if]:
2...?! ...|.!.i.v.uuv.. n.
v7.3

. .
:2 51.1
22.1}. .119
.>..:..2. .1212...
.1. :13}...
.3 .. 2

on fiat-hum; vl—z‘v‘

4

renown—wanna. “a

w. -u-.....,.u. nu...

-
v
.
.
u
u
._

J 1 I“.
.. . :5.
1&3...

 

 

 

LIBRARY
Miclégan State

University

 

 

 

This is to certify that the
dissertation entitled

Human Campy/obacter and Salmonella infections in Michigan:

Environmental Drivers

presented by

Tiffiani Joy Onifade

has been accepted towards fulfillment

of the requirements for the
PhD degree in Comparative Medicine and
Integrative Biology

 

 

 

V / Major Professor's Signature

Iz~I$—07

Date

 

MSU is an Affinnative Action/Equal Opportunity Employer

 

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5108 KzlProj/AccalPresIClRClDateDueindd

 

HUMAN CAMPYLOBACTER AND SALMONELLA INFECTIONS IN MICHIGAN:
ENVIRONMENTAL DRIVERS

By

Tiffiani Joy Onifade

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Comparative Medicine and Integrative Biology

2009

ABSTRACT

HUMAN CAMPYLOBACTER AND SALMONELLA INFECTIONS IN MICHIGAN:
ENVIRONMENTAL DRIVERS

By

Tiffiani Joy Onifade
BACKGROUND: Campylobacteriosis and Salmonellosis are common
gastrointestinal infectious diseases primarily associated with foodbome routes of
infection. Given the high incidence and health burdens of these diseases, much
research has been done to reduce human transmission via that route. However,
some of the national reported trends in the diseases (incidence peaks in the
summer months and geographic variation) have not been fully explained and
may suggest that other factors are driving these disease trends. OBJECTIVES:
This study will explore these driving factors through three major objectives: 1) to
analyze the incidence of historical human Campylobacter and Salmonella
infections in Michigan with respect to demographic, geographic, and temporal
trends (including evaluation of the seasonal high reporting period; obj 1b), 2) to
evaluate the role of environmental and climatological factors in relation to
changes in human incidence of Campylobacter and Salmonella infections in
Michigan, and 3) to evaluate methods for analyzing Campylobacter and
Salmonella environmental prevalence. METHODS: Objectives 1a & 1b:
Retrospective study design along with linear modeling statistical techniques were
used to evaluate the Michigan historical case data (1992-2005). Objective 2:

Retrospective study design was used to evaluate historical case and

environmental data creating Poisson Mixed regression models for statistically
significant relationships between the incidence of disease and specific
environmental factors. Objective 3: Published environmental sampling and
laboratory culture methods were evaluated and hybridized to create cost
effective, time efficient, reliable culture enumeration techniques. FINDINGS: 1a)
Though exhibiting similar trends, incidences of these diseases in Michigan were
substantially lower than national reports. 1b) Parameters of the seasonal high
reporting period in Michigan varies by geography. 2) Environmental and weather
related factors significantly explained some of the variation in incidence of the
diseases. 3) The hybrid culture enumeration methods evaluated produced
inconsistent results. CONCLUSIONS: This study aimed to add to the literature by
explaining and filling a gap in the chain from animals to humans. By focusing on
the environmental connections that may explain some of the variation in human
rates of these diseases, the study was able to begin evaluating a missing link.
The food route has been explained, and the water route has also been explained,
but here the environmental contamination link between animals and the
abundance of the bacteria through the environment was explored. By modeling
the environmental effects on transport and prevalence of these bacteria in the
farming and surrounding environments this has been a major step in
understanding trends in prevalence and will potentially provide insight into how to

lessen transmission between these animals and on to humans.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................. ix
LIST OF FIGURES ................................................................................................ x
INTRODUCTION .................................................................................................. 1
Purpose ............................................................................................................. 1
Specific Aims and Hypotheses .......................................................................... 2
Descriptive Studies ........................................................................................ 2
Ecological Retrospective Study ..................................................................... 4
Field study ..................................................................................................... 5
Overview ........................................................................................................... 6
CHAPTER 1 ......................................................................................................... 8
Drivers of Campylobacter and Salmonella Infections: Known and Suspected .......
A Review ................................................................................................................
1.1 Introduction ............................................................................................. 8
1.2 Objectives ............................................................................................. 10
1.3 Epidemiology of Campylobacteriosis and Salmonellosis ....................... 1O
Campylobacter spp. ..................................................................................... 13
Salmonella spp. ........................................................................................... 14
RESERVOIRS AND TRANSMISSION OF CAMPYLOBACTER AND
SALMONELLA .......................................................................................... 15
DISEASE: CAMPYLOBACTERIOSIS AND SALMONELLOSIS ............... 17
TRENDS IN DISEASE .............................................................................. 19
1.4 Environmental Associations .................................................................. 21
Precipitation Effects on Pathogen Loading in Watersheds ....................... 22
Temperature and Environmental Survival ................................................ 23
Landuse .................................................................................................... 24
Sewage disposal ...................................................................................... 24
Drinking Water Source ............................................................................. 25
Review: Campylobacter and Salmonella seasonality temperature,
precipitation, and environmental associations ............................................. 27
Temperature association .......................................................................... 27
Precipitation association ........................................................................... 29
Environmental association ........................................................................ 29
1.5 CONCLUSIONS .................................................................................... 30

iv

CHAPTER 2 ...........................................................................................................

The Epidemiology of Campylobacteriosis and Salmonellosis in Michigan ..... 33
2.0 STRUCTURED ABSTRACT ................................................................. 33
2.1 INTRODUCTION ................................................................................... 35
2.2 HYPOTHESES ...................................................................................... 36
2.3 METHODS ............................................................................................ 37

a. Study Design ......................................................................................... 37
b. Sources of Data .................................................................................... 37
c. Data Analysis ........................................................................................ 38
2.4 RESULTS .............................................................................................. 39
2.5 DISCUSSION ........................................................................................ 43
Temporal ...................................................................................................... 43
Demographic ............................................................................................... 46
Geographic .................................................................................................. 48
2.6 CONCLUSIONS .................................................................................... 49

CHAPTER 3 ...........................................................................................................
Campylobacter and Salmonella Infections in Michigan: .........................................
Evaluation of Seasonal and Geographic Trends Reporting (1992-2005) ........... 64
3.0 STRUCTURED ABSTRACT: ................................................................ 64
3.1 INTRODUCTION ................................................................................... 65
Rationale: ..................................................................................................... 65
Background: ................................................................................................. 66
3.2 HYPOTHESIS ....................................................................................... 68
3.3 METHODS: ........................................................................................... 69
a. Case Data ............................................................................................. 69
b. Evaluation of Specific Aims ................................................................... 69
c. Statistical Analysis ................................................................................ 71
3.4 RESULTS .............................................................................................. 72
3.5 DISCUSSION ........................................................................................ 73
3.6 CONCLUSIONS .................................................................................... 75

CHAPTER 4 ...........................................................................................................

Environmental Factors Influencing Rates of Campylobacter and Salmonella

Infections in Michigan ......................................................................................... 84
4.0 STRUCTURED ABSTRACT ................................................................. 84
4.1 INTRODUCTION ................................................................................... 85
4.2 HYPOTHESES ...................................................................................... 88
4.3 METHODS ............................................................................................ 88

a. Campylobacteriosis and Salmonellosis Rates ....................................... 89
b. County-level Geographic Analyses ....................................................... 90
0. Seasonal and Monthly Analysis ............................................................ 91
d. Watershed and Climate Division Geographic Analyses ........................ 91
e. Meteorological Factors .......................................................................... 91
4.4 RESULTS .............................................................................................. 93
a. Campylobacteriosis and Salmonellosis Rate$ ....................................... 93
b. County-Level Geographic Analyses ...................................................... 94
c. Seasonality ............................................................................................ 94
c. Meteorological Analysis ......................................................................... 95
4.5 DISCUSSION ........................................................................................ 96
4.6 CONCLUSIONS .................................................................................. 100

CHAPTER 5 ...........................................................................................................

Campylobacter and Salmonella on Michigan Dairy Cattle Farms: Culture Isolation

and Enumeration from Environmental Soil and Water ...................................... 109
5.0 STRUCTURED ABSTRACT: .............................................................. 109
5.1 INTRODUCTION: ................................................................................ 1 12

Background ................................................................................................ 1 12
Rationale: ................................................................................................... 1 13
5.2 HYPOTHESES .................................................................................... 1 14
5.3 SPECIFIC AIMS .................................................................................. 115
5.4 METHODS .......................................................................................... 1 15
a. Study Design ....................................................................................... 115
b. Sampling and Laboratory Procedures ................................................. 116
c. Validation Study .................................................................................. 117
d. Microcosm Temperature Study ........................................................... 119
e. Environmental Dairy Farm Sampling Study ......................................... 119
5.5 RESULTS: .......................................................................................... 120

vi

a. Validation Study .................................................................................. 120

b. Microcosm Study ................................................................................. 121

c. Field Study .......................................................................................... 121

5.6 DISCUSSION: .................................................................................... 122
a. Validation study ................................................................................... 122

b. Microcosm Temperature Study ........................................................... 124

c. Farm Sampling Study .......................................................................... 124

5.7 CONCLUSIONS: ................................................................................ 124
OVERALL DISCUSSION AND CONCLUSIONS .............................................. 135
REFERENCES ................................................................................................. 139

vii

LIST OF TABLES

Table 2.1 Incidence (per 100,000 people) within counties by quartiles of

population density. ............................................................................................. 50
Table 2.2.a. Cambylobacter Incidence (per 100,000 people) by percent
distribution of county by race .............................................................................. 51
Table 2.2.b. Salmonella Incidence (per 100,000 people) by percent distribution of
county by race .................................................................................................... 52
Table 2.3.a. Campylobacter Incidence (per 100,000 people) by percent
distribution of county by age group. ................................................................... 53
Table 2.3.b. Salmonella Incidence (per 100,000 people) by percent distribution
of county by age group. ..................................................................................... 54
Table 3.1. Michigan Counties High Reporting Period Parameters and Geographic
Data ................................................................................................................... 80
Table 3.2.Michigan High Reporting Period Parameters. ................................... 83
Table 4.1. Michigan Counties with Watershed and Climate Division Classification
......................................................................................................................... 104
Table 4.2.Description of Model Variables ....................................................... 108
Table 4.3 Univariate analysis results for climate variables with respect to disease
cases. .............................................................................................................. 109
Table 4.4. Univariate analysis results for environment/geographic variables with
respect to disease cases ................................................................................. 110
Table 5.1. Media Volume and Shaking of Soil Samples ................................. 128
Table 5.2. Campylobacter Processing: Days in Incubation ............................ 129
Table 5.3 Time Til Evaluation of Soil Samples with Background Organism
Consideration .................................................................................................. 130
Table 5.4 Time Til Evaluation of Water Samples with Background Organism
Consideration ................................................................................................. 131
Table 5.5. Range of Detection ........................................................................ 132

viii

Table 5.6 Range of Detection in Water with Background Organism Consideration
......................................................................................................................... 1 33

Table 5.7 Range of Detection in Soil with Background Organisms Consideration
......................................................................................................................... 1 34

Table 5.8 Recovery of Campylobacter and Salmonella with Respect to Length of
Time at Varying Storage Temperatures ......................................................... 135

Table 5.9 Sample Farm Results and County Characteristics .......................... 136

ix

LIST OF FIGURES
Figure 1.1 Model of bacterial transport with Environmental Influence ................ 32

Figure 2.1.a. Michigan Counties with high and low incidences of human
Campylobacter infections. ................................................................................. 55

Figure 2.1.b. Michigan Counties with high and low incidences of human
Salmonella infections. ........................................................................................ 56

Figure 2.2 Annual incidences of Campylobacter and Salmonella infections in
Michigan 1992-2005. ......................................................................................... 57

Figure 2.3 Monthly incidences of Campylobacter and Salmonella infections in
Michigan 1992-2005. .......................................................................................... 58

Figure 2.4 Campylobacter and Salmonella Incidence by Monthly average ....... 59

Figure 2.5 Average Campylobacter and Salmonella Incidence by Season ....... 60
Figure 2.6 Campylobacter and Salmonella Incidence by Age group ................. 61
Figure 2.7.a. Campylobacter Incidence by Gender ........................................... 62
Figure 2.7.b. Salmonella Incidence by Gender .................................................. 62
Figure 2.8 Campylobacter and Salmonella Incidence by race .......................... 63

Figure 3.1.a. Weekly reported cases of Human Campylobacter cases in Michigan
........................................................................................................................... 76

Figure 3.1.b. Weekly reported cases of Human Salmonella cases in Michigan 77

Figure 3.2.a. Weekly averages for Campylobacter cases in Michigan .............. 78
Figure 3.2.b. Weekly averages for Salmonella cases in Michigan .................... 79
Figure 4.1. Michigan Climate Divisions and Counties ...................................... 103

INTRODUCTION

Purpose

In recent years, there has been a surge in research on the effects of
environmental factors on, and their relations to, diseases given our changing
climate and environment on macro (global warming) and micro (urbanization)
levels, and some of this research has focused on the environmental effects on
foodbome diseases. Campylobacteriosis and Salmonellosis are common
infectious diseases caused by infection with Campylobacter spp. and Salmonella
spp and are primarily associated with foodbome routes of infection. The growth
and transport of the responsible bacteria can be influenced by the weather and
several studies have evaluated the effect of ambient temperature on foodbome
diseases that included campylobacteriosis and salmonellosis. Most causes of
campylobacteriosis and salmonellosis are foodbome, however given previous
research there is evidence that food may not fully explain the reported seasonal
increases. The role of environment and climate in the transmission of these
bacteria is complex, as there are many factors to consider (temperature,
precipitation, landuse (agricultural, urban), water source, ect.). Given all this,
there has not been a comprehensive analysis on the effects of a sum of
environmental factors on the transport of Campylobacter and Salmonella and the
association with human incidence of these infections. This study aims to add to
the literature offering that analysis by retrospectively evaluating the associations

of recorded environmental factors with reported human cases of Campylobacter

and Salmonella infections and evaluating culturable environmental prevalence of

these bacteria.

Specific Aims and Hypotheses

This research project, “Campylobacter and Salmonella Infections in
Michigan: Environmental Drivers,” consists of four smaller studies: two
descriptive studies, (The Epidemiology of Campylobacteriosis and Salmonellosis
in Michigan and Campylobacter and Salmonella Infections in Michigan:
Evaluation of Seasonal and Geographic trends in Reporting (1992-2005), an
ecological retrospective study, (Environmental Factors Influencing Rates of
Human Campylobacter and Salmonella Infections in Michigan), and an
environmental prevalence study (Campylobacter and Salmonella on Michigan
Dairy Cattle Farms: Culture Isolation and Enumeration from Environmental Soil
and Water). The long term objectives of this project were to identify and describe
trends in rates of human Campylobacter and Salmonella infections and to
evaluate the environmental and weather factors that statistically influence or
explain them. The specific aims and hypotheses for each part of the study are

given below.

Descriptive Studies
Descriptive Study 1: The Epidemiology of Campylobacter and Salmonella

Infection Rates in Humans in Michigan (Chapter 2)

. Hypotheses
o lncidences of Campylobacter and Salmonella infection in humans in
Michigan will be comparable to those nationally with respect to
temporal and demographic trends.
0 There will be geographic variation in the incidence of Campylobacter
and Salmonella infections in Michigan.
. Specific Aims
0 To analyze historical trends in Campylobacter and Salmonella rates
in Michigan with respect to temporal, demographic, and geographic
trends
0 To identify counties with consistently high and consistently low
incidence of Campylobacter and Salmonella infections
Descriptive Study 2: Campylobacter and Salmonella Infections in Michigan:
Evaluation of Seasonal and Geographic Trends in Reporting (1992-2005)
(Chapter 3)
. Hypotheses
0 There will be variation in the parameters of the seasonal reporting
peak that will we explained by geographic variables.
. Specific Aims
0 To statistically evaluate the seasonal trend in case reporting in
Michigan, identifying parameters peak-week, start and end-week,

and duration of the high reporting period

0 To evaluate the relation of geographic location and county incidence
to high reporting period parameters
Ecological Retrospective Study
Environmental Factors Influencing Rates of Human Campylobacter and
Salmonella Infection in Michigan (Chapter 4)
o Hypotheses
0 Environmental and weather related factors will be associated with
changes in rates of Salmonella and Campylobacter infections in
Michigan such that:

- As temperature increases incidence of these diseases will
increase.

- As precipitation increases incidence of these diseases will
increase.

- Areas with more agricultural land use sources will be areas
with higher incidences of these diseases.

- As percentages of homes with non-municipal water and
sewage disposal increase rates so these diseases will
increase.

. Specific Aims
0 To evaluate the role of environmental and climatological factors in
relation to changes in incidences of human Campylobacter and

Salmonella infections in Michigan.

Field study

Campylobacter and Salmonella on Michigan Dairy Cattle Farms: Culture

Isolation and Enumeration from Environmental Soil and Water (Chapter 5)

Hypotheses

Culture methods can be used in conjunction with MPN to
enumerate Campylobacter and Salmonella from environmental soil
and water samples.

Recovery of Campylobacter and Salmonella will vary with
temperature.

There will be variation in the amounts of Campylobacter and
Salmonella present in the farm soils and surrounding waters that
will relate to human incidence of these infections in county.

There will be a relationship between temperature and the amount of
Campylobacter and Salmonella in cattle farm soils and surrounding
waters such that higher prevalence of the bacteria in the

environment corresponds to warmer temperatures.

Specific Aims

0

To identify and validate methods for the enumeration of
Campylobacter and Salmonella from environmental soil and water
samples.

To evaluate the effect of temperature on Campylobacter and

Salmonella recovery from soil and water.

0 To evaluate Michigan dairy farms for the presence of

Campylobacter and Salmonella in soil and water.

Overview

This research has been conducted with the goal of describing particular
trends and geographical patterns with the hopes that this research can be used
to predict variability in incidence of campylobacteriosis and salmonellosis in
Michigan so that in future public health measures can be put into place to lessen
the transmission. This dissertation is arranged into five major sections with a
Literature Review: Drivers of Campylobacter and Salmonella Infections: Known
and Suspected (Chapter 1), an analysis of the Epidemiology of
Campylobacteriosis and Salmonellosis in Michigan (Chapter 2), the Evaluation of
Seasonal and Geographic Trends in Reporting (Chapter 3), the modeling of
Environmental Factors Influencing Incidence of Campylobacteriosis and
Salmonellosis in Michigan (Chapter 4), and Culture Isolation and Enumeration of
Campylobacter and Salmonella from Michigan Dairy Farm Environmental Soil
and Water (Chapter 5).

The literature review details the background and historical reporting on
campylobacteriosis and salmonellosis to better understand the known national
trends and drivers influencing the burden of the disease. Chapter 2 progresses
into the series of research studies by evaluating historical case data to describe
these epidemiological trends for Michigan (a state that has not been included in
the national extrapolations). The seasonal peak in reporting for both of these

infections has also been noted in the literature but, Chapter 3 goes on further

defining and describing the parameters associated with these reporting trends
and evaluating possible geographic relationships. Chapter 4 evaluates
environmental influences with the goals of explaining additional variability in
incidence of campylobacteriosis and salmonellosis in Michigan that is not related
to demographics. After the detailed evaluation of historical Campylobacter and
Salmonella infection data and relating the trends to environmental factors
(Chapters 2-4), Chapter 5 takes aim at the environmental prevalence. This
chapter evaluates methods for the culture isolation and enumeration of these

bacteria from environmental soil and water samples.

CHAPTER 1
Drivers of Campylobacter and Salmonella Infections: Known and Suspected

A Review

1.1 Introduction

The notion that climate and health are linked was suggested as far back
as Hippocrates, where he related the two around 400 BC (Rees, 1996). Through
the years the idea lingered and in the Middle Ages herbalists would prescribe
different remedies depending on the season. Today, we no longer believe that
weather itself causes disease, but we are beginning to understand how it can
create conditions for disease-causing organisms to thrive and migrate into areas
where human exposure may occur, such as water sources. These types of
relationships and the links to a changing global climate have been identified for

diseases ranging from malaria and dengue fever to cholera (Lipp et al., 2002).

Campylobacter and Salmonella are commonly reported causes of
bacterial enteritis in the United States, and throughout the world (Altekruse et al.,
1999,0berhelman and Taylor, 2000, Coker et al., 2002, USDA, 2003). They are
generally considered food borne pathogens, but waterborne outbreaks are also
known to occur (Blaser et al., 1979, Blaser et al., 1983, Palmer et al., 1983,
Skirrow, 1991, Fahey et al., 1995, Ashbolt, 2004). It is estimated that 1% of the
US. population is infected yearly by these disease, which primarily cause
gastrointestinal illness including diarrhea, nausea, and bloody stool, but could
lead to life threatening illnesses (WHO, 2003, Buzby and Roberts, 1997,

Medema et al., 1996, Nachamkin, 2002, USDA, 2003, Mead et al., 2004). The

8

health burden for these diseases is great, as there are sequales to infection for
Campylobacter and Salmonella, Guillian-Barré syndrome and reactive arthritis,
respectively, and death can occur from infection in vulnerable populations,

immune compromised and the elderly.

Campylobacter and Salmonella transmission and trends for the diseases
have been studied and noted in the literature for the United States and around
other parts of the world. Poultry (Blaser et al., 1983, Skirrow, 1991), milk (Blaser
et al., 1979, Fahey et al., 1995), and water have been implicated as major
sources of infection (Blaser et al., 1983, Palmer et al., 1983, Ashbolt, 2004). This
is of particular concern in Michigan, given the importance of the state’s

agriculture industry.

Many efforts have been made to curb transmission via the food borne
route, and nationally a decline in Campylobacter cases has been seen (Van
Gilder et al., 1999, Samuel et al., 2000, Samuel et al., 2004). This decline is
significant 23% between 1996 and 2000 (Samuel et al., 2004, CDC, 2004)
however, in recent years the national decline has slowed and Salmonella rates
have only decreases slightly. A strong seasonal effect has been observed in the
United States and elsewhere, where Campylobacter and Salmonella cases peak
in the summer months and even with the overall decline in rates, the seasonal
peaks remain (Padungton and Kaneene, 2003, Miller et al., 2004, Louis et al.,
2005, Nylen, 2002, Lindback and Svensson, 2001, Potter et al., 2002). The
relationship between infection rates and season of the year suggests a possible

link to weather patterns. Determining the factors that influence transmission of

9

the disease provides critical information for understanding the epidemiology of

these diseases.

The aim of the literature review is to formulate the background and history
of campylobacteriosis and salmonellosis to better understand the known trends

and possible drivers influencing the burden of the disease.

1.2 Objectives
The objectives of this literature review were to:
1. Explore published literature on the epidemiology of human Campylobacter
and Salmonella infections around the transmission of the diseases.
2. Evaluate studies on the environmental associations with reported cases of

these infections.

1.3 Epidemiology of Campylobacteriosis and Salmonellosis
Campylobacter spp. and Salmonella spp., causative agents of
Campylobacteriosis and Salmonellosis, are microbial pathogens known to be
transmitted through eating foods that have not been properly prepared as these
bacteria are commonly found on raw and undercooked meats (Blaser et al.1983,
Skirrow, 1991). Currently, Campylobacter is the most commonly reported cause
of acute bacterial gastroenteritis in developed countries (Mead et al. 1999) and
Salmonella annually causes 1.3 billion cases world-wide. The health burden for
these diseases is great, as there are sequels to infection for Campylobacter and
Salmonella, Guillian-Barré syndrome and reactive arthritis, respectively, and

death can occur from infection in vulnerable populations, immune compromised

10

and the elderly. Given the high prevalence and health burdens for these
diseases much research has been done examining the foodbome route to
understand the trends in transmission and to minimize infection. However, there
are other transmission routes for these bacteria that may be under appreciated in
their contributions to the trends in disease.

Globally, microbiological contamination of water is the most common source
of diarrhea-causing pathogens and may be the number one cause of childhood
mortality. In the United States alone nine million cases of waterborne disease
may occur annually (Mead et al. 1999). Campylobacter, frequently associated
with poultry and other livestock, is often transmitted by the water route (Blaser et
al.1983, Skirrow, 1991, Kapperud et al. 1992, Ashbolt, 2004). Likewise,
Salmonella, a common zoonotic agent that can be transmitted by water, is one of
the top three causes of waterborne disease (Ashbolt, 2004, Altekruse et al.
1999). However, this water route of infection has not been as extensively studied
to evaluate environmental sources of contamination and environmental
prevalence and transport of the bacteria that may lead to increased transmission
to humans.

Many factors contribute to the human incidence of Campylobacter and
Salmonella infection, and there is growing evidence that temporal and
environmental factors are associated with incidence of these diseases (Rosef
and Kapperrud 1983, Louis et al. 2006). Seasonality in human cases and
environmental prevalence of both pathogens has been noted in the literature with

peaks reported generally in the summer months (Louis et al. 2006). In the UK,

11

Campylobacter detections in watersheds increased with or just prior to peaks in
human cases in the late spring and early summer (Louis et al. 2006, Eyles et al.
2003). Highest frequency of Salmonella isolations from humans occurred in late
summer months, and was also associated with increased rainfall. Dairy cattle
infection with Campylobacter and Salmonella also show a seasonal trend in
shedding rates (Jones, 2001). Given the associations of these bacteria with
seasonal trends, temperature, and rainfall, there may be significant
environmental factors driving these relationships.

Many epidemiological studies of foodborne disease in the United States have
relied on FoodNet data (beginning in 1996) for analysis and description of the
trends in the United States (Tauxe et al. 2004). Ten states participate in the
FoodNet Program and although the states are diverse, all the data in some
states come from only a few counties in that state. FoodNet provides the basis
for a limited sample with cases only recorded from 1996 and the use of select
regions of the country. This database has been used to track the effectiveness of
foodborne pathogen control programs. In 1996, Hazard Analysis and Critical
Control Points (HACCP) rules for poultry processing were implemented and
FoodNet studies identified an annual decline in rates of Campylobacter infection
from 1996 to the present. The reports attributed this decline to the new rule
implementations, however, the seasonal trends (cases peaking in the summer

months) remained (Buchanan and Whiting, 1998, Samuel et al. 2004).

12

Campylobacter spp.

Campylobacter has been reported as a cause of human enteric disease
for over 100 years and it has probably existed for many centuries (Kist, 1985).
The first mention of a Campylobacter-like bacterium occurs in 1886 when
Theodor Escherich isolated a spiral bacterium from the intestinal mucus of
people who died of diarrheal disease and from the stool of others with enteric
diseases (Kist, 1985). Because of its similar comma-shaped appearance it was
classified in the Vibn'o genus (Sebald, 1963). In 1913, the same bacterium was
isolated from bovine fetuses (Kist, 1985). It was not until 1957, that King
described this “Vibn’o” as the agent for the enteritis and later linked it to animals
(Kist, 1985). With further investigation, Sebald and Veron (1963) found that the
metabolism of this Vibn'o was very different from others in the genus; this lead to
the classification of a new genus, Campylobacter (from Greek meaning curved
rod) (Sebald and Veron, 1963). In 1968, a technique was developed to isolate
this microphile, and then improved in 1977, to isolate Campylobacter from feces.
(Kist, 1985) This procedure allowed for further study of Campylobacter leading
to more diagnoses and treatment. For more than thirty years, Campylobacter
has been the leading cause of diarrheal illness in the United States, causing
more disease than Shigella spp. and Salmonella spp. combined (USDA, 2003,

Altekruse et al., 1999). It still persists as an important cause of enteritis.

Campylobacter spp. are motile, Gram-negative slender bacteria with a
curved rod-shape. They range in size from 0.2-0.9 pm in width to 0.5-5 pm in

length (Nachamkin, 2002). The flagellum can be monotrichous or amphitrichous

13

and moves by corkscrew motion. Members of this genus are microaerophillic and
thrive in an environment with 3-5% oxygen, 2-10% carbon dioxide, and 85%
nitrogen. They are also thermopihlic, with optimum growth conditions at
temperatures between 37 and 42° C, with better growth at the upper end of this
range. Campylobacter are susceptible to environmental stresses such as

freezing, drying, acidic conditions, and salinity (Altekruse, 1999).

The campylobacteria (which includes the genera Campylobacter and
Arcobacter), also known as campylobacters can be divided into two classes
based on a positive or negative catalase reaction. Catalase-negative
campylobacters are sensitive to oxygen and require lower oxygen content (3%
02) for growth. They are also able reduce nitrates and nitrites. Catalase-positive
campylobacters can thrive in environments with higher oxygen content (5% 02)
and are able to reduce nitrates but not nitrites. Campylobacterjejuni and C. cell,
the major causes of campylobacteriosis, are both catalase-positive

campylobacters (Butzler, 1984).

§a_lmonella spa.

Salmonella was discovered by Theobald Smith in 1885 when it was
isolated from pigs, and was named for Daniel Elmer Salmon. It was later
discovered to be relevant as a human infectious agent. In 1920, Sir William
Savage published the book “Food Poisoning and Food Infections” that detailed
the previous 40 years of food poisoning outbreaks, many due to Salmonella
(Savage 1920). Since then non-typhoidal Salmonella has been recognized as a
leading cause of gastrointestinal illness world wide.

14

Salmonella are gram-negative, non spore forming rod shaped bacteria.
They are usually motile with peritrichous flagella ranging in size from .7-1.5um
diameter to 2-5um in length. They are also thermophilic with optimum growth
conditions at temperatures between 35 and 37 C and pH between 7 and 7.5.
Salmonella are susceptible to stresses of disinfectants and high temperatures.
Salmonella are members of the family Enterobacteriaceae the genus Salmonella
and one of the two species Enterica or Bongori. The over 2400 identified
serotypes of Salmonella are further classified into two groups based on the O
(somatic/cell wall) antigens or the H (flagellar) antigens. In the United States,
serotypes S. Typhimurium, S. Enteritidis, and S. Newport make up over 50% of

serotypes isolated from infected humans.

RESERVOIRS AND TRANSMISSION OF CAMPYLOBACTER AND
SALMONELLA

Various warm-blooded animals serve as Campylobacter and Salmonella
reservoirs including poultry, cattle, swine, sheep, dogs, cats, and rodents (Atwill,
1995, Cummings et al. 2009, Stanley and Jones, 2003, Farzan et al 2009). The
USDA estimates that between 20 and 100% of retail chicken is contaminated
with Campylobacter and a recent study found 22% contaminated with Salmonella
(USDA, 2003, Lestari et al. 2009). Additionally, natural waters, sediment and
sewage sludge have been found to contain these pathogens (Droppo et al. 2009,
Ahmed et al. 2009, Lucey et al., 2000, Ashbolt, 2004, Sahlstrom et al., 2004,

Jones,2001)

15

Both campylobacteriosis and salmonellosis can occur as sporadic cases
or as outbreaks. Outbreaks come from a single source such as in the town of
Bennington, VT where 200 people were infected with Campylobacter by
consumption of non-chlorinated drinking water (USDA, 2003) or the California
Salmonella outbreaks associated with unregulated interactive water fountains
(Kirian et al. 2008). Sporadic cases are often thought to have a foodbome or
waterborne origin such as eating undercooked poultry or drinking unpasturized
milk or untreated water (CDC, 2004). The infectious dose is low for both
Campylobacter (400-500 bacteria) so one drop of juice from raw meat can cause
infection (CDC, 2004) and for Salmonella where the infectious dose can vary

based on the vector food source (Dunlop 1985).

Because the ideal growing environment for Campylobacter and
Salmonella are in warm environments and they require hosts (Altekruse et al.,
1999), it does not proliferate easily outside of the gut (Ketley, 1997). Therefore,
reservoirs provide critical links to human disease. Livestock, domestic animals,
and birds are some of the commonly known reservoirs for Campylobacter spp.
(Atwill, 1995, Lefebvre et al. 2008, Stanley and Jones, 2003) and are shed in the
feces of these animals in various concentrations throughout the year.
Campylobacterjejuni can be isolated year round from slurry tanks around sheep
farms (Stanley and Jones, 2003) and year round in varying amounts from
environmental pig slurry (Mannion et al. 2007). Land application of fecal waste
could lead to further contamination of the environment and possible runoff into

nearby waterways.

16

DISEASE: CAMPYLOBACTER/OSIS AND SALMONELLOSIS

The disease caused by any member of the Campylobacter genus is
termed campylobacteriosis or Campylobacter enteritis. Campylobacterjejuni
causes over 99% of human cases (CDC, 2004). Campylobacter enteritis is a
disease of interest to public health because of its high frequency in the
population and potential chronic effects. The symptoms of the disease include
mild or severe diarrhea often accompanied with fever and traces of blood in the
stool. Symptoms often appear within two to five days of exposure and persists
usually for one week. In immunocompromised persons, the bacteria can spread
to the bloodstream and cause life-threatening infection. Campylobacter infection
is also believed to be a precursor to Guillian-Barré Syndrome, an autoimmune
disorder that can cause paralysis (Nachamkin, 2002, Takanhashi, 2005). One in

1,000 campylobacteriosis cases lead to Guillian-Barré syndrome (Alice, 1997).

Campylobacteriosis patients are treated with antibiotics and generally
recover within one to two days. Without treatment, Campylobacter continues to
be excreted even after a patient has recovered; cells may be shed in the feces
for days to several weeks post-infection (Bulzer, 1984). Due to the amount of
time that the organism is excreted there are potential environmental ramifications
such that if sewage is not properly treated further transmission of disease is

possible.

Campylobacteriosis is a global health concern. In developing countries,
rates for Campylobacter infection are high, with 5% to 20% of the population

infected annually, depending on the country (Oberhelman and Taylor, 2000). The

17

incidence of campylobacteriosis for children under five years old in developing
countries is 40,000 cases per 100,000 children under five (40% of the <5
population; Coker et al. 2002). In general, there is an increasing incidence of the
disease in developing countries and an expanding spectrum of related diseases
caused by Campylobacter. With the high incidence of HIV in developing
countries there is consequently a greater potential for HIV-related deaths due to

Campylobacter (Coker et al. 2002).

In developed countries, the rate of infection is lower, for example 1% of
the United States population is infected each year (WHO, 2004). In the United
States and other developed countries, Campylobacter remains the most
frequently isolated bacterial enteric pathogen from clinical samples (WHO, 2004).
In 1997, the reported incidence of campylobacteriosis in the United States was
25.2 people for every 100,000 people; however, it is estimated that about 1% of
the US population are actually infected each year with Campylobacter (WHO,
2004). In the US, UK, Canada, Denmark case rates are declining (Samuel et
al., 2004, FDSCG, 2002, Samuelsson, 2004); however, in Australia cases have
risen dramatically (CDA, 2005). The prevalence of the disease among children
under 5 is also noted in developed nations but this peak is less dramatic and the
disease is still common among other age groups (Coker et al. 2002, Padungton

and Kaneene, 2003).

Salmonellosis is the disease caused by infection from Salmonella.
Serotypes S. Typhimurium, S. Enteritidis, and S. Newport make up over 50% of

serotypes isolated from infected humans in the United States. Symptoms of this

18

disease include diarrhea, fever, vomiting, and abdominal cramps. Healthy
individuals often recover within 5-7 days without treatment but occasionally
severe dehydration will require hospitalization for intravenous hydration and
antibiotics when the infection spreads from the intestines. Chronic infection can
lead to Reiter’s syndrome (pain in the joints or eyes) or chronic reactive arthritis.

There is an estimated 400 deaths annually attributed to salmonellosis.

TRENDS IN DISEASE
Demographics

Gender. In general, Campylobacter prevalence is higher in males (Potter
et al., 2002, Samuel et al. 2004, Hopkins and Olmsted, 1985). This trend is not
well understood; however, it has been suggested that this is due to poor food
handling practices more common among men or physiological differences
between the genders (Altekruse et al, 1999, Louis et al., 2005). Recently,
Younus et al. (2006) found higher rates of Salmonella infection for Michigan

females.

Age. Prevalence of Campylobacter and Salmonella infections is
distributed across age groups with the greatest number of cases reported for
children under the age of five and with Campyobactera second, smaller, peak in
the 20-29 age group (Younus et al. 2006, Potter et al., 2002, Samuel et al.,
2004). Several hypotheses have been proposed to explain this trend. First,
parents are more likely to take their young children and infants to the doctor for
symptoms of gastroenteritis (Friedman et al., 2000). Furthermore, children get

sick more frequently due to an immature immune system. Subsequently,

19

infections in childhood act to build immunity such that infection is less likely in
later years (Perez-Perez and Blaser, 2005). The second peak in

campylobacteriosis among the 20-29 year olds has not been explained.

Race. There has been little research in the area of race and
campylobacteriosis. However, in one study in the US, Blacks were noted to
have significantly lower rates than Whites, Hispanics, and Asians (Samuel et al.
2004). In the Younus et al. (2007) study there was no racial association for
salmonellosis. Arshad et al. (2007) reported a higher average annual incidence

of salmonellosis for Blacks.

Seasonality

In the US. and other parts of the world, there is a distinct peak in cases in
the summer months (Miller et al., 2004, Louis et al., 2005, Nylen, 2002, Lindback
and Svensson, 2001, Potter et al., 2002). The cause of this apparently universal
seasonal trend is not fully understood. Some hypotheses have included
increased risk of infection during peak summer travel times (Miller et al. 2004),
increased consumption of poultry products in warmer weather and a higher
likelihood of eating outdoors and outside of the home, in general (Friedman et al.
2000), and spread of Campylobacter via flies (Hald et al., 2004). In other
systematic analyses, Louis et al. (2005) found a significant relationship between
temperature change in England and Wales and seasonal campylobacteriosis
rates and Naumova et al. (2007) found similar relationships with

campylobacteriosis and salmonellosis and temperature in Massachusetts, USA.

20

Declining Cases in the United States

Data from states participating in the Centers for Disease Control and
Prevention (CDC) FoodNet program show large declines in Campylobacter and
from 1996-2000, with similar declines across all races, age groups, and genders
(Van Gilder et al., 1999, Samuel et al., 2000, Samuel et al., 2004). This is
noteworthy because the incidence rates are on the rise in other countries
(Altekruse, 1999), particularly Australia and New Zealand. Between 1996 and
2005, the national (U.S) averages for Campylobacter and Salmonella were 16.9
and 14.5 cases per 100,000 people, respectively (CDC, 2000, 2001, 2002, 2003,
2004, 2005, 2006). The steepest decline in Campylobacter infections occurred
prior to 2001, with rates declining 43% (an average of 8.7% decline annually from
1997 through 2001) then leveling off and only declining an additional 4% since
(less than 1% annually). Since 1996, the decline in Salmonella incidence has
been small (CDC, 2005). Possible explanations for these declines include
improvements in the meat processing and poultry industries due to Hazard
Analysis and Critical Control Points (HACCP) and Pathogen Reduction (PR) rule
implementations (Buchanan and Whiting, 1998, Hariharan et al., 2004, Keener,
2004). These rules, which went into effect in 1997, require the use of more water

when processing and disinfection of that water with trisodium phosphate.

1.4 Environmental Associations
One proposed environmental model for the transmission of

campylobacteriosis to humans (Skelly and Weinstre, 2003) suggests that

21

humans are exposed to the pathogen through feces, food, and aquatic
environments. While Campylobacter and Salmonella have been found in all
these environments, the modes of movement between them are not fully
understood. Figure 1 illustrates the some of the transmission modes mentioned
for Campylobacter and Salmonella that could be influenced by environmental
factors. In this section we examine some environmental factors (Weather
[precipitation and temperature], Landuse [agricultural], Sewage disposal, and

Water source) and their possible associations with enteric diseases.

Precipitation Effects on Pathogen Loading in Watersheds

Changes in precipitation can affect the loading of enteric pathogens in
waterways. Significant runoff and subsequent contamination of watenlvays after
extreme rain events is a common occurrence (e.g., Lipp et al. 2001, Lipp et al.
2002, Leeming et al. 1998; Patz, 2001). The presence of waterborne disease
agents, including Giardia cysts, Cryptospon'dium oocysts, and enteric viruses,
have been positively correlated with rainfall (Graczyk et al. 1999, Patz, 2001,
Kristemann et al., 2002, Lipp et al., 2001). Microbial contamination in drinking
water reservoirs in parts of Germany has been shown to increase by as much as
1- to 2-logs during extreme rainfall and runoff events (Kristemann et al., 2002).
In areas, such as Florida, where wet winters are correlated with El Nino events, a
direct relationship between the El Nino Soulthern Oscillation (ENSO) state and
water quality (measured by fecal coliform bacteria) has been noted (Lipp et al.

2001). This is one of the only studies that has been able to relate ENSO events

22

to the change in local weather patterns and then to discrete changes in water

quality (Tampa Bay, FL).

Temperature and Environmental Sun/ival

Despite the host requirement, Campylobacter and Salmonella are
routinely found in environmental sources such as water, sediment, and sewage
(Haley et al. 2009, Droppo et al. 2009, Buswell et al., 1998, Lucey et al., 2000,
Ashbolt, 2004, Sahlstrom et al., 2004, Jones, 2001). There has been the
discovery of a potentially environmentally adapted strain of Campylobacterjejuni
that is prevalent in northwest England surface waters in late spring (Sopwith et
al. 2008). Experimentally, for short periods of time, Campylobacter spp. can
survive in sterile water but their survival increases when associated with a biofilm
and at lower temperatures (Bruswell et al. 1998). In sterile water at 37° C
Campylobacter survived an average of 21 .8 hours while at lower temperatures
the survival times went up with highest survival in sterile water at 4° C (201.6
hours). When autochthonous microflora were added to the microcosms to better
represent the natural environment, survival rates increased significantly to ~ 200
hours at 30° C and ~ 550 hours at 4° C (Bruswell et al. 1998). Furthermore, by
infecting protozoa (i.e., Acanthamoeba polyphaga) C. jejuni is able to prolong its
survival in the environment and outside of a vertebrate host (Axelsson-Olsson et
al., 2005). In a Salmonella almond soil microcosm, Salmonella recovery
decreased more quickly with spiked samples stored at 35 C as compared to
those at 20 and at 180 days could still be recovered from samples stored at 20°C

but could not be detected in 35 ° C samples (Danyluk et al. 2008).

23

Landuse

The transport of pathogens via runoff can increase concentrations of
waterborne pathogens in impacted watersheds (Ferguson el al., 2003); in turn
the amount and quality of run off is directly related to land use. Runoff is affected
by the amount and intensity of precipitation, surrounding land use, soil type, and
topography (USGS, 2005, Tsubo, 2005, Sheresta, 2003). Sherestha (2003)
suggested that urban land use resulted in the highest level of runoff followed by
residential (village) areas, agricultural land, pasture land, and forests. These are
related to land cover by impervious surfaces. Changes in land use have been
associated with the emergence of pathogenic diseases in many regions of the
world (Patz, 2001). Some of the land use changes include human settlement,
commercial development, and road construction. Combinations of these types of
changes have been linked with emergence of diseases such as malaria and
schistosomiasis (Patz, 2001). Several studies have further implicated land use in
the contamination of waterways (Interlandi & Crockett, 2002). Significant
concentrations of fecal indicator microbes are found in waters that drain from
confined livestock farming operations (Crowther et al., 2001 ). This information
suggests that along with weather factors, the use of the land is an important

factor in the amounts of pathogens in watersheds.

Sewage disposal
Proper disposal of wastewater is also an important consideration when
investigating modes of disease transmission. Public means of sewage disposal is

regulated by local, state, or federal agencies. The remainder of the State uses

24

other means of disposal, usually a private on-site disposal system (OSDS; e.g.,
septic systems and cess pits), which do not include a mechanism for disinfection
of waste. Septic systems include a tank which allows solid material to collect and
scum to surface while the liquid portion is allowed to go into a leach field where
the soil can assist in the filtration of microbes and organics from the waste water
(American Ground Water Trust, 2005). Cesspools are less common and are
simply pits where sewage is dumped. Local ordinances provide guidelines on
how to properly locate these private systems but beyond that it is the
homeowner’s responsibility to ensure it is working properly. This is of particular
importance because of the known links between sewage-contaminated water
and human illness (Haflinger, 1999, Kambole, 2003, Exner, 2001). Public
sewage treatment facilities have more stringent guidelines; however, all facilities
are not required to perform tertiary levels of treatment which may be necessary
to kill many microbial contaminants. Sahlstrom et al. (2004) found that 55% of
sludge samples treated by common methods for secondary treatment
(sedimentation, mesophilic or thermophilic aerobic digestion, composting, and
storage) were positive for Salmonella and other potentially harmful microbes.
Sludge, also known as biosolids, is often applied directly to land for use as

fertilizer and may present a risk for infectious diseases (Sahlstrom et al., 2004).

Drinking Water Source
Waterborne disease agents have been identified as a major concern for
human health (Patz, 2001). It has been estimated that in North America, 15-30%

of gastrointestinal disease is a result of contaminated water (Ashbolt, 2004).

25

Once pathogens are in the watersheds, proper treatment of the water is required
before consumption to prevent human infection and disease, including,

campylobacteriosis and salmonellosis (Ashbolt, 2004).

Waterborne disease outbreaks have been a problem in the United States
for many years. The US Environmental Protection Agency (EPA) regulates
public drinking water systems that serve over 25 people; nationwide, these public
systems serve 90% of the population (US Census, 1990). For those that are not
served by public sources, individual wells are used. These wells are not
regulated by the EPA but suggestions are given to prevent contamination of the
water and each state determines the exact ordinances for that state. Some of
the EPA’s suggestions are for wells to be placed at least 50 feet from septic
tanks and leach fields, silos, and livestock yards, 100 feet from petroleum tanks,
liquid tight manure storage, and fertilizer storage and handling, and 250 feet from
manure stocks (EPA, 2005). The regulation of these water sources are the
responsibility of the homeowner who must carry out any testing to ensure water
safety. The depth of private wells can also indicate likelihood of becoming
contaminated. Drilled wells (deep wells of 100-1000 feet), are drilled below the
bedrock and get water from confined ground water sources, while dug wells (10-
30 feet deep) and bored or driven wells (30-100 feet deep), tap water from the
saturated zone above the bedrock (an unconfined water source) which is more

easily contaminated (EPA, 2005).

26

Review: Campylobacter and Salmonella seasonalitv temperature, precipitation,
and environmental associations

A systematic review of the literature was conducted to examine how
seasonal patterns of campylobacteriosis and salmonellosis, both primarily
foodbome diarrheal illnesses (though the seasonal pattern is not fully explained
through this route), relate to precipitation and temperature fluctuations and the
environment (proxy measured by geographic distribution). Searches were
conducted through the PubMed database and Google scholar and through a
further snowball effort (researching the references of relevant articles to identify
others) using key search terms precipitation, ambient temperature, seasonality,
campylobacteriosis, salmonellosis, climate, environment, spatial distribution and
any variation of those terms. The aim of this search was to identify studies where

associations of temperature, precipitation, or geography were evaluated.

Temperature association

Incorporating a 4 week lag, Patrick et al. (2004), in the Denmark study
found 68% of variation in human Campylobacter incidence could be explained by
maximum temperature in a univariate model. England and Wales Tukey
transformed data analyzed with autoregression techniques showed increased
Campylobacter rates were correlated with temperature (Louis et al. 2006). In
another English study a one degree rise in temperature corresponded to a 5%
increase in the number of Campylobacter reports (Tam et al. 2006). Generalized
linear models and additive models were used by Fleury et al. (2005) in the

Canadian study to identify a non linear association between weekly

27

Campylobacter cases and temperature such that log relative risk increased by
2.2% for every degree increase in weekly mean temperature in Alberta and 4.5%
in NeMoundland-Labrador. A similar temperature association was seen in
Massachusetts, USA where Campylobacter daily incidence peaks 2-14 days
following ambient temperature peaks (Naumova et al. 2007). Spanning Europe,
Canada, Australia, and New Zealand, a slight 3 month lag association was
reported (Kovats et al. 2005) and Australian study showed some unique findings
with inverse associations reported in Adelaide and positive associations in

Brisbane (Bi et al. 2008).

Several of the identified studies evaluated associations between
Salmonella reported cases or incidence and temperature. Generalized linear
models and additive models were used by Fleury et al. (2005) in the Canadian
study to identify a non linear association between weekly salmonella cases and
temperature such that log relative risk increased by 1.2% for every degree
increase in weekly mean temperature. In Australia, D’souza et al. (2004) reported
a positive association (1 month lag) between salmonellosis notifications and
mean monthly temperatures through a log-linear model. In the Naumova et al.
(2007) study, daily salmonellosis incidence peaked 2-14 days after the ambient
temperature peak in Massachusetts, USA. Similarly, Zhang et al. (2008) found a
2 week lag associated with increase in cases in Adelaide, Australia. One study
reported a 1 week time-series lag association consistent across six European
countries, however, this was the only study to suggest that this a link to food

handling practices. All of these studies found a positive relationship between

28

ambient temperatures and cases of Salmonella infections however the specifics
of the findings (strength of association and lag), methods of analysis, and
location of the study vary. This may suggest that there is a universal temperature
association but more research would need to be done to better define what that

association is and what it means.

Precipitation association

Patrick et al. (2004), Denmark study also found precipitation to explain 6%
of variation in human Campylobacter incidence with a 3 week lag in a univariate
model. When seasonality of Campylobacter incidence was examined in England
and Wales variations of measures for precipitation were used (continuous
amount of rain vs. dichotomous rain yes or no) and only with the inclusion of
temperature were small amounts (1%) of the variation in incidence explained and
only for certain regions (Louis et al, 2006).

In the Zhang et al. (2008) study, when using a seasonal autoregressive
integrated moving average (SARIMA) model rainfall was inversely related to the

number of Salmonellosis cases in Adelaide, Australia.

Environmental association

As noted in earlier sections (1.4) there are many environmental factors
that could be associated with variation in transmission or incidence of Salmonella
and Campylobacter infections. Under the assumption that environmentally
related illness will show geographic clustering, studies assessing incidence
distribution were evaluating (the spatial variation is used as proxy measure for
environmental influence).

29

Many studies noted the spatial variation in Campylobacter and Salmonella
infection incidences. Of particular note, the study by Jepsen et al. (2009)
attempted to evaluate the clustering of Campylobacter incidences in a Danish
county. This study clusters data based on space and time under the premise that
data without environmental influence should be randomly distributed in space,
finding that there was clustering around the northwestern portion of the study
area. The researchers note that this may indicate an environmental “cause” in

that area and further research should be done to identify it (Jepsen et al. 2009).

The Louis et al. (2006) study not only reports on the geographic variation
in Campylobacter rates, but attempts to explain some of that variation by
examining the high rate areas (rural and agricultural). The study was not able to

link surface water with incidence.

Studies were not identified that specifically examined the environmental

associations with human incidence of Salmonella infection.

1.5 CONCLUSIONS

Published literature has shown the relationship of climate to health and
disease. Large climatic events affect global and local weather patterns resulting
in increased precipitation and runoff. Based on the type of land use this runoff
can be great and can contain pathogens such as Campylobacter or Salmonella.
These pathogens are able to persist in natural waters, where humans may be
exposed. The source of the drinking water may also be a key factor in the

transmission of the disease. The goal of the research, presented in this

30

dissertation, is to evaluate significant weather and other environmental factors for
their association with and potential driving force influencing changes in Michigan

Campylobacter and Salmonella case rates.

31

Figure 1.1: Model of Bacterial transport with Environmental Influence

Food

   

  

Humans

   
 

Animal - soil

carriers
g)

k____/
Weather factors ‘

This figure shows the various known infection pathways for human infection with

 

 

 

 

 

   
 

 

   
 

  

water

 
 

Environment

 

Campylobacter and Salmonella Infections. The larger circle suggests the

variables that could be influenced by environment and weather factors.

32

CHAPTER 2

The Epidemiology of Campylobacteriosis and Salmonellosis in Michigan

2.0 STRUCTURED ABSTRACT

Background - Campylobacter spp. and Salmonella spp. are some of the most
commonly reported causes of bacterial enteritis in the United States; however,
relatively little is known about regional and local scale variability of these
diseases.

Specific Aims - To describe demographic, temporal, and geographic trends of
campylobacteriosis and salmonellosis in Michigan, US 1) analyzing historical
trends in Campylobacter and Salmonella rates in Michigan with respect to
temporal, demographic, and geographic trends and 2) identifying counties with
consistently high and consistently low incidence of Campylobacter and
Salmonella infections.

Design - Retrospective descriptive study

Methods and Results - Data were analyzed on culture-confirmed cases of
campylobacteriosis and salmonellosis from 1992-2005. The average annual
incidence of these diseases in Michigan were 4.3 cases per 100,000 people
(ranging from a high of 6.3 cases per 100,000 in 2004 to a low of 3.1 cases per
100,000 in 1997) for campylobacteriosis and 4.5 cases per 100,000 people
(ranging from a high of 5.8 cases per 100,000 in 1998 and 2004 to a low of 3.5
cases per 100,000 in 1997) for salmonellosis. Incidence among the 0-4 age
group for both diseases (7.7 Campylobacter spp. and 13.7 Salmonella spp.

cases per 100,000) were significantly higher than all other age groups. There

33

were no significant differences for males and females. A marked seasonal trend,
for both diseases, was also evident with rates peaking in the summer months.
Geographically, incidence varied across the state among counties with no
campylobacteriosis cases reported in Baraga and Lake to a mean annual high of
22.1 Campylobacter cases per 100,000 in Menominee and from 0.28 Salmonella
cases per 100,000 in Cheboygan to 17.3 in Wexford. Case rates state-wide for
both diseases were significantly higher in counties with intermediate population
densities, 55-144 people/miz. Counties most frequently noted as the highest case
rate counties in the state were identified and evaluated to find no distinguishable
geographic trend, however, demographically these counties are all >90% white,
have a large elderly population, and have a low population density. Overall,
Michigan rates are lower than the national rates (12.7 Campylobacter cases per
100,000 and 14.6 salmonella nationally) and have no annual trend while
nationally annual rates have been on the decline (for the period of this study).
Conclusions - There is geographic and seasonal variation in reporting of
campylobacteriosis and salmonellosis and demographics of the high reporting
areas do not follow expected trends. These results may suggest that non-
demographic factors, including environmental influences, may affect the rates of
campylobacteriosis and salmonellosis in these areas.

Significance - This study adds to the growing body of information on the
epidemiology of Campylobacter and Salmonella in the United States. Through
this retrospective descriptive study, data from Michigan available beginning in

1992 (four years prior to FoodNet) will provide information comparable to

34

FoodNet, as Michigan is not a FoodNet state and has not previously been

represented in the national reported data.

2.1 INTRODUCTION

Campylobacter and Salmonella are commonly reported causes of
bacterial enteritis in the United States, and throughout the world (Altekruse et al.,
1999,0berhelman and Taylor, 2000, Coker et al., 2002, USDA, 2003). They are
generally considered a foodbome pathogen, but waterborne outbreaks are also
known to occur (Blaser et al., 1979, Blaser et al., 1983, Palmer et al., 1983,
Skirrow, 1991, Fahey et al., 1995, Ashbolt, 2004). Campylobacter and
Salmonella transmission and trends for the diseases have been studied and
noted in the literature for the United States and around other parts of the world.
These trends include temporal (declining incidences beginning in 1996 and
seasonal peaks in the summer months), demographic (men having higher
incidences than women and children under five years old having the highest
incidences for age groups), and geographic (rates vary from rural to urban areas)
aspects (Buchanan and Whiting, 1998, Van Gilder et al., 1999, Samuel et al.,
2000, Lindback and Svensson, 2001, Allos et al., 2004, Hariharan et al., 2004,
Keener, 2004, Samuel et al., 2004, CDC, 2004, USDA, 2006). However in the
United States, national rates and the trend in rates are determined based on the
Food Net Program, which only began in 1996 and may not be generalizable to
the entire country (Hardnett et al., 2004). Michigan is a non-Food Net state and
has complete data archived from 1992. Michigan also has a diverse population to

compare to the nation demographically, distinct seasons to examine seasonal

35

trends, and geographic variation with urban centers and large agricultural areas.
Here we evaluate the trends in reported cases in Michigan with respect to
national findings.

This study is descriptive and includes a retrospective analysis of
Campylobacter and Salmonella infection incidences in Michigan from 1992-2004.
This study aims to examine historical data on reported cases of Campylobacter
and Salmonella infection with respect to geographic (state and county), temporal
(year, season, and month), and demographic trends (gender and age). This
study also aims to compare these trends to reported national data and identify
Michigan counties with consistently high and low incidence of disease for future

study.

2.2 HYPOTHESES
The specific hypotheses tested in the study were:

1) incidences of human Campylobacter and Salmonella infection in Michigan will
be comparable to those nationally with respect to temporal and demographic
trends; and that

2) there will be geographic variation in the human incidence of Campylobacter

and Salmonella infections in Michigan.

2.3 METHODS

a. Study Design

A retrospective study design was used to evaluate the aforementioned
hypotheses. Historical data were collected from public data sources and linked

for further analysis based on associated data identifiers.

36

b. Sources of Data

Case Data: Michigan local health departments are responsible for
supervising the collection and reporting of notifiable disease data from health
boards, practitioners, and laboratories in theirjurisdiction. The Michigan
Department of Community Health (MDCH) receives all these reports of culture
confirmed laboratory human cases of Campylobacter and Salmonella infections.
Information from these records was abstracted from the period of 1991-2006.
These abstracted de-identified (per Human Subjects Exempt Research
Protocols) data included race, county, organism (Campylobacter or Salmonella),
gender, age, and year, month, and day of disease onset. Annual and monthly
state and county incidence and population densities were calculated using these
case data, population data, and county area data (Incidence = (# cases/
population)*100,000).

Population Data: To evaluate the relationship between county-level case
rates and demographic characteristics of Michigan counties, state and county
level population data were collected from the US Census Bureau. This data were
collected from the annual estimates for Michigan for the demographic subsets by
race, county, gender, age group, and state from the years of 1992-2005.

County land area data were collected from the US Geological Survey and
assessed with the population data to calculate annual state and county

population densities.

37

c. Data Analysis

Descriptive analyses of annual case rates (per 100,000 people) among
categories of gender, age group, race, month, and season were performed for
the entire state of Michigan and for each of 83 counties. Monthly case rates for
age group, gender, and race were also calculated for each county. In instances
where there was a missing identifier, the case was excluded from calculation
based on that identifier. For seasonal analyses, months were collapsed into
seasons defined as spring (March, April, and May), summer (June, July, and
August), fall (September, October, and November), and winter (December,
January, and February). Data were analyzed for differences in mean rates
between counties and for temporal trends.

The distributions of cases among all demographic variables were
calculated for each county and compared to that county’s case rates in the
general population. Counties were grouped according to the distribution of
demographic variables (deciles) by age, gender, race, and geographic location.
Population density was calculated as the number of people per square mile in a
given county. Quartiles for population density were determined by evaluating the
distribution of total number of data points (all counties for all years). Percent
distributions of county populations were calculated for each age group, gender,
and race. Differences in county case rates were analyzed among deciles for
population demographic distribution and quartiles for population density analysis.

To determine statistically significant differences in case rates among study

variables, an analysis of variance (PROC ANOVA) was performed using SAS

38

software (v.9.1, ESRI, Redlands, CA) and post hoc Least Significant Difference
(LSD) or Student-Newman-Keuls (SNK) tests were used to determine the pair-
wise differences. In all measures, statistical significance was declared when
p<0.05.

High and Low incidence counties: To avoid outliers a method was
developed to identify counties that consistently have the highest and lowest
incidence of Campylobacter and Salmonella infections over the period of study.
This method takes into account consistency and overall rates for evaluation.
Consistently high or low incidence counties are defined as counties with
incidence of disease that shows up with high frequency in the upper 25% of
counties for highest or lowest monthly incidence over the period of study. The
overall rates were evaluated by identifying counties with the highest overall
annual mean and monthly mean. Counties that showed up on all these lists were

determined to be the “High and Low incidence counties”.

2.4 RESULTS

Campylobacter: Geographic. The mean annual incidence of
Campylobacter infections in Michigan was 4.28 cases per 100,000 people. There
were nine counties with significantly higher incidences of Campylobacter
infections. These counties are Menominee (22.1 annual cases / 100,000 people),
Marquette (16.7), Missaukee (14.2), Leelanau (13.6), Wexford (13.4), Isabella
(13.0), Hillsdale (12.8), Berrien (12.1), and Alcona (12.0). The counties of

Mackinac (0.65 cases I 100,000 people), Cheboygan (0.63), Presque Isle (0.55),

39

and Gogebic (0.45) had the lowest incidences with the exceptions of Baraga and
Lake that reported no cases over the period of record (Figure 2.1a).

Temporal. Incidence of campylobacteriosis in Michigan was on a marked
decline from a high of 6.16 cases/ 100,000 in 1992 to a low of 3.1 per 100,000 in
1997. Thereafter, the rates rose to 6.29 in 2004 (Figure 2.2). Evaluation of the
monthly incidence shows the cyclical pattern of reported disease incidence
where cases were significantly highest in the summer months peaking in July
with a mean monthly incidence of 0.72 cases per 100,000 people and the cases
were significantly lowest in the winter and spring from November through April
(monthly incidences of 0.26, 0.20, 0.19, 0.17, 0.21, and 0.22, consecutively)
(Figure 2.3 and 2.4). Seasonally, the summer was statistically higher than all
other seasons (Figure 2.5).

Demographic. The 0-4 age group had the highest incidence of disease
(7.7 cases I 100,000 people) while the 5-19 age groups had the lowest (5-9 age
group: 2.9, 10-14 age group: 2.3, and 15-19 age group: 2.5 cases / 100,000
people) (Figure 2.6). There was no significant difference in incidence between
males and females (4.5 and 4.0 cases I 100,000 people, respectively) (Figure
2.7a). African Americans had significantly lower incidence than all other groups
(Figure 2.8).

Salmonella: Geographic. The mean annual incidence of Salmonella
infections in Michigan was 4.50 cases per 100,000 people. All counties reported
cases of Salmonella infections over the period of record with Wexford (17.3

annual cases I 100,000 people), Benzie (14.1), Missaukee (13.6), Marquette

4O

(13.2), Lenawee (13.2), and Menominee (12.6) having the highest mean
incidences over the period. Mackinac (.79 cases I 100,000 people), Midland
(.68), and Cheboygan (.28) had the lowest mean incidences (Figure 2.1b).

Temporal. Incidence of salmonellosis in Michigan shows no significant
overall trends. The incidence in 1992 was 4.5 and remained between 4.5 and
the overall minimum reached in 1997, 3.5 cases I 100,000 people until 1998,
when it increased to 5.8 cases I 100,000 people. From than rates declined
through 2003 to a local minimum of 3.9 and increased again in 2004 to 5.8 once
again (Figure 2.2). Evaluation of the monthly incidence shows the cyclical pattern
of reported disease incidence where cases were significantly the highest in the
summer months peaking in July with a mean monthly incidence of .67 cases I
100,000 people and the cases were lowest in the winter from November through
February (monthly incidences of 0.29 cases / 100,000 people, 0.23, 0.23, and
0.21, consecutively) (Figure 2.3 and 2.4). Seasonally, incidence of disease was
statistically higher in the summer months (Figure 2.5).

Demographic. The incidence of disease for the 0-4 age group was
significantly higher than all other groups (13.7 casesl 100,000 people). The
lowest incidences were seen in the 50-59 age group (3.2 cases I 100,000
people) and the 10-14 age group (3.0) (Figure 2.6). There was no significant
difference in incidence between males and females (4.2 and 4.7 cases / 100,000
people, respectively) (Figure 2.7b). African Americans had statistically

significantly lower incidence of disease than other groups (Figure 2.8).

41

County Demographic Qistribgtions: There were significant differences in
incidence for counties based on population density. Counties with population
densities in the 3rd quartile (population density of 55 - 144 people per square
mile) had significantly higher rates than other population density county
categories (Table 2.1). County racial distribution did not show significant
differences between county groups (Table 2.2a, 2.2b). County age distribution
did not show counties with significantly skewed populations having
disproportionate incidence (Table 2.3a, 2.3b).

High and Low Incidence Cognties: The high incidence counties were
determined to be counties that consistently were in the upper or lower quartiles
for high or low annual incidence. The counties that appeared the most frequently
in the upper quartile for Campylobacter are Menominee, Marquette, Isabella,
Wexford, Leelanau, Emmet, Alcona; and for Salmonella are Wexford, Benzie,
Branch, Kenweenaw, Oscoda (Figure 2.1). Counties most frequently appearing
in the lower quartile for Campylobacter are Arenac, Baraga, Cheboygan,
Genesee, Gogebic, Lake, Lapeer, Mecosta, Presque Isle, Saginaw, Shiawassee,
St Clair, Wayne; and for Salmonella are Arenac, Bay, Cheboygan, Eaton,
Gladwin, Grand Traverse, Macomb, Mecosta, Midland, Montclam, Ogemaw,

Sanilac, Shiawassee, St Clair, Tuscola, Wayne (Figure 2.1).

2.5 DISCUSSION
In this study historical incidences of Campylobacter and Salmonella cases
in Michigan were analyzed with respect to temporal variation, geographic area,

and demographic variation. It was expected that the incidences of

42

Campylobacter and Salmonella infections in Michigan would be comparable to
national data in terms of averages and with respect to temporal and demographic

trends and that there would be geographic variation in rates across the state.

Temporal

During the 14 year period of analysis (1992-2005), 6,111 culture-
confirmed cases of Campylobacter infection and 6,483 Salmonella cases were
reported in Michigan, at means of 437 and 463 cases per year and mean
incidences of 4.3 and 4.5 Campylobacter and Salmonella cases per 100,000
people, respectively. Between 1996 and 2005, the national (U.S) averages for
Campylobacter and Salmonella were 16.9 and 14.5 cases per 100,000 people,
respectively (CDC, 2000, 2001, 2002, 2003, 2004, 2005, 2006). Nationally, there
was a steep decline in Campylobacter infections prior to 2001, with rates
declining 43% (an average of 8.7% decline annually from 1997 through 2001)
then leveling off and only declining an additional 4% since (less than 1%
annually). In Michigan during the same time period, there is no distinguishable
corresponding overall trend (Figure 2.2). However, there was a steep decline that
ended in 1997 and began at the beginning of the recorded period for an overall
decline of 49% (9.9% annually 1992 through 1997). Campylobacter incidence
then increased dramatically throughout the remainder of the record period. Since
1996, the decline in Salmonella incidence has been small (CDC, 2005). In
Michigan, over the reporting period, there have been no significant trends in

Salmonella incidence.

43

Nationally, the decline in Campylobacter and Salmonella rates has been
observed since 1996 with similar declines across all races, age groups, and
genders (Samuel, 2004). This is significant due to the fact that around the world
in both developed and developing countries the incidence of campylobacteriosis
has risen substantially over the past 20 years (Coker et al., 2002). Several
possible explanations for this disparity have been suggested, including
improvements in the meat processing and poultry industries due to Hazard
Analysis and Critical Control Points/ Pathogen Reduction (HACCPIPR) rule
implementations (Buchanan and Whiting, 1998, Allos et al., 2004, USDA, 2006).
These rules require the use of more water when processing and disinfection of
that water with trisodium phosphate; however, these were implemented in 1997
and does not explain the significant decline noted in Michigan that begins at the
beginning of the recorded period (1992). Another possible reason for the decline
is better education of the public on food safety (Samuel, 2004).

Another finding of this study is that the mean rates of both Campylobacter
and Salmonella infections in Michigan (4.3 and 4.5 cases I 100,000 people,
respectively) over the period of study were far below the national averages (12.7
and 14.6) (CDC, 2006). Further study will be needed to fully explain this
difference, however it is possible that climate factors may explain some of this
difference as the expected seasonal temporal trends are seen in Michigan with
high incidence in the summer months and low incidence in the winter months, but
due to Michigan’s high latitude the winter low lasts for six months while the

summer high is only present for one month.

44

Campylobacteriosis and Salmonellosis cases peak in the summer months
(from May to July in the northern hemisphere) across the US and around the
world (Miller et al., 2004, Louis et al., 2005, Nylen, 2002, Lindback and
Svensson, 2001, Potter et al., 2002). The cause of this apparently universal
seasonal trend is not fully understood. Some hypotheses have suggested an
increased risk of infection during peak summer travel times (Miller et al., 2004,
Coker et al., 2002; Louis, 2005), increased consumption of poultry products in
warmer weather and a higher likelihood of eating outdoors and outside of the
home, in general (Friedman et al., 2000). Jones (2001) suggests that the
seasonal trends are due to variations in Salmonella and Campylobacter
infections of livestock and poultry flocks, which could be associated with
increased Campylobacter transmission by flies in the summer months along with
variations in environmental loading (Rosef and Kapperud, 1983; Hald et al.,
2004; Nichols, 2005). In this study, highest rates in Michigan were also reported
in the summer, particularly in July, followed by June and August. In a systematic
analysis, Louis et al. (2005) found a significant relationship between temperature
change in England and Wales and seasonal campylobacteriosis rates suggesting

that environmental factors such as climate affect case rates.

Demographic

There was a high degree of variability in campylobacteriosis rates within
the state over the period of record, which was not related to the demographic

makeup of the county as Michigan counties followed all expected reported

45

demographic trends; high rates in the 0-4 age group, men higher rates than
women, and Blacks lower rates than other racial groups.

The distribution of Campylobacter and Salmonella infection rates peaked
in the 4 year and under group and is similar to previous reports (Friedman et al.,
2000, Louis et al., 2005, Perez-Perez and Blaser, 2005). Campylobacter and
Salmonella, as well as other pathogens that cause gastroenteritis are vastly
underreported in the general population (Gillespie et al., 2002). It has been
speculated that high rates noted in the under five age group may be a reporting
bias, with parents being more likely to take their young children to the doctor for
symptoms of gastroenteritis (Friedman et al., 2000). Because of this Louis et al.
(2005) suggested that this group may better represent the actual case load. In
addition to a reporting bias, children get sick more frequently due to an immature
immune system (Perez-Perez and Blaser, 2005). Subsequently, infections in
childhood act to build immunity such that infection is less likely in later years
(Perez-Perez and Blaser, 2005). It is unclear what factors may have lead to the
second peak in campylobacteriosis among the 20-29 year olds.

In this study we also found a slight difference between the rates of
salmonellosis and campylobacteriosis in males and females in Michigan, with
males having higher rates. This is consistent with previous studies although this
trend is not well understood (Potter et al., 2002, Samuel et al., 2004). It has
been suggested that this is due to poor food handling practices more common
among men; however this seems unlikely given that the trend was evident

among all age groups including those <5 (Samuel et al., 2004). Therefore,

46

physiological differences between the sexes may explain this trend (Altekruse et
al., 1999).

There was a significant difference in incidence of campylobacteriosis and
salmonellosis based on race, with rates for blacks significantly lower than for
whites and other races. This is consistent with previous findings; however it is
unclear what may drive this trend. One explanation may be cultural differences
that result in different consumption and food preparation patterns. Samuel et al.
(2004) speculate that blacks may be less likely to be seen by a physician for mild
gastrointestinal illnesses like Campylobacter or Salmonella infections. This is
suspected because blacks have the highest rates of hospitalizations due to
Campylobacter infection (Samuel et al., 2004).

High case rates were found in the counties with disproportionate
populations of these high rate groups. Noting the state-wide demographic trends,
the counties with the highest mean case rates for the total population have no
apparent similarities. This suggests that other factors not identified here may be
affecting the case rates in these areas. Case rates in these consistently high
counties averaged 15.8 campylobacter and 11.8 salmonella cases per 100,000
people, which is well above the Michigan average but is still lower than the
national average. One avenue for further exploration would involve the evaluation
of environmental factors known to be associated with the presence of these
bacteria. One such environmental factor is farming/agriculture where animal
shedding and slurry spreading has lead to known environmental bacterial

contamination (Jones, 2001). This factor along with water source (many Michigan

47

homes acquire their drinking water from private (often untested) wells (EPA,
2005)) and the spike in Michigan summer recreational activity (increasing
likelihood of environmental exposure) it is likely that some of these influences

may affect the incidence of these diseases in Michigan.

Geggraphic

There was variation in rates across the state as expected. Many of the
counties with high rates of infection show up as the top high infection rate
counties for both Campylobacter and Salmonella (Wexford, Missaukee,
Marquette, and Menominee Counties). These counties are grouped in two areas
of the state, which may suggest some geographic factors influencing the

variations in rates.

2.6 CONCLUSIONS

Many studies to date have relied on Food Net data to describe and identify
trends in campylobacteriosis and salmonellosis (CDC, 2005). However, these
data only represent a very small portion of the US population (5.4% in 1996 and
9.5% in 1999), the state samples are not representative of the entire state, and
the data are only available starting in 1996. Although the surveyed population is
similar it may not be generalizable to the entire US population (Hardnett et al.,
2004). With these data, extrapolations have been made to describe many trends,
particularly the decline in Campylobacter incidence (Samuel et al., 2004). This
decline in case rates from Fodd Net states has been noted from 1996-2003 and
has been attributed to improvements in the poultry industry (Buchanan and

Whiting, 1998, Allos et al., 2004, Samuel et al., 2004). This expected declining
48

temporal trend was not seen in Michigan, as there was no trend in annual rates.
There was also no change in rates after the HACCP rule implementations in
1996. This would suggest that the rates in Michigan may be less strongly
influenced by food routes. Michigan’s rates are also much lower than national
rates. This makes Michigan an important site to evaluate as all other
demographic factor trends are consistent with the national data.

Further studies would need to be done to evaluate the relation of
geographic and environmental factors in order to understand the geographic
variability in rates from county to county. These findings also suggest that there
may be drivers other than food that may explain the variation in rates of these
diseases. The following studies will further define the seasonal variation and

evaluate possible environmental drivers.

49

Table 2.1. Incidence (per 100,000 people) within counties by quartiles of

population density.

 

 

 

 

 

 

 

 

 

County Population Campylobacter Salmonella
Quartile
Density people/mi2 Incidence Incidence
>75%-ile >144 4.92 5.03
51-75%-ile 55 -144 6.93" 6.82*
26-50%-ile 34-55 5.18 5.59
s25%-ile <34 5.37 4.97

 

* This incidence was significantly higher than all other quartiles for the

aforementioned disease.

Population density quartiles were based on intervals of all data points. Density is

reported as the number of people per square mile for all counties and all years.

This table reports the mean human Campylobacter and Salmonella incidences of

counties within a certain a certain population density range.

50

 

Table 2.2.a. Campylobacter Incidence (per 100,000 people) by percent

distribution of county by race.

 

Percentage of County

County Racial groups: Campylobacter rates

 

 

 

Population
White Black Other
>90% 6.04* - -
81 -90% 4.76 - -
71-80% 1.82 - -
61 -70% - - -
51-60% 0.73 - -
41 -50% - 0.73 -
31-40% - - -
21-30% - 1.14 -
11-20% - 5.23 1.51
510% - 5.77 5.76*

 

 

 

 

*This incidence was significantly higher than all others in that racial group.

-There were no counties with populations with populations for this racial group

within this percentage range.

This table reports the mean human Campylobacter incidence of counties with a

certain percentage of its population composed of a particular race.

51

 

Table 2.2.b. Salmonella Incidence (per 100,000 people) by percent distribution of

county by race.

 

 

 

Percentage of County County Racial groups: Salmonella rates
Population White Black Other
>90% 5.72 - -
81-90% 5.20 - -
71-80% 4.86 - -
61-70% - - -
51-60% 0.81* - -
41-50% - 0.81 * _
31-40% - - -
21-30% - 5.56 -
11-20% - 5.86 4.61

310% - 5.58 5.58

 

 

 

 

 

 

* This incidence was significantly lower than all others in that racial group.
-There were no counties with populations for this racial group within this

percentage range.

This table reports the mean human Salmonella incidence of counties with a

certain percentage of its population composed of a particular race.

52

Table 2.3.a. Campylobacter Incidence (per 100,000 people) by percent

distribution of county by age group.

 

Percentage

County Age Groups: Campylobacter Incidence

 

of County

Population 0-4

5-9 10-14 15-19 20-29 30-39 40-49 50-59 260

 

 

>30% -
21-30% -

10-20% -

- - - 5.75 - - -

- 5.24 5.82 5.55 5.55 5.38

<10% 5.55 5.55 5.55 5.55 4.66 5.24 - -

 

4.80
4.23
6.38

6.14

 

-There were no counties with populations for this age group within this

percentage range.

This table reports the mean human Campylobacter incidence of counties with a

certain percentage of its population composed of a particular age group.

53

 

Table 2.3.b Salmonella Incidence (per 100,000 people) by percent distribution of

 

 

 

county by age group.
Percentage County Age Groups: Salmonella Incidence

of County

0-4 5-9 10-14 15-19 20-29 30-39 40-49 50-59 260

Population

>30% - - - - - - - - 4.80
21-30% - - - - 5.75 - - - 4.23
10-20% - - - 5.24 5.82 5.55 5.55 5.38 6.38
<10% 5.55 5.55 5.55 5.55 4.66 5.24 - - 6.14

 

 

 

 

-There were no counties with populations for this age group within this
percentage range.
This table reports the mean human Salmonella incidence of counties with a

certain percentage of its population composed of a particular age group.

54

Figure 2.1.a. Michigan Counties with high and low incidences of human
Campylobacter infections.

     

,. miss “I Mackinac
u. A“: ' 7’ u

 

Gogebic
Menominee \ Cheboygan
m Presque Isle
_ Leelanau W Montmorency
Missaukee ll mu mm W
Alacona
High Incidence I m 954' m '9 Arenac

 

 

Low Incidence safe

 

 

 

‘m m Saginaw

 

 

 

 

 

 

 

 

 

® cu:
Consistently High“ , W m 6909568
I Lapeer
c ‘st tl L * W _
am an y W Wexford , mm all» St. Clair
Lake on m m mm lg
Mecosta .
m m lawn Imam “w” .Shlawassee
Isabella 4 I W
. {A % mm m m Wayne
Berrien s, .
as m m ,m; Hillsdale

 

 

 

*Consistently High or Low counties are counties that frequently showed up in the

top or bottom quartile for annual mean county Campylobacter incidence.

High and low incidence counties had significantly higher or lower mean overall

Camp ylobacter incidences.

This figure illustrates the geographic relationships of high and low incidence

counties and the overlap between those and counties with consistently high and
low human Campylobacter incidence.

55

Figure 2.1b. Michigan Counties with high and low incidences of human
Salmonella infections.

Keweenaw

      
  

\e Cheboygan

Menominee Grand Travege “’9 Oscoda

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“n“ ' ' . Ogemaw
Benzie “m" w GI cl ~
' a Win
Wexford . 8‘3 IQ‘ Arenac
High Incidence - .. m. m Midland
Low Incidence
. . W W“ M Bay
Consrstently High” QB 41““ . Salinac
Lu .
Consistently Low" . , Tuscola
. a. W m Shiawassee
Missaukee m .m w (W mm .
+:~' St. Clair
MeCOSta/Fa WP Imam ”W m
Montclam mm mm mm Macomb
Eaton fl a g, ,0 Wayne
as n... 1....
Branch Lenawee

 

 

 

 

*Consistently High or Low counties are counties that frequently showed up in the
top or bottom quartile for annual mean county Salmonella incidence.

High and low incidence counties had significantly higher or lower mean overall
Salmonella incidences.

This figure illustrates the geographic relationships of high and low incidence
counties and the overlap between those and counties with consistently high and
low human Salmonella incidence.

56

458-39 :5: Sam .6 cocoa 05 BB

.288 coed? Lou 595:2 5 380 33282me new LosomooSQEmo 8:82 So 8:022: .355 m5 9505 93m: wEH

 

Lao>

moom voom NOON ooom mom? mag 32 mag ommv

mfl \ .l/ > M.
\ <\ \7I/D/

 

 

 

m6

e I» \ / M.
we 2 /. .

 

OOO‘OOI Jed eoueprouj

 

m6

 

 

cocoEoE m__ocoE_mw ts... cocoEoc. LoyomnsEEmoI

 

 

 

.moommmmr .5955 E 22.0er m=ocoEEw new LouomnoaaEmo 96 805205 .2.ch «d p.59".

57

563% m5 *0 cocoa
or: .06 8590:. 36:9:me ocm LoeoonRQEmo cmEac cmoE >_£coE $952.2 Co motom mE: m5 9505 93m: mEh

Eco—2 28>

v9-5... modem 5A6: 3.30 8-94 3-3—2 3462 5.3% «5-90... 5-50
mo. 0

mmd
8:865
mmd

 

m__o:oE_mw IT cosmootfiEmo T

moow-mmmF :mmEoS E m__mcoE_mm can Loaomnoinmo V6 30520:. 25:22 MN 959".

58

.mcozoohE goggmm ucm :QomnoEQEmO :35: 59:25. Go oo:ou_o:_ 25:9: :moE 9: 2:95 9:9: wEh

:oQEowqom. b:m .3393: €qu :9: ..eo:oEo:.: :cBmQOEQEmO :0: $38 =m 5:139: A::mo.:.Em..m

mSo: mm $.23: 33. b:m Am: .mo:ob.5:.: m=m:o§mm use :ohomeoSQEmO :0: 6:9: AEmoSah m: 5:: :0 5:8: och

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.3:
O 1 so.V
o o a v»
o 4 O 1 b Os en
039 co co as so A on s 94 «o co
.99 s9 so 599 2% % so ea 4.0 e. c... as
.2, m w. o
m__o:oE_wws .. .. m... Nd m.
LosomnoSQEmol , . m.o m
. . i.
.o
T ...., md W
.a.... - 0
W
m. up
No 8
ad

 

 

omflozw 25:22 3 00:028. EBBEEW ucm 6830:3950 YN p.59“.

59

Figure 2.5. Average Campylobacter and Salmonella Incidence by Season

 

0.7
0.6
0.5
0.4
0.3
0.2

0.1

Incidence per 100,000 people

0

 

 

 

 

 

 

Spnng

 

Summer

Season

Fall

 

Winter

I Campylobacter
sir; salmonella

 

 

The summer is significantly higher for both Campylobacter and Salmonella

seasonal mean incidence.

This figure graphs the mean seasonal incidence of Michigan human

Campylobacter and Salmonella infections.

60

 

.965 can 3 22625 w:o:o§mm ecsfiomeoEQEmO :mEsc 59:25. :0 85205 .msccm :moE or:
292m 2%: 2:: $8538.: 3950: :EmotEQm o£ $8.: 330% 2-3 b:m .14: .a-: 3: 8:93.63 :oBmQOSQEmO
:ou: .oo:ob.6:.: :meE 33:02.58. use :oBonSQEmO 59:05 :8: :0: 6:9: :::mo.:.::m:m w: 35% cum :8» To oi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2: So om<
mmn mmém med: mmdm mmém m 79 w P -o F m-m v-0
1 a
- m
l ”H m
D.
- m m.
m
r w .m
- E .m
o
m
- NF
3
or
a 8:865 Seesaw. 88:62 8:302:88-

 

 

 

 

 

qzem om< E 8520:. m:o:o§mm. b:m :EQonXQEmO 6N 2:2".

61

Figure 2.7a Campylobacter Incidence by Gender

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1990 1992 1994 19% 1998 2(ID 2002 2W4 2%

This figure graphs the mean annual incidence of Michigan human Campylobacter
infections by gender.

Figure 2.7b Campylobacter and Salmonella Incidence by Race

 

 

 

 

 

 

 

 

 

 

199) 1992 1994 19% 19% 2M) 2(132 2W 2%

This figure graphs the mean annual incidence of Michigan human Salmonella
infections by gender.

62

Figure 2.8 Campylobacter and Salmonella Incidence by race.

 

 

 

 

I Campylobacter
Salmonella

Incidence per 100,000 people

 

 

White Black Other

Race

 

 

 

Blacks were significantly lower than Whites and Other races for both
Campylobacter and Salmonella infections.

Data collected from the Michigan Department of Community Health reported the
race as Black or African American, White, Asian, American Indian/Alaskan
Native, or Multiracial. These groups were collapsed into three groups Black
which included Black/African American reports, White which included White
reports, and other which included Asian, American Indian/Alaskan Native, and
Multiracial.

This figure graphs the mean annual incidence of Michigan human Campylobacter

and Salmonella infections by race.

63

CHAPTER 3
Campylobacter and Salmonella Infections in Michigan:
Evaluation of Seasonal and Geographic Trends Reporting (1992-2005)

3.0 STRUCTURED ABSTRACT:

Background - Campylobacteriosis and Salmonellosis are common
gastrointestinal infectious diseases in the US and world-wide. Given the high
incidence and health burdens of these diseases, much research has been done
and trends in incidence of these diseases (seasonal and geographic) have been
reported, but not fully explained.

Specific Aims - To statistically evaluate the seasonal trend in case reporting in
Michigan, identifying parameters peak-week, start and end-week, and duration of
the high reporting period. This study also evaluates the relation of geographic
location and county incidence to high reporting period parameters.

Design - Retrospective analysis

Methods - Historical case data on Campylobacter and Salmonella infections in
Michigan (1992-2005) were evaluated, time series techniques applied to identify
the parameters, and linear modeling techniques were used to evaluate
relationships of geographic location and county incidence to disease reporting
trends.

Results - Time series shows regular behavior with high reporting for both
campylobacteriosis and salmonellosis in the summer and early fall. This period
varies between counties with respect to all parameters. Duration of the high

reporting period was the only variable consistently associated with geographic

64

location and incidence.

Significance — Given the previously reported associations of these diseases with
temperature and the changing global climate (global warming) there is increasing
urgency to fully understand the seasonal trend of these diseases. By
geographically defining the parameters of seasonality, more detailed models can
be created incorporating variables for which geography may be serving as a
proxy, and uncover the drivers of campylobacteriosis and salmonellosis

seasonality.
3.1 INTRODUCTION

Rationale:

Disease seasonality is the systematic recurrence of a compact cluster of
cases followed by a long interval of low incidence forming a typical pattern for a
specific pathogen in a given population in a given locality (Naumova, 2006).
Seasonality is characterized by 1) a point in time when the seasonal curve
reaches its maximum, 2) the amplitude from peak to nadir, and 3) the duration of
the increase defined by a shape of a curve (Naumova, 2006).
Campylobacteriosis and Salmonellosis are known to exhibit seasonality with
cases peaking in the summer months. This seasonality has been evaluated but
the causes remain unclear. Further investigation into the shapes and parameters
of the seasonal patterns and possible associates of those parameters could
inform future research into drivers of the diseases and whether there seasonal
variations in host susceptibility, pathogen survival and transmissibility, or

environmental load.

65

This study evaluates the high reporting period (seasonality) for these
diseases, identifying the critical weeks (start and end week, peak weak, and
duration) and their associations with incidence level and geographic location.
This study highlights time frames and patterns for use in further studies

identifying environmental factors that could be driving these trends.

Background:

Campylobacteriosis and Salmonellosis are common infectious diseases
caused by infection with Campylobacter spp. and Salmonella spp. and are
primarily associated with foodbome routes of infection. Currently, Campylobacter
is the most commonly reported cause of acute bacterial gastroenteritis in
developed countries (Mead et al., 1999) and non-typhoid Salmonella annually
causes 1.4 million cases annually in the United States (CDC 2004). The health
burden for these diseases is great, as there are sequales to infection for
Campylobacter, Guillian-Barré syndrome (Kalra et al., 2009, Vucic et al., 2009),
and for Salmonella, reactive arthritis (Girschick et al., 2008). In vulnerable
populations such as the immune compromised and the elderly, death can occur
from infection. There are commonly reported demographic, geographic, and
temporal trends and given the high prevalence and health burdens for these
diseases much research has been done examining the foodbome route to
understand the trends in transmission and to minimize infection. The reported
temporal trend consists of two parts, the overall declining trend that has been
studied and attributed to improvements in the meat processing industry (Samuel

et al., 2004), and the seasonal trend which is not fully explained.

66

Many factors contribute to the human incidence of Campylobacter and
Salmonella infections, and there is growing evidence that temporal and climate
factors are associated with incidence of these diseases (Louis et al., 2006,
Jepson et al., 2009, Naumova 2006). In efforts to understand seasonality, studies
conducted, using various methods and data from developed countries all around
the world, have assessed the relationships with ambient temperature and
changing incidence relationships. These studies consistently report seasonal
increases in incidence of these diseases exhibit a lag relationship with increases
in temperature (Zhang et al., 2006, Patrick et al., 2004). The lag in these studies
ranged from 2 days to 5 weeks. The authors reporting longer lag times suggest
that factors (food prep and handling) close to the time of the reported infection
may not be the most important step in transmission (D’Souza et al., 2004) and
other routes should be explored. Authors report shorter lags suggest that food
handling could be the cause (Kovats et al., 2004). Given the varying results,
methods, and locations of these studies, further investigation is needed to tease
out the consistencies and differences in the seasonal trends.

The study by Lindback and Svensson on Campylobacter infections in
Sweden attempted to define seasonality for reported cases (2001). This study
found, there was variation in the high reporting period with respect to the start-
week and peak-week between counties and that there was a relationship
between the high reporting period parameters and geography (north/south
position) such that more southern counties had earlier start and an earlier peak

than in northern counties. This study highlighted the regularities and variations in

67

reporting patterns between geographic areas that make unique comparisons
possible.

Michigan Department of Community Health (MDCH) receives reports on
culture confirmed human cases of Campylobacter and Salmonella infection in
Michigan and has data archived from 1991 through the present. Previous studies
in Michigan have reported a seasonal peak and geographic variation that have
not been explained (Arshad et al., 2007, Younus et al., 2007). Using some of the
Lindback and Svensson methods, our study evaluated the seasonal peak in
reporting of Campylobacter and Salmonella infections in Michigan with respect to
peak-week, start and end-weeks, and duration. Further, relationships between
the high reporting period parameters, geographic location, and county incidence
level were evaluated. It is expected that results of such analysis will be used in
modeling the high reporting period and identifying critical time points for future
analysis in conjunction with environmental factors that possibly drive these

disease rates.

3.2 HYPOTHESIS
The specific hypotheses tested in the study were:
1) there will be variation in the seasonal peak parameters (start week, end week,
peak week, and duration) with respect to geographic location; and that
2) there will be variation in the seasonal peak parameters with respect to

incidence.

68

3.3 METHODS:

a. Case Data

Reported cases of Campylobacter and Salmonella infections in Michigan
were collected from the Michigan Department of Community Health (MDCH).
These data were collected as a part of a mandatory reporting system for
communicable diseases of interest in Michigan. Cases are reported to local
health departments by doctors, hospitals, and laboratories once the bacteria
have been culture confirmed. To protect the identity of individuals the data were
requested with limited identifiers. These data were collected with the identifiers of
state, county, onset date, race, gender, and age, but for this study the relevant
identifiers are state, county, and onset date. Cases that did 1) not report the state
as Michigan, 2) not report a county in Michigan, or 3) did not report a date in the
range of 1/1/1992 to 12I31l2005 were excluded. Of the 6148 Campylobacter and
6508 Salmonella infections reported, 6113 cases (99%) and 6041 cases (93%)
respectively, were used in this study. The cases were sorted and counted by
year, month, and week based on the onset date provided for the aggregated

state and county data.

b. Evaluation of Specific Aims
Aim 1: Evaluate high reporting period

Data Management: SAS 9.1 software was used to manage the data for
calculations. All calculations were made for the Campylobacter and Salmonella
data sets. The weekly average over the 1992-2005 period for each week (1-52)

was calculated. The sum of reported cases in the ith week (i is the designation for

69

one week in the set of 1-52 weeks) of each year is averaged yielding the ith week
average. The overall weekly average was calculated; all reported cases are
summed and averaged over the weeks (reported cases/728 weeks). A 9-week
moving average was also calculated for the state and counties (centered on the
mid-week).

Identify Dependent Variables: All variables were identified for
Campylobacter and Salmonella data sets on the county and state levels. Peak-
week (continuous variable) was determined based on the week with the highest
associated 9-week moving average. The high reporting period is defined as the
period when reporting, defined by 9-week moving average, is higher than the
overall weekly average, and that period contains the peak-week. The start-week
(continuous) of the high reporting period is the first week (in a period containing
the peak-week) where the 9-week moving average is above the overall weekly
average value. The end-week (continuous) is the last week in that period where
the 9-week moving average is above the overall weekly average value. The
duration (continuous) is the range of weeks between the start and end-week

(difference, start-week from end-week).

Aim 2: Evaluate geographic relation to seasonal high reporting period

 

Data Collection gnd Management of Independent Vfiariables: Data were
collected on county latitude and longitude measures (continuous). Counties
were ranked and ordered 1-83 based on the latitude and longitude, for the
variables lat rank (ordinal) and long rank (ordinal). Counties were also grouped

based on latitude and longitude quartile. The ranges of latitude and longitude

70

values were 41.89-47.42 and 82.59-89.80. These ranges were divided into
fourths to group the counties into four geographic lat regions (ordinal) and long
regions (ordinal). Data were also collected on Michigan climate zones and
counties were grouped based on the climate zone (categorical).
Aim 3: E_wa_l¢_rate incidence in relation to seasonalhigh reporting period

Data collection and Management of Ind_ewent Variaples: Population
data estimates were collected from the US Census Bureau on Michigan and
Michigan counties from 1992-2005. This data along with MDCH case data for
Campylobacter and Salmonella infections in Michigan was used to calculate the
state and county annual incidence for each disease, for variable incidence
(continuous). The counties were divided based into quartiles (defined by 25%

groups of the highest annual incidence), for the variable incidence rank (ordinal).

c. Statistical Analysis

General linear modeling (proc Glm code in SAS) procedures were used to
evaluate the dependent variables peak-week, start-week, end-week, and
duration with respect to the independent variables latitude, longitude, lat rank,
long rank, lat region, long region, climate zone, and incidence rank. Models were
constructed for each dependent variable with each individual independent
variable and with combinations of non-correlated independent variables.

Statistical significance was determined at the p<0.05 level.

71

3.4 RESULTS
Weekly reporting of Campylobacter and Saimonella Infections in Michigfl

The time trends of reported cases of Campylobacter and Salmonella
infections in Michigan by week are shown in Figures 3.1a and 3.1b. The
seasonal pattern seen here is also present in the county level data (Table 3.1).
Peaks and durations of these seasonal fluctuations vary between counties.
Temporal Distribution

The weekly average (grey), the overall weekly average (blue horizontal
line), and the 9 week moving average (black) for Campylobacter and Salmonella
infections in Michigan are shown in Figures 3.2a and 3.2b. The area above the
overall weekly average and below the 9 week moving average represents the
model (definition) of the high reporting period. These calculations were made for
Michigan and its counties (Table 3.1).
L-l_igh Rflorting Period Parameters

The parameters for the high reporting period, peak-week, start-week, end-
week, and duration for the aggregate Michigan case report data were calculated
and are reported in Table 3.2. The parameters are similar for the Campylobacter
and Salmonella high reporting periods with the Salmonella period starting three
weeks earlier in the year than for Campylobacter and lasting four weeks longer.
These measures were calculated for Michigan and its 83 counties and the values
varied by county (Table 3.1). Of counties with annual incidence over 1 per
100,000 people, the Campylobacter peak-week ranged from week 8 (Luce

county) to week 47 (Alger), start-week 8-43 (Luce, Alger), end-week 11-52 (Luce,

72

Alger), and duration 0-28 (Mecosta, Wexford) for Salmonella the peak-week 8-48
(Ogemaw, Marquette), start-week 8-43 (Ogemaw, Arenac), end-week 14-52
(Ogemaw, Allegan), duration 6-28 (Ogemaw, Marquette).
Relation of Geoggaphv and Inwnce to High ReportingPeriod Parameters

The duration of the period consistently significantly (p<0.05) correlated
with all evaluation variables for Campylobacter incidence rank and all except
longitude derived variables for Salmonella infections. The parameter end-week
was associated with climate zone and incidence for Campylobacter infections
and lat region and incidence for Salmonella. Parameter peak-week was
associated with lat region and incidence for Salmonella. Parameter start-week

was associated with lat region*long region.

3.5 DISCUSSION

Time series shows regular behavior with high reporting for both
campylobacteriosis and salmonellosis in the summer and early fall. This period
varies between counties with respect to all parameters. Duration of the high
reporting period for campylobacteriosis and for salmonellosis was the only
variable consistently associated with geographic location and incidence. For
Campylobacter reporting, the longer periods of high reporting were associated
with more southern and more eastern geographic locations and Salmonella’s
longer reporting period (duration) showed associations with more southern
locations. The Lindback and Svensson (2001) Campylobacter study found
geographic variation (north south) related to in the start and peak weeks such

that the more southern counties had earlier start and an earlier peak than

73

northern counties. Our study showed similar results. This north to south
association could be an artifact of temperature zones and the known temperature
disease relationship following the assumption that southern regions of Michigan
would be warmer longer; however, the climate zones did not show any relation to
the high reporting period parameters. Also, the eastern trend seen with
Campylobacteriosis reporting may be related to the higher population density in
the eastern portion of the state.

Our study went further to evaluate the relationships between the high
reporting period parameters and county incidence level. We found that the
incidence variable was associated with all of the high reporting period
parameters at the county level (start and end week, duration, and peak week). As
incidence increased the reporting parameters were affected as such: 1) start
week was earlier, 2) and week was later, 3) duration was longer, and 4) peak
week was earlier. These findings have not been seen in the literature and require
further investigation.

This evaluation is an essential step toward modeling the high reporting
period and identifying critical time points for future analysis in conjunction with
environmental factors and seasonal changes in human behavior that possibly
drive these disease rates. This study adds to the literature giving more
information on the previously reported, but not fully explained, seasonal peaks in
reported cases of these diseases, and moves toward offering a greater
understanding of these peaks. The larger aim of this study is to eventually lead to

the reduction of incidence and health burdens through understanding the

74

mechanisms of transmission that can prevent future Campylobacter and

Salmonella infections in humans.

3.6 CONCLUSIONS

This study takes a step toward understanding the seasonal trends of
campylobacteriosis and salmonellosis through defining the parameters of
seasonality by geographic location. This is significant due to the relationship of
geography to environmental factors (climate and weather, land use, water
source). By understanding how the parameters of seasonality vary we can begin
to explain and relate that variability to other seasonal variations. These may
have discernable relationships with a specific start or end time, peak week, or
duration that may suggest a link with host susceptibility, pathogen survival and
transmissibility, or environmental load as drivers for that parameter of
seasonality. Future research is needed to evaluate the differences that these
geographic locations represent and modeling those differences to identify the

drivers of seasonality.

75

Figure 3.1.a. Weekly reported cases of Human Campylobacter cases in

 

 

Michigan
40
35
30
25
(I,
8
m 20
O
15
1o
5
o
\‘bfoACbN'bfefloJ'x‘bbAQi
ekexerfirtg’e‘ebvag’e'ké‘emét‘I’
Week

 

 

 

This figure shows the weekly number of Campylobacter cases reported in

Michigan, over the period of study from 1992-2005.

76

Figure 3.1.b. Weekly reported cases of Human Salmonella cases in Michigan

 

‘ . :I 'i ‘j ‘ r » I
f ;l v. w. j
. , Ir . . ,. ,
.' i V I J ‘l' , *‘i-v---O-
' i I . 4 -
y r . .. g .
I :1 '. V : ' '
, t I I 0
j t
‘. : II I
- T 1..in V
‘ .I . , . - 7; -
. ._ r n I . '1 _
‘i _ . . '. _|, > ‘
.- 1 , .J ‘ .. . . ‘
.1. - . . ;; : . '-l '3 .-.‘ . t ‘ I , . t
v ‘. I l1 f "v , v F _x , 7 raffle-d." , A . ; ~"\_- ~. r~~~—---
“ ‘ I: - .‘r ‘- :' '.' I i . ‘2 '-' ~i . . . = ' . 'z "i
2. r ' .'- ‘. ‘- ;.; ~14 4'3 .. . ‘ -‘ Z ‘ r ‘. ‘ '
_- ~ ‘ -, . 2‘ . a.- . .' | | I . ‘ .
I i . . - , v_ . .. . , v . , . 4 5 r

 

 

 

 

. -nr . -—-..p___._.. .. . .

 

Cases

 

f “a...“ _ ,ruu ....- _..~.— ——.~- .~.

. . v n...
. . .... .
«nu—n.

 

 

 

r
.. it I I ’.
.‘ . {. j 7; _ , ‘ ' I _ ' .4 I
I l .‘ I I ’1 E I _ ‘
I 2i . ' ‘ I r
5 3 V ‘ f s . . .I . , ‘ ' r l‘ . . 1 . , ‘ . 1 , -
—}. -~-—.-._—4.-~ 9M - v: .____... :*-—~. - -.. . A i——— .4-.. . .w— ,.. . 4..." -.~.-.|,.g..,-.4~.o.s.‘,- ._...-_,_. ~....-....v.. ...-- ....-
I “' ' ‘ I ' '1 I . ' I - I ‘ “‘ I ‘ ‘ . I I " '~ i ' . .-
I I i. j I I I

l "F
lt-~' ll

 

 

0

r
O N
L0

468 ~ :
520 -
572 —
624 -
676 +
728 ~

Week

 

 

 

 

This figure shows the weekly number of Salmonella cases reported in Michigan,

over the period of study from 1992-2005.

77

Figure 3.2.a. Weekly averages for Campylobacter cases in Michigan.

20

 

 

 

 

 

 

 

 

 

 

 

 

m 1: +meancase m

g 14 f )<V\

8 12 fl / \

2 10 / W \

8

E 8

2 6
4- -
2 \
O: I r l r I . . 1 r 1 I 1 .

0 4 8 12 16 20 24 28 32 36 40 44 48 52

Week

This figure shows the average number of weekly Campylobacter cases reported
in Michigan, over the period of study from 1992-2005. The horizontal line
represents the overall average and the smoothed line represents a nine week

moving average trend line.

78

Figure 3.2.b. Weekly averages for Salmonella cases in Michigan.

20
18
16
14
12
10

+ meancase

—9 . Mov.

Number of Cases

 

omhcnoo

0 4 81216202428323640444852
Week

This figure shows the average number of weekly Salmonella cases reported in
Michigan, over the period of study from 1992-2005. The horizontal line
represents the overall average and the smoothed line represents a nine week

moving average trend line.

79

 

 

 

NO.¢O O0.0v
NOOO ON.¢¢
O0.00 OYOV
NYVO O0.0v
ON.OO POOV
FO.¢O OYOV
NO.NO NOOv
NONO N0.0v
O0.00 O0.0v
VOéO OO.¢¢
NO.VO OONv
OO.¢O N0.0v
OO.VO N0.0¢
_.O.VO OYOV
ONOO ONO»~
NOOO NO. ;
O0.00 ON.Nv
O0.00 OO. 3
OvOO NO. 3
VOOO VOé.‘
N0.00 VOOv
NOOO FONV
OYOO FNOv
N0.00 Ooév
OfiOO FOOV
VOOO O0.0v
O0.00 OONv
OOOO O0.0¢
O0.00 OOéV

85 25: 8.. 8223

905:0

OYN

NO.N_.
00:00_0:_

 

 

a
3:300 00:05:... 0:05
3:300 050O00
3:300 5300.0
3:300 0000:00
3:300 30EEm
3:300 :030m
3:300 :00:_v_0_o
3:300 0:00
3:300 05:50.0
3:300 :9:__0
3:300 0.0.0
3:300 03000_:0
3:300 :0ON000:0
3:300 x_0>0:0:0
3:300 0000
3:300 :30500

 

3:300 5:05
3:300 :0E0m.
3:300 0_N:0m
3:300 3mm.
3:300 3:0O
3:300 0O0:0m,
3:300 00:03
3:300 EEc<
3:300 0:003.
3:300 :0O0__<,
3:300 :0O_<
3:300 0:0o_<

 

ON Nv NN PO VOO 3 3 NN NO
3 ; NN VO OOe O VN O F O _.
O Ov NO NO OOO O OO NN NN
v O r O or O0.0 O O _. F _. O r
NN Ov ON 0O OYO Ow Ov 0N OO
NF Ov ON NO N0.0 _‘N Nv _‘N ON
ON Ov 0N OO NN.v : NO ON NO
OF Ov OO Ov NN.¢ ON ON 9. O_.
3 NO ON ON Of: OF O». «N NO
O ON VN VN NO 3 NO ON OO
NN NV ON OO ON.N OF Ov NN OO
ON OO OF NO VOO 5 OO NN OO
ON O... #N 3 NN.N ON OO OO OO
O F O OO Ov ON.O O NN O F O F
Or 3 ON NO NN.O O 00 NO NO
_‘N Nv ON OO OO.N N—. 00 ON VO
O_. OO ON ON O0.0 ON O0 ON OO
Ow Ov ON OO 0O. 2‘ O 3. OO OO
NF 3 VN VO OOO NN Ov OP @O
O—. OO ON O OO.: O_‘ Ov NN OO
Or 3 ON OO NN.N Or 3 ON ON
VN OO ON OO VNO O_‘ OO ON OO
0 NO O0 O0 ONO NV Ov OO Ov
OF NO OF ON ONN O O O? Nv
3. NO ON ON ONO ON vv OF .N
OF Ov NN ON VNN ON NO ON OO
O_‘ Ov VN _.O ON.O 3 NO OO Ov
O NO Ov Nv ONO O OO OO OO
O NO NN ON OVO N O». VN OO
£035.50 x00; x0¢>> x062, QUCQEOC— £059.30 x003 x003 xmm>>

0:0 020 x000

0:0 0030 x000

 

0.00 of 0: 000

 

03000032200

 

03002500

3:300

 

 

 

0300 03305000 0:0 2030.00.00 00:00 @5080”. :9: 00::300 59:25. ._..O 0.03

O

8

 

 

 

 

 

0.00 0200:0000

 

00.030.300.00

000008.00

 

 

3 000 5.0.. Nvé NN Nv 0N NN EN 0 N 0. N0 3:300 C508:0...
0? 00.00 N0. 3 00.0 0? 00 N 0 00.0 0N F... N F0 3:300 09:05.
0F 0000 00.3 0N.: 00 00 ON 00 00.0.. 0 v0 0N 0N 3:300 0003000....
0F 00.00 00.0.. N; 0N Nv NN 00 00.0 N. 00 NN 0.0 3:300 0:0.0..2
0F N0.N0 00.00 N0.NN N 3 0N ..0 N0.NF NF 00 0N 00 3:300 0055805.
0? 00.00 00.03. 00.? 0 0N 0N 0N 0N.? 3 00 00 :V 3:300 0.00005.
0F 00.00 00.0.. N04 5 00 NN 0 «0.0 NF 00 F0 00 3:300 :0005.
OF F0.N0 00.0.0 NN.0F 0N Nv NN 0N VN.0F 0N N0 .N 00 3:300 00030.0...
0F v.00 N000~ 00... 0 N0 0N 0N NN... FN N0 «0 N0 3:300 00.0.:05.
.1. 00.N0 00.Nv 00.0 0? 0.0 NN N0 NON 0? N0 0N v0 3:300 08000.2
0? 00.00 00.0... 00.0 0 0N N N 00.0 0 NN E. 3 3:300 00:200.).
0.. 00.00 0.200~ 00; 0 E. 0 0 N: 0 00 F0 F0 3:300 003..
1 .000 N0.Nv 0N.N 0F 00 0N N0 NN.0 0N 0v 0N 00 3:300 5.00:3...
3 3.00 N0. 3 0.3 0? 0w 0N 00 0N.0.. 0N 3 N 00 3:300 00305..
0? NN.00 00.3 00.0? 0F 00 NN 00 00.0 0N Nv NN NN 3:300 30:0.00..
0.. 0N.00 N0.0v 00.0 0N 00 0N 00 N06 N 0.0 0N 00 3:300 0000..
0? 00.00 00.00 00.0 0 0 0 0 00.0 0 3 00 00 3:300 00.0..
NF «N00 00.Nv 0N.N 0 v0 0N 0N 0N.N 0 N0 00 V0 3:300 30:00.53.
3 90.00 00.N... 00.0 0N 00 NN 0N 3.0 0N 0v NN 00 3:300 E0v.
0F N00 {.3 0N. S 0N F0 F0 0.. N00 3 00 0N 00 3:300 00.0053.
0.. 00.00 0N.Nv 00.N NP Z. .N N0 00.0 .N 00 FN 0N 3:300 03050.3.
3 00.00 VNNV 00.0 0N N0 0? N0 00.0 E 0N 0F NN 3:300 500.00...
0F 00.00 00.0.. 90.0? ON 0 0N ..0 00.N 0? 00 NN 00 3:300 0.0000.
NF 00.00 3.00 00.N 0 00 0N N0 0N.? NF 0N 0.. NF 3:300 :2.
0? N000 00.3 N00 NN 00 0 NN 000 OF 00 ON N0 3:300 0000.
3 00.00 00.N... 0N... I. 00 ON N0 00.0 N? 0v 0 00 3:300 0.:0.
3 N30 N0.Nv 00.... .N 00 .N _.0 0N0 0F 3. 0N N0 3:300 £9.05
3 3.00 00.00 00.0 E 00 00 N0 N0.N 0 NN 0' FN 3:300 :23:
N.. 5.00 00.2. 00.0 0? 00 0N 00 N00 NN 00 ON 00 3:300 52030...
3 00.00 00. S. 0N.NF NN 00 0N ..0 00.0 NN Nv 0N 0N 3:300 0000......
0:0N 002.30.. 003.00.. 00:00.05 :00050 0.003 0.003 0.003 00:00.05 :00050 0.003 0.003 0.003
0.0:...0 0:0 000 0.000 0:0 000 0.000 3:300

 

 

3035.80 00 033

81

 

 

0. 00.00 00... 00.0. 0N 0. .N .N N0... NN 0. 0N N0 0:80 2282.
0. 8.00 .0.N. 0.... 0. .. 0N 00 .0... m. .. 0N 00 0:80 8.02.
0. 0.00 0N.N. 00.0 NN 0. .N N0 00.0 0N 0. .N .0 0:80 382:0...)
.. .000 0N.N. .00 .. 00 0N .0 00.0 0. 00 0N 00 0:80 850 :0>
.. 0.00 0.0.. 00.. 0. 0. 00 .. 8.0 0. 00 .. NN 0:80 0.80:.
.. 00.00 8. .. N00 0. 0. .N N0 0N.. .N N. .N 0N 0:80 808.. .0
.. 8N0 SN. 0... 3N 0. 0N :0 00.: .. 0. N0 0. 0:80 8.0 .0
.. 0.... .0N. 8.0 0. N0 .0 0. 00.0 N. .0 00 .. 0:80 80008.5
0. .000 ...0. 0.0 0 0N 0. 0. 00.: 0 0N 0. 0. 0:80 8.2880
.. 0..N0 00.0. 00.0 0. 00 .N .0 0 ... .. 0. .0 o. 0:80 8.80
.. 8.0 00.0. ...N .N 0. .N 00 0... 0 00 0N .0 0:80 30:80
0. 00.0 00... 00.N ON 0. 0N :0 No.0 0. .. 0N 00 0:80 855803.
0. 00.00 .00. 00.: 0. N0 0. .N 8.. 0 00 :0 :0 0:80 20. 0809.
.. 00.00 00.N. 0N. .. 0N N. NN 0N 0.0. 0N o. .. .0 0:80 08:0
0. 00.0 .00. ...N .. N0 .N .N N00 0N .. 0. .N 0:80 800.0
0. .0... 00... ...0 0. .. N0 0. 00.. 0 N0 .N 0N 0:80 8800
0. 00.00 00.0. SN 0 30 NN NN .00 0 .0 0N .0 0:80 8800
.. 0N8 00.0. .mN .N .0 .N .0 00.. 0 00 :0 :0 0:80 88:2:0
0. 0 0.0 00... .... N. N0 0. .. 8.. 0 .. 0 0 0:80 8800
0. 0N8 .00. 0..N 0N 00 0. 0N 8.. .N 0. 0. .. 0:80 0:080
.. .000 00.N. 0.0 .N 0. NN N... 0.0 0N 0. ON N0 0:80 0:00.00
0. .000 00.0. .0. 0. 00 ON 0N 8... 0 00 0N .0 0:80 8882
.. NN.00 0N0. N00 .N .. 0N .0 N0. 0. N. 0N 0N 2:80 809.022
0. 0 ...0 No.0. 0.0 0 .0 0N 0N .0. .N 0. .N 00 0:80 8:99:85.
0:0N mvatgi. 00:50.. 00:00.05 00:05.0 x002. 0.00.... x00? 00:00.05 02.0.50 0.00.... x003 x003
0.0:...0 0:0 00.0 .000 0:0 00.0 0.000 3:300

 

 

0.00 0.5.0.300

 

00.0000003000

 

000005.00

 

82

 

 

88:80 .0 0.8.

Table 3.2. Michigan High Reporting Period Parameters

 

 

 

 

 

 

 

 

Start Week Peak Week End Week Duration
Campylobacter 23 32 44 21
Salmonella 20 33 45 25

 

This table reports the high reporting period parameters for the state of Michigan

over the course of the study for Campylobacter and Salmonella reported human

infections.

83

 

CHAPTER 4

Environmental Factors Influencing Rates of Campylobacter and Salmonella

Infections in Michigan

4.0 STRUCTURED ABSTRACT

Objective- To evaluate Campylobacter and Salmonella infection data collected
in Michigan from 1992-2005 in conjunction with environmental factors across
various scales from county level, to watershed, to climate division.

Specific Aims- To evaluate the role of environmental and climatological factors
in changes in incidences of human Campylobacter and Salmonella infections in
Michigan.

Design- Retrospective Analysis

Methods- Data were analyzed on multiple scales including county for localized
effects due to land-use and water resources, watersheds, and by climate division
for climate related variables. Statistical analyses included mixed methods to
account for temporal relationships in the data.

Results- Counties with large percentages (greater than 40%) of their land in
agricultural production were significantly associated with higher rates of
campylobacteriosis. Areas with mid ranges 21-40% in agricultural production
were associated with the lowest levels of campylobacteriosis and salmonellosis.
While source of potable water and sewage disposal method both were significant
factors in predicting rates in counties, they explained small amounts of the

observed variability in both Campylobacter and Salmonella infection rates.

84

Significant differences in rates were seen between the climate divisions with the
highest rates in division 3 (north west lower peninsula).

Conclusion- Time (year and month) and daily maximum temperature were the
best predictors of human Campylobacter and Salmonella incidence by county
across the climate divisions. These types of models may help to explain excess
case rates in high rate counties which do not fit demographic trends, as

described earlier.

4.1 INTRODUCTION

The gram-negative bacteria, Campylobacter and Salmonella have been
recognized as leading causes of diarrheal illness in the United States and
worldwide (USDA, 2003, Altekruse et al., 1999, CDC, 2005). These bacteria are
commonly associated with poultry, and other livestock, and consumption of
poultry is a major risk factor for both campylobacteriosis and salmonellosis
(Kapperud et al., 1992, Wysok and Uradzinski, 2009). Campylobacter and
Salmonella have also been transmitted by the water route (Kussin et al., 2005,
Kirian et al., 2008), which may be related to contamination from animal reservoirs
(including poultry and cattle) especially in rural areas and human behavior.

In US an estimated 1% of the population is infected annually with
campylobacteriosis, with an average 12.7 reported cases per 100,000 people per
year due to underreporting (CDC, 2006). For salmonellosis, an average of 14.6
cases per 100,000 people is reported annually (CDC, 2006). In many regions of
the world, cases for both these diseases are on the rise (Coker et al., 2002). In

the US, case rates of campylobacteriosis have declined since the beginning of

85

coordinated surveillance in 1996 (Samuel et al., 2004); however, distinct
seasonal trends continue to cause a high burden of disease in summer months.
Rates for salmonellosis have only decreased slightly (CDC 2005). Given the high
prevalence of campylobacteriosis (Oberhelman and Taylor, 2000, Lindback and
Svensson, 2001,Coker et al., 2002, Samuel et al., 2002) and the potential for the
infection to lead to more serious illnesses (Guillian-Barré Syndrome (1 in 1,000
cases) (CDC 2004), which costs the United States up to 1.8 billion dollars
annually (Buzby et al., 1997), it is important to identify and understand factors
that influence the incidence of the disease both spatially and temporally (between
years and seasons). This is also important for salmonellosis as it too has great
impact on health with the possibility of leading to reactive arthritis or Reiter’s
syndrome and an economic impact. A recent study in Spain calculated the
economic burden of salmonellosis related hospitalization to find an average
annual cost of 12.4 million Euros (Gil Prieto et al., 2009).

In analyses of campylobacteriosis and salmonellosis in developed nations
(e.g., US and UK.) consistent demographic trends in disease incidence have
been observed, most notably a peak in cases among young children (<5 years
old) and males (CDC, 2005, Samuel et al., 2004, Louis et al., 2006). Additionally,
a distinct seasonal pattern with cases peaking in the summer months has been
noted world-wide. While this seasonal trend is often attributed to food preparation
issues related to picnics and eating outside of the home (Coker et al., 2002,
Miller et al., 2004, Louis et al., 2005, Fullerton et al., 2008), this does not explain

why the same pattern exists among cultures with different summertime customs.

86

Others have suggested environmental factors may drive this seasonal pattern,
including higher loading of the bacteria in livestock and increased Campylobacter
transmission among poultry flocks by flies (Rosef and Kapperud, 1983, Hald et
al., 2004, Nichols, 2005) that could lead to an increase in food animal carriage
and potential for greater human exposure and infection.

As evaluated and reported in Chapter 2, demographic analyses in
Michigan for case data collected between 1992 and 2005 revealed that counties
with a high population density, young populations (<5 year age group), and
largely white populations were positively correlated with case rates; however, in
this study several counties reporting the highest case rates historically (annual
county averages up to 22.1 Campylobacter cases and 17.3 Salmonella cases per
100,000) did not follow these trends. These high incidence counties were often
located in rural areas of Michigan. This disparity in population density is
particularly interesting given that in a nation-wide study of campylobacteriosis in
the UK, Louis et al. (2006) found that case rates were negatively correlated with
population density and positively correlated with agricultural land use. Louis et al.
(2006), and others, have demonstrated that both environmental and weather
related factors influence contamination of surface waters with enteric pathogens
and human campylobacteriosis and salmonellosis disease patterns (Patz, 2001,
Lipp et al., 2002, Kambole, 2003, Ashbolt, 2004, D’ Souza et al. 2004, Fleury et
al. 2005, Zhang et al. 2008). Given the rural and suspected agrarian nature of the
‘anomalous’ high case rate counties in Michigan and the potential for

environmental transmission to explain these rates, we hypothesized that non-

87

demographic factors including environmental and weather-related variables may
be important influences in disease incidence. Here we evaluate the role of
agricultural land-use, water and sewage disposal resources, and weather
variability on a campylobacteriosis and salmonellosis patterns in Michigan over a

13 year period (1992 — 2005).

4.2 HYPOTHESES

Environmental and weather related factors are associated with changes in
rates of Salmonella and Campylobacter infections in Michigan such that:

0 As temperature increases incidence of these diseases will increase.

0 As precipitation increases incidence of these diseases will increase.

0 Areas with more agricultural land use sources will be prone to higher

incidences of these diseases.
0 As percentages of homes with non-municipal water and sewage

disposal increase rates of these diseases will increase.

4.3 METHODS

In order to capture multiple scales of possible influence on rates of
campylobacteriosis and salmonellosis, data were analyzed at the county,
watershed, and climate division levels. Data were obtained for each of the 83
counties and aggregated into ten climate divisions (defined by NCAA) (Figure
4.1) and into 62 watersheds (defined by USGS eight digit HUC codes) for
analysis (Ml DEQ 2009). Geographic analyses were conducted at the county
level to examine local-level land use factors that may contribute to

campylobacteriosis and salmonellosis rates, and at the watershed and climate

88

division level to evaluate the role of regional impacts that may be associated with
regional trends in reported campylobacteriosis. All data variables are described

in Table 4.2 and all model outcomes are described in Tables 4.3 and 4.4.

a. Campylobacteriosis and Sa_lmonel|osis Rates

Over the thirteen year period of study, records of all culture confirmed
Campylobacter and Salmonella cases in Michigan from 1992 to 2005 (from onset
date) were provided by the Michigan Department of Community Health (MDCH).
All culture confirmed cases of campylobacteriosis and salmonellosis in Michigan
must be reported to public health officials when diagnosed. Variables that were
extracted from the State database include the reportable condition, case status,
state and county of residence, onset, confirmation, diagnosis, and referral dates,
age, race, and gender. In instances where there was a missing identifier, the
case was excluded from the calculation of incidence for that variable. (For
example, if a case did not report the county of residence for the individual but did
give the state the case would be included in the state rate calculation but would
not be included with any county level calculations.) Population estimates were
obtained from the US Census Bureau (http:/Iwww.census.gov/popestl
datasets.html.). Monthly county incidence rates were calculated by dividing the
number of cases reported in a county in the study month by the number of
people in that demographic group in the county and multiplying by 100,000 to
give the number of individuals infected per 100,000 people per month.
Campylobacter and Salmonella rates were also evaluated using the Freeman

Tukey Square Root Transformation (transformed rate = (100,OOO)1’2{[C/N]"2

89

+[(C+1 )/N]1’2} where C is the number of Campylobacter or Salmonella cases and

N is the population) to accommodate model assumptions (Cressie, 1993).

b. County-level Geographic Analyses

Agricultural land-use data values were approximated from Michigan
Agricultural Statistics land in agricultural production report (MASS, 2005). The
land in agricultural production was reported in acres and this value was divided
by the county’s total land area to get the percent of agricultural land. No data
were reported for the counties of Luce and Keweenaw. For counties reporting
land in agriculture production, percentages ranged from 79% to less than 1%.
This range of percents was broken into the following category clusters: 0-20%,
21-40%, 41-60%, and <60%.

lnforrnation on the number of occupied homes, the source of drinking
water (public or non-public), and type of waste water disposal (centralized sewer
or on-site disposal, i.e., septic system or cess pit) for households in each county
was obtained from the US Census Bureau. The percent of homes using public
water sources and percent using public sewage disposal within each county were
calculated and these variables are described in Table 4.2. Percents ranged from
2.5% to 99.5% of homes in the counties on public water and 2.6% to 98.1% of
homes on public sewage. These were broken into 10% incremental categories
(deciles) to create 10 groups for public water source and 10 for public sewage
disposal.

Statistical associations between monthly county case rates and land-use

classification were assessed using the Generalized Estimating Equation, GEE,

90

(PROC GENMOD) analyses (SAS v.8, Cary, NC) with a repeated (county) option
(Table 4.2). All associations were considered significant at psO.10. The
GENMOD model (SAS v.8, Cary, NC) was also used to model homes using non-
public water sources and using septic systems in relation to disease incidence

rates.

0. Seasonal and Monthly Analysis

For the state, case rates were analyzed with respect to month of the year.
At the climate division level of analysis, an autocorrelation procedure using
month and incidence were performed for each of the divisions to determine the

seasonal patterns.

g. W_a__tershed and Climate Division Ggggraphic Analyses
County data were assigned to watersheds (defined by USGS eight digit

hydrologic unit codes [HUC]) based on the location of the county center using
geographic information software (ARCGIS v.9.1, ESRI, Redlands, CA). County
data were assigned to climate divisions per the National Oceanic and

Atmospheric Administration (NOAA) division boundaries.

e. Meteorolggical Factors

Daily average precipitation, average maximum daily temperature (F),
average minimum daily temperature, average mean daily temperature, high
temperature, and low temperature monthly data were obtained from the National
Climate Data Center (NCDC) for all weather stations in Michigan for 1992 -
2005. These variables are further described in table 4.1. Data were available

from stations state-wide. The counties Wayne, Washtenaw, St. Clair, Oakland,

91

Newaygo, Monroe, Macomb, Livingston, Lenawee, Lapeer, Genesee, Clare, and
Antrim did not have any weather station data reported from NCDC over the study
period. As these counties were missing these variables, they were excluded from
county level analysis for the missing variables. Data from all stations within an
individual county, watershed, or climate division were compiled taking the
average values from all stations with in the geographic area. Precipitation data
were analyzed as the amount of precipitation or the amount of snow (in inches)
for the month. Monthly average daily maximum and minimum temperatures were
evaluated for all stations within the counties. These values were used in
conjunction with the precipitation data in a GEE model (SAS v.8, Cary, NC) to
predict county case rates (Table 4.2).

The GEE analysis required evaluation of the variables in a two step
process. The first step was to perform a univariate analysis evaluating climate,
and environmental/geographical factors individually with respect to human cases
of Campylobacter and Salmonella infections. Variables were considered
significant at the 0.10 level (p<0.10). lf variables were significant in this univariate
analysis, they were then combined with other significant variables in a
multivariate analysis. The multivariate analysis was run with a variable removal
criterion of p>0.05. The final multivariate model contained all variables with
significance at the 0.05 level. Given that cases were count data and zero counts
were frequently recorded for study counties, the Campylobacter and Salmonella
cases were evaluated using the GEE model with a Poisson distribution with and

offset (log of population) function.

92

4.4 RESULTS

a. Campylobacteriosis and Salmonellosis Rates

Between 1992 and 2005, the average annual incidence of culture
confirmed Campylobacter infections in Michigan (statewide) was 4.28 cases per
100,000 people. There were 5,864 reported Campylobacter cases and an
average of 419 cases each year. The annual incidence of Salmonella infections
were 4.50 cases per 100,000 people. There were 6,263 reported Salmonella
cases and an average of 447 cases each year.

There were significant differences between climate divisions. For the rates
of Campylobacteriosis, division 3 (northwest lower peninsula) was significantly
higher than all other divisions (9.28 cases per 100,000 people) and divisions 7
(mideast lower peninsula) and 5 (midwest lower peninsula) the lowest (3.6 and
2.8 cases per 100,000 respectively). For Salmonellosis, the highest rates were in
climate division 3 (8.0 per 100,000) and the lowest in divisions 6 (middle lower
peninsula), and 7 (3.8 and 3.5 per 100,000) (Figure 4.1) (http:l/www.cpc.ncep.
noaa.gov/products/analysis_monitoring/regional_monitoring/CLIM_DlVS/michiga
n.gif). Throughout Michigan, annual incidences for both Campylobacter and
Salmonella infections showed no trend. For campylobacteriosis, climate divisions
3, 6, 7, 8 (southwest lower peninsula), and 9 (lower mid lower peninsula) showed
significant relationships with temporal measures month and year. Divisions 10
(southwest lower peninsula) and 3 showed significant temporal relationships with

salmonellosis rates.

93

There were significant differences in rates of human disease between 62
watersheds evaluated. Watershed 42 (Cedar) had the highest rates of
campylobacteriosis followed by 46 (Escanaba) and 28 (Platte)
(httpzllwww.michigan.gov/documents/deqllwm-mi-watersheds_202767_7.pdf).
Lowest rates were seen in 25 (Pere Marquette) and 56 (Presque Isle). Highest
salmonellosis rates were seen in 46, 42, and 4 (Betsie) and the lowest rates in 56

and 39 (Au Train).

b. County-Level Geggraphic Analvsefis

At the county level, land use, potable water source and method of waste

 

water disposal were evaluated to determine their relationship to incidence of
campylobacteriosis and salmonellosis. There were significant differences in land
in agriculture categories with respect to incidence rates. Counties with greater
than 40% of land area in agriculture had significantly higher rates of
campylobacteriosis than other counties. Counties with 21 to 40% of land area in
agriculture had the lowest rates for both campylobacteriosis and salmonellosis.
For salmonellosis, the 21 to 40% group was the only group significantly different
from the others. Lower rates of campylobacteriosis were associated with counties
with greater than 70% of homes on public sewage and counties with 20-30% on
public sewage. The counties with the highest percentages of homes using
municipal sewage had the lowest rates of salmonellosis. Counties with greater
than 80%, 50—60%, and 20-30% of homes using public water sources have the

lowest campylobacteriosis and salmonellosis rates statistically.

94

c. Seasonality

State-wide between 1992 and 2005, case rates peaked in the summer
months for both campylobacteriosis and salmonellosis. Mean rates in June, July
and August were significantly greater than those reported in other months
(Chapter 2, Fig. 2.4). For all the climate districts campylobacteriosis case rates
peaked in June and July and exhibited a second, lower peak in the late fall, but
this was not statistically significant. Salmonellosis rates peaked in July and
August. When regression techniques were use to evaluate the amount of
variation in Campylobacter and Salmonella explained by time (year and month),
the amount of Campylobacter variation explained ranged from 1.5% in division 8
to 9.5% in division 10 with division 5 having no significant temporal relationship,
and the amount of Salmonella variation explained ranged from 2.1% in division 1
to 10.4% in division 8. The results of analysis for significant seasonal
autocorrelation of case rates varied among the different climate divisions. Using
Tukey transformed case rates, division 4 was the only climate division to show no
monthly autocorrelations with campylobacteriosis rates, while all other divisions
showed autocorrelations for various monthly cycles with transformed
Campylobacter and Salmonella infection rates. Most climate divisions showed
autocorrelations at cycles 10 or greater suggesting an interannual trend, while a
many showed autocorrelations of the first or second order. This short period

suggests relationships between cases reported within one month of each other.

95

g. Meteorological Analvsis

Case rates in all climate divisions were evaluated for temperature (5
measures) and mean precipitation over the 14 year period of record. There was
no climate data reported for climate division 10. In the univariate analysis,
average temperature variables consistently significantly explained a small
percentage of the variation in Campylobacter and Salmonella infection rates
(Campylobacter. up to 10.5% of the variation in rates were explained by the
mean daily minimum temperature for that month in climate division 8 (southwest
lower peninsula) Salmonella: up to 5.8% explained by the same factor in the
same division (8)). Precipitation significantly explained some of the
Campylobacter rate variability in 4 of the 9 divisions with available climate data
(divisions 9, 8, 6, and 3). Precipitation explained Salmonella rate variability in
divisions 9 and 3. In all climate zones, average maximum daily temperature and
time were the most significant predictors of incidence using the Tukey

transformed data and produced the best fitting models predicting incidence.

4.5 DISCUSSION

During the 13 year period of analysis (1992-2005), rates of Campylobacter
and Salmonella infection in Michigan averaged annually 4.28 and 4.5 cases per
100,000 people respectively and varied across the state from 3.5 cases per
100,000 people in division 7 to 8.1 cases in climate division 3 for
campylobacteriosis and for salmonellosis cases ranged from 2.8 cases per
100,000 people in division 5 to 9.3 cases in division 3. During this same period

the national (U.S) averages for Campylobacter and Salmonella infections were

96

16.9 and 15.4 cases per 100,000 people, respectively. To explain the variations
in rates that are not explained by demographic factors, this study examined case
data by county and by climate district to evaluate the possible environmental
influences on rates of campylobacteriosis and salmonellosis.

Counties in Michigan with consistently high Campylobacter rates have
been identified (Chapter 2) including Menominee, Marquette, Isabella, Wexford,
Leelanau, Emmet, Alcona; and for Salmonella Wexford, Benzie, Branch,
Keweenaw. and Oscoda. These counties are dispersed throughout the state with
the largest grouping of these consistently high rate counties in climate division 3.
Most of these counties followed demographic trends of low population density
and tended to trend toward the older age groups. In contrast, national reported
trends suggest that high density counties with young populations should have
higher rates. These observations combined with recent studies associating
Campylobacter infection rates with climate and agricultural land use (Patrick,
2004; Kovats, 2005; Louis et al., 2005) make it evident that previous research
using only demographic variables have not adequately explained the variation in
Campylobacter infection rates in the US. This study examined possible
environmental factors to explain the variations in the Michigan Campylobacter
and Salmonella infection rates.

Environmental variables of interest include: land-use, potable water
source, method of sewage disposal, and meteorological factors (temperature
variables and average precipitation) as these factors are closely tied to water

quality and can affect large geographic areas. In order to examine these factors,

97

the scale of effect must be considered. The data were first examined on the
county level to ascertain the relationship between land-use, potable water
source, and method of sewage disposal to rates of Campylobacter infection. As
weather patterns often affect large geographic areas, the variables daily
maximum and minimum temperatures and daily precipitation were evaluated at
both the county and climate divisions level.

County. It has previously been shown that land use has a great effect on
the local environment and human health (urban areas allowing for more runoff;
forest lands allowing the least) (Interlandi and Crockett, 2003, Sherestha, 2003).
The source of run-off can in turn affect the types and amounts of pathogens
found in the waterways with agricultural and farmlands often associated with
fecal pathogens (Atwill, 1995, Mallin et al., 2000, Graczyk et al., 2000, Crowther
et al., 2002, Stanley and Jones, 2003, Ferguson et al., 2003, Kelsey et al., 2004).
Indeed, Potter et al. (2002) has found an association between high
concentrations poultry/farmland and high rates of campylobacteriosis (Potter et
al., 2002). In this study we found that counties with greater percentages of land in
agriculture had the highest incidence of campylobacteriosis while an intermediate
group, 21-40% of land in agriculture, had the lowest incidence of
campylobacteriosis and salmonellosis, suggesting that significantly more rural
areas may be more prone to high rates of Campylobacter infection. However, this
presents a disparity between these findings and the demographic analyses that
revealed that counties with intermediately higher population densities

(presumably more urban) had higher rates (Chapter 2). Despite these

98

differences, the trends noted here for agricultural land-use may suggest that
areas with more agrarian lands had higher disease burden, which is consistent
with other reports (e.g., Louis et al. 2006).

Statistically significant relationships were noted between the percentage of
homes using non-municipal potable water source and on-site sewage disposal
and rates, which indicated that these factors were predictors of lower case rates.
However, only small amounts of the variability in the case data could be
explained by either of these variables, suggesting that they are of low value in
studying the epidemiology of this disease.

Climate Division. The climate division level analysis allowed for
examination of large scale factors, such as climate, on disease rates. It has been
shown that variation in precipitation affects the local environment and human
health. Extreme changes in precipitation are known to be associated with
decreased water quality (Leeming et al., 1998, Lipp et al. 2001, lnterlandi and
Crockett, 2003) and increased gastrointestinal disease (Curreno et al., 2001,
Lipp et al, 2002), including campylobacteriosis and salmonellosis (Louis et al.,
2006, Zhang et al. 2008). Temperature has also been shown to have a great
effect on Campylobacter and Salmonella survivability such that lower
temperatures are favorable (Buswell, 1998, Danyluk et al. 2008); however,
published studies have found that higher temperatures and hours of sunshine are
significantly associated with the campylobacteriosis and salmonellosis incidences
(D’Souza et al. 2004, Patrick, 2004, Fleury et al. 2005, Kovats, 2005; Louis,

2006). Both high temperature and the number of hours of sunlight in the summer

99

help to explain the consistent seasonality of this disease; but still do not provide a
mechanism for the trend. These associations are consistent with our findings.
Models that included precipitation or either maximum daily temperature or
minimum daily temperature explained a significant percentages of the variability
in incidence of campylobacteriosis and of salmonellosis by climate division.

Maximum daily temperature alone was best able model most climate divisions.

4.6 CONCLUSIONS

In Chapter 2 we presented results that demonstrated demographic
patterns associated with high case rates including high population density, young
populations (<5 age group), and largely white populations; however, in the
counties with the most frequent high case rates over time this trend was not
observed. These high incidence counties were scattered throughout the state but
were primarily rural areas. We hypothesized that other factors were influencing
the distribution of rates in Michigan, particularly in these regions.

Here the environmental variables land use, source of potable water,
sewage disposal method, and climatological factors were evaluated for their
contribution to case rates in Michigan. As the major trend in rates across the
state is seasonality, as expected, time was often among the best predictors of
Campylobacter and Salmonella case rates in time series analyses. While
precipitation had some influence on certain climate divisions, temperature
(maximum daily temperature) was the most important environmental predictor of

climate division-wide variability in case rates.

100

The counties of interest identified in chapter 2, which had high rates
despite a low population density, included several counties in climate division 3
(north west lower peninsula). This division had the highest rates over all others
and showed significant associations with precipitation and all temperature
measures for both Campylobacter and Salmonella incidences. While differing
from the demographics of the state, the trends in case rates in this division better
correspond to previous research that suggest that rural and agricultural areas are
more prone to high incidence of campylobacteriosis (e.g., Patrick, 2004; Kovats,
2005; Louis et al., 2006). Furthermore, the primary role of temperature in
explaining the case rate variability is also consistent with reports from other areas
(D’Souza et al. 2004, Patrick, 2004, Kovats, 2005; Louis et al., 2006; Zhang et al.
2008). Therefore, it seems that in Michigan, environmental factors should be
further explored as a driver of disease in some rural areas whereas
demographics may play a greater role in highly metropolitan areas. These
observations are difficult to interpret, however, because of the incompatibilities
between the various units of analyses and the lack of understanding of the
natural history of campylobacteriosis and salmonellosis among these different
demographic and geographic units of analysis. This study highlights the poorly
understood environmental ecologies of these diseases and suggests that there
are multiple risk factors of disease at the individual level that are modified by
large scale and regional environmental impacts on pathogen presence. This

information on the difference between urban and rural centers is important in

101

attempts to prevent and understand this disease and suggests that different

strategies may be needed.

102

Figure 4.1. Michigan Climate Divisions and Counties.

'0. NOV

 

counties and climate division

boundaries.

103

Table 4.1 Michigan Counties with Watershed and Climate Division classification

 

 

 

COUNTY Watershed CLIMATE DIVISION
ALCONA 2 4
ALGER 39 2
ALLEGAN 17 8
ALPENA 36 4
ANTRIM 1 3 3
ARENAC 30 7
BARAGA 48 1
BARRY 14 9
BAY 18 7
BENZIE 4 3
BERRIEN 34 8
BRANCH 34 9
CALHOUN 17 9
CASS 34 8
CHARLEVOIX 10 3
CHEBOYGAN 1 1 4
CHIPPEWA 62 2
CLARE 32
CLINTON 14 9
CRAWFORD 2 4
DELTA 58 2
DICKINSON 50 1
EATON 14 9
EMMET 1 1 3
GENESEE 32 1O
GLADWIN 32 6
GOGEBIC 56 1
GRAND TRAVERSE 9 3

 

104

 

Table 4.1 Continued

 

 

 

COUNTY Watershed CLIMATE DIVISION
GRATIOT 32 6
HILLSDALE 21 9
HOUGHTON 55 1
HURON 26 7
INGHAM 14 9
IONIA 14 9
IOSCO 1 4
IRON 50 1
ISABELLA 32 6
JACKSON 14 9
KALAMAZOO 17 8
KALKASKA 20 3
KENT 14 8
KEWEENAW 55 1
LAKE 25 5
LAPEER 32 10
LEELANAU 28 3
LENAWEE 29 10
LIVINGSTON 32 10
LUCE 60 2
MACKINAC 41 2
MACOMB 12 10
MANISTEE 20 3
MARQUETTE 46 1
MASON 5 5
MECOSTA 22 6
MENOMINEE 42 1
MIDLAND 32 6

 

105

 

Table 4.1 Continued

 

 

 

COUNTY Watershed CLIMATE DIVISION
MISSAUKEE 22 3
MONROE 29 1 0
MONTCALM 14 6
MONTMORENCY 36 4
MUSKEGON 22 5
NEWAYGO 37 5
OAKLAND 12 1O
OCEANA 37 5
OGEMAW 30 4
ONTONAGON 53 1
OSCEOLA 22 6
OSCODA 2 4
OTSEGO 1 1 4
OTTAWA 14 8
PRESQUE ISLE 23 4
ROSCOMMON 22 4
SAGINAW 32 7
SANILAC 6 7
SCHOOLCRAFT 49 2
SHIAWASSEE 32 9
ST. CLAIR 27 1O

 

106

 

Table 4.1 Continued

 

 

 

COUNTY Watershed CLIMATE DIVISION
ST. JOSEPH 34 9
TUSCOLA 32 7
VAN BUREN 34 8
WASHTENAW 1 5 1 O
WAYNE 31 10
WEXFORD 20 3

 

Case data were collected at the county level and to evaluate watershed and
climate division levels the county case data had to be assigned to the respective
units. This table shows which counties were assigned to each watershed and to

each Climate division.

107

 

.mcfimtg owe... ho EoucEEma 65 5.3 92m .636 9: 2. 693.90 83.2.9 9.. 950%.. 2an

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:mcSm - 85:53 99%
9.650 m: moom No? .952. 2986 5:300 cozmiaon. Ema cozmfiaon.
a 5.8... meow .8

758250 .ancocn. mommo m.mo__ccoE_mw

.o EoEtwaoQ - Name 2.8 9986 Eczoo 8mm. 0.30
$922.2 .r beacon wowmo m_mo_.ofimno_>quo

mw<s. a

8.2mm 8.6.66 - .mZoE. toughen. .9339: <
53.3.5.4 moom Nam? .macc< msosczcoo 3:30 5 96.. .0 E861 ”cm: new.
$96.:
6.9: 9mm.
2an 9%.: 326: 3.5380 .mEmEco._>cm

:moSm . 98> .95.... .0 E983 ”8.30m Ecum>>

2650 w: ooom cm? or 29m. £635.50 3:300 .mwommfi
6.33 9%: 8:6: 83:80
go Benton. ”.mmoamfi. cmchm
aEcH
26.. 280
anh
82.8 we: :9... 2.3
20.an05 aEc...
6002. team. macaflw >__mo Es. c>< 9392.5”;
.250 Ema 2.8 £635.80 Saga
o.mE__o .98sz 80.5% 5533 anh
. . 0E8 ”202 >__mn. xms. o><
88-89 9:3 :82
2.8 w><
:wwflmuwmmhn. cozmggooi
sweet .93
cocaom 0E... 322:. _o>c.. nEtonom mo.na_._a> on: c_na..a>

 

 

 

 

 

 

108

 

 

mozsa> 682 .o 8.3580 N... 293

.353: 05 05:000.: “Em .:om0

:. 8.5.0:. 3.90:? 05 0:50.“. £6058? 05.00% :80 063.90 2 .650 05 :. 50500.0 0.08:. 055 009.0005 059 0:...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

$36 28.0 5.6.5.85 25:22 8me 080% 8885
58.9 58.9 25cm 258.2 3.825% E. 8628.
. . :o_ m 560:. > :o 8890:. «00:02...
Fomo O rooo CV .U— u. . Eu 5— mwwmo mwmwwfi :Otmtgomkq m<
$3.... 58.? 26:0 2585. 6.88258
55.0 Fooodv E:E.:..>. 5:22
mwmod Fooodv :58.me 5:22
. . 0008 000006
ammo o 58 ov 2.00 E:E.:..>_ :005. m..0:oE.mm
mde Soodv 2.00 E:E.xm.2 :005.
93.0 Soodv 2.8 :85. 00005:. .3:
3:023. 00803:.
mmwod Soodv E:E.:..>. 5:05. Co. 0559380.. m<
memo... Soodv E:E.st. 5:22
. . 0008 00,000.“.
Name o 58 ov 2.8 E:E.:.s. :005. cetmnoinmo
mmwod Boodv 2.0m. £35.55. :02).
RS... 885v 2.8 50.2
m : m0.nm:m> 0030:;
5:030:35 5:03:33 0.350%...

938m 58:. H6

 

 

 

 

 

.0008 000005 2 6050.0. 5.3 0030:? 20:50 :8 5.300. 0.0205 20:02:: me 0.53

109

0050.... 05 02000.0 0:0 .0000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:. 00020:. 00.00..0> 05 0.0.00 0.005003. 0.0.0000 0000 0.0:.0>0 0. >030 05 :. 00000.0 0.000E 05 000.8000 0.00. 0.: ..
-- o F .OA 00:0.0.0>>
. 00000 00000.0 .
l or 0A .0505 5.00... . 00000.0
0..0:oE.0m .0 00:00.05
I o _..OA :o_m_>_O 052.50 btm 99659.9;
-- o FdA 00:0.0..0>> 0.305000 :00500
00000 00000.0 5:08.50.
1 027 8.50.0 5.00: 0.000003500 0 00 ...>. 0.0...“
500.0- 0000.0 :0.0.>.Q 0.0:...0

.. of? .000005 000300 00000 00000.0 0000:05
. ....s 00:00.05
-- o F .7 00.000 .0.0>> 0.. 0:05. 0m 0000.05 000505
-- . .000Q0.0 000.00
or CA .000005 000.50w 00000 00000.0 0:0 0.05. 0.05.0
.. o :7 00.000 .0.0>> .0.0000.>.QE00 5.... 0050: .0 we 0...
. . :0500006 00000 00000.0 00000:. .5:
BR o $8 o 0.0. :. 0:0. .0 0005:0200 050:0..500. 00:00.05 0000005
550000....
.. . :0503006 00000 00000.0 0:053:00
0.. 0A m< :. 0:0. .0 0005:0200 0.000003500 5 0:0. .0 R. 0.0.

m. 0 00.00..0> 00_00_.0> 0.00 “on;

E00:0000:. 5:00:000o . 0 I

0.380 .008. 000

 

 

 

 

.0008 00000.0 0. .00000. 5.3 00.00..0> 0.:00.mo0m0:0E:0..>:0 .0. 0.....00. 0.0>.0:0 0.0..0>.:... v.v 0.00 ..

 

CHAPTER 5

Campylobacter and Salmonella on Michigan Dairy Cattle Farms: Culture

isolation and Enumeration from Environmental Soil and Water

5.0 STRUCTURED ABSTRACT:

Objective: To isolate and enumerate, via culture methods, Campylobacter and
Salmonella from Michigan dairy cattle farm environments in counties with varying
rates of human disease and relate recoverability to temperature.

Specific Aims: 1) To evaluate culture methods for the enumeration of
Campylobacter and Salmonella in environmental soil and water samples, 2) To
experimentally evaluate the effect of temperature on Campylobacter and
Salmonella recovery from soil and water, and 3) To evaluate Michigan dairy
farms in counties, with varying rates of human disease, for the presence of
Campylobacter and Salmonella in soil and water.

Methodology: 1) Methods were derived from available literature on Most
Probable Number (MPN) enumeration techniques and environmental soil and
water isolation methods for Campylobacter and Salmonella. These methods
were straightforward for water but required modification for soil as there was a
gap in the literature on isolation of Campylobacter and Salmonella from soil. For
water, standard methods were used. For soil, published methods for enumeration
of these bacteria in food and detection in soil were evaluated for: a) the necessity
of shaking the sample during incubation, b) volume of media, c) number of
replicates for MPN, d) number of days for Campylobacter incubation, e) effects of

background organisms, f) range of detection, and 9) time untill evaluation. 2)

111

Microcosm experiments were carried out with sampling and processing of soil
and water from varied temperate environments (Freezing 3.5 C, Cool 32 C, and
Room temp 70.8 C) every 24 hours for 72 hours. 3) Soil and Water samples were
collected from Michigan dairy cattle farms. The water samples were filtered at 2
volumes (10ml and 25ml) and soil samples processed at 2 measures (59 and
19), pre-enriched in Campylobacter and Salmonella selective media (Preston,
BPW), replicated and diluted in enrichment media (Preston, Tetra), Plated
(CCDA, XLD), and biochemical tests were performed on presumed positive
samples.

Results: 1) Results from soil and water methods validation procedures did not
follow expected patterns. Methods used to evaluate autoclaved spiked soil
samples showed sensitivity to Salmonella from as low as 5 cfulg and
Campylobacter 50 cfulg, while non-autoclaved spiked samples showed
inconsistent results. Methods used to evaluate autoclaved spiked water samples
showed sensitivity to Salmonella as low as 0.5 cfu/ml and only 500 cfulml for
Campylobacter. Non-autoclaved spiked water samples showed inconsistent
results. 2) Autoclaved spiked soil and water samples stored in the cool
environment had the greatest recovery rates for both Campylobacter and
Salmonella (all positive). Lowest recovery rates for both Campylobacter and
Salmonella came from those stored at room temperature. 3) Fourteen farms
were sampled over the course of the study (sampled for soil and water once or

twice during the study sampling period October 2008-June 2009). One water

112

sample (sampled during the October sampling cycle) was positive for
Salmonella. All other samples were negative for Campylobacter and Salmonella.
Conclusions: MPN techniques in conjunction with culture isolation methods for
Campylobacter and Salmonella from environmental soil and water samples were
insufficient to allow for enumeration in raw samples. it is suspected that
background organisms play a major role in competition for nutrients during the
culture processes as in very dilute spiked non-autoclaved samples there was
some recovery of Campylobacter. While in more concentrated samples there
was growth of other organisms on the plates. This interference is also suspected
to have influenced the Salmonella results, as recovery of Salmonella did not
always decrease with dilutions. As expected, per the literature, the lowest
recovery of Campylobacter and Salmonella was associated with the warmer
temperature. Further research would have to be done to determine if there is a
relationship between the 1) prevalence (enumeration) of Campylobacter and
Salmonella on Michigan dairy cattle farms and temperature and 2) prevalence
and human incidence of disease in that county due to the low recovery of
Salmonella (1 positive sample) and no recovery of Campylobacter.
Significance: Campylobacter and Salmonella are common infectious bacteria
often associated with the foodbome route of infection in humans.
Environmentally, these bacteria are closely linked with poultry and cattle
environments. This study aimed to evaluate the amounts of these bacteria in
Michigan dairy cattle environments in counties with varying rates of human

disease and the relation to environmental temperature. By modeling the

113

 

environmental effects on transport and prevalence of these bacteria in the
farming and surrounding environments this study aimed to provide a major step
towards understanding trends in prevalence and gain insight into how to lessen
transmission between these food animals and on to humans. The actual findings
of this study were inconclusive in this regard; however strides were made in the
field of laboratory methods. in the literature, there were many methods published
for isolation of Campylobacter and Salmonella from water and food, however
there was a lack of adequate instruction for MPN techniques and culture isolation
methods of these bacteria from soil. This study was able to identify the barriers to
this process and to discover that culture techniques for isolation and enumeration
of these bacteria from environmental soil samples is not likely to yield reliable

results given the competition of other organisms in the samples.

5.1 INTRODUCTION:

Backgrognd: Campylobacteriosis and Salmonellosis are common
infectious diseases often associated with food routes of infection and cattle serve
as a major reservoir for Campylobacter and Salmonella with high carrying rates
in these animals (Stanley et. al, 1998, Madden et. al, 2007). Through these and
other food animals, Campylobacter spp. and Salmonella spp. go on to infect
humans causing high rates of gastrointestinal disease, world-wide. Given the
high incidences and health burdens for these diseases, much research has been
done to better understand the routes of transmission and minimize infection;
however this has mainly been done through the food route (food to mouth). As a

result, safeguards were put into place which leads to a decrease in incidence of

114

these diseases, however the seasonal trends (human cases peaking in the
summer months) remained (Buchanan and Whiting, 1998, Allos et al., 2004,
USDA, 2006, CDC, 2006). Numerous factors likely contribute to the disease
burden and there is growing evidence that environmental factors such as climatic
variability, including changes in temperature and precipitation, are associated
with outbreaks of intestinal diseases.

Seasonality in human outbreaks and in environmental prevalence of both
pathogens has been noted in the literature with peaks reported generally in the
summer months. in the United Kingdom and other areas, Campylobacter spp.
detections in watersheds increase with or just prior to peaks in human cases in
the late spring and early summer (Louis ., 2005, Eyles et al., 2003, Arvanitidou et
al., 2005). The highest frequency of Salmonella isolations from humans occurs
in the late summer months also associated with increased rainfall (Haley et al.,
2009, Geather et al., 2009). These links should be further studied to fully
understand and prevent future cases of Campylobacter and Salmonella
infections.

With this study, we aim to add to the literature by exploring and reporting
on the environmental presence of these bacteria in the dairy farm environment of

counties with varying human rates of Campylobacter and Salmonella infections.

Rationale: Research has been done to show the trends in human infection
with Campylobacter and Salmonella showing an overall declining rate noted
since 1996 (attributed to improvements in the poultry and food processing

industries) and a distinct seasonal trend that remains. Research has also been

115

done to show the trends in food animal infection with Campylobacter and
Salmonella showing a seasonal trend in shedding rates. There is also research
showing seasonal trends in water prevalence of these bacteria. However, there
has not been much research done to model this potential pathogen transport
route (Skelly Weinstein, 2003). This study aims serve as a preliminary step in
bridging the gap in the literature. These studies explores through laboratory
evaluation, experimentation, and field sampling the culturable recoverability of
these bacteria in the dairy farm environments along with possible influences
(temperature) that could affect the environmental survival and recovery of these
pathogens. This study should lead to larger studies where tracers or molecular
typing will be used to confirm the path of transport for these bacteria from farm
animals, to the environment and water, and then on to humans (\Mlson et al

2008).

5.2 HYPOTHESES
0 Culture methods can be used in conjunction with MPN to
enumerate Campylobacter and Salmonella from environmental soil

and water samples.

0 Recovery of Campylobacter and Salmonella will vary with
temperature.
0 There will be variation in the amounts of Campylobacter and

Salmonella present in the farm soils and surrounding waters that
will relate to human incidence of these infections in county.

0 There will be a relationship between temperature and the amount of

116

Campylobacter and Salmonella in cattle farm soils and surrounding
waters such that higher prevalence of the bacteria in the

environment corresponds to warmer temperatures.

5.3 SPECIFIC AIMS
0 To identify and validate methods for the enumeration of
Campylobacter and Salmonella from environmental soil and water
samples.
0 To evaluate the effect of temperature on Campylobacter and
Salmonella recovery from soil and water.
0 To evaluate Michigan dairy farms for the presence of

Camp ylobacter and Salmonella in soil and water.

5.4 METHODS

a. Stud Desi n
This study will employed a three tiered design that included 1) a

validation of laboratory procedures, 2) a microcosm experiment in varied
temperate simulated environments, and 3) field sampling from Michigan dairy

cattle farms.

b. Sampling and Laboratory Procedures

_Collection of Soil and Water samples:

 

Soil. Samples from areas of exposed soil (near lagoon/manure storage
area) were collected; about 209 of surface soil will be collected and stored in the

Whirl-Pak bags using methods described by Johnson et al.,(1997).

117

Water. Water samples were collected from waters in the direct draining
area (as determined by the drainage commission) of the sampled farm (2
samples for each bacteria). At each site, samples were collected in sterile
bottles from standing water on the farm. The sample was taken from the top

20cm of water using methods described by Sayah et al.,(2005).

Qampvlobacter enrichment and culture:

For processing the soil, 1g and 59 samples were measured and pre-
enriched in 45ml of Preston Enrichment (PE) broth, incubated at 42°C for 48
hours under microaerophilic conditions. Samples of 10ml and 25mls of water
were filtered through 0.45um membrane filters. These filters were then pre-
enriched in 45ml PE broth and incubated at 42°C for 48 hours under
microaerophilic conditions. From both the water and soil pro-enrichment broths,
secondary enrichments were performed by adding aliquots of the broth to 9ml of
PE and further diluting by adding aliquots from that tube to 9ml of PE. Both of
these broths were done in triplicate and incubated at 42°C for 48 hours under
microaerophilic conditions. All samples were then streaked onto CODA-Preston
agar plates, then incubating them at 42°C for another 48 hours under
microareophilic conditions. Gram-stain, oxidase test, and motility testing were
performed to further confirm the presence of suspected Campylobacter growth

on those plates.

Salmonella enrichment and culture:
For processing the soil, 1g and 59 samples were measured and pre-

enriched in 45ml of Buffered Peptone Water (BPW) and incubated at 37°C for 24

118

hours. Samples of 10ml and 25mls of water were filtered through 0.45um
membrane filters. These filters were then pre-enriched in 45ml BPW broth and
incubated at 37°C for 24 hours. From both the water and soil pre-enrichment
broths, secondary enrichments were performed by adding aliquots of the broth to
9ml of Tetra and further diluting by adding aliquots from that tube to 9ml of Tetra.
Both of these broths were done in triplicate and incubated 37°C for 24 hours. All
samples were then streaked onto Xylose Lysine Desoxychoiate (XLD) agar to be
incubated at 37°C for 24 hours. TSl, Urea, Citrate and LIA testing were
performed to further confirm the presence of suspected Salmonella growth on

those plates.

c. Validation Study

This portion of the study initially required a review of the literature around
laboratory methods for the 1) culture isolation and 2) enumeration of
Campylobacter and Salmonella from environmental soil and water and the 3)
range of detection for these methods with respect to recorded environmental
concentrations. From these findings laboratory procedures were optimized
through experimentation that evaluated the necessity of shaking soil samples,
media volume, time of dilution, order of replication, number of days for
campylobacter incubation, time till evaluation, effects of background organisms,
and the range of detection.

For all optimization experiments soil and water samples were collected
from a dairy farm using methods described above and treated appropriately for

each experiment. To evaluate the necessity of shaking soil samples, autoclaved

119

soil samples were spiked; one set of samples with 50cfulg of Salmonella and
another set with 50cfulg of Campylobacter. These samples were evaluated using
laboratory procedures described above with one set of both Salmonella and
Campylobacter shaken during the pre-enrichment incubation and one set not
shaken. To evaluate the optimum pre-enrichment media volume, spiked
autoclaved soil and water samples were pre-enriched in varying concentrations
of media. Spiked autoclaved samples were evaluated for the effect of diluting the
sample at the pre-enrichment as compared to the enrichment phase of isolation.
To evaluate the effect of replicating the samples at prior to or post dilution, spiked
sample sets were replicated (triplicate) at the pre—enrichment phase and
compared to the set replicated at the enrichment phase. Campylobacter spiked
samples were processed with varying incubation times (24 vs. 48hrs) to
determine the optimum incubation time. Campylobacter and Salmonella spiked
autoclaved soil and water samples were stored in the refrigerator and processed
every 12hrs for 48hrs to assess the maximum time till evaluation. To assess the
effect of background organisms, both autoclaved and not autoclaved spiked
samples were processed and compared. Autoclaved samples were spiked with
varying concentrations of Campylobacter and Salmonella then processed and

compared.

d. Microcosm Temperature Stugy

Microcosm experiments were carried out with sampling and processing of
soil and water from varied temperate environments (Freezing 3.5 C, Cool 32 C,

and Room temp 70.8 C). Autoclaved Campylobacter and Salmonella spiked soil

120

and water samples were stored in a freezer, refrigerator, and desk top with
temperature monitors. A set of the samples were removed from the environment

and processed every 24 hours for 72 hours. Results were recorded.

e. Environmental Daim Farm Sampling Study

Study Area
Michigan is a state in the upper mid-westem area of the United States,

bordering Canada, and has both rural and urban areas. The state has a
population of 10,120,860 people, a land area of 56,804 square miles, and 40,001
square miles of water (US Census, 2006). Michigan has both urban and rural
areas with much of the land used in agriculture and taming. The state is divided
into 83 counties and 59 watersheds (EPA, 2006). Many of the waters of Michigan
are used for recreational activities such as swimming and fishing in the warmer
months.

There are 14,500 cattle operations in the state with 864 dairy farms
(NASS, 2004).

County Selection

The sampling sites were chosen based on identification of Michigan
counties and watersheds with high densities of farms and varying human
incidences of reported Campylobacter spp. and or Salmonella spp. infections.
Wrthin these counties and watersheds, dairy cattle farms were selected for
sampling. A letter describing the research project was sent to a random sample
of cattle farms in the areas of interest and those interested in participating

responded by returning a prepaid postcard. This resulted in the 6 counties to be

121

sampled with 14 farms choosing to participate in the study (Isabella County, 5
farms; Livingston, 2; Mecosta, 1; Missaukee, 4; lngham, 1; and Clinton County, 1

farm).

Sampling and Processing

Soil, environmental, and water samples were collected from all farms
enrolled in the study and their surrounding draining waters. All samples were
stored on ice and analyzed within 24 hours of collection using previously

described methods.

5.5 RESULTS:

a. Validation Study

There was slight variation in results of the shaken and not shaken
autoclaved soil samples (Table 5.1). There was also slight variation in outcome
between the 90 ml pre-enrichment media volume as compared to 45 ml (Table
5.1). There was no difference in outcome when spiked autoclaved samples were
diluted at the pre-enrichment phase when compared to diluting at the enrichment
phase of culture (results not shown). Similar, results were achieved when spiked
sample sets were replicated (triplicate) at the pre-enrichment phase and at the
enrichment phase (results not shown). Campylobacter spiked samples were
processed with varying incubation times to differentiate the recovery between a
24 hour and 48 hour incubation. Greater recovery was seen, with positive
samples at 10‘3 when incubated at 48 hours while no samples were positive at
that concentration when incubated at 24 hours (Table 5.2). When Campylobacter

and Salmonella spiked autoclaved samples were stored in the refrigerator and

122

processed every 12 hours for 48 hours to assess the maximum time till
evaluation the greatest recovery occurs when samples are processed prior to 24
hours (T able 5.3 and Table 5.4). Methods used to evaluate autoclaved spiked
water and soil samples showed sensitivity to Salmonella in the range of 0.1cfulml
of water to a minimum of 2 cfulg of soil. Campylobacter minimum limits were
around 0.4cful.g in soil and 2cfu/g in water (Tables 5.5, 5.6, and 5.7). Non-

autoclaved spiked samples showed inconsistent results.

b. Microcosm §t_ud_y

Autoclaved spiked soil and water samples stored in the cool environment
had the greatest recovery rates for both Campylobacter and Salmonella (all
positive). Lowest recovery rates for both Campylobacter and Salmonella came

from those stored at room temperature (Table 5.8).

c. Field Study

Study Area and Counties Selected

The sample farms in the study were located in Livingston, Missaukee,
Isabella Mecosta, lngham, and Clinton counties. These counties have varying
reported human rates of Campylobacter and Salmonella infections (range in
average incidence of 1.3 to 14.2 for Campylobacter and 1.8 to 13.6 for
Salmonella infections), varying population densities (26 people per mi2 in
Missaukee to 499 in Ingham County), and vary with respect to the number of
dairy farms in the counties (16 dairy farms in Livingston to 119 in Mecosta)

(Table 5.9).

123

Processed Samples

All soil and water samples from all farms were negative for Campylobacter
spp. One water sample from a farm in Livingston County was positive for

Salmonella spp. (T able 5.9).

5.6 DISCUSSION:

a. Validation study

Given the limited budget for the study, we aimed to employ the most cost
effective and time efficient yet still valid and reliable methods for the culture
isolation and enumeration of Campylobacter and Salmonella from environmental
soil and water samples (Carrique-Mas et al, 2009). As such, a process diagram
was created and process evaluation was performed to identify areas for method
modification and validation. Several process steps were identified for evaluation
which included, a) the necessity of shaking soil samples during incubation, b)
volume of media during pre-enrichment, c) replication and dilution (number of
replicates and at which stage in the process to replicate and dilute the samples)
for MPN, d) number of days for Campylobacter incubation, e) effects of
background organisms, f) range of detection, and 9) time till evaluation. The
evaluation and analysis of these process steps made up the validation study.

In this validation study, we found that several steps could be minimized with
respect to volume of media used without compromising the outcome. The results
showed that 45 ml of pre-enrichment media could be used for sample processing
and that samples can be replicated and diluted later in the process (where

smaller volumes of media are required) (this data was not shown). In previous

124

studies with the processing of food samples, it was noted that samples should be
shaken during the pre-enrichment incubation. In our evaluation of soil samples
spiked with Salmonella, we found no difference in detection between the shaken
vs. the non shaken samples when using 45ml pre—enrichment methods. We also
found that at all steps of Campylobacter processing (pre-enrichment, enrichment,
and plating), maximum detection occurs when samples are allowed to incubate
for 48 hours. All samples should also be processed within 24 hours of collection
for maximum detection.

The final variables addressed in the validation study included the range of
detection for the methods employed and the effect of background organisms on
recovery. To evaluate the effect of background organisms, autoclaved and non-
autoclaved samples were processed and compared. It is suspected that
background organisms play a major role in competition for nutrients during the
culture processes as in very dilute spiked non-autoclaved samples there was
some recovery of Campylobacter. While at higher concentrations there was
growth of other organisms on the plates. This interference is also suspected to
have influenced the Salmonella results, as recovery of Salmonella did not always
decrease with dilutions. Due to these findings we determined that MPN
techniques in conjunction with culture isolation methods for Campylobacter and
Salmonella from environmental soil and water samples would be insufficient to
allow for enumeration in raw samples, as such microcosm temperature and farm

studies evaluated samples for recovery and detection.

125

b. Microcosm Temgrature Study
It has been reported in the literature that Salmonella and Campylobacter

survive for longer periods of time it colder temperatures (Buswell et al 1998). As
expected, per the literature, the lowest recoveries of Campylobacter and

Salmonella were associated with the warmer temperature.

c. Farm Sampling Study

It was expected that the sampling areas would have variation in the
environmental presence of Campylobacter and Salmonella such that counties
with higher incidences of these human diseases would have more environmental
presence of the associated bacteria. This finding would begin to strengthen the
possibility that environmental contamination may be a significant source of
human infection. However, in this study there was low recovery of Salmonella (1
positive sample) and no recovery of Campylobacter from any of the farm
samples. So our findings are inconclusive as to whether there is any correlation

between county level human incidence and environmental prevalence.

5.7 CONCLUSIONS:

This study aimed to begin filling in a gap in an alternate route
(environment) in the chain of Campylobacter and Salmonella transmission from
animals to humans. The food route has been explained, and the water route has
also been explained, but here the environmental contamination link between the
dairy cattle and the abundance and transport of the bacteria through the

environment was explored. The study employed a systematic approach to

126

validate the methods, cany out an experimental microcosm study, and finally
evaluate the natural farm environment.

This study found that due to the competition of background organisms,
culture methods were insufficient to accurately and consistently evaluate
environmental samples for the presence or absence of Campylobacter and
Salmonella. Further research would have to be done, possibly repeating the
study using quantitative real-time PCR techniques, to determine if there is a
relationship between the prevalence (enumeration) of Campylobacter and
Salmonella on Michigan dairy cattle farms (in the natural environment) and
temperature (Hadjinicolaou et al, 2009).

Ideally, this study would lead to larger studies where tracers or molecular
typing will be used to confirm the path of transport for these bacteria from animal,
to the environment and water, and then on to humans. This would also significant
public health policy potential, by confirming this path policies can be put into
place to inform animal farming facilities of the biological waste contamination
produced and make efforts to reduce them. lnfonnation can also be given to the
residents of nearby communities where exposure to these contaminants is likely

in efforts to reduce exposure and prevent infection.

127

Table 5.1 Media Volume and Shaking of Soil Samples

 

 

 

 

 

Concentration N“ Soil Results (positive/total")
Media volume
(Scfu/g of Salmons/W10") shaken not shaken

-2 3/3 3/3

-4 3/3 3/3
45ml

-6 3/3 3/3

-8 0/3 013

-2 313 3/3

-4 3/3 3/3
90ml

-6 3/3 0/3

-8 0/3 0/3

 

 

 

 

** The concentration of each sample is the initial spiking volume multiplied by

10", where N is the value listed above in the concentration column.

*Results are given as the number of positive samples out of the three replicates.

128

 

Table 5.2 Campylobacter Processing: Days in Incubation

 

 

 

Concentration N**
1 day 2 days
(500cfu/g of Campylobacter’10")
o 013 013
-1 3/3 3/3
—2 3/3 3/3
-3 0/3 3/3

 

 

** The concentration of each sample is the initial spiking volume multiplied by

10", where N is the value listed above in the concentration column.

129

 

Table 5.3 Time Till Evaluation of Soil Samples with Background Organism

 

 

 

 

 

 

Consideration
Soil Results (positive/total*)
Time till
Non-Autoclaved soil
Evaluation Autoclaved soil amount“
amount
(NS)
59 1g .59 59 1g .59
0 6/6 6/6 6/6 6/6 6/6 6/6
12 6/6 6/6 6/6 2/6 3/6 6l6
Campylobacter 24 6/6 6/6 6/6 0/3 3/3 3/3
36 3/3 3/3 3/3 0/3 3/3 313
48 3/3 3l3 0/3 0/3 0/3 3l3
0 3/3 1l6 0/6 5/6 6/6 0/6
12 6/6 1/6 0/6 0/6 0/6 016
Salmonella 24 6/6 0/6 0/6 0/6 0l6 0/6
36 6/6 0/6 1l6 2/6 3/6 1l6
48 6/6 2/6 6/6 2/6 3l6 2/6

 

 

 

 

 

 

 

 

*Results are reported as the number of positive samples out of the total number.
The six samples include triplicate samples spiked with 218 CPU/ml of
Campylobacter or 2cfu/g of Salmonella and processed in triplicate and each
diluted 1:10 ratio.

“Soil samples were evaluated using 59, 1g, and .59 amounts of spiked soil.

130

 

Table 5.4 Time Till Evaluation of Water Samples with Background Organism

 

 

 

 

 

Consideration
Time till Salmonella Water Results (positiveltotal*)
Evaluation Autoclaved water volume Non-Autoclaved water volume
(hrs) 25ml 10ml 1ml 25ml 10ml 1ml
0 9/9 9/9 2l9 6/9 7l9 5/9
12 9/9 9/9 4/9 9/9 9/9 6/9
24 9/9 9/9 0/9 8/9 7/9 5/9
36 9/9 6/9 0/9 4l9 9/9 7/9
48 9/9 6/9 7l9 5/9 7/9 9/9

 

 

 

 

 

 

 

*Results are reported as the number of positive samples out of the total number.
The nine samples include triplicate samples spiked with 50 CF Ulml of Salmonella

and processed in triplicate and each serially diluted 1:10 and 1:100 ratios.

131

 

 

 

 

 

 

Table 5.5 Range of Detection
Camp ylobacter Salmonella
Sample type Concentration Result Concentration Result
(CF U) (positive/total) (CF U) (positive/total)
102 11/12 25.6 12l12
Soil (per 9)* 10 11/12 2.2 12l12
1 12/12 1.7 6l12
102 12112 25.6 12/12
Water (per
10 12l12 2.2 12l12
ml) *
1 12/12 1.7 12l12

 

 

 

 

 

 

Results are reported as the number of positive samples out of the total number.

*The twelve samples include autoclaved samples spiked with Campylobacter or
Salmonella and processed in triplicate and each diluted 1:10 ratio. This was done

for each concentration and for two soil amounts (59 and 19) and for two water

volumes (25ml and 10ml).

132

a“.
w

 

Table 5.6 Range of Detection in Water with Background Organism

 

 

 

 

 

 

Consideration
Water Results (positiveltotaP)
Concentration Non Autoclaved Autoclaved water
(CF Ulml) water volume volume
25ml 10ml 1ml 25ml 10ml 1ml
384 019 019 019 919 719 919
38 019 019 019 019 019 019
Campylobacter
4 019 019 019 019 019 019
0.4 919 319 019 019 119 019
143 919 819 419 919 919 919
14 619 719 619 919 919 219
Salmonella
1 519 719 119 919 919 519
0.1 419 319 519 119 619 319

 

 

 

 

 

 

 

 

*Results are reported as the number of positive samples out of the total number.
The nine samples include triplicate samples spiked with Campylobacter or

Salmonella and processed in triplicate and each diluted 1:10 and 1:100 ratios.

133

 

Table 5.7 Range of Detection in Soil with Background Organism

 

 

 

 

 

 

Consideration
Soil Results (positiveltotal*)
Concentration Non Autoclaved soil Autoclaved soil
(CF U/g) amount amount
59 19 .59 59 19 .59
218 616 616 616 616 616 616
Campylobacter 20 116 616 316 616 616 616
2 316 016 0/6 516 016 616
2 516 616 016 616 116 016
Salmonella 0.2 016 016 016 016 016 016
0.02 016 016 016 016 016 016

 

 

 

 

 

 

 

 

*Results are reported as the number of positive samples out of the total number.
The six samples include triplicate samples spiked with Campylobacter or

Salmonella and processed in triplicate and each diluted 1:10 ratio.

134

 

Table 5.8

Recovery of Campylobacter and Salmonella with Respect to Length

of Time at Varying Storage Temperatures

 

 

 

 

 

 

Temperature Microcosm
Sample Campylobacter (+1total) Salmonella(+ltotal)
Time
Type Room Room
Freezing Cool Freezing Cool

Temp Temp

24 12112 0112 12112 11112 12112 12112

Soil“ 48 1112 12112 0112 12112 12112 12112
72 1112 12112 0112 1112 12112 12112

24 12112 12112 12112 11112 12112 12112

Water" 48 12112 12112 12112 12112 12112 0112
72 12112 12112 12112 12112 12112 12112

 

 

 

 

 

 

 

 

Results are reported as the number of positive samples out of the total number.

 

*The twelve samples include autoclaved samples spiked with 10cfulg or ml of
Campylobacter or 2.2 cfulg or ml Salmonella, processed in triplicate, and each
diluted 1:10 ratio. This was done for each soil amount (59 and 19) and for each

water volume (25ml and 10ml).

135

 

 

 

 

 

 

 

 

 

 

 

 

2. a. RF €24; m.» 8.9-8 3. .4. .4. E 82.6
9. v.8». 6.2.3 mm. #73 E. .4. .4- m 229..
a: NB 8.1. 3 8.3V 3 .4. .4. m: 9.822
2 3% 6.83.8 3 3.8-0.8 2 +44.. M k c2952..
.4. .4- a
-4. .4. m
8 tom €1.19 3: 8. Fmé «.1 .4- .4- n $35835.
.4. -4- o
.4. .4. m
.4. .4. v
8 to: 43.9 3 5.8 0.2 H H M «:28.
.4. .4. F
58256.“ coumiaow
“swan. ~_E4o_aoon_ £4ctoE4mm 483843.:an £426.53 LQoonoinmo % mm c:
6% Emcee cmEax Nomcmcv 3303 80.8 {moms Econ. : oo

 

 

 

 

 

3903 3520:. cmEax

 

gamma mEEEmw ELmn.

 

 

136

 

 

83.589ch 2:300 new £3.31 Em... macaw

a...» 03¢...

OVERALL DISCUSSION AND CONCLUSIONS

The five major sections of this dissertation included the Literature Review:
Drivers of Campylobacter and Salmonella Infections: Known and Suspected
(Chapter 1),an analysis of the Epidemiology of Campylobacteriosis and
Salmonellosis in Michigan (Chapter 2), the Evaluation of Seasonal and
Geographic Trends in Reporting (Chapter 3), the modeling of Environmental
Factors Influencing Incidence of Campylobacteriosis and Salmonellosis in
Michigan (Chapter 4), and Culture Isolation and Enumeration of Campylobacter
and Salmonella from Michigan Dairy Farm Environmental Soil and Water
(Chapter 5). These sections attempted to make logical connections incorporating
the model of thought (illustrated in figure 1.1) that environmental factors could be
influencing and possibly driving human rates of Campylobacter and Salmonella
infections in Michigan.

Published literature has shown the relationship of climate to health and
disease. Large climatic events affect global and local weather patterns resulting
in increased precipitation and runoff. Based on the type of land use there can be
significant amounts of runoff containing pathogens such as Campylobacter or
Salmonella. These pathogens are able to persist in natural waters, where
humans may be exposed and their survival is related to environmental
conditions. The source of the drinking water may also be a key factor in the
transmission of the disease. The literature review detailed the background

literature around the environment and links to Campylobacter and Salmonella

137

human infections, paying particular attention to routes of transmission that may
be influenced by the environment (Figure 1.1). This section concluded by
reporting on the series of recent articles evaluating the consistent associations of
climate and geography to variability in rates of these diseases.

To further establish the foundation for the necessity of evaluating
environmental factors, chapters 2 and 3 examine historical case data to describe
the epidemiological trends for Michigan (a state that has not been included in the
national extrapolations). The seasonal peak in reporting for both of these
infections had already been noted in the literature but, Chapter 3 goes on further
defining and describing the parameters associated with these reporting trends
(seasonality) and evaluating possible geographic relationships. It has been
suggested that in sporadic cases of Campylobacter and Salmonella infections
when evaluated by geography should occur randomly in space. However, the
reported clustering of cases that do not occur around outbreaks but are
sustained clusters suggests that factors other than food, possibly geographic or
environmental may be driving these clusters. Chapter 2 highlighted these areas
of consistently high and consistently low rates of disease and Chapter 3 began to
associate geography with parameters of the high reporting periods.

Chapter 4 continues to build on the previous work by incorporating
environmental factors that have been described in detail in chapter 1, where the
logic around these specific factors and their relation to variation in rates of
disease is explained. This study found that indeed, in Michigan some of the

variation in rates of campylobacteriosis and salmonellosis can be explained (in

138

various portions of the state) by combinations of these meteorological and
environmental variables.

The final step in logic was covered in chapter 5. After the detailed
evaluation of historical Campylobacter and Salmonella infection data and relating
the trends to environmental factors (Chapters 2-4), Chapter 5 targets at the
environmental prevalence. This chapter evaluated methods for the culture
isolation and enumeration of these bacteria from environmental soil and water
samples and found that future studies should employ alternative techniques
when performing environmental Campylobacter and Salmonella culture isolation
and enumeration due to competition from indigenous microorganisms in
environmental soil and water sample cultures.

This research has been conducted with the goal of describing particular
trends and geographical patterns with the hopes that this research can be used
to predict variability in incidence of campylobacteriosis and salmonellosis in
Michigan so that in future public health measures can be put into place to lessen
the transmission. This study also aimed to add to the literature by explaining and
filling a gap in the chain from animals to humans. By focusing on the
environmental connections that may explain some of the variation in human rates
of these diseases, the study was able to begin evaluating a missing link. The
food route has been explained, and the water route has also been explained, but
here the environmental contamination link between animals and the abundance
of the bacteria through the environment was explored. By modeling the

environmental associations with rates of disease and beginning to evaluate

139

environmental prevalence this has been an informative step towards
understanding trends in prevalence and will potentially provide insight into how to

lessen transmission between these animals and on to humans.

140

REFERENCES

Abulreesh, H.H., Paget, TA. and Goulder, R. (2005) Recovery of then'nophflic
campylobacters from pond water and sediment and the proplem of
interference by background bacteria in enrichment culture. Water Research.
39, 2877-2882.

Ahmed W, Sawant S, Huygens F, Goonetilleke A, Gardner T. 2009. Prevalence
and occurrence of zoonotic bacterial pathogens in surface waters determined
by quantitative PCR. Water Res. Apr 10. [Epub ahead of print]

Allos BM, Moore MR, Griffin PM, Tauxe RV. Surveillance for sporadic foodbome
disease in the 21st century: the FoodNet perspective. Clin Infect Dis
2004;38(Suppl 3):S1 15—20.

Allos, BM, 1997. Association between Campylobacterinfecection and Guillan-
Barré syndrome. Journal of Infectious Diseases. 176:S125-128.

Allos, BM, 2001. Campylobacterjejuni infections: update on emerging issues
and trends. Clinical Infectious Diseases. 32:1201-1206.

Altekruse, S.F., Stern, N.J., Fields, P.|., Swerdlow, D.L., 1999. Campylobacter
jejuni—an emerging foodbome pathogen. Emerging Infectious Diseases.
5:28-35.

American Ground Water Trust, 2005, <http:11www.agvvt.orngepticSystems.htm.>
Online.

Arshad MM, Wilkins MJ, Downes FP, Rahbar MH, Erskine RJ, Boulton ML,
Saeed AM. 2007. A registry-based study on the association between human
salmonellosis and routinely collected parameters in Michigan, 1995—2001.
Foodbome Pathog Dis. 4(1):16—25.

Arvanitidou, M., Kanellou, K. and Vagiona, DC. (2005) Diversity of Salmonella
spp. and fungi in northern Greek rivers and their correlation to fecal pollution
indicators. Environmental Research. 99, 278-284.

Ashbolt, N.J., 2004. Microbial contamination of drinking water and disease
outcomes in developing regions. Toxicology. 198(1-3):229-238.

Atwill, ER, 1995, Microbial Pathogens Excreted by livestock and Potentially
Transmitted to Humans through Water. Veterinary Medicine Teaching and
Research Center, University of California Davis.

Australian Food Safety Centre of Excellence. 2005. The effect of ambient
temperature on foodbome salmonellosis. Food safety and hygiene. March
2005.

141

Axelsson-Olsson, D., Waldenstrom, J., Broman, T., Olsen, 8., Holmberg, M.
2005. Protozoan Acanthamoeba polyphaga as a Potential Reservior for
Campylobacterjejuni. Applied and Environmental Microbiology. 71(2):987-
992.

Banrvick, R., Levy, D., Craun, 6., Beach, M., Calderon, R., 2000. Surveillance for
waterborne disease outbreaks. MMWR. 49(SSO4):1-35.

Bates and Phillips. 2005. Agricultural practices as a source of Campylobacter
spp. in river water. J of Environ Health Research. 4(1):17—23.

Bentham G, Langford IH. 1995. Climate change and the incidence of food
poisoning in England and Wales. lntemational Journal of Biometeorology.
39(2):81-86.

Bi P, Cameron AS, Zhang Y, Parton KA. 2008. Weather and notified
Campylobacter infections in temperate and sub-tropical regions of Australia:
an ecological study. J Infect. 57(4):317-23

Blaser, M.J., Cravens, J., Powers, B.W., Laforce, F.M., Wang, W.L., 1979.
Campylobacter enteritis associated with unpasturized milk, Short
communication. The American Journal of Medicine, 67(4):715-718.

Blaser, M.J., Taylor, D.N., Feldman, RA, 1983. Epidemiology of Campylobacter
jejuni Infections. Epidemiology Review. 5:157-176.

Buchanan, R.L., Whiting, RC, 1998. Risk Assessment: A Means for Linking
HACCP Plans and Public Health. Journal of Food Prot. 61 :1531-1534.

Buswell, C.M., Herlihy, Y.M., Lawrence, L.M., McGuiggan, J.T.M., Marsh, P.D.,
Keevil, C.W., Leach, SA, 1998. Extended Survival and Persistence of
Campylobacter spp. in Water and Aquatic Biofilms and their Detection by
Immunoflourescent-Antibody and -rRNA Staining. Applied and
Environmental Microbiology. 64, 733-741.

Butzler. 1984. Campylobacter infection in man and animals. P 6. Boca Raton
Florida, CRC Press.

Buzby, J.C., Roberts, T., 1997. Estimated Annual Cost of Campylobacter-
Associated Guillain-Barre Syndrome. USDNERS report, Washington DC.

Carrique-Mas, J.J., Barnes, 8., McLaren, l., and Davies, R. (2009) Comparison
of Three Plating Media for the Isolation of Salmonella from Poultry
Environmental Samples in Great Britan Using ISO 6579:2002 (Annex D). J
Appl Microbiol. May 20 [Epub ahead of print]

142

CDA (Communicable diseases Australia). 2005. National Notifiable Diseases
Surveillance System. <
http:11www1.health.gov.aulcda1$ource1Rpt_2_sel.cfm> online.

Centers for Disease Control (CDC). Preliminary FoodNet data on the incidence
of foodbome illnesses—selected sites, United States, 2001. MMWR Morb
Mortal Wkly Rep 2002; 51 (1 5):325-329.

CDC. Preliminary FoodNet Data on the Incidence of Foodbome Illnesses --
Selected Sites, United States, 2002. MMWR 2003;52(15):340-343.

CDC. Preliminary FoodNet data on the incidence of infection with pathogens
transmitted commonly through food—10 sites, United States, 2004. MMWR
Morbidity & Mortality Weekly Report 54(14):352-6, 2005.

CDC. Preliminary FoodNet data on the incidence of foodbome illnesses-—
selected sites, United States, 1999. MMWR 2000;49:201-5.

CDC. Preliminary FoodNet data on the incidence of foodbome illnesses-
selected sites, United States, 2000. MMWR Morb Mortal Wkly Rep 2001;
50(13):241-246

CDC, 2006. Preliminary FoodNet Data on the Incidence of Infection with
Pathogens Transmitted Commonly Through Food -- 10 States, 2005.
MMWR 2006; 55(14);392-395

CDC, 2004. Preliminary Food Net Data on the Incidence of Infection with
Pathogens Transmitted Commonly Through Food -- Selected Sites, United
States, 2003, MMWR. 53(16):338-343.

CDC, 2004. Campylobacter Infections.
Http:llwww.cdc.gov1ncididldbmdldiseaseinfolcampylobacter_t.htrn. [online]

CDC (FoodNet), 2005. http:/1www.cdc.gov1foodnetfindex.htm [online].

CDC (Center for Disease Control), 1991. Campylobacter enteritis—New Zeland,
1990, MMWR. 40(7):116—117,123.

Chen, Z., Grasby,S.E., Osadetz, KG, 2004. Relation between climate variability
and ground water levels in the upper carbonate aquifer, southern Manitoba
Canada. Journal of Hydrology, 290, 43-62.

Coker, A.O., Isokpehi, R.D., Thomas, B.N., Amisu, K.O., Obi, CL, 2002. Human
Campylobacteriosis in Developing Countries. Emerging Infectious Diseases.
8(3):237-243.

Cressie, N., 1993. Statistics for spatial data. IMley-Interscience, New York, N.Y.

143

Crowther, J., Kay, 0., Wyer, MD, 2002. Faecal-indicator concentrations in
waters draining lowland pastoral catchments in the UK: relationships with
land use and farming practices. Water Research. 36 (7), 1725-1734.

Cummings KJ, Wamick LD, Alexander KA, Cripps CJ, Grohn YT, McDonough
PL, Nydam DV, Reed KE. 2009. The incidence of salmonellosis among dairy
herds in the northeastern United States. J Dairy Sci. 92(8):3766-74.

Curriero, Patz, Rose, and Lele. 2001. The association between extreme
precipitation and waterborne disease outbreaks in the United States, 1948-
1994. Am J of Public Health. 91(8):1 194M 199.

Danyluk MD, Nozawa-lnoue M, Hristova KR, Scow KM, Lampinen B, Harris LJ.
2008. Survival and growth of Salmonella Enteritidis PT 30 in almond orchard
soils. J Appl Microbiol. 104(5): 1 391-9.

Droppo IG, Liss SN, Williams D, Nelson T, Jaskot C, Trapp B. 2009. Dynamic
existence of waterborne pathogens within river sediment compartments.
Implications for water quality regulatory affairs. Environ Sci Technol.
43(6):1737-43.

D'Souza RM, Becker NG, Hall G, Moodie KB. 2004. Does ambient temperature
affect foodbome disease? Epidemiology. 15(1):86-92.

Dunlop J. 1985. Salmonella infection in the community. Public Health. 99(2):79-
84.

EL-Askary, H., Sarkar, S., Chiu, L., Kafatos, M., EI—Ghazawi, T., 2004. Rain
gauge derived precipitation variability over Virginia and its relation with the El
Niho Southern Oscillation. Advances in Space Research 33 (3), 338-342.

EPA (Environmental Protection Agency) 2005. Private drinking water wells.
<http:11www.epa.govlsafewater/privatewells>. Online.

EPA (Environmental Protection Agency) 2005. Private drinking water wells.
http:11www.epa.gov/safewaterlprivatewells [online]

EPA, 1998. List of Impaired Waters. The National Water Quality Inventory Report
to Congress.

Environmental protection agency (EPA), 1996. List of Impaired Waters. The
National Water Quality Inventory Report to Congress.

EPA, 2000. List of Impaired Waters. The National Water Quality Inventory Report
to Congress.

EPA, 2002. List of Impaired Waters. The National Water Quality Inventory Report
to Congress.

144

1‘: _'a . .‘T‘Jrr

 

 

Exner, M., Hartemann, P., Kistemann, T., 2001, Hygiene and health—the need
for a holistic approach, Association for Professionals in Infection Control and
Epidemiology, Inc. 29, 228-231.

Eyles, R., Niyogi, D., Townsend, C., Benwell, G. and Weinstein, P. (2003) Spatial
and Temporal patterns of Campylobacter Contamination Underlying Public
health Risks in the Taieri River, New Zealand. Journal of Environmental
Quality. 32, 1820-1 829.

Fahey, T., Morgan, 0., Gunneburg, C., Adak, G.K., Majid, F., Kaczmarski, E.,
1995. An outbreak of Campylobacterjejuni enteritis associated with failed
milk pasteurization. Journal of Infection, 31(2):137-143.

Farzan A, Friendship RM, Cook A, Pollari F. 2009. Occurrence of Salmonella,
Campylobacter, Yersinia enterocolitica, Escherichia coli 0157 and Listeria
monocytogenes in Swine. Zoonoses Public Health. Jul 23. [Epub ahead of
pnnfl

Ferguson, 0., de Roda Husman, A.M., Altavilla, N., Deere, D., Ashbolt, N., 2003.
Fate and Transport of Surface Water Pathogens in Watersheds. Critical
Reviews in Environmental Science and Technology. 33 (3), 299-361.

Fleury M, Charron DF, Holt JD, Allen OB, Maarouf AR. 2006. A time series
analysis of the relationship of ambient temperature and common bacterial
enteric infections in two Canadian provinces. Int J Biometeorol. 50(6):385—91.

F DSCG (Foodbome Disease Strategy Consultative Group). 2002. Incidence and
trends in foodbome disease in the UK in the years 2001-2002.
<www.foodstandards.gov.uklmultimedialpdfsl‘foodbomediseasetrendspdf.>.
Online.

Fossler, C.P., Wells, S.J., Kaneene, J.B., Ruegg, P.L., Wamick, L.D., Eberly,
L.E., Godden, S.M., Halbert, L.W., Campbell, A.M., Bolin, CA. and Geiger
Zwald, AM. (2005) Cattle and environmental sample-level factors associated
with the presence of Salmonella in a muIti-state study of conventional and
organic dairy farms. Preventive Veterinary Medicine. 67, 39—53.

Friedman, C., Neimann, J., Wegener, H., Tauxe, R., 2000. Epidemiology of
Campylobacterjejuni infections in the United States and Other Industrialized
nations, p. 121-138. In: Campylobacter, 2"“ Ed. (Nachamkin, l. and Blaser,
M. eds), ASM Press, Washington, DC.

Fullerton K; OzFoodNet Working Group, 2008. Monitoring the incidence and
causes of diseases potentially transmitted by food in Australia: annual report
of the OzFoodNet Network, 2007. Commun Dis Intell. 32(4):400-24.

Gaertner, J.P., Garres, T., Becker, J.C., Jimmenez, M.L., Forstner, MR, and
Hahn, D. (2009) Temporal Analyses of Salmonellae in headwater Spring

145

Ecosystem Reveals the Effects of Precipitation and Runoff Events. J Water
Health. 7(1), 115-121.

Gillespie, I., O’Brien, 8., Frost, J., Adak, G., Horby, P., Swan, A., Painter, M.,
Neil, K., 2002. A case-case Comparison of Campylobacter coli and
Campylobacterjejuni Infection: A tool for generating Hypotheses. Emerging
Infectious Diseases. 8:937-942.

Gil Prieto R, Alejandra CG, Meca AA, Barrera VH, de Miguel AG. 2009.
Epidemiology of hospital-treated Salmonella infection; data from a national
cohort over a ten-year period. J Infect. 58(3):175-81.

Girschick HJ, Guilhenne L, lnman RD, Latsch K, Rihl M, Sharer Y, Shoenfeld Y,
ZeidIer H, Arienti S, Doria A. 2008. Bacterial triggers and autoimmune
rheumatic diseases. Clin Exp Rheumatol. 26(1 Suppl 48):S12-7

Glantz, 2001. Currents of Change, Cambridge university pres82001 NY, NY pg
1-43.

Graczyk, T., Evans, 8., Shift, 0., Kaneman, H., Patz, J., 2000. Environmental
and Geographical Factors Contributing to Watershed Contamination with
Cryptosporidium parvum Oocysts. Academic Press, Environmental Research
Section A. 82:263-271.

Hadjinicolaou, A.V., Demetriou, V.L., Emmanuel, M.A., Kakoyiannis, OK, and
Kostrikis, LG. (2009), Molecular Beacon-based real-time PCR detection of
Primary Isolates of Salmonella Typhimurium and Salmonella Enteritidis in
environmental and clinical samples. BMC Microbial. 19:9:97.

Hafliger, D., Huber, Ph., Luthy, J. 1999. Outbreak of viral gastroenteritis due to
sewage-contaminated drinking water. International Journal of Food
Microbiology. 54, 123-126.

Hald, B., Skovgard, H., Bang, 0., Pedersen, K., Dybdahl, J., Jespersen, J.,
Madsen, M. 2004. Flies and Campylobacter Infection of Broiler Flocks.
Emerging Infectious Diseases. 10(8):1490-1492.

Haley BJ, Cole DJ, Lipp EK. 2009. Distribution, diversity, and seasonality of
waterborne salmonellae in a rural watershed. Appl Environ Microbiol.
75(5):1248-55.

Hardnett FP, Hoekstra RM, Kennedy M, Charles L, Angulo FJ; Emerging
Infections Program FoodNet Working Group. Epidemiologic issues in study
design and data analysis related to FoodNet activities. Clin Infect Dis
2004;38(Suppl 3):StZ1—8.

146

Hariharan, H., Murphy, G.A, Kempp, l., 2004. Campylobacterjejuni: Public health
Hazards and potential control methods in poultry: a review. Vet Med.
49(11):441-446.

Hitt, K.J., 1991. WSGSWR Landuse Digital map file (GA) of major land uses in
the United States. USGS, Reston, Va.

Hopkins RS, Olmsted RN. 1985. Campylobacterjejuni infection in Colorado:
unexplained excess of cases in males. Public Health Rep. 100(3):333-6.

lnterlandi, S.J., Crockett, CS, 2003. Recent water quality trends in the Schuylkill
River, Pennsylvania, USA: a preliminary assessment of the relative
influences of climate, river discharge and suburban development. Water
Research. 37 (8), 1737-1748.

Jepsen MR, Simonsen J, Ethelberg S. 2009. Spatio-temporal cluster analysis of
the incidence of Campylobacter cases and patients with general diarrhea in a
Danish county, 1995-2004. lntJ Health Geogr. 20;8:11.

Johnson-lfearulundu, Y. and Kaneene, J.B. (1997) Relationship between soil
type and Mycobacterium paratuderculosis. J. Am. Vet. Med. Assn. 210(12),
1-6.

Jones, K., 2001. The Campylobacter Conundrum. Trends in Microbiology.
9(8):365-366.

Jones, K., 2001. Campylobacters in water, sewage and the Environment. Journal
of Applied Microbiology. 90,688-79s.

Kalra V, Chaudhry R, Dua T, Dhawan B, Sahu JK, Mridula B. 2009. Association
of Campylobacterjejuni infection with childhood Guillain-Barré syndrome: a
case-control study. J Child Neurol. 24(6):664-8.

Kambole, MS, 2003. Managing the water quality of Kafue River. Physics and
Chemistry of the Earth. 28, 1105-1109.

Kapperud, G., Skjerve, E., Bean, N., Ostrodd, S., Lassen, J., 1992. Risk factors
for sporadic Campylobacter infections; result of case-control study in
southeastern Norway. J Clin Microbiol. 30(12):3117-3121.

Keener, K.M., Bashor, M.P., Curtis, P.A., Sheldon, B.W., Kathariou, S., 2004.
Comprehensive Review of Campylobacter and Poultry Processing.
Comprehensive Reviews in Food Science and Food Safety. 3:105—116.

Kelsey, H., Porter, 0., Scott, G., Neet, M., White, 0., 2004. Using geographic
information systems and regression analysis to evaluate relationships
between land use and fecal coliform bacterial pollution. Journal of
experimental marine biology and ecology. 298:197-209.

147

Ketley, J.M., 1997. Pathogenesis of enteric infection by Campylobacter.
Microbiology. 143:5-21.

Kirian ML, Meregillano G, Gennette D, Weintraub JM. 2008. Multi-jurisdictional
investigation of interactive fountain-associated cryptosporidiosis and
salmonellosis outbreaks. Epidemiol Infect. 136(1 1):1547-51.

Kist, M. 1985. The historical background of Campylobacter infection: new
aspects p. 23-27. In A.D. Pearson, M.B. Skirrow, H. Lior, and B. Rowe (ed),
Campylobacter III. Public Health Laboratory Service, London, United
Kingdom.

Kistemann, T., Claben, T., Koch, C., Dangendorf, F ., Fischeder, R., Gebel, J.,
Vacata, V., Exner, M., 2002, Microbial load of Drinking Water Reservoir
Tributaries during extreme Rainfall and Runoff. Applied and Environmental
Microbiology. 68, 2188-2197.

Kovats, R., Bouma, M., Hajat, S., Worrall, E., Haines, A., 2003. El Nino and
Health. Lancet. 362: 481-489.

Kovats RS, Edwards SJ, Hajat S, Armstrong BG, Ebi KL, Menne B. 2004. The
effect of temperature on food poisoning: a time—series analysis of
salmonellosis in ten European countries. Epidemiol Infect. 132(3):443-53.

Kovats RS, Edwards SJ, Charron D, Cowden J, D'Souza RM, Ebi KL, Gauci C,
Gemer—Smidt P, Hajat S, Hales S, Hernandez Pezzi G, Kriz B, Kutsar K,
McKeown P, Mellou K, Menne B, O'Brien S, van Pelt W, Schmid H. 2005.
Climate variability and campylobacter infection: an international study. Int J
Biometeorol. 49(4):207-14.

Kussin, M., Nuortin, J., Hannien, L., Koskela, M., Jussila, V., Kala, E., Miettien, I.,
Ruutu, P., 2005. A large outbreak of campylobacteriosis associated with
municipal water supply in Finland. Epidemiol Infect. 133(4):593—601.

Leeming, R., Bate, N., Hewlett, R., Nichols, PD, 1998. Discriminating faecal
pollution: a case study of stonnwater entering Port Phillip Bay, Australia.
Water Science and Technology. 38 (10), 15-22.

Lefebvre SL, Reid-Smith R, Boeriin P, Weese JS. 2008. Evaluation of the risks of
shedding Salmonellae and other potential pathogens by therapy dogs fed
raw diets in Ontario and Alberta. Zoonoses Public Health. 55(8-10):470—80.

Lestari SI, Han F, Wang F, Ge B. 2009. Prevalence and antimicrobial resistance
of Salmonella serovars in conventional and organic chickens from Louisiana
retail stores. J Food Prot. 72(6):1165-72

Lindback, J., Svensson, A., 2001. Campylobacter infections in Sweden-A
statistical analysis of temporal and spatial distributions of notified sporadic

148

campylobacter infections. Mathematical Statistics Stockholm University,
Research report 2001 :4. ISSN 1650-0377.

Lipp. EK., Schmidt, N., Luther, M.E., Rose, J.B., 2001, Determining the effects of
El Nino-Southem Oscillation events on Coastal Water Quality. Estuaries. 24.
491-497.

Lipp, E.K., Huq, A., Colwell, RR, 2002. Effects of Global Infectious Disease: the
Cholera Model. Clinical Microbiology Reviews. 15, 757-770.

Louis, V., Gillespie, I., O’Brien, 8., Russek-Cohen, E., Pearson, A., Colwell, R.,
2006. Temperature-Driven Campylobacter Seasonality in England and
Wales. Applied and Environmental Microbiology. 71(1): 85—92.

Lucey, B., Crowley, D., Moloney, P., Cryan, B., Daley, M., O’Halloran, F.,
Threlfall, E.J., Fanning, S., 2000. Campylobacter spp. of human and Animal
Origin. Emerging Infectious Diseases. 8(1)

Madden, R.H., Murray, KA., and Gilmour, A. (2007) Carriage of four bacterial
pathogens by beef cattle in Northern Ireland at Time of Slaughter. Lett Appl
Microbiol. 44(2), 1 15—1 19.

Mallin, M., Williams, K., Esham, C., Lowe, P., 2000. Effect of Human
development on bacteriological water quality in coastal watersheds.
Ecological Applications. 10(4):1047-1056.

Mannion C, Lynch PB, Egan J, Leonard FC. 2007. Seasonal effects on the
survival characteristics of Salmonella Typhimurium and Salmonella Derby in
pig slurry during storage. J Appl Microbiol. 103(5):1386-92.

McEwen, S.A., Martin S.W., Clarke, R.C., Tamblyn, SE. and McDennott, J.J.
(1988) The Prevalence, Incidence, Geographic Distribution, Antimocrobial
Sensitivity Patterns and Plasmid Profiles of milk Filter Salmonella Isolates
from Ontario Dairy Farms. Can J Vet Res. 52, 18-22.

Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Sharpiro, C.,
Griffin, P.M., Tauxe, R.V., Food-Related Illness and Death in the United
States, Emerging Infectious Diseases. 5(5):607-825.

Medema, G.J., Teunis, P.F.M., Havelaar, A.H., Haas, ON, 1996. Assessment of
the dose-response relationship of Campylobacter Jejuni. International
Journal of Food Microbiology, 30(1-2):101-11 1.

Miller, G., Dunn, G., Smith-Palmer, A., Ogden, I., Strachan, N., 2004. Human
camoylobacteriosis in Scotland: seasonality, reagional trends and bursts of
infection. Epidemiology Infection. 132:585—593.

149

Nachamkin, I., 2002. Chronic effects of Campylobacter Infection. Microbes and
Infection. 4(2002):399-403.

NASS, 2006. Dairy, Milk cows: Number of operations, by size group, 2001-2005.
Michigan Agricultural Statistics. 61.

Naumova EN. 2006. Mystery of seasonality: getting the rhythm of nature. J
Public Health Policy.;27(1):2-12

Naumova EN, Jagai JS, Matyas B, DeMaria A Jr, MacNeiII IB, Griffiths JK. 2007.
Seasonality in six enterically transmitted diseases and ambient temperature.
Epidemiol Infect. 135(2):281-92.

Naumova EN and MacNeill B. (2009) Time—distributed effect of exposure and
infectious outbreaks. Environmetn'cs. 20(3):235-248.

Nichols, G., 2005. Fly Transmission of Campylobacter. EID. 11(3);March 2005.

NOAA (National Oceanic and Atmospheric Association), 1997. Our Changing
Climate. Reports to the Nation on Our Changing Planet. 4, 1-28.

Nylen, G., Dunstan, F., Palmer, 8., Andersson, Y., Bager, F., Cowden, J., Feierl,
G., Galloway, Y., Kapperud, G., Megraud, F., Molbak, K., Petersen, L.,
Ruutu, P., 2002, The seasonal distribution of Campylobacter infection in nine
European countries and New Zealand. Epidemiology infections. 128(3):383-
390.

Oberhelman RA, Taylor DN. Campylobacter infections in developing countries.
In: Nachamkin I, Blaser MJ, editors. Campylobacter, 2nd edition.
Washington: American Society for Microbiology; 2000. p.139—53.

Patrick, M., Christiansen, L., Waino, M., Ethelburg, S., Madsen, H., Wegener, H.,
2004. Effects of climate on incidence of Campylobacter spp. in humans and
prevalence in broiler flocks in Denmark. App Environ microbiol. 70(12):7474-
80.

Padungton, P., Kaneene, J.B., 2003. Campylobacter spp. in human, chickens,
pigs and their antimicrobial resistance. Journal of Veterinary Medical
Science. 65(2):161-170.

Palmer, S.R., Gulley, P.R., White, J.M., Pearson, A.O., Suckling, W.G., Jones,
D.M., Rawes, J.C.L., Penner, J.L., 1983. Water-Bome Outbreak of
Campylobacter Gastroenteritis.. The Lancet. 321 (831 9):287—290.

Patz, J.A., 2001. Public Health Risk Assessment Linked to Climatic and
Ecological Change. Human and Ecological Risk Assessment. 7 (5), 1317-
1 327.

150

Patz, J., 2002. A human disease indicatior for the effects of recent global climate
change. PNAS. 99(20): 12506-12508.

Perez-Perez, G.I., Blaser, M.J., 2005. Campylobacter and Helicobacter.
Medmicro. Ch 23. <http:119sbs.utmb.edu1microbooklch023.htm.>. Online.

Philander, 1990. E1 Niflo, La Nifla, and the Southern Oscillation. Academic Press,
San Diego.

Potter, R., Kaneene, J., Gardiner, J., 2002. A comparison of Campylobacter
jejuni enteritis incidence rates in high and low poultry density counties:
Michigan 1992-1999. Vector Borne Zoonitic Dis. 2(3):137-143.

Rees, R., 1996. Under the weather: climate and disease, 1700-1900 History
Today. 46:35-42.

Rose JB, Epstein PR, Lipp EK, Sherman BH, Bernard SM, Patz JA. 2001.
Climate variability and change in the United States: potential impacts on
water- and foodbome diseases caused by microbiologic agents. Environ
Health Perspect. 109(Suppl2):211-21.

Rosef, O., Kapperrud, G., 1983. House Flies (Musca domestica) as Possible
Vectors of Campylobacter fetus subsp. Jejuni. Applied and Environmental
Microbiology. Febz381-383.

Sahlstrom, L., Aspan, A., Bagge, E., Danielsson-Tham, M., Albihn, A., 2004.
Bacterial pathogen Incidences in Sludge from Swedish Sewage treatment
Plants. Water Research. 38, 1989-1994.

Samuelsson, Susanne. 2004. Zoonotic Enteric Infections 2003. Epi-News. No 9.

Samuel, M. Reilly, K., Shallow, 8., Marcus, R., Segler, S., McGivem, T.,
Kassenborg, H., Hollinger, K., Vose, D., Bartholomew, M., Kennedy, M.,
Vugia, D., 2000. Burden of Campylobacter Infection in United States and
Declining Trend in California, FoodNet 1996-1998. 2"d International
Conference on emerging Infectious Diseases, Atlanta, GA, July.

Samuel, M.C., Vugia, D.J., Shallow, 8., Marcus, R.,Segler, S., McGivem, T.,
Kassenborg,H., Reilly, K., Kennedy, M., Angulo, F., Tauxe, R.V., 2004.
Epidemiology of Sporadic Campylobacter Infection in the United States and
Declining Trend in Incidence, Food Net 1996—1999. CID, 38, $165-$174.

Savill, M.G., Hudson, J.A., Ball, A., Kiena, J.D., Scholes, P., Whyte, R.J.,
McCormick, RE. and Janovic, D. (2001) Enumeration of Campylobacter in
New Zealand recreational and drinking waters. Journal of Applied
Microbiology. 91, 38-46.

151

Sayah, R.S., Kaneene, J.B., Johnson, Y. and Miller R. (2004) Patterns of
Antimicrobial Resistance Observed in Escherichia coli Isolates Obtained from
Domestic- and Wild-Animal Fecal Samples, Human Septage, and Surface
Water. Applied and Environmental Microbiology. 71(3), 1349-1404

Schmidt, N., Lipp. E.K., Rose, J. B., Luther,M. E., 2001. Enso Influences on
Seasonal Rainfall and River discharge in Florida. American Meterological
Society. 14, 615-628.

Sebald, M., and M. Veron. 1963. DNA base content in the classification of
vibrios. Ann. Inst. Pasteur 105: 897-910.

Shrestha, MN, 2003. Spatially Distributed Hydrological Modelling considering
Land-use changes using Remote Sensing and GIS. Map Asia Conference
2003. GIS developmentnet, 1-8.

Skelly, C., Weinstein, R, 2003. Pathogen Survival Trajectories: And eco-
environmental approach to the modeling of Human Campylobacteriosis
Ecology. Environmental Health Perspectives. 111, 19-28.

Skirrow, MB, 1991, Epidemiology of Campylobacter enteritis. International
Journal of Food Microbiology, 12(1):9-16.

Sopwith W, Birtles A, Matthews M, Fox A, Gee S, Painter M, Regan M, Syed Q,
Bolton E. 2008. Identification of potential environmentally adapted
Campylobacterjejuni strain, United Kingdom. Emerg Infect Dis. 14(11):1769-
73.

Stanley, K., Jones, K., 2003. Cattle and Sheep Farms as Reservoirs of
Campylobacter. Journal of Applied Microbiology. 94,1048—1138.

Takahanshi, M., Koga, M., Yokoyama, K., Yuki, N., 2005, Epidemiology of
Campylobacterjejuni Isolated from Patents with Guillian-Barre and Fisher
Syndromes in Japan. Journal of Clinical Microbiology. 43(1):335—339.

Tarn CC, Rodrigues LC, O‘Brien SJ, Hajat S. 2006. Temperature dependence of
reported Campylobacter infection in England, 1989-1999. Epidemiol Infect.
134(1):119-25.

Tauxe RV, Nelson JM, Angulo FJ. (2004) Active public health surveillance for
foodbome infections: The FoodNet Program in the United States.
[www.cdc.gov/enterics/publications/358-PaduaAbstract.pdt].

Tsubo, M., Walker, 3., Hensley, M., 2005. Quantifying risk for water harvesting
under semi-arid conditions: Part I. Rainfall intensity generation. Agricultural
Water Management. 76(2):77-93.

152

US Department of Agriculture,(USDA) Food Safety and Inspection Service.
Progress report on Salmonella testing of raw meat and poultry products,
1998—2005. Washington, DC: US Department of Agriculture; 2006. Available
at
http://www.fsis.usda.gov/sciencelprogress_report_salmonelIa_testing/index.a

sp.

USDA, Food Safety and Inspection Service. Salmonella verification sample result
reporting: agency policy and use in public health protection. Fed Regist
2006;71:9772-7. Available at
http:/1www.fsis.usda.gov10PPDE/rdale RPubs/04-026N.pdf.

USDA (US_Food_and _Drug_Administration), 2003, posting date. Foodbome
Pathogenic Microorganisms and Natural Toxins Handbook
httpzllvm.cfsan.fda.govl~mow1chap4.html [Online]

US Department of Health and Human Services. Healthy people 2010
(conference ed, in 2 vols). Washington, DC: US Department of Health and

Human Services; 2000.

US. Census Bureau, 1990. Georgia-DP-5 Housing Characteristics.
<Http:/1factfinder.census.gov/servletlQTTable?_bm=n&_lang=en&qr_name=

DEC_1990_ST>. Online.

USDA (US Food and Drug Administration), 2003, posting date. Foodbome
Pathogenic Microorganisms and Natural Toxins Handbook.
<http:11vm.cfsan.fda.govl~mow/chap4.html>. Online.

USGS, 2005. The Water Cycle-Surface Runoff.
<http:119a.water.usgs.gov1edu/watercyclerunoff.html>. Online.

Van Gilder, T., Vugia, D., Fiorentino, T., Segler, 8., Carter, M., Smith, K., Morse,
D., Cassidy, M., Angulo, F ., 1999. Decline in Salmonella and Campylobacter
but not E. coli O157 isolation rates in FoodNet sites: Farm, food, or
fluctuation? 37th Annual Meeting of the Infectious Disease Society of
America, Philadelphia, PA, November.

Vucic S, Kieman MC, Comblath DR. 2009. Guillain—Barré syndrome: an update.
J Clin Neurosci. 16(6):733-41.

Waage, A.S., Vardund, T., Lund, V. and Kapperund, G. (1999) Detection of low
numbers of Salmonella in environmental water, sewage and food samples by
a nested polymerase chain reaction assay. Journal of Applied Microbiology.
87(3), 418-428.

WHO (World Health Organization), 2003. Quantifying selected major risks to
health. The world health report 2002. World health Organization. Geneva.

(Chapter 4).

153

Wilson, D.J., Gabriel, E., Leatherbarrow, A.J., Cheesbrough, J., Gee, 8., Bolton,
E., Fox, A., Feamhead, P., Hart, C.A., and Diggle, FJ. (2008) Tracing the
Source of Campylobacteriosis. PLoS Genetics. 4(9):e1000203.

Wysok B, Uradzihski J. 2009. Campylobacter spp.—a significant microbiological
hazard in food. I. Characteristics of Campylobacter species, infection source,
epidemiology. Pol J Vet Sci. 12(1):141-8.

Younus M, Hartwick E, Siddiqi AA, Wilkins M, Davies HD, Rahbar M, Funk J,
Saeed M. 2007. The role of neighborhood level socioeconomic
characteristics in Salmonella infections in Michigan (1997-2007): assessment
using geographic information system. Int J Health Geogr. 19:6:56.

Zhang Y, Bi P, and Hiller J. 2008. Climate variations and salmonellosis
transmission in Adelaide, South Australia: a comparison between regression
models. lntemational Joumal of Biometeorology. 52(3)179—187.

Zimmer, 1999. The El Nino Factor- El nino weather pattern may be beginning of
new weather patterns. Discover 1.

154

         

     

         

    

VERSITY LIBR

I Hlllllll H I II

063 554

ARIES

 

 

 

 

 

MICIH'IIINIIIISIIIIIIIIIIW
3 1293 O3