.3
. .Ir 3459.1!‘1- .«

.3... 2.. 3...... ,. ..
.33 kifiaflgfifflm
a. s... dkw‘dflukficfi.‘
:L . .._ u.

a um
I .10 l

55$»

3.

. I:
11.5.01.“
. 2‘: . 1 13......va
‘» 3.1.. . tr... I
39.. 5 433......
.5 . 1...?
REC... a... l
1“: .
. 3.4.51

wmhm...
: .

i!» .
:EL-

.1
11.5 urine
1.} t. .1... .
hEh... ..

 

:- I51-
: ll. v.0
0:11.181" .
.1»... a

_

(it .
i.?1u:..i:. Gaunt-51!.

3.0.19.6...” .1“. .

nfifild
5...}

 

. :i 51.9.5.5.
. M94 fold... “I"...
9- lp.4r.|>~l¢}

?.I .;

2......
.3 2.
$253....
:5 :3 21 3..
3.8....
3. ‘nflg‘d It.

.vl)V!\~.’.v
Luau}... xi..- . ,
3-.‘u (:1:- VI"- .
.5... «59......» “RR

.. .. 5.1.x?
. f; s
a

”.13

 

 

I ;
£1

 

 

 

 

.z.

:.

.. , . .T .3... .. . . .mwzmmcwsau . ,I. .. , .
. . , , . _ Mfwffiry a.» 3 4 .. .
rush...m . . . .... I . .

 

 

 

This is to certify that the
dissertation entitled

RISK FACTORS FOR SPORADIC NON-TYPHOIDAL
SALMONELLA INFECTIONS IN MICHIGAN CHILDREN: A
POPULATION-BASED CASE-CONTROL STUDY

presented by

 

MUHAMMAD YOUNUS

I]...

has been accepted towards fulfillment
of the requirements for the

 

.=-38RARY
Michlgan State

Ph.D degree in Epidemiology

 

 

 

F“
L

642.2% gaa/

 

Major Professor’s Signature

v/c—er 0f-

Date

 

MSU is an affinnative-action, equal-opportunity employer

 

 

 

I-_!4_-.-I-".?!.-l-I-I-I-l-O—I-O-I-O-I-O-'—I-I-O-Uu-O-I-I-I-D-l-'—' -

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K.IProj/Acc&Pres/ClRC/DateDue,indd

 

RISK FACTORS FOR SPORADIC NON-TYPHOIDAL SALMONELLA
INFECTIONS IN MICHIGAN CHILDREN: A POPULATION-BASED CASE-
CONTROL STUDY

By

Muhammad Younus

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Epidemiology

2008

ABSTRACT
RISK FACTORS FOR SPORADIC NON-TYPHOIDAL SALMONELLA INFECTIONS
IN MICHIGAN CHILDREN: A POPULATION-BASED CASE-CONTROL STUDY
By
Muhammad Younus

Epidemiologic investigations have consistently shown higher incidences of laboratory-
confirmed Salmonella infections (salmonellosis) in children compared to adults. Our
recent work investigating associations between demographic attributes and salmonellosis
based on data from the Michigan Department of Community Health (MDCH) (I 995-
2001) revealed about an approximate 10-fold increase in risk for acquiring salmonellosis
in children aged < 1 year and a 3-fold increase in risk for those aged 1-4 years when
compared to adults aged 15-39 years. The majority (80 - 85%) of sporadic cases of
salmonellosis in adult populations results from exposures to contaminated foods.
However, few analytical studies have addressed the role of contaminated environmental
exposures, which have been suspected in a large proportion of cases of Salmonella
infections in children. We conducted a population-based case-control study of sporadic
cases of non-typhoidal Salmonella infections in Michigan children aged _<_ 10 years to
identify various food vehicles and environmental exposures associated with illness in this
high-risk population. Laboratory-continued cases of Salmonella infections in children
aged 5 10 years reported to MDCH, and healthy control children who did not experience
symptoms of gastrointestinal illness during the past month, were recruited between
December 15, 2006 and October 15, 2007. Controls were obtained using an on-Iine

telephone directory. A pre-tested structured questionnaire, administered through trained

interviewers or self-administered by mail-in questionnaire, was used to gather data from
parent(s) or caretakers. Information was collected on sociodemographic characteristics
of children, child rearing (e. g., daycare, pre-school, elementary school attendance etc),
and various environmental exposures (e.g., contact with animals, contact with a person
having symptoms of gastrointestinal illness). A total of 123' cases and 139 controls were
enrolled during the study period. The final multivariate model, after adjusting for age
group revealed that having salmonellosis was significantly associated with contact with
cats (adjusted odds ratio (AOR) = 2.62, 95% Cl: 1.17 — 5.87) and reptiles (AOR = 8.16,
95% CI: 1.55 — 42.88). Additionally, attending a daycare center (AOR = 4.86, 95% CI:
1.44 — 16.37) and contact with a person having symptoms of gastrointestinal infection
during the 3 days prior to the onset of child’s illness was significantly associated with
Salmonella infections (AOR = 2.27, 95% CI: 1.02 — 5.44). Salmonellosis was not
associated with exposures to other household- and food-related sources. Our study
results suggest that environmental sources significantly contribute to the acquisition of
Salmonella infections in children. This is in contrast to the adult population where a
larger proportion of infections are acquired through food vehicles. Several public health
recommendations have been made to educate parents and caretakers about the risk of
Salmonella transmission to children from infected persons and animals, particularly
reptiles. However, our study demonstrated that exposure to these factors continue to
cause Salmonella infections in children. Additional efforts are needed to educate parents
and caretaker about the risk of Salmonella transmission to children from cats and reptiles,

along with the risk of 2% transmission following exposure to symptomatic individuals.

ACKNOWLEDGEMENTS

I am indebted to my guidance committee members, their contribution made this work
possible. A special thank to my major advisor, Dr. Mahdi Saeed, for his continuous
intellectual advisement, encouragement, and financial support. He is a great mentor,

colleague and friend.

Many thanks to my Ph.D. guidance committee members, Drs. Melinda Wilkins, H. Dele
Davies, M. Hossein Rahbar and Julie Funk for their mentorship. Their suggestions and
feedback while conducting various parts of my research and writing my dissertation were

very helpful.

I gratefully acknowledge the staff of Communicable Disease Division, Bureau of
Epidemiology, Michigan Department of Community Health for their support. In
particular, I would like to thank Paula J ager and Nazneen Syed (Administrative
Assistants, Communicable Disease Division) for providing study related logistical

support at MDCH.

My special thanks to Chau Nguyen, Diane Sinawi, Chiko Obi, Symone Coleman
(Undergraduate students, MSU) for their efforts in data collection, data entry, and

managing study logistics at MSU.

Many friends and colleagues at Michigan State University provided substantial advice,
support and encouragement during this project: Mokhtar Arshad, Amy Steffey, Nicole
Crisp, Kristin Evon, Alyssa DiFilippo, Laya Kevan, , Seongbeom Cho, Nasr Aref , Azfar

Siddiqi, and Edward Hartwick.

iv

Finally, I am thankful to my family for their continuous prayers, encouragement, and

support towards achieving this important milestone of my career.

 

XX

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................. ix
LIST OF FIGURES ................................................................................................ xi
LIST OF ABBREVIATIONS ............................................................................... xii
BACKGROUND ....................................................................................................... 1
Foodbome infections ................................................................................ 1
Salmonella ................................................................................................. 3
Pathogenesis .............................................................................................. 4
Transmission of Salmonella infections ..................................................... 4
Clinical manifestations .............................................................................. 5
Salmonellosis: Burden of the disease ....................................................... 6
Outbreaks of foodbome Salmonella infections ....................................... 11
Selected outbreaks of salmonellosis ....................................................... 13
o Multi-state outbreak associated with contaminated milk,
1985 .................................................................................... 13
o Multi-state Schwan’s ice cream outbreak, 1994 ................ l3
0 Michigan bakery product outbreak, 2002 .......................... 14
o Multi-state raw almond outbreak, 2004 ............................. 14
0 South Carolina Restaurant Outbreak, 2005 ....................... 14
o Multi-state Orchid Island Juice outbreak, 2005.. .. ...1...5
- Multi-state tomato associated outbreak, 2006 ................... 15
o Multi—state peanut butter outbreak 2007 ............................ 16
Selected common Salmonella serotypes ................................................. 16
o Salmonella enterica serotype Typhimirum ........................ 18
o Salmonella enterica serotype Enteritidis ............................ 20
c Salmonella enterica serotype Newport .............................. 22
o Salmonella enterica serotype Heidelberg ........................... 23
Salmonella infections in children ............................................................ 24
STUDY OBJECTIVES ........................................................................................... 28
MATERIAL AND METHODS ............................................................................. 29
Study population ...................................................................... 29
Source of data .......................................................................... 29
Study design ............................................................................. 30
Cases ........................................................................................ 32
Controls .................................................................................... 33
Enrollment of cases and controls by age-groups ..................... 33
Questionnaire development ..................................................... 34

vi

Data collection method ............................................................ 35

Data collection process ............................................................ 37
ETHICAL CONSIDERATIONS .................................................................................... 38
Informed consent ..................................................................... 38
Data confidentiality .................................................................. 38
Approval from the IRB for human research ........................... 38
SAMPLE SIZE CALCULATIONS ................................................................................ 39
DATA MANAGEMENT ................................................................................................. 40
Variable description and transformation .................................. 40
STATISTICAL ANALYSES .......................................................................................... 45
Descriptive statistics ................................................................ 45
Inferential statistics .................................................................. 45
Univariate analysis ....................................................... 45
Multivariate analysis .................................................... 45
Calculation of Population Attributable Risk ........................... 46
Subgroup analyses ................................................................... 47
Analysis-1 .................................................................... 47
Analysis-2 .................................................................... 48
Analysis-3 .................................................................... 48
Analysis-4 .................................................................... 48
RESULTS ......................................................................................................................... 49
Case enrolhnent ....................................................................... 49
Comparison of case participants and non-participants ............ 50
Distribution of Salmonella serotypes among case children ..... 50
Comparison of Salmonella serotypes between cases
aged <10 years and cases aged >11 years ................................ 51
Control enrollment ................................................................... 5]
Descriptive statistics ................................................................ 52
Inferential statistics .................................................................. 53
Univariate analysis ................................................................... 53
Multivariate analysis ................................................................ 55
Population Attributable Risk for selected variables ................. 56
Subgroup analyses ................................................................... 56

vii

Analysis-1 ............................................................................. 56

Analysis-2 ............................................................................. 57
Analysis-3 ............................................................................. 57
Analysis-4 ............................................................................. 57
DISCUSSION ........................................................................................................ 59
Validity of findings ............................................................... 59
Criteria for causation ............................................................. 60
Public health recommendations ............................................ 75
Study strengths ...................................................................... 82
Study limitations ................................................................... 85
CONCLUSIONS ..................................................................................................... 90
FUNDING SOURCE ............................................................................................. 90
TABLES ........................................................................................................ 92
FIGURES ...................................................................................................... 116
APPENDIX .. ....................................................................................................... 123
Case-control study invitation letter ..................................... 124
Case-control study information sheet and consent form ..... 127
Case-control study questionnaire ........................................ 131
REFERENCES ...................................................................................................... 145

viii

LIST OF TABLES

Tablel. Comparison of the 2006 incidence of infections with major enteric pathogens and

the US National Health Objective 2010 ........................................................... 92
Table 2. Selected large foodbome outbreaks where Salmonella serotypes were identified as
etiologic agents (1974-2007) ....................................................................... 93
Table 3. The twenty most common Salmonella serotypes in descending order of isolation
frequency in the US .................................................................................. 95
Table 4. Examples of Salmonella serotypes by host adaptation .............................. 96

Table 5. Percent change (2004 vs. 1996-1998) in the incidence of reported cases of four
most common Salmonella serotypes under surveillance at FoodNet sites .................... 97

Table 6. Distribution of multi-drug resistant (MDR) Salmonella Typhimurium and
definitive phage type 104 (DT104) strains in selected countries, 1992—2001 ................ 98

Table 7. Identified risk factors for Salmonella infections in children (1985-2007). ....... 99

Table 8. Comparison of demographic characteristics between enrolled case children
(participants) and non-enrolled case children (non-participants) in Salmonella case-control
study, Michigan, 2007. ............................................................................. 100

Table 9. Distribution of Salmonella serotypes in case children aged 5 10 years and 2
llyears reported to MDCH during the study period (Dec 15, 2006-October 15, 2007). ..101

Table 10. Socioeconomic characteristics of children aged 5 10 years enrolled in a
population-based case-control study to identify risk factors for Salmonella infections,
Michigan, 2007. .................................................................................... 103

Table 11. Univariate analyses of putative risk factors for Salmonella infections in
children aged 510 years, assessed in a population-based case- control study, Michigan,
2007. ................................................................................................. 104

Table 12. Multivariate analysis of putative risk factors for Salmonella infections in
children aged 310 years, assessed in a population-based case- control study, Michigan,
2007. .................................................................................................. 106

Table 13. Multivariate analysis of selected risk factors for Salmonella infections by age

groups in children population aged 510 years, assessed in a population-based case- control
study, Michigan, 2007 (Subgroup analysis-1). ................................................ 107

ix

Table 14. Assessment of selected putative risk factors for Salmonella infections in children
aged < 1 year, assessed in a population-based case- control study, Michigan, 2007
(Subgroup analysis-2). . .......................................................................... 108

Table 15. Assessment of selected risk factors for Salmonella infections in children aged 1-
10 years, assessed in a population-based case- control study, Michigan, 2007 (Subgroup
analysis-3). ......................................................................................... 109

Table16. Demographic characteristics of cases of Salmonella serotype S. Typhimurium and

controls in children aged 5 10 years, assessed in a population-based case- control study,
Michigan, 2007 (Subgroup analysis-4). ........................................................ 110

Table 17. Univariate analysis of putative risk factors for Salmonella serotype Typhimurium
infections in children aged 5 10 years, assessed in a population-based case- control study,
Michigan, 2007 (Subgroup analysis-4). ........................................................ 111

Table 18. Post hoc power analysis of selected potential risk factors for Salmonella
infections in Michigan children assessed in a population-based case-control study,

2007 .................................................................................................. 112

Table 19. Comparison of controls: neighborhood matched vs. non-neighborhood
matched. ............................................................................................... 113

Table 20. Comparison of controls identified by the case parents (method-1) and from the
landline telephone directory (method-2). ........................................................ 114

Table 21. Sample size calculations for Michigan Salmonella case-control study ........... 115

LIST OF FIGURES

Figure 1. Incidence of Salmonella infections per 100,000 population in England and

Wales, 1981-2004. ................................................................................ 116
Figure 2. Incidence of Salmonella infections per 100,000 population in the United States,

1944-2002. ......................................................................................... 117
Figure 3. Salmonella Enteritidis infections incidence in the United States,

1970-2001. ......................................................................................... 1 18
Figure 4. Age-stratified salmonellosis incidence, Michigan, 1992-2006. ................. 119
Figure 5. Surveillance of Salmonella infections in Michigan. .............................. 120
Figure 6. Michigan Salmonella case-control study, 2007: Enrollment of cases. ......... 121

Figure 7. Michigan Salmonella case-control study, 2007: Enrollment of controls ....... 122

xi

AOR
AVMA
CDC
CDR
CFSAN
CI
CRIRB
DHEC
ESR
FoodNet
FSIS

GI
HACCP
IRB
LHD
LR
MDCH
MDSS
MDR
MMWR
NARMS
OR
PAR
PHLIS
PR-HACCP
RR

SES

US FDA
USDA

LIST OF ABBREVIATIONS

Adjusted Odds Ratio

American Veterinary Medical Association
Centers for Disease Control and Prevention
Communicable Disease Rules

Center for Food Safety and Applied Nutrition
Confidence Interval

Community Research Institutional Review Board
Department of Health and Environment Control
Economic Research Services

Foodbome Disease Active Surveillance Network
Food Safety Inspection Service

Gastrointestinal

Hazard Analysis Critical Control Point
Institutional Review Board

Local Health Department

Logistic Regression

Michigan Department of Community Health
Michigan Disease Surveillance System

Multi Drug Resistant

Morbidity Mortality Weekly Reports

National Antimicrobial Resistance Monitoring System
Odds Ratio

Population Attributable Risk

Public Health Information Laboratory System
Pathogen Reduction-Hazard Analysis Critical Control Point
Rate Ratio

Socioeconomic Status

United States Food and Drug Administration
United States Department of Agriculture

xii

BACKGROUND

Foodbome infections:

Globally, foodbome illnesses are a major public health concern (1, 2). Foodbome
infections are a common, unpleasant and sometimes life-threatening problem for millions
of people worldwide. The Centers for Disease Control and Prevention (CDC) estimates
76 million people experience foodbome illnesses each year in the United States (US),
accounting for 325,000 hospitalizations and more than 5,000 deaths (1, 3). A recent
report noted that an estimated 14 million illnesses, 60,000 hospitalizations, and 1,800
deaths annually are caused by foodbome pathogens where the etiologic agent is known.
More than 250 foodbome infections have been described so far. However, in the US, the
majority of bacterial foodbome infections are caused by Salmonella, Campylobacter,

and Escherichia coli (4).

The symptoms of foodbome illnesses vary widely depending on the etiologic agent,
dosage, and immunologic status of the host. However, diarrhea, vomiting, and
abdominal discomfort are the most common symptoms (5). In the US, regulations for the
control of foodbome and waterborne illnesses have been in place since the early 19005.
The CDC, in partnership with state and local counterparts, has been responsible for the
investigation, control, and prevention of diseases spread by food and water since 1961.

In 1996, the CDC established the Foodbome Disease Active Surveillance Network

(F oodNet). FoodNet is the principal foodbome disease component of CDC’s Emerging
Infections Program (EIP). It is a collaborative project among CDC, state health

departments in EIP sites, the Food Safety and Inspection Service (F SIS) of the United

1“» .J.. 5'". .
|

States Department of Agriculture (USDA), and the Center for Food Safety and Applied
Nutrition (CFSAN) of the US Food and Drug Administration (FDA) (6, 7). The
objectives of FoodNet are to 1) determine the frequency and severity of foodbome
diseases, 2) monitor trends in foodbome diseases over time and 3) study the association
of common foodbome diseases with the consumption of specific foods. To address
these objectives, F oodNet uses active surveillance and conducts epidemiologic studies.
Between 1996 and 2006, the F oodNet surveillance population increased from 14.2
million persons (5% of the US population) in five states to 44.9 million persons (15%

of the US population) in 10 states (8).

F oodNet specifically targets seven bacterial pathogens: Campylobacter, E. coli
01 57:H7, Listeria, Salmonella, Shigella, Vibrio, and Yersinia (7, 8). This report

discusses the epidemiology of Salmonella infections.

SALMONELLA
Salmonella are gram-negative, rod-shaped, nonlactose-fermenting, bacteria belonging
to the family Enterobacteriaceae (9). Salmonella can survive and grow under a variety
of environmental conditions outside of living hosts, ranging from dry surfaces to
indigenous flora of living animals. Salmonella have been recovered from almost all
vertebrate species. Inhibition of Salmonella grth occurs at pH <38 and temperature
<7°C (10).
Salmonella serotypes are a common cause of zoonotic infections and are considered
among the most ubiquitous pathogens, both in humans and animals (11). From the time
of the first Salmonella isolation from a diarrheic pig in 1885 by Salmon and Smith and
the first laboratory confirmed outbreak of salmonellosis in humans due to contaminated
beef in 1888, Salmonella have been considered one of the most important foodbome
‘ pathogens worldwide (12). The large number of foodbome outbreaks associated with
Salmonella infections is testimony to the importance of this bacterial genus (13).
Additionally, the social and economic impact of Salmonella infections is considerable.
They impose significant costs upon the public sector, on industry (especially the food
industries), and upon infected persons and their families. In 1989, the costs of
Salmonella infections were reported at $4 billion in the US and $486 million in Canada
(14). The Economic Research Service (ERS), and the USDA have estimated that the
annual economic costs due to Salmonella infections are $3 billion, and $2.9 billion of
that cost is due to foodbome Salmonella infections. This estimate includes medical
costs and the value of time lost from work due to acute illnesses, and the economic cost

of premature deaths (15).

Pathogenesis:

Salmonella enters the human digestive system through Salmonella-contaminated food,
water, or enviromnental sources (e. g., person-to-person transmission) and survives at
low acidic conditions in the stomach by possessing an adaptive acid-tolerance response
(particularly Salmonella serotype Typhimurium) (16). Salmonella passes into the small
intestine via flagella] movement and swim chemotactically toward the mucosal surface.
Their firnbriae adheres to intestinal epithelium using receptors present on the
epithelium. Afier colonizing the lower intestine (ileum and cecum) (9), Salmonellae
invades the mucosal cell, resulting in an acute inflammation. This inflammation leads
to the activation of adenylate cyclase, increased fluid production, and release of fluid
into the intestinal lumen, which results in diarrhea. Salmonella gastroenteritis has an 8-
72 hours incubation period and may last fiom 2-7 days (17). The clinical presentation
of salmonellosis varies by serotype, infectious dose, nature of the contaminated food,
and host immune status. Certain serotypes are highly pathogenic for humans.

However, the virulence of rare serotypes is not known.

Transmission of Salmonella infections:

Salmonella is typically transmitted through the fecal-oral route. Ingestion of
contaminated food and water is the most important source of human infection.
Although a large number of bacteria (106 Cfil) are usually needed to cause an infection,
the bacteria grow well in most types of food. In foods with a high fat content, such as
chocolate and cheese, the infective dose is very low and only a few bacteria may be

sufficient to cause infection (18). The following food items have been implicated in

outbreaks of human sahnonellosis worldwide (5): meat products (raw meat, corned
beef, salami, ham, cooked turkey meat, salami sticks); milk products (raw milk, infant
dried milk, unpasteurized raw milk); soft cheese products (cheddar cheese, vacherin
Mont d’Or cheese, mozzarella); eggs or products containing eggs (mayonnaise, custard
in bakery goods, ice cream, confectionery products), fresh produce (mung bean sprouts,
cantaloupe, fresh tomatoes, alfalfa sprouts, raw almonds); and other foods (potato salad,
apple cider, roast cuttlefish, unpasteurized orange juice, peanut butter). Person-to-
person transmission of Salmonella in households, daycares, nursing homes, and
healthcare settings has been reported (19, 20). Having pets in the household,
particularly reptiles, have also been associated with transmission of Salmonella
infections in family members as these animals harbor Salmonella (21). Other factors
such as lack of hygienic practices have also been shown to contaminate the
environment with Salmonella (22) and result in indirect transmission of Salmonella
infections. Salmonella can withstand the environment outside its host for a long period
of time, therefore, inanimate objects that are contaminated can serve as a vehicle for the

transmission of infection to a susceptible individual (22).

Clinical manifestations:

Salmonella causes illnesses ranging from mild to severe gastroenteritis, bacteremia,
septicemia, localized infections, and a variety of long-term sequelae such as reactive
arthritis and Reiter’s syndrome (joint pain, irritation of eyes, and painful urination) (5).
While human infection with a host-specific serotype (typhoidal serotypes) such as S.

Typhi is associated with rather severe disease symptoms, the typical symptoms of

salmonellosis attributable to infection with non-typhoidal serotypes may include
nausea, vomiting, fever, diarrhea, and abdominal cramps. Stools are typically loose, of
moderate volume, and usually do not contain blood. Diarrhea is usually self-limited
and subsides spontaneously in 3-7 days (18, 23). The mean duration of carriage of
Salmonella in the stool is 4-6 weeks, but some carriers can be asymptomatic for months
or even years. The susceptibility to infection varies, the critical infective dose is lower
in young children, the elderly, and immunocompromised hosts (e. g., HIV infected
individuals) (24). The likelihood of extraintestinal manifestations of Salmonella
infections such as bacteremia, septic arthritis, cholecystitis, muscle abscesses, and

vascular infection (25) are much higher for immunocompromised individuals (26, 27).

Salmonellosis: Burden of the disease:

Despite the implementation of several control and prevention measures, Salmonella
infections remain a major public health problem worldwide (1). In England and Wales,
incidence of Salmonella infections started to rise in the mid 19803, primarily because of
the epidemic of S. Enteritidis infections. Between 1987 and 1992, an overall 83%
increase, from 40 cases per 100,000 population in 1987 to 73 cases per 100,000
population in 1992 was observed (28). However, in recent years, due to several control
and prevention measures, a significant decrease in the overall incidence has been
noticed. In 2004, a Salmonella incidence of 22 cases per 100,000 was reported (29).
Figure 1 shows incidence of Salmonella infections per 100,000 population in England

and Wales, between 1981 and 2004.

In Australia fi'om 1996 to 2003, the average incidence of Salmonella was reported to be
28.99 cases per 100,000 population (30). Japan had a relatively low average incidence
of 3.32 cases per 100,000 population between 1993 and 2004 compared to European
countries (31). In Canada, mean annual incidence (1990-2004) was 19.4 cases per
100,000 population (32). The differences in Salmonella incidence could be partly
explained by differences in the effectiveness of the existing public health measures that
may limit extensive spread of contaminated poultry products and other foods
potentially contaminated with the organism. Additionally, disease surveillance and the
rate of case detection among these countries may also vary.

Figure 2 shows the annual incidence of Salmonella infections per 100,000 population
between 1944 and 2002 in the US. A steady rise in the incidence of Salmonella
infections has been observed between 1944 and 1980. However, a significant increase
in incidence was observed after the 1980s. The incidence increased from 10 cases per
100,000 population in 1980 to 23 cases per 100,000 in 1994. As a result of prevention
programs such as on-farm microbiologic testing for Salmonella and improved bio-
security of food in the early 19905, the incidence of Salmonella infections started to
decline. The overall incidence of salmonellosis decreased fi'om 16.6 cases per 100,000
population in 1996 to 14.2 cases per 100,000 population in 2001, although large

outbreaks and sporadic cases continue to occur (4).

Although most culture-confirmed cases are reported to health authorities, the disease
surveillance system underestimates the actual number of Salmonella infections as a

result of surveillance artifacts. First, a person infected with Salmonella should develop

 

symptoms that are severe enough to seek medical care. Second, the physician must
request and collect a specimen from the patient for bacterial culture. Third, the
laboratory must test the specimen for Salmonella using a sensitive method and forward
the isolate to a State Public Health Laboratory for confirmation. Fourth, the state
laboratory must report results to the CDC. It has been estimated that the number of
reported cases represent just 1% - 5% of the actual number of Salmonella infections
that occur in the population (33). To better estimate the disease burden in the
population associated with the foodbome infections, FoodNet conducts surveys of
laboratories, physicians, and the general population at FoodNet sites. By estimating the
proportion of the population seeking medical care for diarrhea] symptoms, the
proportion of physicians advising bacterial stool culture, and taking into account the
influence of variations in laboratory testing for bacterial pathogens on the yield of
number of culture-confirmed cases, FoodNet estimated in 1997 there were 1.4 million
Salmonella infections, resulting in 113,000 physician office visits. Additionally, in
recent years, salmonellosis has been attributed to 16,000 hospitalizations, and more

than 500 deaths each year in the US (1).

Between 1996 and 2004, a decline in overall Salmonella incidence was observed, but
when the data were stratified by the common serotypes, only one of the four most
common (S. Typhimurium, S. Enteritidis, S. Newport, and S Heidelberg) Salmonella
serotypes, S. Typhimurium, declined significantly (34). In contrast, there were marked
increase in the incidence of S. J aviana and monophasic serotype identified as S.

I 4,[5] 12:i:- infections. No substantial declines in the incidence of the other common

Salmonella serotypes, S. Enteritidis, S. Newport, and S. Heidelberg, were observed (4).
A contributing factor to the decline in Salmonella infections was a change in the
industry and regulatory approaches to meat and poultry safety. In the mid 1990’s, the
UDSA-Food Safety Inspection Services (F SIS) implemented Pathogen
Reduction/Hazard Analysis Critical Control Point (PR/HACCP) systems regulations in
meat and poultry slaughter and processing plants. The decline in the incidence of S.
Typhimurium infections in humans may be related to changes in meat processing as
evidenced by a decline in the prevalence of Salmonella isolated from FSIS-regulated
meat and poultry products (35).

In 2005, a total of 6,655 laboratory-confirmed cases of salmonellosis, with an incidence
of 14.55 cases per 100,000 population in FoodNet sites was reported, which accounted
for 38% of all laboratory-confirmed cases of foodbome infections (36).

In 2006, a total of 6,655 laboratory-confirmed cases of Salmonella infections, with an
incidence of 17.4 cases per 100,000 population, was reported by F oodNet surveillance
sites. Of the 6,655 submitted samples, 5,957 (90%) of the Salmonella isolates were
serotyped and seven serotypes accounted for 64% of the infections: Typhimurium,
1,157 (19%); Enteritidis, 1,109 (19%); Newport, 531 (9%); Javiana, 292 (5%);

Montevideo, 250 (4%); Heidelberg, 239 (4%); and S. I 4,[5],12:i:-, 239 (4%) (8).

Healthy People 2010 objectives have been established for four foodbome pathogens
under FoodNet surveillance (37). In recent years (2004-2006), the incidences of
Campylobacter, Shiga-toxin producing E. coli (STEC 0157), and Listeria were

approaching their targets of 12.31, 1.00, and 0.25 cases per 100,000 population,

respectively.‘ However, the incidence of Salmonella infections in 2006 remained much
higher than the goal of 6.8 cases per 100,000 population by the year 2010 (8). Table 1
compares 2006 incidences of selected pathogens with the National Health Objectives

2010.

10

OUTBREAKS OF FOODBORN E SALMONELLA INFECTIONS

As mentioned earlier, despite control and prevention efforts, foodbome outbreaks
(defined by the CDC as two or more illnesses from a common source) caused by
Salmonella have continued to occur. Salmonellosis outbreaks have been associated
with family gatherings, restaurants, and community outbreaks either limited to a
defined population or spread community-wide (3 5). Table 2 summarizes selected large

foodbome outbreaks where Salmonella have been isolated as a cause of the outbreak.

It should be noted that the number of reported outbreaks represents a small proportion
(6%-7%) of the total outbreaks that actually occurred in the population. Most
outbreaks, particularly smaller ones, are never recognized, and those that are
recognized frequently go unreported (38). The likelihood that an outbreak is brought to
the attention of public health authorities depends on many factors, including general
population and physician awareness, and motivation to report the incident as well as the
resources and disease surveillance activities of local health and environmental agencies.
Outbreaks that are most likely to be reported to health authorities include those that are
large, multi-state, restaurant-associated, or that cause serious illness, hospitalization, or

death (38).

During 1993-1997, Salmonella caused 357 (55%) of the 655 bacterial foodbome
outbreaks with a known etiology. Salmonella serotype in the US. Enteritidis was the most
frequently reported cause of foodbome disease outbreaks, accounting for 7% of all

foodbome outbreaks (38).

11

In 2004, of the 6,498 Salmonella cases ascertained, 352 (5%) were identified as being
outbreak-related. Of the outbreak-associated Salmonella cases, 78% were food-related,
20% were not food-related (e. g., person-to-person transmission) and for 2%, the mode of
transmission was unknown. In the same year, Salmonella was responsible for 23% of
nationally reported foodbome outbreaks in which an etiology was confirmed (4). In
2006, outbreak-associated cases accounted for 404 (6.1%) of 6,655 Salmonella cases

ascertained at FoodNet sites, compared with 296 (4.6%) of 6,505 cases in 2005 (8).

12

SELECTED OUTBREAKS OF SALMONELLOSIS

Multi-state outbreak associated with contaminated milk, 1985: (35)

In 1985, an outbreak associated with S. Typhimurium causing 16,000 confirmed cases
in 6 states was due to the consumption of milk from a dairy plant in Chicago, Illinois.
Located in an industrial area of Chicago, the dairy plant was the sole supplier of milk to
217 supermarkets in Illinois, Indiana, Iowa, and Michigan. The dairy plant at one time
processed about 1.5 million pounds of milk a day. This was the largest reported
outbreak of foodbome salmonellosis in the US. Investigators from CDC and FDA
could not identify the actual deficiency that led to the contamination of the pasteurized
milk shipments but suspected the failing valves between the towers that contained the

raw milk and those that hold the pasteurized milk.

Multi-state Schwan’s ice cream outbreak, 1994: (39)

In September and October of 1994, Minnesota's Health Department observed a sharp
rise in the reported cases of salmonellosis. Investigators from the FDA and state
officials determined that contaminated ice cream was the likely cause of a S. Enteritidis
outbreak that may have sickened more than 3,000 people in as many as 41 states.
Health officials suggested that Salmonella contamination occurred when raw,
unpasteurized eggs were hauled in trucks that later carried pasteurized ice cream
premix to the Schwan's plant. CDC received reports from 41 states of illness associated
with Schwan's products: 740 cases from 30 states were confirmed by cultures, and 41

states reported an additional 3,423 suspected cases, however, no deaths were reported.

13

Analysis of egg samples from a tanker, ice cream from the plant, and stool specimens
fi‘om infected consumers

revealed that all were contained with Salmonella of the same genetic type.
Michigan bakery product outbreak, 2002: (40)

In May 2002, S. Enteritidis associated outbreaks with the consumption of bakery
products in Macomb County, Michigan resulted in 196 reported illnesses, among those
24 individuals were hospitalized. The state health authority, Michigan Department of
Community Health (MDCH), concluded that black forest cakes and pastries were the

vehicles for Salmonella transmission.

Multi-state raw almond outbreak, 2004: (41)

On May 12, 2004, the Oregon State Public Health Laboratory identified a cluster of
five patients infected with Salmonella serotype Enteritidis. Further investigation led to
the identification of at least 29 patients in 12 states and Canada as part of the
Salmonella outbreak that began in September 2003. After an investigation by public
health officials, the illnesses were linked to the consumption of raw almonds distributed

by Paramount Farms and roughly 18 million pounds of raw almonds were recalled.

South Carolina restaurant outbreak, 2005: (42)

In May, 2005, the Department of Health and Environmental Control (DHEC) was
informed about a possible outbreak of foodbome illness at Old South restaurant in
Camden, South Carolina. The outbreak turned out to be one of the largest foodbome

outbreaks in South Carolina history. Laboratory results from DHEC documented the

14

presence of S. Enteritidis in roasted turkey that had been consumed at an event catered
by Old South. During the course of investigation, investigators determined that the
convection oven used to cook the contaminated turkey had malfimctioned, thereby
preventing the turkey fi'om reaching a temperature sufficient to destroy Salmonella. A
total of 304 laboratory-confirmed and suspected cases were identified and one person

died as a result of these infections.

Multi-state Orchid Island Juice outbreak, 2005: (43)

In early July 2005, the FDA issued a nationwide warning to consumers against drinking
unpasteurized orange juice products distributed under a variety of brand names by
Orchid Island Juice Company of Fort Pierce, Florida. Fifteen cases had been directly
linked to the product and at least 16 states had reported cases of S. Typhimurium
infections that matched the outbreak strain. On July 15, 2005, with an increasing
number of Salmonella illnesses traced to unpasteurized orange juice being reported to
state health departments, the company agreed to issue a recall of all fresh and fiozen

juices.

Multi-state tomato associated outbreak, 2005-2006: (44)

During 2005-2006, four large multi-state outbreaks of Salmonella infections (with
multiple serotypes) associated with eating raw tomatoes at restaurants were reported.
These outbreaks resulted in 459 culture-confirmed cases of salmonellosis in 21 states.
The investigation revealed that the tomatoes had been supplied to restaurants either

whole or pre-cut from tomato fields in Florida, Ohio, and Virginia. Irnplicated

15

tomatoes were traced to a single packinghouse in Ohio supplied by three tomato
growers fi'om 25 fields in three counties. Tomato production had ended by the

time the packinghouse was implicated.

Multi-state peanut butter outbreak, 2007: (45)

On February 14, 2007, the CDC and FDA announced that there had been 290 cases of
S. Tennessee infections in 39 states that were linked to the consumption of Peter Pan
and Great Value brand peanut butter that was manufactured in ConAgra's Georgia
peanut butter plant. Peter Pan and Great Value brand peanut butter beginning with a

particular product code was recalled in response to the outbreak investigation.

SELECTED COMMON SALMONELLA SEROTYPES

Using the Kauffmann-White scheme based on antisera prepared to group and individual
somatic and flagellar antigens (O and H antigens) fi'om representative serotypes, over
2,500 different Salmonella serotypes have been identified in different parts of the world
(46, 47). Salmonella can be classified into two main groups 1) S. enterica and 2) S.
bongori. Salmonella enterica is further divided into six subspecies 1) S. enterica subsp.
arizonae, 2) S. enterica subsp. diarizonae, 3) S. enterica subsp. houtenae, 4) S. enterica
subsp. indica, 5) S. enterica subsp. Salamae, and 6) S. enterica subsp. enterica. The
latest subspecies include all of the ~25,000 Salmonella serotypes. However, over 80%
of all Salmonella isolated from humans and animals belong to about 20 serotypes

(Table 3) (4).

16

Another way of classifying Salmonella serotypes is ecologically based on host
adaptation (48, 49). Salmonella serotypes can be divided into two groups 1) host
adapted and 2) ubiquitous (non-host adapted). Host-adapted serotypes typically cause
systemic disease in a limited number of host species (Table 4) (49). For example, S.
Typhi, S. Gallinarum, and S. Abortusovis are almost exclusively associated with
systemic disease in humans, poultry, and sheep, respectively. However, some host
adapted serotypes can also cause disease in more than one host species: Dublin and
Choleraesuis, for example, are generally associated with disease in cattle and pigs,
respectively but may also infrequently cause disease in other mammalian hosts
including humans. Examples of non-host adapted serotypes include S. Typhimurium
and S. Enteritidis, which are the most common serotypes that have been isolated from

humans, animals, and environmental sources (49, 50).

Over periods of several years, incidences of certain Salmonella types have varied
within large geographic regions. In 2001 , approximately 60% of human cases reported
to the CDC were caused by four serotypes, namely S. Typhimurium (22.1%), S.
Enteritidis (17.7%), S. Newport (10.0%), and S. Heidelberg (5.9%). Of the 5,957
(90%) Salmonella isolates serotyped in 2006, the same four serotypes accounted for
51% of Salmonella infections: S. Typhimurium, 1,157 (19%); S. Enteritidis, 1,109

( 19%); S. Newport, 531 (9%) and S. Heidelberg, 239 (4%).

Table 5 compares the percent change (2004 vs. 1996-1998) in reported incidence of the

four common Salmonella serotypes (4). A significant decrease (41%) occurred in the

17

incidence of S. Typhimurium, while there was an increase in the other two common

serotypes, S. Heidelberg, and S. Newport.

A brief description of the epidemiology of the four common Salmonella serotypes, S.

Typhimurium, S. Enteritidis, S. Heidelberg, and S. Newport, is presented here.

Salmonella enterica serotype Typhimurium:

S. Typhimurium is among the most prevalent human Salmonella serovars worldwide
(4). S. Typhimurium accounted for 23% of all human Salmonella isolates reported in
2000. In regard to animals, S. Typhimurium is primarily a pathogen for cattle, but other
species such as sheep, goats, pigs, and birds can be affected (51). Additionally, it has
been isolated from various environmental sources.

Several risk factors have been identified for S. Typhimurium infections. Outbreaks have
been associated with consumption of raw or undercooked ground beef (52), lamb
kebabs (53), commercially processed salad (CDC, 2004b), unpasteurized milk (3 5),
eating raw or undercooked eggs (54), cheese made of raw milk (55), chocolate (56),
and salami sticks (57). In 2006, a multi-state outbreak of S. Typhimurium infections
associated with tomatoes accounted for 58 (14%) of the outbreak-associated Salmonella
cases identified by F oodNet (44).

Sporadic S. Typhimurimn infections have been associated with the consumption of
undercooked ground beef (58) and undercooked eggs or egg-product (59). Contact

with farm animals and pets has also been associated with infection (21, 60).

18

Phage typing has enabled differentiation of S. Typhmurium into more than 200
Definitive phage Types (DTs). The recent public health concern related to S.
Typhimurium was due to the emergence of multi-drug resistant (MDR) S.
Typhimurium DT 104 (61). As a result of several control and prevention programs for
salmonellosis, the incidence of S. Typhimurium infection decreased 24% from 1996 to
2001. However, an increasing proportion of isolates are resistant to multiple

antimicrobial agents (62).

Multi-drug resistant Salmonella Typhimurium DT104:

Globally, the prevalence of S. Typhimurium resistant strains has increased several folds
in the past few decades and caused a considerable number of outbreaks in North
America since 1996. It has been suggested that drug-resistant Salmonella serotypes
have emerged, primarily in response to antimicrobial use in food animals and the
international trade of animals (51, 63). Among Salmonella serotypes, S. Typhimurium
has shown the highest incidence of antibiotic resistance. The most frequent S.

Typhimurium phage type associated with a multi-drug resistance pattern is DT104.

S. Typhimurium DT104 is commonly resistant to 5 antibiotics—arnpicillin,
chloramphenicol, streptomycin, sulfonarnides, and tetracycline (64). The first resistant
strain of S. Tyiphimurium DT04 was detected in the UK in cattle during the late 19805,
and since then it has become common in other animal species such as poultry, pigs, and

sheep (61).

19

In epidemiologic investigations, human infections with multiple antibiotic resistant
DT104 isolates have been associated with the consumption of undercooked meat,
poultry, beef, and pork. Table 6 shows the distribution of MDR and S. Tyiphimurium
and definitive phage type 104 (DT04) in selected countries from 1992 through 2001. In
most European countries and North America, MDR started to increase from the mid-

19905(61,65)

Antimicrobial resistance to Salmonella serotypes, particularly DT104, appear to pose a
significant health risk. In a recent review, it has been suggested that antimicrobial
resistance in Salmonella results in about 300 excess hospitalizations and 10 deaths in
the US each year (66). Additionally, a study conducted in a Danish population found
that persons with resistant Salmonella infections had a higher mortality rate compared

to those with antibiotic-susceptible Salmonella infections (67).

Salmonella enterica serotype Enteritidis:

S. Enteritidis is the second most common Salmonella serotype in the US (50). An
epidemic of S. Enteritidis began in the late 19705 in the northeast region of the country
and some areas of Europe (68). In 1976, 1,207 S. Enteritidis isolates were detected
nationwide with an incidence of 0.6 cases per 100,000 population. The incidence
reached 2.4 cases per 100,000 by 1985. From 1980 to 1996, an increase of S.
Enteritidis isolation fiom 5% to 25% of all Salmonella cultures was reported (38).
Figure 3 shows the overall S. Enteritidis infection incidence in the US between 1970

and 2001. According to the CDC, 677 S. Enteritidis-related foodbome outbreaks were

20

reported between 1990 and 2001, resulting in 23,366 illnesses, 1988 hospitalizations,
and 33 deaths (38, 69). During 1994, 1995, and 1996, S. Enteritidis surpassed S
Typhimurium to become the most common Salmonella serotype isolated in the US
(70). The incidence of S. Enteritidis infection increased markedly from 1980 to 1995,

but has decreased 22% from 1996 to 2001 (71).

Epidemiologic investigations of sporadic cases and outbreaks of S. Enteritidis
infections have demonstrated that contaminated eggs and egg-products are major risk
factors and that about 80-85% of all S. Enteritidis infections cases can be attributed to
the consumption of contaminated egg products (1, 50). However, recent studies have
identified that poultry has been associated with the transmission of S. Enteritidis
infections. In a population-based case-control study, Kimura et al. reported that eating
chicken outside the house doubled the risk of acquiring S. Enteritidis infection (OR =

2.6, 95% CI, 1.6-4.4) (72).

Moreover, in 2003, 12.8% of chickens sampled in slaughter plants in the F SIS-
PR/HACCP Verification Testing Program were contaminated with Salmonella.
USDA-FSIS reported an increase in the frequency of isolation of Salmonella,

particularly S. Enteriditis, in chicken broiler during 2000-2005.

Furthermore, in 2005, an outbreak of S. Enteritidis associated with eating raw almonds

was identified.

21

Unlike the increase in incidence of antibiotic-resistance seen in S. Typhimurium, S.

Enteritidis has remained sensitive to most antibiotics (73).

Salmonella enterica serotype Newport

S. Newport is the third most common Salmonella serotype in humans and has recently
been named an emerging disease by the American Association of Veterinary
Laboratory Diagnosticians. From 1973 through 1997, 149 S. Newport outbreaks
caused 7,159 cases of illnesses. The median number of cases per outbreak was 17
(range: 2 to 700 cases) and 11.6% of case patients were hospitalized, while 0.1% died
(62).

Unlike S. Typhimurium, the incidence of S. Newport increased in recent years—a 32%
increase was observed fiom 1996 to 2001 (62).

Outbreaks and sporadic cases of S. Newport have been associated with consumption of
hamburger (74), peanuts (75), cured ham (76), alfalfa sprouts(77), undercooked eggs
(78), and pork sandwiches (79).

In addition to an overall increase in the incidence, the emergence of multi-drug resistant
strains of S. Newport in recent years (80) is a significant health problem. Since 1996,
the National Antimicrobial Resistance Monitoring System (NARMS) has identified an
increasing number of S. Newport isolates that are resistant to at least nine of 17
antimicrobial agents tested: moxicillin/clavulanate, ampicillin, cefoxitin, ceftiofur,
cephalothin, chloramphenicol, streptomycin, sulfamethoxazole, and tetracycline (80).

The increase of S. Newport multi-drug —resistant infections in humans has been

22

associated with exposure to animal products: ill cattle (81, 82) cheese made from

unpasteurized milk (83), and raw or undercooked ground beef (80).

Salmonella enterica serotype Heidelberg:

S. Heidelberg was the fourth most commonly reported Salmonella serotype in the US
fiom 1993 through 1997. An average of 2,180 cases of S. Heidelberg infections were
reported annually, accounting for about 6% of all culture-confirmed Salmonella
infections in the US. However, culture-confirmed cases may represent only about
M2.6% of all illnesses of salmonellosis. Therefore, the actual burden associated with S.
Heidelberg is estimated at 84,000 cases of illnesses annually with an incidence of 27.1
cases per 100,000 population (62).

Like other Salmonella serotypes, S. Heidelberg has largely been associated with food
vehicles. However, reports of person-to-person or direct animal-to-person transmission
have been reported (84). Investigations of outbreaks caused by S. Heidelberg have
identified undercooked chicken, pork and cheddar cheese as food vehicles associated
with S. Heidelberg (85, 86). Additionally, S. Heidelberg has been isolated from
eggshells and shown to grow in eggs in in-vitro studies (87). In a population-based
case-control study conducted at the F oodNet sites, Hennessy et a1. (2004) reported
eating eggs prepared outside the home at commercial food establishments as a
significant risk factor (OR = 6.0; 95% CI, 1.2—29.6) for S. Heidelberg infections with a

population-attributable fraction of 37% (88).

23

SALMONELLA INFECTION IN CHILDREN

As stated earlier, the implementation of several control and prevention measures for
salmonellosis across the nation in the mid 19905 resulted in a significant decrease in the
overall incidence of Salmonella infections in recent years (1996-2003) (70). However,
age-stratified analysis has shown a relatively stable infection rate in children at the
national level (47). A large number of epidemiologic investigations, including reports
from the CDC’s FoodNet, have shown consistently higher incidences of laboratory-
confirrned Salmonella infections in children compared to other age groups (3, 50, 89-
91). Children aged < 5 years account for a large proportion of reported cases in the
majority of the US disease surveillance systems (62). Between 1998 and 2003, the
incidences of laboratory-confirmed cases of Salmonella infections reported at FoodNet
sites were 122.7 cases per 100,000 population for children aged < 1 year and 50.6 cases
per 100,000 population for children aged 1-4 years, compared to 10.8 cases per 100,000
population for those aged :5 years. Our recent work investigating associations
between demographic attributes and the distribution of salmonellosis based on data
from MDCH (1995-2001) revealed an approximately 10-fold increased risk for
acquiring salmonellosis in children age < 1 year and a 3-fold increased risk for those
aged 1-4 years when compared to adults aged 15-39 years (70, 92). Based on 13,877
salmonellosis case reports in Michigan, age-stratified analysis of recent MDCH data
(1992-2006) showed an incidence of 70 cases per 100,000 population for children aged
< 1 year, 22.2 cases per 100,000 population for children aged 1-4 years, and 10 cases
per 100,000 population for children aged 5-9 years, compared to 4 cases per 100,000

population for those aged 10-34 years (Figure 4).

24

r-u-ugLu
..

In addition to a several fold increased risk for acquiring salmonellosis, pediatric cases
account for substantially more morbidity and mortality compared to adults (93, 94).
The reasons for the observed higher incidence of salmonellosis in young children are
not well understood. It is suggested that the higher incidence of salmonellosis in
children could be due to their host irnmunoincompetent status, which makes them
vulnerable to many infections including salmonellosis (94). Another reason cited in the
literature is the increased case detection of Salmonella infections in children. In young
children with gastrointesitinal symptoms, it is more likely that 1) the parents will seek

medical attention and 2) the healthcare provider will submit a sample for culture (95).

As discussed earlier, the majority (80 - 85%) of sporadic cases of salmonellosis in the
adult population results fi'om consumption of contaminated food (96, 97). Another
major risk factor in the adult population identified is traveling to areas (e. g., South
America, Asia) where Salmonella is more prevalent. However, risk factors for
salmonellosis in children have not been extensively evaluated in population-based
epidemiologic studies. The majority of the studies in children have been conducted in
response to salmonellosis outbreaks, particularly in daycare centers and nurseries.
These investigations usually identify a point source as a cause for the outbreak. The
risk factors identified in outbreaks may not be similar to those of sporadic cases of

salmonellosis since exposures in sporadic cases vary widely.

Some analytical studies have shown a strong association between Salmonella infections
and the consumption of contaminated food including raw or undercooked eggs in

children aged < 5 years (59, 98-100), while a number of other investigations have failed

25

to implicate food as a source of infection for children < 5 years of age (93, 101-103).
Additionally, in the majority of studies where investigators addressed the implication of
specific food vehicles for salmonellosis, they did not explore the influence of mode of
food preparation, food handling methods, and family kitchen hygiene practices as
potential factors associated with Salmonella infections. There is an overall dearth of
epidemiological data on the influence of these practices and the risk of sporadic
Salmonella infections in households with children (96).

The risk factors for salmonellosis in children may be substantially different from adults
because of markedly different behaviors and exposures. Table 7 shows the identified
risk factors for Salmonella infections in children. It has been suggested that
contaminated environmental sources contribute more than contaminated food vehicles
in the acquisition of Salmonella infections in children. However, limited evidence
exists to explain the influence of various environmental exposures in acquiring
Salmonella infections in children (104). The majority of the literature showing an
association between various environmental factors and Salmonella infections consists
of case reports and case series. The few observational studies conducted, mainly
outside the US, have shown mixed results (97, 102). Some investigations suggest that
children have been infected from contaminated environmental sources(104). In
contrast, other studies have not found associations between environmental sanitation or

ownership of pets and salmonellosis in the pediatric population (97).

In a population-based case-control study, Delarocque-Astagneau et a1. (1998) suggested

that the predominant mode of Salmonella transmission differs in children by age (96).

26

 

They found that for children < 1 year of age, Salmonella infections are mainly related
to exposure to an infected family member, whereas for children 1-5 years, infections
are associated with the consumption of raw or undercooked egg products or chicken.
However, the investigators did not collect the data on several established risk factors
for Salmonella infections in children (e.g., exposure to reptiles, travel to endemic
Salmonella zones). Hence confounding of the effect size by these known risk factors

for salmonellosis in children cannot be determined (96).

In a few studies conducted in Guam, Salmonella have been isolated from kitchen
counters (22), household dust and soil close to the house entrance (104). Moreover, in
assessing an association between environmental contamination at shopping centers and
Salmonella infections in children, a recent population-based study conducted using
FoodNet sites identified that riding in a shopping cart next to meat and poultry was a
risk factor for salmonellosis in children (89). However, this newly identified factor has

yet to be validated.

State and local health authorities nationwide address Salmonella infections risk factors
primarily for enteric outbreaks, which constitute about 6%-7% of all reported cases of
salmonellosis (62). In the US, a few studies have been conducted to identify risk
factors for sporadic infections of Salmonella infections. Furthermore, to date, no
systematic population-based epidemiological study to identify risk factors for

Salmonella infections in children has been conducted in Michigan.

27

 

Objectives:

Primary objective:

To assess the role of potential risk factors in the etiology of Salmonella infections in
Michigan children, we conducted a population-based case-control study to determine 1)
household-related factors (e. g., household density, family kitchen practices), 2) selected
environmental exposures (e. g., contact with a person having symptoms of GI infection,
contact with pets, and 3) consumption of various food vehicles (e. g., eating eggs,
poultry and meat) in the etiology of sporadic non-typhoidal Salmonella infections in

Michigan children aged S 10 years.

Secondary objectives:

To assess select exposures (food vehicles and environmental exposures) by age group:
1) <1 year and 1-10 years.

To identify factors (food vehicles and environmental exposures) associated with

Salmonella serotype Typhimurium infections.

28

 

MATERIAL AND METHODS
Study setting and population:
The study was conducted in the state of Michigan. With an area of 96,810 square
miles, Michigan is the 11th largest and eighth most populous state in the US.
Michigan's major industries include car manufacturing, farming (corn, soybeans, and
wheat), timber, and fishing. According to the 2000 US Census, Michigan has a total
population of 9,956,] 11 and an average per capita income of $22,168. The target
population for this study was all children aged 5 10 years with a permanent Michigan
residential address. This age group constitutes about 15.6% (1,560,702) of the total

population (105) .

Source of Data:

Michigan Department of Community Health:

Salmonellosis is a notifiable disease and under Michigan’s Communicable Disease
Rules, which require the reporting of the occurrence or suspected occmrence of all
certain serious diseases and conditions (106). Therefore, physicians and laboratories
across Michigan submit disease reports to their local health department (LHD) in either
the jurisdiction where the individual with suspected or confirmed salmonellosis resides
or where the reporting facility is located. The LHDs then submit these reports to the
statewide communicable disease reporting system, Michigan Disease Surveillance
System (MDSS), which is maintained by the MDCH. Figure 5 describes the
surveillance of Salmonella infections in Michigan. MDSS is a centralized, statewide,

web-based database of reportable diseases. In addition to reporting to the MDCH,

29

LHDs and clinical laboratories send clinical specimens to the Bureau of Laboratories,
MDCH for testing. For all cases of salmonellosis, the Salmonella isolates are serotyped

at MDCH Bureau of Laboratories and results are entered into MDSS.

Study design:

A case-control design was used to achieve study objectives. Case-control studies are
used to identify factors that may contribute to a disease or condition by comparing a
group of individuals who have a particular disease or condition to a group of
individuals that do not. In this design, patients who have developed a disease are first
identified, and their past exposure to suspected etiological factors are compared with
that of controls or referents who do not have the disease. Case-control design allows
for investigation of multiple exposures potentially associated with the given disease or
condition. The use of a case-control approach facilitates rapid and cost-effective
collection of data and allows scientific evaluation of the risk factors contributing to
disease occurrence (107, 108). The other observational study design is a cohort or
follow up study. In a cohort study, subjects who presently have a certain condition
(exposed) and another group who are not affected by the condition (un-exposed group)
are followed up longitudinally and compared for a defined period of time or till
outcome of interest(s) (e. g., disease) occur (108). In our study, we wanted to study
multiple potential risk factors associated with a disease (salmonellosis), cohort study
design was not appropriate.

Unmatched vs. matched case-control study:

Matching refers to the selection of a reference (control) series- unexposed subjects in a

case-control of cohort study- that’ is identical, or nearly so, to the index series (cases)

30

 

with respect to the distribution of one or more potentially confounding factors (108).
Matching control selection strategies are primarily done in case-control studies. When
properly applied, matching may provide improved study efficiency and precision.
Matching may be performed by subject to subject (individual matching) or for groups
of subject (frequency matching). Individual matching usually involve one or more
control subjects with matching-factor similar to those of the case subject (e.g., age, sex,
race). Frequency matching involves selection of an entire stratum of control subjects
with matching-factor similar to that of stratum of case subjects. Although matching
does not offer advantages over unmatched control selection with regard to study
validity under the case-control design, gains in study precision of results may be
improved in matched design. Greater precision produces a smaller effect size (odds
ratio) variance, and narrower confidence intervals. Disadvantages of matching may
include 1055 of statistical efficiency and logistical issues with the enrolhnent of matched
controls, particularly if the cases are matched with controls on multiple factors (difficult
to find controls matched on several variables). Additionally, if the controls are matched
on factors that are affected by exposure or disease (over matching), such as symptoms
or sign of the exposure or the disease, such matching can distort the study data and
yield biased estimates (107). In our study, we began enrolling participants using a
neighborhood-matched design, but discontinued the matched design early in the course
of data collection because of the lack of apparent differences between neighborhood-
matched controls and non-neighborhood controls with regard to SES attributes and

difficulty in finding the neighborhood matched controls (Table 19).

31

 

Cases:
In this study, cases were defined as Michigan children aged S 10 years with laboratory-
confinned Salmonella infections, except Typhi or Paratyphi, reported to MDCH

between December 15, 2006 and October 15, 2007.

Inclusion and exclusion criteria:
All individuals of age $10 years infected with any Salmonella serotypes other than
Typhi or Paratyphi, isolated from any clinical specimen (e. g., stool, urine, blood,

cerebrospinal fluid), were eligible for inclusion.

Cases were excluded from the study if 1) the case had a reported congenital
malformation (e. g., birth defect), serious medical condition (e.g., leukemia, lymphoma),
or a concomitant infection at the time of the Salmonella infection, 2) the case was
reported as part of a salmonellosis outbreak investigated by public health officials, 3)
more than one eligible cases were reported from the same household, the youngest case

was selected for the study.

Additionally, case children were not included 1) if the child’s family could not be
contacted after two consecutive mailed invitation letters were returned to sender due to
a wrong residential address in the MDSS and/or after about 20 phone call attempts,
including evening and weekend calls, and 2) the contact established with the family but

the caretaker (e.g., parent(s), grandparents) refused to participate in the study.

32

Controls:

Controls were children aged 5 10 years who were not diagnosed with any enteric
infections (e. g., salmonellosis, carnpylobacteriosis) by a healthcare provider and did not
experience any enteric disease symptoms (e. g., diarrhea, vomiting, nausea) during the
30 days prior to the interview day. After infection with Salmonella, stool shedding of
the bacteria may last for up to 30 days, therefore, we choose the 30 days of symptoms-

free period to exclude individuals with asymptomatic Salmonella carriage.

Enrollment of cases and controls by age-groups:

The incidence of salmonellosis and its risk factors vary by age within the S 10 year old
age group. To ensure enrollment of an adequate number of cases and controls in each
age category, we used age-stratified sampling. We divided participants into three age
categories: <1 year, 1-5 years, and 6-10 years, and enrolled cases and controls by these
age categories.

Two methods were used to enroll controls:

Method 1: Case parent(s) were asked to identify a child of similar age to their own

within their county of residence.

Method 2: To obtain potential controls, we used the on-line telephone directory
available at www.whitepages.com. This website has a reverse address function that
allowed us to find a list of household phone numbers within the same county. We

obtained a list of potential control household phone numbers using case’s addresses.

33

Each listed household was called, and after explaining the study’s objective, we asked
if they have a child of age S 10 years. We interviewed consenting parent(s) or
caretakers. If efforts to enroll a control from the compiled list were not successful,
additional phone numbers were obtained using the same method working outward from
the case’s home address. This process was repeated to enroll the required number of

controls.

Control exclusion criteria:
Controls were excluded if 1) the child had a congenital malformation (e.g., birth defect)
or serious medical condition (e. g., leukemia, lymphoma) and/or 2) their caretaker (e. g.,

parent(s), grandparents) refused to participate in the study.

Questionnaire development:

We developed a structured questionnaire to collect data on sociodemographic
characteristics (e. g., age and sex, household income, parental education), child feeding
practices (e.g., breast feeding, formula milk, use of pacifier), child rearing (e.g.,
daycare, pre-school, or elementary school attendance), and various environmental
exposures (e. g., contact with animals, contact with a person having GI symptoms).
Additionally, we collected data on household kitchen practices. We included questions
that had previously been used in similar research (e.g., food frequency questionnaire).
The questionnaire was pilot-tested on volunteer parents, issues identified in this
exercise were addressed, and appropriate changes were incorporated into the final

version of the questionnaire (Appendix 4).

34

Data collection methods:

Several data collection methods are available for epidemiologic studies. The choice of
data collection method is determined by several factors, including study population,
response expectations, available resources, and the preferences of investigators (109).

The methods of data collection we used, along with the rationale, are outlined below:

Face-to-face (in-person) interview:

To conduct face-to-face interviews, the fieldwork and its organization requires more
resources compared to interviewing by telephone or using mail-in questionnaires and
can be associated with interviewer bias. The advantages of face-to-face interview
include a higher response rate, use of a longer survey instrument with complex skip
patterns, more accurate recording of responses, low non-response on questions, and

more appropriate for hard to reach populations (e. g., illiterate, institutionalized).

Telephone interview:

The telephone interview for data collection is very commonly used in epidemiologic
research. The advantages include: less costly than face-to-face interview, higher
response rates than mailed in questionnaires, relatively quicker access to participants,
supervision of interviewers is feasible, and a better response rate with sensitive
questions. The disadvantages include selection bias (e.g., persons without phones are

not included) and a relatively high refusal rate.

35

 

Mail-in-questionnaire:

Mail-in-questionnaires are becoming more popular in public health research.
Advantages include a lower cost compared to in-person interviews and anonymity (no
threat of interviewer bias). Additionally, a self-administered questionnaire may
produce more reliable and reduce non-response to sensitive questions. However, mail-
in questionnaires are associated with a high rate of missing data, a low response rate,

and are not suitable for population with low literacy (110).

Electronic data collection:

Computer-assisted data collection methods such as self-administered electronic
questionnaires are being used in epidemiologic research. The instrument may either be
an Internet questionnaire or an electronic questionnaire sent as an email attachment
(111). This method facilitates the use of tailored questions and question branching and
for different responses is easy to implement. It hastens data collection and also reduces
the data entry error since the interviewer (or respondent) enters the data directly into a
computer during the interview and can check for correctness. However, a potential
problem of one type of electronic data collection is that not every household or

individual has lntemet access (110).

In our study because of the time and budget constraints we could not use face-to-face
interview for data collection. Additionally, because of limited accessibility associated

with electronic data collection methods, their use was not feasible. Therefore, parents

36

 

were given the option to fill out a self-administered mail-in questionnaire or participate

in a 15-20 minute phone interview with a trained interviewer.

Data collection process:

An introductory cover letter describing the study’s aim, methods, anticipated length of
the interview, and potential risks and benefits associated with the study was mailed to
the parents or guardians of each potential case. An informed consent form and self-
addressed stamped envelope were also included (Appendix 2 & 3). If the informed
consent form was not returned within one week after mailing the invitation letter, a
follow-up phone call was made to inquire about the willingness to participate in the
study. Cases with a missing residential address but a valid phone number were
contacted over the phone. Either a signed consent form or oral agreement was secured

from each case’s legal guardian before the child was enrolled.

Duration of exposure assessment:

Based on the perceived period of (8 hours —72 hours) incubation before symptoms of
Salmonella infections, food history and other exposures were collected for the three
days preceding the illness onset date for cases. Similarly, for control exposure
assessment, we gathered data for the three days prior to the interview day. This
approach is similar to the data collection efforts of the CDC-FoodNet during their
national investigation of the risk factors for foodbome diseases as well as case control

studies conducted by other investigators (87, 112).

37

ETHICAL CONSIDERATIONS
Informed Consent:
Parents or caretakers of children were adequately informed about the study’s aims,
methods, risks, and benefits. Informed consent was obtained from parent(s) or

caretakers before enrolling the children into the study (Appendix 3).

Data confidentiality:

Data were de-identified by removing names and home addresses soon after the
completion of the interviews. An identification code was assigned to each study
subject before entering the information into a database. Data were stored in a
password-protected computer. The parents were assured of the confidentiality of the

information gathered.

Approval from the Institutional Review Boards (IRB) for human research:
The study protocol was reviewed and approved by the Community Research
Institutional Review Board (CRIRB) at Michigan State University (MSU) and the

Institutional Review Board (IRB) at the MDCH (Appendix 1).

38

SAMPLE SIZE CALCULATIONS

A sample size calculation to achieve at least 80% study power, while restricting the
probability of type I error to 5% (alpha), was based on the assumptions of prevalence of
select exposures in the reference (control) population: family member had symptoms of
GI prior to child’s illness onset and exposures to pet, particularly reptiles. A case to
control ratio of 1:1 was used.

In order to detect an unadjusted OR of 2.5 between cases and controls for select
exposures having a prevalence ranging fi'om 15% to 20% in the reference (control)

population, we calculated the following sample sizes (5 8) (Table 21).

Since sample size of 124 x2=248 was large enough to detect the stated OR at desired
precision at any presumed prevalence, we chose this sample size and recruited the
required number of cases accordingly.

We assumed that 15% of reported cases would not have accurate contact information
available (15% of 124=~17). Further, we anticipated a ‘decline to participate’/ non-

consenting rate of 19% based on published literature (19% of 124=~24).

Thus to enroll 124 cases, we needed to approach a minimum of 165, (124+17+24= 165)
salmonellosis cases. Based on the past five-year MDCH data (2000-2004), an average
25 cases of salmonellosis of ages 10 or younger were reported to MDCH each month.
The anticipated duration to enroll the required number of cases was about 8 months.

Efforts to enroll controls were performed in parallel following each case’s enrollment.

39

 

DATA MANAGEMENT

Epi-info program was used to enter the data (version 6.04, Atlanta, GA; Centers for
Disease Control and Prevention, 1995). All of the data were entered by one operator.
About 10% of the entered questionnaires were randomly selected and checked for data
entry errors. The error rate was < 3%. Data were cleaned to remove data entry and
other logical error using Statistical Program for Social Sciences SPSS (version 10.0).
All questions were coded into numerical terms. Continuous variables, if necessary,

were categorized on biologically plausible or logical grounds.

A brief description of select variables follows:

Variable description and transformation:

Dependant variable:

Incident laboratory-confirmed Salmonella infections cases of human origin in children
aged 5 10 and reported to MDCH via MDSS from December 15, 2006 through October

15, 2007 were included in this study.

Independent variables:

Age:

For cases, age of the child was extracted from the MDSS and was also verified by the
parent(s) or caretakers. For control children, age reported by the caretaker was
recorded. We categorized age into two groups for statistical analysis: <1 year, 1-10

years.

40

 

Race:

Race (of child) was reported by the parent(s) or caretaker of the case or control
according to the race with which they most closely identify. Based on the distribution,
we categorized race into Caucasian, African-American, Other (any race other than
Caucasians and Afiican-Americans e. g., Alaskan Indian, Middle Eastern, Asian and
Pacific Islander etc) and Multiracial. Children born to parents of different races were

classified as ‘Multiracial.’

Parental education:
The highest level of schooling completed by the interviewing parents was recorded into
one of the following categories: 1) no formal education to high school, 2) some college

to a four year college degree, and 3) higher than college degree.

Household income:

Participants were asked to identify the income category in to which their household
income would fall. The responses were categorized into the following income
categories: 1) < $20,000, 2) $20,000 - $35,000, 3) $35,001 - $50,000, 4) $50,001 -
$75,000, 5) $75,001 - $100,000, and 6) > $100,000. Annual household income

included wages, salary, bonuses, or earnings from self-employment.

41

Area of residence:

Based on zip code level median household income reported by the US Bureau of
Census 2000, we divided the area of residence into high income (income $ > 60,000),
medium income (income between $38,000 and $ 60,000) and low-income ($< 38,000)

neighborhoods.

Breast-feeding:

In this study, exclusive breast-feeding was defined as breast-feeding only with no
formula or consumption of semi-solid or solid food in the three days before illness
onset. A binary variable was created ‘exclusively breast-fed vs. ‘non-exclusively breast

fed.’

Formula milk used:

Formula use was categorized into ‘no formula use’ and ‘formula use’ (if the caretaker
used formula in addition to breast feeding, or formula alone) to feed the child during
the 3 days prior to illness onset, the variable was categorized into ‘formula use’ and no

‘formula use.’

Pacifier use:
Parents were asked if the child used a pacifier during the 3 days prior to illness onset,
regardless of the duration of use. The responses were coded as ‘used a pacifier’ and

‘did not use a pacifier’.

42

 

Daycare / pre-school / school-related questions:

Parents were asked if their children attended day care or school, and if they did, then
greater detail such as number of hours per week child spent in daycare, total number of
children enrolled in daycare, number of children sharing the same room in daycare, etc.

were also asked.

Food exposures:

Several questions were used to collect information about the consumption of foods
known to be associated with Salmonella infections. Egg consumption in the 3 days
prior to the child’s illness onset was categorized into three dummy variables: 1) ate
firlly cooked egg, 2) ate partially cooked egg, 3) ate firlly cooked and partially cooked
egg, and 4) did not eat egg. Similarly, parents were asked if the child had consumed
poultry in the 3 days before illness onset, and the responses were grouped into: 1) ate at
home; 2) ate outside home, 3) ate at home and outside, 4) did not eat, and 5) do not
remember. In a separate question, similar inquiry was made about the consumption of
meat other than poultry. In addition to capturing the data for 3 days prior to illness
onset, we asked parent(s) about the average frequency of eating food at commercial

food establishments and child’s preferred food at these places.

Kitchen practices:
Family kitchen practices were assessed by asking multiple questions. The variable

‘cleaning kitchen counter after preparing raw chicken and meat’ was categorized into:

43

never, sometimes, and always. Questions regarding ‘how do you clean the counters
was categorized into 1) with soap and water, 2) with a disinfectant, and 3) both.

We also asked whether the ‘family kept eggs in a refrigerator’, which was grouped into
never, sometimes, and always.

Contact with animals:

Animals, particularly reptiles, are known to carry Salmonella bacteria (21). Parents
were asked if the child had contact with any pet (household pet, someone else’s pet, or
an animal in a petting zoo) during the 3 days prior to illness onset. If the response was
yes, then a second question asked about the type of animal (i.e., dog, cat, reptile, bird,
hamster, gerbil, or ferret). The response was coded as a binary variable (e.g., contact

with a reptile vs. no contact).

Contact with a person with symptoms of GI upset:

In two separate questions, parents were asked if the child had contact with a household
member or house visitor with symptoms of stomach upset (e.g., diarrhea, vomiting)
during the 3 days prior to the child’s illness onset. Based on these responses, a
dichotomous variable was created: ‘contact with a person having symptoms of GI

upset’ and ‘no contact’.

Parent(s) or caretakers who did not provide information in response to any of the above

questions were coded as ‘refused to answer’.

44

STATISTICAL ANALYSES

Descriptive statistics:
Counts and percentages for each categorical variable were computed. Socioeconomic
and demographic characteristics between cases and controls were compared using the

chi-square test for two proportions (107).

Inferential statistics:

Univariate analysis:

Exposure variables were categorized into two or more levels, using the category with
(or expected to have) the lowest risk of infection as the reference. Logistic Regression
(LR) was used to examine associations between predictor variables (cases’ socio-

demographic characteristics and hypothesized risk factors) and Salmonella infections.

Multivariate analysis:

Since we did not match cases with controls on potential confounders, we used the
unconditional multivariate statistical analysis to obtain the independent effect of
exposures on the outcome variable while controlling for potential confounders.
Adjusted odds ratios (AORs) with their respective 95% confidence intervals (95% CIs)

were computed (107).

All variables with estimates of a p value < 0.25 on univariate analyses, along with those
hypothesized a priori as putative risk factors for Salmonella infections, were considered

for inclusion in the multivariate model. We checked the correlations between potential

45

correlated variables in our dataset. After identifying the variables, we examined the
impact of co-linearity by separately entering the correlated variables (e. g., income,
education, area of residence, and race) into the multivariate regression model. We
developed several multivariate models and obtained the adjusted effect size for each
collinear variable. Beside main effects, we also included several two way interaction
terms in the model such as household income and race, household income and reptile
ownerships, and age group and reptile ownership. We used the backward elimination
method to obtain a parsimonious but yet plausible model. Exposure variables reported
as risk factors for salmonellosis in previous studies such as consumption of eggs/egg-
containing products, poultry,meat were kept in the final model regardless of their
statistical association with the outcome variable. The final model was adjusted for age-

group. The goodness-of-fit model was checked by using the Pearson chi-square test.

Calculation of Population Attributable Risk (PAR%):
Population attributable risk (PAR) for selected exposures was estimated
using adjusted 0R5 from the final multivariate model (Table 12). We used Levin's

formula (see below) for the calculation of PAR (113).

190-0
p(r — 1) +1

 

Where:
p = Proportion of the population with exposure

I“. = Adjusted Odd ratios

46

 

Levis formula has been shown by Leviton (114) to be algebraically identical to the

formula:

Incid_ence in total population — Incidence in non-exposed group

Incidence in total population
Since, our study was a population-based study, had the control children developed
Salmonella infections, they would have reported to MDCH. Therefore, our cases and
controls arose from the same source population, which is a requirement for the

calculation of PAR% from case-control design.

Subgroup analyses:
To better understand the relative contribution of certain exposures to Salmonella
infections in children, we restricted our data to selected variables and performed the

following subgroup analyses:

Analysis 1:
A few investigations have suggested that the magnitude of association between certain
exposure variables and Salmonella infections varies by age within the pediatric
population. To estimate the effect of selected risk factors (those with significant
associations in multivariate analyses: contact with a person having GI upset and contact
\with reptiles and cats), we computed the AORs for the following age categories: < 1

year, and 1-10 years.

47

Analysis 2:
To identify associations between selected food related exposures and Salmonella

infections in children aged <1 year

Analysis 3:
To identify associations between selected food-related exposures and Salmonella

infections in children aged 1-10 years.

Analysis 4:
To study the risk factors for the most common Salmonella serotype Typhimurium, we
restricted the data to cases (n=3 6) with Salmonella serotype Typhimurium only and

compared it with controls (n=139) for selected exposures.

48

RESULTS
Case enrollment:
During the 10 months of data collection (December 15, 2006 — October 15, 2007), a
total of 862 cases of salmonellosis were reported to the MDSS, of which 228 (26.45%)
occurred in children aged 5 10 years. Figure 6 describes the enrollment of cases. Of
the 228 cases in children, 29 (12.72%) were infected with typhoidal Salmonella
serotypes (S. Typhi and S. Paratyphi). Twenty-nine cases (12.72%) could not be
included due to incomplete mailing addresses and/or missing home phone numbers. A
letter of invitation was sent to 170 (74.56%) potential case households. The letter
included a self-addressed stamp envelope and consent form. One case (0.59%) was
ineligible on initial screening because of a coexisting chronic condition. Of 169 cases
we attempted to enroll, 10 (5.91%) declined to participate: 4 (2.35%) on mail-in
invitation letter and 6 (3.53%) over the phone. Of 159 (93.53%) eligible cases who
agreed to participate either by returning the signed consent form via mail or on a follow
up phone call, 36
(22.64%) could not be re-contacted again to complete the questionnaire despite
repeated attempts. A total of 123 of 159 eligible cases were enrolled during the study
period yielding a participation rate of 72.35%. Of the 123 interviewed cases, 102
(82.94%) were interviewed over the phone and 21 (17.06%) were interviewed through

a mail-in questionnaire.

49

Comparison of case participants and non-participants:

Since public health officials routinely collect demographic information from all
reported cases during the disease investigation process, we were able to compare the
demographic characteristics (age, sex, race) of enrolled cases with non-enrolled
children.

A total of 76 cases of salmonellosis in children aged 5 10 years reported during the
study period were not enrolled in our study (Total cases in children aged 5 10 years —
cases with Typhoidal serotypes — cases could not be contacted and refused to
participate, 228-29-123=76). The enrolled cases (participants) did not differ
significantly from non-enrolled cases (non-participants) with respect to demographic
characteristics, including age group (p = 0.42), sex (p = 0.55), and race (p = 0.78).
Table 8 shows a comparison of demographic characteristics between participants and

non-participants in Salmonella case-control study, Michigan, 2007.

Distribution of Salmonella serotypes among case children:

Salmonella serotype information was available for 199 of 228 cases in children aged S
10 years reported to MDCH during the study period. Among these cases, the four most
conunon Salmonella serotypes included S. Typhimurium (20.10%), S. Enteritidis

(8.54%), S. Newport (3.52%), and S. Heidelberg (3.52%) (Table 9).

50

Comparison of Salmonella serotypes between cases aged 5 10 years and cases aged
2 11 years:

The most common Salmonella serotype isolated among case children (participants and
non-participants), S. Typhirnurim, made up 20.10% of cases compared to 13.50% in
cases aged 2 11 years during the study period. S. Enteritidis was the second most
common serotype (8.54%) isolated in case children, while S. Enteritidis was the most
common serotypes in cases aged 2 11 years accounting for 25.21% of the cases. Table
9 compares the distribution of Salmonella serotypes between children aged 5 10 years

and reported cases 2 11 years during the study period.

Control enrollment:
Figure 7 describes the enrolhnent of controls. A total of 139 control children were

obtained using one of the following two methods:

Method 1:

A total of 37 potential controls were identified by the interviewed case parents.
Twenty-eight (75.68%) of the 37 households were contacted and interviewed, while 9
(24.32%) either could not be reached despite repeated phone calls or refused to

participate in the study (Figure 7).

Method 2:

A total of 2,463 control addresses and phone numbers were obtained from the on-line

white pages using the second method (Figure 7). Of these, 445 (18.07%) were

51

disconnected phone numbers and 53 (2.15%) were commercial phone numbers. After
excluding these phone numbers, there were 1,965 potential control households used to
identify appropriate controls. Of these potential controls, 1,134 (46.04%) could not be
reached due to receiving an answering machine on repeated calling including evening
calls, no answer, or a busy phone line. Additionally, 338 (13.72%) declined, hung up
the phone, or said they were not interested in the study. Four hundred and ninety-three
phone numbers were left of which 371 (15.06%) households had no children or no
children aged 5 10 years of age. Of the 122 that scheduled an interview call, 1 1(0.45%)
could not be contacted again after multiple calls. Therefore, a total of 111 parents

(4.51%) of control children aged _<_ 10 years were interviewed using the second method.

Descriptive statistics of cases and controls:

Table 10 compares the socioeconomic and demographic characteristics of the enrolled
123 cases and 139 controls. The enrolled cases and controls did not differ by
socioeconomic characteristics including parental education (p = 0.94) and annual
household income (p = 0.34). Similarly, there were no significant age group (p=0.54)
and gender differences (p = 0.86). However, significant differences in the distribution
of racial composition (p < 0.01) were noted between enrolled cases and controls, which
warranted control of this variable in the analysis stage. F orty-seven study subjects (22
cases and 25 controls) were aged < 1 year, 132 children (66 cases and 66 controls) were
aged 1-5 years, and 83 subjects (35 cases and 48 controls) were aged 6-10 years.
Overall, the rate of refusal to the question about parental education attainment was low

(< 2%) among case and control households. However, the refusal rate for the question

52

regarding household income was lower (10.66%) among case households compared to
refusal rate of 17.39% in control households.

One hundred and ninety-nine (75.95%) study subjects were Caucasian (107 cases and
92 controls), 36 (13.71%) were African-American (9 cases and 27 controls), and 24

(9.16%) belonged to other minority groups (6 cases and 18 controls).

Inferential statistics
Univariate analyses:
The results of the univariate analysis of putative risk factors for Salmonella infections

in children aged _<_10 years are presented in Table 11 and briefly summarized here.

Household related variables:
Case and control subjects did not differ significantly with regard to household
characteristics including number of people residing in the house, presence of other

siblings aged :10 years, and type of family room flooring.

Daycare and/or school-related exposures:

Data showed that salmonellosis was significantly associated with attending a daycare
facility (OR = 2.31, 95% CI: 1.01 - 5.40). However, no statistically significant
association was found between attending a school and salmonellosis in the univariate

analysis (OR = 0.84, 95% CI: 0.51 - 1.36).

53

Food consumption and family kitchen practices:

When studying associations between food consumption and Salmonella infections,
univariate analysis did not show an association with the consumption of eggs or
products containing eggs, eating poultry or meat, eating at commercial food
establishments, and source of drinking water during the 3 days prior to the illness onset
date (in cases) or interview date (in controls). Moreover, none of the family kitchen
practices related variables was identified as a factor associated with the outcome

variable.

Person-to—person transmission:

A total of 27 cases (21.95%) and 16 controls (11.5%) had contact with a sick person
(having symptoms of GI upset) within 3 days prior to the onset of illness for cases and
within 3 days before the interview day for controls. Our data suggest that contact with
a person having symptoms of GI infection increases the odds for contracting
salmonellosis (OR = 2.16, 95% CI: 1.10 - 4.24). Persons with symptoms of GI upset
that had contact with the cases were either a family member, a visitor of the child’s

home, or someone the child visited.

Contact with animals:

Exposure to an animal during the 3 days prior to the child’s illness onset was
significantly associated with salmonellosis (OR = 2.69, 95% CI: 1.62 - 4.45).
Analyses showed that having contact with reptiles (OR = 4.29, 95% CI: 1.53 - 12.02)

was significantly associated with Salmonella infections. A total of 14 cases were

54

exposed to reptiles. The reptiles to which case children reported exposure included
iguanas [1 (6.25%)], lizards [3 (18.75%)], snakes [6 (37.5%)], and turtles [7 (43.75%)].

In addition, 3 cases had contact with frogs and 1 case had contact with an alligator.

Contact with cats showed an association with salmonellosis (OR = 2.23, 95% CI: 1.21 -
4.10). Of the 138 cases interviewed, there were 35 cases that had contact with cats and
34 (89.4%) of these case children contacted cats aged > 1 year, while 4 (10.6%)
contacted cats aged 5 1 year. A total of 53 cases had contact with dogs and, 33 (91.7
%) of these case children had contact with dogs aged >1 year, while 3 (8.3 %) had
contact with dogs aged 5 1 year. Having contact with dogs, birds or hamsters did not

show an association with salmonellosis.

Other environmental exposures:

Caretakers’ handling packages of raw meat or chicken without gloves or plastic bags
during grocery shopping with a child did not show an association with the outcome
variable in our data. Additionally, there was no association found between the variable

‘placing child on the floor without a blanket’ and Salmonella infections.

Multivariate analysis:

Table 12 shows the results of a multivariate logistic regression model. The final
multivariate model, after adjusting for age group revealed that having salmonellosis
was significantly associated with contact with cats (adjusted odds ratio (AOR) = 2.62,

95% CI: 1.17 — 5.87) and reptiles (AOR = 8.16, 95% CI: 1.55 — 42.88). Additionally,

55

attending a daycare center (AOR = 4.86, 95% CI: 1.44 — 16.37) and contact with a
person having symptoms of gastrointestinal infection during the 3 days prior to the
onset of child’s illness was significantly associated with Salmonella infections (AOR =
2.27, 95% CI: 1.02 — 5.44). None of the two-way interactions tested were statistically
significant (p > 0.05). Therefore, the final multivariate model only included the main
effects. The model was adjusted for age category and race. Pearson chi-square

goodness-of-fit showed a good model fit (p = 0.69).

Population Attributable Risk (PAR%) for selected variables:
The PAR% of 19.98%, 19.65%, 20.45% and 12.75% was estimated for attending a
daycare center, contact with cats, contact with reptiles and contact with a person having

symptoms of GI infection respectively (Table 12).

Findings from subgroup analyses:

Analysis 1:

Contact with reptiles (AOR = 3.57, 95% CI: 1.04 - 12.25) and cats (AOR = 2.28, 95%
CI: 1.01 - 4.28) were significantly associated with the Salmonella infections in children
aged 1-10 years, while having contact with a sick person during the 3 days prior to
illness onset did not show association in children aged 1-10 years.

None of the three variables were significantly associated with Salmonella infections in

children aged < 1 year (Table 13).

56

Analysis 2:

Table 14 shows selected potential risk factors for Salmonella infections in children
aged < 1 year. Salmonellosis was significantly associated with attending a daycare
facility (OR = 2.31, 95% CI: 1.07 - 5.40) in children aged < 1 year. Additionally, a
greater number (> 6 vs. S 6) of children in the room of a daycare center increased the
odds for Salmonella infections by 16-fold (OR = 15.99, 95% CI: 1.38 - 185.39).

Other daycare-related variables including hours spent in a daycare per week, number of
children in the daycare, number of children in diapers in the same room, having a child
with symptoms of GI upset, and having a separate diaper changing area in the daycare,
along with food-related exposures such as formula use and pacifier use did not show

associations with the Salmonella infections.

Analysis-3:
Table 15 shows the results of a univariate analysis of selected risk factors associated
with Salmonella infections in children aged 1-10 years. None of the school- or food-

related risk factors were found significant with Salmonella infections.

Analysis-4:

Among case children, 36 (29.20%) were infected with S. Typhimurium, the most
common Salmonella serotype in the US. To study the association between selected
exposures variables and S. Typhimurium, we restricted our data to serotype S.
Typhimirum cases and calculated the ORs with 95% CIs. Table 16 shows the

demographic characteristics of cases and controls and Table 17 shows the results of

57

univariate analyses of selected risk factors for S. Typhimurium in children aged 1-10
years. Having contact with any animal (OR = 3.22, 95% CI: 1.47 - 7.08), reptiles (OR
= 5.40, 95% CI: 1.37 - 21.29), and birds (OR = 12.45, 95% Cl: 1.25 - 123.56) was

significantly associated with S. Typhimurium.

58

DISCUSSION

This population-based case-control study was designed to identify potential risk factors

for Salmonella infections in Michigan children aged S 10 years.

Validity of findings:

The validity of our study results, or degree to which the results are free from error for
the study sample being studied has been evaluated by: 1) reviewing the study design, 2)
selection of subjects, 3) comparability of cases with controls, 4) exposure assessment,
and 5) statistical analysis based on selected variable specific to the case-control study
design (interval validity) to identify if the results are threatened by any systematic

(bias) or random error.

The case-control design was selected as the main approach because we wanted to study
multiple potential risk factors associated with salmonellosis (prediction model). The
case-control design was therefore the most appropriate design; investigation begins
with diseased (cases) and non-diseased (controls) and retrospectively ascertains

exposures between the two groups.

One of the central issues in case-control studies is the comparability of control subjects
to case subjects. Since we used the MDCH database to identify cases, our cases were
picked from across the state of Michigan in accordance with the times of their
reporting. For the enrollment of controls, we adopted a method that would provide
controls from the same base population. We have compared the recruited control
subjects with cases and did not found that they were significantly different on selected

socioeconomic characteristics.

59

In order to minimize the measurement error, we used validated questions used in other
similar research studies, and our questions were pre-tested before their use on study
subjects. Additionally, we used a detailed questionnaire to obtain data on plausible
sources and a priori risk factors for salmonellosis reported in earlier studies, and control
for confounders was possible in the multivariate analysis. Therefore, the results of our
study may serve as fair and statistically unbiased estimate for the risk factors of

salmonellosis in Michigan children.
Criteria for causal relationship between exposure and outcome:

Despite the inherent weakness of the case-control study design in establishing causal
relationship between exposures and an outcome, a carefully designed case-control study
can still be valuable in supporting the causal relationship between exposures and
outcomes. We therefore evaluated our study findings using the Asutin Bradford Hills
(1897-1991) Criteria of Causation - the conditions proposed to improve the likelihood

of a causal relationship between risk factors of interest and disease being studied.

Although they are not rigid criteria the fulfillment of all may not be accomplished, they

still give positive support to inferences about causality.

Biological plausibility:

It has been established that humans and animals harbor Salmonella serotypes and an
infected person or animal can shed the bacteria in their feces. In a study conducted by
Shutze et al in Arkansas n 1999, Salmonella has been isolated from household

members, pets including cats and reptiles, and various places in the household such as

60

kitchen counters, bathrooms, and flooring. Through the fecal-oral route, direct or

indirect transmission of an infection can occur.

Magnitude of association:

In this study, the magnitude of statistical association between the predictors and
outcome variables is measured by the odds ratio. The stronger the association, the
more likely it is that the relationship between the two variables is causal. In this study,
the odds of exposure to the respective risk factors was at least 2 folds or higher among
cases compared to controls, supporting a causal relationship between the specific

exposure and sahnonellosis.

Consistency of findings:

This criterion implies that if a relationship is causal, we would expect to find it
consistently in different studies and in different populations. This is why numerous
studies have to be done before meaningful statements can be made about the causal
relationship. The associations between exposure variables and development of
salmonellosis observed in our study are consistent with the results of similar studies
using different settings, populations, and methods. A recent large population-based
case-control study conducted by Jones and colleague at the FoodNet sites (89) showed
that contact with reptiles and infected persons are significant risk factors for Salmonella
infection in children. Similarly, our results are consistent with the findings of studies

carried out outside the US., in France and Island of Guam (96,104).

61

Coherence:

The association should be compatible with existing theory and knowledge. In other
words, it is necessary to evaluate claims of causality within the context of the current
state of knowledge within a given field. As mentioned earlier that both animals and
humans harbor Salmonella and infected humans and animals can transfer infection
through direct contact and indirectly by contaminated environment. Therefore, in our
study the observed results are compatible with the existing theory of and knowledge of

Salmonella infection transmission.

Temporality:

According to this criterion, exposure always precedes the outcome. Hill emphasizes
the criterion of temporality as necessary, because in order for exposure to cause disease,
exposure must precede disease in time. However, this criterion is usually restricted to
prospective studies where we follow a cohort of exposed and unexposed people and
look for the disease of interest within the cohort. However, in the majority of case-
control studies temporal relationship between an exposure and an outcome cannot be
ascertained. In our study, ascertainment of the exposures was made retrospectively for
the 3 days prior to the child’s illness onset. Based on the formulated questions to gather
the exposure data in our study, it is likely that the exposure preceded the diseases

particularly in cases of contact with animals and person with symptoms of GI infection.

62

 

Dose-response relationship:

An increasing amount of exposure increases the risk. If a dose-response relationship is
present, it is strong evidence for a causal relationship. In our study, because of the
limited sample size, we could not assess the dose- response relationship between

exposure variables and Salmonella infections.

Specificity:

Specificity is established when a single putative cause produces a specific effect
(outcome). This is considered as the weakest of all causal criteria. When specificity of
an association is found, it provides additional support for a causal relationship.
However, absence of specificity does not negate a causal relationship. Diseases are
often caused by multiple factors, and it is rare to find a one-to-one cause-effect
relationship between an exposure and a disease. Since humans can acquire
salmonellosis from multiple sources, such as contaminated food, water, or

environmental sources, this criterion is not met in our study.

Although the general findings of this study were not unexpected and in many ways in
agreement with other studies conducted to identify risk factors for salmonellosis in
children (89), the risk factors for Salmonella infections identified in this study differ
fiom those identified in the adult population reported by previous studies. The
differences may be partially attributable to markedly different dietary and other
environmental exposures. Data from this study demonstrated that reported cases of

laboratory-confirmed salmonellosis in children are associated with some potentially

63

 

modifiable risk factors and thus can be used as the base for strengthening the

implementation of the disease prevention efforts.

Person-to-person transmission:

Data from this study suggest that children with Salmonella infections were more likely
than controls to have been in contact with a person with symptoms of GI infection
during the three days before to illness onset. Evidence of person-to-person
transmission of Salmonella infections among family members of different age groups
(115), nursing home residents (116, 117), children attending daycare (118, 119),
children in schools (120) and hospital patients (121-124) has been well established in
epidenriologic studies. A population-based case-control study conducted using
FoodNet sites reported about a l3-fold increase for salmonellosis in children residing in
households where a member had diarrhea in the 4 weeks before illness onset (125).
Similarly, in France, Delarocque-Astagneau et al. (1998) reported an association
between diarrhea] symptoms in a household member and Salmonella infections in
children aged < 5years (58, 96). Among household members, transmission may occur
directly or indirectly (i.e., fomites). It has been suggested that an innoculum of about
107-108 colony forming unit (cfu) of non-typhoidal Salmonella is usually required for
person-to-person transmission (126). In our data, 27/123 (22%) cases reported contact
with a person with GI upset symptoms during the 3 days prior to illness onset and 13
(48%) of these cases had family members that had been diagnosed with Salmonella
infections. In a Dutch study in children, household members of cases revealed that

about 42% of the families had at least one family member with a culture positive for

64

Salmonella (67). Similarly home investigations conducted in family members of
children aged < 4 years diagnosed with Salmonella infections in Arkansas showed that
about 14% of case children had a family member with a positive culture for Salmonella

(101).

Contact with animals:

Salmonella is a well-recognized zoonosis (49). Animals are the predominant reservoirs
for the bacteria, and the prevalence of Salmonella carriage varies by species. (11, 127).
Salmonella serptypes have been isolated from most vertebrates including dogs and cats
that have reported carriage rates of up to 36% and 18%, respectively (128). However, a
much higher (up to 94%) Salmonella carriage rate has been observed in reptiles and
amphibians (21). After infection, dogs and cats can remain asymptomatic and may tend
to shed Salmonella for prolonged periods of time (129). In our study, children who had
contact with cats during the three days before their illness onset were more likely to
acquire Salmonella infections than their counterparts who did not have contact with
cats. Most reports that have addressed the association between human salmonellosis
and cats are case series where investigators reported the proportion of cats with positive
Salmonella infections (130-133). The risk associated with Salmonella transmission
from cats to humans has not been evaluated in analytical studies. Sources of
Salmonella infections in cats vary and depend on whether the cats reside indoors or
outdoors. For indoor cats, the most likely exposure is the consumption of food
contaminated with Salmonella organisms, whereas outdoor cats may be exposed

through scavenging and hunting prey, especially birds (128). Salmonella serotypes that

65

 

have been isolated from infected cats include S. Typhimuirum, S. Enteritidis, S. Anatum

and S. Derby (128).

Exposures to reptiles:

Our study suggests that having contact with reptiles was significantly associated with
Salmonella infections in children (AOR=8.16, 95% CI, 1.55-42.88). Additionally,
reptile exposure had the highest PAR% (20.45) among the risk factors identified for
salmonellosis. A large number of epidemiologic studies have repeatedly shown that
exposure to turtles, lizards, and snakes have been associated with an increased risk of
human salmonellosis, particularly for young children (134-138). In a recent case-
control study conducted using F oodNet sites, Jones TF reported that reptile ownership
is associated with a more than 5-fold increased risk of salmonellosis in children aged <
1 year (89). Similar or even higher magnitudes of association between children’s
contact with reptiles or amphibians and Salmonella infections have been observed in
other epidemiological investigations (137).

The prevalence of Salmonella infections in exotic animals kept as pets is reportedly
highest in reptiles and amphibians. An estimated 90% of all reptiles, in particular,
turtles and iguanas, carry and shed Salmonella in their feces intermittently (139).
Attempts to treat reptiles with antibiotics to eliminate Salmonella carriage have been
unsuccessful and can increase the development of antibiotic resistance (140).
Salmonella survives well in the environment (141) and can be isolated from surfaces
contaminated by reptile feces for prolonged periods of time. Direct transmission of

Salmonella occurs by handling of a reptile, and indirect transmission by contact with an

66

environment contaminated by reptile feces. However, it is more likely that they were

infected indirectly through a Salmonella-contaminated environment.

Although our study sample consisted of children within a narrowly defined range (aged
5 10 years), transition of a child from an infant to toddler brings about significant
changes in activities of children, that can profoundly affect their exposure to various
known risk factors of Salmonella, particularly contact with household pets. We
therefore performed an additional sub-group analysis of children < 1 year and those
aged 2 1 year separately (Table 13). The analysis showed that exposures to cats and
reptiles were significantly associated with salmonellosis in older children (2 1 year).
This is a plausible finding because the mobility of older children (2 1 year) allows
greater chances of direct contact with these pets and results in an increase acquisition of
infection. The subgroup analysis in children aged < 1 years showed less significant
associations between exposure to reptiles and contact with a person having symptoms
of GI infection and salmonellosis. Although plausible, this finding may also be
explained by the smaller sample size. In a recent case-control study, Jones and
colleagues, identified both of these exposures as important risk factors for Salmonella
infections in children aged <1 year. It is conceivable that this age group acquire
Salmonella infections by indirect transmission from contaminated home enviromnent or

via parents or other family members.

Salmonella serotypes that are commonly isolated from exotic pets, particularly iguanas
and turtles, include S. Java, S. Stanley, S. Poona, S. Litchfield, S. Manhattan, S. Miami,

S. Jangwani, S. Tilene, S. Arizonae, and S. Rubislaw (142).

67

During the 19705, small pet turtles were identified as a major source of Salmonella
infections in the US. In 1975, the FDA banned the sale of small (i.e., < 4 in. long)
turtles. This resulted in a substantial decrease in cases of salmonellosis (143).
However, reptiles remain popular pets in the US. The increase in pet reptile popularity
has been paralleled by an increase in the number of reptile-related Salmonella serotypes
isolated from humans (137, 144). According to the American Veterinary Medical
Association (AVMA), as many as 2.8 million reptiles were owned as pets in 2001 and
1.5- 2.5 million US households (1.6%) had a pet reptile (145). Applying these
estimates to the Michigan population, 60,570 Michigan households would have reptiles
as pets (142) .

Frogs and toads also carry Salmonella and have been linked to salmonellosis outbreaks
in humans (21, 146). In one reported case-control study, ownership of amphibians was
independently associated with Salmonella infections. It has been estimated that reptile
and amphibian contact account for 74,000 (6%) of the approximately 1.2 million
sporadic Salmonella infections each year in the US (21). In our study, however,
associations between amphibian and salmonellosis were inconclusive because only

three children reported contact with frogs.

Other animals, such as horses and cattle, have also been recognized as potential sources
of Salmonella for exposed individuals (e.g., veterinary clinicians and students, and
farmers) (147, 148). A recent study noted that inadvertent contamination of household
carpets with Salmonella serotypes can occur when veterinarians have occupational

exposure to cattle on farms (149). Additionally, pigs have been identified as a source

68

of Salmonella Choleraesuis infection associated with high mortality in humans (150).
However, exposure to these animals did not show significant association with

salmonellosis in our study.

Daycare-related exposures:

Children attending childcare centers experience a greater number of illnesses associated
with infections, particularly enteric infections, compared to children cared for at home
(151, 152). A study found that children who attended childcare required 40% to 80%
more medical consultations for acute infections than their counterparts who remained at
home (153). Wald and colleagues (1988) reported that children attending these centers
had 51% more episodes of infection than their counterparts who were cared for at home
(154). A large number of enteric infectious agents including Salmonella have been
associated with attending daycare (152, 155, 156). Concurring with these studies, our
data suggest that attending a daycare significantly increases the odds of salmonellosis
(Table 11). Furthermore, our subgroup analysis in children aged <1 year (Table 14)
showed that staying in a crowded room (> 6 children in the same room) at a daycare
increases the odds of contracting Salmonella infections (OR=15.99, 95% CI, 1.38-
18539). The spread of infections in daycare centers is facilitated by crowding and
microbial contamination of the childcare environment, as well as a greater
susceptibility of young children to infections. It has been suggested that direct contact
(person-to-person) is the major route of transmission in the majority of enteric
infections in daycare settings. However, indirect contact through contaminated fomites,

including toys and other shared items, can occur (157). Children, particularly young

69

infants, have habits that facilitate the dissemination of infection, such as putting their
hands and objects in their mouth (158). A study of bacterial contamination in daycare
centers found that the prevalence of infectious agents on the hands of daycare workers,
daycare surfaces, and in air samples was inversely related to the age of the children
attending the daycare. The likelihood of fecal contamination was greatest on the hands
of young infants and their caretakers, and least on those of the older children (159,
160)

It has been suggested that the transmission rate of an enteric infectious agent within a
childcare center is influenced by the characteristics of the daycare (number of children
enrolled, room size), children attending (age, length of time enrolled, immunological
status), and daycare workers (number of workers per child and workers’ level of
training) (155). Epidemiologic studies examining risk factors for diarrheal illness have
found that daycare centers with non-toilet trained infants and those in which food-
handling staff also changed diapers had higher diarrheal rates (155, 159).

Among identified risk factors for enteric infections in daycares, diaper changing is
considered the procedure with the highest risk for transmission between children and
workers (161). Although in a subgroup analysis, we evaluated the relationship between
several daycare-related questions and Salmonella infections, only one factor, daycare
crowding, achieved statistical significance with the outcome variable, likely because of

our sample size (Table 14).

7O

Food-related risk factors:

In the adult population, a wide variety of food-related potential sources of Salmonella
infections have been identified, which vary by serotype. Consrunption of contaminated
poultry and meats, particularly ground beef, has been identified for S. Typhimurium
(52) and S. Heidelberg, whereas eating eggs and egg products has been associated with
S. Enteritidis (5 9, 99). In children, acquiring infections through environmental

contamination is thought to be more common than via food vehicles (58, 96, 101).

In our study, food-related exposures such as consumption of chicken, meat, and
eggs/egg-containing products within three days before the child’s illness onset in cases
and prior to the interview day in controls did not show a significant association with
Salmonella infections. It is conceivable the actual magnitude of food exposure related
risk for Salmonella infections was not evident in this study due to several factors. First,
we obtained exposure information from interviewing surrogate sources—parents or
caretakers. While certain exposures such as contact with animals and sick persons are
more likely to be recalled, recall of consumption of specific foods is difficult and thus
prone to measurement error. We suspect the use of surrogate sources in our study may
have contributed to measurement error in the food-related exposures for older children.
It is conceivable that some of the older children in our study population might have
consumed food outside the home without informing their caretakers. This potential for
inaccurate measurement could have resulted in the absence of an association between

food exposures and Salmonella infections seen in our study.

71

Furthermore, the results of our subgroup analyses (2 & 3) examining the role of food-
related sources and salmonellosis separately in children aged < 1 year and those aged 1-
10 years did not show association with salmonellosis (Tables 13 & 14). However, our
sample size calculations were based on the prevalence of mainly environmental
exposures related to salmonellosis in the reference population. These null findings
could have resulted because of the limited sample size our study had to detect food-
related exposures between cases and controls. For example, breast feeding in children
aged <1 year has been shown to prevent salmonellosis in epidemiologic studies. There
is sufficient biologic evidence to support the protective effect of breast milk, that it can
provide host defense for the breast-fed infant against the majority of infections
including Salmonella (162).

Another possible explanation for the protective effect of exclusive breast-feeding could
be the characteristics of the environment of breast-fed infants. For example, non—
breast-fed infants often drink powdered formula, which has been shown to be an
important risk factor for infectious agents including Salmonella. Infant formula has
been shown to support the growth of Salmonella in epidemiologic studies but few
investigations have demonstrated that infant formula is associated with Salmonella
infections (163). Contamination of formula with Salmonella usually occurs during
preparation and handling, and growth of bacteria is highly likely if contaminated
formula milk is kept at room temperature for several hours. In our study, however
infant formula use did not show significant association with salmonellosis.

Pacifier contaminated with pathogens could also transmit infections to children. Some

studies have demonstrated an association between pacifier use in young children and

72

infectious agents. In our study pacifier use did not show an association with

salmonellosis.

When the data was restricted to serotype Typhimurium (subgroup analysis 4, Table 17),
consumption of meat other than poultry both at home and restaurants during the three
days before child’s illness onset or interview day showed a significant association at the
univariate level with S. Typhimurium infections (OR = 5.11, 95% CI: 1.07 — 24.30) in
children aged 1-10. This finding corroborates with the results of other investigations
that demonstrated an association between eating contaminated meat (164) and S.
Typhimurium. A large number of reports of foodbome disease outbreaks have been
traced back to the contamination of food during preparation and handling at restaurants.
Epidemiologic studies of both sporadic and outbreak-associated enteric disease cases
suggest that restaurants are an important source of foodbome disease in the US. During
1998-2004, the CDC reported there were 349 restaurant-associated outbreaks (165). In
a recent case-control study, consumption of chicken prepared outside the home was
associated with Salmonella serotype Enteritidis infection (72). Another case-control
study identified eating eggs that were prepared outside of the home as a risk factor for

Salmonella serotype Heidelberg infection (88).
In accordance with the findings of a case-control study conducted in the UK to identify

risk factors of Salmonella infections in kitchens, none of the household-related

variables in our study showed associations with the outcome variable (22, 141).

73

Drinking water is a major source of microbial pathogens in developing countries due to
poor sanitation and hygiene practices (166).

In industrialized countries, drinking municipal tap water or water from

private wells has not been identified as an independent risk factor for

enteric infections. Accordingly, our study did not find an association between drinking
tap or well water and salmonellosis. A few studies conducted in the US have shown
association between drinking untreated water from a lake, river, or stream and enteric

infections (167, 168). Few children drank water from these sources in our study.

Other environmental exposures:

A large number of studies have demonstrated that Salmonella can be efficiently
transferred from contaminated environments to infect humans, particularly children (89,
97). A recent study reported that children aged < 1 year who ride in a shopping cart
with meat or poultry placed next to them have a 4-fold increased risk for salmonellosis
(89). It has been demonstrated that substantial levels of contamination with foodbome
pathogens exist on the packaging of meats and poultry (169). In assessing the role of
the contaminated environment in our study, we asked parents if the child accompanied
a caretaker while grocery shopping and the whether the caretaker touched the packages
of meat and/or poultry without gloves or plastic. However, this variable did not show

an association with Salmonella infections in our study.

We also attempted to evaluate the role of in-house contamination and the risk of
salmonellosis in children aged < 1 year. There was no significant association detected

between the variable ‘placing child on the floor without a blanket’ and salmonellosis

74

(Table 14). Although some studies (104, 141) have reported the isolation of Salmonella
from household dust, soil samples near the home, and samples from bathrooms, we did
not evaluate contamination of the household environment through these means.
Travel-associated salmonellosis:

Individuals who travel to places where foodbome infections like salmonellosis are
prevalent (e. g., South America, Asia) are at a greater risk of contracting enteric
diseases. In contrast to the findings of other case-control studies conducted in adult
populations that have identified travel-associated risk factors for salmonellosis and
many other enteric pathogens, our study found no association between illness among
children aged 5 10 years and travel. Pathogens such as enterotoxigenic Escherichia
coli, Campylobacterjejuni, and Salmonella serotypes account for the majority of
diarrheal disease cases associated with travel. Nontyphoidal salmonellosis is mostly
caused by the Salmonella serotypes Enteritidis and Typhimurium; however, other

serotypes have been isolated from individuals with a history of recent travel.

Public health recommendations:

Exposure to cats:

Our study suggested a significant association between contact with cats and Salmonella
infections. Additionally, exposure to cats had a PAR of 19.65% for salmonellosis.

This finding should be viewed as a significant public health problem. Cats are among
the most widely kept pets in the US—it is estimated that about 34% of households have
at least one cat as a pet. The AVMA estimates that between 1996 and 2001, the US

population of cats increased 16 %, reaching 78 million cats in 2001 (170).

75

Educating pet owners about the safe handling of their pets, disinfection of contaminated
areas in the household, and restriction of contact with the family members who might
be at greater risk for developing the disease, particularly young children. Additionally,
cat owners should be informed of asymptomatic cat carriers of Salmonella. Older
children (aged > 5 years) should be educated about hygiene practices such as hand
washing after touching cats.

As stated earlier, risk of Salmonella transmission has not been examined in analytical
studies, additional epiderrriologic studies are needed to quantify the risk of Salmonella

transmission from infected cats to humans.

Exposure to reptiles:

Numerous public health recommendations regarding ownership and care of reptiles and
the potential risks of Salmonella exposure to children have been made (142). In 1999,
the CDC recorrunended that children aged < 5 years and irnmunocompronrised persons
should avoid contact with reptiles and that reptiles should not be kept in homes where
immunocompromised people or children < 5 years old reside (60, 135, 136). However
reptile-associated salmonellosis continues to be a major public health problem in the
US (60). Legislation requiring pet store owners to communicate the increased risks of
salmonellosis to customers who wish to purchase reptiles exists in several States (142).
Michigan requires consumer education regarding the risk of salmonellosis for the sale
of turtles (135). In 1999, the National Association of State Public Health Veterinarians
and the Council of State and Territorial Epidemiologists recommended that state and

local authorities adopt regulations to prohibit the sale of reptiles without written point-

76

of-sale education to consumers about the risks for and prevention of reptile-associated
salmonellosis (145). In 2003, the CDC gathered information from the health
departments in all 50 states and New York City to determine whether such regulations
existed. Among the 49 health departments responding, four states (Colorado, Illinois,
Kansas, and Texas) required pet stores to provide information about salmonellosis to
persons purchasing any reptile, and five states (California, Connecticut, Maryland,
Michigan, and New York) required providing salmonellosis information to persons
purchasing a turtle but not other reptiles. Tennessee prohibited the sale of all turtles,
while NYC prohibited the sale of certain reptiles, including iguanas, small turtles, and
boas, and required posting of information about reptile-associated sahnonellosis where
other kinds of reptiles were sold.

Pet-store owners, health-care providers, and veterinarians should educate owners and
potential purchasers of reptiles and amphibians about salmonellosis prevention
measures. A study reported that less than 50% of the families having iguanas as pets
realize their pets may carry Salmonella, demonstrating an inadequate knowledge about
potential Salmonella transmission (171).

It has been widely accepted that pets offer advantages in terms of providing
companionship for lonely individuals, and helping children develop a sense of care and
compassion. However, some pets, particularly exotic pets such as turtles and iguanas
are known for their carriage status of enteric pathogens (e. g., E. Coli, Campylobacter
and Salmonella), and thus can pose a risk to humans who contact them, especially
children. The risk increases when such pets are mishandled, for example by placing

turtles and iguanas in the bathtubs and failing to sanitize before human use (ref).

77

Mitigation of the risk for salmonellosis that is associated with exposure to pet animals
has been the subject of intensive efforts by federal and state agencies along with the

AVMA.

Recommendations for the prevention of reptile-associated salmonellosis:
Educating parents and caretakers regarding the risk of salmonellosis associated with

exposure to animals can help reduce the disease burden.

At the Federal or State level:

. Periodic assessment of compliance with Federal laws regarding the sale of small
sized turtles (<4 inches)

. Evaluation of the effectiveness of mandated point-of-sale education in reducing
amphibian- and reptile-associated salmonellosis.

0 Using mass media to educate parents regarding the risk of Salmonella
transmission from reptiles

. Prohibition of day care centers and preschools to house reptiles or amphibians.

- Integration of hmnan and veterinary surveillance systems and education of the

veterinary community on its role in public health

For veterinarians and healthcare providers:

. Encourage pet retailers (pet store owners), veterinarians, and healthcare
professionals to educate owners of reptiles or amphibians regarding the risk of

Salmonella infections associated with reptiles.

78

0 Provision of education by veterinarians to animal owners about the risk of
Salmonella transmission whether or not pets are exhibiting symptoms of

salmonellosis

For pet owners:

0 Additional efforts to educate reptile and amphibian owners of the potential for
Salmonella transmission fi'om pets using mass media; also educate older
children, aged > 5 years

0 Encourage parents with young children not to keep reptiles and amphibians in
the household

. Do not allow reptiles or amphibians to roam free in the living areas, particularly
the kitchen

. Wear gloves when cleaning cages and treating animals, and wash hands
thoroughly with soap and water each time a reptile or amphibian or its
equipment is handled

0 Do not clean reptile and amphibian cages and equipment in the kitchen or
bathroom sinks or tubs

0 Use designated tubs for cleaning equipment or bathing reptiles and disinfect
with a bleach solution after use

0 Immediately clean and disinfect areas contaminated with animal feces

79

Attending a daycare:

Illness in the daycare setting is a great concern of parents and a significant public health
problem worldwide. Simple measures to control and prevent infections such as
workers washing hands with soap and water after changing diapers, after assisting
children with the toilet, and before handling food would help to substantially reduce the
incidence of infections related to daycare. The effectiveness of hand washing has been
illustrated by a study that showed a markedly reduced incidence of diarrhea among
young children in child care centers after the introduction of an intensive hand washing
program for the workers (172). Disposable gloves should be worn for changing the
diapers and the changing station should be cleaned after each use. If possible, daycare
workers who handle diapers should not prepare the food. Another key measure in
controlling the spread of infections in daycares is to thoroughly clean the children’s

toys at the end of each day with hot water or disinfectant.

Food-related exposures:

Instituting safer food preparation practices in commercial kitchens could reduce much
of the risk associated with eating at commercial restaurants. Commercial food
establishments should take measures to ensure that meat, produce, and other foods are
obtained from high-quality suppliers. Educating and training restaurant workers is
important to ensure that safe food handling procedures are consistently followed.
Public health authorities should also regularly perform inspections of food
establishments and enforce regulatory policies. In addition, consumers should avoid

consumption of high-risk foods such as undercooked eggs and meat in commercial food

80

establishments.

81

STRENGTHS OF THE STUDY
Incident enrollment of cases:
We enrolled new cases of salmonellosis reported to MDCH. Parents of cases were
contacted for an interview soon after they were reported in the disease surveillance
system. Interviewing parents of case children soon after the onset of the disease and
questioning control parents about the 3 previous days allowed for better recall of food

history and related exposures.

Laboratory-confirmed cases of salmonellosis:

Bacterial cultures of samples obtained from patients with suspected Salmonella
infections are cultured for Salmonella in local laboratories across Michigan and then
sent to MDCH for confirmation and serotyping. We enrolled only laboratory-
confirmed (ie., on stool, urine, cerebrospinal fluid, or blood culture) cases of
salmonellosis reported to MDCH. Since microbiological laboratory culture is
considered the gold standard for diagnosis of salmonellosis, the use of this highly
sensitive and specific method for testing Salmonella infections minimized the chance of

misclassification of the disease status.

Generalizability of findings (external validity):

The MDCH surveillance system collects information from the entire state of Michigan,
therefore the participants in this case-control study are representative of all Michigan
children. Although not all reported cases were enrolled in the study, enrollees

(participants) and non-enrollees (non- participants) were drawn from the same

82

population base, and the two groups were similar with respect to demographic
characteristics including sex, and race (Table 8). Additionally, cases and controls did
not differ on socioeconomic and demographic characteristics including sex, parental
education, annual household income and area of residence (Table 10). The enrollment
of population-based community controls allows the generalization of our results to all

Michigan children and possibly to children in similar neighboring states.

Controlling for confounding variables:

Our detailed questionnaire allowed us to collect information on a number of food- and
environment-related variables to study their association with Salmonella infections.
While building our statistical model, we were able to control for numerous potential
confounders, including those reported in previous studies. However, chances of

unmeasured (unknown) confounders could not be eliminated.

Participation rate

Our overall participation rate of 72% is higher than other population-based studies
conducted to answer similar research questions. In a case-control study conducted at
the FoodNet sites, Rowe SY reported a participation rate of 59% (162). Another recent
population based case-control study carried out at the FoodNet sites (2002-2004),
reported a response rate of 67% (89), while a study conducted in France by Delarocque-
Astagneau and colleague (1995) reported a 60% response rate (5 8). The refusal rate
among cases was much lower in our study compared to what has been reported in larger

studies reported by Jones TF et a1 (2006) (11% vs. 19%) (89). The reason for higher

83

participation rate in our study was likely the result of our repeated attempts to contact

parents of the study children (~ 20 phone calls), including evening and weekend calls.

Post hoc power analysis:

The post hoc power analysis for selected exposures showed that attending a daycare
center, and having contact with cats or reptiles had a power of 80% at 5% probability of
type-I error (Table 18). Therefore for the variables (risk factors) of interest, our study
had sufficient power to detect significant association between the exposure variables in
question and salmonellosis as the outcome, if such association really existed. Egg
consumption, which is reported in the contemporary literature to be associated with
Salmonella infections (88, 98), was not significantly associated with salmonellosis in
our study. For factors that are highly prevalent in a population (and therefore in
controls), detection of statistical significance requires large sample sizes. Our study
was not powered to reveal statistically significant associations between salmonellosis
and most of the prevalent risk factors such as consumption of eggs and poultry.
Additionally, the high-risk foods are also commonly consmned by the general
population. However, it is conceivable that cases may have consumed similar foods as
controls but the contaminated one. This could only have been ascertained if we have
had a microbiological testing of the food items listed by the cases and controls during

the 3 days prior to the illness onset.

84

STUDY LIMITATIONS

Selection bias in enrolling cases:

Our data were limited to laboratory-diagnosed cases, and are thus biased by factors that
affect the probability of an illness being reported (173, 174). Cultures are not obtained
in all cases of suspected foodbome diseases, including salmonellosis, for several
reasons (1). First, the majority of foodbome illnesses are self-limiting and resolve
spontaneously in about a week. Therefore, individuals with mild to moderate disease
symptoms may not seek medical care and hence do not get reported. Second,
physicians may not request stool or other specimen cultures for patients seeking care
for gastrointestinal disease symptoms (1, 70, 95, 173, 174). Laboratory testing for
salmonellosis is largely dependent upon a patient’s presenting symptoms. In this study,
we have only enrolled children _<_ 10 years, an age group likely to receive closer medical
attention when they manifest gastrointestinal disease symptoms compared to adults.
Recall bias:

As an inherent weakness in case-control design, recall bias may be present in the
measurement of some exposure variables (107). Parents or caretakers knew the disease
status of their children prior to the interview and this may have influenced their
responses (5 8, 162). Parents of case patients may recall exposures more accurately than
parents of control subjects. In addition to questions related to specific exposures prior
to illness onset, we asked questions about children’s food preferences and the presence
of common exposures (e. g., ownership of pets, family kitchen practices) in the

household. This approach allowed us to study the association between these common

85

exposures and salmonellosis, in addition to measured exposures in the three-day period

that may have caused the illness.

Misclassification of outcome variable:

We did not obtain specimens for culture fiom controls to exclude asymptomatic cases
of salmonellosis. It is possible that some of our control children were asymptomatic
Salmonella carriers and thus may have been misclassified. However, given the very
low (1%) prevalence of chronic carriers of Salmonella in healthy populations, we
expect very few to none misclassified controls (23, 175). Furthermore, there is no
reason to assume that this rrrisclassification may have been dependent on the presence
or absence of any risk factor of interest for salmonellosis. Therefore, any
misclassification that might have occurred because of inclusion of a few ‘asymptomatic
cases’ as controls must necessarily be non-differential in nature. This may have yielded
somewhat conservative (biased towards null) estimates of effect measure, but the

validity of estimates of exposure, and their relationship with disease is not threatened.

Interviewer bias:

A total of 3 interviewers conducted telephone interviews fiom both cases and controls.

Since we called parents of each of our study participants multiple times and at different
times and day of the week, for logistical purpose we assigned separate interviewers for

cases and controls. One person interviewed the cases and two persons interviewed the

majority of controls. Because of the design of our study, we were not able to blind

interviewers to the disease status of the study participants (107). This might have

86

introduced some bias in the interviewing process. However, prior to conducting the
study, all interviewers received standard training for conducting phone interviews.
They were informed about ways to prevent the introduction of bias during the interview
process when assessing exposures. Interviewers were provided written instructions on
how to administer the questionnaire to all participants following a similar
approach/protocol. They read from a common script irrespective of the disease status
of the interviewee. All pre-testing, which also served as practical training of the

interviewers, was supervised by at least one of the investigators at all times.

Selection bias in control enrollment:

For control selections, we used the on-line white pages, which only provide landline
phone numbers and do not include cell phone or Internet phones. It is possible that our
enrolled controls (households that have landline phones) might have been different
fi'om household that did not have a landline phone in regard to certain socioeconomic
attributes. However, the National Health Interview Survey (2005) estimated that about
2% of households do not have any telephone service (wireless or landline) and only
about 7.8% of adults lived in households with only a cell phone. Moreover, the
majority of cell phone only households belonged to younger and single individuals

(3 7). Since our study enrolled households with children, it is less likely that we have
missed a large proportion of households without a landline phone while enrolling
control children. Therefore, it is likely that our case households are similar to the

control households with regard to landline phone status.

87

Two methods of enrolling controls:

We used two methods to enroll control subjects. In method 1, only 28 appropriate
subjects suggested by cases caretakers based on their familiarity with them (fiiends or
relatives), were included among the control sample of this study. This approach could
yield a control group similar to the case group with regard to certain socioeconomic and
life style characteristics (unplanned matching) (108). In method 2, we enrolled controls
using an online telephone directory, which provide a community or population control
group. Comparing controls enrolled using the two methods, based on household
income and parental education attainment revealed that the two groups were
significantly different with regard to selected socioeconomic attributes (Table 20)
perhaps due to the difference in sample size of the compared groups (28 vs. 111).
However, overall comparison of cases and controls (Table 10) did not reveal significant
difference with regard to socioeconomic difference, except the racial distribution

between the two.

Subgroup analyses:

Some of the subgroup analyses were based on a small numbers of observations,
resulting in a lack of adequate statistical power. Therefore, findings based on subgroup
analyses should be interpreted with caution. This however does not jeopardize the
main findings of the study. Additionally, results of subgroup analyses (II, IH, IV) were

based on univariate analysis and did not control for potential confounders.

88

Age and area of residence as risk factors:

We used age-stratified sampling to enroll a sufficient nrunber of cases and controls in
each age stratum. Moreover, as mentioned earlier, we started enrolling our controls
using a neighborhood matched design. However, early in the course of data collection
based on apparent lack of difference between the neighborhood matched and non-
matched controls with regard to certain socioeconomic attributes (Table 19), we
dropped the neighborhood matching design. Therefore, we did not analyze age and
area of residence as potential risk factors for salmonellosis in our study. However, the

final multivariate model was adjusted for age group.

Our cases and controls differed with regard to race. Although the reason for this
difference is not very well understood, food consumption, handling, and preparation,
along with lifestyle factors, have been reported to vary among different racial and
ethnic groups (176) . Additional studies are needed to clearly delineate the risk of

Salmonella infections associated with these populations.

89

CONCLUSIONS
This is the frrst population-based case-control study designed at identifying risk factors
for Salmonella infections in children in Michigan. Our study revealed that contact with
cats and reptiles within three days before the onset of child's illness is a risk factor for
infection with Salmonella serotypes. Additionally, attending a daycare center and
contact with a person with symptoms of GI upset is also associated with significant risk
for Salmonella infections in children. In agreement with other studies aimed at
examining factors associated with Salmonella infections in children, our data suggest
that the contribution of environmental sources plays an important role in the acquisition
of Salmonella infections in children, compared to the adult population where a larger
proportion of infections are acquired through food vehicles. Several recommendations
have been made to educate parents and caretakers about the risk of Salmonella
transmission to children from infected persons and household pets, particularly reptiles.
However, our study showed that exposure to these factors continued to cause
Salmonella infections in children. Additional efforts are needed to educate parents and
caretakers about the risk of Salmonella transmission to children from cats and reptiles,

along with individuals having GI symptoms.

FUNDING SOURCES

This research was supported by the NIH, Contract No. N01-AI-3005 8 (Microbiology
Research Unit, MSU), while supplemental funds were provided by the Office of the
Research Associate Dean of the College of Veterinary Medicine at Michigan State

University. Additionally, M. Younus received a ‘Food, Nutrition, and Chronic Disease

90

Fellowship’ fiom the College of Human Medicine, ‘Graduate Research Enhancement

Award’ and ‘Dissertation Completion Fellowship’ from the Graduate School, MSU.

CONFLICT OF INTEREST

No conflict of interest to declare

91

Table 1. Comparison of the 2006 incidences of infections with major enteric pathogens

and the US National Health Objective 2010 [Annual disease summary, CDC, 2004

 

 

(modified)].
' Pathogen ' 2006 incidence* 2010 objective**
Shiga toxin-producing Escherichia coli 0.80 1.00
015 7
Campylobacter 12.80 12.3 1
Listeria 0.27 0.25
Salmonella 14.61 6.80

 

 

 

*Reported cases per 100,000 population
M2010 Healthy People Objective

92

Table 2. Selected large foodbome outbreaks where Salmonella serotypes were

identified as etiologic agents (1974-2007). [Source: compiled from various MMWR

and FoodNet reports]

 

     
 

2007
2006-2007
2006
2006
2006

2005
2005

2004

2004
2003
2002
2002
2002
2001

2001
2001
1998
1998
1998
1998
1997
1996
1996
1996
1995

1995
1995

 

 

 

. . Number or ,. . " " ' ' ' '- Location L
60 Veggie Booty Nationwide
425 Peanut butter Nationwide
183 Restaurant tomatoes Nationwide
84 Deli IN
29 Frozen chicken MN
dinners
3 1 Orange juice Nationwide
300 Under-cooked SC
turkey
300 Roma tomatoes PA, OH, MD, VA,
WV
29 Raw almonds Canada and US
99 Hospital cafeteria MO
47 Unpasteurized milk OH
141 Roma tomatoes FL
27 Cantaloupe Western US
1000 Bakery products MI
20 Cantaloupe CA
225 Deli sandwiches VA
209 Toasted oats cereal Nationwide
58 Chile relleno AZ
50 Mexican cake MD
71 Ziti NV
79 Cheese / raw milk CA
44 Chile relleno GA
52 Roast beef SD
66 Chicken MA
62 Orange juice FL
(unpasteurized)
241 Alfalfa sprouts 6 States & Finland
133 Alfalfa sprouts OR, BC

 

93

 

 

(Table 2

 

 

continued)
1994 158 Raw ground beef WI
1994 224,000 Ice cream 41 states
1993 19 Egg rolls TX
1993 23 Hollandaise and CA
béarnaise sauce
1993 22 Mayonnaise CA
1990 690 Bread pudding IL
1989 164 Mozzarella and MN, WI. NY
shredded cheese
from a single plant
1985 16,000 Milk IL, MI, IN, IA
1974 3,400 Potato salad Not known

 

94

 

Table 3. The 20 most frequently reported Salmonella serotypes
from human sources reported to CDC in 2004 [Annual disease summary,

CDC, 2004 (modified)].

 

 
   

 

 

 

Twenty most common" ' '%_': ,
; Sdlnmnella sero ' e8 ' '

1 Typhimurium 19.2
2 Enteritidis l4. 1
3 Newport 9.3
4 Javiana 5.0
5 Heidelberg 4.9
6 Montevideo 2.4
7 I4,[5],12:i:- 2.1
8 Muenchen 2.1
9 Saintpaul 1.9
10 Braenderup 1.9
11 Infantis 1.6
12 Mississippi 1 .6
l3 Oranienburg 1.4
14 Thompson 1.4
15 Berta 1.1
16 Agona ll
17 Paratyphi B var. L(+) tartrate+ 354 1.0
18 Typhi 306 0.9
19 Hadar 277 0.8
20 Anatrrrn 250 0.7

Total 26568 74.5

 

 

95

Table 4. Examples of Salmonella serotypes by host adaptation. [Source: Uzzau S, 2000]

 

 

 

Serotype Natural host Other host(sL
S. Typhi Human -

S. Paratyphi A Human -

S. Paratyphi C Human -

S. Sendai Human -

S. Abortusovis Ovine -

S. Gallinarum Poultry -

S. Typhisuis Swine -

S. Abortusequi Equine -

S. Choleraesuis Swine Human

S. Dublin Bovine Human, Ovine

 

96

 

Table 5. Percent change (2004 vs. 1996-1998) in the incidence of reported cases of
four of the most common Salmonella serotypes under surveillance at F oodNet sites.

[Annual disease summary, CDC FoodNet, 2005 (modified)]

 

 

 

, . Pathogn ’ , Percent change in incidence
Salmonella Typhimurium - 41
Salmonella Enteritidis 0
Salmonella Heidelberg 3
Salmonella Newport 40

 

 

97

Table 6. Distribution of multi-drug resistant (MDR) Salmonella Typhimurium and

definitive phage type 104 strains in selected countries, 1992—2001. [WHO, 2003]

 

  
 

  

 

 

 

 

 

 

linfly :15 :1992—1993 1994—1995 1996—1997 1998—1999 2000—2001
Ireland 40.3, 38.9 66.8, 61.2 76.1, 70.0 70.7, 65.4 63.3, 45.2
Scotland NA, NA NA, NA NA, NA 75.0, 63.1 79.7, 56.3
Denmark NA, NA 1.5, 1.5 5.0, 3.0 21.9, 15.8 23.1, 12.7
Austria NA, 17.0 NA, 14.6 13.7, 32.7 13.1, 28.9 35.8, 29.6
Germany 14.3, 3.1 30.2, 9.2 44.3, 32.1 49.0, 32.1 57.1, 44.0
Netherlands 10.3, 6.7 8.9, 15.3 26.3, 23.5 29.5, 29.6 33.9, 37.2
Canada NA, 17.7 NA, 27.3 16.2, 46.1 44.1, 43.8 47.8, 35.5
USA NA, NA 19.8, NA 45.7, 29.1 40.5, 34.0 39.0, NA
Japan NA, 2.1 NA, 3.1 NA, 8.8 NA, 13.7 NA, 9.8
Australia 3.8, 0.1 2.7, NA 1.5, NA 1.8, 0.2 3.2, 0.7
New NA, 0.8 NA, 0.5 NA, 0.3 NA, 0.4 NA, 0.1
.Zealand

 

NA: data not available.

98

 

Table 7. Identified risk factors for Salmonella infections in children.

 

 

 

 

Risk factors Location(sj I InveLtigator/Yeari' (Reference)
Food-related exposures
Raw or undercooked egg France Delarocque-Astagneau E/1998
Netherlands (96)
Doorduyn Y/2005 (177)
Consumption of ground France Delarocque-Astagneau E/2000
beef (5 8)

Infant formula

Environmental exposures
Person-to-person
transmission

FoodNet sites, US
Gangwon, Korea
Island of Guam

France

France

FoodNet sites, US
Wisconsin, US

JonesTF/2006 (89{Rowe, 2004
#3)

Park J/2002 (178)

Haddock RL/1991 (163)

Delarocque-Astagneau E/1998
(96)

Delarocque-Astagneau E/2000
(58)

Rowe SY/2004 (162)

Wilson R/198l (67)

 

99

Daycare attendance F oodNet sites, US J onesTF /2006 (89)

Contact with reptiles F oodNet sites, US J onesTF/2006 (89)

Contaminated home Island of Guam Haddock RL/1994 (104)

environment

Riding a shopping cart F oodNet sites, US J onesTF/2006 (89)

during grocery shopping

Playing in a sandbox Netherlands Doorduyn Y/2005 (177)

Travel outside US FoodNet sites, US JonesTF/2006 (89)
*Year of reporting

 

Table 8. Comparison of demographic characteristics between enrolled case children

(participants) and non-enrolled case children (non-participants) in Salmonella case-

 

 

 

 

 

 

 

control study, Michigan, 2007
' ' " j children" f: ”j
’ n=761 V
No. No. %
Age (year) 0.42
<1 22 (17.89) 17 (22.37)
1-5 66 (53.66) 37 (48.68)
6-10 35 (28.46) 22 (28.93)
Sex 0.55
Male 61 (49.60) 41 (53.95)
Female 54 (43.90) 28 (36.84)
Unknown* 8 (6.50) 7 (9.21)
Race 0.78
Caucasians 55 (7.32) 29 (38.16)
African- 9 (2.44) 6 (7.89)
Americans 3 (44.71) 2 (2.63)
Asian 4 (3.25) 6 (7.89)
Other 52 42.28) 33 (43.42)
UnknownM

 

 

 

 

 

 

 

* Computed using the Chi-square test for two proportions

“Information was missing in the MDSS.

Public health officials at Local Health Departments collect data on demographic

characteristics (e.g., age, sex, and race) fiom reported cases during the disease
investigation process and report to MDCH. The information for this table was

obtained from the Michigan Disease Surveillance System (MDSS). Cases infected
with S. Typhi and S. Paratyphi were excluded.

100

 

Table 9. Distribution of Salmonella serotypes in children aged S 10 years and Z llyears

reported to MDCH during the study period (Dec 15, 2006 - October 15, 2007).

 

     

 

 

 

 

S 10 years a Z 11 years
n=199 n=602

No. % No. % P-valued
Typhimurium 40 20.10” 88 13.50 0.06
Enteritidis 17 8.54 164 25.21 c <0.1
Newport 7 3.52 34 5.20 0.23
Heidelberg 7 3.52 34 5.20 0.23
Oranienburg 5 2.51 8 1.20 -
Braenderup 4 2.0 1 7 l .00 -
Stanley 3 1.51 7 1.00 -
Saintpaul 2 1.01 8 1 .20 -
Pomona 2 1.01 2 0.30 -
Infantis 2 1.01 7 1.00 -
Hartford 2 1.01 6 0.90 -
Cotharn 2 1.01 0 - -
Thompson 5 2.51 15 2.30 -
Soerenga 1 0.50 0 - -
Rough O's:[e,h:l,5] 1 0.50 4 0.60 -
Norwich 1 0.50 2 0.30 -
Muenchen 3 1.51 9 1.38 -
Montevideo 5 2.51 3 0.46 -
Meleagridis 1 0.50 2 0.30 -
Mbandaka 3 1.51 3 0.46 -
Kentucky 1 0.50 1 0.15 -
Hadar 3 1.51 7 1.00 -
Bovismorbificans 1 0.50 1 0.15 -
Bareilly 1 0.50 0 - -
Adelaide 1 0.50 0 - -
Sp., 4,5,12:b:- 4 2.01 8 1.20 -
Sp., 4,5,12:i:- 15 7.54 36 5.50 -
Abony 0 0.00 1 0.15 -
Agona 2 1.01 6 0.90 -
Anatrrm 6 3.02 13 2.00 -
Anecho 0 1 0.15 -

 

 

 

 

101

 

 

 

 

 

(Table 9 continued)

Baildon 0 - 1 0.15 -
Berta 2 1.01 4 0.60 -
Chester 4 2.01 11 0.15 -
Corvallis 0 — 1 0.15 -
Derby 1 0.50 4 0.60 -
Dublin 1 0.50 2 0.30 -
Gnesta 0 - 1 0.15 -
Group B,4,12:i:- 2 1.01 5 0.77 -
Group B,4,5,12:-:l,2 0 - 2 0.30 -
Group B,4,5,12znonmotile 0 - l 0.15 -
Group C1 3 1.51 7 1.00 -
Havana 0 - 1 0.15 -
Javiana 4 2.01 9 1.40 -
Kiarnbu 0 - 1 0.15 —
Kottbus 2 1.01 1 0.15 -
Litchfield 2 1.01 5 0.77 -
Manhattan 0 - 1 0.15 -
Miami 0 - 1 0.15 -
Muenster 0 — 2 0.30 -
Ohio 0 - 1 0.15 -
Oslo 0 - 1 0.15 -
Reading 0 - 2 0.30 -
Sanjuan 0 - 1 0.15 -
Schwarzengrund 0 - 2 0.30 -
Subgroup 1 0 - 2 0.30 -
Subgroup IIIA 0 - 1 0.15 -
Subgroup 111B 0 - 2 0.30 -
Subgroup IV 0 - 1 0.15 -
Telelkebir 0 - 1 0. 1 5 -
Tennessee 5 2.51 16 2.50 -
Virchow 0 - 2 0.30 -
Weltevreden 2 1 .01 5 0.77 -
Not named 3 1 .51 0 - -
Serotype was not listed” 21 10.55 74 11.35 -

 

IInformation was missing in the Michigan Disease Surveillance System
a All cases of salmonellosis in children aged 5 10 years (participants and non-
lparticipants)
S. Typhimurium was the most common Salmonella serotypes reported in children
aged _<_ 10 years
° S. Enteritidis was the most common Salmonella serotype reported
in agedz 11 years.
d P-value calculated for selected
Cases infected with S. Typhi and S. Paratyphi were excluded

102

Table 10. Socioeconomic characteristics of children aged 5 10 years enrolled in a

population-based case-control study to identify risk factors for Salmonella infections,

 

 

 

Michigan, 2007.
Socioeconomic characteristics Cases Controls p-value"
n=123 =l39
No. (%) No. (%)

Age (Year) 0.54

< 1 22 (17.89) 25 (17.99)
1-5 66 (53.66) 66 (47.48)
6-10 35 (28.46) 48 (34.53)

Sex 0.86
Female 66 (53.66) 76 (54.68)

Male 57 (46.34) 63 (45.32)

Race < 0.01 *
Caucasians 107 (87.70) 92 (67.15)
African-Americans 9 (7.38) 27 (19.71)

Other minorities" 6 (4.92) 18 (13.14)

Parental education 0.94
Elementary to High school 30 (24.39) 33 (23.91)

Some college to college degree 72 (58.54) 85 (61.59)
Post-graduate degree 19 (15.45) 18 (13.04)
Refused to answer"" 2 (1.63) 2 (1.45)

Annual income household 0.34
<$ 35,000 26 (21.31) 25 (18.12)
$35,001- $50,000 17 (13.93) 17 (12.32)
$50,001- $75,000 28 (22.95) 39 (28.26)
>$75,000 38 (31.15) 33 (23.91)

Refused to answer 13 (10.66) 24 (17.39)

Area of residence 0.31
High income: $>60000 31 (25.20) 35 (25.18)

Medium income: $38000 - $60000 56 (45.53) 74 (53.24)
Low income: $<38000 5 36 (29.27) 30 (21.58)

 

 

 

*Significant at P < 0.05; computed using Chi-square test for two proportion "Asian,
Middle Eastern, Alaskan Indian and other racial minority groups

"*Participants refused to provide the answer/response

6 Categorized based on zip code level median household income obtained from the US
Bureau of Census, 2000.

103

Table 11. Univariate analyses of putative risk factors for Salmonella infections in

children aged 510 years, assessed in a population-based case- control study, Michigan.

 

 

 

ngousehold related variables

 

Number of people in household

 

      

 

_ 82 (66.67) 77 (55.40) Reference
> 4 41 (33.33) 62 (44.60) 1.61(0.97-2.66)
Number of children aged 510
years
1 40 (35.52) 58 (41.73) Reference
2-3 73 (59.35) 72 (51.80) l.47(0.87-2.46)
2 3 10 (8.13) 9 (6.47) l.61(0.60-4.32)
Number of bedrooms
> 3 36 (29.27) 67 (48.20) Reference
2-3 70 (56.91) 46 (33.09) 2.83 (1.63-4.90)
S 2 17 (13.82) 26 (18.71) 1.21 (0.58-2.53)
Family room flooring
Wood 17 (13.82) 12 (8.63) Reference
Carpet 88 (71.54) 117 (84.17) 0.53 (0.24-1.16)
Other and combination 18 (14.63) 10 (7.19) 1.27 (0.43-3.70)
Attend a daycare
No 106 (86.18) 130 (93.53) Reference
Yes 17 (13.82) 9 (6.47) 2.31 (1.01-5.40)
Child attends school other than
daycare
69 (56.10) 72 (51.81) Reference
54 (43.92) 67 (48.20) 0.843051-136)
Ate eggs/egg containing
product
No 44 (40.74) 60 (44.44) Reference
Yes 64 (59.26) 75 (55 .56) 1.16(0.69-1.94)
Ate poultry
No 13 (11.30) 17 (12.50) Reference
Yes 78 (67.83) 95 (69.85) 1.07(0.49-2.34)
Ate poultry at:
Home 15 (41.67) 78 (56.52) Reference
Outside home at a restaurant 3 (8.33) 3 (2.17) 2.08 (0.57-7.51)
Both home and outside home 4 (11.11) 10 (7.25) 5.20 (095-2826)
Ate meat
No 36 (30.51) 51 (37.78) Reference
Yes 58 (49.15) 60 (44.44) 1.36 (0.78-2.39)
Ate meat at:
Home (25.00) 46 (33.33) Reference

(5.56) 6 (4.35) 1.70 (0.29-9.30)
(11.11) 4 (2.90) 5.11(1.07-24.30)

Outside home at a restaurant
Both home and outside home

QNO

 

 

 

104

    

 

 

 

 

 

 

 

 

 

 

 

 

 

n=13 (%)
9
Drinking water source
Bottled 25 (20.33) 38 27.34 Reference
Municipal tap 73 (59.35) 81 (58.27) l.04(0.63-1.71)
Private well water 25 (20.33) 20 (14.39) 1.5 1(0.79-2.89)
'-Faiiiil¥.Kitche.n Practices
Keeps eggs refrigerated
Always l 18 (95.93) 136 (97.84) Reference
Never or sometimes 5 (4.07) 3 (2.16) 1.92 (0.44-8.20)
Clean kitchen counters with
Soap and disinfectant 45 (36.89) 70 (50.36) Reference
Soap and water only 26 (21.31) 25 (17.99) 1.61(0.83-3.14)
Disinfectant only 51 (41.80) 44 (31.65) 1.80 (.84-3.12)
How often clean kitchen counter
Daily 109 (90.08) 130 (93.53) Reference
More than once a week/once a
Week/less than once a week 12 (9.92) 9 (6.47) 1.59 (0.64-3.91)
Other environmental exposure
Handled packages of raw
meat/eggs while shopping
with child 64 (52.03) 87 (62.59) Reference
Did not go to shopping with child 15 (12.20) 16 (11.51) l.27(0.58-2.76)
Handled packages with
plastic/gloves 44 (35.77) 36 (25.90) 1.66(0.96-2.86)
Handled packages without
plastic/gloves
Contact with a person having
GI upset 96 (78.05) 123 (88.49) Reference
No 27 (21.95) 16 (11.51) 2.16(l.10—4.24)
Yes
Contact with animal (vs. no
contact)
Any animal contact 81 (65.85) 58 (41.73) 2.69 (1.62-4.45)
Dogs 53 (43.09) 47 (33.81) 1.48 (0.89-2.44)
Cats 35 (28.46) 21 (15.11) 2.23 (1.21-4.10)
Reptiles 17 (13.82) 5 (3 .60) 4.29 (1.53-12.02)
Birds 4 (3.25) 1 (0.72) 4.63 (0.51-42-07)
Hamster 1 (0.81) 2 (1.44) 0.56 (005-626)
90 (73.17) 106 (76.26) Reference
31 (25.20) 33 (23.74) 1.10 (0.62-1.94)

 

*Odds ratio, *"‘ Confidence interval. All exposure data were gathered for during the 3
days of child’s illness onset for cases and 3 days before the interview for controls

105

 

Table 12. Multivariate analysis of putative risk factors for Salmonella infections in

children aged 5 10 years, assessed in a population-based case-control study, Michigan,

 

 

 

 

2007.
. ' _3Adjusted OR'; ' .,

Ate eggs / egg-containing product

No Reference

Yes 1.52 (0.73-3.15) 22.41 (-17.43-83.61)
Ate poultry

No Reference

Yes 1.57 (0.60-4.09) 32.36 (-50.55-72.17)
Ate meat

No Reference

Yes 1.14 (0.57-2.30) 7.03 (-30.24-41.24)
Attended a daycare

No Reference

Yes 4.86 (1.44-16.37) 19.98 (2.72-49.84)
Attended a school

No Reference

Yes 0.89 (0.35-2.26) -5.60 (-45.34-37.77)
Contact with cats

No Reference

Yes 2.62 (1 .17-5.87) 19.65 (243-4236)
Contact with reptiles

No Reference

Yes 8.16 (1.55-42.88) 20.45 (193-6006)
Contact with a person having
symptoms of gastrointestinal
infection Reference

No 2.27 (1.02-5.44) 12.75 (0.22-33.77)

Yes

 

 

' Odds ratio, b Confidence interval, L Population attributable risk estimate ranges based
on the adjusted OR and the 95% CI. All exposure data were assessed for the periods:
3 days prior to child’s illness onset for cases and 3 days before the interview for

controls. Model adjusted for age category and race.

106

 

Subgroup analysis-1

Table 13. Multivariate analysis of selected risk factors for Salmonella infections by age

groups in children population aged $10 years, assessed in a population-based case-

control study, Michigan, 2007.

 

Exposures

Age groups

 

aContact with a person having symptoms of GI upset

No
Yes

Contact with reptiles
No
Yes

Contact with cats
No
Yes

 

< 1 year 1-10 years
AOR" (95% Cl“) AOR (95% CI)

Reference Reference
2.20 (0.22-21.60) 2.06 (0.92-4.59)
Reference Reference

4.33 (0.23-80.26) 3.57 (1.04-12.25)

Reference Reference
0.55 (0.05-6.01) 2.08 (1.01-4.28)

 

*Adjusted odds ratio, "Confidence interval

Model adjusted for race; All exposures data were gathered for during the 3 days prior
to child’s ilhress onset for cases and 3 days before the interview for controls

107

 

Subgroup analyses-2

Table 14. Assessment of selected putative risk factors for Salmonella infections in

children aged < 1 year, assessed in a population-based case- control study, Michigan,
2007.

    

No. ' M. No.7 i 3%.

 

Attend a daycare

No 17 (77.27) 16 (84.00) Reference

Yes 6 (27.27) 4 (16.00) 2.31 (1.07-5.40)
Hours spent in daycare per week

5 5 (4.07) 3 (2. 16) Reference

> 15 12 (9.76) 6 (4.32) 1.20 (0.21-6.80)
Number of children attending
daycare 6 (4.88) 3 (2. l 6) Reference

5 15 9 (7.32) 5 (3.60) 0.90 (0.15-5.25)

> 15

Number of children in the same
room as the enrolled child

5 6 1 (0.81) 4 (2.88) Reference

> 6 16 (13.01) 4 (2.88) 15.99 (1.38-185.39)
Number of children in diapers I
11 same room

S 1 9 (7.32) 1 (0.72) Reference

> 1 7 (5.69) 2 (5.04) 0.11 (0.01-1.12)
Placing child on floor/carpet
without blanket

 

 

Once a day 3 (2.44) l (0.72) Reference
More than once a day/other 10 (8.13) 14 (10.07) 0.23 (0.02-2.63)
Exclusively breast fed
Yes 7 (5.69) 7 (5.04) Reference
No 19 (15.45) 17 (12.23) 1.11 (0.32-3.84)
Formula fed
No 5 (4.07) 2 (1 .44) Reference
Yes 20 (16.26) 22 (15.83) 0.36 (0.06-2.08)
Pacifier used
No 9 (7.32) 9 (6.47) Reference
Yes 15 (12.20) 15 (10.79) 0.86 (0.32-2.25)
Child ate food containing eggs
No 17 (13.93) 21 (15.11) Reference
Yes 6 (4.92) 3 (2.16) 2.47 (0.53-11.36)

 

*Odds ratio, "Confidence interval All exposure data were gathered for during the 3
days of child’s illness onset for cases and 3 days before the interview for controls.

108

Subgroup analyses-3
Table 15. Univariate analyses of selected risk factors for Salmonella infections in
children aged 1-10 years, assessed in a population-based case- control study, Michigan,

2007.

 

   

 

 

 

 

 

" posures.(aged 2.1 year) Cases ' ' “.Contrbl's ' H ' ((95%, CI“) '1 ‘
“ 2.2511,- . ~ n=101/ 1 _' n=114, - ' . f _ -
No. (%) No. (° 0)
Consumed unpasteurized
milk or cheese
No 95 (77.24) 1 10 (79.14) Reference
Yes 4 (3.25) 5 (3.60) 0.92 (0.24-3.55)
Ate food containing eggs
Ate at home 72 (58.54) 65 (46.76) Reference
Ate outside of home 6 (4.88) 25 (17.99) 0.21 (0.08-0.56)
Ate both at home and outside 3 (2.44) l (0.72) 2.70 (0.27-26.68)
Ate poultry
Ate at home 62 (50.41) 79 (56.83) Reference
Ate outside 8 (6.50) 10 (7.19) 1.01 (0.38-2.73)
Ate both at home and outside 8 (6.50) 3 (2. 16) 3.39 (0.86-13.3)
Ate meat other than poultry
Ate at home 41 (33.33) 46 (33.09) Reference
Ate outside 6 (4.88) 6 (4.32) 1.12 (0.33-3.75)
Ate both at home and outside 5 (4.07) 4 (2.88) 1.40 (0.35-5.57)
Frequency of eating at
commercial food establishments
Daily to more than once a week 6 (4.88) 9 (6.47) Reference
Once a week 45 (36.59) 51 (36.69) 1.20 (0.33-4.36)
Never to once a month 48 (39.02) 55 (39.57) 1.58 (0.66-3.79)
Preferred food at fast food
establishment
(vs. never to once a month)
Hamburger 11 (12.79) 22 (19.13) 0.65 (0.27-1.54)
Chicken 36 (41.86) 39 (33.91) 1.20 (0.62-2.31)
Other/combination 9 (10.47) 15 (13.04) 0.78 (0.30-2.02)
Child attends school '
other than daycare
No 69 (56.10) 72 (51.81) Reference
Yes 54 (43.92) 67 (48.20) 0.84 (0.51-1.36)
School food usually
prepared by:
Family member 8 (8.08) 12 (10.53) Reference
School cafe/cook 14 (14.14) 21 (18.42) 1.00 (0.32-3.06)
Other/combination/
do not eat at 521201 8 (8.08) 8 (7.02) 1.50 (0.39-5.65)
"Odds ratio, "Confidence interval

All exposure data were gathered for during the 3 days of child’s illness onset for cases
and 3 days before the interview for controls.

109

Subgroup analysis-4

Table 16. Demographic characteristics of cases of Salmonella serotype S. Typhimuriun

and controls in children aged S 10 years, assessed in a population-based case- control

study, Michigan, 2007.

 

 

, Demographic characteristics Cases Controls
_. , . _ n=36 n=139 '
N o. (%) No. (%) p-vglu;

Age (year) 0.69
< 1 6 (16.67) 25 (18.12)

1-5 20 (55.56) 66 (47.83)
6-10 10 (27.78) 47 (34.06)

Sex 0.08
Female 14 (38.89) 76 (55.07)

Male 22 (61.11) 62 (44.93)

Race 002“
Caucasians 31 (86.1 1) 91 (66.91)
Minorities "‘ 5 (13.89) 45 (33.09)

Parental education 0.15
Elementary to High school 6 (16.67) 33 (24.09)

Some college to college degree 29 (80.56) 85 (62.04)
Post-graduate degree 1 (2.78) 17 (12.41)
Refused to answer 7‘ 0 ' 2 (1-46)

Annual income household 0.13
S $ 35,000 9 (25.00) 25 (18.25)
$35,001- $50,000 7 (19.44) 17 (12.41)
$50,001- $75,000 8 (22.22) 39 (28.47)
>$75,000 l 1 (30.56) 32 (23.36)

Refused to answer x l (2.78) 24 (17.52)

Area of residence 5 0.04"
High income: $>60000 9 (25.00) 35 (25.36)

Medium income: $38000 - $60000 10 (27 .78) 73 (52.90)
Low income: $<38000 17 (47.22) 30 (21.74)

 

 

 

*Significant at p < 0.05 (p-value obtained using a chi-square test for two proportions)
"African-Americans, Asian, Middle Eastern, Alaskan Indian and other racial groups
1 Parents refused to provide the answer/response
a Categorized based on zip code level median household income obtained from the US
Bureau of Census, 2000.

110

Table 17. Univariate analyses of putative risk factors for Salmonella serotype

Typhimurium infections in children aged $10 years, assessed in a population-based

case- control study, Michigan, 2007.

 

' " coll???" " ‘V

     

No.1 ' "tr/o "NaQ ‘°/.;' "

 

Ate eggs/egg containing
product
No 4 (11.11) 21 (15.22) Reference
Yes 2 (5.56) 3 (2.17) 3.50 (0.43-28.13)
Ate poultry
Ate at home 15 (41.67) 78 (56.52) Reference
Ate outside home 3 (8.33) 3 (2.17) 2.08 (0.57-7.51)
Ate both home and outside
home 4 (11.1 1) 10 (7.25) 5.20 (0.95-2826)
Ate meat
Ate at home 9 (25.00) 46 (33.33) Reference
Ate outside home 2 (5.56) 6 (4.35) 1.70 (0.29-9.80)
Ate both home and outside 4 (1 1.11) 4 (2.90) 5.11 (1.07-24.30)
home

Frequency of eating at
commercial food establishments

Never 3 (8.33) 17 (12.32) Reference
Once a month 12 (33.33) 37 (26.81) 1.83 (0.45-7.37)
Once a week 13 (36.1 1) 51 (36.96) 1.44 (0.36-5.68)

Daily to more than once a week 2 (5.56) 9 (6.52) 1.25 (0.17-8.96)

Contact with animal

 

 

Any 25 (69.44) 57 (41.30) 3.22 (1 .47-7.08)
Dogs 17 (47.22) 46 (33.33) 1.78 (0.85-3.76)
Cat 9 (25.00) 21 (15.22) 1.85 (0.76-4.50)
Reptiles 5 (13.89) 4 (2.90) 5.40 (1.37-21.29)
Birds 3 (8.33) 1 (0.72) 12.45 (1.25-123.56)
Contact with a person with G1
symptoms
No 30 (83.33) 122 (88.41) Reference
Yes 6 (16.67) 16 (11.59) 1.52 (0.55-4.22)
Travel outside the states
No 26 (72.22) 106 (76.8 1) Reference
Yes 10 (27.78) 32 (23.19) 1.27 (0.55-2.92)

 

*Odds ratio, "Confidence interval. All exposures data were gathered for during the
3 days of child’s illness onset for cases and 3 days before the interview for controls

lll

Table 18. Post hoc power analysis of selected potential risk factors for Salmonella

infections in Michigan children assessed in a population-based

case-control study, 2007

 

 

Ate eggs/egg- ' 55.55 H I 103 I 1.52 29%
containing product

Ate poultryM 83.92 91 1.57 13%
Ate meat“ 54.05 94 1.14 5%

Attended a daycare 6.47 123 4.86 97%
Attended a school 48.20 123 0.89 5%

Contact with cats 15.10 123 2.62 86%
Contact with 3.59 123 8.16 99%
reptiles

Contact with a 11.51 123 2.27 62%
person having

symptoms of

gastrointestinal

infection

 

 

' Adjusted odds ratio; Not assessed in children aged < 1 year

112

Table 19. Comparison of controls: neighborhood matched vs.

non-neighborhood matched

 

 

 

Neighborhood Unmatched P-
Controls Controls value“
(n=36) (n=36)
Annual household 0.82
income 4 (11.11) 3 (8.33)
Some high school 5 (13.89) 7 (19.44)
High school or GED 12 (33.33) 8 (22.22)
Some college 9 (25.00) 11 (30.56)
Four year college 4 (11.11) 2 (5.56)
degree 1 (2.78) 3 (8.33)
Graduate degree 1 (2.78) 2 (5.56)
Post Graduate degree
Refused to answer
Parental education 0.63
5 $20,000 3 (8.33) 6 (16.67)
$20,000-$35,000 2 (5.56) 3 (8.33)
$35,001-$50,000 7 (19.44) 1 (1.78)
$50,001-$75,000 7 (19.44) 10 (27.78)
$75,001-$100,000 4(11.11) 4(11.11)
>$100,000 5 (13.89) 5 (13.89)
Refused to answer 8 (22.22) 7 (19.44)

 

 

 

 

 

 

*Computed using Chi-square test for two proportions

and from the landline telephone directory (method-2)

e parents (method-1)

 

 

 

 

Variables Controls P-
value"
Method-1 Method-2
(n=28) (n=lll)

Race“ <0.01
Caucasians 10 (35.71) 83 (74.77)
African-Americans 16 (57.14) 11 (9.91)

Other minorities 2 (7.14) 2 (15.32)

 

 

 

 

 

113

 

 

 

 

 

 

Parental education 0.04
High school 4 (14.29) 29 (26.13)
Some college 15 (53.57) 29 (26.13)
Four year college 6 (21.43) 35 (31.53)
degree 3 (10.71) 18 (16.22)
Post Graduate
dame
Annual household 0.02
income 7 (25.00) 18 (16.22)
5 $35,000 7 (25.00) 10 (9.01)
$35,001-$50,000 11 (39.29) 28 (25.23)
$50,001-$75,000 2 (7.14) 31 (27.93)
>$75,001

 

*Computed using chi-square test for two proportions, “Asian, Middle Eastern,

and Alaskan Indian

114

 

Table 21 . Sample size calculations for Michigan Salmonella case-control study

 

 

 

 

 

 

 

 

Power Exposure in Odds Ratio Number of cases
control group required
80% 15% 2.5 124
80% 18% 2.5 1 l 1
80% 20% 2.5 105

 

115

 

Figure 1. Incidence of non-typhoidal Salmonella infections per 100,000
population, England and Wales, 1981-2004.

 

W
O

 

1
\
M1

 

M
O

 

 

 

&
O
1

 

 

 

 

30

 

 

 

N
O

Salmonellosis incidence per 100,000 population

 

 

—a
O
l

 

 

 

O I I r I I I I T 7 I I I I I I I I I I T I I I
eeeeeeeeee‘fee‘eeeeeeeefafaqfe

Year

116

Figure 2. Incidence of non-typhoidal Salmonella infections per 100,000 population,

US, 1944-2002.

 

 

 

25 .b MW -_._. __ __ _

 

 

 

 

15 —- ~-

 

 

 

 

 

 

 

 

 

 

Salmonellosis Incidence per 100,000 population

 

 

0 TTIIIIIIFWIIIIIIIIITTTTIIIIIIIIIIIIIIIIIIIIIIIIIIIIIITIIIIIII

¢05°95852065tvahl0c§<8>4°®”9545‘
edeeeeeee‘eeeeeeedeea

Year

 

 

 

117

Figure 3. Salmonella Enteritidis infections incidence in the United States, 1970-

2001

 

 

S. Enteritidis incidence per 100,000 population

 

 

 

 

 

 

 

 

 

 

 

 

\°’ \q \°’ \°’ \°’\°’ \°’ \°’\q \°’ \qxqqequqqxqq

Year

I I I T I I I I I r I I I I I I I I I I F I I T I I I I I I F

o
o
'19

 

118

 

Figure 4. Age-stratified salmonellosis incidence, Michigan, 1992-2006.

(11 = 13,877)

 

I
l
l
1
l
l
l
1
l
l
I
l

50~

t»
O
1

 

N
O
1

Salmonellosis incidence per 100,000 population

 

 

i
I)
I

<1 year 1-4 years 5-9 years 10-34 years 35-64 years
Age categories

119

60- - ~ ,v *7 i-v 7-7 7_-#-a-14

265 years

 

Figure 5. Surveillance of Salmonella infections in Michigan

 

Centers for Disease Control and Prevention

A

Michigan Department of Community

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Health
Local Health Department
I i 4
Physicians Laboratories

 

 

 

 

 

 

120

Figure 6. Michigan Salmonella case-control study, 2007: Enrollment of cases

(12/15/06 - 10/15/2007)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

All reported cases between
12/15/06 and 10/15/2007
n=862
Cases in aged >11 years ‘
“=65! Cases in children
aged $10 years
n=228
Typhoidal cases T
n=29 . .
Non-typhordal cases 1n
children aged 510 years
n=199
Incomplete address/phone j
number ¢ *
n=29 Invitation letter sent
n=l70
Comorbid condition
n=1 ‘
(1/170)=0.59% V
Eligibility criteria met
n=169
Declined to participate
n=10 | |
(10/169)=5.9l% *
Written or oral consent
provided
n=159
Answering machine/busy call /could
not be contacted at all ‘
n=36 .
(3 6/ 1 59)=22.64% lntervrewed
n=l23
(123/169)=72.35%
Phone interview Mail-in-
n=102 Questionnaire
(l 02/123)=82.94% “=21

 

 

 

 

 

 

121

Figure 7. Michigan Salmonella case-control study, 2007: Enrollment of controls
(12/15/06 - 10/15/2007)

Disconnected
phone numbers
n=445
(445/2463)=18.07%

Commercial
phone numbers
n=53
(53/2463)=2.15%

Answering
machine
n=l 134
( l 134/2463 )=46.04%

Hung up/not
interested
n=338
(338/2463)=l3.72%

No children or
children <10
years
n=37l
(371/2463)=15.06%

Could not be
contacted again
n=11
(1 1/2463)=0.45%

Method-2

 

 

Method-l

 

 

 

l

 

 

Total phone numbers
obtained
n=2,463

 

 

4

l
V

 

 

Valid phone numbers
n=2,018

 

 

 

‘7

 

Y

 

 

Household phone
numbers
n=1,965

 

 

 

4

 

V

 

 

Received phone call
=83]

 

 

 

1

 

V

 

 

Screened for potential
control children
n=4

 

 

4

 

 

V

 

 

Scheduled interview
n=122

 

 

 

«

 

V

 

 

Interviewed
n=111
(l l l/2463)=4.51%

 

 

122

l

 

Potential controls
obtained from case
parents
n=37

 

 

 

Could not be 4—
contacted
again/refused
n=9
(9/37)=24.32%

 

V

 

Interviewed
n=28
(28/37)=75.68%

 

 

 

APPENDIX

Contents:

1. Case-control study invitation letter
2. Case-control study information sheet and consent form
3. Case-control study questionnaire

4. References

123

CASE-CONTROL STUDY INVITATION LETTER

MICHIGAN STATE
u N IV E R 5 IT Y

 

Date: / /

Name:

 

Address:

 

Dear

 

I am writing to you because your child is a possible candidate for a statewide Salmonella
Study, which is being conducted by researchers from Michigan State University (MSU)
and Michigan Department of Community Health (MDCH). The main objective of this
study is to identify the risk factors and conditions for Salmonella infections in Michigan

children less than ten years of age.

We need to contact parents of children who have experienced recent Salmonella
infection. Hospitals and Physicians are required by law to report diagnosed cases of
Salmonella infection to Michigan Department of Community Health (MDCH). The
MDCH records are being used to identify contact information of reported cases. In this
research, each child’s parent will be interviewed regarding food intake history, food
handling and cooking practices, and household sanitation. The phone interview will take
about 15-20 minutes. Alternatively, you may choose to fill out the same questionnaire

that can be mailed back in a provided self-addressed, stamped envelope.

124

This survey is a very common way to study the causes of Salmonella infection. It is an

important study since it will contribute to the understanding of the risk factors related to
Salmonella infection in young children. Michigan State University and MDCH Internal
Review Boards have reviewed and approved this research project according to the most

recent patient rights and privacy rules.

Enclosed with this letter, you will find an information sheet that describes the study as
well as a consent form to participate in this study. If you need more information on this

study, you may contact professor A. Mahdi Saeed, Ph.D. at 517-432-9517.

Although your participation is voluntary, it is very important for the success of this
study. We will appreciate the return of the consent form even you elect not to

participate. This confirms that you were successfully contacted. Please respond

promptly.

Sincerely,

Dr. Mahdi Saeed Dr. Melinda Wilkins

Professor of Public Health, Director, Division of Communicable Diseases,
College of Human Medicine, Bureau Of Epidemiology.

Michigan State University Michigan Department of Community Health

125

Please return via the included self-addressed, stamped envelope to:
Dr. Melinda Wilkins/Michigan Salmonella study

Room #508

Capital View Building,

Michigan Department of Community Health

201 Townsend Street, Lansing, MI 48913

126

CASE-CONTROL STUDY INFORMATION AND CONSENT FORM

Michigan State University and the Michigan Department of Community Health (MDCH)
are conducting research to identify factors that increase the risk of illness from

Salmonella in children.

Importance of the study:

Salmonella is a bacterium, which can cause illness of the digestive system. Among
adults it is commonly caused by eating contaminated food. Our study attempts to
determine the role of kitchen and household practices that may contaminate food and

objects, and cause illness in younger children.

Description of the study:

You are being contacted either because your child’s illness was reported to the health
department (Salmonella illness is legally reportable to the health department) or just at
random as part of a comparison group who were not ill. If you agree to participate you
will be asked questions about your child’s food intake and things like your household
food handling, cooking, and cleaning/sanitation practices. The interview will take about

15-20 minutes and you do not have to answer any questions you don’t want to answer.

Risk/Benefits:
There is no direct benefit to you for participation, but we hope it will help us learn more

about this illness. The only potential risk is to your confidentiality.

127

Confidentiality:

The information from your questionnaire will be put into an electronic file that does not
identify you or your child. Once that is done we will destroy study records that could
identify you or your child. The confidentiality of your information will be protected to

the maximum legal extent.

Contact details:
If you have any questions regarding the study, you may contact Dr. Mahdi Saeed at 517-

432-9517.

For information about your/your child’s rights as a research subject you may contact:

Peter Vasilenko, Ph.D.

Director of Human Research Protections
Michigan State University

202 Olds Hall Lansing,

East Lansing, MI 48823-1047

Phone: (517) 355-2180

Fax: (517) 432—4503

E-mail: irb@msu.edu

128

CASE-CONTROL CONSENT FORM

1 agree to allow a researcher to contact me to complete the study questionnaire.

 

 

 

 

 

 

I agree I don’t agree

 

 

If you agree:

Please indicate if you prefer to complete the interview:

 

 

 

Over telephone

 

 

 

 

Myself (you will be mailed the questionnaire)

 

If you chose telephone interview, please write:

Home telephone number:

 

Best Time of Day to Call:

 

Best Time of Week to Call:

 

Best Days of Week to Call:

 

Dates to Avoid or on which you are Unavailable over the next 60 days:

 

Thank you in advance for your contribution in this project. We will be happy to send

you a summary of the study findings upon your request.

129

Sincerely,

Please return via the enclosed self-addressed, stamped envelope to:

Michigan Salmonella Study

Michigan Department of Community Health, Division of Communicable Disease
Division, Bureau of Epidemiology, Capitol View Building,

201 Townsend Street, Lansing, Michigan

130

Interview starts (time)- -------- Interview ends (time)- ----------

 

 

Michigan Salmonella Case-Control Study

 

STUDY QUESTIONNAIRE

[Telephone interview Form]

 

   

Date Ofmtervlew' " .. ..

 

 

 

 

Study Introduction:

Hello, my name is and I work for Michigan
State University (MSU).
Are you the parent or guardian of ( )?

Insert child '5 name
MSU is conducting a study, in collaboration with the Michigan Department of
Community Health (MDCH), to identify factors and conditions that make some
children more likely than others to get salmonellosis, a foodbome illness. Children are
also at a higher risk of getting salmonellosis compared to adults. Therefore, we are
trying to study the causes of this higher infection rate. Salmonella infection is a
reportable disease by law in Michigan. Your contact information was obtained with the
permission of our collaborator, Michigan Department of Community Health.
We are very hopeful that you will be willing to participate in this project to enable us to
generate very much needed information on the conditions associated with the disease

in Michigan children.

131

 

Your participation is voluntary. However, we are asking for your help because the
knowledge gained through this study may contribute to the control and prevention efforts

of Salmonella infections in Michigan’s children.

The type of effort needed from you, as a participant, is to complete a short questionnaire.
You can answer the questions over the phone or by filling out the questionnaire mailed to
you. There are no known physical and/or psychological side effects associated with these
questions. The questionnaire will only take about 10 minutes for older children, and 15-

20 nrinutes for younger ones.

All information gathered from you will remain confidential. Data will be reported in a
summary form and no individually identifiable responses will be presented or published.
You may decide to withdraw from the study even after the interview, and you can decline

to answer any question that makes you uncomfortable. Do you have any questions?

Are you willing to take part in this research? Yes “Thank you in advance for
your contribution to this project. “

No “Thank you for your time”

 

 

Singed consent: Yes No

 

 

 

 

Verbal consent Yes ‘ No

 

 

 

 

 

 

(Please read the consent form over telephone)

132

How would you like to fill out the questionnaire phone or by mail?

By mail (Confirm the address):

 

By phone interview

9

Is now a good time to talk to you? Yes “Thanks, we will now begin the questionnaire’
No “When can I call you back?

 

 

 

 

 

Day and date: Time:
Eligibility Criteria: To radar... theweligibility or? i ) for this study,
could you please tell us if ( ) has any serious medical conditions (e.g

cancer: leukemia, lymphoma) or birth defects?

Yes [We apologize, we cannot enroll ( ) as a participant in this
particular study because having a serious medical condition will complicate the
understanding of Salmonella infection risk factors.

 

 

 

No [please proceed with the interview]

 

 

Below is the information that will be obtained from the MDSS" database

 

 

 

DEMOGRAPHIC INFORMATION

1) Illness Onset date / / (dates three days prior to illness onset
_l_ to _l )

For example illness onset 7/1 700% so three days prior would be 7/14 — 7/16, use the
above dates throughout questionnaire

For controls for this case, use the three-day period prior to interview date throughout
the questionnaire

 

2) Child’s name

 

3) Age of the child at time of illness onset (in completed months): '
Days/Months/Y ears (number of days if the child is less than a month)

4) Child’s gender: Male Female

 

*Michigan Disease Surveillance System maintained by the Michigan Department of
Community Health

133

5) Zip code of child’s permanent residence
6) County of child’s permanent residence
7) Child’s race

Caucasian\White
African American\Black

[I
E

Pacific Islander
Unknown

8) Child’s ethnicity

Hispanic/Latino
Non-Hispanic/Latino
Unknown

Other (specify):

9) Case is reported as part of an outbreak
(Only for cases)

_Y

as No Unknown

    

10) What is your relationship to

child’s name
Mother
Father
Other (with parent or guardian’s permission)

HOUSEHOLD INFORMATION

 

First, I will ask some questions about your household during the 3 days

prior to I ) illness, from ( to )

 

 

 

l! or cases: All questions should refer to the 3 days preceding the illness onset date]
IEor controls: All Questions should refer to the 3 days time preceding the interview
date]

11) How many people live in your home?

134

12) How many children in your home are less than 10 years of age?
13) How many children in your home are in diapers?

14) How many bedrooms are in your house?

15) What kind of flooring do you have in the family room?

Carpet

Wood

Tile/linoleum

Does not have a family room

Other (specify): indicate
rugs here

16) What kind of flooring is in ( ) bed room?
child’s name
Carpet
Wood
Tile/linoleum
Other (specify): indicate rugs

 

here
CHILD CARE

 

Now, I Will ask you about ( ) Childcare during the 3 days prior to
(his/her) illness, from ( to ). (insert same date’s as above)

 

 

 

17) Does L ) attend a day care outside of your home? Yes
No (IfNo. £0 to 0#1&

 

17a) How many hours per week does (he/she) usually spend in day care?
hours/wk

17b) How many total children attend ( ) daycare? children
Child’s name

17c) About how many children share the same room as ( )?

children Child’s name

17d) About how many children in your child’s room are in diapers? children

17c) About how many day care workers attend to this room? workers

135

17f) Is there a separate room for changing diapers in the day care?

17g) Is there a sink with soap and water next to the diaper-changing area in the
day care?
_Yes _No _Don’t know

17h) In the day care, approximately how far in feet is the diaper-changing area from
the area where food, milk, and other beverages are handled? it

171) Are you aware of any child at the daycare who experienced vomiting, diarrhea,

or abdominal cramps during the 3 days prior to ( ) illness?
Yes No Don’t know (If No or Don ’t know, go to Q #17k)

17 j) How many children had nausea, vomiting, diarrhea or abdominal cramps
during the 3 days prior to ( ) illness? children
Child’s name

17k) Who usually prepares the food (child’s name) eats while at the daycare?
(Mark all that apply)
Mother
Father
Other family member
Daycare personnel

Other (specify):

 

18) Does ( ) attend a preschool, kindergarten, or elementary school?
Yes No
(If No, go to Q #19)

 

 

Can be in addition to daycare — such as before or after school care programs.

18a) Who prepares the food that ( ) eats while at school?
(Check all that apply)

Mother

Father

Other family member

Cafeteria/cook
Other (specify):

 

 

19) Did you take ( ) with you while grocery shopping during the 3
days prior to (his/her) illness?

Yes No Don’t know (If No or don’t know, go to Q #20)

136

19a) Did you use gloves or plastic bags when handling packages of raw chicken,
meat, and egg products while grocery shopping that time?

Yes No Don’t know _ Did not handle meat or egg products

 

This part of the questionnaire asks you about (child’s name) food
history and activities (skip if over age 1 year)

 

 

 

20) Did you put ( ) on the floor or carpet without a blanket during
the 3 days prior to ( ) illness?

 

Yes No (If No, go to Q #21)

20a) About how often was ( ) placed on (or played on) the floor or
carpet without a blanket in the 3 days prior to his/her illness?

Never

Once a day
More than once a day

Other (specify):

 

21) Was L ) breast-fed during the 3 days prior to (his/her) illness?
Yes No Don’t know

 

22.) Did you use formula to feed ( ) during the 3 days prior to (his/her)
illness?
Yes No _Don’t know (If No or Don ’t know, go to Q# 23)

22a) What type (e.g., milk, soy, rice-based) and brand of formula did you feed
(child ’5 name) during the 3 days prior to (his/her) illness?

 

Please record exact brand and type if known.
If not known, use list below to prompt recall.
(Check all that apply)

137

Isomil

 

 

 

 

Enfamil
Bright Beginnings
Nestle
Similac
Store brand (e.g. Meijers, Krogers etc)
Other (specify):
23) Did ( ) use a pacifier during the 3 days prior to (his/her) illness?
Yes No _Don’t know
24) Did ( ) eat egg during the 3 days prior to (his/her) illness?
Yes (If yes, how it was prepared? fully cooked partially cooked)
No Don’t know
25) Did ( ) eat any food that contained eggs during the 3 days prior
to (his/her) illness?
Yes No _Don’t know

 

U1 “Food’History’
‘_ . (Skip rfless than 1 year of age and go to

 

 

 

26) Did ( ) eat or drink any unpasteurized milk, or cheeses such as
queso fresco made with unpasteurized milk during the three days before your
illness?

 

Yes Probably yes Probably not No Don’t know

26a) Did ( ) eat egg during the 3 days prior to (his/her) illness?
child ’5 name
Yes (If yes, how it was prepared? fully cooked half cooked .7)

_No _ Don’t know

138

26b) Did ( ) eat any food that contained eggs (such as: cookie dough,
salad dressings, mayonnaise, ice cream, custard, cake mix) during the 3 days prior
to (his/her) illness?

Yes if yes, prepared at home: Yes No

No Don’t know

26c) Did ( J eat any food that contained poultry (such as chicken, or
turkey) during the 3 days prior to (his/her) illness?

 

Yes if yes, prepared at home: Yes No

No Don’t know

26d) Did ( ) eat any food that contained meat other than poultry (such
as hamburger) during the 3 days prior to (his/her) illness?

 

Yes if yes, prepared at home: Yes No

No Don’t know

26e) In the three days before ( ) illness, did he/she eat at any of the
following types of commercial food establishment? (mark all that apply)

 

Restaurant

 

 

I If don’t remember then ask Q26 f and g

 

Fast-food establishment
Cafeteria

Deli

_Read-to-eat food served in a supermarket or department store
_Street-vended food

_Concession stand at sporting event

_Snack bar

Gas station

139

_Other (specify)

261) How often does ( ) eat at fast food restaurants?
child ’5 name '

_Daily
More than once a week
Once a week
Once a month

Never

0ther(Specifi/):

 

26g) What is ( ) preferred food at fast food places?
child ’5 name

Hamburger
Chicken

Other (specify

 

Question about source(s) of drinking water

 

 

 

27) Now I am going to ask about the types of water sources (child’s name) drank
during the 3 days prior to (his/her) illness? Did (child’s name) drink water from:

Municipal tap water Yes No Don’t know

Private well water Yes No Don’t know

Untreated surface water Yes No Don’t know
(river, pond, lake)

Bottled water Yes No Don’t know

Other:

 

140

INTRAFAMILIAL TRANSMISSION OF SALMONELLA

’ Thrspartof the questionnaire asks you about your family’s possible
, g ‘ exposure to Salmonella during the 3 days prior to illness .

 

28) Was anyone in your household ill with symptoms of stomach upset, which may
include nausea, vomiting, diarrhea, and abdominal cramps during the 3 days prior

to ( ) illness?
Yes No Don’t know (If No or Don ’t know, go to Q# 29)

28a) Did (he or she) seek medical care for these symptoms?
Yes No Don’t know (If No or Don ’t know, go to Q29)

28b) What was the diagnosis? diagnosis or Don’t know
29) During the 3 days prior to ( ) illness, did (he/she) visit any friends
or relatives who had symptoms of stomach upset, which may include nausea,

vomiting, diarrhea, and abdominal cramps?
Yes No Don’t know

30) During the 3 days prior to ( ) illness, did anyone who had
symptoms of stomach-upset visit your home?

Yes No Don’t know

FAMILY KITCHEN PRA TICES

 

This part of the questionnaire asks you about your family’s kitchen
practices

 

 

 

31) Do you keep your eggs in a refrigerator?
Never Sometimes Always

32) Do you wash your kitchen counters, sinks, and cutting boards after preparing
raw chicken?
Never Sometimes Always (If never go to Q#35)

33) How do you clean your kitchen counters?

with soap and water with a disinfectant

141

34) How often do you clean your kitchen counters?

Less than once a week

Once a week

More than once a week

Daily

ANIMAL EXPOSURE

 

This section of the questionnaire asks you about pets

 

 

h“. run-‘-—r —-n-v flrufifl't -u.» \-_ -‘Lu- 4.- - on» r.- “I” -—m .1 “a -

 

35) During the 3 days prior to ( ) illness, did (he/she) have contact with
any type of pet, your pet or someone else’s pet or animals in a petting zoo setting?
Yes No Don’t know (If No or Don ’t know, go to Q# 36)

35a) What kind of pet(s) did ( ) have contact with during the 3 days
prior to (his/her) illness?
(Check all that apply, get as much detail as possible)

Dogs (if yes, how many?) _# Dog(s) age(s) _weeks, months, adult
Cat (if yes, how many?) _# Cat(s) age(s) _weeks, months, adult
Reptiles (if yes, how many) _# describe (iguana, cornsnake etc)
Birds (if yes, how many) _# describe

(chicken, duckling, parakeet etc.)
Hamster
Gerbil
Ferret

Other (specify):

 

 

35b) Were any of these animals noticeably ill with diarrhea?
Yes No Don’t know

TRAVEL HISTORY

 

This section of the questionnaire asks you about your child’s travel
history

 

 

 

36) Did ( ) travel anywhere during the 3 days prior to (his/her)
illness?

142

Yes No Don’t Know (If No or Don ’t know, go to Q #3 7)

 

36a) Did ( J meet any person with symptoms of stomach upset
during your visit?
Yes No Don’t know
SOCIOECONOMIC HISTORY

 

Just a couple more questions about your income and education, you don’t
need to answer if you are uncomfortable

 

 

 

37) What is the highest level of education you have completed?
Some High School
High School or GED
Some college or technical training
4 year college degree
Graduate degree
Post graduate degree

37a) What is your total annual household income?

less than $20,000
$20,000 - $35,000
$35,001 - $50,000
$50,001 - $75,000
$75,001 - $100,000
more than $100,000
Refirsed to answer

“That’s it! Thank you so much for your time, we really appreciate that you have shared
this important information with us as we try to research this important childhood disease
If you have any questions related to the study you may contact Dr. Mahdi Saeed, the
principal investigator of this research, at 517-432-9517.”

 

Investigators:

Dr. Mahdi Saeed

Professor, Department of Epidemiology, College of Human Medicine
Michigan State University

Tel: 517-432-9517

E-mail: saeeda@msu.edu

Dr. Melinda Wilkins

Director, Communicable Disease Division
Michigan Department of Community Health
Tel: 517-335-8165

143

E-mail: wilkinsm@michigan.gov

Michigan State University Community Research Institutional Review Board (MSU
CRIRB)

202 Olds Hall, East Lansing, MI 48824

Phone: (517) 355-2180; Fax: (517) 432-4503; E-mail: ucrihs@msu.edu

144

REFERENCES

1. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee J S, Shapiro C, et a1. Food-
related illness and death in the United States. Emerg Infect Dis. 1999 Sep-Oct; 5(5):607-
25.

2. Gomez TM, Motarjemi Y, Miyagawa S, Kaferstein FK, Stohr K. Foodbome
salmonellosis. World Health Stat Q. 1997;50(1-2):81-9.

3. Voetsch AC, Van Gilder TJ, Angulo FJ, Farley MM, Shallow S, Marcus R, et al.
FoodNet estimate of the burden of illness caused by nontyphoidal Salmonella infections
in the United States. Clin Infect Dis. 2004 Apr 15;38 Suppl 3:S127-34.

4. CDC. Centers for Disease Control and Prevention. Summary of Notifiable
Diseases, United States, 2003. Morb Mortal Wkly Rep. 2005;52(54):1-85.

5. Bell C KA. Salmonella. In: Blackburn, C., McClure, P.J. (Eds), Foodbome
pathogenes. Cambridge. CRC Press, Woodhead Publishing, pp 307—33 1. 2002.

6. Hardnett FP, Hoekstra RM, Kennedy M, Charles L, Angulo FJ. Epidemiologic
issues in study design and data analysis related to F oodN et activities. Clin Infect Dis.
2004 Apr 15;38 Suppl 3:Sl21-6.

7. Scallan E. Activities, achievements, and lessons learned during the first 10 years
of the Foodbome Diseases Active Surveillance Network: 1996-2005. Clin Infect Dis.
2007 Mar 1;44(5):718-25.

8. CDC. Centers for Disease Control and Prevention. Preliminary FoodNet data on
the incidence of infection with pathogens transmitted commonly through food-10 states,
2006. Morb Mortal Wkly Rep. 2007;56(14):336-9.

9. Pegues DA OM, Miller SI. Salmonella species, including Salmonella Typhi.
Principles and practice of infectious diseases 6th edition Elsevier, Churchill Livingstone;
2005.

10. Gray JT F-CP. Salmonella. In: Cliver DO, Riemann HR , ed. F oodbome diseases
Boston: Academic press; 2002.

11. Velge P, Cloeckaert A, Barrow P. Emergence of Salmonella epidemics: the
problems related to Salmonella enterica serotype Enteritidis and multiple antibiotic
resistance in other major serotypes. Vet Res. 2005 May-Jun;36(3):267-88.

12. Roberts J, Sockett, PN The socio-economic impact of human Salmonella
enteritidis infection. Int J Food Microbiol 1994;21(1-2):117-29.

13. Levine WC SJ, Archer DL, Bean NH, Tauxe RV. Foodbome disease outbreaks in
nursing homes, 1975 through 1987. JAMA. 1991;266:2105-09.

145

14. Todd E. Preo preliminary estimates of costs of foodbome disease in Canada and
costs to reduce salmonellosis. J Food Prot. 1989;52:586-94.

15. Frenzen PD RT, Buzby J C, Breuer T, Roberts T, Voetsch D. Salmonella cost
estimate updated Using F oodNet Data. Food Review. 1999;22:10-5.

l6. Riesenberg-Wilrnes MR, Bearson B, Foster J W, Curtis R, 3rd. Role of the acid
tolerance response in virulence of Salmonella typhimurium. Infect Irnmun. 1996
Apr;64(4): 1085-92.

17. Darwin KH, Miller VL. Molecular basis of the interaction of Salmonella with the
intestinal mucosa. Clin Microbiol Rev. 1999 Jul;12(3):405-28.

18. Blaser MJ, Newman LS. A review of human salmonellosis: I. Infective dose. Rev
Infect Dis. 1982 Nov-Dec;4(6): 1096-106.

19. Rushdy AA, Wall R, Seng C, Wall PG, Stuart JM, Ridley AM, et a1. Application
of molecular methods to a nosocomial outbreak of Salmonella enteritidis phage type 4. J
Hosp Infect. 1997 Jun;36(2):123-31.

20. Wall PG, Ryan MJ, Ward LR, Rowe B. Outbreaks of salmonellosis in hospitals in
England and Wales: 1992-1994. J Hosp Infect. 1996 Jul;33(3):181-90.

21. Mennin J, Hutwagner L, Vugia D, Shallow S, Daily P, Bender J, et a1. Reptiles,
amphibians, and human Salmonella infection: a population-based, case-control study.
Clin Infect Dis. 2004 Apr 15;38 Suppl 3:S253-61.

22. Parry SM, Palmer SR, Slader J, Humphrey T. Risk factors for salmonella food
poisoning in the domestic kitchen-a case control study. Epidemiol Infect. 2002
Oct; 129(2):277-85.

23. Buchwald DS, Blaser MJ. A review of human salmonellosis: 11. Duration of
excretion following infection with nontyphi Salmonella. Rev Infect Dis. 1984 May-
Jun;6(3):345-56.

24. Miller SL PD. Salmonella species, including Salmonella typhi. In: Mandel GL,
Bennet JE, Dolin R, eds. Principle and practice of infectious diseases, 5th ed. New York:
Churchill Livingstone. 2000:2344-59.

25. Cohen J1 BJ, Corey GR. Extra-intestinal manifestations of Salmonella infections.
Medicine (Baltimore). 1987 Sep;66(5):349-88.

26. Gruenewald R, Blum S, Chan J. Relationship between human immunodeficiency

virus infection and salmonellosis in 20- to 59-year-old residents of New York City. Clin
Infect Dis. 1994 Mar;18(3):358-63.

146

27. Fisk TL LB, Guest JL, Ray S, Barrett TJ, Holland B, Starney K, Angulo FJ,
Farley MM. Invasive infection with multidrug-resistant Salmonella enterica serotype
typhimurium definitive type 104 among HIV-infected adults. Clin Infect Dis
2005;40(7): 101 6-21 .

28. CDR. Communicale Disease Reort Weekly, Health Protection Agency. Available
at: http://www.hpa.org.uk/cdr/pages/enterichtm. [Accessed on June 27, 2007].

29. Cogan TA, Humphrey TJ. The rise and fall of Salmonella Enteritidis in the UK. J
Appl Microbiol. 2003;94 Suppl:114S-9S.

30. CDA. Communicable Diseases Australia. National Notifiable Diseases
Surveillance System. Department of Health and Aging, Australia. Available at:
http://www.healthconnectgov.au/intemet/wcms/publishing.nsf/

Content/health-pubhlth-strateg-communic-index.htm. [Accessed on September 02, 2007].

31. MHLW. Ministry of Health, Labour and Walfare. Salmonellosis in Japan as of
June 2003. Infectious agen surveillance report. Available at:
ht_tp://www.mh1w.go.jp/english/index.html. [Accessed on October 9, 2007].

32. CIDPC. Center for Infectious Disease Prevention and Control. Notifiable
Diseases on-line. Public Health Agency of Canada. Available at: http://www.p11a_c;
aspcgcca/centres e.html. [Accessed on September 10, 2007].

33. CDC. Salmonella Surveillance: Annual summary, 2004. Atlanta, Georgia: US
Department of Health and Human Services, CDC 2005. (Avail at:

http://wwwcdegov/ncidod/dbmd/phlisdata/salmonella.htm.

 

34. CDC. Centers for Disease Control and Prevention. Salmonella Surveillance:
Annual summary, 1995 - 2001. Atlanta, Georgia: US Department of Health and Human
Services. 2002.

35. CDC. Centers for Disease Control and Prevention. Summary of F oodNet Annual
Report, 2004. 2003 [cited; Available from:
http://www.cdc.gov/foodnet/annual/2003/2003 reportpdf

36. CDC. Centers for Disease Control and Prevention. Preliminary FoodNet Data on
the Incidence of Infection with Pathogens Transmitted Commonly Through Food, 10
States, United States, 2005. Morb Mortal Wkly Rep. 2006;55(14):392-95.

37. Blumberg SJ, Luke JV, Cynamon ML. Telephone coverage and health survey
estimates: evaluating the need for concern about wireless substitution. Am J Public
Health. 2006 May;96(5):926-31.

38. CDC. Centers for Disease Control and Prevention. Surveillance for foodbome
disease outbreaks-United States, 1993-1997. Morb Mortal Wkly Rep. 2000;49:1-72.

147

39. CDC. Centers for Disease Control and Prevention. Outbreaks of Salmonella
enteritidis associated with nationally distributed ice cream products -Minnesota, South
Dakota, and Wisconsin. Morb Mortal Wkly Rep. 1994;43(40):740-1.

40. CDC. Centers for Disease Control and Prevention. Summary of Salmonella
serotype Enteritidis Outbreak Reported in 2002. Available at:
http://www.cdcgov/foodborneoutbreaks/salm_sum/SE2002.fin_al.pdf. [Accessed on
Januay 10, 2007].

41. CDC. Centers for Disease Control and Prevention. Outbreak of Salmonella
serotype Enteritidis infections associated with raw almonds-United States and Canada,
2003-2004. MMWR Morb Mortal Wkly Rep. 2004;53(22):484-7.

42. CIDRAP. Center for Infectious Disease Research and Policy. Turkey blamed in
South Carolina Salmonella outbreak. Available at:

http://www.cidrap.umn.edu/cidrap/content/fs/food-
disease/news/iune03055almonella.html. [Accessed on June 27, 2007].

43. ISID. International Society for Infectious Diseases. Salmonellosis, unpasteurized
orange juice - USA (multistate): alert. Available at:
http://www.promedmai1.0rg/pls/askus/f?p=2400: 1001 :420::::F2400_P1 001‘BACK_PAG
E,F2400 P1001_ARCHIVE_NUMBER,F2400_P1001_USE_ARCHIVE:1001,20050709
.1951 Y. [Accessed on June 23, 2007].

 

44. CDC. Centers for Disease Control and Prevention. Multistate outbreaks of
Salmonella infections associated with raw tomatoes eaten in restaurants-United States,
2005-2006. Morb Mortal Wkly Rep. 2007;56(35):909-11.

45. CDC. Centers for Disease Control and Prevention. Multistate outbreak of
Salmonella serotype Tennessee infections associated with peanut butter-United States,
2006-2007. Morb Mortal Wkly Rep. 2007;56(21):521-4.

46. Popoff MY BJ, Brenner F W, Gheesling LL. Supplement 2000 (no. 44) to the
Kauffrnann-White scheme. Res Microbiol. 2001;152(10):907-9.

47. Brenner FW, Villar RG, Angulo FJ, Tauxe R, Swaminathan B. Salmonella
nomenclature. J Clin Microbiol. 2000 Jul;38(7):2465-7.

48. Baumler AJ, Tsolis RM, F icht TA, Adams LG. Evolution of host adaptation in
Salmonella enterica. Infect Immun. 1998 Oct;66(10):4579-87.

49. Uzzau S, Brown DJ, Wallis T, Rubino S, Leori G, Bernard S, et al. Host adapted
serotypes of Salmonella enterica. Epidemiol Infect. 2000 Oct;125(2):229-55.

148

50. Patrick ME, Adcock PM, Gomez TM, Altekruse SF, Holland BH, Tauxe RV, et
a1. Salmonella enteritidis infections, United States, 1985-1999. Emerg Infect Dis. 2004
Jan;10(1):1-7.

51. WHO. World Health Organization. Salmonella infections. Fact sheet no. 139.
2003.

52. Roels TH, Frazak PA, Kazrnierczak JJ, Mackenzie WR, Proctor ME, Kurzynski
TA, et a1. Incomplete sanitation of a meat grinder and ingestion of raw ground beef:
contributing factors to a large outbreak of Salmonella typhimurium infection. Epidemiol
Infect. 1997 Oct;119(2):127-34.

53. Evans MR, Salmon RL, Nehaul L, Mably S, Wafford L, Nolan-Farrell MZ, et al.
An outbreak of Salmonella typhimurium DT170 associated with kebab meat and yogurt
relish. Epidemiol Infect. 1999 Jun;122(3):377-83.

54. Hall R. Outbreak of gastroenteritis due to Salmonella typhmurium phage type 1
35a following consumption of raw egg. Commun Dis Intell. 2002;26(2):285-7.

55. De Valk H, Delarocque-Astagneau E, Colomb G, Ple S, Godard E, Vaillant V, et
al. A community-wide outbreak of Salmonella enterica serotype Typhimurium infection
associated with eating a raw milk soft cheese in France. Epidemiol Infect. 2000

Feb;124(1):1-7.

56. Kapperud G, Gustavsen S, Hellesnes I, Hansen AH, Lassen J, Hirn J, et a1.
Outbreak of Salmonella typhimurirun infection traced to contaminated chocolate and
caused by a strain lacking the 60-megadalton virulence plasmid. J Clin Microbiol. 1990
Dec;28(12):2597-601.

57. Cowden JM, O'Mahony M, Bartlett CL, Rana B, Smyth B, Lynch D, et al. A
national outbreak of Salmonella typhimurium DT 124 caused by contaminated salami
sticks. Epidemiol Infect. 1989 Oct; 103(2):2 1 9-25.

58. Delarocque-Astagneau E, Bouillant C, Vaillant V, Bouvet P, Grimont PA,
Desenclos J C. Risk factors for the occurrence of sporadic Salmonella enterica serotype
typhimurium infections in children in France: a national case-control study. Clin Infect
Dis. 2000 Aug;31(2):488-92.

59. Hedberg CW, David MJ, White KE, MacDonald KL, Osterholm MT. Role of egg
consumption in sporadic Salmonella enteritidis and Salmonella typhimurium infections in
Minnesota. J Infect Dis. 1993 Jan;167(1):107-11.

60. CDC. Centers for Disease Control and Prevention. Turtle-associated

salmonellosis in humans-United States, 2006-2007. Morb Mortal Wkly Rep.
2007;56(26):649-52.

149

61. Threlfall EJ, Frost JA, Ward LR, Rowe B. Epidemic in cattle and humans of
Salmonella typhimurium DT 104 with chromosomally integrated multiple drug
resistance. Vet Rec. 1994 May 28;134(22):577.

62. CDC. Centers for Disease Control and Prevention. Preliminary FoodNet Data on
the Incidence of Foodbome Illnesses - Selected Sites, United States, 2001. Morb Mortal
Wkly Rep. 2002;51(15):325-9.

63. Angulo F JK, Tauxe R, Cohen M. Significance and sources of antirnicrobial-
resistant nontyphoidal Salmonella infections in the United States. Microbial Drug
Resistance. 2000;6(1):77-83.

64. CDC. Centers for Disease Control and Prevention. Multistate outbreak of
Sahnonella serotype Typhimurium infections associated with drinking unpasteurized
milk — Illinois, Indiana, Ohio, and Tennessee, 2002-2003. Morb Mortal Wkly Rep.
2003;52(26):613-15.

65. CDC. Centers for Disease Control and Prevention . Outbreaks of multidrug-
resistant Salmonella Typhimurium associated with veterinary facilities — Idaho,
Minnesota, and Washington 1999. Morb Mortal Wkly Rep. 2001;50(53):701-4.

66. INFOSAN. International Food Safety Authorities Network. Antimicrobial-
resistant Salmonella. INF OSAN Information note no. 3/2005 - Salmonella. 2005.

67. Wilson R FR, Davis J, LaVenture M. Salmonella in infants: The importance of
intrafarrrilial transmission. Pediatrics. 1982;69(4):436-38.

68. St Louis ME MD, Potter ME, DeMelfi TM, Guzewich JJ, Tauxe RV, et a1. . The
emergence of grade A eggs as a major source of Salmonella Enteritidis infections: new
implications for the control of salmonellosis. JAMA. 1988;259:2103-7.

69. CDC. Centers for Disease Control and Prevention. Summary of Notifiable
Diseases - United States, 2001. Morb Mortal Wkly Rep. 2003;50(53):1-108.

70. Younus M, Wilkins MJ, Arshad MM, Rahbar MH, Saeed AM. Demographic risk
factors and incidence of Salmonella enteritidis infection in Michigan. F oodbome Pathog
Dis. 2006 Fall;3(3):266-73.

71. Olsen SJ, MacKinnon LC, Goulding J S, Bean NH, Slutsker L. Surveillance for
foodbome-disease outbreaks-United States, 1993-1997. Morb Mortal Wkly Rep. 2000
Mar 17;49(1):1-62.

72. Kimura A, Reddy, V, Marcus, R. Chicken consumption is a newly identified risk

factor for sporadic Salmonella enterica serotype Enteritidis infections in the United
States: a case-control study in F oodNet sites. Clin Infect Dis. 2004;38(Suppl 3):S244-52.

150

73. CDC. Centers for Disease Control and Prevention.CDC/FDA/USDA National
Antimicrobial Monitoring System. 1996 Annual Report. Atlanta, GA. 1996.

74. Spika J S, Waterman SH, Hoo GW, St Louis ME, Pacer RE, J arnes SM, et a1.
Chloramphenicol-resistant Salmonella newport traced through hamburger to dairy farms.
A major persisting source of human salmonellosis in California. N Engl J Med. 1987 Mar
5;316(10):565-70.

75. Kirk M, Little, CL, Lem, M, Fyfe,M Genobile, D,Tan, Paccagnella, J, . An
outbreak due to peanuts in their shell caused by Salmonella enterica serotypes Stanley
and Newport- sharing molecular information to solve international outbreaks. Epidemiol
Infect. 2004;132(4):571-7.

76. Lyytikainen O, Koort J, Ward L, Schildt R, Ruutu P, Japisson E, et al. Molecular
epidemiology of an outbreak caused by Salmonella enterica serovar Newport in Finland
and the United Kingdom. Epidemiol Infect. 2000 Apr; 124(2): 1 85-92.

77. Van Beneden CA KW, Strang RA, Werker DH, King AS, Mahon B, et a1.
Multinational outbreak of Sahnonella enterica serotype Newport infections due to
contaminated alfalfa sprouts. JAMA. 1999;281(2):158-62.

78. Aseffa A, Mengistu G, Tiruneh M. Salmonella newport: outbreak of food
poisoning among college students due to contaminated undercooked eggs. Ethiop Med J.
1994 Jan;32(l):1-6.

79. Narain JP, Lofgren, J. P. Epidemic of restaurant-associated illness due to
Salmonella newport. South Med J. 1989;82(7):837-40.

80. CDC. Centers for Disease Control and Prevention. Outbreak of multidrug-
resistant Salmonella Newport -United States, J anuary-April 2002. Morb Mortal Wkly
Rep. 2002;51(25):545-8.

81. Gupta A, Fontana J, Crowe C, Bolstorff B, Stout A, Van Duyne S, et a1.
Emergence of multidrug-resistant Salmonella enterica serotype Newport infections
resistant to expanded-spectrum cephalosporins in the United States. J Infect Dis. 2003
Dec 1;188(11):1707-16.

82. Rankin SC AH, Munson R. Transmission of the blaCMY-2 gene from Sahnonella
enterica serotype Newport in the dairy farm environment. Atlanta, Georgia: Presented at
3rd International Conference on Emerging Infectious Diseases. 2002.

83. McCarthy T PQ, Mshar P, Mshar R, Howard R, Hadler JL. Outbreak of
multidrug-resistant Salmonella Newport associated with consumption of Italian-style soft
cheese, Connecticut. Atlanta, Georgia: Presented at International Conference on
Emerging Infectious Diseases. 2002.

151

84. Rice PA, Craven C, Wells JG. Salmonella heidelberg enteritis and bacteremia. An
epidemic on two pediatric wards. Am J Med. 1976 Apr;60(4):509-16.

85. F ontaine RE, Cohen ML, Martin WT, Vernon TM. Epidemic salmonellosis from
cheddar cheese: surveillance and prevention. Am J Epidemiol. 1980 Feb;111(2):247-53.

86. Bokanyi RP, Jr., Stephens JF, Foster DN. Isolation and characterization of
Salmonella from broiler carcasses or parts. Poult Sci. 1990 Apr;69(4):592-8.

87. Jones FT, Rives DV, Carey JB. Salmonella contamination in commercial eggs
and an egg production facility. Poult Sci. 1995 Apr;74(4):753-7.

88. Hennessy TW, Cheng LH, Kassenborg H, Ahuja SD, Mohle-Boetani J, Marcus R,
et a1. Egg consumption is the principal risk factor for sporadic Salmonella serotype
Heidelberg infections: a case-control study in FoodNet sites. Clin Infect Dis. 2004 Apr
15;38 Suppl 3:S237-43.

89. Jones TF, Ingram LA, Fullerton KB, Marcus R, Anderson BJ, McCarthy PV, et al.
A case-control study of the epidemiology of sporadic Salmonella infection in infants.
Pediatrics. 2006 Dec;l 18(6):2380-7.

90. Olsen SJ, DeBess EE, McGivem TE, Marano N, Eby T, Mauvais S, et al. A
nosocomial outbreak of fluoroquinolone-resistant Salmonella infection. N Engl J Med.
2001 May 24;344(21):1572-9.

91. Schutze GE, Flick EL, Pope SK, Lofgren JP, Kirby RS. Epidemiology of
salmonellosis in Arkansas. South Med J. 1995 Feb;88(2):195-9.

92. Arshad MM, Wilkins MJ, Downes F P, Rahbar MH, Erskine RJ, Boulton ML, et
al. A registry-based study on the association between human salmonellosis and routinely
collected parameters in Michigan, 1995-2001. Foodbome Pathog Dis. 2007

Spring;4(1): 16-25.

93. Schutze GE, Kirby RS, Flick EL, Stefanova R, Eisenach KD, Cave MD.
Epidemiology and molecular identification of Salmonella infections in children. Arch
Pediatr Adolesc Med. 1998 Jul;152(7):659-64.

94. Schutze GE KR, Flick EL, Stefanova R, Eisenach KD, Cave MD. The
epidemiology and molecular identification of Salmonella infections in children less than
four years of age. Arch Pediatr Adolesc Med. 1998;52:659-64.

95. Arshad MM WM, Downes FP, Rahbar MH, Erskine RJ, Boulton ML, Younus M,

Saeed AM. Epidemiologic attributes of invasive non-typhoidal Salmonella infections in
Michigan, 1995-2001. Int J Infect Dis. 2007;8z76-82.

152

IIIIIIIIIIIIIIIIIIIIIII

1111311)) 11121)

       

130 1111

9 56