hemolytic-uremic syndrome, asymptomatic subjects, animal
reservoirs, and food products. J Clin Microbiol, 1999. 37(3): p. 778-81.

64.

Joris, M.A., et al., Longitudinal follow-up of the persistence and dissemination of EHEC
on cattle farms in Belgium. Foodborne Pathog Dis, 2013. 10(4): p. 295-301.

65.

Singer, R.S., W.M. Sischo, and T.E. Carpenter, Exploration of biases that affect the
interpretation of restriction fragment patterns produced by pulsed-field gel
electrophoresis. J Clin Microbiol, 2004. 42(12): p. 5502-11.

111

CHAPTER 4
Shiga toxin-producing E. coli (STEC) cases in Michigan, U.S., 2001-2012

112

ABSTRACT

Infection with Shiga toxin-producing Escherichia coli (STEC) by serotypes other than
O157 (non-O157) have been increasingly reported as the cause of human illness in the U.S. This
increase in reporting is likely primarily due to the improvements in diagnostic tests. To better
understand the demographic characteristics of STEC cases and the odds of hospitalization
associated with STEC infections, specifically non-O157 STEC, this study reviewed STEC cases
reported to the Michigan Department of Community Health (MDCH) through the Michigan
Disease Surveillance System (MDSS) from 2001 through 2012. An increasing trend of nonO157 STEC cases was observed in this 12-year period, and the incidence rates were similar for
O157 and non-O157 STEC cases in 2012. No demographic characteristics were significantly
different between O157 and non-O157 STEC cases. However, the odds of hospitalization were
2.36 times higher in O157 STEC cases than in non-O157 STEC cases when adjusted for age and
gender. The information enhances our understanding in epidemiology of non-O157 STEC in
Michigan, and future research is warranted to understand these pathogens in order to improve
prevention and control efforts.

113

INTRODUCTION

Shiga toxin-producing Escherichia coli (STEC) represent a subset of E. coli strains that
are capable of producing one or more Shiga toxins (Stx). STEC infections, mostly acquired from
consuming contaminated food or water, are associated with outbreaks and sporadic cases of
diarrhea, hemorrhagic colitis (HC), and hemolytic uremic syndrome (HUS) (1, 2). Among the
HC cases, 3-9 % of sporadic cases to 20 % of epidemic cases develop the life-threatening HUS
(3-5), which is one of the leading causes of acute renal failure in young children worldwide (3, 6,
7). Therefore, STEC infections are a critical public health concern, and more than 170,000
illnesses are attributed to STEC in the U.S. annually (8).
While STEC strains belonging to serotype O157 are more common among outbreaks and
severe clinical cases (9), infections with non-O157 STEC strains have been increasingly
documented (9-12). Non-O157 STEC infections have been viewed as a public health concern.
Since the year 2000, non-O157 STEC has become a nationally notifiable disease in the U.S. (13).
Six non-O157 STEC serogroups, namely O26, O103, O111, O121, O45, O145, are the most
prevalent in human infections and thus, are viewed as a major public health concern. Recently,
these six O serogroups joined O157 in their classification as adulterants in raw, non-intact beef
products in the U.S. (14).
The lower reported rates of non-O157 STEC cases in the past may in part be explained by
the challenges of non-O157 STEC detection methods. Over time the Center for Disease Control
and Prevention (CDC) has updated recommendations for clinical laboratory practices to better
detect non-O157 STEC infections (15-17). According to the CDC’s FoodNet, which is an active
surveillance system for laboratory-confirmed STEC cases in 10 states in the U.S., increasing

114

detection of non-O157 STEC infections has been associated with the improvement of laboratory
detection protocols (18). Similar observations were also reported in studies conducted in states
not part of the FoodNet, for example, in Washington (19) and a surveillance study conducted in
20 clinical laboratories in Michigan (20).
The association of non-O157 STEC with milder clinical symptoms may also contribute to
the lower reported rates of non-O157 STEC (21). Having milder clinical symptoms decreases the
chance of seeking medical attention, and this leads to a smaller chance of having the specimen
collected and tested. Therefore, these cases with milder clinical symptoms are less likely to be
diagnosed and reported to the public health agencies (22). Compared with O157 STEC cases,
non-O157 STEC infections are less likely to lead to severe clinical symptoms, specifically
bloody diarrhea, HUS, and resulting in hospitalization (11, 21, 23). However, it was estimated
that more than 60% of STEC infections are attributed to non-O157 STEC in the U.S. (8).
Besides clinical symptoms, similar demographic characteristics associated with nonO157 STEC cases have been described in a number of studies. In general, more female cases
were observed in both O157 STEC and non O-157 STEC infections (13, 20, 21, 23, 24). The
age distribution was similar between O157 STEC cases and non-O157 STEC cases in the CDC’s
FoodNet data (13). In contrast, one study suggested that non-O157 STEC infections were more
commonly identified in children younger than 5 years old (24). Relative to O157 STEC
infections, however, little is known about the epidemiology of non-O157 STEC infections in
human clinical cases.
Current disease control and prevention plans can be improved by enhancing our
understanding of host characteristics associated with STEC infections and disease trends over
time. Because Michigan is not included in the CDC’s FoodNet records, this study was conducted

115

to examine the epidemiology of STEC cases reported to the Michigan Department of
Community Health (MDCH) over a 12-year period. Trends associated with STEC case were
examined per year, as were the demographic and clinical data associated with each STEC case.

116

MATERIAL AND METHODS

Data source

The STEC case information was retrieved from the Michigan Disease Surveillance
System (MDSS), managed by the MDCH (25). MDSS is a web-based surveillance system in
which healthcare providers and clinical laboratories report communicable diseases to the public
health system. Within 24 hours of STEC identification, the case must be reported to the local
health department via MDSS. The MDCH is simultaneously notified via the MDSS when the
case is reported to the local health department. The MDCH then reports the case to the CDC. All
STEC serotypes are reportable, and an STEC isolate (if available) or non-culture positive broth
and stool sample must be submitted to MDCH Bureau of Laboratories for isolation, confirmation
and genotyping (26).
The MDSS has been designed according to U.S. national data standards. An electronic
case investigation form includes the following information (if available): patient contact
information, demographic, hospitalization status, clinical symptoms, laboratory data (e.g.
serotype, virulence gene profiles, Shiga toxin information), collection date, length of illness,
outbreak status, and notes containing additional clinical and/or follow-up information, for
example, updated STEC isolate serotyping results.

117

Study design

The following inclusion criteria were followed when collecting the case reports from
2001 through 2012 in the MDSS. A case of STEC was designated under the reportable
conditions “Escherichia coli O157:H7”, “Shiga toxin E. coli, unspecified”, “Shiga toxin E. coli,
non-O157”, and “Shiga toxin-producing E. coli (STEC)”. Notably, the reportable conditions of
STEC infections in Michigan have undergone updates during the past decade. Before 2010, the
three terms “E. coli O157:H7”, “STEC non-O157”, and “STEC unspecified” were used.
Beginning in 2010, these three terms were consolidated into one reportable condition called
“STEC”. Subgrouping of STEC by serotypes was not included under the reportable condition. In
addition to the reportable conditions, the case inclusion criteria also included case status,
investigation status, and dates of cases reported to the MDCH. Cases were included in the study
if case status was “confirmed”, the investigation status was “completed”, and the report dates
were from 01/01/2001 until 12/31/2012.
The demographic, clinical, laboratory, and epidemiologic data were extracted from each
case record. Basic demographic information was documented, including age at the onset of
disease, gender, race, ethnicity, and date of case was reported. The clinical information included
hospitalization, death, HUS status, bloody diarrhea, and non-bloody diarrhea. Presence of other
clinical symptoms were also documented, which included fever, chills, abdominal pain, fatigue,
body aches, nausea, vomiting, and other health conditions. The laboratory information collected
included STEC serotypes and the Shiga toxin (Stx) gene profile. Epidemiologic investigation
information included food history (up to seven-day food history), drinking water source (e.g.

118

municipal, well), animal contact, travel history within one month before the disease onset, and
swimming within one month before the disease onset.

Data management and analysis

O157 STEC cases represented all cases with serotype O157 infections, while non-O157
STEC cases were defined as cases infected with all serotypes other than O157. If serotype
information was not noted in the case report (stated “unknown”), then the serotype was coded as
missing data. The same was true for Shiga toxin type (Stx1 or Stx2) information stated
“unknown”. Because the symptom onset date or specimen submission date was not available for
every case, the “year” and “month” of each case was defined according to the date the case was
reported. The season was categorized based on the date of case report: winter (December,
January, February), spring (March, April, May), summer (June, July, August), and fall
(September, October, November).
Additional data were retrieved from the case reports but were not analyzed further. In
general, if the information was not reported in over 50% of the cases, then the information was
not included in the analysis. For example, race and ethnicity were “unknown” in 50% of all the
cases. Similarly, some clinical symptom data were not analyzed because the definition of the
symptoms was unclear in the case reports. The timeline was not available in the reports to
distinguish whether the symptoms occurred before the STEC diagnosis or hospitalization, during
the hospitalization, or after the STEC diagnosis or discharge from the hospital. These symptoms
included fever, chills, abdominal pain, fatigue, body aches, nausea, vomiting, and other health
conditions.

119

The data in the epidemiologic information section were not analyzed further because of
the following limitations. First, some of the information was sparse and thus, it was challenging
to categorize the information entered as descriptive text (e.g. food history and drinking water
source). Second, some of the data were missing in up to 60% of the cases (e.g. animal contact).
Lastly, it was difficult to specify the time duration of the potential exposures to the onset of the
symptoms (i.e., the potential incubation period). For example, any travel history and swimming
information occurred within one month before the disease onset was documented in the case
report, but the exact time of the travel and swimming was not available in every case. Therefore,
travel history and swimming information could not be accurately evaluated in the analysis.
The annual age-adjusted incidence rates (case per 100,000 population) of STEC cases
were computed based on the population estimates of Michigan from the Bridged-Race
Population Estimates 1990-2012 dataset (27). The standard population was based on the U.S.
2000 standard population (28, 29). The data management was performed in both Microsoft Excel
(Microsoft Corporation, Redmond, WA, U.S.) and SAS 9.3 (SAS Institute, Cary, NC, U.S.).
Chi-square tests were performed to assess the association between STEC O serotypes and
the clinical outcomes. The statistical analyses were performed in SAS 9.3. P-values <0.05 were
considered to be significant in all statistical tests unless stated otherwise. For assessing the
association between characteristics of the reported STEC cases and the clinical outcomes,
hospitalization status was used as an indicator for a severe clinical outcome associated with
STEC infection. The dependent variable was hospitalization status of the STEC cases (yes: the
STEC cases who were hospitalized; no: the STEC cases who were not hospitalized). The
variables assessed were STEC serotypes (O157; non-O157), Shiga toxin types (stx2-only; stx1only or both stx1 and stx2), age groups (<5; 5-10; 11-59; >=60), gender (female; male), and

120

season of report (fall and winter; spring and summer). Univariate analysis was performed, and
any independent variable with a p-value <0.25 and considered to be biologically plausible was
selected as a candidate for inclusion in a multivariate model using logistic regression. The
variable selection method followed the “purposeful selection” steps described by Hosmer and
Lemeshow (30). A variable was kept in the model when the p-value was less than 0.1 and
removal of this variable caused the values of other variables’ effect sizes to change >10%. The
univariate analysis and the multivariate analyses were performed using PROC GENMOD
function in SAS 9.3.

121

RESULTS

Descriptive epidemiology of STEC cases

From 2001 to 2012, a total of 1,497 confirmed STEC cases were reported to the MDCH.
The highest numbers of STEC cases were reported in 2008 (n=186) and 2012 (n=189), and the
lowest was in 2006 (n=88) (Figure 4.1.a). Among these 1,497 STEC cases, 138 (9.2%) cases
were confirmed to be associated with outbreaks (n=29). The highest number of outbreak
associated cases was 54 in 2012 followed by 47 in 2008. In other years, the outbreak-associated
case numbers ranged from 0 between 2001 and 2003 to 12 in 2010. Overall, the age-adjusted
incidence rates of STEC cases remained relatively constant before 2007 (ranging from
0.86/100,000 in 2006 to 1.31/100,000 in 2002) (Figure 4.2). The incidence rate peaked for all
STEC cases (1.9/100,000) and O157 STEC cases (1.23/100,000) in 2008. No non-O157 STEC
cases were reported in 2001. Notably, since 2002, there has been an increasing trend of incidence
rate of non-O157 STEC infection. The year 2010 was the first time that incidence of non-O157
STEC cases was greater than O157 STEC cases in Michigan. For all years combined, among
O157 STEC cases, the peak in the number of cases (20.4%, 183/895) occurred in September,
while non-O157 STEC cases peaked (14.6%, 58/398) in August (Figure 4.3.a.b).
Among all the 1497 cases, STEC O serotypes were not available in 13.6% (203/1497) of
the cases. O157 was the most commonly reported serotype among all the STEC cases (69.2%,
895/1497) followed by O45 (11.3%, 146/1497), O103 (5.5%, 71/1497), O26 (4.6%, 60/1497),
O111 (2.6%, 33/1497), O145 (1.8%, 23/1497), and O121 (0.8%, 11/1497) (Table 4.1). The
remainder of the cases were infected with STEC belonging to 35 additional serotypes. Moreover,

122

the Shiga toxin gene profiles were available for 1294 (86%) cases. Within the 792 O157 STEC
cases with data available, 12 (1.5%) had stx1, 284 (35.9%) had stx2, and the remaining 496
(63%) had both stx1 and stx2. The distribution varied among the 399 non-O157 STEC cases as
most (n=321; 80%) had stx1, followed by stx2 (n=46; 12%), and both stx1/stx2 (n=25; 6%).
Examining the demographic data demonstrated that there were more female STEC cases
versus male, which did not vary by serotype, except in O145 cases (Table 4.1). The overall
median age in these STEC cases was 22, ranging from 1 day old to 102 years old. The lowest
median age by serogroup was 17 years old (0.92-66) in O111 STEC cases, and the highest
median age was 27 years old (0.33-85) in cases infected with STEC in O serogroups other than
O157, O45, O103, O26, O111, O145, and O121. The age distribution of O157 STEC and nonO157 STEC cases was similar (Figure 4.4.a.b).
Frequencies of clinical outcomes (hospitalization, HUS, bloody diarrhea, and non-bloody
diarrhea) in the STEC cases varied by serotype (Table 4.2). Two out of the 1497 (0.1%) cases
died within the time of epidemiologic investigation and both were infected with O157 STEC.
The proportion of cases hospitalized ranged from 14.1% in O103 STEC cases to 48.2% in O157
STEC cases (Chi-square p<0.05). The highest proportion of cases hospitalized was in cases older
than 85 years old (100% for both O157 STEC cases and non-O157 STEC cases), and the lowest
proportion was in cases younger than 5 years old (7.5% in O157 STEC cases and 25% in nonO157 STEC cases) (Figure 4.5). HUS was reported in 26 (1.7%) out of the 1497 cases, and 22
(84.6%) of the 26 HUS cases were infected with O157 STEC. The proportion of cases that
developed bloody diarrhea ranged from 32.7% in STEC cases in non-O157 serotypes other than
the six most prevalent non-O157 serotypes to 74.6% in O157 STEC cases (Chi-square p<0.05).
The proportion of cases with non-bloody diarrhea ranged from 13% in O145 STEC cases to

123

54.5% in cases infected with STEC in non-O157 serotypes other than the six most prevalent nonO157 serotypes (Chi-square p<0.05).
Twenty six cases reported co-infection with more than one STEC serotype or another
pathogen; most (n=11; 42.3%) cases were co-infected with Cryptosporidium or Camplybacter
(n=7; 26.9%). One case was co-infected with Clostridium difficile, and another was co-infected
with Klebsiella pneumoniae. One case was co-infected with Chlamydia and Campylobacter,
another was co-infected with Salmonella, O157 STEC, and O145 STEC, and another was coinfected with O91:NM STEC and O157 STEC. Cases with co-infections were excluded from the
statistical analysis as it was not clear which pathogen contributed to the clinical data highlighted
in the MDSS records.

The odds of hospitalization in STEC cases

All together, a total of 880 STEC cases were analyzed to identify covariates associated
with hospitalization; 617 cases were excluded from the analysis due to missing data for the
outcome variable (hospitalization status) and/or other covariates. All covariates but gender met
the initial screening significance level of p<0.25 in the univariate analysis (Table 3.3). Among
these covariates, the greatest univariate effect sizes were observed in O serotypes and age. The
odds of hospitalization were 2.3 times higher in O157 STEC cases when compared with nonO157 STEC cases (95% confidence interval (CI): 1.73, 3.09; p<0.0001). Moreover, the odds of
hospitalization were in cases aged 60 years or older when compared with cases aged 11 to 59
years (95% CI: 1.79, 3.98; p<0.0001).

124

Because of reporting convention, we elected to force gender in the multivariate analysis.
The final model included STEC O serotype (O157; non-O157), age, and gender (Table 4.4).
Shiga toxin gene profile and season were removed because they were not significantly associated
with hospitalization in the multivariate model. The odds of hospitalization in O157 STEC cases
were 2.36 times higher than in non-O157 STEC cases (95% CI:1.76, 3.18; p<0.0001). The odds
of hospitalization in cases aged 60 years or older old were 2.61 times higher than cases aged 11
to 59 years (95% CI:1.73, 3.93; p<0.0001). Cases aged younger than 5 years had odds of
hospitalization 0.62 times lower than cases aged 11 to 59 years (95% CI:0.39, 0.98; p<0.0411).

125

DISCUSSION

This study describes STEC cases reported to the MDCH from 2001 through 2012. From
both the numbers of cases and the age-adjusted incidence rates, there is a clear increasing trend
of non-O157 STEC cases in Michigan over this 12-year period. This increasing trend is likely
associated with the improvement of diagnostic tests for non-O157 STEC nationally (17, 31) and
in Michigan (20). A peak in the age-adjusted incidence rate was observed among O157 STEC
cases in 2008, and this peak may in part be attributed to a large number of outbreak-associated
cases (n=47 of a total of 138 outbreak-associated cases) in 2008. The year 2010 was the first time
that incidence of non-O157 STEC cases was greater than O157 STEC cases in Michigan. In
2012, the age-adjusted incidence rates were similar between O157 (0.93/100,000) and non-O157
STEC (0.94/100,000) cases. This agrees with the estimation by Scallan et al. that more than 50%
of total the STEC infections are attributed to non-O157 STEC (8). The CDC have updated the
recommendations for detecting non-O157 STEC cases over time (15-17), and the current
recommendations are to simultaneously test all stool samples from acute community-acquired
diarrhea for STEC O157 by selective and differential media and non-O157 STEC by Shiga toxin
assay (17). Testing in parallel or increased diagnostic effort clearly improved sensitivity of
detection of STEC infections, and it is essential for the clinical laboratories to continuously
follow the current diagnostic recommendations from CDC.
Among all the non-O157 serotypes, O45 was the most reported O serotype, followed by
O103, O26, O111, O145, and O121 in Michigan. In the CDC’s FoodNet data, O26 was the most
commonly reported non-O157 STEC, followed by O103, O111, O121, O45, and O145 (13).
Moreover, O111 and O26 were responsible for 66% of the non-O157 STEC outbreaks reported

126

to the CDC up to 2010 (32). Although the frequency of O serotypes differed in Michigan
compared with the FoodNet sites, these six O types predominated in all sites. The differential
ranking based on frequency of cases associated by serotype in Michigan may indicate differential
risk of exposures regionally in Michigan as compared to FoodNet sites. Other observations in
Michigan are consistent with results from FoodNet sites. The demographic characteristic (age
and gender) were similar between O157 STEC cases and non-O157 STEC cases in Michigan,
and these agree with previous studies in the FoodNet sites (13). However, there were some mild
differences. In Michigan, the median age (in years) reported per O serotype ranged from 17 for
O111 STEC to 27 for all other non-O157 STEC. This median age range was older than what
was reported in the FoodNet reports which, ranged from 9 for O26 to 29 years of age for O121
(13). Although it was difficult to pinpoint factors that contributed to these discrepancies, the
number and characteristics of different outbreaks may play a role (13). Future studies are
therefore warranted to identify risk factors associated with different non-O157 STEC serotypes.
These data may assist in public health interventions through identifying risk factors that may be
differential by serotype.
The odds of hospitalization were higher in O157 STEC cases than non-O157 STEC cases
in Michigan, which is consistent with data from Minnesota (21) and Connecticut (23). Similarly,
one German study has also reported that the risk of hospitalization in O157 STEC cases was
equal to or greater than 2 times that in non-O157 STEC cases, except the O104 serotype (33).
Interestingly, our model suggested that children younger than 5 years old were associated with
lower odds of hospitalization when compared with cases aged 11 to 59 years. The similar age
association was reported in the surveillance study in Michigan (20) and one German study (33).
This observation suggests other risk factors associated with younger age may affect the risk of

127

hospitalization, but there is limited information regarding age and odds of hospitalization. Age
effect to the risk of STEC infection has been reported elsewhere. In contrast to most reports
suggesting similar age distributions between O157 and non-O157 STEC (13), both more likely
affecting the young and the old, one study reported that non-O157 STEC infections were more
commonly detected in children younger than 5 years old when compared with O157 STEC (24).
Age-specific risk factors associated with STEC infections have been reported. For example,
eating hamburgers was associated with STEC infections in people aged younger than 12 years
old (34). Another study suggested than contact with ruminants was associated with the highest
odds of STEC infection in children aged younger than 3 years (35). Whether age associated
severity of disease (represented by the proxy measure of hospitalization in this study) is related
to differential exposures or a biological component of age warrants further investigation as it
may allow for more targeted and cost efficient public health interventions.
The cases included in this study are reported to the MDCH through the public health
surveillance system, and they represent a subset of the actual total STEC infections in Michigan.
For instance, it is likely that some people infected with STEC developed milder clinical
symptoms; and therefore did not seek medical care and were never reported to the public health
department (36). Although O157 STEC has been associated with a more severe clinical outcome
as suggested by the results of this study, the degree of under-reporting of O157 STEC cases has
been estimated to be 20-fold (37). Moreover, the challenges of diagnosing non-O157 STEC in
clinical laboratories are widely described (16). A 106.8 fold degree of under-diagnosis was
estimated for non-O157 STEC cases, and 26.1 fold of which for O157 STEC cases (8). As a
result, the cases reported to the MDCH are an underestimated number of actual STEC incidence
in Michigan due to the effect of under-reporting or under-diagnosis. In addition, public health

128

surveillance systems are not designed for research, but for monitoring trend of diseases,
identifying outbreaks and assessing the impact of public health prevention and control measures
to disease frequencies (22, 36, 38). Therefore, there are inherent limitations when it comes to the
use of surveillance data for risk factor analysis. For example, some data have more than 50% of
missing values, e.g. animal contact information. Despite these limitations, reviewing surveillance
cases may reveal at-risk populations as these cases may be valuable for understanding STEC
cases with more severe clinical outcomes or individuals with higher risk of infection. They are
certainly a biased representation of all STEC cases and therefore likely overestimate the
magnitude of effect of the risk factors evaluated in this study.
Notably, non-O157 STEC contains STEC in heterogeneous serotypes. Some reports have
suggested that STEC in different serotypes have different epidemiologic characteristics. For
example, O111, O103 and O26 were more common among cases reporting history of
international travel (13, 23). Treating a heterogeneous group of bacteria (non-O157 STEC) as a
single serotype category limits our abilities to fully understand risk factors for disease, and
therefore may prevent more effective interventions. Future studies are needed to examine the
epidemiologic characteristics in different O serotypes, for example, the age effect to the severity
of illness in different O types or differential risk of exposure to certain serotypes geographically
in Michigan.
Whether increasing in actual incidence or through an increase in our ability to diagnose
cases, non-O157 STEC, a heterogeneous and not fully understood group of pathogens, represent
a significant public health concern in Michigan (as in the U.S.) and is not only in disease
estimates but in cases reported likely to surpass O157:H7. Although non-O157 STEC were less
likely to result in hospitalization, the increasing incidence of this heterogeneous group of

129

pathogens demonstrates a significant public health cost. Investment in research to more fully
elucidate risk factors for STEC infection, including serotype and more complete human risk
factor data, are needed.

130

ACKNOWLEDGEMENTS

Sincere thanks to Dr. James Rudrik, James Collins and Tiffany Henderson of the
Michigan Department of Community Health for assisting with collection of epidemiologic and
laboratory data as well as the National Institues of Health Enterics Research Investigational
Network (ERIN) Cooperative Research Center at Michigan State University [grant number
U19AI090872 (SDM)] and in part by the United State Department of Agriculture, National
Institute of Food and Agriculture [grant number 2011-67005-30004(SMD)].

131

APPENDIX

132

Table 4.1. Demographic characteristics of STEC cases by STEC O serotypes, Michigan, 2001-2012
O serotypesb
Total no. of

O157

O45

O103

O26

O111

O145

O121

Othersd

cases (%)

(n=895,

(n=146,

(n=71,

(n=60,

(n=33,

(n=23,

(n=11,

(n=55,

69.2%c)

11.3%)

5.5%)

4.6%)

2.6%)

1.8%)

0.8%)

4.3)

Age (year)a
<5

181(12.1)

108(12d)

13(8.9)

12(16.9)

6(1)

9(27.3)

0

0

8(14.5)

5-10

161(10.8)

102(11.3)

14(9.6)

5(7)

5(8.3)

4(12.1)

1(4.3)

3(27.3)

6(10.9)

11-69

1050(70.1)

627(70.1)

110(75.3)

52(73.2)

46(76.7)

20(60.6)

20(87)

8(72.7)

34(61.8)

>70

105(7)

58(6.5)

9(6.2)

2(2.8)

3(5)

0

2(8.7)

0

7(12.7)

Male

691(46.5)

421(47)

61(41.8)

33(46.5)

21(35)

15(45.5)

13(56.6)

3(27.3)

26(42.3)

Female

795(53.5)

469(52.4)

85(58.2)

38(53.5)

38(63.3)

16(48.5)

10(43.5)

8(72.7)

28(50.9)

Gendera

a

Gender was not available in n=11 cases.

b

Serotype was not available in 13.6% (203/1497) cases. O157 includes O157, O157:H7, O157:NM (non motile), O157:H rough,

O157:H undetermined, O157:H unknown. O45 in cludes O45:H undetermined, O45:H2. O103 includes O103, O103:H11, O103:H2,
O103:H25. O26 includes O26:H11, O26:NM. O111 includes O111, O111:H8, O111:NM. O145 includes O145, O145:NM. O121
133

Table 4.1. (cont’d)
includes O121, O121:H19, O121:H7, O121:H9. Others include the following serotypes: O rough:H2, O rough:NM, O
undertermined:H undertermined, O undertermined:H11, O undertermined:H19, O undetermine:NM, O undetermined:H11, O
undetermined:H25, O undetermined:H52, O undetermined:NM ,O104:H4, O123:H11, O128:H25, O128:NM, O123:H11,O128:NM,
O147:H7, O156:H25, O156:NM, O163:H19, O165:NM, O168:H8, O174:H21, O177:NM, O185:H28, O186:H11,O186:H2, O39:H49,
O49:NM, O5:NM, O69:H11, O71:H11, O71:NM, O76:H19, O91:NM, two cases with two STEC strains isolated (one with O157,
O145; one with O91:NM, O157).
c

% represents the % of total no. of cases with O serogroup information (n=1294)

d

% represents the % of total no. of cases within each O serogroup.

134

Table 4.2. Clinical outcomes of STEC cases by STEC O serotypes, Michigan, 2001-2012
Hospitalization

HUS

Bloody diarrhea

Non-bloody
diarrhea

O serotypesa

Total

No. of cases (%b)

No. of cases (%)

No. of cases (%)

No. of cases (%)

O157

895

432(48.2)

22(2.5)

668(74.6)

145(16.2)

O45

146

60(41.1)

0(0)

94(64.3)

42(28.7)

O103

71

10(14.1)

1(1.4)

31(43.6)

31(43.7)

O26

60

13(21.6)

0

34(56.7)

20(33.3)

O111

33

12(36.3)

0

13(39.4)

11(33.3)

O145

23

7(30.4)

0

19(82.6)

3(13)

O121

11

5(45.5)

0

6(54.5)

4(36.4)

Others

55

18(32.7)

1(1.8)

18(32.7)

30(54.5)

Unknown

203

54(26.6)

2(1)

67(33)

44(21.6)

Total

1497

611(40.8)

26(1.7)

956(63.9)

330(22)

a

Serotype information was not available in 13.6% (203/1497) cases (O serogroup “unknown” ). O157 includes O157, O157:H7,

O157:NM (non motile), O157:H rough, O157:H undetermined, O157:H unknown. O45 in cludes O45:H undetermined, O45:H2.
O103 includes O103, O103:H11, O103:H2, O103:H25. O26 includes O26:H11, O26:NM. O111 includes O111, O111:H8, O111:NM.
135

Table 4.2. (cont’d)
O145 includes O145, O145:NM. O121 includes O121, O121:H19, O121:H7, O121:H9. Others include the following serotypes: O
rough:H2, O rough:NM, O undertermined:H undertermined, O undertermined:H11, O undertermined:H19, O undetermine:NM, O
undetermined:H11, O undetermined:H25, O undetermined:H52, O undetermined:NM ,O104:H4, O123:H11, O128:H25, O128:NM,
O123:H11,O128:NM, O147:H7, O156:H25, O156:NM, O163:H19, O165:NM, O168:H8, O174:H21, O177:NM, O185:H28,
O186:H11,O186:H2, O39:H49, O49:NM, O5:NM, O69:H11, O71:H11, O71:NM, O76:H19, O91:NM, two cases with two STEC
strains isolated (one with O157, O145; one with O91:NM, O157).
b

c

% of total cases within each O serogroup.

Hospitalization status was missing in 9.9% (148/1497) case reports. HUS status was stated “unknown” in 10.8% (162/1497) case

reports. Death status was stated “unknown” in 0.5% (8/1497) case reports. Bloody diarrhea and diarrhea status was stated “unknown”
in 11% (165/1497) case reports.

136

Table 4.3. Univariate analysis of characteristics associated with hospitalization in STEC cases, Michigan, 2001-2012
Characteristic

No. with

No (%)

ORa

95% CIb

p-valuec

2.3

(1.73-3.09)

<0.0001

1.5

(1.15-2.07)

0.0041

characteristic hospitalized
O serogroup
O157

574

299(52.1)

Non-O157

306

98(32)

stx2-only

248

131(52.8)

stx1-only, stx1 and stx2

632

266(42.1)

(reference)
Shiga toxin profiles

(reference)

137

Table 4.3. (cont’d)
ORa

95% CIb

p-valuec

32(33)

0.64

(0.49-1.2)

0.247

97

36(37.1)

0.77

(0.41-1.01)

0.056

11-59 (reference)

555

241(43.4)

>=60

131

88 (67.2)

2.67

(1.79-3.98)

<0.0001

Female

475

222(46.7)

1.15

(0.883-1.506)

0.2948

Male (reference)

405

175(43.2)

Characteristic

No. with

No. (%)

characteristic

hospitalized

<5

97

5-10

Age groups

Gender

138

Table 4.3. (cont’d)
Characteristic

No. with

No. (%)

characteristic

hospitalized

Fall and winter

422

203(48.1)

Spring and summer

458

194(42.4)

ORa

95% CIb

p-valuec

1.26

(0.97-1.65)

0.0872

Season

(reference)
a

Odds ratio

b

95% confidence interval

c

Walds Chi-square test

139

Table 4.4. Multivariate logistic regression model for characteristics associated with hospitalization in STEC cases, Michigan, 20012012
Parameter

Regression

Standard error

ORa

95% CIb

p-valuec

coefficient
Intercept

-0.8798

0.1566

<0.0001

O157 (reference: non-O157)

0.8601

0.152

2.36

(1.76-3.18)

<0.0001

Female (reference: male)

0.1093

0.152

1.12

(0.84-1.47)

0.4423

Age <5

-0.4838

0.2368

0.62

(0.39-0.98)

0.0411

Age 5-10

-0.3425

0.2325

0.71

(0.45-1.12)

0.1407

Age >60

0.9602

0.2098

2.61

(1.73-3.93)

<0.0001

(Age reference: 11-59)
a

Odds ratio

b

95% confidence interval

c

Walds Chi-square test

140

140

120

Numbers of STEC cases

100

80
O157 STEC
60

non-O157 STEC

40

20

0
2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Year

Figure 4.1.a. Numbers of O157 STEC cases and non-O157 STEC cases, Michigan, 2001-2012 (representing total n=1294; excluding
n=203 with no serotype information)

141

100

Proportions of STEC cases (%)

90

80
70
60
50

O157 STEC

40

non-O157 STEC

30
20

10
0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
a

Proportions of O157 (non-O157) STEC cases out of all STEC cases reported in each year

Figure 4.1.b. Proportionsa of O157 STEC cases and non-O157 STEC cases, Michigan, 2001-2012 (representing total n=1294;
excluding n=203 with no serotype information)

142

2

1.8

Incedence per 100,000 population

1.6
1.4
1.2
O157 STEC

1

non-O157 STEC
0.8
0.6
0.4
0.2

0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Figure 4.2. Age-adjusted incidence rates of STEC cases per 100,000 population by year, Michigan, 2001-2012 (representing total
n=1294; excluding n=203 with no serotype information)

143

25

Proportions of STEC cases (%)

20

15
O157 STEC
10

non-O157 STEC

5

0
Jan

Feb

Mar

Apr

May

Jun
Jul
Month

Aug

Sep

Oct

Nov

Dec

Figure 4.3.a. Numbers of STEC cases by month of report, Michigan, 2001-2012 (representing total n=1294; excluding n=203 with no
serotype information)

144

25

Proportions of STEC cases (%)

20

15
O157 STEC

10

non-O157 STEC

5

0
Jan

a

Feb

Mar

Apr

May

Jun
Jul
Month

Aug

Sep

Oct

Nov

Dec

Proportions of O157 (non-O157) STEC reported in each month out of all O157 STEC (non-O157) cases 2001-2012

Figure 4.3.b. Proportionsa of STEC cases by month of report, Michigan, 2001-2012 (representing total n=1294; excluding n=203 with
no serotype information)

145

120

Numbers of STEC cases

100

80

60
O157 STEC
non-O157 STEC

40

20

0

Age

Figure 4.4.a. Numbers of O157 STEC cases and non-O157 STEC by age groups, Michigan, 2001-2012 (representing total n=1294;
excluding n=203 with no serotype information)

146

25

Proportions of STEC cases (%)

20

15

O157 STEC
10

non-O157 STEC

5

0

Age
a

Proportions of numbers of O157 STEC (non-O157) cases in each age group out of all O157 STEC (non-O157) cases.

Figure 4.4.b. Proportionsa of O157 STEC cases and non-O157 STEC by age groups, Michigan, 2001-2012 (representing total n=1294;
excluding n=203 with no serotype information)

147

100

Proportions of hospitalized STEC cases (%)

90
80
70

60
50
O157 STEC

40

non-O157 STEC

30
20
10
0

Age

Figure 4.5. Proportion of hospitalized STEC cases by age group and serotype, Michigan, 2001-2012 (representing total n=1294;
excluding n=203 with no serotype information)

148

REFERENCES

149

REFERENCES

1.

Nataro, J.P. and J.B. Kaper, Diarrheagenic Escherichia coli. Clin Microbiol Rev, 1998.
11(1): p. 142-201.

2.

Karmali, M.A., et al., The association between idiopathic hemolytic uremic syndrome
and infection by verotoxin-producing Escherichia coli. J Infect Dis, 1985. 151(5): p. 77582.

3.

Tarr, P.I., C.A. Gordon, and W.L. Chandler, Shiga-toxin-producing Escherichia coli and
haemolytic uraemic syndrome. The Lancet, 2005. 365(9464): p. 1073-1086.

4.

Zoja, C., S. Buelli, and M. Morigi, Shiga toxin-associated hemolytic uremic syndrome:
pathophysiology of endothelial dysfunction. Pediatr Nephrol, 2010. 25(11): p. 2231-40.

5.

Noris, M. and G. Remuzzi, Hemolytic uremic syndrome. J Am Soc Nephrol, 2005. 16(4):
p. 1035-50.

6.

Karmali, M.A., V. Gannon, and J.M. Sargeant, Verocytotoxin-producing Escherichia coli
(VTEC). Vet Microbiol, 2010. 140(3-4): p. 360-70.

7.

Williams, D.M., et al., Acute kidney failure: a pediatric experience over 20 years. Arch
Pediatr Adolesc Med, 2002. 156(9): p. 893-900.

8.

Scallan, E., et al., Foodborne Illness Acquired in the United States—Major Pathogens.
Emerging Infectious Diseases, 2011. 17(1): p. 7-15.

9.

Gyles, C.L., Shiga toxin-producing Escherichia coli: an overview. J Anim Sci, 2007.
85(13 Suppl): p. E45-62.

10.

Henderson, H., Direct and indirect zoonotic transmission of Shiga toxin-producing
Escherichia coli. Journal of the American Veterinary Medical Assocciation, 2008.
232(6): p. 848-59.

11.

Law, D., Virulence factors of Escherichia coli O157 and other Shiga toxin-producing E.
coli. J Appl Microbiol, 2000. 88(5): p. 729-45.

150

12.

Bettelheim, K.A., The non-O157 Shiga-toxigenic (verocytotoxigenic) Escherichia coli;
under-rated pathogens. Crit Rev Microbiol, 2007. 33(1): p. 67-87.

13.

Gould, L.H., et al., Increased recognition of non-O157 Shiga toxin-producing
Escherichia coli infections in the United States during 2000-2010: epidemiologic features
and comparison with E. coli O157 infections. Foodborne Pathog Dis, 2013. 10(5): p. 45360.

14.

FSIS, Food Safety and Inspection Service. Shiga toxin-producing Escherichia coli in
certain raw beef products. Federal Register, 2011. 76: p. 58157-58165.

15.

CDC, Importance of culture confirmation of Shiga toxin-producing Escherichia coli
infection as illustrated by outbreaks of gastroenteritis---New York and North Carolina,
2005. Morbidity and Mortality Weekly Report, 2006. 55(38): p. 1042-1045.

16.

CDC, Recommendations for diagnosis of Shiga toxin-producing Escherichia coli
infections by clinical laboratories. Morbidity and Mortality Weekly Report, 2009.
58(RR-12): p. 1-14.

17.

Gould, L.H., Update: Recommendations for Diagnosis of Shiga Toxin-Producing
Escherichia coli Infections by Clinical Laboratories. Clinical Microbiology Newsletter,
2012. 34(10): p. 75-83.

18.

Hoefer, D., et al., Laboratory practices for the identification of Shiga toxin-producing
Escherichia coli in the United States, FoodNet sites, 2007. Foodborne Pathog Dis, 2011.
8(4): p. 555-60.

19.

Stigi, K.A., et al., Laboratory practices and incidence of non-O157 shiga toxinproducing Escherichia coli infections. Emerging Infectious Diseases, 2012. 18(3): p. 4779.

20.

Manning, S.D., et al., Surveillance for Shiga toxin-producing Escherichia coli, Michigan,
2001-2005. Emerging Infectious Diseases, 2007. 13(2): p. 318-21.

21.

Hedican, Erin B., et al., Characteristics of O157 versus Non‐O157 Shiga Toxin–
ProducingEscherichia coliInfections in Minnesota, 2000–2006. Clinical Infectious
Diseases, 2009. 49(3): p. 358-364.

151

22.

Tauxe, R.V., Real burden and potential risks from foodborne infections: the value of
multi-jurisdictional collaborations. Trends in Food Science & Technology, 2008. 19: p.
S18-S25.

23.

Hadler, J.L., et al., Ten-year trends and risk factors for non-O157 Shiga toxin-producing
Escherichia coli found through Shiga toxin testing, Connecticut, 2000-2009. Clinical
Infectious Diseases, 2011. 53(3): p. 269-76.

24.

Lathrop, S., K. Edge, and J. Bareta, Shiga toxin-producing Escherichia coli, New Mexico,
USA, 2004-2007. Emerging Infectious Diseases, 2009. 15(8): p. 1289-91.

25.

MDCH. [cited 2013, 2014; Available from: http://www.michigan.gov/mdch/0,4612,7132-2945_5104_31274---,00.html.

26.

MDCH (2014) Reportable diseases in Michigan-2014.

27.

United States Department of Health and Human Services (US DHHS), C.f.D.C.a.P. and
N.C.f.H.S.N. (CDC). Bridged-Race Population Estimates, United States July 1st resident
population by state, county, age, sex, bridged-race, and Hispanic origin. 2013 13 June
2013 [cited 2014 06 March]; Compiled from 1990-1999 bridged-race intercensal
population estimates (released by NCHS on 7/26/2004); revised bridged-race 2000-2009
intercensal population estimates (released by NCHS on 10/26/2012); and bridged-race
Vintage 2012 (2010-2012) postcensal population estimates (released by NCHS on
6/13/2013). :[Available from: http://wonder.cdc.gov/bridged-race-v2012.html

28.

Anderson, R.N. and H.M. Rosenberg, Age standardization of death rates: implementation
of the year 2000 standard. National Vital Statistics Reports, 1998. 47(3): p. 1-16, 20.

29.

Klein RJ, S.C., Age Adjustment Using the 2000 Projected U.S. Population. . 2001,
National Center for Health Statistics: Hyattsville, Maryland.

30.

David W. Hosmer, S.L., Model-building strategies and methods for logistic regression,
in Applied Logistic Regression 2000, John Wiley & Sons, Inc: Hoboken, NJ. p. 91-142.

31.

CDC (2013) National Shiga toxin-producing Escherichia coli (STEC) surveillance
annual report, 2011.

32.

Luna-Gierke, R.E., et al., Outbreaks of non-O157 Shiga toxin-producing Escherichia coli
infection: USA. Epidemiology and Infection, 2014: p. 1-11.
152

33.

Preussel, K., et al., Shiga toxin-producing Escherichia coli O157 is more likely to lead to
hospitalization and death than non-O157 serogroups--except O104. PLoS One, 2013.
8(11): p. e78180.

34.

Voetsch, A.C., et al., Risk factors for sporadic Shiga toxin-producing Escherichia coli
O157 infections in FoodNet sites, 1999-2000. Epidemiology and Infection, 2007. 135(6):
p. 993-1000.

35.

Werber, D., et al., Shiga toxin-producing Escherichia coli infection in Germany: different
risk factors for different age groups. American Journal of Epidemiology, 2007. 165(4): p.
425-34.

36.

Tauxe, R.V., et al., Evolving public health approaches to the global challenge of
foodborne infections. International Journal of Food Microbiology, 2010. 139 Suppl 1: p.
S16-28.

37.

Mead, P.S., et al., Food-related illness and death in the United States. Emerging
Infectious Diseases, 1999. 5(5): p. 607-25.

38.

Crim, S.M., et al., Incidence and trends of infection with pathogens transmitted
commonly through food - foodborne diseases active surveillance network, 10 u.s. Sites,
2006-2013. MMWR. Morbidity and Mortality Weekly Report, 2014. 63(15): p. 328-32.

153