the role of drinking water as a source of …

THE ROLE OF DRINKING WATER

AS A SOURCE OF TRANSMISSION

OF ANTIMICROBIAL RESISTANT ESCHERICHIA COLI

by

Brenda Lee Coleman

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Public Health Sciences

University of Toronto

© Copyright by Brenda Lee Coleman, 2008

ii Drinking water as a source of antimicrobial resistance

The Role of Drinking Water as a Source of Transmission of Antimicrobial Resistant

Escherichia Coli

Doctor of Philosophy, 2008

Brenda Lee Coleman Department of Public Health Sciences

University of Toronto

Abstract

Antimicrobial resistance is a serious threat to the treatment of infectious diseases and a leading public

health concern of the 21st century. Antimicrobial resistant E. coli has been detected in many places

including domestic livestock, humans, food items, surface water, and drinking water. Although the use

of antibiotics is a major contributor to the emergence of resistance, the ingestion of water

contaminated with antimicrobial resistant bacteria may contribute to the prevalence of antimicrobial

resistance in humans. Purpose: The objectives of the research were to determine the prevalence of

human faecal carriage of antimicrobial resistant E. coli in people residing in southern Ontario who used

private water sources and to determine whether the use of water contaminated with antimicrobial

resistant E. coli was associated with human carriage of same. Method: The study population consisted

of people living in Ontario households that submitted water samples from private water sources for

bacteriological testing between May 2005 and September 2006. Respondents completed a questionnaire

and submitted a self-collected rectal swab. Results: Antimicrobial resistant E. coli were detected in

the swabs of 41% of the 699 respondents, with 28% resistant to ampicillin, 25% to tetracycline, and 24%

to sulfisoxazole, and 29% that were multi-drug resistant. Subjects from households using untreated

water contaminated with antimicrobial resistant E. coli were 40% more likely to carry antimicrobial

resistant E. coli in their gastrointestinal system than people from households using uncontaminated

water, even after adjusting for the effect of other variables. Implications: The association between

the consumption of water contaminated with antimicrobial resistant E. coli and human carriage of

resistant E. coli highlights the ongoing risks associated with water contamination and antimicrobial

resistance in Ontario. The high rates of resistant E. coli in healthy non-institutional persons provides

further rationale for public health programs to reduce antibiotic use in medicine and agriculture.

Drinking water as a source of antimicrobial resistance iii

Acknowledgements

A sincere thank you to my thesis supervisor, Dr. Allison McGeer. Allison has been a terrific

advisor as well as a true inspiration to me. She is a brilliant researcher who effectively manages, much

to my continued admiration, her “day job” as the Director of Infection Control at Mount Sinai Hospital,

numerous research projects, publications, teaching commitments, speaking engagements, and

committee work - along with a family! She has the ability to see both the “big picture” and the little

details. Allison, I can only gawk in amazement and reiterate my thanks for mentoring me through my

PhD and introducing me to the world of research in infectious diseases. I have learned so much, but

have so much more to learn.

I can’t adequately thank the other members of my committee, Dr. Susan Bondy, Dr. Iris

Gutmanis, and Dr. Ian Johnson. The fact that you took the time from your already overloaded

schedules to help me through the seemingly never-ending process of completing a PhD thesis is very

much appreciated. I can’t imagine how you managed to read and re-read the never-ending litany of

drafts of the thesis. You have been supportive and effective supervisors every step of the way…from

hypothesis generation and questionnaire development through the protocol defence and on to the

analysis and thesis-writing.

Ian, right from our first meeting, you asked the “so what” questions and so often directed me

back to the public health and policy pieces of the research. Thank you for picking up on things I was

taking for granted and for keeping the focus on public health practice. Sue, you were invaluable for

your insight into the methods, statistics, and policy as well as the requirements of the public health

department. Thank you for nudging me in the right direction of continued learning. Iris, thank you for

your tireless support, your expertise in both public health and methods, and your ability to see the

veins on the leaves of the trees in that forest! It was your belief in my abilities that started this

adventure. Thank you.

Edward Woods, my soul mate. I would not have had the self-confidence to apply for my

undergraduate degree without your unfailing support and belief in me. It has been a long road and you

have been there every step of the way….even, and often most importantly, for the wobbly steps!

Thank you too for being Mr. Mom and the cruise director, for being my sounding board, and for reading

so many papers and drafts of theses that I’m sure you’ve lost count. I know that I have. Thanks, hon.

Hey kids, thank you for your indulgence and support through the oh-so-many years of school.

Jeff and Kristy, I appreciate that you read and commented on my dissertation. Kristy, you were a huge

help working in the lab, making calls and doing interviews but I think it was your unfailing belief in me

that I appreciate the most. And, just think…if my old brain could complete a dissertation, imagine what

you can do! Jamie and Natalie, your interest and concern has been appreciated and I thank you for

your patience and understanding through the years. All four of you have grown and matured into

beautiful, intelligent young adults. I love you and I’m proud of all of you.

Thanks Mom and Dad! Thanks for everything, starting with…thanks for teaching me that “things

half-done are never done right” and for helping me through my post-divorce return-to-school by being

iv Drinking water as a source of antimicrobial resistance

my day care centre, the best grandparents in the world, my coffee house of choice, and the loving

support system that I always knew would be there for me. Thanks…for everything from rescuing me

when my car broke down to being interviewed numerous times when I trained new staff to

understanding that the long stretches between visits were not because I didn’t want to see you but

because I was trying to finish this work before yet another birthday passed.

Research is rarely completed in isolation and this project relied on the good-will and

participation of a number of people across Canada. A sincere thank you to all of the subjects who

participated in the study. Without you, this research would never have happened. Your willingness to

answer the survey questions as well as provide a sample for laboratory analysis was very much

appreciated and I want you to know that the information we learn from this study is already being

disseminated and will be used to promote health.

A big thank you to the staff at the public health laboratories in Ontario and Alberta that

provided thousands of E. coli slants for susceptibility testing - with a special thanks to Nancy Latimer

for also answering all of my questions about water testing in Ontario. I would also like to thank the

staff at the Safe Water Unit for working with us to contact the people who were selected for interview,

with a special thanks to Dr. Zsuzsanna Rajda for being the go-between for this study. Thanks too, to

the staff in the Ontario public health units and Alberta regional health authorities for making this

research successful.

And last, but not least, thank you to Dr. Marie Louie, Dr. Marina Salvadori, Dr. Scott McEwen,

and all of the researchers involved in the studies that made this part of the project possible. I would

also like to thank Caroline Guénette, Bryanne Crago, Kristen McLeod, Danielle Daignault, the staff at

the provincial laboratory in Calgary and the laboratory for Foodborne Zoonosis, the students who

worked on associated projects, and the staff of the Ontario Well Water Study. Your participation and

hard work are appreciated and without you, my piece of the puzzle would not exist.

Drinking water as a source of antimicrobial resistance v

Table of Contents

Abstract................................................................................................................... ii Acknowledgements .................................................................................................... iii Table of Contents ....................................................................................................... v List of Tables ............................................................................................................vi List of Figures............................................................................................................vi List of Appendices.......................................................................................................vi An Introduction to Antimicrobial Resistance........................................................................ 1

1.1 Background ...................................................................................................... 1 1.2 Antimicrobial Resistance in E. coli .......................................................................... 3

1.2.1 The Epidemiology of E. coli ............................................................................. 3 1.2.2 Epidemiology of Antimicrobial Resistance in E. coli ................................................ 6

Mechanisms of antimicrobial resistance in E. coli ....................................................... 6 Prevalence of antimicrobial resistant E. coli in Canadian studies..................................... 8 Prevalence across time....................................................................................... 9 Prevalence across place...................................................................................... 9 Prevalence across persons .................................................................................. 10 Factors affecting prevalence ............................................................................... 11

1.3 Objectives and Hypotheses .................................................................................. 21 Research Methods ...................................................................................................... 23

2.1 Study Population .............................................................................................. 23 2.1.1 Inclusion and Exclusion Criteria.................................................................... 25

2.1.2 Subject Recruitment .................................................................................... 26 Approvals and funding ....................................................................................... 27

2.2 Data Collection ................................................................................................ 27 2.2.1 London and Hamilton Regions ......................................................................... 27 2.2.2 Ottawa, Kingston, Peterborough, Orillia, and Toronto Regions.................................. 28 2.2.3 Questionnaires ........................................................................................... 29 2.2.4 Laboratory Testing of Water and Rectal Swab Samples ........................................... 30

2.3 Data Analyses .................................................................................................. 31 2.3.1 Objective 1: Prevalence of Antimicrobial Resistant E. coli ....................................... 31 2.3.2 Objective 2: Association Between Water Consumption and Human Carriage of Antimicrobial Resistant E. coli.................................................................................................. 33

Model-building strategies. .................................................................................. 35 Focal relationship analysis. ................................................................................. 36

2.3.3 Student’s Role............................................................................................ 37 Results ................................................................................................................... 39

3.1 Characteristics of Water Samples, Households, and Respondents..................................... 39 3.1.1 Water samples ........................................................................................... 39 3.1.2 Participation.............................................................................................. 39 3.1.3 Households................................................................................................ 45 3.1.4 Respondents .............................................................................................. 48

3.2 Prevalence of Carriage of Antimicrobial Resistant E. coli .............................................. 51 3.2.1 Ampicillin Resistant E. coli Prevalence .............................................................. 54

3.3 Association of Human Carriage and Consumption of Contaminated Water .......................... 56 Discussion................................................................................................................ 62

4.1 Prevalence of Resistant E. coli .............................................................................. 62 4.1.1 Ampicillin Resistant E. coli............................................................................. 62 4.1.2 Antimicrobial Resistance to Other Agents ........................................................... 64

4.2 Association of Human Carriage and Consumption of Contaminated Water .......................... 66 4.2.1 Strengths and Limitations of the Study .............................................................. 67

4.3 Conclusions ..................................................................................................... 72 4.3.1 Contaminated water .................................................................................... 72 4.3.2 Prevalence of Carriage of Resistant E. coli.......................................................... 73 4.3.2 Next steps................................................................................................. 75

References .............................................................................................................. 77 Appendices .............................................................................................................. 97

vi Drinking water as a source of antimicrobial resistance

List of Tables

Table 1.1 Studies of E. coli contamination of private water sources, Ontario 5 Table 2.1 Sampling and data collection for antimicrobial-resistance studies . . . . . 24 Table 2.2 Susceptibility break-points for resistance screening and NARMS panels for

enteric bacteria, by antibiotic . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Table 3.1 Details of household recruitment and participation for surveillance project and case-control study . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

Table 3.2 Proportion of E. coli-positive water samples with antimicrobial resistant E. coli, by antibiotic and class of antibiotic . . . . . . . . . . . . . . . . . . . .

41

Table 3.3 Details of subject participation in prevalence study . . . . . . . . . . . . 41 Table 3.4 Comparison of households participating in prevalence and case-control

studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Table 3.5 Comparison of participants who completed personal questionnaires and submitted a swab and subjects who completed questionnaires only . . . .

44

Table 3.6 Descriptive statistics of participating households . . . . . . . . . . . . . . . . 47 Table 3.7 Descriptive statistics of participants . . . . . . . . . . . . . . . . . . . . . . . . 50 Table 3.8 Proportion of human rectal swabs with antimicrobial resistant E. Coli, by

antibiotic and class of antibiotic . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Table 3.9 Intra-class resistance of antimicrobial resistant E. coli isolates from human rectal swabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53

Table 3.10 Inter-class resistance of antimicrobial resistant E. coli isolates from human rectal swabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

Table 3.11 Proportion of human rectal swabs with ampicillin resistant E. coli, by selected covariates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

Table 3.12 Bivariate associations between carriage of antimicrobial resistant E. coli and covariates. Poisson regression . . . . . . . . . . . . . . . . . . . . . . . . .

59

Table 3.13 Multivariable model of association between carriage of antimicrobial resistant E. coli, use of water contaminated with antimicrobial resistant E. coli, and covariates. Poisson regression . . . . . . . . . . . . . . . . . . . .

61

Table 4.1 Comparison of rates of antimicrobial resistance in E. coli from Canadian studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66

List of Figures

Figure 1.1 Possible routes of transmission of E. coli to humans . . . . . . . . . . . . . . 13 Figure 2.1 The theorized relationship between human carriage of E. coli, use of

contaminated and untreated water, and other study variables used for multivariable model building . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

List of Appendices

Appendix A Studies comparing E. Coli antimicrobial resistance rates: Canadian subjects 97 Appendix B Studies comparing E. coli antimicrobial resistance rates for human faecal

carriage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Appendix C Sample size calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Appendix D Ontario Well Water Study, information sheet . . . . . . . . . . . . . . . . . . . . 105 Appendix E Telephone script for contact from Safe Water Unit . . . . . . . . . . . . . . . . . 106 Appendix F Telephone script for initial study contact . . . . . . . . . . . . . . . . . . . . . . . 109 Appendix G Ethics board approvals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Appendix H Information and consent form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Appendix I Household questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Appendix J Personal questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Appendix K Rectal swab collection information sheet . . . . . . . . . . . . . . . . . . . . . . . 137 Appendix L Variables derived from personal and household questionnaires . . . . . . . . . 138

Background 1

Chapter 1

An Introduction to Antimicrobial Resistance

1.1 Background

Resistance to penicillin was first noted in bacteria just four years after mass production of this

antibiotic began in 1943 and antimicrobial resistance has since been detected in bacteria, viruses,

fungi, and parasites (1;2). Since the 1940s the use of antibiotics has dramatically reduced death and

morbidity due to infectious diseases caused by bacteria. Yet the very use of antibiotics promotes

resistance through selective pressure. Antimicrobial resistance is considered one of the most serious

threats to the treatment of infectious diseases and is one of the leading public health concerns of the

21st century (2;3).

Antimicrobial resistance occurs when the organism is able to survive and reproduce in the

presence of the concentrations of an agent that can be achieved in target tissues/sites during therapy

(4;5). Bacteria are particularly predisposed to developing resistance due to the speed with which they

reproduce: a single bacterium can multiply within hours with several generations of bacteria being

created within days. They are also proficient at exchanging the genetic material that confers

resistance. Resistance to antibiotics has become a concern for a wide variety of bacterial species, both

pathogenic and commensal. It has been an issue for Neisseria gonorrhoea and Mycobacterium

tuberculosis for decades (6). More recently, methicillin-resistant Staphylococcus aureus (MRSA),

penicillin-resistant Streptococcus pneumoniae, and vancomycin-resistant enterococci (VRE) have been

troubling the health care systems of the world, particularly in acute and long-term care facilities (7;8).

There is a growing concern about resistance in enteric pathogens including Salmonella species, Shigella

species, Vibrio cholerae, Campylobacter species, and Escherichia coli (E. coli) (6). Antimicrobial

resistance in these enteric pathogens is of particular interest in developing countries where diarrhoeal

diseases are a leading cause of illness and death, but are also of importance in developed nations as a

reservoir for the transmission of resistance.

The health consequences of antimicrobial resistance include increased morbidity and mortality

due to delays in starting effective treatment or outright treatment failure (9). People infected with

antimicrobial resistant species of bacteria are twice as likely to be hospitalized, up to twice as likely to

die (10;11), and have as much as double the length of hospital stay as patients with susceptible strains

2 Drinking water as a source of antimicrobial resistance

of bacteria (12). Other consequences include increased health care costs due to hospitalization or

prolonged stays (10;13), and additional costs for: supplementary laboratory investigations, isolation of

patients, special cleaning of health care facilities, and the use of more potent and expensive

antibacterial agents (12;14;15). It is estimated that resistant infections add $200-700 million annually

to the cost of health care in Canada (16;17).

In addition to the health and health care system consequences, no new class of antibiotic has

been introduced for several years thereby reducing the number of agents capable of combating

resistant organisms (3;18-20). Antibiotics are the foundation upon which the treatment of infectious

diseases builds and antibiotic resistance threatens not only the current management of pathogenic

bacteria, but the long-term efficacy of antimicrobial agents (21). The continuing emergence of

antimicrobial resistant Gram-negative pathogens, in particular, has not been matched by the

development of new classes of antimicrobial agents (22).

There is a general consensus that the emergence of antimicrobial resistance is largely due to

the use of antimicrobial agents (23-27). This occurs through a process called selective pressure in which

bacteria with mutations capable of surviving in the presence of the antimicrobial agent(s) persist and

pass these changes on to their offspring (28-32). The use of antimicrobial agents may promote selective

pressure in a second way: by eliminating susceptible bacteria, even of a different species, thus

permitting the resistant bacteria to multiply in the absence of competition (22;24;33;34).

The dissemination of antimicrobial resistance in bacteria is largely due to the genetic exchange

of resistance genes. Resistance genes are carried on chromosomes which enable vertical spread from

mother to daughter. However, they can also be transferred horizontally between bacteria and bacterial

species if the resistance genes are located on small segments of deoxyribonucleic acid (DNA) called

plasmids, on segments of plasmids called transposons, and on integrons which are found within

transposons and plasmids and in bacterial chromosomes (15;24;35). A resistance gene on the

chromosome of one strain of bacteria would have limited dissemination. However, resistance genes

that evolve on plasmids or transposons can be transferred to other bacteria, including other strains of

the same bacteria and other species of bacteria, making the dissemination much more efficient (24).

Resistance to one antibiotic can be conferred by one or more genes (e.g. tetA, sul1) that may

be localized on plasmids, transposons, or integrons (36). In comparison, other chromosomal mutations

Background 3

(e.g. mar) are capable of confering resistance to a large number of antibiotics (37;38). Resistance

genes can be spread to other bacteria by transformation, conjugation, or transduction. Transformation

is the uptake and incorporation of naked DNA and incorporation into the genome or plasmids.

Conjugation involves the donation of a copy of the plasmid (which may contain genes for resistance to

several antibiotics) to a recipient cell. Transduction is the transfer of resistance genes from one

bacterium to another through viruses (39;40). The transfer of antimicrobial resistance from one

bacterium to another and the subsequent vertical transfer to daughter bacteria is core to the

persistence and dissemination of antimicrobial resistance, including multi-drug resisitance (40).

Multiresistance integrons are important components of antimicrobial resistance in Gram-

negative bacteria. Integrons are capable of acquiring, incorporating, and expressing the responses of

gene cassettes that encode for resistance against antibiotics, heavy metals, and detergents (41;42).

One integron may contain many gene cassettes with different resistance genes allowing one bacterium

to resist the effects of antimicrobial agents with varying mechanisms of action.

Although the emergence and persistence of resistance is largely due to selective pressure,

transmission of resistant bacteria and/or resistance genes contributes greatly to the overall prevalence

(26;33;43). Studies report that being hospitalized, attending a day care centre, or living with another

person who is colonized with antimicrobial resistant bacteria are risk factors for colonization or

infection with resistant bacteria suggesting transmission between humans (44-46). It is also possible

that people can become colonized after ingesting resistant strains of bacteria from food or water (47-

49). This thesis examines the prevalence of antimicrobial resistant E. coli in the human gastrointestinal

tract and the role of drinking water in the transmission of antimicrobial resistant E. coli to humans.

1.2 Antimicrobial Resistance in E. coli

1.2.1 The Descriptive Epidemiology of E. coli

Escherichia coli are Gram-negative bacteria of the Enterobacteriaceae genus. They are the

predominant facultative organism (capable of living in both aerobic and anaerobic environments) of the

human gastrointestinal system. There are 700 or more strains of E. coli bacteria, the majority of which

do not cause disease if they remain in the intestinal lumen. In fact, many live commensally with

humans suppressing the growth of pathogenic species of bacteria and producing vitamin K and B


complex vitamins that are necessary for health. However, some strains of E. coli are pathogenic and all

strains will cause disease if they infect areas that are normally sterile (e.g. the urinary tract) (50).

Escherichia coli are clinically significant bacteria, not only for their ability to cause

gastrointestinal disease but because they are the leading cause of Gram-negative infections.

Escherichia coli are responsible for 85-95% of urinary tract infections, 60-70% of hospital-acquired

pneumonia, and 17-37% of nosocomial bacteraemia in Europe and North America (51). Escherichia coli

also cause a high proportion of neonatal meningitis cases and abdominal, pelvic, and surgical site

infections (51). Between 2000 and 2002, E. coli comprised 13% of pathogens isolated from clinical

specimens from intensive care unit patients in Canada making it the third most common bacterial

pathogen and the most common Gram-negative pathogen (52).

Diarrhoeal infections caused by E. coli are treated with fluid and electrolyte replacement with

the role of antibiotics being uncertain. Extraintestinal infections, however, are treated with

antibiotics. The growing prevalence of antimicrobial resistance in E. coli is complicating the treatment

of all infections caused by these bacteria (53;54).

Escherichia coli are common in all warm-blooded animals meaning that humans, domestic

livestock, pets, wild animals, and birds are all reservoirs. Escherichia coli are spread by the faecal-oral

route. Transmission to humans from these reservoirs may occur in a variety ways including contact

with: other humans or animals, contaminated surfaces (e.g. door handles), manure or sewage, or

contaminated meat/poultry products, vegetables, raw milk, or untreated water (7;55-57).

Water is a known vehicle for the transmission of bacteria, including E. coli, in concentrations

large enough to cause illness in humans (58;59). In Canada and the United States, several large

outbreaks of E. coli O157:H7 have implicated drinking water as the source of infection, including a

large outbreak in Walkerton, Ontario in 2000 (60).

Escherichia coli are often used as an indicator of faecal contamination of drinking water.

Ontario microbiological drinking water standards for E. coli are set at zero colony forming units per 100

mL of water (61). In 2005-2006, E. coli were detected in only 20 of 20,000 (0.1%) routine post-

treatment bacteriological tests of municipal water systems (62). In comparison, 20 of 250 tests (8%) of

small water systemsi in Ontario (62) and 4% of water samples from private water supplies in southern

i Small water system is one that serves a designated facility (school, nursing home, day care).

Background 5

Ontario (63) were determined to have E. coli contamination in the same period. Table 1.1 outlines the

proportion of private water supplies contaminated by E. coli according to published articles and

reports.

Table 1.1 Studies of E. coli Contamination of Private Water Sources, Ontario Geographic area (reference)

Year(s) of

testing

Proportion of E. coli-positive

sources* Ontario (64) 1950-1954 15% Brantford (65) 1987 & 1989 2% Ontario (64) 1991-1992 17-24% Ontario (66) 1992 20% Middlesex (67) 1993 2% Halton (68) 2000 14-16% Kingston, Frontenac, Lennox, & Addington (69)

2000

16%

Haldimand-Norfolk (70) 2004 4-11% * Ranges are presented for results reported separately by time period or geographic region.

Four million or more Canadians rely on private water systems that are strictly the responsibility

of the residence owner (71). While private water sources serve only about 13% of the population in

Canada, they are implicated in about 20% of waterborne outbreaks (71).

In Ontario in 2006, about 80% of the 12.1 million residents were served by municipal water

systems with another 250,000 serviced by small water systems, leaving 1.5 to 2 million residents reliant

on private water sources (62;72;73). There are a variety of systems that supply water for people using

private sources including drilled or dug wells, sand or well points, cisterns, springs, and shore or lake

wells (71;74).

Although the province of Ontario provides bacteriological testing without direct cost to the

owners of private systems, many residents do not avail themselves of the service. A postal survey of

households using private water sources in Hamilton, Ontario determined that 28% of respondents were

not aware that the testing was provided without direct cost. Further, only 8% of households tested

their water at least three times per year as recommended in Ontario (75;76). Data from the Rapid Risk

Factor Surveillance System (2001-2005) determined that 61% of surveyed Ontario households with a

private water source had their water tested for bacteriological contamination at least once in the

previous year and 17% had tested it three or more times (63).


Bacteria in drinking water may be killed by boiling, ultraviolet light, chlorine or iodine, ceramic

or glass candle filters, or ozone. Reverse osmosis, carbon filters, and softeners are not effective

against bacterial contamination (77;78). Although 56% of Hamilton, Ontario-area households using

private water sources treated their drinking water, only 26% used a device capable of killing bacteria

(75). Similar results were reported in the Households and Environment Survey conducted in 2006. Of

the households in that study that did not primarily use bottled water, 25% treated their drinking water

with the goal of removing possible bacterial contamination (73). A survey of Halton, Ontario households

using private water sources determined that only 8% of households used chlorine or ultraviolet

treatments despite the finding that 14-16% of these water supplies were contaminated with E. coli

(68). Thus, even to this day, tens of thousands of Ontario residents are regularly exposed to E. coli

from their drinking water. These residents consistute a population in which we can assess the

relationship between the carriage of antimicrobial resistant E. coli and the use of water contaminated

with same.

1.2.2 Epidemiology of Antimicrobial Resistance in E. coli

Mechanisms of antimicrobial resistance in E. coli

Escherichia coli resist the effects of antibiotics in four ways. First, by producing antibiotic-

inactivating (or -modifying) enzymes, such as beta-lactamases, which deactivate the drug by cleaving

to the core beta-lactamase ring structure. Extended-spectrum beta-lactamases (ESBLs) mediate

resistance to penicillins and third-generation cephalosporins. Escherichia coli also resist the effects of

antibiotics by restricting the concentration of antibiotics within the cell by reducing membrane

permeability and resisting entry of a wide range of antibiotics. A third mechanism of resistance is

alteration of the antibiotic target site so the antibiotic is unable to bind properly. This type of

mechanism is used by some strains of E. coli to resist quinolones and macrolides. The fourth type of

resistance mechanism is to eliminate the antibiotic target site entirely, either through overproducing

the target enzyme or producing an alternative target enzyme (15;79). Some resistance mechanisms

confer resistance to only one class of antibiotics while other mechanisms confer resistance to many

classes of antibiotics (7). One strain of bacteria may possess several different types of resistance

mechanisms making it multi-drug resistant (79).

Background 7

Escherichia coli are often used in antimicrobial resistance studies because a) they are found in

high numbers in warm-blooded mammals, including humans; b) they are a common human pathogen;

c) resistance is found in both pathogenic and non-pathogenic strains; d) they have the ability to

transfer resistance between different strains of E. coli; and e) they have the ability to transfer

resistance between different strains and species of bacteria within the gastrointestinal tract (80;81).

Also, E. coli reside primarily in mammalian hosts, thus being subjected to the pressures of

antimicrobial use and other environmental pressures. This makes them an ideal agent for surveillance

and research into factors that may contribute to the selection and spread of resistant bacteria (6;82-

84). Further, these bacteria are abundant in the environment making them a predominant vehicle for

the transmission of resistance genes (35;55;85;86).

Antimicrobial resistance has been detected in a variety of E. coli strains including numerous

shiga toxin-producing E. coli (including O157:H7), enteropathogenic, enterohaemorrhagic,

enterotoxigenic, enteroaggregative, enteroinvasive, and others (87). Several researchers have reported

that strains of E. coli that are antimicrobial resistant are often, although not always, less virulent than

susceptible strains (88-92). This appears to be related to the particular strain of E. coli rather than

some relationship between the resistance and virulence factors per se, as non-B2 strains are much

more likely to exhibit multi-drug resistance than B2 strains. Although, Johnson et al. state that the

difference in antimicrobial resistance by virulence may be due to the tendency of non-B2 strains to

acquire and/or retain resistance genes, they hypothesize that the difference in the prevalence of

antimicrobial resistance by strain may be due to differential antibiotic selection pressure: B2 strains

are less likely to be commensal flora than non-B2 strains, thus less likely to be colonizing the

gastrointestinal system during antibiotic therapy as well as being less abundant in human

gastrointestinal systems (90).

There are no differences between pathogenic and non-pathogenic bacteria in the basic cellular

processes that affect antibacterial resistance (93) and the remaining text will not differentiate

between the two. Commensal flora have the same plasmids, transposons, and integrons as their

disease-producing brethren, thus acting as a reservoir for resistance genes for pathogenic bacteria (37).


Prevalence of antimicrobial resistant E. coli in Canadian studies

The direct comparison of studies is hampered by the wide variety of protocols, break-points,

and analytic techniques. The prevalence of antimicrobial resistance in E. coli also varies depending on

the source of the isolate. Bagger-Skjøt et al. (92) and others have reported that antimicrobial

resistance rates in clinical isolates were significantly higher than those from faecal isolates from non-

institutionalized subjects from the same country and in the same year. Although the rates of resistance

are higher from clinical isolates, the patterns of resistance are similar (i.e. higher in penicillins and

tetracycline than in fluoroquinolones).

In clinical isolates from Canadian patients, the reported rates of resistance range from 34-46%

to ampicillin (94;95), 15-19% to trimethoprim-sulfamethoxazole (52;95), and 27% to tetracycline (95)

(see Appendix A). Urinary tract infections are one of the most common infections in women (96) and E.

coli is the most common pathogen; responsible for 74-84% of urinary tract infections in Canada (97;98).

Rates of resistance detected in urinary tract infections caused by E. coli range from 30-45% for

ampicillin (99;100), 12-28% for trimethoprim-sulfamethoxazole (100;101), and 0-11% for nitrofurantoin

(98;99;102). Multi-drug resistance is also common in E. coli causing urinary tract infections.

One recently published study reported that 25% of urinary tract isolates from adult Canadian

women were resistant to two or more antibiotics (97). In another study, 20% of urinary isolates from

Canadian children were resistant to both ampicillin and trimethoprim-sulfamethoxazole (100).

Similarly, the most common pattern of resistance found in urinary tract isolates collected for the ECO-

SENS study was ampicillin and sulfamethoxazole (101). In a separate study, 36% of ampicillin resistant

E. coli isolates and 100% of ciprofloxacin-resistant isolates from Canadian outpatients were also

resistant to trimethoprim-sulfamethoxazole (98), the antibiotic recommended for empiric treatment of

uncomplicated urinary tract infections in adults (103).

Two studies, both conducted in the 1990s, used faecal samples from healthy non-

institutionalized Canadian subjects to determine the prevalence of antimicrobial resistant E. coli. The

prevalence of resistant E. coli in 154 volunteers from St. Johns, Newfoundland was highest for

amoxicillin (22%), followed by oxytetracycline (16%), and trimethoprim (10%) (104). Similar rates of

carriage of antimicrobial resistant E. coli were reported for 115 residents of pork producing farms in

Background 9

Ontario and British Columbia. In these subjects the rates of resistance were 16% for ampicillin, 24% to

tetracycline, and 5% to trimethoprim-sulfamethoxazole (105).

Prevalence across time

Antimicrobial resistance in E. coli has, in general, increased over time with older antimicrobial

agents like penicillin tending to have higher rates of resistance than newer drugs like cephalosporins

(44;106-108). Houndt and Ochman compared the antimicrobial resistance patterns of two E. coli

collections: one from the pre-antibiotic era (strains collected between 1885 and 1941) and one with

strains collected between 1972 and 1982. The authors compared the minimum inhibitory

concentrations for kanamycin, ampicillin, tetracycline, and chloramphenicol and determined that there

was a higher incidence of high-level antibiotic resistance for all antibiotics tested in the more recently

collected strains of the bacteria. For example, 4 of the 32 (13%) samples collected during the pre-

antibiotic era were resistant to kanamycin at ≥16 µg/mL compared to 21 of 72 (29%) of those collected

between 1972 and 1982. Similarly, 1 of 32 (3%) of pre-antibiotic era samples were resistant to

ampicillin while none were resistant to tetracycline or chloramphenicol. In the samples collected after

the introduction of antibiotics, 10 of 72 (14%) strains were resistant to ampicillin, 12 of 72 (17%) to

tetracycline, and 1 of 72 (1%) were resistant to chloramphenicol at ≥16 µg/mL (82).

The increased prevalence of antimicrobial resistant E. coli can be tracked in studies of urinary

tract infections as well (96;109). A study published in 1971 determined that only 19% of E. coli strains

from English women with urinary tract infections were resistant to one or more of the antibiotics

tested; 11% to tetracycline and 3% to ampicillin (110). In comparison, 39-45% of urinary tract isolates

collected between 2000 and 2004 from outpatients living in the United Kingdom were resistant to

amoxicillin (36-40%) or trimethoprim (7-17%) (111;112).

Prevalence across place

Rates of antimicrobial resistance in E. coli vary widely across the globe. In 2001, non-

susceptibility to ampicillin ranged from 20% to 40% in Canadian, Swedish and Japanese hospital isolates

to over 60% in isolates from Mexico, Poland, Israel, Spain, Turkey, Hong Kong, Philippines, South Africa,

and Taiwan (113). Similar results were reported for hospital isolates collected between 2004 and 2006,


with ampicillin resistant E. coli detected in 67-69% of isolates from Asia, the Pacific rim, and Latin

America compared with 53-59% of isolates from the United States, Canada, and Europe (114).

Appendix B outlines studies of faecal carriage of antimicrobial resistant E. coli. Similar patterns

of resistance are noted in faecal isolate studies as in those of hospital isolates: rates of ampicillin

resistance were lower in Canada, the United States, Japan, China, and Europe (12-53%) than in Mexico

(73-94%), several countries of Africa (49-89%), and Central and South America and the Phillipines (73-

97%). The prevalence of multi-drug resistance was also lower in developed countries than developing

ones (43;80;84;104;105;115-127).

Differences between countries may be related to the unique health care systems, regulations,

and guidelines regarding antimicrobial prescribing as well as differing standards and practice for

antimicrobial use in animal husbandry and agriculture in each country. Greater environmental exposure

to antimicrobial resistant bacteria and resistance genes through crowding, lack of sanitation, and

contamination of food and water may also be associated with higher rates of carriage of resistant

organisms.

Prevalence across persons

In studies of healthy children, a higher proportion of children under two years of age had

antibiotic resistant E. coli detected in faecal samples than children two to seven years old (45;84;115).

Similar findings were reported in studies of antimicrobial resistance in E. coli causing urinary tract

infections where higher rates of resistance were detected in isolates from younger children than older

ones (102;109;128). The frequent use of antimicrobial agents in younger children, their less than

fastidious hygienic practices, as well as high contact rates between children may account for the higher

rates in young children.

In adults, there is no clear pattern of differences in the prevalence of antimicrobial resistance

by the age of the person. Using isolates from faecal samples, no differences in colonization with

antimicrobial resistant E. coli have been reported for adults of varying ages (46;129;130). Meanwhile,

in one study using clinical isolates, higher rates of resistance were reported for isolates from older

adults for gentamicin, piperacillin, and tobramycin while isolates from younger adults were more likely

to be resistant to ciprofloxacin (113). In studies using isolates from urinary tract infections, older

Background 11

adults had higher rates of fluoroquinolone resistant E. coli (97;102;131) and trimethoprim-

sulphamethoxazole-resistant E. coli (97).

The prevalence of antimicrobial resistant E. coli is similar for males and females. Three studies

describing gastrointestinal carriage of antimicrobial resistant E. coli by the sex of the individual

reported no difference by sex (46;130;132) with only one reporting a higher rate in male subjects

(133). Several studies of antimicrobial resistant E. coli from urinary tract and clinical infection isolates

report higher rates of resistance for isolates from males than from females (102;109;113;131;134-136),

while others detected no difference (137), or higher rates in isolates from females (111).

Factors affecting prevalence

Factors affecting prevalence: Antibiotic use.

Antimicrobial resistance is generally agreed to be a result of the use of antimicrobial agents in

human medicine possibly contributed to by the use of antimicrobial agents in veterinary medicine,

agriculture, and aquaculture (1;25-27;35;138). At an ecological level, higher rates of resistance are

reported in countries with high consumption of antibiotics. Kenyan children’s isolates had a

significantly higher rate of resistance (68%) to trimethoprim-sulfamethoxazole than isolates from

Japanese children (2%) which was attributed to the antibiotic’s use in Kenya as a first-line drug for the

treatment of diarrhoea and enteric infections (139). Similar findings were reported from a study done

in Burkina Faso (west Africa), in which 80% of E. coli isolates from people with diarrhoea were resistant

to trimethoprim-sulfamethoxazole (140). In Mexico, high rates of resistance in diarrheogenic E. coli

were reported for trimethoprim-sulfamethoxazole (65%) and ampicillin (73%); antibiotics commonly

used for children with diarrhoea. In comparison, they detected no E. coli isolates resistant to

ciprofloxacin or cefotaxime, antibiotics rarely used in non-hospitalized children (141). High rates of

ampicillin resistant E. coli (90%) have also been detected in Mexican children without diarrhoea (119).

Children in these developing countries which have high rates of antibiotic use have correspondingly

high rates of antimicrobial resistant E. coli.

However, ecological comparisons between and within developed countries ascertained that

antibiotic consumption alone does not account for all of the differences in the prevalence of

antimicrobial resistant E. coli. For instance, although the overall consumption of antibiotics, defined as


the number of daily doses per 1,000 residents, was similar in Newfoundland and Athens, the rates of

resistance in faecal E. coli isolates from people in Athens was significantly higher (104). The authors

contend that the differences are related to the colonization pressure of living in more densely

populated regions.

Differences also exist within countries where antibiotic availability is uniform suggesting that

some mechanism beyond antibiotic use is responsible for at least part of the differences in prevalence.

Rates of antimicrobial resistant E. coli from urinary tract isolates varied significantly by region in

England (142) and by state in the United States (125). There are also studies that show high rates of

antibiotic resistance in populations with low rates of antibiotic use (117;120;143).

At an individual level, antibiotic use during hospitalization is believed to be responsible for the

higher prevalence of antimicrobial resistant E. coli reported at discharge when compared to rates on

admission for the same patients (132;144-148). In other studies, hospital patients and residents of long-

term care facilities were more likely to have antimicrobial resistant E. coli infections if they had

previously been treated with antibiotics (137;149). Antibiotic use in ill community-dwelling subjects is

also associated with increased antimicrobial resistance following use of antibiotics in many studies of

adults (30;46;122;150-156). In contrast to these findings, three studies of children attending day care

centres in the United States found no association between individual consumption of antibiotics and

the carriage of antimicrobial resistant E. coli (45;157;158). These findings indicate that although

antibiotic use is likely involved in the emergence of resistance, there are other factors involved in the

persistence and dissemination.

Escherichia coli are transmitted directly from person-to-person or animal-to-person and

indirectly via contaminated surfaces, food, or water (159-163). In studies where the infectious agent

has been identified as E. coli, the most common exposures identified include beef, unpasteurized dairy

products, swimming, untreated drinking water, fresh produce, and travel to a developing country

(160;164;165). Figure 1.1 depicts some routes of transmission of E. coli to susceptible hosts. Since

there is no evidence to suggest that antimicrobial resistance confers adaptive advantages or

disadvantages to E. coli (7;166;167), the study of the spread of antimicrobial resistant E. coli must

consider the same pathways of transmission.

Background 13

Figure 1.1 Possible Routes of Transmission of E. coli to Humans

Reservoir: Human or animal

Portal of exit: faecal

• Direct contact • Environmental

o Surfaces o Soil o Recreational water

• Foods o Meat & poultry products o Dairy products o Field & orchard crops o Fish & seafood o Drinking water*

Portal of entry: oral

Susceptible Host

* Focus of this research

Factors affecting prevalence: Person-to-person transmission.

The spread of E. coli from person-to-person is well-documented and, since most E. coli strains

are species-specific, humans serve as a main reservoir for the transmission to other humans

(33;161;162;168). As highlighted earlier, 16-22% of community-dwelling subjects in Canada were

colonized with penicillin-resistant E. coli in two recent studies (104;105). This represents a substantial

reservoir of antimicrobial resistant bacteria and resistance genes from which transmission to other

people is possible.

Infants’ intestinal tracts are sterile at birth but are colonized with bacteria, including E. coli,

within hours of birth (169). One of the most convincing pieces of evidence for person-to-person

transmission of commensal strains of antimicrobial resistant E. coli is the detection of resistant strains

within the faecal matter of infants that had never received antibiotics. Ampicillin resistant E. coli were


detected in the faecal matter of 8% of Swedish neonates and 28% of Turkish infants (8-10 weeks of age)

who had never taken an antibiotic in their lifetime (147;170).

Similarly, there was no difference in the rates of antimicrobial resistant E. coli from the stools

of young children from the United States, Venezuela, and China that had never had antibiotics

compared to those that had (127). Also, transmission from adult to child is one explanation for the

existence of fluoroquinolone and doxycycline resistant E. coli in young children and tetracycline

resistant E. coli in infants who are rarely, if ever, prescribed these antimicrobial agents (84;120;171-

173). Whether this transmission occurs directly from person-to-person or through less direct routes,

such as the ingestion of contaminated foods or water, was not determined.

Person-to-person transmission was likely responsible for the high rates (67%) of antimicrobial

resistance detected in faecal E. coli isolates of residents of a remote rural village in Bolivia where the

use of antimicrobials for humans was quite limited; only 7% reported antibiotic use in the previous 12

months compared to 35-40% of residents of the United States (46;158). Further, the use of antibiotics

for animal medicine was absent in this rural village (117). In a follow-up article, molecular

characterization of the isolates revealed a notable variety of resistance genes so the authors concluded

that the relatively high prevalence of resistance was likely due to the introduction of resistant strains

into the community by travellers and/or animals with subsequent horizontal gene transfer to and

between the village residents (174). Whether the transmission occurred directly between people or

indirectly through contaminated surfaces, foods, or water was not determined.

Some researchers believe that the use of antibiotics, in conjunction with high colonization

pressure (i.e. underlying rates of colonization with the organism) increases the probability of

colonization (116;175). Hospitals and long-term care facilities are sites that would have high

colonization pressure. Several studies have reported hospitalization as a significant predictor of

colonization with antimicrobial resistant E. coli, even after adjusting for antibiotic use

(44;122;155;176;177). The rates of antimicrobial resistance in faecal E. coli isolates are also higher in

residents of long-term care facilities than in community-dwelling subjects (80;149).

Day care centres are another site of enhanced transmission of enteropathogens, including E.

coli (53). Trimethoprim and ampicillin resistant E. coli have been reported to be more prevalent in

children within specific day care centres (157) and more prevalent in children attending day care

Background 15

centres compared to those not attending a centre (45). Further, person-to-person transmission was

implicated in an outbreak of antimicrobial resistant E. coli O26 in a day care centre in Japan (178).

Transmission of antimicrobial resistant E. coli likely occurs at the household level as well (179).

Having one household member with resistant E. coli was determined to be a risk factor for colonization

for other household members in several studies completed in developed countries (46;158;180-182).

Although people living within one household are more likely to share antimicrobial resistance patterns,

whether this is due to sharing a common source of resistant bacteria or resistance genes (i.e. food,

water, animals/pets) or is due to person-to-person transmission has not been determined.

Travel to areas with a high prevalence of human carriage, typically developing countries, has

been linked to colonization with antimicrobial resistant bacteria (15). Three studies of healthy college

students from the United States describe an increased prevalence of resistant E. coli after the students

spent several weeks in Mexico, a country in which a high proportion of residents are colonized

(152;183;184). Similarly, people who had travelled to a developing country in the previous year were

significantly more likely to carry antimicrobial resistant E. coli than other subjects (32% vs. 9%) in a

cross-sectional study completed in the United States (185). Exposure to the large reservoir of

antimicrobial resistant bacteria that exists in developing nations likely increases the probability of

carriage of antimicrobial resistant bacteria for visitors and residents alike. Whether higher prevalences

in people who have travelled is due to person-to-person or other form(s) of transmission is unknown.

Factors affecting prevalence: Animal-to-human transmission.

Although not all strains of E. coli are capable of colonizing the gastrointestinal tracts of both

humans and other animals, many are. Humans do become ill with pathogenic strains of E. coli whose

normal reservoir is animals after direct contact with the animals or their manure (161;163;186-188).

There are numerous studies highlighting the occurrence of antimicrobial resistant E. coli in

cattle, horses, pigs, sheep, goats, and companion animals such as dogs. Microbiological studies show

that several types of integrons (DNA element involved in antimicrobial resistance) are shared among E.

coli isolated from humans, dogs, and domestic livestock like cattle, pigs, and poultry, making the

transmission between species likely (189-191). There are also several reports of antimicrobial


resistance in E. coli O157:H7, an animal strain of E. coli, detected in human isolates (192-194). Thus, it

is possible that resistant bacteria are transmitted to humans from animals (26).

The Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) reported

that 33% of beef, 88% of swine, and 84% of chicken caecal samples from Canadian abattoirs were

positive for antimicrobial resistant E. coli in 2003 (195). Dogs, cats, and horses are also a potential

source of resistant bacteria and/or resistance genes (189;196-198). In one study, 25% of dogs in

breeding kennels and 12% of individually-housed dogs carried antimicrobial resistant E. coli (199).

People involved in livestock farming have been shown, in some studies, to have higher rates of

carriage of antimicrobial resistant bacteria than non-farming controls (130;200-203). However, other

studies have shown no association between farming and higher rates of carriage of antimicrobial

resistant E. coli (46;204).

A study of Canadian pig farmers determined that farmers who used in-feed antimicrobial

agent(s) were more likely to carry antimicrobial resistant E. coli than those who did not. The personal

use of antibiotics by the farmer and the number of hours he/she spent in the pig barn were positively

associated with human carriage, but, interestingly, having animals that carried antimicrobial resistant

E. coli was not (105). Similarly, pig farmers in the Netherlands had higher rates of faecal carriage of

antimicrobial resistant E. coli than urban-dwelling subjects (200). However, only 4% of E. coli from

faecal samples of farmers matched the resistance patterns for pigs from the same farm and laboratory

studies showed they had distinctly different plasmid DNA (205;206). Further, a study done in the United

States found no difference in human carriage of antimicrobial resistant E. coli for people working for

swine production facilities and people who did not (204).

A community-based study of healthy Dutch subjects six years of age and older determined that

people living on cattle farms had higher rates of carriage of tetracycline resistant E. coli than subjects

not involved in cattle farming. Of note, no differences were found for rates of carriage of ampicillin or

sulphamethoxazole resistant E. coli between the two groups (130).

Poultry workers in the United States were significantly more likely to carry E. coli resistant to

gentamicin, an antibiotic of limited human use, than community controls who were not involved with

poultry production (207). Similarly in Saudi Arabia, poultry farmers had a higher prevalence of

gentamicin resistant E. coli than hospitalized patients (38% versus 22%, respectively) (208). In Holland,

Background 17

poultry farmers had a higher prevalence of carriage of ciprofloxacin resistant E. coli (17%) than

subjects participating in other studies (<1% to 3%) in the 1990s (104;209;210).

The direct and prolonged exposure to animals, as experienced by farmers and farm-workers, is

linked with higher rates of carriage of antimicrobial resistant E. coli in some, but not all, studies. It

may be that the association is specific to the animal species, whether the animals are given antibiotics

as growth promoters, the antibiotic(s) under study, or other unmeasured variables such as the

consumption of water contaminated with bacteria originating from the animals.

Factors affecting prevalence: Environmental and foodborne transmission.

Food crops fertilized or otherwise contaminated by animal manure or human excrement are

another possible source of transmission of resistant bacteria and/or resistance genes to humans

(47;211-214). Antimicrobial resistant E. coli have been detected on a variety of foods in a number of

countries including Finland (214), Spain (198), and the United States (215;216). Escherichia coli were

detected on 9% of vegetables, fruit, and other foods purchased in stores in the United States and 27%

of these isolates were found to be antimicrobial resistant (211;217). Antimicrobial resistant E. coli has

also been detected in several different brands of ready-to-eat shrimp that was purchased in the United

States but caught and packaged in a variety of countries (218). The Canadian Integrated Program for

Antimicrobial Resistance Surveillance reported that 25% of raw beef, 60% of pork, and 70% of chicken

sampled from Canadian retail sources in 2003 were contaminated with antimicrobial resistant E. coli

(195). And, an Ontario-based study detected ampicillin resistance in 25% of E. coli isolates recovered

from the milk of cattle with mastitis (219).

Although antimicrobial resistant E. coli has been detected in numerous foods, only one study

has shown an association between consumption of meat and colonization with resistant E. coli. The

transfer of antimicrobial resistant E. coli from a raw chicken carcass to a human who handled, cooked,

and ate it was demonstrated in 1 of 14 trials within the same study, showing that transmission is

possible but infrequent (220). Several other studies from developed countries have found no

differences in the prevalence of carriage of antimicrobial resistant E. coli between people who

consume meat or poultry infrequently or not at all compared with those who eat it more often

(46;155;180;185). The lack of differentiation in the frequency of meat consumption may be a factor in


the lack of findings. Other reasons for the lack of findings include the fact that meat is one of myriad

possible sources of antimicrobial resistant bacteria and resistance genes. Additionally, foods are often

prepared (i.e. washed or cooked) which may dislodge and/or destroy the bacteria before consumption.

Factors affecting prevalence: Waterborne transmission.

Ground and surface water contaminated by animal or human waste is a reservoir of

antimicrobial resistant bacteria (7;221). Antimicrobial resistant E. coli have been detected in surface

waters in a number of countries and various water sources. Antimicrobial resistant E. coli, often multi-

drug resistant, has been detected in surface water (222-226), rivers and streams (227-234), farm

retention ponds (235), and ground well water (234).

Studies have detected antimicrobial resistant E. coli in the effluent of sewage treatment plants

as well as in the river water into which the effluent is discharged and in the sludge that is either

spread on agricultural land or dumped into landfill sites (212;230;236). A Canadian study completed in

1979 detected transferable drug resistance factors in 6-13% of faecal coliforms from the final effluent

of sewage treatment facilities in the Northwest Territories, Alberta, Manitoba, and Saskatchewan

(237). A study done in the late 1970s in Manitoba detected antibiotic resistant faecal coliforms in 47%

and 59% of sewage and river water samples, respectively (238). Antimicrobial resistant E. coli carrying

known resistance genes was detected in 11% of isolates from water sampled from the St. Clair and

Detroit rivers in Southern Ontario (239).

In Ontario, 95% of E. coli isolates from surface water sampled within a ten kilometre radius of

the Hamilton harbour in 2002 were resistant to one or more antibiotics (236) as were 81% of bacterial

isolates (not limited to E. coli) from sewage treatment effluent in Ottawa, Ontario (240). Escherichia

coli-positive beach water isolates submitted to the public health regional laboratories as part of the

requirements for surveillance of recreational water safety were tested for antimicrobial resistance as

part of the surveillance project connected with this research. Antimicrobial resistance was detected in

10% of isolates submitted from Alberta, 15% from Quebec, and 27% from Ontario public beaches for the

summer of 2004. Surface water is of importance not only for its recreational use but also because it

provides water for most large municipal drinking water systems in Ontario.

Background 19

Untreated water used for drinking could be a source of transmission for antimicrobial resistant

strains and genes (48;49;241). A study of drinking water from both surface and ground water sources in

the United States determined that 34% of the bacteria (not just E. coli) were multi-drug resistant in

1980 (242). In 1984, 13% of E. coli isolates from a public water supply in Connecticut (detected during a

period of elevated total coliform count) were antimicrobial resistant (243). In the same era, 10 of 18

wells tested in the United States contained antimicrobial resistant bacteria, with 16% of faecal

coliforms being multi-drug resistant (241). A study of 44 private groundwater supplies in West Virginia

determined that 46% of E. coli isolates (n=28) were resistant to one or more antibiotics (241).

Antimicrobial resistant E. coli was also isolated from drinking water in Montana with 70% and 55% of

isolates resistant to carbenicillin and tetracycline, respectively (244). In a pilot study for this project,

14% and 16% of E. coli-positive water samples submitted from Ontario and Alberta private water

sources were antimicrobial resistant. The highest rates of resistance were to tetracycline (11%),

sulfamethoxazole (6%), streptomycin (6%), and ampicillin (5%) (245). These findings emphasize the

potential magnitude of the problem of antimicrobial resistant E. coli in drinking water sources and

further support the role of water as a source of antimicrobial resistant bacteria.

One positive link between the consumption of antimicrobial resistant E. coli-contaminated

water and carriage of resistant E. coli comes from an animal study. One flock of grazing sheep in

Oregon had a significantly higher prevalence of multi-drug resistant E. coli than the other nine flocks

studied. Upon investigation, it was determined that their water source was contaminated with multi-

drug resistant E. coli likely originating from a nearby human septic system. No further study of the

issue was completed as this finding was coincidental to the original research (246). However, it raises

the possibility of the transmission of antimicrobial resistant E. coli to mammals through the ingestion

of contaminated drinking water.

A connection between drinking contaminated water and human carriage of antimicrobial

resistant E. coli was reported as a result of the investigation of three outbreaks that occurred on cruise

ships in 1997-1998. In all three outbreaks, the ingestion of the ship’s tap water or ice made from tap

water was significantly associated with diarrhoeal illness and implicated as the likely cause of the

outbreaks. Faecal samples collected from ill passengers were cultured for a variety of pathogenic

organisms. The investigators determined that 38% of the enterotoxigenic E. coli-positive isolates from


these ill passengers were resistant to three or more antibiotics (247). The authors hypothesized that

the ship’s potable water was contaminated with antimicrobial resistant E. coli thus infecting the

passengers. Unfortunately, by the time the investigations were conducted, it was not possible to

isolate E. coli from the water supplies of the cruise ships. Thus, no direct association could be made

between ingestion of water contaminated with antimicrobial resistant E. coli and infection with

resistant strain(s).

Shanahan et al. reported that there was no association between consuming water

contaminated with bacteria and human carriage of resistant bacteria in a 1992 study conducted with

healthy South Africans (248). The rural and urban dwelling subjects had similar rates of carriage of

resistant enterobacteria for three antibiotics (ampicillin, trimethoprim, and nalidixic acid) but rural

dwelling subjects had a higher prevalence of gentamicin (11% versus 4%) and chloramphenicol (60%

versus 46%) resistant enterobacteria. Although the authors reported no association between the use of

water contaminated with bacteria and the carriage of resistant enterobacteria, water was tested for

bacterial contamination at only one of two urban locations and two of four rural locations (a school and

a clinic). No tests were completed on the household water supplies, either rural or urban, and no tests

of antimicrobial resistance were completed on the bacteria that were isolated. Thus, it was not

possible to determine whether there was exposure to water contaminated with bacteria let alone

exposure to water contaminated with antimicrobial resistant bacteria. Therefore, we cannot support

nor refute the conclusion of no association between consumption of water contaminated with bacteria

and human carriage of resistant bacteria.

The prevalence of carriage of resistant aerobic faecal bacteria, not E. coli specifically, was

studied by Amyes et al. in 1989 in four villages in India (249). In this study, urban and rural subjects

were equally likely to carry bacteria resistant to ampicillin (98%), trimethoprim (98%), and

chloramphenicol, (97%), but urban dwelling subjects were more likely to be colonized with bacteria

resistant to nalidixic acid (35%) than the rural dwelling subjects (13%). The authors reported that the

rural village wells and the municipal drinking water both carried antimicrobial resistant bacteria

although no further details were given. The authors concluded that the high rates of human carriage of

resistant bacteria were due to high use of antibacterial drugs combined with the ingestion of faecal

bacteria from contaminated water supplies. However, without further details and with no statistical

Background 21

tests of association reported, the author’s conclusion that the ingestion of contaminated water was

associated with human carriage of resistant bacteria cannot be substantiated.

Despite the lack of more direct evidence, it is biologically feasible that humans will become

infected with antimicrobial resistant E. coli after consuming water contaminated with antimicrobial

resistant bacteria. In support of this idea is the fact that several outbreaks of antimicrobial resistant E.

coli have been reported indicating that the transmission of antimicrobial resistant bacteria is

occurring. These include an outbreak of antimicrobial resistant E. coli O26 in a day care nursery (178),

outbreaks of extended beta-lactamase producing E. coli in long-term care facilities (250) and acute

care hospitals (251), a foodborne outbreak caused by a resistant strain of E. coli O153:H45 (87).

Antimicrobial resistant E. coli have been found in numerous water sources (241;242;244) and E. coli is

known to be transmitted in large enough quantities in drinking water to cause disease (58;252-255).

Since there are no differences in the basic cellular processes that affect antimicrobial resistance in

pathogenic and non-pathogenic bacteria (93) and the acquisition of antimicrobial resistance does not

appear to confer adaptive advantage or disadvantages to E. coli (7;166), it is reasonable to assume that

antimicrobial resistant E. coli may be transmitted through water.

Whether the ingestion of water contaminated with antimicrobial resistant E. coli is associated

with human colonization with resistant strains of E. coli has not been definitively determined and was

the focus of this thesis.

1.3 Objectives and Hypotheses

The objectives of the research were to determine the prevalence of human faecal carriage of

antimicrobial resistant E. coli in people residing in southern Ontario who used private water sources

and whether the use of water contaminated with antimicrobial resistant E. coli was associated with

human carriage. We proposed to:

1. measure the prevalence of faecal carriage of antimicrobial resistant strains of E. coli in people

using private water sources in southern Ontario, by antimicrobial agent, with a focus on

ampicillin; and

2. determine whether the consumption of water contaminated with antimicrobial resistant E. coli

was associated with faecal carriage of antimicrobial resistant strains of E. coli.


We were most interested in examining the prevalence of resistance specific to ampicillin based

on the fact that penicillins are the most frequently prescribed antibacterial agents in Canada and

ampicillin is the first line choice for treatment of urinary tract infections due to E. coli when resistance

is below 20% (256). Ampicillin was also the only non-combination penicillin used in the 2002 or 2004

National Antimicrobial Resistance Monitoring System (NARMS) panels for enteric bacteria (the antibiotic

panels used to assess for antimicrobial resistance in this study) (257). Ampicillin is used for both

children and adults to treat urinary tract infections, bacterial respiratory tract infections,

gastrointestinal infections, bacterial meningitis, septicaemia, and endocarditis caused by E. coli,

enterococci, S. pneumoniae, H. influenza, P. mirabilis, and S. epidermis (258).

Research hypotheses were:

1. The prevalence of human carriage of ampicillin resistant E. coli in southern Ontario residents

using private water sources will be greater than or equal to the prevalence rate of 22%ii as

described in previous studies (104;105); and

2. The use of water contaminated with antimicrobial resistant E. coli will be associated with

faecal carriage of antimicrobial resistant E. coli in this population.

ii The hypothesized prevalence of ampicillin resistant E. coli was determined using an estimate from the paper by Bruinsma et al. (104). We hypothesized that the prevalence in Ontario, a province that is more densely populated and which is more intensively cultivated, 6-7 years after the collection of data in Newfoundland would be at least, if not greater than, the 22% prevalence of amoxicillin resistant E. coli (breakpoint concentration ≥25μg/mL) reported in that study. Another reason for using the 22% hypothesized prevalence was based on the Infectious Diseases Society of America recommendation that empirical treatment of urinary tract infections change when resistance to the recommended antibiotic reaches 20% (103).

Methods 23

Chapter 2

Research Methods

This study was one component of a three-part project starting with of a multi-province

surveillance project investigating the prevalence and geospatial distribution of antimicrobial resistant

E. coli in beach water samples submitted to provincial laboratories in Ontario, Alberta, and Quebec in

2004 and 2005, and private water samples submitted from participating public health laboratories in

Ontario and Alberta between April 2004 and September 2006.

In the second year of the surveillance project, April 2005 to September 2006, a case-control

study was conducted to determine the risk factors for contamination of private water sources with

antimicrobial resistant E. coli. The case-control study used water sample results from the surveillance

study to identify case and control water sources (households).

The research for this thesis was comprised of a cross-sectional human prevalence study of

antimicrobial resistant E. coli colonizing the gastrointestinal tract of healthy human adults and

adolescents. It used a convenience sample of people living in households who agreed to participate in

the case-control study (see Table 2.1). All consenting household members who were 12 years of age

and older and capable of completing the questionnaire in English were eligible for inclusion. This

sample was then used to determine whether consumption of water contaminated with antimicrobial

resistant E. coli was associated with faecal carriage of antimicrobial resistant E. coli. Sample sizes

calculated for the protocol are available in Appendix C.

2.1 Study Population

The sampling frame consisted of all suitableiii water samples submittediv for bacteriological

testing at the participating public health laboratories in Ontario between May 1, 2005 and September

30, 2006. All E. coli-positive water samples tested at the London and Hamilton laboratories and a

monthly quota of E. coli-positive samples, selected at random, from the Ottawa, Kingston,

Peterborough, Orillia, and Toronto laboratories were screened for antimicrobial resistance.

iii Water submitted from a private water source, tested within 48 hours of collection, containing adequate contact information, and sent in an approved bottle that is not broken, leaking, or frozen iv Submission of water samples from private water sources was voluntary in Ontario


Table 2.1 Sampling and Data Collection for Antimicrobial-Resistance Studies; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Water samples tested for bacteriological contamination by participating public health laboratories

⇓

Surveillance project (Water tests)

E. coli positive water samples (All from London & Hamilton; Random selection from Kingston, Orillia, Ottawa, Peterborough, & Toronto*)

⇓ ⇓ Antimicrobial resistant

E. coli Antimicrobial susceptible

E. coli

⇓

All non-repeated samples from

surveillance project

(Case)

Random sample from surveillance project

(A control)

Random sample of bacteria-free water

tests from lab data base

(B control) Eligible & agreed to share contact information

Site visit Telephone interview Interviewer called to explain study

& arrange site visit

Case-control study (Water source of households)

Household interview at site visit

Survey company explained study & completed household interview by

telephone ⇓ ⇓

Personal interview(s) by telephone Personal interview(s) & rectal swab(s) from eligible individuals in household Swab kits mailed & follow-up call

Prevalence study (Human subjects) Interviewer mailed swab(s) Subject mailed swab & consent form * A random selection of E. coli-positive water samples were submitted from these laboratories due to workload issues

Cases for the case-control study included households with E. coli-positive water samples that

were resistant to one or more of the antibiotic agents included in the National Antimicrobial Resistance

Monitoring System (NARMS) panel for enteric bacteria (Table 2.2). “A” controls were households

randomly selected from the E. coli-positive samples submitted to the study that were susceptible to all

antibiotics on the screening panel. “B” controls were households that were randomly selected from the

provincial database of all water submissions that tested negative for contamination with E. coli or non-

E. coli coliform. Both “A” and “B” controls were frequency matched by laboratory region from samples

submitted within one month of the date of the case submission, with an average of 1.3 controls per

case to ensure that an adequate number of controls were available.

Methods 25

Table 2.2 Susceptibility Breakpoints for Screening and NARMS Panels for Enteric Bacteria, by Antibiotic; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Screening1 NARMS Panels 2002

breakpoints (259)

2004 breakpoints

Antibiotic Concentration (260) Aminoglycoside Amikacin 64 μg/mL 64 μg/mL Gentamicin 16 μg/mL 16 μg/mL 8 μg/mL Kanamycin 64 μg/mL 64 μg/mL Streptomycin 64 μg/mL 64 μg/mL 32 μg/mL Beta-lactam Amoxicillin/clavulanic acid 32/16μg/mL 32/16μg/mL Ampicillin 32 μg/mL 32 μg/mL 8 μg/mL Cefoxitin 32 μg/mL 32 μg/mL Ceftiofur 8 μg/mL 8 μg/mL Ceftriaxone 64 μg/mL 64 μg/mL Fluoroquinolone Ciprofloxacin 4 μg/mL 4 μg/mL Nalidixic acid 32 μg/mL 32 μg/mL 4 μg/mL Sulphonamide Trimethoprim/sulphamethoxazole 4/76 μg/mL 4/76 μg/mL

--- Sulphamethoxazole 512 μg/mL 128 μg/mL 2 Sulfisoxazole --- 256μg/mL

--- Cephalothin 32 μg/mL 32 μg/mL Chloramphenicol 32 μg/mL 32 μg/mL Tetracycline 16 μg/mL 16 μg/mL 4 μg/mL

1 Screening of E. coli-positive samples was done using antibiotic concentrations lower than or equal to the NARMS breakpoints to reduce the number of samples requiring full NARMS panel analysis (samples susceptible on screening were not tested using NARMS) 2 Breakpoint concentration was 512 μg/mL in NARMS, but 256 μg/mL was the highest concentration available on the test plate

The case-control and human prevalence studies were initially conducted only in the London and

Hamilton public health laboratory regions of Ontario. It was expanded in January 2006 to include five

other laboratory regions: Ottawa, Kingston, Peterborough, Orillia, and Toronto. Escherichia coli-

positive water samples were already being submitted from these laboratory regions as part of the

surveillance project so that the same time period of water sample submission (May 2005 to September

2006) was used for all regions.

2.1.1 Inclusion and Exclusion Criteria

Inclusion in the case-control study was limited to households that a) submitted a suitable water

sample to a participating Ontario public health laboratory for bacteriological testing during the study


period, b) consented to have their contact information disclosed to the study group, c) resided on the

property from which the water was submitted, d) had at least one household member who was 18 years

of age or older, e) spoke English, and f) could be contacted by telephone.

Water samples were not eligible if they were from a) households that had already been

contacted by the study and submitted a subsequent water sample, b) sites outside the study area,

c) real estate brokers, d) commercial properties, or e) households from which the submitter had

moved. Also ineligible were households selected as “B” controls that had a water sample that tested

positive for bacterial (E. coli or other coliform) contamination within the previous twelve months.

All people living in case and control households that were 12 years and older were eligible for

inclusion in this human prevalence study. Only people who agreed to submit a rectal swab were

interviewed for the prevalence study.

2.1.2 Subject Recruitment

The case-control and prevalence studies were first introduced to potential participants with an

information sheet (Appendix D) that was mailed, along with the bacteriological test results, to all

households that submitted a water sample to a participating laboratory during the study period. In the

Hamilton and London areas, information sheets were also attached to water sample kits that were

distributed between May 2005 and September 2006.

Following public health laboratory testing, screening, and final resistance testing, case and “A”

control records were forwarded to the study assistant who was located at the Safe Water Unit of the

Ministry of Health and Long-Term Care. The study assistant matched the cases and “A” controls’

records (accession number, laboratory, and date of sample submission) with the provincial database

and then randomly selected “A” controls from within the matches. “B” controls were randomly

selected from results without bacterial contamination from within the provincial database. Households

selected for inclusion in the case-control study were telephoned by the study assistant.

Using an approved script (Appendix E), the study assistant solicited the consent of the submitter

of the water sample to share their contact information with the study investigators. When submitters

agreed to be contacted, the study assistant screened them for eligibility and sent their contact

Methods 27

information to the study coordinator. The study coordinator stripped the case-control designation from

the contact information and sent it to the appropriate interviewer.

The interviewer called consenting households and, using an approved script (Appendix F),

described the research in more detail. The person submitting the water sample, or his/her designate,

was asked to participate in both the case-control and prevalence sections of the study. They were

informed that they would be requested to provide a rectal swab as part of the prevalence study and, if

they refused, were given the option of completing the case-control study (household questionnaire)

only. For those households that participated in the case-control section of the study, the interviewer

asked if other eligible household members would be willing to participate in the prevalence section of

the study.

The study assistant, interviewers, and the telephone survey company made a minimum of ten

attempts to contact each submitter. These calls were made at various times of day, between 9:00 a.m.

and 8:00 p.m., throughout the week and on weekends. Response files were kept by each caller

detailing the date and time of attempted calls, eligibility to participate, reason for no contact, and

response.

Approvals and funding

Ethics approvals were granted by the Universities of Western Ontario and Toronto (Appendix G).

Approval was also granted by the Ontario Ministry of Health and Long-Term Care. The research was

funded by the Canadian Institute of Health Research-Health Canada and the Physicians’ Services

Incorporated Foundation (Ontario).

2.2 Data Collection

2.2.1 London and Hamilton Regions

Visits were made to many of the study participants’ homes to collect data in the London and

Hamilton regions. Trained interviewers called the person who submitted the water sample to explain

the study, arranged a site visit with those who consented to participate, and solicited assent to speak

with other household members during the visit. Parental or custodial assent was obtained before

approaching youths 12 to 15 years of age. The submitter completed a written consent (Appendix H) and


was interviewed using the household questionnaire (Appendix I). All household members who agreed to

submit a rectal swab provided written consent and were interviewed using the personal questionnaire

(Appendix J). Responses of participants were recorded on paper copies of the questionnaires with

responses entered into a database by one data entry professional.

At site visits, the participant was given the labelled rectal swab, a swab collection information

sheet (Appendix K), a bio-hazard specimen bag, and asked to collect the swab. The interviewer mailed

the swabs to the study laboratory within 24 hours of collection. Rectal swabs were chosen over stool

samples as the preferred method of collection for several reasons. At the site visits, interviewers could

request the sample be collected immediately, the interviewers could demonstrate how to collect the

swab, and we expected, given results of other research, that requesting the subject collect the swab

at the time of the interview would increase the proportion of swabs submitted versus having the

subject collect and mail the swab to the study laboratory at a later date (261-264). For subjects

interviewed by telephone, swab kits were more easily mailed to participants than stool collection kits.

Previous research has shown that rectal swabs are a good tool for the detection of antimicrobial

resistant E. coli. Lautenbach studied the sensitivity and specificity of rectal and peri-rectal swabs for

the detection of fluoroquinolone resistant E. coli compared to stool samples. In their sample, 90% of

the patients with positive stool samples also had positive rectal swabs and 90% had positive peri-rectal

swabs. The authors noted that the two samples in which the fluoroquinolone resistant E. coli were not

detected using the rectal or peri-rectal swabs had low concentrations of E. coli (168).

Site visit interviewers were allowed to follow the protocol for telephone surveys (see Section

2.2.2) for subjects living in remote areas and for individuals who refused a site visit.

2.2.2 Ottawa, Kingston, Peterborough, Orillia, and Toronto Regions

Data for respondents living in the Ottawa, Kingston, Peterborough, Orillia, and Toronto public

health laboratory regions were collected by telephone interview. Trained professional interviewers

from a telephone survey company called the submitters. After explaining the study, they completed

household interviews with consenting submitters. Personal questionnaires were completed for

submitters and other eligible household members who verbally consented and agreed to submit a rectal

Methods 29

swab. Telephone surveys were entered directly into a database using a computer assisted telephone

interview (CATI) system.

vRectal swab sampling kits were mailed to all participants who completed personal

questionnaires by telephone. Swabs and consent forms were mailed to the laboratory and study

coordinator, respectively, by the participant. Participants were contacted by telephone if the rectal

swab had not arrived at the laboratory within four weeks of mailing it to the participant.

2.2.3 Questionnaires

The household questionnaire was completed by an adult, either the person who submitted the

water sample for bacteriological testing or their designate (e.g. spouse/partner). The household

questionnaire included information about the water source as well as several questions about the

household including the number and age of household members, occupations of household members,

county of residence, pets, livestock, water treatment(s), and household income. It took about 20

minutes to complete. When data were collected by site visit, physical measurements were made by the

interviewer to determine distances between the water source and possible sources of contamination.

People completing questionnaires by telephone were asked to estimate these distances.

Individual (‘personal’) questionnaires were completed by household members, 12 years of age

and older, who consented to participate and to provide a rectal swab for analysis. The questionnaire

collected information about the participant’s age, sex, length of residence at the site, underlying

disease processes, hospitalization, antibiotic use, and contact with or consumption of possible sources

of antimicrobial resistant E. coli including animals, meat, water, and human or animal waste.

Questionnaires took about 10 minutes to complete and were not necessarily collected in isolation from

other household members. No proxy interviews were allowed.

Household and personal questionnaire items were developing following a review of the literature

to determine the topic areas that needed to be covered to illicit information on risk factors for

bacterial contamination of well water and the development and transmission of antimicrobial resistant

enteric bacteria as well as variables that might confound the associations under study. Items were

fashioned after items from a variety of questionnaires including the Canadian Community Health Survey

v rectal swab sampling kits contained a letter from the study coordinator, swab collection instruction sheet, stamped self-addressed envelopes, swabs, labels, and consent forms


(265), the Canadian Census (72), the draft Households and Environment Survey (73), the Rapid Risk

Factor Surveillance System (266), and an enteric outbreak case management questionnaire from a local

public health unit (267). The questionnaires were reviewed by several content experts including M.

Louie, S. McEwen, A. McGeer, I. Johnson, I. Gutmanis, and S. Bondy. They were pilot tested with 12

volunteers, including 8 adults and 4 adolescents of both sexes who lived on farm and non-farm rural

properties.

2.2.4 Laboratory Testing of Water and Rectal Swab Samples

Water samples were processed and E. coli-positive water samples identified by public health

laboratory personnel according to standard methods (268). All E. coli colonies on each selected plate

were picked onto a swab creating a single pool of colonies to enhance the likelihood of identifying

antimicrobial resistant isolates from an individual water sample. Swabs were shipped to the study

laboratory (Provincial Laboratory, Calgary) on trypicase soy agar slants. Screening for antibiotic

resistance was performed by the agar screen plate method using antibiotic concentrations listed in

Table 2.1. Isolates growing on agar screen plates were confirmed as E. coli using standard biochemical

testsvi. Escherichia coli isolates that screened positive for resistance to any antibiotic were sent to the

Laboratory for Foodborne Zoonoses, St. Hyacinthe, Quebec to have broth microdilution susceptibility

testing using the 2002 and 2004 NARMS antimicrobial susceptibility panel for enteric bacteria (259;260).

Rectal swabs collected by study participants were mailed to the study laboratory in Carey Blair

transport media. Upon receipt, swabs were inoculated into tryptic soy broth and incubated overnight.

A 1 mL aliquot of the overnight broth was archived and frozen at -70°C for future susceptibility testing.

A swab of the frozen sample was inoculated onto MacConkey agar with crystal violet. Lactose-

fermenting colonies growing on the antibiotic screen plates were screened to identify E. coli isolates.

Antimicrobial resistant E. coli were sent to the Laboratory for Foodborne Zoonoses to be tested using

the 2004 NARMS panel.

Susceptibility testing was limited to water and faecal samples that were not susceptible to

antibiotics used in the screening panel and also confirmed as being E. coli using API-E20 (BioMerieux

Canada Inc.) assay. Antimicrobial susceptibility results were interpreted using resistance breakpoints

vi Citrate, indole, malonate metabolism

Methods 31

relevant to human health as outlined by the National Committee for Clinical Laboratory Standards

(269).

2.3 Data Analyses

These analyses were based on the prevalence study conducted with a convenience sample of

people living in households that participated in the case-control study (Table 2.1). Questionnaires of

people who did not submit a swab or submitted a swab that was not screened for antimicrobial

resistance were excluded from analyses. Because of the change from the 2002 to the 2004 NARMS

panels part way through the study, with the substitution of sulphamethoxazole with sulfisoxazole and

the elimination of cephalothin, we excluded households that had water samples that were resistant

only to sulphamethoxazole or cephalothin and individuals that had rectal swabs that were resistant

only to sulfisoxazole.

Data from six different sources were ultimately joined to create one data file. These sources

included the Calgary laboratory files for rectal swab and water screening results, the Laboratory for

Foodborne Zoonoses results files for rectal swabs and water tests, the Compustat Consultants’ file of

telephone interviews, and the file of site interviews.

A comparison of households participating in the case-control study only (household questionnaire

only) and the prevalence/cross-sectional study (household questionnaire and one or more personal

questionnaires with rectal swabs) was completed using Pearson chi square tests of independence.

Similarly, a comparison of participants who completed a questionnaire without submitting a swab (or

who submitted a swab not eligible for inclusion) and subjects who completed a questionnaire and

submitted a swab that was used in the analyses was completed using Pearson chi square without

adjustment for household clustering.

2.3.1 Objective 1: Prevalence of Antimicrobial Resistant E. coli

Prevalence of carriage of antimicrobial resistant E. coli in this study sample was determined

using the total number of isolates (1 per rectal swab) in which the E. coli was resistant to at least one

antibiotic included in the 2004 NARMS for enteric bacteria panel (see Table 2.2 for breakpoints) divided

by the total number of isolates that grew E. coli. Antibiotic specific prevalence was calculated as the


proportion of all E. coli-positive isolates that were resistant to the individual antibiotic. Multi-drug

resistance was defined as resistance to two or more classes of antibiotics included in the panel.

Estimate intervals were presented with 95% probability confidence limits. A Z-approximation was used

to test the first hypothesis. Chi-square tests were used to determine if there were differences in the

proportion of subjects carrying ampicillin resistant E. coli, by risk factor. The variances for these

estimates were adjusted to account for the non-independence of observations within households using

the household identifier as the primary sampling unit in the Stata survey command which restricts the

degrees of freedom to the number of clusters rather than the number of observations and adjusts the

variance for non-independence of observations.

Inter-class resistance is a measure of bacterial resistance between two classes of antibiotics

while intra-class antimicrobial resistance is resistance to two or more drugs within the same class of

antibiotics. Both were calculated by taking the sum of the observed resistances to both antibiotics (or

classes) divided by the observed resistances in the antibiotic (or class) of interest (e.g. inter-class

resistance of tetracycline with ß-lactam: number of samples resistant to ß-lactam and tetracycline

divided by number resistant to tetracycline).

Resistance scores (the total number of observed resistances divided by the total number of

possible resistances) were reported to reflect the burden of multi-drug resistance (184;197). Resistance

scores are presented to enable readers to compare the extent of multi-drug resistance when the

studies use a different number of antibiotics in their panels.

The observations were re-weighted to provide a prevalence estimate removing the impact of

the sampling strategy. Since “case” and “A” control households were over-represented in the sample,

the households were given weights that represented their probability of selection from the sampling

frame. The weights were based on the known results of water tests completed during the study period

but could not be adjusted for repeat submissions from the same household. The variances for these

estimates were also adjusted to account for non-independence of observations within households

(clustering).

The direct age-standardized prevalence was also calculated. Since the vast majority of

residences supplied with private water sources are located in rural areas, this was done using the 2006

estimated population for rural Ontario residents as the standard population (270).

Methods 33

2.3.2 Objective 2: Association Between Water Consumption and Human Carriage of Antimicrobial

Resistant E. coli

The dependent variable was the subject’s laboratory results regarding whether the E. coli

detected in their rectal swab was either (0) susceptible to all or (1) resistant to one or more of the

antibiotics included in the 2004 NARMS susceptibility panel for enteric bacteria.

The independent variable of primary interest was the use of water that was either (1)

contaminated with E. coli that was resistant to one or more antimicrobial agents included in both the

2002 and 2004 NARMS panels and which was not treated for bacterial contamination or (0) water that

was not contaminated with E. coli, contaminated with E. coli that was susceptible to all antibiotics in

the panel, or contaminated with antimicrobial resistant E. coli but treated for bacterial contamination

for one year or longer from the date of the interview. Treatment for bacterial contamination was

defined as drinking water that was boiled or treated with chlorine, ultraviolet light, or ozone (77;78).

We determined that one year of treatment was required to take into account the possible lag time

between beginning the water treatment and potential elimination of antimicrobial resistant E. coli

from the user’s gastrointestinal system (carriage in humans was detected for up to 10 months in one

longitudinal study (149)).

The covariates of interest as possible predictors of human carriage of antimicrobial resistant E.

coli or confounders of the association of interest are presented in Figure 2.1. The covariates, including

how variables were derived, are outlined in Appendix L.


Figure 2.1 The Theorized Relationship between Human Carriage of E. coli, Use of Contaminated and Untreated Water, and Other Study Variables Used for Multivariable Model Building; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Primary predictor: - Water used

Outcome: - AR E. coli carriage

Potential confounders: - Age - Antibiotic use - Sex - Hospitalization - Household education - Travel - Household income - Child in day care - Laboratory region - Household size - Mode of data collection - Contact with livestock - Days between water - Farming property sample & interview - Contact with dog/cat

- Contact with raw meat

Potential effect modifier: - Used bottled water

AR: antimicrobial resistant (to any antibiotic in the both the 2002 and 2004 panels) Primary predictor: Use of untreated, antimicrobial resistant E. coli-contaminated water versus uncontaminated water; contaminated, treated water; or water with susceptible E. coli (water sample) Outcome: Faecal carriage of antimicrobial resistant E. coli versus susceptible E. coli (rectal swab) Potential effect modifier: Use of tap water versus use of both tap and bottled water versus use of bottled water Potential confounders:

- Demographic variables: age, sex, education, income - Based on study design: laboratory region, mode of data collection, days between water sample

and personal interview (rectal swab collection) - As identified by literature review: antibiotic use, hospitalization, travel, child in day care,

household size, contact with livestock, farming property, contact with pets, contact with raw meat/poultry

Methods 35

Although the dependent variable was binary, logisitic regression was not ideal for the analyses

since data collection was cross-sectional. The preferred measure was the relative risk or prevalence

ratio rather than the odds ratio. Poisson regression has been identified as an alternative that can

provide correct estimates of relative risk while adequately representing the association between the

dependent and independent variables (271-273) and hence was chosen for these analyses.

The observations in this data set were not independent as several individuals from one

household were eligible to participate in the prevalence study. To account for the statistical non-

independence of the observations, we analyzed the data using a generalized linear model with the

“cluster” estimation option in Stata, version 9.2 (274). This estimation option specifies that

observations are independent across groups but not necessarily independent within groups/clusters.

The equation uses a robust (or Huber-White) estimate of variance that produces correct standard errors

even if observations are correlated (275) and adjusts the Poisson regression error variance to account

for its conservative results when the dependent variable is binary (273).

Multivariable Poisson regression was used to determine the relationship between the

dependent and independent variables while taking into account other covariates; the second objective.

How a regression equation is constructed drives the process and depends on the goal of the analysis

(276).

Model-building strategies.

There are generally three distinct forms of research objectives that can be met through the

development of regression models and each argues for a different approach to selection of the best

model. The first is exploratory (277-280). This approach is used to determine multiple important

predictors of an outcome. The emphasis of model-building is on what covariates to include in the

model.

In contrast to the exploratory approach is the predictive approach. This second reason for

developing regression models is to predict the outcome of future observations (278-283). This approach

is usually more concerned with the accuracy of the prediction of the overall model rather than the

value of the individual coefficients. The goal is to include as many variables as necessary to accurately


predict the outcome of current and/or future observations. Cross validation and/or data splitting is

often used to determine the predictive ability of the covariates (281).

Kleinbaum, Klein, Austin, Aneschensel, Stahel, Sauerbrei, Vittinghoff, and others describe an

explanatory model-building process. Explanatory model-building assesses the association between an

exposure and an outcome of primary interest with the goal of producing an accurate, unbiased

estimate of the relationship between them (277-283). This is a hypothesis-driven process that takes

into account apriori-identified confounders, moderators, and effect modifiers. The goal is to develop a

model that rules out confounding of the focal relationship by other variables.

Focal relationship analysis.

The goal of our analysis was explanatory: to determine whether the focal relationship

(between human carriage and use of water contaminated with resistant E. coli) remained after

adjusting for confounding and effect modification. We used the strategy outlined by Vittinghoff et al.

(280) who suggest including all potential confounders including those established through previous

studies, those hypothesized to matter on substantive grounds, and any variables that act like

confounders during statistical analysis. This is done to rule out confounding as much as possible.

Confounding occurs when the crude association fails to reflect the size and/or direction of the

exposure effect because of different distributions of a third variable among exposed and unexposed

subjects. A variable is confounding to the association if it is a) predictive of the outcome, b) associated

with exposure, and c) not affected by exposure or outcome (284). Meaningfully different

interpretations of the focal relationship occur when a confounding variable is included or excluded

from the analysis. Confounding can be controlled for through stratification, matching, or regression

analyses. We used regression analysis as it provides an estimate of the focal relationship while holding

the other variables constant, which is the main goal of explanatory model building (280).

Variables were entered into the original (full) model if the association with the dependent

variable had a p-value of ≤ 0.25 on bivariate analysis (276). The product term of the primary predictor

and the theorized effect modifier were also included in the full model, along with the product term of

the primary predictor and biologically-plausible confounding variables. Interaction, or effect

modification, exists when the association between the outcome and exposure is different at different

Methods 37

levels of a third variable (effect modifier). The inclusion of biologically-plausible combinations of

variables are included in multivariable regression analyses to determine whether effect modification

exists (276;280).

The model was reduced by removing variables with the largest p-value, one at a time, until

only variables with a p-value of ≤ 0.20 (276;283;285-287) and those that changed the estimate of the

primary predictor by more than 10% (280;282;288) remained. To reduce the potential bias of missing a

confounding variable, all variables were added back into the model, one at a time, to identify variables

that were important to the model in the presence of other variables (276). Continuous variables were

assessed for linearity and transformed or categorized, if necessary. Analyses were run again after

variables were transformed.

Regression diagnostics were also performed. Tables of residual values were examined and

residual values were graphed to determine if there were covariate patterns that were outliers or had

high leverage influence values (276;280). The statistical impact of outlying or high leverage

observations was examined by analyzing the data after removing the observations from the data set.

The final model was assessed to determine if the Poisson distribution adequately described the

association between the dependent and independent variables. The response variance was not

expected to be meaningfully different than the mean as using Poisson regression with binary data

(denominator of 1) can not produce over-dispersed errors (280;289). The specification test, a summary

measure of the model’s goodness of fit, and the Akaike and Bayesian information criteria were

reviewed for the final model.

The data were also analyzed using generalized estimating equations with exchangeable

correlation within households (clusters), a Poisson distribution, robust standard error estimates, and

population-averaged equations. The parameter estimates, standard errors, and model-building results

were compared with those resulting from the generalized linear model equation.

2.3.3 Student’s Role

The candidate was involved in the CIHR-Health Canada-funded surveillance study since

September 2004. She was instrumental in determining the design of the case-control and prevalence

studies. Her original proposal was pared down, due to budget constraints, by excluding a three-month


follow-up of wells and participants. She worked closely with other members of the research group in

writing the proposal for additional funding, which was granted by the Physicians’ Services Incorporated

Foundation, Ontario. The proposal was mirrored for funding received from the Alberta Heritage

Foundation for an expansion to the Ontario study completed in Alberta.

The candidate was instrumental in the implementation of the Ontario case-control and

prevalence studies. She worked with the laboratories, public health units, and Ministry of Health and

Long-Term Care to ensure their cooperation in the study. Further, she drafted and pre-tested the

questionnaires, scripts, consent forms, and information sheets including necessary changes for the

telephone surveying component of the study. The candidate determined the need to implement

changes to the protocol and implemented the resulting expansion and extension of the study. She

wrote the applications for ethics review and amendments to the approvals received from the University

of Western Ontario and University of Toronto research ethics boards.

The candidate was the coordinator of the Ontario site of the well water study, which included

hiring, training (including the development of a training manual), and supervising interviewers, callers

at the Ministry of Health and Long-Term Care, and the telephone survey company. She managed the

flow of information between the laboratories, callers, interviewers, and data entry clerk and managed

the databases. The candidate was also the contact person listed on the information sheet and consent

forms making her the contact for the public.

The candidate was responsible for the cleaning, editing, and analysis of the case-control and

prevalence studies and will also be responsible for providing data to other students and researchers

involved in the CIHR-Health Canada surveillance study. She has been and will be involved in the

preparation and submission of the case-control findings and will be lead author on publications

submitted from the prevalence study.

Results 39

Chapter 3

Results

3.1 Characteristics of Water Samples, Households, and Respondents

3.1.1 Water samples

In total, 342,009 water samples submitted to the seven participating public health laboratories

during the study period (May 2005 to September 2006) were tested for bacterial contamination.

Recalling that multiple water samples may be submitted from one household, 15,238 (4.5%) water

samples were contaminated with E. coli, 60,540 (18%) were contaminated with non-E. coli coliform

bacteria, 12,139 (3.5%) were overgrown with non-E. coli bacteria, and 254,092 (74%) had no bacterial

contamination (see Table 3.1). Of the E. coli-positive samples, 6,492 were sent to the study laboratory

for susceptibility testing and 645 (10%) were antimicrobial resistant. Assuming the samples sent to the

study laboratory were representative of all samples, about 0.4% of all water samples tested by the

participating public health laboratories were contaminated with antimicrobial resistant E. coli.

Due to selection of households for participation in the case-control study, 22% of households

participating in the study had water contaminated with antimicrobial resistant E. coli (Table 3.2).

3.1.2 Participation

Dwelling questionnaires were completed by 880 of the 1,717 (51%) eligible households in

southern Ontario (Tables 3.1 & 3.3). Personal questionnaires were completed by 1,007 individuals from

671 households (average 1.5 interviews per household). Three-quarters of individual respondents (752

or 75%) submitted a rectal swab and the laboratory was able to detect E. coli on 703 (94%) of the 746

swabs that were screened. Four swabs were not eligible for the final analysis because the E. coli were

resistant only to sulfisoxazole (which was not tested for in all submitted water samples). The final

analyses were conducted using the resulting 699 swabs and personal survey responses with the

associated 488 household questionnaires.


Table 3.1 Details of Household Recruitment and Participation for Surveillance Project and Case-Control Study, Antimicrobial Resistant E. coli in Private Water Sources, Southern Ontario, 2005-2006

1 342,009 water samples2Water samples: total testedWater samples : results of bacteriological testing

15,238 (4.5%) 254,092 E. coli-positive No bacteria

Water samples: screened for resistance 6,492 water samples Water samples: results of antimicrobial resistance testing

645 (9.9%) resistant

5,856 susceptible

645 840 846 (Case) (A control) (B control) Water samples: selected for study TOTAL

Households: not eligible for study3 202 180 232 614 eligible for study 443 660 614 1,717

143 151 164 458 Households: refused unable to contact 29 29 27 85 agreed to be called by study 272 480 423 1,175 Households: refused survey 88 107 92 287 completed survey 184 369 327 880 Households: eligible for analysis 108 196 184 488 1 Includes all water samples tested for bacteriological contamination, including multiple samples from the same household 2 Includes 72,679 not eligible for the study (contaminated with non-E. coli coliform bacteria [n=60,540] or overgrown with no E. coli detected [n=12,139]) 3 Unique and eligible households (see section 2.1.1 for criteria)

Results 41

Table 3.2 Proportion of E. coli-positive Water Samples with Antimicrobial Resistant E. coli, by Antibiotic and Class of Antibiotic, Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

All water samples Participating households only Antibiotic1

Number2 positive

% Number positive

% CI CIResis-

tant Resis-tant

95% 95%

(N=6,492) (N=488)

Resistant to one or more antibiotics 645 10% 9, 11 108 22% 18, 26 Multi-drug resistance (2+ classes) 394 6% 5, 7 64 13% 10, 16

236 4% 3, 4 38 Aminoglycoside 8% 5, 10 Amikacin 0 --- --- 0 --- --- Gentamicin 25 0.4% 0, 1 6 1% 0, 2 Kanamycin 69 1% 1, 1 11 2% 1, 4 Streptomycin 209 3% 3, 4 33 7% 5, 9 Beta-lactam 268 4% 3, 5 49 10% 7, 14 Amoxicillin/clavulanic acid 58 1% 1, 1 10 2% 1, 3 Ampicillin 268 4% 4, 5 49 10% 7, 13 Cefoxitin 15 0.2% --- 9 2% 1, 3 Ceftiofur 38 0.6% 0, 1 7 1% 0, 3 Ceftriaxone 3 0 --- 1 0.2% 0, 1

1% 0, 2 Fluoroquinolone 0, 1 5 30 0.5% 0.2% 0, 1 --- 1 Ciprofloxacin 5 0.1%

0, 2 5 1% 0.5% 0, 1 Nalidixic acid 30 295 Sulphonamide 5% 4, 5 53 11% 8, 14 133 Trimethoprim-sulfamethoxazole 2% 2, 2 22 5% 3, 6

Sulfisoxazole (n=3602) 1022 3% 2, 3 20 10% 6, 14 Sulphamethoxazole (n=2890) 1792 6% 5, 7 29 10% 7, 14 Tetracycline 498 8% 7, 8 86 18% 14, 21 Chloramphenicol 50 1% 1, 1 13 3% 1, 4

2Cephalothin (n=2890) 56 2% 1, 2 14 5% 2, 7 CI: Confidence interval (95%) 1 Using the 2004 NARMS panel for enteric bacteria 2 Number of E. coli-positive water samples tested is 6,492 unless otherwise stated (to account for change from 2002 to 2004 NARMS panel)

Table 3.3 Details of Subject Participation in Prevalence Study, Survey of Households using Private Water Sources, Southern Ontario, 2005-2006 Households Individuals Eligible for case-control study 1,717 Completed household survey 880 Household survey only 209 Completed household & personal surveys 671 1,007 Submitted rectal swab 752 Lost or damaged in transit 6 Swabs screened for E. coli 746 No E. coli 43 Swabs screened for resistance 703 Not eligible (only sulfisoxizole resistant) 4 Swabs used in analysis 488 699


As shown in Table 3.4, there were few differences between the households that participated in

the prevalence study (completed a household questionnaire and one or more individuals completed a

personal questionnaire and submitted a swab) and those that participated only in the case-control

study (completed only a household survey). Households that participated in the prevalence study had

fewer household residents, with an average of 2.6 household members compared to 3.0 members in

households that took part only in the case-control study (p=0.001). Households that participated in the

prevalence study were also more likely to state their household income than households than only took

part in the case-control section. Households in the London and Hamilton regions were more likely to

participate in the prevalence study, which is associated with the mode of interview: people

interviewed face-to-face were significantly more likely to participate in both the case-control and

prevalence studies than people interviewed by telephone and asked to post the swabs at a later date.

On crude analysis (not adjusted for household clustering), individuals who participated in the

prevalence study by completing a personal questionnaire and submitting a rectal swab were older

(mean: 58 years) than individuals who completed a personal questionnaire without submitting a swab

or for whom the swab was not usable (mean: 52 years; p<0.001). As shown in Table 3.5, individuals 49

years of age and younger were less likely to submit a swab after completing a personal questionnaire

than older individuals (p<0.01). Similarly, individuals in recent contact with horses were less likely to

submit a swab than those without recent contact but who completed a personal questionnaire

(p=0.03). In comparison, individuals who used tap water exclusively were more likely to submit a swab

than people who used bottled water (p=0.01). Subjects who travelled outside Canada in the previous

year were more likely to submit a swab than those who did not travel (p=0.05), and participants who

were in contact with poultry in the previous three months were more likely to submit a swab than

those without contact but who completed a questionnaire (p=0.02).

Results 43

Table 3.4 Comparison of Households Participating in Prevalence and Case-control Studies; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Prevalence with swab

Case-control only p-value

Household-level item χ2 % of house-holds

% of house-holds

† Number Number (N=488)* (N=392)*

Water source Contaminated with resistant E. coli 108 22% 76 19% Contaminated with susceptible E. coli 196 40% 173 44% No bacterial contamination past year 184 38% 143 36% NS

1Water treated (n=487; 390) 243* 50% 177* 45% NS Household size (n=488; 390)

53 36 1 11% 9% 2 273 56% 163 42% 3 or 4 24% 115 127 33% 5 to 10 10% 47 64* <0.01 16%

Children in household (1 or more) 20% 77 13% 64 12-19 years 19% 76 11% 56 4-11 12% 48 7% 32 <4 11% 45 8% 40 In diapers

NS 4% 17 2% 9 Attend day care Farming property 134 27% 92 23% NS

Highest education in household 26 7% 38 8% Less than high school 67 17% 74 15% Graduated high school 31% 122 30% 148 College or trade school 39% 44% 153 University 214

NS 6% 3% 24 Not stated 14 Household income2

10% 41 16% 80 <$40,000 annually 27% 104 32% 156 $40,000-79,999 31% 121 29% 140 $80,000 or more

<0.01 32% 126 23% 112 Not stated Laboratory region3

38% 150 46% 223 London 15% 59 27% 131 Hamilton 13% 50 8% 37 Kingston 10% 39 7% 34 Orillia 11% 44 7% 34 Peterborough 6% 25 4% 18 Ottawa

<0.01 6% 25 2% 11 Toronto Mode of interview

356 91% 63% 308 Telephone interview 37% 36 9% <0.01 180 Site visit (face-to-face interview)

* N=488 and 392 households unless otherwise stated (i.e. incomplete data for item) 1 Water treated by boiling, chlorination, ultraviolet light, or ozonation 2 Annual household income, before taxes, for all members in household (including net farm income) 3 Laboratory to which the water sample was submitted (not necessarily household location) † Chi square test of independence between case and control households


Table 3.5 Comparison of Participants who Completed Personal Questionnaires and Submitted a Rectal Swab and Subjects Who Completed Questionnaires Only or Submitted an Ineligible Swab, Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Questionnaire & swab

Questionnaire only or ineligible

swab

p-value

Item χ2 † Number (N=699)*

% of subjects

Number (N=308)*

% of subjects

Respondent’s age at time of interview 9% 30 4% 31 12-29 years 10% 32 7% 49 30-39

74 24% 14% 97 40-49 73 24% 25% 50-59 175 51 17% 30% 60-69 207 35 11% 17% 70-79 122

<0.01 10* 3% 3% 18* 80 & older (n=699; 305) Respondent’s sex 46% 143 49% 346 Female

54% NS 165 51% 353 Male Hospitalized (over night) in past year 55 8% 34 11% NS Antibiotic used, past 3 months 85 12% 33 11% NS Currently using antibiotic 6 1% 2 1% NS Chronic condition Diabetes mellitus 41 6% 16 5% NS Crohn’s, celiac, IBS, etc. 51 7% 25 8% NS Heart disease/high blood pressure 206 29% 79 26% NS Arthritis/rheumatism 233 33% 85 28% NS Tap or bottled water1

52% 161 62% 433 Tap water only 26% 79 19% 133 Tap and bottled

Bottled only 22% 0.01 68 19% 133 Raw milk consumed, regularly (n=698) 48* 7% 18 6% NS Travelled outside Canada, past year 376 54% 145 47% 0.05

Contact in past 3 months NS 82% 250* 81% 562* Raw red meat (n=697; 306) NS 79% 241* 78% 543* Raw poultry products (n=696; 305) NS 79% 243 76% 528* Dogs (n=698; 308) NS 56% 171* 56% 392 Cats (n=699; 307)

0.03 19% 57* 13% 94 Horses (n=699; 306) 0.02 7% 20* 11% 80 Poultry (chicken,turkey)(n=699; 307) NS 12% 36* 11% 79* Cattle (dairy or beef) (n=698; 306) NS 7% 20* 6% 39 Sheep or goats (n=699; 306) NS 3% 9* 4% 27* Pigs (n=697; 305)

* N=699 and 308 subjects unless otherwise stated (i.e. incomplete data for item) 1 Tap water only: do not used bottled water at home on regular basis; Tap and bottled: glasses

of water [total] > glasses bottled; Bottled water only: glasses of water [total] = glasses bottled † Unadjusted Chi square test of association between respondents submitting swabs and those

who only completed the personal questionnaire or submitted an ineligible swab

Results 45

3.1.3 Households

Table 3.6 highlights the demographic information about the 488 households in which the survey

respondents lived. One hundred eight households (22%) had water that was contaminated with

antimicrobial resistant E. coli, 196 (40%) had water that was contaminated with E. coli that was

susceptible to the tested antibiotics, and 184 (38%) had water than had not been contaminated with E.

coli or other coliform bacteria for one year or longer. This finding mirrors the selection of households

for the case-control section of the project, not the probability of contamination of private water

sources (see Section 3.1.1).

One-half of the households in this study (243 of 488) used a treatment to kill water-borne

bacteria (boiling, chlorine, ozone, or ultraviolet light). Households that had E. coli-contaminated water

were significantly more likely to treat their water than households with no recent history of

contamination (54% versus 43%; p=0.02).

Of the 108 households that had water contaminated with antimicrobial resistant E. coli, 54

households (50%) met the criteria to be classified, for this study, as treating their water for bacterial

contamination (chlorine, ultraviolet light, ozone, or boiled). However, 15 of these households had been

treating their water for less than one year. Thus, 69 households, representing 94 individuals, were

classified as exposed to untreated, contaminated water for the analyses. The only difference between

case and control households was the proportion of households that treated their water for bacterial

contamination.

Household size ranged from 1 to 10 people, with the median size of 2 people per household

(mean 2.7). Sixty-four households (13%) had youths 12 to 18 years of age living in them, 11% had

children 4 to 11 years of age, and 7% had children under 4 years old. Forty households (8%) had one or

more children that used diapers while only nine (2%) of these largely rural households had a child that

attended day care.

Three hundred sixty-two households (74%) had one or more household members with a college or

university education. Two hundred ninety-six households (61%) earned $40,000 or more per year

although 112 (23%) did not answer the questions regarding household income. One hundred thirty-four

households (27%) were located on farming properties.


As expected, given the sampling strategy used for the case-control study, most households (73%)

were located in the London (n=223) and Hamilton (n=131) public health laboratory regions. Three

hundred eight households (63%) were interviewed by telephone and the remaining 180 were

interviewed face-to-face in their home (site visit). Households in the London and Hamilton regions

were less likely to be interviewed by telephone than households in the Ottawa, Kingston,

Peterborough, Orillia, and Toronto regions (49% versus 99%; p<0.01).

Results 47

Table 3.6 Descriptive Statistics of Participating Households; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

All households Resistant E. coli Susceptible Household-level item Number

(N=488)*

% of house-holds

Number (N=69)*

% of house-holds

Number (N=419)*

% of house-holds

Water source Contaminated with resistant E. coli Contaminated with susceptible E. coli No bacterial contamination past year

108 196 184

22% 40% 38%

69

100%

39 196 184

9% 47% 44%

Water treated1 (n=487) 243* 50% 15 22%† 228 54%† Household size2 1 2 3 or 4 5 to 10

53 273 115 47

11% 56% 24% 10%

11 31 19 8

16% 45% 27% 12%

42 242 96 39

10% 58% 23% 9%

Children in household (1 or more) 12-19 years 4-11 <4 In diapers Attend day care

64 56 32 40 9

13% 11% 7% 8% 2%

10 9 6 5 0

14% 13% 9% 7% 0

54 47 26 24 9

13% 11% 6% 6% 2%

Farming property3 134 27% 24 35% 110 26% Highest education in household4

Less than high school Graduated high school College or trade school University Not stated

38 74 148 214 14

8% 15% 30% 44% 3%

5 11 20 32 1

7% 16% 29% 46% 1%

33 63 128 182 13

8% 15% 31% 43% 3%

Household income5

<$40,000 annually $40,000-79,999 $80,000 or more Not stated

80 156 140 112

16% 32% 29% 23%

11 23 21 14

16% 33% 30% 20%

69 133 119 98

16% 32% 28% 23%

Laboratory region6

London Hamilton Kingston Orillia Peterborough Ottawa Toronto

223 131 37 34 34 18 11

46% 27% 8% 7% 7% 4% 2%

29 23 5 4 2 4 2

42% 33% 7% 6% 3% 6% 3%

194 108 32 30 16 30 9

46% 26% 8% 7% 4% 7% 2%

Mode of interview Telephone interview Site visit (face-to-face interview)

308 180

63% 37%

38 31

55% 45%

270 149

64% 36%

* N=488, 69, or 619 households unless otherwise stated (i.e. incomplete data for item) 1 Water treated by boiling, chlorination, ultraviolet light, or ozonation 2 Household size (all): mean: 2.6; median: 2; range: 1-10 people 3 As described by person completing household questionnaire 4 Highest level of education attained by any member in household 5 Annual household income, before taxes, for all members in household (including net farm income) 6 Laboratory to which the water sample was submitted (not necessarily household location) † Chi square test of independence of households with/without contaminated water: p-value of ≤0.05


3.1.4 Respondents

As detailed in Table 3.7, the age of the 699 respondents ranged from 12 to 87 years (median 59;

mean 58). Three hundred eighty-two respondents (55%) were 50 to 69 years of age with only 31

respondents (4%) 12 to 19 years old. Three hundred fifty-three (51%) respondents were male and 346

(49%) were female. The ages and sex of respondents were similar for those from households that had

water that was or was not contaminated with antimicrobial resistant E. coli.

The number of days between the submission of the water sample to the regional public health

laboratory for bacteriological testing and the date of the interview (proxy for date of rectal swab

collection) ranged from 3 to 439 days. The number of days between water submission and interview

had a skewed distribution with the median lower than the mean (125 and 142, respectively). The

respective median (122 and 125 days) and mean (155 and 140 days) lag times were similar for

respondents from households that had water that was and was not contaminated with E. coli that was

antimicrobial resistant (p=0.12).

Fifty-five respondents (8%) reported being hospitalized in the previous year and 85 (12%) reported

using an antibiotic in the previous three months (6 were using antibiotics at the time of the interview).

Forty-one respondents (6%) reported having been diagnosed with diabetes mellitus, 51 (7%) with an

underlying gastrointestinal condition (e.g. Crohn’s disease, celiac disease, colitis, ileitis, or irritable

bowel syndrome), 206 (29%) with heart disease, and 233 (33%) with arthritis or rheumatism.

Four hundred thirty-three respondents (62%) did not use bottled water regularly (most days of

the week) at home. Of the 266 respondents (38%) who reported using bottled water at home on a

regular basis, half (n=133) used bottled water almost exclusivelyvii while the other half (n=133) used

both bottled and tap water. Respondents from households with E. coli-contaminated water (n=429),

not antimicrobial resistant E. coli specifically, were significantly more likely than respondents from

households with no recent history of bacterial contamination (n=270) to use bottled water exclusively

(24% and 15%, respectively) or to use both bottled and tap water (21% and 12%, respectively) (p<0.001).

Similarly, respondents in households using water contaminated with antimicrobial resistant E. coli were

significantly more likely than subjects in other households to use bottled water (p=0.02).

vii Glasses of water consumed (total) = glasses of bottled water consumed

Results 49

Although only 185 respondents (27%) had direct contact with farm livestock or their manure, 528

(76%) had direct contact with dogs and 392 (56%) had contact with cats in the previous three months.

Of interest, although males were more likely than females (14% versus 8%) to have had direct contact

with cattle (or cattle manure) in the previous three months (p=0.02), there were no significant

differences between the sexes regarding contact with any other animals.

Forty-eight subjects (7%) drank raw milk or ate dairy products made with unpasteurized raw milk

in the previous three months. Raw dairy product consumption was more common among people living

on a farming property (n=30 or 15%) than subjects not living on a farm (n=18 or 4%; p<0.001).

Five hundred sixty-two respondents (81%) had handled raw red meat (beef, pork, or lamb) or raw

poultry (78%). Females were significantly more likely than males (95% vs. 75%) to have handled raw

meat or poultry in the previous three months (p<0.001).

Over one-half (n=376 or 54%) of respondents had travelled outside Canada in the previous year.

One hundred seventy-one (25%) subjects travelled outside both Canada and the United States with 73

respondents visiting countries in the Caribbean and Central or South America, 34 visiting Mexico, and

the remainder visiting Europe, Australia, New Zealand, Asian or Indonesian countries, or Russia.


Table 3.7 Descriptive Statistics of Participants, Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

All subjects Resistant E. coli Susceptibile Individual-level item

Number (N=699)*

% of subjects

Number (N=94)*

% of subjects

Number (N=605)

% of subjects

Respondent’s age at time of interview 12-29 years 30-39 40-49 50-59 60-69 70-79 80 & older

31 49 97 175 207 122 18

4% 7% 14% 25% 30% 17% 3%

5 12 15 22 29 11 0

5% 13% 16% 23% 31% 12% 0

26 37 82 153 178 111 18

4% 6% 14% 25% 29% 18% 3%

Respondent’s sex Female Male

346 353

49% 51%

48 46

51% 49%

298 307

49% 51%

Lag: water submission to interview 3-59 days 60-99 100-139 140-199 200-439

130 90 114 130 141

21% 15% 19% 21% 23%

5 27 18 16 28

5% 29% 19% 17% 30%†

135 117 132 146 169

19% 17% 19% 21% 24%†

Hospitalized (over night) in past year 55 8% 9 10% 46 8% Antibiotic used, past 3 months Currently using antibiotic

85 6

12% 1%

10 0

11% 0

75 6

12% 1%

Chronic condition Diabetes mellitus Crohn’s, celiac, IBS, etc. Heart disease, high blood pressure Arthritis/rheumatism

41 51 206 233

6% 7% 29% 33%

4 3 20 27

4% 3% 21% 29%

37 48 186 206

6% 8% 31% 34%

Tap or bottled water3

Tap water only Tap and bottled Bottled only

433 133 133

62% 19% 19%

49 17 28

52% 18% 30%†

384 116 105

63% 19% 17%†

Raw milk consumed, regularly (n=698) 48* 7% 7 7% 41 7% Travelled outside Canada, past year Travelled outside Canada & USA

376 171

54% 25%

47 23

50% 24%

329 148

54% 24%

Contact in past 3 months Raw red meat (n=697) Raw poultry products (n=696) Dogs (n=698) Cats Horses Poultry (chickens, turkeys, etc) Cattle (dairy or beef) (n=698) Sheep or goats Pigs (n=697,93,604)

562* 543* 528* 392 94 80 79* 39 27*

81% 78% 76% 56% 13% 11% 11% 6% 4%

81 76 79 59 19 28 17 7 7*

86% 82% 84% 63% 20% 30% 18% 7% 7%

481 467 449 333 75 52 62 32 20*

80% 77% 74% 55% 12% 9% 10% 5% 3%

* N=699, 94, or 605 subjects unless otherwise stated (i.e. incomplete data for item) 2 Age (all): mean: 58 years, median: 59 years, range: 12-87 years 3 Lag (all): days from water submission to interview: mean:142; median:124; range:3-439 4 Tap water only: do not used bottled water at home on regular basis; Tap and bottled: glasses

of water [total] > glasses bottled; Bottled water only: glasses of water [total] = glasses bottled † Chi square test of independence of subjects using/not contaminated water: p-value of <0.05

Results 51

3.2 Prevalence of Carriage of Antimicrobial Resistant E. coli

The prevalence of antimicrobial resistant E. coli detected on the rectal swabs, defined as

resistance to one or more of the antibiotics in the 2004 NARMS panel, was 41% (95% confidence

interval, CI95%, 37, 45). Since the weighting strategy was rudimentary, and because the prevalence

estimates were comparable, the un-weighted estimates of prevalence, adjusted for non-independence

of observations within households, are presented in the text.

The highest rates of resistance were to ampicillin (n=194 or 28%), tetracycline (n=176 or 25%),

and sulfisoxazole (n=164 or 24%). The prevalence of streptomycin resistant E. coli in this population

was 17% and trimethoprim-sulfamethoxazole resistance was 14%. The most common patterns of

resistance were between ampicillin and sulfisoxazole with 120 (17%) isolates resistant to both of these

antibiotics. Resistance to both tetracycline and sulfisoxazole (n=114 or 16%) and ampicillin and

tetracycline (n=110 or 16%) was also common.

As shown in Table 3.8, the classes of drugs with the highest rates of resistance were the beta-

lactam (n=194 or 28%), tetracycline (n=176 or 25%), sulphonamide (n=166 or 24%), and aminoglycoside

(n=127 or 18%) agents. Escherichia coli from the faecal swabs were less likely to be resistant to the

fluoroquinolone and chloramphenicol classes of antibiotics (6% (n=42) and 5% (n=32) respectively). The

resistance score (observed resistances divided by total possible resistances), a summary measure of the

prevalence of multi-drug resistance, was 8.5% in this sample.

Table 3.9 highlights the intra-class resistance detected in the E. coli from human faecal samples.

Within the beta-lactam class of antibiotics, all E. coli isolates that were resistant to amoxicillin-

clavulanic acid (n=10), cefoxitin (n=8), and ceftiofur (n=7) were also resistant to ampicillin. All

ciprofloxacin resistant E. coli (n=18) were resistant to nalidixic acid. Most (98 of 99 or 98%) of the

trimethoprim-sulphamethoxazole resistant isolates were resistant to sulfisoxazole. There was slightly

lower intra-class resistance in the aminoglycoside agents with 9 of 14 (64%) of gentamicin resistant

isolates and 9 of 12 (75%) kanamycin resistant E. coli isolates being resistant to streptomycin as well.

For this study, multi-drug (or inter-class) resistance was defined as resistance to two or more

classes of antibiotics included in the NARMS panel. Multi-drug resistance was detected in 204 of the 699

swabs (29%; CI95% 26, 33). As shown in Table 3.10, greater than 70% of aminoglycoside resistant E. coli

isolates were simultaneously resistant to a sulphonamide (108 of 127 or 85%), a beta-lactam (97 of 127


or 76%), or tetracycline (91 of 127 of 72%). Similarly, greater than 70% of fluoroquinolone (31 of 42 or

74%) or sulphonamide (121 of 166 or 73%) resistant isolates were resistant to one of the beta-lactam

agents included in the 2004 NARMS enteric bacteria panel. Further, more than 80% of chloramphenicol

resistant E. coli isolates were also resistant to one of the sulphonamide (29 of 32 or 91%) drugs or

tetracycline (26 of 32 or 81%). In fact, 18 of 32 (56%) chloramphenicol resistant isolates were resistant

to tetracycline, a sulphonamide, and a beta-lactam agent.

Table 3.8 Proportion of Human Rectal Swabs with Antimicrobial Resistant E. coli, by Antibiotic and Class of Antibiotic, Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Design-adjusted estimate2

Weighted Estimate2,3

Number resistant

(N=699) Antibiotic1 Resistant CI Resistant CI95% 95%

Resistance to one or more antibiotics 285 41% 37, 45 39% 33, 45 Multi-drug resistance (2+ classes) 204 29% 26, 33 28% 23, 33

14, 23 18% 15, 21 18% 127 Aminoglycoside --- 0 --- 0 0 Amikacin

0, 3 2% 1, 3 2% 14 Gentamicin 1, 5 3% 1, 3 2% 12 Kanamycin

13, 22 17% 14, 20 17% 119 Streptomycin 194 21, 32 28% 24, 31 27% Beta-lactam 10 0, 1 1% 1, 2 0.4% Amoxicillin/clavulanic acid 194 21, 32 28% 24, 31 27% Ampicillin 8 0, 1 Cefoxitin 1% 0, 2 0.4%

Ceftiofur 7 1% 0, 2 1% 0, 2 Ceftriaxone 0 --- 0 0 ---

3, 8 5% 4, 8 6% 42 Fluoroquinolone 1, 5 3% 1, 4 3% 18 Ciprofloxacin 3, 8 5% 4, 8 6% 42 Nalidixic acid

16, 25 21% 20, 27 24% 166 Sulphonamide 10, 17 13% 12, 17 14% 99 Trimethoprim-sulfamethoxazole 16, 25 20% 20, 27 23% 164 Sulfisoxazole

Tetracycline 176 25% 22, 29 25% 20, 30 Chloramphenicol 32 5% 3, 6 5% 3, 8 CI: Confidence interval (95%) 1 Using the 2004 NARMS panel for enteric bacteria 2 Variances adjusted to account for non-independence of observations within households 3 Weighted by household’s probability of selection from Safe Water Unit data base

Results 53

Table 3.9 Intra-class Resistance of Antimicrobial Resistant E. coli Isolates from Human Rectal Swabs; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

(a) Beta-lactam Amoxicillin-clavulanic

acid Ampicillin Cefoxitin Ceftiofur Ceftriaxone Resistant isolates (n) n=10 n=194 n=8 n=7 n=0 (%) (1%) (28%) (1%) (1%) Amoxicillin- --- Clavulanic acid --- 5% 100% 86% Ampicillin 100% --- 100% 100% --- Cefoxitin 80% 4% --- 86% --- Ceftiofur 60% 4% 75% --- ---

(b) Aminoglycoside Amikacin Gentamicin Kanamycin Streptomycin Resistant isolates (n) n=0 n=14 n=12 n=119 (%) (2%) (2%) (17%)

--- Gentamicin --- 0 8% --- Kanamycin 0 --- 8% --- Streptomycin 64% 75% ---

(c) Fluoroquinolone Ciprofloxacin Nalidixic acid

n=18 n=42 Resistant isolates (n) (%) (3%) (6%) Ciprofloxacin --- 43% Nalidixic acid 100% ---

(d) Sulphonamide Trimethoprim-

sulphamethoxazole Sulfisoxazole n=99 n=164 Resistant isolates (n)

(%) (14%) (24%) Trimethoprim-sulphamethoxazole --- 59% Sulfisoxazole 98% ---

Column identifies antibiotic used in denominator: it specifies the percentage of isolates resistant to that antibiotic that are also resistant to the antibiotic in the corresponding row. E.g. Gentamicin: of the 14 isolates resistant to gentamicin, none were resistant to kanamycin but 64% (n=9) were resistant to streptomycin.


Table 3.10 Inter-class Resistance of Antimicrobial Resistant E. coli Isolates from Human Rectal Swabs; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Amino-glycoside

Beta-lactam

Fluoro-quinolone

Sulphon-amide

Tetra-cycline

Chloram-phenicol

Resistant isolates (n) n=127 n=194 n=42 n=166 n=176 n=32 (%) 18% 28% 6% 24% 25% 4%

Aminoglycoside -- 50% 48% 65% 52% 44% Βeta-lactam 76% -- 74% 73% 63% 63% Fluoroquinolone 16% 16% -- 17% 14% 9% Sulphonamide 85% 62% 67% -- 65% 91% Tetracycline 72% 57% 57% 69% -- 81% Chloramphenicol 11% 10% 7% 18% 15% -- Column identifies antibiotic used in denominator: it specifies the percentage of isolates resistant to that class of antibiotic that are also resistant to the class of antibiotic in the corresponding row (e.g. aminoglycoside: of the 127 isolates resistant to an aminoglycoside, 76% (n=97) were resistant to a beta-lactam).

3.2.1 Ampicillin Resistant E. coli Prevalence

The prevalence (variance adjusted for non-independence of observations) of ampicillin resistant

E. coli from rectal swabs of non-institutionalized people living in southern Ontario and using private

water sources was 28% (CI95% 24, 31). This corresponds to the hypothesized prevalence of greater than

or equal to 22% (p<0.001). The age-standardized prevalence (30%; CI95% 26, 33) and weighted

prevalence (27%; CI95% 21, 32) of ampicillin resistant E. coli were not significantly different than the

original estimate of 28% (Table 3.8).

As shown in Table 3.11, several variables were associated with the prevalence of carriage of

ampicillin resistant E. coli in these subjects. The prevalence in the population of respondents who used

water contaminated with antimicrobial resistant E. coli was significantly higher than for those who did

not use water that was contaminated with antimicrobial resistant E. coli; 38% versus 26% (p=0.01).

Males were significantly more likely than females to carry ampicillin resistant E. coli (32% versus 23%;

p=0.007) and people who had travelled outside of Canada in the previous year were more likely to

carry resistant E. coli than people who had either not travelled or had travelled only within Canada

(32% versus 23%; p=0.008).

Results 55

Table 3.11 Proportion of Human Rectal Swabs with Ampicillin Resistant E. coli by Selected Covariates, Variances Adjusted for Household Clustering; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

p-value1 Number 2) (N=699)* (χCovariate Percent CI95%

Water Contaminated with resistant E. coli, untreated 94 38% 29, 48 No resistant E. coli or treated 12+ months 605 26% 23, 30 0.01 Respondent’s sex Female 346 23% 19, 28 Male 353 32% 27, 37 0.007 Travel outside Canada past year 376 32% 27, 37 No travel or only within Canada 323 23% 18, 27 0.008 Household size 1 53 17% 7, 27 2 410 28% 24, 32 3 or 4 160 27% 20, 34 5 to 10 76 36% 25, 46 0.13 Hospitalized past year 55 20% 11, 32 Not hospitalized 644 28% 25, 32 0.18 Child in day care 14 43% 21, 68 No child or no child in day care 685 27% 24, 31 0.20 Contact with sheep or goats past 3 months 39 18% 8, 36 No contact 660 28% 25, 32 0.22 Interviewed face-to-face (site visit) 266 30% 25, 37 Telephone interview 433 26% 22, 31 0.23

2 Lag: water submission to interview 16, 30 23% 135 3-59 days 24, 41 32% 117 60-99 25, 41 33% 132 100-139 19, 34 27% 146 140-199

0.26 18, 31 25% 169 200-439 Household income in previous year

105 < $40,000 25% 17, 34 $40,000-79,999 229 24% 19, 30 $80,000 or more 209 32% 26, 39 Not stated 156 30% 23, 37 0.27

Respondent’s age at interview 25, 39 32% 177 12-49 years 23, 37 30% 175 50-59 19, 31 25% 207 60-69

0.30 17, 31 24% 140 70 & older Contact with poultry past 3 months (n=696) 80 33% 23, 44 No contact 619 27% 24, 31 0.34

85 24% 15, 34 Antibiotic used past 3 months No antibiotic 614 28% 25, 32 0.37 Tap or bottled water3

23, 32 27% Tap only 433 19, 34 26% 133 Tap and bottled

0.42 25, 41 32% 133 Bottled only Contact with cattle past 3 months (n=698) 79 32% 22, 43 No contact 619 27% 24, 31 0.42 Table continued on next page


Table 3.11, continued p-value1

Covariate Number Percent 2) (χCI95%

Property Farming 194 30% 24, 37 Not farming 505 27% 23, 31 0.45

Contact with pigs past 3 months (n=697) 27 22% 11, 39 No contact 670 28% 24, 31 0.48 Laboratory region London 318 30% 25, 35 Hamilton 192 30% 23, 37 Kingston 55 25% 16, 38 Orillia 50 22% 12, 36 Ottawa 27 26% 11, 49 Peterborough 45 16% 8, 28 Toronto 12 33% 14, 60 0.51 Contact with horses past 3 months 94 26% 17, 36 No contact 605 28% 25, 32 0.62 Contact with cats past 3 months 392 27% 23, 32 No contact 307 29% 24, 34 0.65 Contact with dogs past 3 months (n=698) 528 28% 24, 32 No contact 170 26% 20, 34 0.67

4 Highest education in household 51 Less than high school 25% 15, 40

High school graduate 99 30% 21, 41 College or trade school 213 27% 22, 34 University 319 28% 23, 33

17 Not stated 24% 9, 48 0.96 CI: Confidence interval (95%) *N=699 unless otherwise stated 1 Variances adjusted for non-independence of observations within households 2 Lag: days between water sample collection for submission to public health laboratory for bacteriological testing and date of interview (proxy for the date of rectal swab collection) 3 Tap water only: do not used bottled water at home on regular basis (most days) Tap and bottled: glasses of water [total] > glasses of bottled water Bottled water only: glasses of water [total] = glasses of bottled water 4 Highest level of education attained by any member in household

3.3 Association of Human Carriage and Consumption of Contaminated Water

The prevalence of carriage of antimicrobial resistant E. coli, adjusted for non-independence of

observations within households, was higher in subjects living in households with untreated,

antimicrobial resistant E. coli-contaminated water (50 of 94 or 53%) than in other subjects (235 of 605

or 39%) (p=0.01). Of note, there was no difference between subjects living in households with no

history of E. coli or coliform bacterial contamination of their water for one year or longer (i.e. people

living in “B” control households), those with water that was contaminated with E. coli that was

Results 57

sensitive to the antibiotics in the NARMS panel (i.e. people living in “A” control households), or people

living in households with antimicrobial resistant E. coli-contaminated water that was treated for

bacterial contamination for one year or longer (39%, 40%, and 32%, respectively; p=0.54).

Poisson regression, adjusted for non-independence of observations of people living within the

same household, was used to determine that people living in households that used water contaminated

with antimicrobial resistant E. coli were 1.4 times (CI95% 1.1, 1.7) more likely to be colonized with

antimicrobial resistant E. coli than people living in households without antimicrobial resistant E. coli-

contaminated water sources (p=0.007; n=699 respondents from 488 households; AIC:1081; BIC:1091).

Variables that were associated with the dependent variable (carriage of resistant E. coli) at a

p-value of 0.25 or less on bivariate analyses were used to construct the multivariable models (see Table

3.12). The initial (full) model included the following variables: the focal independent variable: water

source (contaminated/not); the potential confounders: contact with cattle, respondent’s sex, travel,

contact with poultry, property type (farming/non-farming), contact with horses, household income,

antibiotic use, hospitalization history, contact with raw red meat; the hypothesized effect modifier

(tap or bottled water use) and its product term (tap or bottled water use by water source); and the

product terms of the focal independent variable by property type and by household income.

As discussed in section 2.3.2, the final model included the focal independent variable (water

source) and the following potentially confounding variables: travel outside of Canada in the previous

year, respondent’s sex, and contact with dairy or beef cattle in the previous three months (see Table

3.13). Although the potential interaction term and product term (tap or bottled water use by water

source) were also entered into the final model, they were removed as they did not improve the fit of

the model (p=0.72 for partial F-test; AIC:1081; BIC:1117). No other variables were detected that

affected the focal relationship or improved the fit of the model when the variables were added back

into the model after it was reduced.

The final model had no high leverage or outlying residual patterns and used 698 observations

from 487 clusters/households. Only one observation was not included in the final model due to item-

specific missing data. The goodness of fit test for applicability of the Poisson distribution was not

significant showing that the Poisson distribution was not over-dispersed and was appropriate for this

analysis (289). Also, the information criteria (Akaike and Bayesian) confirmed that the full model fit


the data as well as or better than the bivariate association (AIC:1076; BIC:1099 for n=698 respondents

and n=487 households).

The data were also analyzed using generalized estimating equations with exchangeable

correlation within households (clusters), a Poisson distribution, robust standard error estimates, and

population-averaged equations. The parameter estimates, standard errors, and model-building results

were virtually identical (< 0.001% difference) to those achieved using the generalized linear model

equation as described above. We chose to present the generalized Poisson regression results since the

statistical package provided a wider array of model-testing tools and because the results did not differ.

After adjusting for the effect of other variables, it was determined that people living in

households that used water contaminated with antimicrobial resistant E. coli were 1.4 times (CI95% 1.1,

1.7) more likely to be colonized with antimicrobial resistant E. coli than people living in households

without antimicrobial resistant E. coli-contaminated water sources (p<0.0001).

The risk difference or attributable risk between the subjects who were exposed and subjects

who were not exposed to water contaminated with antimicrobial resistant drinking water was 14%. The

attributable fraction, both crude (based on the crude prevalence rates) and adjusted (based on the

multivariable model), for subjects exposed to water contaminated with antimicrobial resistant E. coli

was 26%.

Based on samples tested in the participating public health laboratories during the study period,

4.5% of water samples were contaminated with E. coli, 8-50% of water sources have treatment systems

for eradicating bacteria, and 10% of the E. coli-positive samples were resistant to one or more of the

antimicrobial agents in the NARMS panel. Thus, about 8-14 cases of antimicrobial resistant E. coli

carriage per 10,000 Ontario residents using private water sources (1-2 cases per 10,000 Ontario

residents) would be attributable to drinking untreated water contaminated with antimicrobial resistant

E. coli.

Results 59

Table 3.12 Bivariate Associations between Carriage of Antimicrobial Resistant E. coli and Covariates. Poisson Regression, Variances Adjusted for Household Clustering; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006 Prevalence Standard

error

1 p-value Covariates Ratio CI95%

Focal independent variable Water source

2 Referent Not contaminated or treated Contaminated & not treated 1.37 0.16 0.007 1.09, 1.72 Potential confounders No contact with cattle Referent Contact in past 3 months 1.39 0.16 0.003 1.11, 1.74 Respondent’s sex Female Referent Male 1.25 0.11 0.009 1.06, 1.48 No travel, or only within Canada Referent Travel outside Canada, past year 1.29 0.13 0.011 1.06, 1.56 No contact with poultry Referent Contact in past 3 months 1.23 0.16 0.127 0.94, 1.59 Non-farming property Referent Farming property 1.16 0.12 0.135 0.95, 1.42 No contact with horses Referent Contact in past 3 months 1.17 0.15 0.199 0.92, 1.50 Household income, past year3 <$40,000 Referent $40,000-79,999 0.97 0.15 0.878 0.71, 1.33 $80,000 or more 1.24 0.18 0.154 0.92, 1.66 Not stated 1.16 0.18 0.356 0.85, 1.57

overall test (3 df) 0.200 No antibiotic used Referent Antibiotic in past 3 months 0.82 0.13 0.211 0.60, 1.12 No hospitalization4 Referent Hospitalized in past 12 months 0.79 0.15 0.228 0.54, 1.16 No contact with raw red meats Referent Contact in past 3 months 1.16 0.14 0.243 0.90, 1.48

5 Household education, highest Less than grade 9 Referent High school 1.24 0.33 0.429 0.73, 2.09 College or trade 1.39 0.34 0.175 0.86, 2.23 University 1.47 0.35 0.103 0.92, 2.34 Not stated 1.80 0.60 0.076 0.94, 3.45

overall test (4 df) 0.332 Respondent’s age at interview 12-49 years Referent 50-59 1.02 0.12 0.842 0.81, 1.30 60-69 0.87 0.11 0.259 0.68, 1.11 70 & older 0.85 0.12 0.268 0.64, 1.13

overall test (3 df) 0.379 CI: Confidence interval (95%) 1 Robust variance estimate, adjusted for household clustering 2 Water source not contaminated with antimicrobial resistant E. coli –OR- contaminated but treated with ultraviolet light, chlorine, boiling, or ozone for 12 months or longer –OR- contaminated with susceptible E. coli 3 Annual household income, before taxes, for all members in household (including net farm income) 4 Hospitalized for one night or longer in past 12 months Table continued on next page


Table 3.12, Continued Prevalence Standard

error1 p-value CICovariates Ratio 95%

8Lag: water submission to interview Referent 3-59 days

1.15 0.19 0.391 0.83, 1.60 60-99 1.28 0.19 0.104 0.95, 1.72 100-139 1.10 0.18 0.561 0.80, 1.50 140-199 0.20 0.18 0.240 0.89, 1.62 200-439 overall test (4 df) 0.379

Household size 1 person Referent 2 1.32 0.28 0.181 0.88, 1.99 3 or 4 1.17 0.26 0.485 0.75, 1.81 5 to 10 1.39 0.33 0.166 0.87, 2.23

overall test (3 df) 0.385 6 Interview mode

Site visit Referent Telephone interview 1.07 0.134 0.551 0.84, 1.37 Laboratory region7 London Referent Hamilton 1.14 0.13 0.218 0.92, 1.43 Kingston 1.00 0.18 0.993 0.70, 1.42 Orillia 0.85 0.17 0.410 0.58, 1.25 Ottawa 1.11 0.29 0.687 0.66, 1.87 Peterborough 0.83 0.16 0.348 0.57, 1.22 Toronto 0.83 0.32 0.634 0.40, 1.76

overall test (6 df) 0.583 No contact with raw poultry product Referent Contact in past 3 months 0.95 0.10 0.653 0.77, 1.17 No contact with pigs Referent Contact in past 3 months 1.09 0.23 0.670 0.72, 1.66 No contact with dogs Referent Contact in past 3 months 1.03 0.12 0.811 0.82, 1.28 No child or not in day care Referent Child in day care centre 1.05 0.32 0.869 0.57, 1.93 No contact with sheep/goats Referent Contact in past 3 months 1.01 0.22 0.976 0.65, 1.55 No contact with cats Referent Contact in past 3 months 1.00 0.09 0.979 0.83, 1.20 Potential effect modifier Tap or bottled water9

Referent Tap only 0.86 0.11 0.244 0.67, 1.11 Tap and bottled 1.10 0.13 0.416 0.88, 1.37 Bottled only

overall test (2 df) 0.278 CI: Confidence interval (95%) 1 Robust variance estimate, adjusted for household clustering 5 Highest level of education attained by any member in household 6 Mode of interview: site visit to home versus telephone interview 7 Laboratory to which the water sample was submitted (not necessarily household location) 8 Lag: days between water sample collection for submission to public health laboratory for bacteriological testing and date of interview (proxy for the date of rectal swab collection) 9 Tap water only: Do not used bottled water at home on regular basis (most days) Tap and bottled: Glasses of water [total] > glasses of bottled water Bottled water only: Glasses of water [total] = glasses of bottled water

Results 61

Table 3.13 Multivariable Model of Association between Carriage of Antimicrobial Resistant E. coli, Use of Water Contaminated with Antimicrobial Resistant E. coli, and Covariates. Poisson Regression, Variances Adjusted for Household Clustering; Survey of Households using Private Water Sources, Southern Ontario, 2005-2006

Prevalence ratio

Standard error1 p-value CIVariable 95%

Water source 2 Referent Not contaminated or treated

Contaminated & not treated 1.35 0.15 0.009 1.08, 1.69 Travel in past year No travel or only within Canada Referent Travelled outside Canada 1.30 0.13 0.007 1.08, 1.58 Contact with cattle No contact past 3 months Referent Contact 1.33 0.15 0.011 1.03, 1.45 Respondent’s sex Female Referent Male 1.22 0.11 0.019 1.03, 1.45

CI: Confidence interval (95%) 1 Robust variance estimate, adjusted for household clustering 2 Water source not contaminated with antimicrobial resistant E. coli –OR- contaminated but treated with ultraviolet light, chlorine, boiling, or ozone for 12 months or longer


Chapter 4

Discussion

4.1 Prevalence of Resistant E. coli

4.1.1 Ampicillin Resistant E. coli

The prevalence of faecal carriage of ampicillin resistant E. coli in 699 non-institutionalized

subjects who lived in southern Ontario, Canada and used private water sources was 28% (CI95% 24, 31),

which was greater than the hypothesized prevalence of 22%. This estimate was quite stable, with little

change noted after weighting the observations to account for the over-sampling of households with

contaminated water sources or after age-standardizing the estimate to reflect the age of the rural

Ontario population.

The prevalence estimates of this study were not statistically different than two smaller Canadian

studies, both completed in the 1990s, that also estimated the prevalence of antimicrobial resistant E.

coli by using faecal samples. Akwar et al. determined that 16% of E. coli isolates from 115 subjects who

lived on swine farms located in Ontario and British Columbia were resistant to ampicillin (105).

Bruinsma et al. reported that 22% of 154 non-institutionalized subjects living in St. John’s,

Newfoundland carried amoxicillin resistant E. coli (104). Ampicillin and amoxicillin are comparable in

that they are both bactericidal aminopenicillin agents that are used to treat systemic and urinary tract

infections caused by E. coli and other Gram-negative and Gram-positive organisms (290).

The prevalence of carriage of ampicillin resistant E. coli in our study was lower than rates from

studies of clinical and urinary tract infections. Reported rates of ampicillin resistance originating from

clinical isolates (including urinary tract isolates) ranged from 30-46% (94;95;97-99;102;291). The lower

rate of resistance in our study likely reflects the fact that the samples were drawn from a non-

institutionalized population. It is expected that higher rates of resistance would be reported from

clinical isolates because many clinical isolates originate from hospitalized individuals who are more

likely to be exposed to antimicrobial agents and to the nosocomial transmission of antimicrobial

resistant strains of bacteria (92). Also, since most uncomplicated urinary tract infections are treated

empirically, without submission of urine for laboratory analysis (97;103), laboratory-based studies

likely reflect isolates of individuals who failed initial treatment or had complicated infections, thus

biasing reported rates of antibiotic resistance (292). In addition, Laupland et al. determined that rates

Discussion 63

of antimicrobial resistance from urine samples were about 30% higher when samples from the same

patient were not excluded for at least one year of the two-year study period (293). For these reasons,

the estimated prevalence of the carriage of ampicillin resistant E. coli from our study is likely a more

accurate reflection of the prevalence of resistance in the community than that from studies based on

clinical isolates.

Although not part of the hypothesis, a preliminary examination of the factors associated with

ampicillin resistance was made. Three subgroups had different prevalences of ampicillin resistant E.

coli in this sample. First, a higher proportion of people from households with untreated water supplies

that were contaminated with antibiotic resistant E. coli carried ampicillin resistant E. coli than people

from households with water that was either not contaminated or was contaminated but treated with

chlorine, ultraviolet light, ozone, or by boiling (38% versus 26%; p=0.01). This association is explored

further in the following discussion (see Section 4.2).

Second, subjects who travelled beyond the borders of Canada within the previous year had a

higher prevalence of carriage of ampicillin resistant E. coli than subjects who did not travel or

travelled only within Canada (32% and 23%, respectively; p=0.008). In several other studies, travel to

areas with a high prevalence of human carriage of antimicrobial resistant E. coli was linked to higher

rates of colonization with antimicrobial resistant bacteria upon return from these areas (15;152;183-

185). Similary, travel has been identified as a risk factor for infection with a resistant strain of E. coli

in women with urinary tract infections (109;294). The prevalence of faecal carriage of ampicillin

resistant E. coli ranges from 73-94% in Mexico, a common destination for many of our subjects who

travelled in the previous year (106;141). It is likely that people travelling to foreign countries become

colonized with the predominant strains of that country while visiting.

Also, a higher proportion of males than females carried ampicillin resistant E. coli (32% and

23%, respectively; p=0.007). Although some other studies of ampicillin or amoxicillin resistance in E.

coli have detected a higher prevalence in males than females (102;136), the opposite has also been

reported (111), and others report no difference by sex (132;134). In our study, males and females were

equally likely to be exposed to water contaminated with antimicrobial resistant E. coli and equally

likely to have travelled outside of Canada. Although it is possible that this association was found purely


by chance, it may be that there are unmeasured factors driving the difference in the prevalence of

ampicillin resistance by sex in these subjects.

Of note, neither antibiotic use in the previous three months nor hospitalization in the previous

year were associated with a higher prevalence of carriage of ampicillin resistant E. coli in our study.

These factors have been identified as risk factors for carriage or infection with resistant E. coli in

several other studies (129;146;295-298). However, our sample size was not large enough to refine the

items further to detect an association; only 55 respondent had been hospitalized and 85 respondents

had taken an antibiotic (28 of whom had taken penicillin). The power to detect a two-fold difference in

the association between prior antibiotic use and the prevalence of ampicillin resistance was only 11%

(α = 0.05) in this study.

The results of our study, that 28% of healthy non-institutionalized people living in southern

Ontario carried E. coli that was resistant to ampicillin, is probably a less biased estimate than that

available from clinical laboratory based studies. Further, the estimate is quite stable, showing little

change when standardized to the age of the underlying population or weighted to reflect the

underlying risk of exposure to antimicrobial resistant E. coli from consumption of household water.

Although the current estimate of ampicillin resistance was not statistically significantly different than

the estimates from the earlier Canadian studies using faecal samples, the difference is clinically

relevant and likely reflects an increased prevalence of carriage of ampicillin resistant E. coli over time.

Guidelines written for the Infectious Diseases Society of America recommend changing the agents

used for the empiric treatment of urinary tract infections when the rates of resistance to an agent

reach 20% in the geographic region (103). Following these guidelines, ampicillin should not be used for

the empiric treatment of infections suspected to be caused by E. coli among residents of southern

Ontario.

4.1.2 Antimicrobial Resistance to Other Agents

The prevalence of faecal carriage of antimicrobial resistant E. coli in our study was not

statistically significantly different, when examined by individual antibiotic, than that reported in two

other studies of non-institutionalized Canadian subjects that estimated prevalence of colonization

(104;105). As shown in Table 4.1, however, the prevalence of resistance to all antibiotics except one

Discussion 65

(kanamycin) is higher in the current study. This trend of increased prevalence makes the findings

clinically relevant.

Tetracycline is not recommended for empiric use against infections caused by E. coli due to the

high rates of resistance (258). However, it is one of the three most frequently dispensed classes of

antibiotics for human use in Canada (256) and is commonly prescribed for the treatment of acne. The

finding, that 25% of subjects carried E. coli resistant to tetracycline, confirms the recommendation not

to use tetracycline against infections caused by E. coli since there is a high probability that the strain

causing the infection will be resistant to this antibiotic.

Trimethoprim-sulfamethoxazole is recommended for empiric treatment of uncomplicated

urinary tract infections in adult women as long as the prevalence of resistance to the antibiotic agent

remains below 20% (103). The prevalence in our study was below that threshold at 14% (CI95% 11, 16)

but, comparing results from this and other studies (97;299), appears to be increasing over time. Since

the use of this antibiotic is associated with the development of antimicrobial resistance (136;298), and

because urinary tract infections are quite common in women, recommendations regarding its use as

the first-line empiric treatment of uncomplicated urinary tract infections in women should be reviewed

before it reaches this threshold.

Patients with multi-drug resistant infections are at higher risk of morbidity and mortality

because of delayed or incorrect treatment with effective antimicrobial therapy (14) and there are

fewer treatment options for their treatment. Multi-drug resistance (resistance to two or more

antibiotic classes) was detected in 29% (CI95% 26, 33) of E. coli isolates in this study of residents of

households using private water sources. The most common pattern of inter-class/multi-drug resistance

in our study was between sulphonamide and beta-lactam agents, with over 70% of the sulphonamide

resistant E. coli isolates being resistant to one of the beta-lactam agents included in the study. This

pattern of resistance was also noted in other Canadian studies where 72-80% of the trimethoprim-

sulphamethoxazole resistant E. coli isolates from outpatient urine samples were also resistant to

ampicillin (98;291). Howard reported that co-resistance to ampicillin and trimethoprim-

sulphamethoxazole was associated with the use of either antibiotic (136). If this relationship is borne

out in other studies, it will need to be considered in future recommendations of treatment of

infections caused by E. coli. Although much can be learned about resistance when researching one


antimicrobial agent in isolation, the high rates of multi-drug resistance demand epidemiological and

microbiological research into this perplexing problem. Ultimately, the prevalence of multi-drug

resistance and the associated treatment issues argue for strong action to slow the rise in the

prevalence of antimicrobial resistance for all antibiotics, not just those that have high rates of

resistance at present.

Table 4.1 Comparison of Rates of Antimicrobial Resistance in E. coli from Canadian Studies 1Antibiotic Current study Akwar (105) Bruinsma (104) UTI Studies

Ampicillin 28% 16% 22% (AMX) 30-42% Tetracycline 25% 24% 16% (OXT) 19% Sulfisoxazole 24% 17% (SMX) NA NA Streptomycin 17% 10% NA NA Trimethoprim-sulphamethoxazole

14% 5% 10% (TMP) 12-19%

Nalidixic acid 6% <1% 1% 1% Chloramphenicol 5% 3% 1% NA Ciprofloxacin 3% 0 1% 0-7% Gentamicin 2% 1% 1% 1-3% Kanamycin 2% 4% NA NA Cefoxitin 1% NA NA 4% Ceftiofur 1% 0 NA NA Ceftriaxone 0 0 NA NA Amikacin 0 0 NA 1% UTI: urinary tract infection; AMX: amoxicillin; OXT: oxytetracycline; SMX: sulfamethoxazole; TMP: trimethoprim 1 (97-99;102;291;300)

4.2 Association of Human Carriage and Consumption of Contaminated Water

The prevalence of carriage of antimicrobial resistant E. coli, adjusted for non-independence of

observations, was 40 percent higher in subjects living in households that used water contaminated with

antimicrobial resistant E. coli than for people living in households that used uncontaminated or treated

water (53% vs 39%, respectively). This relationship persisted when associated covariates (travel outside

Canada in the previous year, direct contact with cattle in the previous three months, and respondent’s

sex) were included in the multivariable regression model. This finding shows that, for the people

involved in this study, the use of water contaminated with antimicrobial resistant E. coli increased the

likelihood of carriage of resistant E. coli even after accounting for other associated factors.

This research is the first to show a link between the consumption of water contaminated with

antimicrobial resistant E. coli and human carriage of resistant E. coli. Two other studies attempted to

Discussion 67

determine if an association existed between the consumption of water contaminated with, and human

carriage of, antimicrobial resistant bacteria. However, neither of the studies looked at E. coli

specifically and the findings of the studies were inconclusive.

As detailed in the literature review, Amyes et al. compared the faecal carriage of antimicrobial

resistant aerobic faecal bacteria in subjects from four villages in India in 1989 (249). The authors

concluded that the consumption of water contaminated with antimicrobial resistant bacteria coupled

with the high use of antibiotics was likely responsible for the high rates of carriage of resistant bacteria

in the inhabitants of the four villages. Although the authors presented details about the differences in

human carriage of resistant bacteria by village, no details were given about the resistance profiles of

the water supplies. Thus, no direct association between water consumption and human carriage was

made and the authors’ conclusion could not be substantiated.

Shanahan et al. studied the faecal carriage of antimicrobial resistant enterobacteria of

residents of South Africa in 1992 (248). The authors concluded that there was no association between

consuming water contaminated with bacteria and the carriage of resistant enterobacteria. However,

the water was not tested in all study locations and the bacteria present in the water that was collected

were not tested for antimicrobial resistance. Thus, it was not possible to determine if there was an

association between consuming water contaminated with antimicrobial resistant bacteria and human

carriage of resistant bacteria.

In our study, the prevalence of carriage of antimicrobial resistant E. coli was 40% higher for

individuals using water contaminated with antimicrobial resistant E. coli than for subjects consuming

uncontaminated water or water contaminated with E. coli sensitive to antibiotics. Although this

association was not unexpected, the fact that the association was still significantly associated even

after accounting for the effect of other associated variables highlights the importance of contaminated

water as a risk factor for colonization with antimicrobial resistant E. coli.

4.2.1 Strengths and Limitations of the Study

As with all studies, this one had a number of strengths and limitations. One of the strengths of

this study was that it provided a sample of people living across a wide geographic area in southern

Ontario. The results included responses from people living in Elgin County (southwestern Ontario),


Peterborough (central Ontario), Ottawa (eastern Ontario) and all counties in between. The survey

included people living in a variety of settings including farms, non-farm rural residences, villages, and

small towns. In this population, we were able to estimate the prevalence of antimicrobial resistant E.

coli in non-institutionalized individuals. This group of subjects were more representative of the general

population than studies based on clinical isolates which largely rely on isolates from individuals

screened for infections.

Another strength of this study was that the E. coli isolates tested for resistance were not

necessarily pathogenic. It is important to know the rates of resistance in all strains of E. coli since they

are all able to cause urinary tract and bloodstream infections and are able to transfer resistance to

other strains of E. coli as well as other types of bacteria. Of similar importance, commensal bacteria

are the largest reservoir for the transmission of antimicrobial resistant bacteria to other humans,

animals, and the environment (33;301).

The design of the study allowed for the systematic collection of household and personal level

variables hypothesized to be confounding or modifying to the relationship between human carriage of

resistant bacteria and the consumption of water contaminated with antimicrobial resistant E. coli. The

questionnaires were designed to collection information on variables a) identified in the literature as

being associated with antimicrobial resistance, b) influenced by the design of the study, and c) that

were common epidemiological confounders. The use of regression analyses allowed for the control of

these variables.

The sample size for this study was large enough to estimate the prevalence of antimicrobial

resistant carriage with reasonable accuracy; enough to detect clinically meaningful and statistically

significant differences in the rates of carriage. It also provided enough data for more complex

statistical analysis with the power to detect factors associated with the carriage of antimicrobial

resistant E. coli. Because our study was nested within the case-control project, we were able to

accurately locate a large enough sample of people using water contaminated with antimicrobial

resistant E. coli to allow the study of the effects of consuming contaminated water. Since the

identification of water sources contaminated with antimicrobial resistant E. coli is not routine, and

because of the low prevalence of contamination (4 per 1,000 water tests), the cost of identifying these

households would have been prohibitive for a traditional cohort or cross-sectional study.

Discussion 69

Also, our choice of control groups for the case-control project allowed us to compare subjects

who used water contaminated with antimicrobial resistant E. coli with those using water contaminated

with E. coli susceptible to antibiotics and people who used water not contaminated by coliform

bacteria for one year or longer. By determining the use of water treatment(s) and the date the

treatment was started, we were able to further refine the exposure of participants. This allowed us to

conclude that it was not the consumption of E. coli in general that was associated with carriage, but

the consumption of antimicrobial resistant E. coli, specifically.

Another strength of the study was that all subjects and interviewers in the study were blinded

to the specific exposure status of the subjects. Although subjects knew whether their household water

source was contaminated with E. coli, no information was given to them or the interviewer, about the

results of the laboratory analyses for antimicrobial susceptibility. This prevented biased responses

specific to the antimicrobial resistant status of the E. coli in the water source. Also, since subjects did

not know the antimicrobial resistance status of the E. coli, we could reasonably assume that the

differences between the two groups with E. coli contaminated water were not related to some other

unmeasured factor (e.g. disinterest in the health-related concerns of using E. coli contaminated

water).

One of the limitations of the study was that it was based on a convenience sample of people

living in households participating in the case-control project. The case-control project itself was

limited by having access only to households that submitted water samples to a participating regional

public health laboratory for bacteriological testing during the study period. Since not all households

with private water sources submit their water for bacteriological testing, the sampling frame itself may

not accurately reflect the target population: that of all Ontario households using private water sources

(302;303). However, although our response rate was only 49%, it was similar for case and control

households and for individuals within case and control households. Furthermore, we had questionnaires

and rectal swabs from people between the ages of 12 and 87 years, from a wide geographic area of

southern Ontario including both farming and non-farming households located in rural areas, villages,

and small towns without public water systems, and from households with E. coli-contaminated and

uncontaminated water.


In the same vein, we found that the mean age of the respondents in this study (58 years) was

significantly older than the 45 year average age reported for residents in rural Ontario (270). Since

there was no sampling frame of households using private water sources, it was not possible to

determine whether our sample was representative of the underlying population. However, we did

determine that younger subjects were less likely to submit a rectal swab than older participants. In our

study, there was a non-significant trend of a lower prevalence of carriage of antimicrobial resistant E.

coli in the older age groups. However, the direct age-standardized (to the 2006 estimate for rural

Ontario residents-our best estimate of the underlying population) prevalence was not significantly

different than the original estimate. Further, the prevalence of antimicrobial resistant E. coli was not

statistically different by age group in other studies estimating the prevalence of faecal carriage in

adults (46;129;130).

Another limitation of the study was that the questionnaire items, although pilot-tested,

reviewed by content experts, and based on items from other surveys, were not necessarily validated.

Although it would be ideal to validate the items, it was not an objective of this thesis to do so.

The time lag between the collection of the water sample for bacteriological testing and the

date of interview (proxy for the date of collection of the rectal swab) was another limitation of the

study. This time lag represented a potential separation between the subject’s exposure to water

contaminated with antimicrobial resistant E. coli and measurement of the outcome (rectal swab to

assess carriage of resistant E. coli). Earlier studies have reported varying lengths of colonization with

antimicrobial resistant E. coli; from days (181;213) to months (34;120;304;305). In one follow-up study,

the median time for clearance of fluoroquinolone-resistant E. coli was five months (range 2 to 10)

(149). Thus, if some of the subjects were no longer exposed to water contaminated with antimicrobial

resistant E. coli they may have cleared resistant strains of E. coli from their system before the swab

was collected. However, we know that 8.4% of the samples of antimicrobial resistant water submissions

from Hamilton and London households had already been contacted by the study signifying that the

contamination of the water sources with antimicrobial resistant E. coli was not a one-time occurrence.

Thus, although we had no measure of the length of exposure, exposure did not likely stop at the time

of the water submission. Also, statistical analysis showed that the number of days between water

Discussion 71

sampling and interview was not associated with the prevalence of human carriage. Nevertheless, we

recommend that future studies cut the time lag to reduce this potential bias.

Another limitation with our study was the inability to determine the exposure dose (i.e. the

concentration of antimicrobial resistant E. coli per quantity of water and/or the length of exposure).

Similarly, we were unable to determine the efficacy of the water treatment: were they correctly

installed and maintained? Thus, some subjects living in households with water contaminated with

antimicrobial resistant E. coli may have been incorrectly classified as using treated water (i.e.

classified as not exposed when they were exposed). We recommend that future studies collect samples

of the water from the point of consumption, as well as the source, and to do so over a period of weeks

or months to determine whether antimicrobial resistant E. coli is detected in repeated measures. It

would also be informative to determine the concentration of E. coli in the water. We note however,

given the results of this study, it would not be ethical to knowingly allow continued consumption of

water contaminated with antimicrobial resistant E. coli.

Finally, although we analysed both the water and human faecal swabs for the specific

resistance patterns to determine their relationship, it was not possible to determine causality. It is

possible that the contaminated water did not transmit resistance to humans, but rather, that both

were contaminated from a separate source (or sources) such as other humans or animals. However, we

have included a number of variables in the regression models to help rule out confounding, including

variables hypothesized to be associated with carriage of resistant bacteria in other research (antibiotic

use, hospitalization, travel, child in day care, household size, contact with livestock, farming, contact

with pets, contact with raw meat), demographic variables (age, sex, education, income), and study

design variables (laboratory region, mode of data collection, days between water sample and

interview).

The probability of reverse causation; that the people colonized with antimicrobial resistant E.

coli contaminated their water supply rather than the opposite (that the people carrying antimicrobial

resistant E. coli were infected through contaminated water), is low. First, the association between

human carriage of antimicrobial resistant E. coli and the use of water contaminated with resistant E.

coli was not significant (relative risk: 1.1; p=0.19); the association was dependent on the use of

untreated water (see section 3.3). Also, for human sewage to contaminate the water source, the


bacteria would have to be transported from the septic system to the water source. Analysis of factors

associated with water source contamination (not presented in this thesis) determined that there was

no association between the distance between the water source and the septic system and the presence

of antimicrobial resistant E. coli in the water sample (p=0.61). Thus, it is more likely that the

contaminated water was the source of colonization for the subjects in this study than visa versa.

4.3 Conclusions

4.3.1 Contaminated water

In this study, the consumption of water contaminated with antimicrobial resistant E. coli was

associated with carriage of antimicrobial resistant E. coli. Although this association may seem obvious,

our findings confirm the association. In Canada, the microbiological quality of drinking water is

determined through testing for the presence of E. coli since it indicates faecal contamination of the

water supply. The Canadian drinking water guidelines state that all systems should be tested and that

the maximum acceptable concentration is zero colony forming units of E. coli per 100 mL for public,

semi-public, and private drinking water systems (306;307). In our study, 4.5% of water samples

submitted for bacteriological testing were contaminated with E. coli and an additional 18% were

contaminated with coliform bacteria above the accepted limit. Yet, according to earlier studies, only

8-26% of the households with private water supplies treat the water to destroy bacteria (68;73;75).

Since only 50-60% of households in recent surveys tested their water in the year before the interview

(73;266), it is likely that many household residents don’t even know they are consuming contaminated

water, which may also be contaminated with other pathogenic bacterial species such as Salmonella,

Shigella, Campylobacter, or Legionella (308;309). This would put them at increased risk of contracting

a gastrointestinal illness and for acquiring antimicrobial resistant bacteria, including E. coli. However,

since there is no registry of private water sources, and thus no denominator data, it makes it very

difficult to quantify the size of the problem with any accuracy.

There are several steps required to reach the goal of treating all contaminated water sources.

First, is a registry of all public and private water systems. Second, is the routine monitoring of these

systems through bacteriological testing of the water. Third, is the follow-up of systems that have

significant levels of bacterial contamination. Not only would this reduce the number of cases of

Discussion 73

gastrointestinal illness caused by waterborne bacteria, but would reduce the transmission of

antimicrobial resistant E. coli and, perhaps, the transmission of other resistant waterborne bacteria.

Preventing people from becoming colonized with antimicrobial resistant E. coli is important for

several reasons. First, people carrying antimicrobial resistant E. coli have the potential to develop

antimicrobial resistant infections if the bacteria infect the carrier’s urinary tract, a wound, or blood

system. Second, the carriage of resistant E. coli increases the probability of transferring resistance

genes to other types of bacteria within the individual’s gastrointestinal system. In addition, carriage of

resistant bacteria creates a reservoir for the transmission of resistant bacteria and resistance genes to

other humans, other mammals, and the environment.

The impact of the transmission of antimicrobial resistance through the ingestion of

contaminated drinking water is not inconsequential. To put this issue in perspective, 1.5 to 2 million

Ontario residents and over 4 million Canadians rely on private drinking water sources (62;71;73). In

Ontario only about 1 case of antimicrobial resistant E. coli carriage per 1,000 residents who use private

water sources can be attributed to drinking untreated water contaminated with antimicrobial resistant

E. coli. However, there are hundreds of millions of people across the globe, many of them in less

developed countries without the resources to treat drinking water, who are exposed to water

contaminated with E. coli and other bacteria that are resistant to frequently-prescribed antibiotics

(234;241-244;249). If the attributable fraction was similar in a developing country as it was in Ontario,

the population attributable risk fraction would be 60 to 260 per 1,000 resident of India, for example,

where even the “treated” water is often contaminated with antimicrobial-resistant bacteria (310).

Since people who ingest water contaminated with antimicrobial resistant E. coli are more likely to

carry resistant E. coli than people who consume uncontaminated water, the list of strategies to reduce

the prevalence and transmission of antimicrobial resistance needs to include the adequate treatment

of contaminated water.

4.3.2 Prevalence of Carriage of Resistant E. coli

The prevalence of antimicrobial resistant E. coli appears to be increasing in Canada, a nation

with relatively strict regulations on the distribution of antimicrobial agents (see Table 4.1). The

findings, that 28% of rural Ontario residents were colonized with ampicillin resistant E. coli and over


40% were colonized with E. coli resistant to one or more antibiotics, is a clear indication that antibiotic

resistance has made its way from the clinical setting to the rural population.

Although the prevalence of carriage of antimicrobial resistant E. coli in this study was

estimated among subjects that used private drinking water sources, and thus from rural areas, the

findings may be generalizable to the Ontario population in general. Although rural and urban residents

have somewhat varied colonization pressures due to different human and animal population densities,

the subjects included in our sample were not from remote rural areas. Thus, they likely had similar

access to medical treatment, antimicrobial agents, travel, and food sources as residents of urban

Ontario. Study subjects also included people living in urban areas who used private water sources at

their cottage and people living in rural areas and urban fringes but who worked in larger urban centres.

The colonization pressures for these subjects would not differ significantly from residents of urban

centres. Also, the majority of subjects were not exposed to water contaminated with antimicrobial

resistant E. coli (605 of 699 were not exposed). As noted above, the attributable risk fraction is low, at

1 per 1,000 people using private water sources. Thus, the drinking water exposure of most subjects was

similar to that of residents of urban centres who use municipally treated water. It is probable that the

prevalence of carriage of antimicrobial resistant E. coli is comparable for urban and rural residents of

southern Ontario.

It is troubling that the prevalence has reached this height in Ontario and that it is increasing

despite the plethora of literature about the issue. Antibiotic resistance threatens the management of

infectious diseases through increased morbidity and mortality. People infected with resistant strains of

bacteria are more likely to be hospitalized, have longer hospital stays, and are more likely to die than

people with susceptible strains of the same bacteria (9-13).

We strongly encourage a reduction in the use of antibiotics for animal, agricultural, and human

use to reduce the effects of selective pressure on the development of antimicrobial resistance. This,

along with infection control techniques to reduce the transmission of resistant bacteria will help slow

the rising prevalence of resistant bacteria (16;311;312). Physicians, nurse practitioners, dentists, and

veterinarians must prescribe antibiotics only for bacterial infections, use laboratory tests to confirm

the cause of the infection, and change the prescription to a more effective (often more narrow

spectrum) antibiotic if necessary (22). The development of rapid diagnostic tests may aid in the early

Discussion 75

diagnosis of disease thereby reducing unnecessary and/or ineffective antibiotic use (8;55). Also,

consulting infectious disease specialists would improve the rational use of antimicrobial agents, help

prevent infections, and reduce the transmission of infectious agents (23).

Health care providers must also communicate effectively with patients to ensure that

antimicrobial agents are used properly (i.e. completing the full course of antibiotics and not using

other people’s prescriptions)(15). Guidelines for the use of antibiotics should also be written,

distributed, and updated to help clinicians treat their patients effectively in the ever-changing world of

infectious disease medicine (27;313).

Research needs to continue into the development of new antimicrobial agents. It is also

essential that we learn the factors associated with the emergence, persistence, and transmission of

antimicrobial resistance. Ongoing surveillance of antibiotic use and antimicrobial resistance in humans,

animals, and the environment could provide timely information on the state of antimicrobial resistance

in Canada (25;83;314). Research into vaccines and other preventive mechanisms, including public

health measures to reduce the transmission of infectious organisms, would also help reduce the need

for antimicrobial agents, thereby reducing our dependence on an ever-dwindling collection of effective

antibiotics (8;55).

4.3.2 Next steps

We plan to publish several articles based on the results of this thesis including an already-

drafted article on the difference in participation by mode of interview (face-to-face versus telephone).

We plan to submit articles on the prevalence of antimicrobial resistant E. coli in this sample of healthy

subjects using private water sources and one on the association between carriage of resistant bacteria

and the consumption of water contaminated with antimicrobial resistant E. coli. A submission has been

accepted to present a poster at the Interscience Conference on Antimicrobial Agents and

Chemotherapy conference in October 2008. The PhD student will also be involved in the publication of

the case-control project and other articles based on the personal and household questionnaires.

Beyond academic outputs, there are several avenues available to influence policy development

within the province. The student has already presented results to staff at the Public Health Laboratory

and will be drafting a brief to be given to Fred Ruf, the director of the Safe Water Unit in the Ministry


of Health and Long-Term Care and to all medical officers of health of Ontario. Since drinking water is

under the jurisdiction of more than one ministry in Ontario, we will also provide a synopsis of the

findings to the Ministry of the Environment and Health Canada. The goal is to provide information that

will help guide informed decision-making and aid in the provision of safe water to all residents in the

province of Ontario and Canada.

References

References

(1) Lewis R. The rise of antibiotic-resistant infections. US Food and Drug Administration 1995 [cited 2005 Mar 23];Available from: URL: http://www.fda.gov

(2) Statement of J.P. Koplan, Director, Centers for Disease Control and Prevention, Department of Health and Human Services before the Committee on appropriations: Hearings Before the SubComm. on Labor, Health, and Human Services of the House Comm. on Appropriations, Koplan JP, (Sep 20, 2000).

(3) Weber JT, Courvalin P. An emptying quiver: Antimicrobial drugs and resistance. Emerging Infectious Diseases 2005;11(6):791-3.

(4) Mehrotra M, Dougherty J, Poppe C. Antimicrobial resistance: What is it? Ottawa, Ontario: Health Canada; 2003 Jun. Report No.: 6.

(5) World Health Organization. WHO global strategy for containment of antimicrobial resistance. Switzerland: World Health Organization; 2001.

(6) Okeke IN, Laxminarayan R, Bhutta ZA, et al. Antimicrobial resistance in developing countries. Part I: Recent trends and current status. Lancet Infection 2005;5:481-93.

(7) Kümmerer K. Resistance in the environment. J Antimicrob Chemother 2004;54:311-20.

(8) McCormick JB. Epidemiology of emerging/re-emerging antimicrobial-resistant bacterial pathogens. Current Opinions in Microbiology 1998;1:125-9.

(9) Lautenbach E, Metlay JP, Bilker WB, Edelstein PH, Fishman NO. Association between fluoroquinolone resistance and mortality in Escherichia coli and Klebsiella pneumoniae infections: The role of inadequate empirical antimicrobial therapy. Clinical Infectious Diseases 2005;41:923-9.

(10) Cosgrove SE, Carmeli Y. The impact of antimicrobial resistance on health and economic outcomes. Clinical Infectious Diseases 2003;36:1433-7.

(11) Linares L, Cervera C, Cofán F, et al. Epidemiology and outcomes of multiple antibiotic-resistant bacterial infection in renal transplantation. Transplantation Proceedings 2007;39:2222-4.

(12) French GL. Clinical impact and relevance of antibiotic resistance. Advanced Drug Delivery Reviews 2005;57:1514-27.

(13) Cosgrove SE. The relationship between antimicrobial resistance and patient outcomes: Mortality, length of hospital stay, and health care costs. Clinical Infectious Diseases 2006;42:S82-S89.

(14) Peralta G, Sánchez MB, Garrido JC, et al. Impact of antibiotic resistance and of adequate empirical antibiotic treatment in the prognosis of patients with Escherichia coli bacteraemia. J Antimicrob Chemother 2007;60(4):855-63.

(15) Gin A, Zhanel GG. Antimicrobial resistance: Endgame for antibiotics? A CE lesson…. Canadian Pharmaceutical Journal 2001.

(16) Conly JM. Antimicrobial resistance in Canada. Canadian Medical Association Journal 2002;167(8):885-91.

(17) Emerson DL. Government of Canada invests in research to combat antibiotic resistance. In: Fox C, ed.Vancouver, B.C.: Industry Canada, 2005.

77

http://www.fda.gov/

Appendices

(18) Levy SB. Antibiotic resistance - The problem intensifies. Advanced Drug Delivery Reviews 2005;57:1446-50.

(19) Norrby SR, Nord CE, Finch R. Lack of development of new antimicrobial drugs: A potential serious threat to public health. Lancet Infection 2005;5:115-9.

(20) Mølbak K. Human health consequences of antimicrobial drug-resistant Salmonella and other foodborne pathogens. Clinical Infectious Diseases 2005;41:1613-20.

(21) Singh KV, Reves R, Pickering L, Murray BE. Comparative in vitro activities of amoxicillin-clavulanic acid, cefuroxime, cephalexin, and cephalothin against trimethoprim-resistant Escherichia coli isolated from stools of children attending day-care centers. Antimicrobial Agents and Chemotherapy 1990;34(11):2047-9.

(22) Patterson JE. Multidrug-resistant Gram-negative pathogens: Multiple approaches and measures for prevention. Infection Control and Hospital Epidemiology 2006;27(9):889-973.

(23) Jansen WTM, van der Bruggen JT, Verhoef J, Fluit AC. Bacterial resistance: A sensitive issue. Complexity of the challenge and containment strategy in Europe. Drug Resistance Updates 2006;(9):123-33.

(24) O'Brien TF. The global epidemic nature of antimicrobial resistance and the need to monitor and manage it locally. Clinical Infectious Diseases 1997;24(S1):S2-S8.

(25) Beovic B. The issue of antimicrobial resistance in human medicine. International Journal of Food Microbiology 2006;112(3):280-7.

(26) vonBaum H, Marre R. Antimicrobial resistance of Escherichia coli and therapeutic implications. International Journal of Medical Microbiology 2005;295:503-11.

(27) Levy SB. Factors impacting on the problem of antibiotic resistance. J Antimicrob Chemother 2002;49:25-30.

(28) Pop-Vicas AE, D'Agata E. The rising influx of multidrug-resistant Gram-negative bacilli into a tertiary care hospital. Clinical Infectious Diseases 2005;40:1792-8.

(29) Samore MH, Magill M, Alder S, et al. High rates of multiple antibiotic resistance in Streptococcus pneumoniae from healthy children living in isolated rural communities: Association with cephalosporin use and intrafamilial transmission. Pediatrics 2001;108(4):856-65.

(30) Putnam SD, Sanders JW, Tribble DR, et al. Posttreatment changes in Escherichia coli antimicrobial susceptibility rates among diarrheic patients treated with ciprofloxacin. Antimicrobial Agents and Chemotherapy 2005;49(6):2571-2.

(31) Arason VA, Kristinsson K, Sigurdsson J, Stefansdottir G, Molstad S, Gudmundsson S. Do antimicrobials increase the carriage rate of penicillin resistant pneumococci in children? Cross sectional prevalence study. British Medical Journal 1996;313(387):391.

(32) Seaton RA, Steinke D, Phillips G, MacDonald T, Davey PG. Community antibiotic therapy, hospitalization and subsequent respiratory tract isolation of Haemophilus influenzae resistant to amoxicillin: A nested case-control study. J Antimicrob Chemother 2000;46:307-9.

(33) Donskey CJ. The role of intestinal tract as a reservoir and source for transmission of nosocomial pathogens. Healthcare Epidemiology 2004;39:219-26.

(34) Vollaard EJ, Clasener HAI. Colonization resistance. Antimicrobial Agents and Chemotherapy 1994;38(3):409-14.

78

References

(35) O'Brien TF. Emergence, spread, and environmental effect of antimicrobial resistance: How use of an antimicrobial anywhere can increase resistance to any antimicrobial anywhere else. Clinical Infectious Diseases 2002;34(Suppl 3):S78-S84.

(36) Bryan A, Shapir N, Sadowsky MJ. Frequency and distribution of tetracycline resistance genes in genetically diverse, nonselected, and nonclinical Escherichia coli strains isolated from diverse human and animal sources. Applied and Environmental Microbiology 2004;70(4):2503-7.

(37) Chopra I, Roberts M. Tetracycline antibiotics: mode of action, applications, molecular biology, and epidemiology of bacterial resistance. Microbiology and Molecular Biology Reviews 2001;65(2):232-60.

(38) Zhao S, White DG, Ge B, et al. Identification and characterization of integron-mediated antibiotic resistance among shiga toxin-producing Escherichia coli isolates. Applied and Environmental Microbiology 2001;67(4):1558-64.

(39) Hawkey PM. Molecular epidemiology of clinically significant antibiotic resistance genes. British Journal of Pharmacology 2008;153:S406-S413.

(40) Carattoli A. Importance of integrons in the diffusion of resistance. Veterinary Research 2001;32:243-59.

(41) Sunde M. Prevalence and characterization of class 1 and class 2 integrons in Escherichia coli isolated from meat and meat products of Norwegian origin. J Antimicrob Chemother 2005;56:1019-24.

(42) Normark BH, Normark S. Evolution and spread of antibiotic resistance. Journal of Internal Medicine 2002;252(2):91-106.

(43) Bonten M, Stobberingh E, Philips J, Houben A. High prevalence of antibiotic resistant Escherichia coli in faecal samples of students in the south-east of the Netherlands. J Antimicrob Chemother 1990;26:585-92.

(44) Shannon K, French GL. Increasing resistance to antimicrobial agents of gram-negative organisms isolated at a London teaching hospital, 1995-2000. J Antimicrob Chemother 2004;53:818-25.

(45) Reves R, Fong M, Pickering L, Bartlett A, Alvarez M, Murray BE. Risk factors for fecal colonization with trimethoprim-resistant and multiresistant Escherichia coli among children in day-care centers in Houston, Texas. Antimicrobial Agents and Chemotherapy 1990;34(7):1429-34.

(46) Hannah EL, Angulo F, Johnson JR, Haddadin B, Williamson J, Samore MH. Drug-resistant Escherichia coli, rural Idaho. Emerging Infectious Diseases 2005;11(10):1614-7.

(47) Threlfall EJ, Ward L, Frost J, Willshaw G. The emergence and spread of antibiotic resistance in food-borne bacteria. International Journal of Food Microbiology 2000;62:1-5.

(48) Levy SB. Starting life resistance free. New England Journal of Medicine 1990;323(5):335-7.

(49) Sørum H, L'Abée-Lund TM. Antibiotic resistance in food-related bacteria a result of interfering with the global web of bacterial genetics. International Journal of Food Microbiology 2002;78:43-56.

(50) Weir E. Escherichia coli O157:H7. Canadian Medical Association Journal 2000;163(2):205.

(51) Russo TA, Johnson JR. Medical and economic impact of extraintestinal infections due to Escherichia coli: Focus on an increasingly important endemic problem. Microbes and Infection 2003;5:449-56.

(52) Jones ME, Draghi D, Thornsberry C, Karlowsky JA, Sahm DF, Wenzel R. Emerging resistance among bacterial pathogens in the intensive care unit - a European and North American surveillance study (2000-2002). Annals of Clinical Microbiology and Antimicrobials 2004;3(14).

79

Appendices

(53) Control of Communicable Diseases Manual. 16 ed. Washington: American Public Health Association, 1995.

(54) Percival SL, Chalmers RM, Embrey M, Hunter P, Sellwood J, Wyn-Jones P. Escherichia coli. Microbiology of Waterborne Diseases.New York, NY: Academic Press, 2004.

(55) Cohen ML. Epidemiology of drug resistance: Implications for a post-antimicrobial era. Science 1992;257(5073):1050-5.

(56) Lipsitch M, Samore MH. Antimicrobial use and antimicrobial resistance: A population perspective. Emerging Infectious Diseases 2002;8(4):347-54.

(57) Linton AH. Flow of resistance genes in the environment and from animals to man. J Antimicrob Chemother 1986;18(Supp C):189-97.

(58) Jackson SG, Goodbrand R, Johnson R, et al. Escherichia coli O157:H7 diarrhoea associated with well water and infected cattle on an Ontario farm. Epidemiology and Infection 1998;120:17-20.

(59) Schets FM, During M, Italiaander R, et al. Escherichia coli O157:H7 in drinking water from private water supplies in the Netherlands. Water Research 2005;39:4485-93.

(60) Bruce-Grey-Owen Sound Health Unit. The investigative report on the Walkerton outbreak of waterborne gastroenteritis May-June 2000. Owen Sound, Ontario: Bruce-Grey-Owen Sound Health Unit; 2000 Oct 10.

(61) Government of Ontario. Safe drinking water act, 2002. Ontario regulation 169/03 ed. 2002.

(62) Ontario Ministry of the Environment. Chief drinking water inspector annual report 2005-2006. Toronto, ON: Queen's Printer for Ontario; 2007.

(63) Coleman BL. Private water source tests for bacteriological contamination. 2007.

(64) Goss MJ, Barry D, Rudolph D. Contamination of Ontario farmstead domestic wells and its association with agriculture: 1. Results from drinking water wells. Journal of Contaminant Hydrology 1998;32:267-93.

(65) Brant County Health Unit. State of the environment report: water quality. Brantford, ON: Brant County Health Unit; 2007.

(66) Ontario farm groundwater quality survey. Summer 1992. Waterloo, ON; 1993.

(67) Lonc.W., Parkhouse P, Carson RM. Middlesex-London migrant farm housing drinking water quality study. London, ON: Middlesex-London Health Unit; 1994. Report No.: 2.

(68) Halton Region Health Department. Halton Region Health Department Rural Drinking Water Survey. Oakville, Ontario: Halton Region Health Department; 2000. Report No.: MO-03-01.

(69) Moccio P. Status of drinking water for the Kingston, Frontenac and Lennox & Addington area. Kingston, ON: Kingston, Frontenac and Lennox & Addington Health Unit; 2000.

(70) Tries D. Water quality report for Haldimand-Norfolk, 2004. 2007.

(71) Corkal D, Schutzman W, Hilliard C. Rural water safety from the source to the on-farm tap. Journal of Toxicology and Environmental Health, Part A 2004;67:1619-42.

(72) Statistics Canada. Community profiles. 2006 Census. Statistics Canada 2008 [cited 2008 Jan 16];Available from: URL: www.statscan.ca

80

http://www.statscan.ca/

References

(73) Statistics Canada. Households and the environment. 2006. 2007.

(74) Simpson H. Promoting the management and protection of private water wells. Journal of Toxicology and Environmental Health, Part A 2004;67:1679-704.

(75) Jones AQ, Dewey C, Dore KA, et al. Public perceptions of drinking water: A postal survey of residents with private water supplies. BMC Public Health 2006;6(1):94-104.

(76) Government of Ontario. Get acquainted with your well.Toronto, Ontario: Government of Ontario, 2002.

(77) Government of Ontario. Choosing a water treatment system.Toronto: Government of Ontario, 2002.

(78) Health Canada. Water treatment devices for disinfection of drinking water. Health Canada 2008 January 7 [cited 2008 Jan 16];Available from: URL: http://www.hc-sc.gc.ca

(79) Brueggemann AB. Antibiotic resistance mechanisms among pediatric respiratory and enteric pathogens. The Pediatric Infectious Disease Journal 2006;25(10):969-73.

(80) Österblad M, Hakanen A, Manninen R, et al. A between-species comparison of antimicrobial resistance in enterobacteria in fecal flora. Antimicrobial Agents and Chemotherapy 2000;44(6):1479-84.

(81) Oppegaard H, Steinum T, Wasteson Y. Horizontal transfer of a multi-drug resistance plasmid between coliform bacteria of human and bovine origin in a farm environment. Applied and Environmental Microbiology 2001;67(8):3732-4.

(82) Houndt T, Ochman H. Long-term shifts in patterns of antibiotic resistance in enteric bacteria. Applied and Environmental Microbiology 2000;66(12):5406-9.

(83) National Academy of Sciences. Challenges for the development of new antimicrobials - rethinking the approaches: Report of a workshop. 2005.

(84) Bartoloni A, Pallecchi L, Benedetti M, et al. Multidrug-resistant commensal Escherichia coli in children, Peru and Bolivia. Emerging Infectious Diseases 2006;12(6):907-13.

(85) vandenBogaard AE, Stobberingh E. Antibiotic usage in animals. Impact on bacterial resistance and public health. Drugs 1999;58(4):589-603.

(86) Erb A, Stürmer T, Marre R, Brenner H. Prevalence of antibiotic resistance in Escherichia coli: Overview of geographical, temporal, and methodological variations. European Journal of Clinical Microbiology & Infectious Diseases 2007;26(2):83-90.

(87) Roels TH, Proctor ME, Robinson LC, Hulbert K, Bopp CA, Davis JP. Clinical features of infections due to Escherichia coli producing heat-stable toxin during an outbreak in Wisconsin: a rarely suspected cause of diarrhea in the United States. Clinical Infectious Diseases 1998;26:898-902.

(88) Moreno E, Prats G, Sabaté, Pérez T, Johnson JR, Andreu A. Quinolone, fluoroquinolone and trimethoprim/sulfamethoxazole resistance in relation to virulence determinants and phylogenetic background among uropathogenic Escherichia coli. J Antimicrob Chemother 2006;57(2):204-11.

(89) Jauréguy F, Carbonnelle E, Bonacorsi S, et al. Host and bacterial determinants of initial severity and outcome of Escherichia coli sepsis. Journal of Infectious Diseases 2007;190(10):1739-44.

(90) Johnson JR, Kuskowski MA, Gajweski A, Sahm DF, Karlowsky JA. Virulence characteristics and phylogenetic background of multidrug-resistant and antimicrobial-susceptible clinical isolates of Escherichia coli from across the United States, 2000-2001. Journal of Infectious Diseases 2004;190(10):1739-44.

81

http://www.hc-sc.gc.ca/

Appendices

(91) Grude N, Potaturkina-Nesterova NI, Jenkins A, et al. A comparison of phylogenetic group, virulence factors and antibiotic resistance in Russian and Norwegian isolates of Escherichia coli from urinary tract infection. Clinical Microbiology and Infection 2007;13(2):208-11.

(92) Bagger-Skjøt L, Sandvang D, Frimodt-Møller N, et al. Association between antimicrobial resistance and virulence genes in Escherichia coli obtained from blood and faeces. Scandanavian Journal of Infectious Diseases 2007;39:724-7.

(93) Summers AO. Generally overlooked fundamentals of bacterial genetics and ecology. Clinical Infectious Diseases 2002;34(S3):S85-S92.

(94) London Laboratory Services Group. Antimicrobial susceptibilities 2003. London, Ontario: London Laboratory Services Group, Department of Microbiology; 2003.

(95) Pfaller MA, Jones R, Doern G, Kugler K, and the Sentry Participants Group. Bacterial pathogens isolated from patients with bloodstream infection: Frequencies of occurrence and antimicrobial susceptibility patterns from the SENTRY antimicrobial surveillance program (United States and Canada, 1997). Antimicrobial Agents and Chemotherapy 1998;42(7):1762-70.

(96) Gupta K, Hooten TM, Stamm WE. Increasing antimicrobial resistance and the management of uncomplicated community-acquired urinary tract infections. Annals of Internal Medicine 2001;135:41-50.

(97) McIsaac WJ, Mazzulli T, Permaul J, Moineddin R, Low DE. Community-acquired antibiotic resistance in urinary isolates from adult women in Canada. The Canadian Journal of Infectious Diseases & Medical Microbiology 2006;17(6):337-40.

(98) Zhanel GG, Karlowsky JA, Harding GKM, et al. A Canadian national surveillance study of urinary tract isolates from outpatients: Comparison of the activities of trimethoprim-sulfamethoxazole, ampicillin, mecillinam, nitrofurantoin, and ciprofloxacin. Antimicrobial Agents and Chemotherapy 2000;44(4):1089-92.

(99) Kahlmeter G. An international survey of the antimicrobial susceptibility of pathogens from uncomplicated urinary tract infections: the ECO-SENS Project. J Antimicrob Chemother 2003 Jan 1;51(1):69-76.

(100) Allen UD, MacDonald N, Fuite L, Chan F, Stephens D. Risk factors for resistance to "first-line" antimicrobials among urinary tract isolates of Escherichia coli in children. Canadian Medical Association Journal 1999;160(10):1436-40.

(101) Kahlmeter G, Menday P. Cross-resistance and associated resistance in 2478 Escherichia coli isolates from the Pan-European ECO-SENS project surveying the antimicrobial susceptibility of pathogens from uncomplicated urinary tract infections. J Antimicrob Chemother 2003;52:128-31.

(102) Zhanel GG, Hisanaga T, Laing N, et al. Antibiotic resistance in outpatient urinary isolates: Final results from the North American urinary tract infection collaborative alliance (NAUTICA). International Journal of Antimicrobial Agents 2005;26(5):380-8.

(103) Warren JW, Abrutyn E, Hebel JR, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clinical Infectious Diseases 1999;29:745-58.

(104) Bruinsma N, Hutchinson J, vandenBogaard A, Giamarellou H, Degener J, Stobberingh E. Influence of population density on antibiotic resistance. J Antimicrob Chemother 2003;51(2):385-90.

(105) Akwar TH, Poppe C, Wilson J, et al. Risk factors for antimicrobial resistance among fecal Escherichia coli from residents on forty-three swine farms. Microbial Drug Resistance 2007;13(1):69-76.

82

References

(106) Nys S, Okeke I, Kariuki S, Dinant G, Driessen D, Stobberingh E. Antibiotic resistance of faecal Escherichia coli from healthy volunteers from eight developing countries. J Antimicrob Chemother 2004;54:952-5.

(107) Hsueh P, Chen W, Luh K. Relationships between antimicrobial use and antimicrobial resistance in Gram-negative bacteria causing nosocomial infections from 1991-2003 at a university hospital in Taiwan. International Journal of Antimicrobial Agents 2005;26:463-72.

(108) McGeer AJ, Fleming CA, Green K, Willey BM, Low DE. Antimicrobial resistance in Ontario - are we facing defeat? Quality Management Program - Laboratory Services News. 2004 Oct. Report No.: 83.

(109) Burman WJ, Breese P, Murray BE, et al. Conventional and molecular epidemiology of trimethoprim-sulfamethoxazole resistance among urinary Escherichia coli isolates. The American Journal of Medicine 2003;115(5):358-64.

(110) Brumfitt W, Faiers MC, Reeves DS, Datta N. Antibiotic-resistant Escherichia coli causing urinary-tract infection in general practice: Relation to faecal flora. Lancet 1971;297(7694):315-7.

(111) Hay AD, Thomas M, Montgomery A, et al. The relationship between primary care antibiotic prescribing and bacterial resistance in adults in the community: A controlled observational study using individual patient data. J Antimicrob Chemother 2005;56(1):146-53.

(112) Butler CC, Hillier S, Roberts Z, Dunstan F, Howard A, Palmer S. Antibiotic-resistant infections in primary care are symptomatic for longer and increase workload: outcomes for patients with E. coli UTIs. British Journal of General Practice 2006;56:686-92.

(113) Stelling JM, Travers K, Jones R, Turner P, O'Brien R, Levy SB. Integrating Escherichia coli antimicrobial susceptibility data from multiple surveillance programs. Emerging Infectious Diseases 2005;11(6):873-82.

(114) Reinert RR, Low DE, Rossi F, Zhang X, Wattal C, Dowzicky MJ. Antimicrobial susceptibility among organisms from the Asia/Pacific rim, Europe and Latin and North America collected as part of TEST and the in vitro activity of tigecycline. J Antimicrob Chemother 2007;60:1018-29.

(115) Vatopoulos AC, Varvaresou E, Petridou E, et al. High rates of antibiotic resistance among normal fecal flora Escherichia coli in children from Greece. Clinical Microbiology and Infection 1998;4(10):563-9.

(116) Bonten M, Stobberingh E, Philips J, Houben A. Antibiotic resistance of Escherichia coli in fecal samples of healthy people in two different areas in an industrialized country. Infection 1992;20(5):258-62.

(117) Bartoloni A, Bartalesi F, Mantella A, et al. High prevalence of acquired antimicrobial resistance unrelated to heavy antimicrobial consumption. The Journal of Infectious Diseases 2004;189:1291-4.

(118) Domínguez E, Zarazaga M, Sáenz Y, Briñas L, Torres C. Mechanisms of antibiotic resistance in Escherichia coli isolates obtained from healthy children in Spain. Microbial Drug Resistance 2002;8(4):321-7.

(119) Calva J, Sifuentes-Osornio J, Cerón C. Antimicrobial resistance in fecal flora: Longitudinal community-based surveillance of children in urban Mexico. Antimicrobial Agents and Chemotherapy 1996;40(7):1699-702.

(120) Karami N, Nowrouzian F, Adlerberth I, Wold AE. Tetracycline resistance in Escherichia coli and persistence in the infantile colonic microbiota. Antimicrobial Agents and Chemotherapy 2006;50(1):156-61.

(121) London N, Nijsten R, vandenBogaard A, Stobberingh E. Carriage of antibiotic-resistant Escherichia coli by healthy volunteers during a 15-week period. Infection 1994;22(3):187-91.

83

Appendices

(122) Levy SB, Marshall BJ, Schluederberg S, Rowse D, Davis J. High frequency of antimicrobial resistance in human fecal flora. Antimicrobial Agents and Chemotherapy 1988;32:1801-6.

(123) van de Mortel HJ, Jansen EJ, Dinant GJ, London N, Palacios Prü E, Stobberingh EE. The prevalence of antibiotic-resistant faecal Escherichia coli in healthy volunteers in Venezuela. Infection 1998;26(5):292-7.

(124) Walson JL, Marshall BJ, Pokhrel B, Kafle K, Levy SB. Carriage of antibiotic-resistant fecal bacteria in Nepal reflects proximity to Kathmandu. The Journal of Infectious Diseases 2001;184:1163-9.

(125) Karlowsky JA, Kelly LJ, Thornsberry C, Jones ME, Sahm DF. Trends in antimicrobial resistance among urinary tract infection isolates of Escherichia coli from female outpatients in the United States. Antimicrobial Agents and Chemotherapy 2002;46(8):2540-5.

(126) Karlowsky JA, Thornsberry C, Jones ME, Sahm DF. Susceptibility of antimicrobial-resistant urinary Escherichia coli isolates to fluoroquinolones and nitrofurantoin. Clinical Infectious Diseases 2003;36:183-7.

(127) Lester SC, Del Pilar Pla M, Wang F, Perez Schael I, Jiang H, O'Brien TF. The carriage of Escherichia coli resistant to antimicrobial agents by healthy children in Boston, in Caracas, Venezuela, and in Qin Pu, China. New England Journal of Medicine 1990;323(5):285-9.

(128) Storby K, Österlund A, Kahlmeter G. Antimicrobial resistance in Escherichia coli in urine samples from children and adults: A 12 year analysis. Acta Peaediatrica 2004;93:487-91.

(129) Degener JE, Smit ACW, Michel MF, Valkenburg HA, Muller L. Resistance of Escherichia coli in the faeces and the use of antimicrobial agents in the treatment of hospital patients. Journal of Hygiene 1983;91:491-8.

(130) Degener JE, Smit ACW, Michel MF, Valkenburg HA, Muller L. Faecal carriage of aerobic Gram-negative bacilli and drug resistance of Escherichia coli in different age-groups in Dutch urban communities. Journal of Medical Microbiology 1983;16:139-45.

(131) Becnel-Boyd L, Atmar RL, Randall GL, Hamill RJ, Steffen D, Zechiedrich L. Increased fluoroquinolone resistance with time in Escherichia coli from >17,000 patients at a large county hospital as a function of culture site, age, sex, and location. BMC Infectious Diseases 2008;8(4).

(132) Leegaard TM, Daugant D, Frøholm L, et al. Do HIV-seropositive patients become colonized with drug-resistant microorganisms? European Journal of Clinical Microbiology and Infectious Diseases 2002;21(12):856-63.

(133) Olofsson SK. Relation between drug exposure and selection of antibiotic resistant bacteria. Uppsala Universitet; 2006.

(134) Sahm DF, Thornsberry C, Mayfield DC, Jones ME, Karlowsky JA. Mulitdrug-resistant urinary tract isolates of Escherichia coli: Prevalence and patient demographics in the United States in 2000. Antimicrobial Agents and Chemotherapy 2001;45(5):1402-6.

(135) Sotto A, deBoever CM, Fabro-Peray P, Gouby A, Sirot D, Jourdan J. Risk factors for antibiotic-resistant Escherichia coli isolated from hospitalized patients with urinary tract infections: A prospective study. Journal of Clinical Microbiology 2001;39(2):438-44.

(136) Howard AJ, Magee J, Fitzgerald K, Dunstan F. Factors associated with antibiotic resistance in coliform organisms from community urinary tract infections in Wales. J Antimicrob Chemother 2001;47(3):305-13.

(137) Lepelletier D, Caroff N, Reynaud A, Richet H. Escherichia coli: Epidemiology and analysis of risk factors for infections caused by resistant strains. Clinical Infectious Diseases 1999;29(3):548-52.

84

References

(138) Ministry of Health and Long-Term Care. Antibiotic resistance: Antibiotic review and Ontario drug benefit formulary listing changes. 2001.

(139) Bii CC, Taguchi H, Ouko TT, Muita LW, Wamae N, Kamiya S. Detection of virulence-related genes by multiplex PCR in multidrug-resistant diarrheogenic Escherichia coli isolates from Kenya and Japan. Epidemiology and Infection 2005;133:627-33.

(140) Bonfiglio G, Simporé J, Pignatelli S, Musumeci S, Solinas ML. Epidemiology of bacterial resistance in gastro-intestinal pathogens in a tropical area. International Journal of Antimicrobial Agents 2002;20:387-9.

(141) Estrada-Garcia T, Cerna JF, Paheco-Gil L, et al. Drug-resistant diarrheogenic Escherichia coli, Mexico. Emerging Infectious Diseases 2005;11(8):1306-8.

(142) Livermore DM, Stephens P, Weinberg J, et al. Regional variation in ampicillin and trimethoprim resistance in Escherichia coli in England from 1990 to 1997, in relation to antibacterial prescribing. J Antimicrob Chemother 2000;46:411-22.

(143) Steinke DT, Seaton RA, Phillips G, MacDonald TM, Davey PG. Factors associated with trimethoprim-resistant bacteria isolated from urine samples. J Antimicrob Chemother 1999;43:841-3.

(144) Jonkers D, Swennen J, London N, Driessen C, Stobberingh E. Influence of cefazolin prophylaxis and hospitalization on the prevalence of antibiotic-resistant bacteria in the faecal flora. J Antimicrob Chemother 2002;49:567-71.

(145) Perea S, Hidalgo M, Arcediano A, et al. Incidence and clinical impact of fluroquinolone-resistant Escherichia coli in the faecal flora of cancer patients treated with high dose chemotherapy and ciprofloxacin prophylaxis. J Antimicrob Chemother 1999;44:117-20.

(146) Bruinsma N, Filius P, vandenBogaard A, et al. Hospitalization, a risk factor for antibiotic-resistant Escherichia coli in the community? J Antimicrob Chemother 2003;51:1029-32.

(147) Tullus K, Berglund B, Burman LG. Emergence of cross-resistance to ß-lactam antibiotics in fecal Escherichia coli and Klebsiella strains from neonates treated with ampicillin or cefuroxime. Antimicrobial Agents and Chemotherapy 1990;34(2):361-2.

(148) Carratala J, Fernandez-Sevilla A, Tubau F, Dominguez MA, Gudiol F. Emergence of fluoroquinlone-resistant Escherichia coli in fecal flora of cancer patients receiving norfloxacin prophylaxis. Antimicrobial Agents and Chemotherapy 1996;40(2):503-5.

(149) Maslow JN, Lee B, Lautenbach E. Fluoroquinolone-resistant Escherichia coli carriage in long-term care facility. Emerging Infectious Diseases 2005;11(6):889-94.

(150) Datta N, Faiers MC, Reeves DS, Brumfitt W, Ørskov I. R factors in Escherichia coli in faeces after oral chemotherapy in general practice. Lancet 1971;297(7694):312-5.

(151) London N, Nijsten R, Mertens P, vandenBogaard A, Stobberingh E. Effect of antibiotic therapy on the antibiotic resistance of faecal Escherichia coli in patients attending general practitioners. J Antimicrob Chemother 1994;34:239-46.

(152) Murray BE, Rensimer ER, DuPont HL. Emergence of high-level trimethoprim resistance in fecal Escherichia coli during oral administration of trimethoprim or trimethoprim-sulfamethoxazole. New England Journal of Medicine 1982;306(3):130-5.

(153) Hirsch DC, Burton GC, Blenden DC. Effect of oral tetracycline on the occurrence of tetracycline-resistant strains of Escherichia coli in the intestinal tract of humans. Antimicrobial Agents and Chemotherapy 1973;4(1):69-71.

85

Appendices

(154) Møller JK, Leth-Bak A, Stenderup A, Zachariae H, Afzelius H. Changing patterns of plasmid-mediated drug resistance during tetracycline therapy. Antimicrobial Agents and Chemotherapy 1977;11(3):388-91.

(155) Stürmer T, Erb A, Marre R, Brenner H. Prevalence and determinants of antibiotic resistance in faecal Escherichia coli among unselected patients attending general practitioners in Southwest Germany. Pharmacoepidemiology and Drug Safety 2004;13:303-8.

(156) Raum E, Lietzau S, von Baum H, Marre R, Brenner H. Changes in Escherichia coli resistance patterns during and after antibiotic therapy: a longitudinal study among outpatients in Germany. European Journal of Clinical Microbiology and Infectious Diseases 2008;14(1):41-8.

(157) Reves R, Murray BE, Pickering L, Prado D, Maddock M, Bartlett A. Children with trimethoprim- and ampicillin-resistant fecal Escherichia coli in day care centers. Journal of Infectious Diseases 1987;156(5):758-62.

(158) Fornasini M, Reves RR, Murray BE, Morrow AL, Pickering LK. Trimethoprim-resistant Escherichia coli in households of children attending day care centers. Journal of Infectious Diseases 1992;166:326-30.

(159) Committee on Infectious Diseases. Red book: 2006 report of the committee on infectious diseases. 27th ed. Elk Grove Village, IL: American Academy of Pediatrics, 2006.

(160) Control of communicable diseases manual. 18th ed. Washington, DC: American Public Health Association, 2004.

(161) Rangel JM, Sparling P, Crowe C, Griffin P, Swerdlow DL. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982-2002. Emerging Infectious Diseases 2005;11(4):603-9.

(162) Doyle ME, Archer J, Kaspar CW, Weiss R. Human illness caused by E. coli O157:H7 from food and non-food sources. Madison, WI: Food Research Institute, University of Wisconsin; 2006.

(163) Warshawsky B, Henry B, Gutmanis I, et al. An E. coli O157:H7 outbreak associated with an animal exhibit: The Middlesex-London Health Unit investigation and recommendations. London, Ontario: Middlesex-London Health Unit; 1999.

(164) Eklund M, Nuorti JP, Ruutu P, Siitonen A. Shigatoxigenic Escherichia coli (STEC) infections in Finland during 1998-2002: A population-based surveillance study. Epidemiology and Infection 2005;133:845-52.

(165) Locking ME, O'Brien S, Reilly W, et al. Risk factors for sporadic cases of Escherichia coli O157 infection: The importance of contact with animal excreta. Epidemiology and Infection 2001;127:215-20.

(166) Khachatryan AR, Hancock DD, Besser TE, Call DR. Antimicrobial drug resistance genes do not convey a secondary fitness advantage to calf-adapted Escherichia coli. Applied and Environmental Microbiology 2006;72(1):443-8.

(167) Wang C, Wu H, Lee L, et al. Pasteurization is effective against multidrug-resistant bacteria. American Journal of Infection Control 2006;34:320-2.

(168) Lautenbach E, Harris A, Perencevich E, Nachamkin I, Tolomeo P, Metlay JP. Test characteristics of perirectal and rectal swab compared to stool sample for detection of fluoroquinolone-resistant Escherichia coli in the gastrointestinal tract. Antimicrobial Agents and Chemotherapy 2005;49(2):798-800.

(169) Penders J, Thijs C, Vink C, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006;118(2):511-21.

(170) Duman M, Abacioglu H, Karaman M, Duman N, Özkan H. β-lactam antibiotic resistance in aerobic commensal fecal flora of newborns. Pediatrics International 2005;47:267-73.

86

References

(171) Lietzau S, Raum E, von Baum H, Marre R, Brenner H. Household contacts were key factor for children's colonization with resistant Escherichia coli in community setting. Journal of Clinical Epidemiology 2007;60(11):1149-55.

(172) Oteo J, Campos J, Faquero F, and Spanish members of the European Antimicrobial Resistance Surveillance System. Antibiotic resistance in 1962 invasive isolates of Escherichia coli in 27 Spanish hospitals participating in the European Antimicrobial Resistance Surveillance System (2001). J Antimicrob Chemother 2002;50:945-52.

(173) Qin X, Razia Y, Johnson JR, et al. Ciprofloxacin-resistant Gram-negative bacilli in the fecal microflora of children. Antimicrobial Agents and Chemotherapy 2006;50(10):3325-9.

(174) Pallecchi L, Lucchetti C, Bartoloni A, et al. Population structure and resistance genes in antibiotic resistant bacteria from a remote community with minimal antibiotic exposure. Antimicrobial Agents and Chemotherapy 2007;51(4):1179-84.

(175) Bonten M, Slaughter S, Ambergen AW, et al. The role of "colonization pressure" in the spread of vancomycin-resistant enterococci. Archives of Internal Medicine 1998;158(10):1127-32.

(176) Karki T, Truusalu K, Vainumae I, Mikelsaar M. Antibiotic susceptibility patterns of community- and hospital-acquired Staphylococcus aureus and Escherichia coli in Estonia. Scandanavian Journal of Infectious Diseases 2001;33:333-8.

(177) Oteo J, Lázaro E, deAbajo FJ, Baquero F, Campos J, and Spanish members of the EARSS. Antimicrobial-resistant invasive Escherichia coli, Spain. Emerging Infectious Diseases 2005;11(4):546-53.

(178) Hiruta N, Murase T, Okamura N. An outbreak of diarrhoea due to multiple antimicrobial-resistant Shiga toxin-producing Escherichia coli O26:H11 in a nursery. Epidemiology and Infection 2000;127:221-7.

(179) Hawkey PM. Resistant bacteria in the normal human flora. J Antimicrob Chemother 1986;18(Supp C):133-9.

(180) Lietzau S, Raum E, von Baum H, Marre R, Brenner H. Clustering of antibiotic resistance of E. coli in couples: Suggestion for a major role of conjugal transmission. BMC Infectious Diseases 2006;6:119-23.

(181) Petrocheilou V, Richmond MH, Bennett PM. Spread of a single plasmid clone to an untreated individual from a person receiving prolonged tetracycline therapy. Antimicrobial Agents and Chemotherapy 1977;12(2):219-25.

(182) Millar MR, Walsh TR, Linton CJ, et al. Carriage of antibiotic-resistant bacteria by healthy children. J Antimicrob Chemother 2001;47:605-10.

(183) Huang DB, Jiang Z, Ericsson C, Adachi J, DuPont HL. Emergence of tripmethoprim-resistant Escherichia coli in healthy persons in the absence of prophylactic or therapeutic antibiotics during travel to Guadalajara, Mexico. Scandanavian Journal of Infectious Diseases 2001;33:812-4.

(184) Murray BE, Mathewson J, DuPont H, Ericsson C, Reves R. Emergence of resistant fecal Escherichia coli in travelers not taking prophylactic antimicrobial agents. Antimicrobial Agents and Chemotherapy 1990;34:515-8.

(185) Sannes MR, Belongia EA, Kieke B, et al. Predictors of antimicrobial-resistant Escherichia coli in the feces of vegetarians and newly hospitalized adults in Minnesota and Wisconsin. The Journal of Infectious Diseases 2008;197:430-4.

(186) Coia JE, Sharp C, Campbell D, Curnow J, Ramsay C. Environmental risk factors for sporadic Escherichia coli O157 infection in Scotland: Results of a descriptive epidemiology study. Journal of Infection 1998;36:317-21.

87

Appendices

(187) Crampin M, Willshaw G, Hancock R, et al. Outbreak of Escherichia coli O157 infection associated with a music festival. European Journal of Clinical Microbiology & Infectious Diseases 1999;18:286-8.

(188) Guan TY, Holley RA. Pathogen survival in swine manure environments and transmission of human enteric illness - a review. Journal of Environmental Quality 2003;32:383-92.

(189) Van Duijkeren E, Box.A.T.A., Schellen P, Houwers DJ, Fluit AC. Class 1 integrons in enterobacteriaceae isolated from clinical infections of horses and dogs in the Netherlands. Microbial Drug Resistance 2005;11(4):383-6.

(190) Box ATA, Mevius DJ, Schellen P, Verhoef J, Fluit AC. Integrons in Escherichia coli from food-producing animals in the Netherlands. Microbial Drug Resistance 2005;11(1):53-7.

(191) Singh R, Schroeder CM, Meng J, et al. Identification of antimicrobial resistance and class 1 integrons in Shiga toxin-producing Escherichia coli recovered from humans and food animals. J Antimicrob Chemother 2005;56:216-9.

(192) Willshaw GA, Cheasty T, Smith HR, O'Brien SJ, Adak GK. Verocytotoxin-producing Escherichia coli (VTEC) O157 and other VTEC from human infections in England and Wales: 1995-1998. Journal of Medical Microbiology 2001;50:135-42.

(193) Ahmed AM, Kawamoto H, Inouye K, et al. Genomic analysis of a multidrug-resistant strain of enterohaemorrhagic Escherichia coli O157:H7 causing a family outbreak in Japan. Journal of Medical Microbiology 2005;54:867-72.

(194) Bettelheim KA, Hornitzky M, Djordjevic S, Kuzevski A. Antibiotic resistance among verocytotoxigenic Escherichia coli (VTEC) and non-VTEC isolated from domestic animals and humans. Journal of Medical Microbiology 2003;52:155-62.

(195) Health Canada. Canadian integrated program for antimicrobial resistance surveillance (CIPARS) 2003. Ottawa: Queen's Printer; 2005.

(196) Van Immerseel F, Pasmans F, De Buck J, et al. Cats as a risk for transmission of antimicrobial drug-resistant Salmonella. Emerging Infectious Diseases 2004;10(12):2169-74.

(197) Skurnik D, Ruimy R, Andremont A, et al. Effect of human vicinity on antimicrobial resistance and integrons in animal faecal Escherichia coli. J Antimicrob Chemother 2006.

(198) Sáenz Y, Zarazag M, Briñas L, Lantero M, Ruiz-Larrea F, Torres C. Antibiotic resistance in Escherichia coli isolates obtained from animals, foods and humans in Spain. International Journal of Antimicrobial Agents 2001;18:353-8.

(199) de Graef EM, Decostere A, Devriese LA, Haesebrouck F. Antibiotic resistance among fecal indicator bacteria from healthy individually owned and kennel dogs. Microbial Drug Resistance 2004;10(1):65-9.

(200) Nijsten R, London N, van den Bogaard AE, Stobberingh E. Resistance in faecal Escherichia coli isolated from pig farmers and abattoir workers. Epidemiology and Infection 1994;113:45-52.

(201) Aubry-Damon H, Grenet K, Sall-Ndiaye P, et al. Antimicrobial resistance in commensal flora of pig farmers. Emerging Infectious Diseases 2004;10(5):873-9.

(202) Levy SB. Emergence of antibiotic-resistant bacteria in the intestinal flora of farm inhabitants. The Journal of Infectious Diseases 1978;137(5):688-90.

(203) Siegel D, Huber WG, Drysdale S. Human therapeutic and agricultural uses of antibacterial drugs and resistance of the enteric flora of humans. Antimicrobial Agents and Chemotherapy 1975;8(5):538-43.

88

References

(204) Scott HM, Campbell L, Harvey R, et al. Patterns of antimicrobial resistance among commensal Escherichia coli isolated from integrated multi-site housing and worker cohorts of humans and swine. Foodborne Pathogens and Disease 2005;2(1):24-37.

(205) Nijsten R, London N, vandenBogaard A, Stobberingh E. Antibiotic resistance among Escherichia coli isolated from faecal samples of pig farmers and pigs. J Antimicrob Chemother 1996;37:1131-40.

(206) Nijsten R, London N, van den Bogaard AE, Stobberingh E. In-vitro transfer of antibiotic resistance between faecal Escherichia coli strains isolated from pig farmers and pigs. J Antimicrob Chemother 1996;37:1141-54.

(207) Price LB, Graham JP, Lackey LG, Roess A, Vailes R, Silbergeld EK. Elevated risk of carrying gentamicin-resistant Escherichia coli among U.S. poultry workers. Environmental Health Perspectives 2007;115(12):1738-42.

(208) Al-Ghamdi MS, El-Morsy F, Al-Mustafa ZH, Al-Ramadhan M, Hanif M. Antibiotic resistance of Escherichia coli isolated from poultry workers, patients and chicken in the eastern province of Saudi Arabia. Tropical Medicine and International Health 1999;4(4):278-83.

(209) vandenBogaard AE, London N, Driessen C, Stobberingh E. Antibiotic resistance of faecal Escherichia coli in poultry, poultry farmers and poultry slaughterers. J Antimicrob Chemother 2001;47:763-71.

(210) Bruinsma N, Filius P, deSmet P, et al. Antibiotic usage and resistance in different regions of a Dutch community. Microbial Drug Resistance 2002;8(3):209-14.

(211) Johnson JR, Kuskowski M, Smith KE, O'Bryan T, Tatini S. Antimicrobial-resistant and extraintestinal pathogenic Escherichia coli in retail foods. Journal of Infectious Diseases 2005;191:1040-9.

(212) Reinthaler FF, Posch J, Feierl G, et al. Antibiotic resistance of E. coli in sewage and sludge. Water Research 2003;37:1685-90.

(213) Corpet DE. Antibiotic residues and drug resistance in human intestinal flora. Antimicrobial Agents and Chemotherapy 1987;31(4):587-93.

(214) Österblad M, Pensala O, Peterzéns M, Heleniusc H, Huovinen P. Antimicrobial susceptibility of enterobacteriaceae isolated from vegetables. J Antimicrob Chemother 1999;43:503-9.

(215) Schroeder CM, Zhao C, DebRoy C, et al. Antimicrobial resistance of Escherichia coli O157 isolated from humans, cattle, swine, and food. Applied and Environmental Microbiology 2002;68(2):576-81.

(216) Agricultural Research Service. Bacterial epidemiology and antimicrobial resistance E. coli - 2002 Report (NARMS). US Department of Agriculture; 2005.

(217) Johnson JR, Delavari P, O'Bryan T, Smith KE, Tatini S. Contamination of retail foods, particularly turkey, from community markets (Minnesota, 1999-2000) with antimicrobial-resistant and extraintestinal pathogenic Escherichia coli. Foodborne Pathogens and Disease 2005;21(1):38-49.

(218) Durán GM, Marshall DL. Ready-to-eat shrimp as an international vehicle of antibiotic-resistant bacteria. Journal of Food Protection 2005;68(11):2395-401.

(219) Hariharan H, Barnum DA, Mitchell WR. Drug resistance among pathogenic bacteria from animals in Ontario. Canadian Journal of Comparative Medicine and Veterinary Science 1974;38(3):213-21.

(220) Linton AH, Howe K, Bennett PM, Richmond MH. The colonization of the human gut by antibiotic resistant Escherichia coli from chickens. Journal of Applied Microbiology 1977;43:465-9.

(221) McEwen SA, Fedorka-Cray P. Antimicrobial use and resistance in animals. Clinical Infectious Diseases 2002;S3:S93-S106.

89

Appendices

(222) Sayah RS, Kaneene J, Johnson Y, Miller R. 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 2005;71(3):1394-404.

(223) Anderson ME, Sobsey MD. Detection and occurrence of antimicrobially resistant E. coli in groundwater on or near swine farms in eastern North Carolina. Water Science and Technology 2006;54(3):211-8.

(224) Kaspar CW, Burgess J. Antibiotic resistance indexing of Escherichia coli to identify sources of sources of fecal contamination in water. Canadian Journal of Microbiology 1990;36:891-4.

(225) Parveen S, Murphree R, Edmiston L, Kaspar CW, Portier K, Tamplin M. Association of multiple antibiotic-resistance profiles with point and nonpoint sources of Escherichia coli in Apalachicola Bay. Applied and Environmental Microbiology 1997;63(7):2607-12.

(226) Niemi M, Sibakov M, Neimela S. Antibiotic resistance among different species of fecal coliforms isolated from water samples. Applied and Environmental Microbiology 1983;45(1):79-83.

(227) Selvaratnam S, Kunberger JD. Increased frequency of drug-resistant bacteria and fecal coliforms in an Indiana creek adjacent to farmland amended with treated sludge. Canadian Journal of Microbiology 2004;50:653-6.

(228) Roe MT, Vega E, Pillai SD. Antimicrobial resistance markers of class 1 and class 2 integron-bearing Escherichia coli from irrigation water and sediments. Emerging Infectious Diseases 2003;9(7):822-6.

(229) Schwartz T, Kohnen W, Jansen B, Obst U. Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. FEMS Microbiology Ecology 2003;43:325-35.

(230) Iwane T, Urase T, Yamamoto K. Possible impact of treated wastewater discharge on incidence of antibiotic resistant bacteria in river water. Water Science and Technology 2001;43(2):91-9.

(231) Boon PI, Cattanach M. Antibiotic resistance of native and faecal bacteria isolated from rivers, reservoirs and sewage treatment facilities in Victoria, south-eastern Australia. Letters in Applied Microbiology 1999;28:164-8.

(232) Biyela PT, Bezuidenhout CC. The role of acquatic ecosystems as reservoirs of antibiotic resistant bacteria and antibiotic resistance genes. Water Science and Technology 2004;50(1):45-50.

(233) Ogan MT, Nwilka DE. Studies on the ecology of aquatic bacteria of the lower Nile delta: Multiple antibiotic resistance among the standard plate count organisms. Journal of Applied Bacteriology 1993;74:595-602.

(234) Antai SP. Incidence of Staphylococcus aureus, coliforms and antibiotic-resistant strains of Escherichia coli in rural water supplies in Port Harcourt. Journal of Applied Bacteriology 1987;62(371):375.

(235) Parveen S, Lukasik J, Scott TM, et al. Geographic variation in antibiotic resistance profiles of Escherichia coli isolated from swine, poultry, beef and dairy cattle farm water retention ponds in Florida. Journal of Applied Microbiology 2006;100(1):50-7.

(236) Edge TA, Hill S. Occurrence of antibiotic resistance in Escherichia coli from surface waters and fecal pollution sources near Hamilton, Ontario. Canadian Journal of Microbiology 2005;51:501-5.

(237) Bell JB, Macrae W, Elliot GE. R factors in coliform-fecal coliform sewage flora of the prairies and Northwest Territories of Canada. Applied and Environmental Microbiology 1981;42(2):204-10.

(238) Bell JB, Macrae WR, Elliot GE. Incidence of R factors in coliform, fecal coliform, and Salmonella populations of the Red River in Canada. Applied and Environmental Microbiology 1980;40(3):486-91.

90

References

(239) Hamelin K, Bruant G, El-Shaarawi A, et al. Occurrence of virulence and antimicrobial resistance genes in Escherichia coli isolates from different aquatic ecosystems within the St. Clair river and Detroit river areas. Applied and Environmental Microbiology 2007;73(2):477-84.

(240) Murray GE, Tobin RS, Junkins B, Kushner DJ. Effect of chlorination on antibiotic resistance profiles of sewage-related bacteria. Applied and Environmental Microbiology 1984;48(1):73-7.

(241) McKeon DM, Calabrese J, Bissonnette G. Antibiotic resistant gram-negative bacteria in rural groundwater supplies. Water Research 1995;29(8):1902-8.

(242) Armstrong JL, Shigeno DS, Calomiris JJ, Seidler RJ. Antibiotic-resistant bacteria in drinking water. Applied and Environmental Microbiology 1981;42(2):277-83.

(243) Edberg SC, Piscitelli V, Cartter M. Phenotypic characteristics of coliform and noncoliform bacteria from a public water supply compared with regional and national clinical species. Applied and Environmental Microbiology 1986;52(3):474-8.

(244) Walia SK, Kaiser A, Parkash M, Chaudhry G. Self-transmissible antibiotic resistance to ampicillin, streptomycin, and tetracycline found in Escherichia coli isolates from contaminated drinking water. Journal of Environmental Science and Health, Part A - Toxic/Hazardous Substances & Environmental Engineering 2004;A39(3):651-62.

(245) Louie M. Antimicrobial resistance in private water sources: Preliminary data. 2006 Mar 3.

(246) Krumperman PH. Multiple antibiotic resistance indexing of Escherichia coli to identify high-risk source of fecal contamination of foods. Applied and Environmental Microbiology 1983;46(1):165-70.

(247) Daniels NA, Neimann J, Karpati A, et al. Traveler's diarrhea at sea: three outbreaks of waterborne enterotoxigenic Escherichia coli on cruise ships. The Journal of Infectious Diseases 2000;181:1491-5.

(248) Shanahan PM, Wylie BA, Adrian PV, Koornhof HJ, Thomson CJ, Amyes SG. The prevalence of antimicrobial resistance in human faecal flora in South Africa. Epidemiology and Infection 1993;111:221-8.

(249) Amyes SG, Tait S, Thomson CJ, et al. The incidence of antibiotic resistance in aerobic faecal flora in South India. J Antimicrob Chemother 1992;29:415-25.

(250) Muller M, McGeer A, Willey BM, et al. Outbreaks of multi-drug resistant Escherichia coli in long-term care facilities in the Durham, York and Toronto regions of Ontario, 2000-2002. Canada Communicable Disease Report 2002;28-14.

(251) Naseer U, Natås OB, Haldorsen BC, et al. Nosocomial outbreak of CTX-M-15-producing E. coli in Norway. APMIS 2006;114:120-6.

(252) Craun GF, Nwachuku N, Calderon R, Craun M. Outbreaks in drinking-water systems, 1991-1998. Journal of Environmental Health 2002;65(1):16-23.

(253) Hunter PR. Drinking water and diarrhoeal disease due to Escherichia coli. Journal of Water and Health 2003;1(2):65-72.

(254) Said B, Wright F, Nichols GL, Reacher MH, Rutter M. Outbreaks of infectious disease associated with private drinking water supplies in England and Wales 1970-2000. Epidemiology and Infection 2003;130:469-79.

(255) Schuster CJ, Ellis AG, Robertson WJ, et al. Infectious disease outbreaks related to drinking water in Canada, 1974-2001. Canadian Journal of Public Health 2005;96(4):254-8.

91

Appendices

(256) Health Canada. Canadian integrated program for antimicrobial resistance surveillance (CIPARS) 2005. Public Health Agency of Canada; 2007.

(257) White DG. The National Antimicrobial Resistance Monitoring System (NARMS). Center for Veterinary Medicine 2007 April 10 [cited 2007 Nov 17];Available from: URL: http://www.fda.gov/cvm/NARMSSSB07White.htm

(258) Canadian Pharmacists Association. Compendium of pharmaceuticals and specialties. 9 ed. Ottawa, ON: Canadian Pharmacists Association, 2008.

(259) CDC. National antimicrobial resistance monitoring system for enteric bacteria (NARMS). 2003 human isolates final report. Atlanta, GA: US Department of Health and Human Services; 2006.

(260) CDC. National antimicrobial resistance monitoring system for enteric bacteria (NARMS). 2004 human isolates final report. Atlanta, GA: US Department of Health and Human Services; 2007.

(261) Domeika M, Oscarsson L, Hallén A, Hjelm E, Sylvan S. Mailed urine samples are not an effective screening approach for Chlamydia trachomatis case finding among young men. JEADV 2007;21:789-94.

(262) Klavs I, Rodrigues LC, Wellings K, Keše D, Švab I. Feasibility of testing for Chlamydia trachomatis in a general population sexual behaviour survey in Slovenia. International Journal of STD & AIDS 2002;13(S2):5-8.

(263) Le Marchand L, Lum-Jones A, Saltzman B, Visaya V, Nomura AMY, Kolonel LN. Feasibility of collecting buccal cell DNA by mail in a cohort study. Cancer Epidemiology, Biomarkers & Prevention 2001;10:701-3.

(264) Weeks MF, Kulka RA, Lessler JT, Whitmore RW. Personal versus telephone surveys for collecting household health data at the local level. American Journal of Public Health 1983;73:1389-94.

(265) Statistics Canada. Canadian Community Health Survey. Statistics Canada 2008 [cited 20040 Nov 11];(2005-06)Available from: URL: www.hc-sc.gc.ca

(266) RRFSS Working Group. Rapid Risk Factor Surveillance System (RRFSS). 1999.

(267) Perth District Health Unit. RDIS case management: enteric illness questionnaire - campylobacter study. 2003.

(268) Ministry of the Environment Laboratory Services Branch. Protocol of accepted drinking-water testing methods. Queen's Printer for Ontario; 2003.

(269) National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. Thirteenth informational supplement, M100-S13. Wayne, PA; 2003.

(270) Ontario Ministry of Agriculture FaRA. Ontario Profile 2007. 2007 Jun 13.

(271) Barros AJD, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: An empirical comparison of models that directly estimate the prevalence ratio. BMC Medical Research Methodology 2003;3(21).

(272) Lumley T, Kronmal R, Ma S. Relative risk regression in medical research: Models, contrasts, estimators, and algorithms. University of Washington; 2006. Report No.: 293.

(273) Zou G. A modified Poisson regression approach to prospective studies with binary data. American Journal of Epidemiology 2004;159(7):702-6.

(274) Stata statistical software [computer program]. Version 9.2. College Station, Texas: StataCorp; 2007.

92

http://www.fda.gov/cvm/NARMSSSB07White.htm


References

(275) Stata User's Guide. Release 9 ed. College Station, TX: StataCorp LP, 2005.

(276) Hosmer DW, Lemeshow S. Applied logistic regression. 2 ed. New York: John Wiley & Sons, Inc., 2000.

(277) Aneshensel CS. Theory-based data analysis for the social sciences. Sage Publications, Inc., 2002.

(278) Austin PC, Tu JV. Automated variable selection methods for logistic regression produced unstable models for predicting acute myocardial infarction mortality. Journal of Clinical Epidemiology 2004;57:1138-46.

(279) Stahel WA. On strategies in regression analysis. Zurich, Switzerland: Swiss Federal Institute of Technology; 2001. Report No.: 100.

(280) Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression methods in biostatistics: Linear, logistic, survival, and repeated measures models. New York, NY: Springer Science & Business Media Inc., 2005.

(281) Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Statistics in Medicine 2007;26:5512-28.

(282) Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. 2 ed. Belmont: Duxbury, 1988.

(283) Kleinbaum DG, Klein M. Modeling strategy guidelines. Logistic regression: a self-learning text. Statistics for biology and health. 2002. p. 162-81.

(284) Rothman KJ, Greenland S. Modern epidemiology. 2 ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1998.

(285) Sun GW, Shook TL, Kay GL. Inappropriate use of bivariable analysis to screen risk factors for use in multivariable analysis. Journal of Clinical Epidemiology 1996;49(8):907-16.

(286) Agresti A. An introduction to categorical data analysis. New York: John Wiley & Sons, 1996.

(287) Harrell FE, Lee K, Mark D. Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine 1996;15:361-87.

(288) Greenland S. Modeling and variable selection in epidemiologic analysis. American Journal of Public Health 1989;79(3):340-9.

(289) Hilbe JM. Negative binomial regression. Cambridge University Press, 2007.

(290) Nursing90 Drug Handbook. Springhouse, PA: Springhouse Corporation, 1990.

(291) Zhanel GG, Hisanaga TL, Laing NM, et al. Antibiotic resistance in Escherichia coli outpatient urinary isolates: Final results from the North American urinary tract infection collaborative alliance (NAUTICA). International Journal of Antimicrobial Agents 2006;27:468-75.

(292) Richards DA, Toop LJ, Chambers ST, et al. Antibiotic resistance in uncomplicated urinary tract infection: Problems with interpreting cumulative resistance rates from local community laboratories. New Zealand Medical Journal 2002;115:12-4.

(293) Laupland KB, Ross T, Pitout JDD, Church DL, Gregson DB. Investigation of sources of potential bias in laboratory surveillance for anti-microbial resistance. Clinical Investigative Medicine 2007;30(4):E159-E166.

93

Appendices

(294) Colgan R, Johnson JR, Kuskowski M, Gupta K. Risk factors for trimethoprim-sulfamethoxazole resistance in patients with acute uncomplicated cystitis. Antimicrobial Agents and Chemotherapy 2008;52(3):846-51.

(295) Filius PM, Gyssens IC, Kershof IM, et al. Colonization and resistance dynamics of Gram-negative bacteria in patients during and after hospitalization. Antimicrobial Agents and Chemotherapy 2005;49(7):2879-86.

(296) Kaye KS, Harris AD, Gold H, Carmeli Y. Risk factors for recovery of ampicillin-sulbactam-resistant Escherichia coli in hospitalized patients. Antimicrobial Agents and Chemotherapy 2000;44(4):1004-9.

(297) Barbosa TM, Levy SB. The impact of antibiotic use on resistance development and persistence. Drug Resistance Updates 2000;(3):303-11.

(298) Hillier S, Roberts Z, Dunstan F, Butler C, Howard A, Palmer S. Prior antibiotics and risk of antibiotic-resistant community-acquired urinary tract infection: A case-control study. J Antimicrob Chemother 2007;60:92-9.

(299) Brown PD, Freeman A, Foxman B. Prevalence and predictors of trimethoprim-sulfamethoxazole resistance among uropathogenic Escherichia coli isolates in Michigan. Clinical Infectious Diseases 2002;34:1061-6.

(300) Jones RN, Kugler KC, Pfaller MA, Winokur PL, the SENTRY surveillance group NA. Characteristics of pathogens causing urinary tract infections in hospitals in North America: Results from the SENTRY antimicrobial surveillance program, 1997. Diagnostic Microbiology and Infectious Diseases 1999;35:55-63.

(301) Infante B, Grape M, Larsson M, et al. Acquired sulphonamide resistance genes in faecal Escherichia coli from healthy children in Bolivia and Peru. International Journal of Antimicrobial Agents 2005;25:308-12.

(302) Coleman BL. Testing of private well water. St. Thomas, Ontario: Elgin-St. Thomas Health Unit; 2005.

(303) Coleman BL. Bacteriological water testing in Southern Ontario. 2008.

(304) Hartley CL, Richmond MH. Antibiotic resistance and survival of E. coli in the alimentary tract. British Medical Journal 1975;4:71-4.

(305) Anderson MA, Whitlock JE, Harwood VJ. Diversity and distribution of Escherichia coli genotypes and antibiotic resistance phenotypes in feces of humans, cattle, and horses. Applied and Environmental Microbiology 2006;72(11):6914-22.

(306) Federal-Provincial-Territorial Committee on Drinking Water. Guidelines for Canadian drinking water quality. Health Canada 2008Available from: URL: http://www.hc-sc.gc.ca

(307) Federal-Provincial-Territorial Committee on Drinking Water. Summary of guidelines for Canadian drinking water quality. Health Canada 2003 [cited 2008 Apr 24];Available from: URL: http://www.hc-gc.

(308) Leclerc H, Schwartzbrod L, Dei-Cas E. Microbial agents associated with waterborne diseases. Critical Reviews in Microbiology 2002;28(4):371-409.

(309) Krewski D, Balbus J, Butler-Jones D, et al. Managing the microbiological risks of drinking water. Journal of Toxicology and Environmental Health, Part A 2004;67:1591-617.

(310) Pathak SP, Gopal K. Prevalence of bacterial contamination with antibiotic resistant and enterotoxigenic fecal coliforms in treated drinking water. Journal of Toxicology and Environmental Health, Part A 2008;71:427-33.

94


http://www.hc-gc/

http://www.hc-gc/

References

(311) FAAIR Scientific Advisory Panel. Policy recommendations. Clinical Infectious Diseases 2002;34(S3):S76-S77.

(312) Larson E. Community factors in the development of antibiotic resistance. Annual Review of Public Health 2007;28:435-47.

(313) McDonald LC. Trends in antimicrobial resistance in health care-associated pathogens and effect on treatment. Clinical Infectious Diseases 2006;42:S62-S71.

(314) Interagency Task Force on Antimicrobial Resistance. Second Annual Progress Report: Implementation of a Public Health Action Plan to Combat Antimicrobial Resistance. Part 1: Domestic Issues. 2004.

(315) Zhanel GG, deCorby M, Laing N, et al. Antimicrobial-resistant pathogens in intensive care units in Canada: Results of the Canadian national intensive care unit (CAN-ICU) study, 2005-06. Antimicrobial Agents and Chemotherapy 2008;52(4):1430-7.

(316) Lin L. Correlating trends in antibiotic consumption and antimicrobial resistance in British Columbia. British Columbia; 2004.

(317) Jones ME, Karlowsky JA, Draghi D, Thornsberry C, Sahm DF, Bradley JS. Rates of antimicrobial resistance among common bacterial pathogens causing respiratory, blood, urine, and skin and soft tissue infections in pediatric patients. European Journal of Clinical Microbiology & Infectious Diseases 2004;23(6):445-55.

(318) Wenzel R, Sahm DF, Thornsberry C, Draghi D, Jones ME, Karlowsky JA. In vitro susceptibilities of gram-negative bacteria isolated from hospitalized patients in four European countries, Canada, and the United States in 2000-2001 to expanded-spectrum cephalosporins and comparator antimicrobials: Implications for therapy. Antimicrobial Agents and Chemotherapy 2003;47(10):3089-98.

(319) Preston CAK, Bruce AW, Reid G. Antibiotic resistance of urinary pathogens isolated from patients attending the Toronto Hospital between 1986 and 1990. Journal of Hospital Infection 1992;22:129-35.

(320) Cooke EM, Shooter R, Breaden A, Farrell S. Antibiotic sensitivity of Escherichia coli isolated from animals, food, hospital patients, and normal people. Lancet 1971;8-10.

(321) Jenkins C, Smith HR, Lawson A, et al. Serotypes, intimin subtypes, and antimicrobial resistance patterns of atypical enteropathogenic Escherichia coli isolated in England from 1993 to 1996. European Journal of Clinical Microbiology & Infectious Diseases 2006;25(1):19-24.

(322) Kronvall G, Larsson M, Borén C, et al. Extended antimicrobial resistance screening of the dominant faecal Escherichia coli and of rare resistant clones. International Journal of Antimicrobial Agents 2005;26:473-8.

(323) Leistevuo T, Leistevuo J, Österblad M, et al. Antimicrobial resistance of fecal aerobic gram-negative bacilli in different age groups in a community. Antimicrobial Agents and Chemotherapy 1996;40(8):1931-4.

(324) Eylan E, Cohen S. Resistance transfer factor in E. coli strains isolated from human and water sources. Contributions to Microbiology and Immunology 1979;6:210-21.

(325) Mora A, Blanco J, Blanco M, et al. Antimicrobial resistance of Shiga toxin (verotoxin)-producing Escherichia coli O157:H7 and non-O157 strains isolated from humans, cattle, sheep and food in Spain. Research in Microbiology 2005;156(7):793-806.

(326) Bartoloni A, Pallecchi L, Fiorelli C, et al. Increasing resistance in commensal Escherichia coli, Bolivia and Peru. Emerging Infectious Diseases 2008;14(2):338-40.

95

Appendices

(327) Okeke IN, Fayinka ST, Lamikanra A. Antibiotic resistance in Escherichia coli from Nigerian students, 1986-1998. Emerging Infectious Diseases 2000;6(4):393-6.

(328) Ahmed AA, Osman H, Mansour A, et al. Antimicrobial agent resistance in bacterial isolates from patients with diarrhea and urinary tract infection in the Sudan. American Journal of Tropical Medicine and Hygeine 2000;63(5,6):259-63.

(329) Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. 3 ed. Hoboken: John Wiley & Sons, Inc., 2003.

96

Appendices

97

Appendices

Appendix A a) Studies Comparing E. coli Antimicrobial Resistance Rates in Canadian Subjects: Clinical Isolates

Study Zhanel (315)

Lin (316) Jones (317)

Jones (52) Wenzel (318)

Anon. (94)

Pfaller (95)

Year 2008 2004 2004 2004 2003 2003 1998 Age of subjects All Children All All All All Subjects Hospital

ICU Inpt & outpt

Hospital B.S.I.

Hospital ICU

Hospital Hospital Hospital B.S.I.

Total isolates 536 6,860 87 2,013 939 Amikacin 0 0.4% 1% Amoxicillin-CA 7% 20% 14% 23%

Ampicillin 37% 35% 34% 46% Cefoxitin 7%

Ceftiofur Ceftriaxone 4% 6% 2% 1% 1%

Cephalothin Chloramphenicol Ciprofloxacin 21% 10% 4% 10% 7% 4% 3%

Gentamicin 3% 5% 7% 5% 5% 9% 6%

Kanamycin Nalidixic acid Nitrofurantoin Piperacillin 30% 40% Sulphamethoxazole Tetracycline 27%

Trimethoprim TMP/SMX 25% 21% 15% 17% 18% 19%

b) Studies Comparing E. coli Antimicrobial Resistance Rates in Canadian Subjects: Urinary Tract Infection Isolates Study Zhanel

(291) McIsaac

(97) Zhanel (102)

Kahlmeter (99)

Zhanel (98)

Jones (300)

Preston (319)

Year 2006 2006 2005 2003 2000 1997 1992 Age of subjects All All All 18-65 yrs All All All Subjects Outpt Comm Outpt Comm Outpt Hosp Total isolates 280 2,199 496 166 1,681 182 716 Amikacin 1% 1% Amoxicillin-CA 4% 21% Ampicillin 33% 32% 42% 30% 41% 33% 31% Cefoxitin 4% Ceftiofur Ceftriaxone Cephalothin 13% Chloramphenicol Ciprofloxacin 1% 7% 7% 0 1% 1% 0 Gentamicin 1% 3% Kanamycin Nalidixic acid 1% 1% Nitrofurantoin 0 1% 12% 1% 0.1% 4% 0 Sulphamethoxazole 25% 30% Tetracycline 19% Trimethoprim 11% 13% TMP/SMX 17% 15% 15% 12% 19% 19% 11% CA: clavulanic acid TMP/SMX: trimethoprim-sulfamethoxazole Comm: community Hosp: hospitalized subjects Outpt: outpatient NOTE: empty cells reflect the fact that the antimicrobial agent was not used in susceptibility testing.

Appendices

98

Appendix B a) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in the United Kingdom and the Netherlands (Holland) Study Cooke

(320) Jenkins (321)

Filius (295)

Bruinsma (146)

Bruinsma (104)

Jonkers (144)

Bruinsma (210)

Year 1971 2006 2005 2003 2003 2002 2002 Country UK UK Holland Holland Holland Holland Holland Subject age All ages Adult Adult Adult Adult Subject Comm Outpt Hosp Hosp Comm Hosp Comm Isolates/subject 5 1 Total subjects 41 241 411 183 129 180 438 Amikacin Amoxicillin 11% 28% 30% 12% 12% Amoxicillin-CA 1% Ampicillin 15% 13% Cefoxitin Ceftiofur Ceftriaxone Cephalothin 4% Chloramphenicol 7% 6% 3% Ciprofloxacin 1% <1% 3% 0 0 <1% Gentamicin <1% 2% 1% 0 1% Kanamycin Naladixic acid 5% 2% 1% 1% Nitrofurantoin 0 0 0 Streptomycin 37% Sulphamethoxazole Tetracycline 39% Oxytetracycline 26% 33% 12% 11% Trimethoprim 17% 23% 9% 8% TMP/SMX b) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in the Netherlands (Holland), continued Study Nijsten

(200) London (151)

London (121)

Bonten (116)

Bonten (43)

Degener (130)

Degner (129)

Year 1994 1994 1994 1992 1990 1983 1983 Country Holland Holland Holland Holland Holland Holland Holland Subject age Adult 1-82 yr Adult Adult 17-32 yr All All Subject Comm Outpt Comm Comm Comm Comm Hosp Isolates/subject 1 7.4 1.5 5 Total subjects 150 168 183 310 172 680 286 Amikacin Amoxicillin 13% 13% Amoxicillin-CA 10% Ampicillin 16% 84% 26% 24% Cefoxitin Ceftiofur Ceftriaxone Cephalothin 13% 10% Chloramphenicol 3% 9% 55% 73% Ciprofloxacin 0 Gentamicin 27% 23% Kanamycin 4% 4% 47% 34% Naladixic acid 0 2% 1% 16% 9% Nitrofurantoin 0 0 1% 10% 9% Streptomycin 24% 24% 26% 34% Sulphamethoxazole 10% 24% 32% 44% 82% 46% 27% Tetracycline 20% 29% 42% 18% Oxytetracycline 8% 16% 21% 8% Trimethoprim 3% 8% 9% 29% 10% 3% TMP/SMX

Appendices

99

Appendix B c) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in Scandanavia Study Bagger-

Skjøt (92) Kronvall

(322) Leegaard

(132) Leistevuo

(323) Osterblad

(80) Osterblad

(80) Osterblad

(80) Year 2007 2005 2002 1996 2000 2000 2000 Country Denmark Sweden Norway Finland Finland Finland Finland Subject age Adult Adult All Adult Subjects Comm Comm Outpt Comm Comm Hospital LTC Isolates/subject 3 3 5 5 5 Total subjects 85 50 83 334 125 159 74 Amikacin 0 Amoxicillin 40% Amoxicillin-CA 0 0 0 3% 1% Ampicillin 26% 31% 38% 12% 14% 25% Cefoxitin Ceftiofur Ceftriaxone 0 Cephalothin 1% 5% 8% Chloramphenicol 4% 12% 0 4% 7% 3% Ciprofloxacin 1% 2% 0 0 0 Gentamicin 0 0% 4% 0 3% 0 Kanamycin 6% Naladixic acid 1% 5% 1% 4% 13% Nitrofurantoin 3% Streptomycin 24% 18% 14% Sulphamethoxazole 27% 24% 16% 13% 28% Tetracycline 16% 23% 21% 14% 13% 25% Oxytetracycline Trimethoprim 9% 14% 9% 12% 40% TMP/SMX 45% 8% 8% 15% d) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in Europe and Israel Study Sturmer

(155) Lietzau (180)

Lietzau (171)

Bruinsma (104)

Vatopo-lous (115)

Eylan (324)

Eylan (324)

Year 2004 2006 2007 2003 1998 1979 1979 Country Germany Germany Germany Greece Greece Israel Israel Subject age Adult Adult <5 yr Adult <7 yr Adult Infant Subjects Outpt Comm Outpt Comm Comm

Outpt

Isolates/subject 3 1 6 6 Total subjects 484 412 492 179 181 100 85 Amikacin Amoxicillin 39% Amoxicillin-CA 2% 8% Ampicillin 17% 17% 17% 41% 8% 40% Cefoxitin Ceftiofur Ceftriaxone Cephalothin Chloramphenicol 13% 7% 24% Ciprofloxacin 1% 3% Gentamicin 2% Kanamycin 1% 11% Naladixic acid 2% 12% 5% Nitrofurantoin Streptomycin 23% 40% Sulphamethoxazole Tetracycline Oxytetracycline 50% Trimethoprim 31% TMP/SMX 9% 10% 9%

Appendices

100

Appendix B e) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in Europe Study Dunman

(170) Mora (325) Domínguez

(118) Saenz (198)

Carratala (148)

Year 2005 2005 2002 2001 1996 Country Turkey Spain Spain Spain Spain Subject age Infant < 2 yr Adult Subjects Hosp Clinical Comm Comm Outpt Isolates/subject 4 1 1 5 Total subjects 118 138 41 40 25 Amikacin 0 0 Amoxicillin 58% 60% Amoxicillin-CA 11% 1% 11% 0 Ampicillin 31% 19% 31%

Cefoxitin 1% 0 0 Ceftiofur Ceftriaxone 0 Cephalothin 6 Chloramphenicol 10 12% 6%

Ciprofloxacin 0 5% 0 12%

Gentamicin 1% 2% 0 10% Kanamycin 8% 10% 6% Naladixic acid 4% 22% 6%

Nitrofurantoin Streptomycin 32% 37% Sulphamethoxazole Tetracycline 32% 51% 25%

Oxytetracycline Trimethoprim 8% TMP/SMX 7% 24% 3% 70% f) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in China, the Phillipines, and Japan Study Lester

(127) Nys (106) Nys (106) Bii (139)

Year 1990 2004 2004 2005 Country China Philli-

pines Philli-pines

Japan

Subject age <7 yr 18-50 yr 18-50 yr < 5 yr Subjects Comm Comm Comm Outpt Isolates/subject 9.9 Total subjects 53 105 106 47 Amikacin Amoxicillin Amoxicillin-CA Ampicillin 26% 87% 75% 38% Cefoxtin Ceftiofur Ceftriaxone Cephalothin Chloramphenicol 19% 68% 60% Ciproflox. 63% 35% Gentamicin 19% 46% 20% 0 Kanamycin 11% Naladixic acid Nitrofurantoin Streptomycin 51% Sulphamethoxazole 58% Tetracycline 75% Oxytetracycline 94% 85% Trimethoprim 38% 93% 86% TMP/SMX

Appendices

101

Appendix B g) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in Mexico and Venezuela Study Nys (106) Nys (106) Estrada-

Garcia (141)

Calva (119)

van de Mortel (123)

Lester (127)

Nys (106)

Year 2004 2004 2005 1996 1998 1990 2004 Country Mexico Mexico Mexico Mexico Venez-

uela Venez-

uela Venez-

uela

Subject age 18-50 yr 18-50 yr < 5 yr Infant All <7 yr 18-50 yr Subjects Comm Comm Hosp Comm Comm Comm Comm Isolates/subject 4.7 13 1 9.6 Total subjects 99 99 62 20 161 41 230 Amikacin Amoxicillin 53% Amoxicillin-CA Ampicillin 94% 78% 73% 90% 32% 43% Cefoxtin Ceftiofur Ceftriaxone Cephalothin Chloramphincol 75% 45% 19% 62% 15% 30% Ciprofloxacin 51% 15% 0 1% 1% Gentamicin 14% 10% 5% 0 3% Kanamycin 12% Naladixic acid Nitrofurantoin 2% 0 Streptomycin 56% Sulphamethoxazole 62% 63% Tetracycline 82% 62% 66% Oxytetracycline 97% 86% 64% 52% Trimethoprim 96% 76% 77% 41% 12% 33% TMP/SMX 65% h) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in South America Study Nys (106) Bartoloni

(326) Bartoloni

(84) Nys (106) Bartoloni

(326) Bartoloni

(84) Bartoloni

(117) Year 2004 2008 2006 2004 2008 2006 2004 Country Curacao Peru Peru Peru Bolivia Bolivia Bolivia Subject age 18-50 yr < 6 yr < 6 yr 18-50 yr < 6 yr < 6 yr All Subjects Comm Comm Comm Comm Comm Comm Comm Isolates/subject ~ 6 ~ 6 ~ 6 ~ 6 Total subjects 149 1593 1590 95 1600 1584 108 Amikacin 0 1% 0.1% Amoxicillin Amoxicillin-CA Ampicillin 48% 96% 92% 95% 97% 97% 58% Cefoxtin Ceftiofur Ceftriaxone 2% 0.1% 2% 0.1% Cephalothin Chloramphenicol 8% 72% 71% 52% 67% 70% 41% Ciprofloxacin 1% 39% 21% 36% 26% 16% Gentamicin 4% 24% 20% 33% 30% 23% Kanamycin 24% 22% 34% 34% Naladixic Acid 62% 38% 51% 36% Nitrofurantoin Streptomycin 91% 79% 94% 92% Sulphamethoxazole Tetracycline 94% 76% 92% 86% 64% Oxytetracycline 56% 93% Trimethoprim 32% 93% TMP/SMX 94% 91% 94% 96% 50%

Appendices

102

Appendix B i) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in the United States and Canada Study Sannes

(185) Price (207)

Hannah (46)

Putnam (30)

Scott (204)

Akwar (105)

Bruinsma (104)

Year 2008 2007 2005 2005 2005 2007 2003 Country USA USA USA USA USA Canada Canada Subject age >13 yrs Adult All Adult Adult Adult Subjects Comm Comm Outpt Comm Comm Comm Isolates/subject 5 5 Total subjects 667 49 517 188 308 115 154 Amikacin 0 Amoxicillin 22% Amoxicillin-CA 2% Ampicillin 61% 46% 16% 16% Cefoxitin 3% 2% Ceftiofur 0 Ceftriaxone 2% 0 0 Cephalothin 28% 2% Chloramphenicol 16% 3% 3% 1% Ciprofloxacin 1% 6% 13% <1% 0 1% Gentamicin 18% 1% 1% 1% Kanamycin <1% 4% Naladixic acid 5% 3% 4% <1 1% Nitrofurantoin 2% <1 0 Streptomycin 12% 10% Sulphamethoxazole 12% 17% Tetracycline 41% 47% 19% 23% Oxytetra. 16% Trimethoprim 10% TMP/SMX 8% 11% 36% 8% 5% j) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in the United States Study Murray

(184) Lester (127)

Levy (122) Levy (122) Siegel (203)

Year 1990 1990 1988 1988 1975 Country USA USA USA USA USA Subject age <7 yr Subjects Comm Comm Hosp Comm Isolates/subject 9.3 9.9 1.5 1.5 2.3 Total subjects 13 39 289 189 130 Amikacin Amoxicillin Amoxicillin-CA Ampicillin 38% 13% 17% 27% 22% Cefoxitin Ceftiofur Ceftriaxone Cephalothin 5% Chloramphenicol 8% 3% 2% 2% <1% Ciprofloxacin Gentamicin 0 0 <1% 1% 0 Kanamycin 0 0 4% 8% Naladixic acid 1% 1% 2% Nitrofurantoin Streptomycin 38% 15% 16% 21% 37% Sulphamethoxazole 13% Tetracycline 46% 15% 19% 23% Oxytetracycline 39% Trimethoprim 0 3% TMP/SMX 0

Appendices

Appendix B k) Studies Comparing Escherichia Coli Antimicrobial Resistance Rates: Faecal Carriage in Africa Study Okeke

(327) Ahmed (328)

Bii (139) Nys (106)

Nys (106)

Bonfiglio (140)

Nys (106)

Shana-han

(248) Year 2000 2000 2005 2004 2004 2002 2004 1993 Country Nigeria Sudan Kenya Kenya Ghana Burkina

Faso Zimb-abwe

S. Africa

Subject age Infant < 5 yr 18-50 yr 18-50 yr All 18-50 yr All Subjects Comm Hosp Outpt Comm Comm Hosp &

outpt Comm Comm

Isolates/subject 1 Total subjects 26 82 100 100 1405 207 361 Amikacin Amoxicillin 38%* 95% Amoxicillin-CA 47%

80% 69% Ampicillin 66% 89% 89% 95% 49% 73% Cefoxtin Ceftiofur Ceftriaxone

Cephalothin 60% 4% Chloramphenicol 45% 82% 19%

Ciprofloxacin 0 1% 8% 1% Gentamicin 6% 2% 2% 1% Kanamycin

3% 4% Naladixic acid Nitrofurantoin

55% Streptomycin Sulphamethoxazole

100% Tetracycline Oxytetracycline 73% 92% 90% 59%

48% Trimethoprim 88% 89% 64% 83% TMP/SMX 80%

Amoxicillin-CA: Amoxicillin clavulanic acid Sulphamethox: Sulphamethoxazole TMP/SMX: Trimethoprim-sulfamethoxazole Comm: Community-based Hosp: Hospital LTC: Long-term care facility Outpt: outpatient NOTE: empty cells reflect the fact that the antimicrobial agent was not used in susceptibility testing

103

Appendices

Appendix C

Sample Size Calculations

Sample size for proportions (329) Hypothesis 1: the prevalence of human carriage of ampicillin resistant E. coli will be greater than or equal to 22% The hypothesized proportion is based on 22% carriage in Canadian residents in Bruinsma, et al’s article (104) and assuming a type I error probability of 5%. H0: P < 0.22 H1: P ≥ 0.22 n = (Z/m)2 * pq

p m n α 0.22 0.05 0.035 539 0.22 0.05 0.03 733 0.22 0.05 0.025 1,055 0.22 0.05 0.02 1,648

Where: Z = ([p – p0] – 1/2n) / √pq/n m = confidence width / 2 p = hypothesized proportion q = 1-p α = probability of Type I error Sample size for logistic regression (276) This was used to estimate the sample size required for the case-control study from which the subjects for the prevalence study were drawn.

and pThe estimates for p0 1 were based on the probability of the contaminated water sources being located on farms that housed livestock. H0: β1 = 0 H1: β1 ≠ 0 n = (Z1-α/2√2pq + Z1-β√p0q0 + p1q1)

2 2 / (p1-p0)

Β p Odds n pα 0 1Ratio (per group)

0.05 0.20 .30 .39 1.5 425 0.05 0.20 .30 .46 2.0 141 0.05 0.20 .35 .45 1.5 401 0.05 0.20 .35 .52 2.0 135 0.05 0.20 .40 .50 1.5 388 0.05 0.20 .40 .57 2.0 133

Where p1 = OR x p0 / [(1-p0)+OR*p0] OR = odds ratio (looking for ability to detect difference) α = probability of Type I error β = probability of Type II error p1 = probability of outcome in group 1 p0 = probability of outcome in group 2

104

1 0 1

Appendix D

Ontario Well Water StudyInformation Sheet

Safe drinking water is important for everyone. In Ontario, about 90% of all drinking water samples that test positive for bacteria are from private drinking water sources such as wells. Some of the bacteria in well water, including certain kinds of bacteria called Escherichia coli (E. coli), can be resistant to a number of the antibiotics used to treat infections. There are many unanswered questions about how antibiotic resistance develops; some scientists think that antibiotic resistant E. coli in water may be one source of antibiotic resistant infections in people. From May 2005 to December 2006, a group of doctors and scientists from the University of Western Ontario and the University of Toronto will be studying how drinking or using contaminated well water affects people. This study will determine possible factors (e.g. type of farm, soil, or well) that could contaminate private wells with antibiotic resistant bacteria. It will also determine whether antibiotic resistant bacteria in drinking water might be a source of antibiotic resistant bacteria in humans. Since you have submitted a water sample for testing, you may be telephoned a few weeks after your water sample is tested to ask if you would be willing to take part in the study. The researchers need to talk to some people with water that tested positive for E. coli and some people whose water tested negative. Only a small fraction of people, who will be selected at random, will be telephoned. If you do receive a call, it will be from a member of the Safe Water Unit at the Ministry of Health and Long-Term Care. This person will ask permission to give your phone number to the researchers. You do not have to take part in the study if you do not want to. If you consent to participate in the study, the Ministry of Health and Long-Term Care will disclose your contact information to the researchers. The researchers will then contact you to arrange an interview. If you decide not to participate in the study, the Ministry of Health and Long-Term Care will not disclose your contact information to the researchers. Whether or not you decide to take part in the study will not affect future water testing in any way. If you have any questions about this study, you may contact Brenda Coleman, the study coordinator, at (519) 631-3159 ext. 265 weekdays between 8:30 a.m. and 4:30 p.m. or e-mail her at [email protected]. If you are interested, more information about the study is available at http://microbiology.mtsinai.on.ca/research/wellwaterstudy/ For information about the results of your water test please call your local public health unit. Thank you.

Appendices

105

mailto:[email protected]

http://microbiology.mtsinai.on.ca/research/wellwaterstudy/

Appendix E

Ontario Well Water Study

Telephone Script for Contact from Safe Water Unit Cases and “A” Controls (E. coli positive)

Hello, my name is <FILL: First and last name>. I am calling from the Safe Water Unit at the Ministry of Health and Long-Term Care. May I please speak with <FILL: name of submitter>. Submitter/designate not available: I am calling to talk to you about the water sample that was sent from your household for testing. Is there someone else that I could talk to? [NO] Is there a good time to call back so I can talk with <FILL: Name of submitter>? [YES] Continue Submitter (or designate) speaking: Thank you. I am calling for a group of doctors and scientists at the University of Western Ontario and the University of Toronto who are studying private well water. They want to know why the water from some wells is contaminated with bacteria that are resistant to antibiotics while water in other wells is not. They are also looking at whether using contaminated water affects the health of people who drink it. To help with this study, they need to talk to people who have had their well water tested. Because you have submitted a sample for testing, they would like to talk with you about taking part in their study. I am phoning to ask if you would agree to allow the Safe Water Unit to give your name and phone number to these researchers. They would then contact you to explain the study and ask if you would be willing to participate. Would it be okay if they called you to explain the study? [NO] Thank you very much for your time. Good bye. [YES] I have a few other questions I need to ask to see if you are eligible to take part in the study. Do you use the water that was sent for testing on <FILL: Date Collected>? [NO – real estate, moved, submitted for someone else] The researchers need to talk with someone who can answer questions about the water source and possible sources of contamination, so you are not eligible for the study. Thank you very much for your time. Good bye. [NO – use bottled/filtered/treated water] Is the water from a well or other water source in Southern Ontario? [YES] Is the water from a well or other water source in Southern Ontario? (Note: see attached) [NO] The researchers are only interested in talking with people who are using water from Southern Ontario, so you are not eligible for the study. Thank you very much for your time. Good bye. [YES] Good. The Ontario Well Water Study will be calling you in the next week or two to explain the study and ask if you will agree to participate. They will also be asking you if other members of your household would be willing to take part in the study. Thank you for your time and good bye.

Appendices

106


Telephone Script for Contact from Safe Water Unit “B” Controls (uncontaminated)

Hello, my name is <FILL: First and last name >. I am calling from the Safe Water Unit at the Ministry of Health and Long-Term Care. May I please speak with <FILL: Name of submitter >. Submitter/designate not available: I am calling to talk to you about the water sample that was sent from your household for testing. Is there someone else that I could talk to? [NO] Is there a good time to call back so I can talk with < FILL: Name of submitter >? [YES] Continue… Submitter (or designate) speaking: Thank you. I am calling for a group of doctors and scientists at the University of Western Ontario and the University of Toronto who are studying private well water. They want to know why the water from some wells is contaminated with bacteria that are resistant to antibiotics while water in other wells is not. They are also looking at whether using contaminated water affects the health of people who drink it. To help with this study, they need to talk to people who have had their well water tested. Because you have submitted a sample for testing, they would like to talk with you about taking part in their study. I am phoning to ask if you would agree to allow the Safe Water Unit to give your name and phone number to these researchers. They would then contact you to explain the study and ask if you would be willing to participate. Would it be okay if they called you to explain the study? [NO] Thank you for your time. Good bye. [YES] I have a few other questions I need to ask to see if you are eligible to take part in the study. Do you use the water that was sent for testing on <FILL: Date Collected>? (Note: i.e. not for real estate, conservation area, or municipal building) [NO – real estate, moved] The researchers need to talk with someone who can answer questions about the water source and possible sources of contamination, so you are not eligible for the study. Thank you very much for your time. Good bye. [NO – use UV/other filter/bottled – or YES] Is the water from a well or other water source in Southern Ontario? (Note: i.e. not their cottage in North Bay.) [NO] The researchers are only interested in talking with people who are using water from Southern Ontario, so you are not eligible for the study. Thank you very much for your time. Good bye. [YES] Have you submitted this water for testing before or since the test on <FILL: Date Collected>? [NO/DON’T KNOW] Good (okay). The Ontario Well Water Study will be calling you in the next week or two to explain the study and ask if you will agree to participate. They will also be asking you if other members of your household would be willing to take part in the study. Thank you for your time and good bye. [YES] And, has this water source ever tested positive for E. coli? (Note: some people send in water samples from before and after a treatment system, e.g. UV lights. If this sample is a post-treatment one and the same water source tested positive for E. coli, we can’t accept them as a “B” control. Similarly, water tests may vary from time-to-time, so if this water source tested positive in the past year, we can’t accept them.) [NO/DON’T KNOW] Good (okay). The Ontario Well Water Study will be calling you in the next week or two to explain the study and ask if you will agree to participate. They will also be asking you if other members of your household would be willing to take part in the study. Thank you for your time and good bye.

3

Appendices

107

[YES] Was the positive test within the past year? [NO/DON’T KNOW] Good (okay). The Ontario Well Water Study will be calling you in the next week or two to explain the study and ask if you will agree to participate. They will also be asking you if other members of your household would be willing to take part in the study. Thank you for your time. Good bye. [YES] The researchers are only interested in talking with people who have had uncontaminated water for a year or longer, so you are not eligible for the study. Thank you very much for your time. Good bye. Southern Ontario: Brant county & Brantford (and surrounding areas in county) Bruce & Grey counties & Owen Sound, Walkerton, Durham, Southampton. Elgin county & St. Thomas, Aylmer Haldimand & Norfolk counties & Simcoe, Dunnville, Caledonia Halton region & Oakville Hamilton city & region Huron county & Clinton, Wingham Lambton county & Sarnia, Forest Middlesex county & London, Strathroy Niagara region & St. Catherines, Welland Oxford county & Woodstock Perth county & Stratford, Listowel Waterloo city & region, Cambridge Wellington & Dufferin counties - Guelph, Mount Forest, Orangeville, Shelburne, Palmerston NEW EXPANDED AREA: Ottawa-Carelton (and surrounding areas in county) Eastern Ontario – Cornwall, Rockland, Hawkesbury, Casselman. Renfrew County – Arnprior, Deep River, Barry’s Bay. Kingston, Frontenac, Lennox & Addington – Napanee, Cloyne, Sharbot Lake Hastings and Prince Edward – Belleville, Bancroft, Picton, Trenton Leeds, Grenville & Lanark – Brockville, Gananoque, Kemptville, Almonte, Smiths Falls Haliburton, Kawartha, Pine Ridge – Lindsay, Campbellford, Brighton Durham Region – Whitby, Ajax, Oshawa, Pickering, Uxbridge Toronto – Scarborough, North York, Etobicoke York region – Newmarket, Unionville, Richmond Hill Peterborough county and city – Cochrane, Hearst, Hornepayne, Moosonee, Matheson Muskoka-Parry Sound – Huntsville, Burk’s Falls Peel – Brampton, Mississauga Simcoe county - Barrie, Orillia, Midland, Collingwood, Cookstown

4

Appendices

108

Appendix F


Telephone Script for Initial Study Contact Hello, my name is <FILL First and last name >. I am calling for the Ontario Well Water Study. May I please speak with <FILL: Name of submitter >. I’m calling for a group of researchers at the University of Western Ontario and the Ministry of Health who are studying bacteria in well water. A few days ago, the Safe Water Unit called you and asked for permission for me to contact you about your participation in the study. You many also have seen information about the study attached to the water sample bottle you submitted for testing. The researchers are studying why the water in some wells becomes contaminated with bacteria that are resistant to antibiotics while water in other wells does not. They are also looking at whether using contaminated water affects the health of the people who drink it. To help with this study we need to talk to people like you who have had their water tested. It is important that we talk to people whose water tested positive for E. coli as well as people whose water tested negative so we can see whether there are differences between the groups. May I take a few minutes to explain what your participation in the study would include? [NO] Arrange for a date/time to return call. [YES] To understand why some wells become contaminated and why some people carry antibiotic resistant bacteria, we are going to interview 900 households in Southern Ontario. If you agree to participate in the study, a trained interviewer will come to your home to interview you about your well and things that might affect whether it becomes contaminated with bacteria. The interviewer will also ask you, and other members of your household, questions about your health and other things that might affect the bacteria in your intestines. Also, to find out which bacteria live in your intestines, we will ask you to give us a simple rectal swab. All personal information will be kept confidential and participation in this study is completely voluntary. However, we think this research is very important, especially considering the growing concern around antibiotic resistant bacteria and we hope you will agree to take part. Do you have any questions about the study? Would you be willing to take part in this study? [NO] Thank you very much for your time. Good bye. [YES] I would like to set up a time when I can visit your home to talk with you. We would also like to ask others in your household if they would agree to take part in the study.

Appendices

109

Would it be okay with you if I ask others in your household if they would agree to take part in the study?

[NO] That is fine, thank you. (Continue with appointment) [YES] Thank you, I will ask them if they are interested in participating at that time.

[ALL] I am in your area on <FILL: First choice of day, date, and morning, afternoon, or evening >. Would this be a good time to catch yourself [and others in your household] at home? [NO] How about <FILL: Second choice: Day, date, and morning, afternoon, or evening? [YES] Thank you, the interview will take about <FILL: 1 hour for household interview plus 30 minutes per personal interview>. I should arrive about <FILL: time >. If I’m going to be late, I will call to let you know. Would you give me your address and directions to your home? _______________________________________________________________________________ _______________________________________________________________________________ Do you have any questions? Thank you very much for agreeing to take part in this important study. I will be there on <FILL: Date and time >. Would you like to take down my name and phone number in case you need to change this meeting or if you have any [further] questions about the study?

Appendices

110

elewis

Cross-Out

Appendices

111

Appendices

112

Appendices

113

Appendices

114

Appendix H Ontario Well Water Study

Information and Consent Form You are invited to participate in a research study looking at bacterial resistance to antibiotics in well water and whether it is linked to antibiotic resistant bacteria in the human gut. Purpose of the study Several doctors and scientists from the London Health Sciences Centre, Mount Sinai Hospital, and the public health system are studying the public health impact of drinking, or being exposed to, well water contaminated with bacteria that are resistant to antibiotics. They want to find out whether well water is a potential source of bacteria that are resistant to antibiotics. The study investigators will compare the environment surrounding wells contaminated with bacteria that are antibiotic resistant to wells that are not contaminated to determine what factors may lead to contamination with antibiotic resistant bacteria.

The doctors and scientists are also studying gut bacteria. Some gut bacteria strains cause no harm and can be found in most healthy humans. Once acquired, humans may carry these bacteria in their gut for a long time. However, these bacteria are among the most common causes of infections in humans and they may carry genes linked with antibiotic resistance. How the study works This study will collect information from over 1,500 people living in Ontario who use private well water and have sent a sample for testing to a public health laboratory between May 1, 2005 and May 31, 2006. Only people who are residents of the household for three months or more before the test was sent, who are 12 years and older, and speak English are eligible to participate in this study. Risks of participating in the study Although it is highly unlikely, there is a small risk of rectal perforation while performing the rectal swab. Benefits of participating in the study Although there is not any known benefit for any individual participating in the study, the results of this investigation may help find out what factors are linked with contamination of private wells with antibiotic-resistant bacteria so it may be prevented in future. This study may also help determine what things are linked with humans carrying antibiotic-resistant bacteria in their gut so they can be avoided. What you will be asked to do if you agree to participate The study consists of two separate parts. The first is a visit to your home to interview you about your property, household, and well. The interviewer will also tour your property to note the location of your well, septic system, open water, pastures, and farm buildings. This should take about one hour. The second part is a personal interview about your health and, to find out which bacteria live in your gut, we will ask you to give us a rectal swab. A rectal swab is a good way to find out which bacteria live in a person’s gut. A rectal swab is a “Q-tip” that is inserted about 1 to 2 cm (less than 1”) into a person’s own anal area to sample stool. If you consent to provide a rectal swab, you also agree to have the sample tested for antibiotic resistant bacteria at the study lab, and to have the bacteria grown from your rectal swab frozen in the lab for up to one year. We would also like to talk to other members of your household for this study. However, we ask you to inquire whether other members of your household would be willing to be contacted by the researchers before giving us their name or contact information.

Page of 2 Participant initials: _________ 101

Appendices

115

Confidentiality Participation in this study is completely voluntary. You may refuse to participate. You are free to withdraw from the study at any time during the study. You may refuse to answer any question or to complete any interview, even if you have already started it. All personal information will be kept confidential and only grouped data will be published or released to the public. Personal information, such as names, will be replaced by a number on all surveys and specimens to protect privacy. Other things you should know If you are already participating in another study at this time, please inform the study person right away to determine if it is appropriate for you to participate in this study. You may call or write the study co-ordinator or the primary investigator listed on this form (at the number below) if you have questions at any time. You do not waive any legal rights by signing this information/consent form. You will be given a copy of this consent form. I have read the information/consent form, have had the nature of the study explained to me, and I agree to participate. All questions have been answered to my satisfaction. Signature ______________________________________ Date ________________________ Name (please print) _______________________________ I have read the information/consent form, have had the nature of the study explained to me, and I agree to give a rectal swab. All questions have been answered to my satisfaction. Signature ______________________________________ Date ________________________ Signature of person obtaining informed consent ____________________________________________ Contact information Study Coordinator and Co-investigator Primary Investigator Brenda Coleman, PhD candidate Dr. Marina Salvadori [email protected] London Health Sciences Centre (519) 631-3159 ext. 265 (519) 685-8500 ext. 52255

If you have any questions about the conduct of this study or your rights as a research subject you may contact Dr. J. Gilbert, VP Research and Development, at London Health Sciences Centre at (519) 685-8500 ext. 76649.

Page of 2 Participant initials: _______

Appendices

116

mailto:[email protected]

Appendix I


Household Questionnaire

Date of interview: - - (dd/mm/yyyy) Interviewer: ________________________________________ Household ID:

Consent acquired I am going to start with a general household questionnaire. It should take about 10 minutes and covers things about people who live here, your water supply and septic system, and even your pets. You are free to refuse to answer any question and to stop the interview at any time. However, your answers are all important and I hope you are able to answer all of the questions I ask you. Do you have any questions? I am going to start with a few questions about you. 1. The respondent is:

Male Female

2. How old are you?

years (999 for don’t know/refused)

3. How long have you lived at this address? (using this well) months years (999 for don’t know/refused)

I am going to ask about people who currently live in your home. For these questions, I would like to know about people who live in your home, whether or not they are related to you, but who live at this address four or more days per week. 4. Including yourself, how many adults, that is people 20 years and older, currently live in your home? (99 don’t know/refused) 5. How many youths 12 to 19 live here?

(If none, enter 0, do NOT leave blank)


Appendices

117

6. And how many children 4 to 11 years?

7. How many children under 4 years of age live in your household? (If zero, skip to Q=8)

7a. Are any of the children still in diapers? (Includes “pull-ups”) Yes No Don’t know / refused

7b. Do any children in your household go to a day care centre? (5 or more children in centre; child in care 1 or more days/week)

Yes No Don’t know

8. Does anyone in the household work at any of the following…

Yes No Day care centre or babysitting service Hospital, nursing home or residential home Sewage treatment plant Any other job where they are in contact with human waste: ______________________ Farm with livestock (any type) Abattoir, butcher shop, or meat processing Animal feed processing plant Nursery or landscaping service Any other job where they are in contact with meat, animals, or animal waste:___________

9. What township and county is this residence a part of? Township: ___________________________________ (Write don’t know/refused as required)

County: _____________________________________ 10. Do you have a swimming pool? (not a pond or swimming hole)

Yes No Don’t know

11. Do you have a hot tub or spa?

No Yes

12. Have many washrooms do you have in your home?

(With toilet; include outhouse; 99 for don’t know/refused)


Appendices

118

Now I would like to ask a few questions about your pets. 13. Do you have any pets?

Yes No (Skip to Q=14) Don’t know/refused

13a. What kind of pets do you have?

Dog(s) Cat(s)

Bird(s) Other: specify: ________________________________

14. In the past 3 months, have any animals spent more than a few minutes inside the house? (Several hours per week. Also include animals that live in house e.g. hamster)

Yes No (Skip to Q=17) Don’t know/refused

14a. What kind of animals have spent time inside the house?

Dog(s) Cat(s)

Bird(s) - (Skip to Q=16) Other: specify: ________________________________ - (Skip to Q=16)

________________________________ - (Skip to Q=16) 15. How often would you say you give your <FILL: cat and/or dog> any of the following. Would you say your pet(s) often, sometimes, rarely or never get(s)… (Read list, if more than one pet, indicate the more extreme measure for each item)

Often Some times

Rarely Never

Commercial dry or canned food Commercial biscuits or dry treats Raw meat (any kind) Cooked meat Raw hide treats

16. Were any of these animals on antibiotics in the past three months?

Yes No (Skip to Q=17) Don’t know/don’t remember

16a. Do you recall what kind of antibiotic were they given? __________________________________ __________________________________ __________________________________ __________________________________


Appendices

119

I’m going to ask a few questions about your drinking water now. 17. Where do you get the water you use for drinking? Is it from a private well, a well used by 6 or more households, a cistern, a municipal system, or some other source?

Private well Communal well (6 or more households) Cistern Municipal (or town) water Bottled (bulk or individual) Other: _____________________________ Don’t know

17a. And… where do you get the water you use for bathing, dental care, and other household uses?

Same as above (Q=17) Private well Communal well (6 or more households) Cistern (Skip to Q=21) Municipal (or town) water (Skip to Q=21) Other: _____________________________ Don’t know

18. What type of well do you have? Is it drilled, dug, bored, or driven, which is also called a sand point or well point?

Drilled Dug or bored Driven (sandpoint or wellpoint) Other: ________________________ Don’t know

19. How deep is your well?

feet (9999 for don’t know) metres 20. How old is it?

months (999 for don’t know) years 21. Have any repairs or maintenance been done on your well or water lines in the past 12 months?

Yes No skip to Q=22 Don’t know

21a. And in the past 3 months have any repairs or maintenance been done on your well or water lines?

Yes No Don’t know


Appendices

120

22. Why did you submit your water for bacteriological testing this most recent time? Do it regularly / routinely Off colour / cloudy Bad / different taste Odour Heavy rain People ill with stomach illness / diarrhoea E. coli in previous test Coliforms in previous test Other: _____________________________________________________ No specific reason Don’t know

23. How many times did you send your water for bacteriological testing in the past 12 months?

Number Many times (don’t know exactly, but more than 10) Don’t know

24. How many times has your well water tested positive for E. coli in the past 12 months? (Use results sheets if available)

Number Many times (don’t know exactly but more than 10) Don’t know

24a. And do you recall how many times it tested positive for coliforms?

Number Many times (don’t know exactly but more than 10) Don’t know

25. Do you currently treat the water you use for drinking? By treating, I mean boiling, adding chlorine or some other treatment to remove bacteria and other contaminants?

Yes No (Skip to Q=26) Don’t know

25a. How do you treat it?

Boil Chlorine Did you “shock” treat it or use a chlorinator ? Filtration Brita or other “filter” system Ultraviolet (UV) Ozone Other: ______________________________ Don’t know


Appendices

121

25b. When did you start treating it? - - (dd/mm/yyyy) (Year only if several years ago)

25c. And do you treat the water you use for dental care, bathing, and other household uses?


25d. How do you treat it?

Same as above (Q=25) Boil Chlorine Did you “shock” treat it or use a chlorinator ? Filtration Brita or other “filter” system Ultraviolet (UV) Ozone Other: ______________________________ Don’t know

25e. When did you start treating it?

- - (dd/mm/yyyy) (Year only if several years ago) Now a few questions about your septic system. Remember that everything you tell me during this survey is confidential. Your name will not be connected to anything you tell me and it will never be shared with anyone outside this study. 26. How is your domestic sewage handled? Do you have a … (Read list)

Septic tank and weeping bed (aka: field or leaching bed) Field tank Holding tank Lagoon Surface discharge (Skip to Q=28) Municipal system (Skip to Q=28) Other: _____________________________________ Don’t know ***Do NOT read***

27. When was the last time you had the tank [lagoon] pumped?

months (999 for don’t know) years (888 for never) 28. How old is your septic system? If Q=26 is “municipal system” ask How long have you been on the municipal sewage system? (Note: Oldest part if renovations completed)

months (999 for don’t know) years


Appendices

122

29. Have any upgrades or maintenance been done on your sewage system in the past 12 months?

Yes No Skip to Q-30 Don’t know

29a. And in the past 3 months have any repairs or maintenance been done on your sewage system??

Yes No Don’t know

Next I would like to ask a few things about your property. 30. How would you describe the soil on your property. Would you say it is predominantly (Read list)

Gravel Sand Loam, or Clay Don’t know ***Do NOT read***

31. How many acres of property do you own or rent at this location?

acres (9999 for don’t know) hectares

32. Would you describe your property as being … (Read list)

Farm Non-farm rural Village or hamlet (<1,000 people) Small town (1,000 to 10,000 people) Other: ________________________________

33. Have livestock been housed on this property in the past 12 months? This includes animals owned and/or cared for by your family or housed here and cared for by other people. (Include pony, chickens, pigeons, etc. but not cats, dogs unless it is a kennel)


33a. What type of livestock have been on this property in the past 12 months? (Check all that apply.)

Dairy cattle Pigs Beef cattle Horses or ponies Sheep (lambs) Chickens Goats Turkeys Other: _________________________________ Other: _________________________________


Appendices

123

33b. What is the largest number of <FILL: type of livestock> that have been housed on this property in the past 12 months?

____________________________ (type) ____________________________ (type)

____________________________ (type) ____________________________ (type)

34. Are livestock currently housed on the property?

Yes No Don’t know

35. Do [or did] you care for the livestock on this property?


36. Have you used antibiotics as a feed supplement for your livestock in the past 12 months?


36a. What types of antibiotics have you used as a feed supplement?

________________________________ (Write in ‘don’t know’ if applicable) ________________________________

________________________________

36b. When did you start using antibiotics in your feed? - (mm/yyyy) (Year is sufficient if several years ago) 36c. Are you still using antibiotics in your feed?

Yes (Skip to Q=37) No Don’t know

36d. When did you stop using it?

- (mm/yyyy) 37. Has manure been stored on your property or spread on your fields in the past 12 months?



Appendices

124

37a. Where, in relation to your well, has manure been stored or spread over past year? Would you say it is stored or spread …(Read list)

(Includes liquid or solid; stored in any way) Within 15 metres (50’) of your well Within 30 metres (100’) of the well (i.e. 16 to 30 metres) Within 100 metres (330’) of your well (31 to 100 metres) More than 100 metres (330’) from the well Don’t know ***Do NOT read***

37b. And within the past 3 months, where, in relation to your well, has manure been stored or spread? Would you say it is stored or spread …(Read list)

Within 15 metres (50’) of your well Within 30 metres (100’) of the well Within 100 metres (330’) of your well More than 100 metres (330’) from the well (Skip to Q=38) Don’t know ***Do NOT read***

37c. When was the last time manure was stored or spread on fields within 30 metres (100’) of your well? Would you say… (Read list)

Within the past month Within the past 3 months Within the past 12 months More than 12 months ago, or Never [not since s/he lived in house] Don’t know ***Do NOT read***

38. How soon is manure usually worked into the ground when it is spread? Would you say it is worked in… (Read list)

Same day (includes injected) Within 1 to 3 days Within 4 to 7 days More than one week after it is spread Don’t know ***Do NOT read***

39. Has a neighbour bordering your property had livestock on their land in the past 12 months? By bordering, I mean a neighbour that shares a fence line with you.

No (Skip to Q=40) Yes Don’t know

39a. What type of livestock were on that property within the past 12 months?

(Check all that apply) Dairy cattle Pigs Beef cattle Horses or ponies Sheep (lambs) Chickens Goats Turkeys Other: _________________________________ Other: _________________________________


Appendices

125

40. Has a neighbour bordering on your property stored manure on their property or spread manure on their fields in the past 12 months? (Stored in any way including piled, barnyard, tank..)


40a. Where, in relation to your well, have neighbours stored or spread manure over past 12 months? Would you say it is spread … (Read list)

Within 15 metres (50’) of well Within 30 metres (100’) of well Within 100 metres (330’) of well More than 100 metres (330’) from well Not stored/spread in past 12 months (Skip to Q=42) Don’t know ***Do NOT read***

40b. Where, in relation to your well, have neighbours stored or spread manure over past three months? Would you say it is spread … (Read list)

Within 15 metres (50’) of well Within 30 metres (100’) of well Within 100 metres (330’) of well (Skip to Q=41) More than 100 metres (330’) from well (Skip to Q=41) Not spread in past 3 months (Skip to Q=41) Don’t know ***Do NOT read***

40c. When was the last time manure was stored or spread on fields within 30 metres (100’) of your well? Would you say… (Read list)

Within the past month Within the past 3 months Within the past 12 months ago More than 12 months ago, or Never Don’t know ***Do NOT read***

41. How soon is manure usually worked into the ground when it is spread? Would

you say… (Read list) Same day (includes injected) Within 1 to 3 days Within 4 to 7 days More than one week Not spread (stored only) Don’t know ***Do NOT read***

42. Do you fertilize your vegetable or flower gardens or fruit orchards with animal manure? This includes manure purchased in bags at a store or garden centre.

Yes No Don’t know


Appendices

126

43. Has sludge from human waste been spread on fields within 90 metres (300 feet) of your well in the past 12 months?

Yes No Don’t know

44. And in the past 12 months, has waste from meat processing been spread within 90 metres (or 300 feet) of your well?

Yes No Don’t know

I am going to ask a few questions that will help us group your information with other households most like your own. Remember that nothing about you, as an individual, will ever be released and you are identified by number in this study. 45. First, what is the highest level of education that has been attained by any adult in the household? Would that be (read list) …

Less than grade 9 Some high school Graduated high school College or trade school University Don’t know ***Do NOT read*** Not stated ***Do NOT read***

46. What is your best estimate of the total off-farm income, before taxes and deductions, of all household members combined, from all sources, in 2005? Was that total household income… (Read list – include income from government sources)

Less than $20,000 $20,000 to less than $40,000 $40,000 to less than $60,000 $60,000 to less than $80,000 $80,000 or more Or do you not have off-farm income Don’t know ***Do NOT read***

Not stated ***Do NOT read*** For farming households only (Q=32=farm)… 47. What is your best estimate of the net income from your farm, before taxes, in 2005? Was that net income … (Read list)

Less than $20,000 $20,000 to less than $40,000 $40,000 to less than $60,000 $60,000 to less than $80,000 $80,000 or more Don’t know ***Do NOT read*** Not stated ***Do NOT read*** Not applicable: Not a farming property


Appendices

127

Thank you. That is all of the questions I have about the household. I would like to take a few measurements around your property to show where things are in relation to your well. However, I would also like to ask you a few questions about your personal water use, use of medicine, travel, and some other things that might influence whether you carry antibiotic resistant bacteria. Would you prefer I do the personal interview(s) now and do the measurements after we are through with it? 48. GPS coordinates from well:

. (heading/degrees/minutes/seconds) Longitude . (heading/degrees/minutes/seconds) Latitude 49. Note the distance between the well and each of the following… Distance M Km Ft Yd Mile NA DK

Septic tank Weeping tile House Vegetable or flower garden Area used to store manure in past 12 months

Stables or kennels Land used as pasture in the past 12 months

Field where manure applied in past 12 months

Fields tilled/worked in past 12 months

Open water Forest/wooded area Sanitary land fill site Nearest property line Neighbour’s septic system Municipal sewage tile – if within 1000 metres

Other: * 9999 = don’t know 8888 = not applicable Thank you very much for your time. Do you have any questions?


Appendices

128

Appendix J

Ontario Well Water Study Personal Questionnaire

Date of interview: - - (dd/mm/yyyy) Interviewer: _________________________________ Personal ID: -

Consent acquired

This interview should take about 10 minutes and covers things about your medical, work, and travel history, as well as some questions about your use of water and some personal habits.

You are free to refuse to answer any question and to stop the interview at any time. However, your answers are all important and I hope you are able to answer all of the questions I ask you.

First, I am going to ask a few questions about yourself so we can group you with others like you. 1. What month and year were you born? (Birthday this month: record next month if not passed)

- (mm-yyyy Enter 99-9999 for don’t know/refused) 2. Male

Female

3. How long have you lived at this address? (meaning at the house with this well) months (Enter 999 for don’t know/refused) years

Now I am going to ask a few questions about your health. 4. Has a doctor ever told you that you have…(Read list) Yes No D.K. Diabetes (type I or type II) Cancer Arthritis or rheumatism Heart disease or high blood pressure Asthma, bronchitis, or emphysema Autoimmune disease like lupus or Grave’s disease Migraines Kidney disease Crohn’s disease, ciliac disease, colitis, ileitis, or IBS Ulcers Any other digestive problems*: __________________ Any other chronic conditions: ____________________

Ontario Well Water Study of 8

Appendices

129

(*e.g. diverticulitis, recurring heart burn, etc.) 5. Have you been hospitalized, for at least one night, in the past 12 months? (Ref: calendar)

Yes No (Skip to Q=6) Don’t know/remember

5b. When were you admitted to hospital? (Probe: any other admissions in past year?) 5c. And how many nights did you stay in hospital in <FILL: month>?

Date of admission dd/mm/yyyy

# of nights admitted

- - - - - - - -

**Enter 99-99-9999 for don’t know/don’t remember

Now I’m going to ask a few questions about medications and other medical treatments you may have used in the past three months. 6. Within the past three months, have you taken any of the following medications or treatments? Have you taken….(Read list – Refer to calendar)

Yes No Steroids like prednisone or cortisone Immunosuppressive drugs like cyclosporine Chemotherapy (for cancer) Radiation therapy Aspirin or ASA for more than a day or two at a time

7. How about antibiotics like penicillin, tetracycline, gentamycin, and other prescriptions for infections. Have you taken any antibiotics in the past three months?

Yes No (Skip to Q=9 if “NO” to all medications and treatments….Skip to Q=8 if no only to Q=7) Don’t remember

7a. What antibiotics have you taken in the past three months? (Probe: any others?)

7b. How long were you on <FILL: name of antibiotic>? ______________________________________ days/weeks (Circle) ______________________________________ days/weeks (Circle) ______________________________________ days/weeks (Circle) ______________________________________ days/weeks (Circle) ______________________________________ days/weeks (Circle) (Enter “don’t know” if applicable) 8. Are you currently taking any of the medications or treatments I asked about?

Yes No (Skip to Q=9) Don’t remember


Appendices

130

8a. What medication(s) [treatments] are you taking right now? 8b. And when did you start taking <FILL: name of medication>?

________________________________________ / / (dd/mm/yyyy) ________________________________________ / / (dd/mm/yyyy) ________________________________________ / / (dd/mm/yyyy) ________________________________________ / / (dd/mm/yyyy) (Use prescription bottle when available - Enter “don’t know” if applicable) 9. When you are prescribed antibiotics, how often do you take the medication exactly as prescribed? By that I mean, taking the right number of pills at the right time of day. Would you say you always, usually, sometimes, rarely or never take it exactly as prescribed?

Always Usually Sometimes Rarely or never Have never been prescribed an antibiotic (Skip to Q=10a) Don’t know

10. When you use antibiotics, how often do you finish the full prescription? Would you say you always, usually, sometimes, rarely, or never finish all of the prescription?

Always (Skip to Q=11) Usually Sometimes Rarely or never Have never been prescribed an antibiotic Don’t know

10a. When you do not finish all medication or when it goes “out of date” or expires, how do you usually dispose of what is remaining? (List all that apply)

Return to pharmacy / drug store Throw in garbage for land fill Throw in the toilet / down sink Always complete medications Other (specify) ______________________________ Don’t know

11. We define diarrhoea as three or more loose stools or bowel movements in any 24-hour period. Have you suffered from diarrhoea in the past three months?


11a. Have you had diarrhoea in the past month?



Appendices

131

11b. How many times have you had diarrhoea in the past month? number (99 for don’t know. If zero, use 0, do not leave blank) 11c. How many days did the last episode of diarrhoea last? hours (99 for don’t remember/don’t know) days 11d. How many days of school or work - including work at home - did you miss

because of it? hours (99 for don’t remember/don’t know)

days (If zero, use 0, do not leave blank) The next set of questions are about some of the foods you eat. 12. Did you drink raw or unpasteurized milk, or eat diary products made from raw milk, in the past three months? (includes cream, butter, yoghurt, cheese, or ice cream)


12a. How often did you drink raw milk or eat dairy products made from raw milk? Would you say… most days of the week, a few times a month, or less often?

Most days A few times a month Less often Don’t know

13. Have you drank unpasteurized cider in the past three months?

Yes No Don’t know

Now I am going to ask you a few questions about your use of water. 14. While at home, approximately how many 8 ounce (240 mL) glasses of water do you drink every day? This would also include water in hot drinks like tea or coffee and for cold drinks like orange juice or Kool Aid. How many 8-ounce glasses do you think you drink every day?

glasses (Refer to glass - enter 99 for don’t know but probe for estimate) 15. Do you regularly use bottled water at home? By regularly, I mean most days of the week. (Individual and bulk)



Appendices

132

15a. Including water used for hot and cold drinks, about how many 8 ounce (240 mL) glasses of bottled water do you drink every day?

glasses (99 for don’t know)

15b. Do you use bottled water to …(Read list) Yes No Wash vegetables or fruit Brushing teeth Washing hands

15c. How long have you been using bottled water at home?

days (99 for don’t know) months

years This section is about meal preparation and some of your personal practices. 16. How often are you involved in meal preparation in your household? Would you say you always, usually, sometimes, rarely or never prepare meals?

Always Usually Sometimes Rarely (Skip to Q=17) Never (Skip to Q=17) Don’t know

16a. When preparing meals, do you touch raw beef, pork, or poultry with your bare

hands? Yes No Don’t know

17. In general, how often do you wash your hands with soap and water for each of the following. Would you say you always, usually, sometimes, rarely, or never wash your hands with soap and water… (Read list)

Always

Usually

Some times

Rarely

Never

Not applic.

Before preparing food After handling raw meat/poultry Before eating meals After using the toilet After playing with pets/animals After changing diapers After handling garbage After caring for sick people


Appendices

133

Now I’m going to ask about your work and leisure activities. 18. Have you travelled outside Canada in the past 12 months? (Includes USA if overnight)


18a. To what country or countries did you travel? (Probe: any others? If many, focus on the

past 3 months. If several to same country, focus on most recent trips) 18b. What dates did you travel to return from <FILL: country>? ______________________________ - - dd/mm/yyyy ______________________________ - - dd/mm/yyyy ______________________________ - - dd/mm/yyyy (Enter “don’t know” if applicable) 18c. Did you get diarrhoea while you were travelling or within a few days of returning from your trip(s)?

Yes No Don’t know

19. Over the past three months have you been swimming in an ocean, lake, river, or pond? (Include foreign and Canadian)

Yes No Don’t know

20. Have you been swimming in a pool in the past three months? (public or private)

Yes No Don’t know

21. And how about a hot tub? Have you used a hot tub in the past three months?

Yes No Don’t know


Appendices

134

I am going to read you a list of activities that might be a part of your day-to-day life. This includes things you might do at work, at home, or during leisure time. 22. Over the past three months, have you been in personal contact with human waste including diapers or bedpans, or while doing plumbing repairs or working at a sewage treatment plant?


22a. Would you say you were in contact with human waste several times per week, several times per month, or less often?

Several times per week Several times per month Less often

23. Have you been in personal contact with antibiotics for either human or animal use over the past three months? This might include at a pharmacy, veterinary clinic, or on a farm.


23a. Would you say you were in contact with antibiotics several times per week, several times per month, or less often?


24. Have you been in personal contact with animal or pet food, either at home or at work, in the past three months?


24a. Would you say you were in contact with animal or pet foods several times per week, several times per month, or less often?


25. Have you touched raw beef, pork or lamb with your bare hands in the past three months?



Appendices

135

25a. Would you say you touched raw beef, pork, or lamb several times per week,

several times per month, or less often? Several times per week Several times per month Less often

26. Over the past three months, have you touched raw poultry with your bare hands? This would include chicken, turkey, or other poultry.


26a. Would you say you have touched raw poultry several times per week, several times per month, or less often?


27. And over the past three months how often have you been in direct contact with any of these animals, meaning actually touching them or their manure? Would you say you were in contact several times per week, several times a month, less often, or not at all with…(Read list)

Several per week

Several per month

Less often Not at all

Dairy cattle Beef cattle Horses Pigs Sheep Goats Chickens Turkeys Other birds (including wild) Dogs Cats Game animals:________________________ Other: _______________________________

28. Do you attend school or work away from home? (incl. unpaid/volunteer work and all applicable)

Attend school Work No (Skip to end)

28a. On average, how many hours per week do you attend school or work away from

home? hours That is all of the questions I have for you at this time. Thank you very much for your help with this study. Do you have any questions for me? ***Provide instructions on how to collect rectal swab.


Appendices

136

Appendix K


RECTAL SWAB COLLECTION INFORMATION How to collect your swab The easiest way to ensure a proper sample is if you take your own rectal swabs and send the swabs for testing. To take the swab:

• Open the plastic covering by peeling back the top edges. • Place the numbered label on the tube with the black gel in it. • Remove the top from the tube with black gel in it.

• Pick up the cotton-tipped swab and roll it around the wrinkled skin of your

anus (the skin around your rectum). • Then insert the swab into your rectum (the cotton tip should be completely

inside your rectum). • Rub around the inside of your rectum twice. • Remove the swab and place in the black gel. Press the swab into the black

gel until the top is on tightly. • Place the swab into the “Specimen Transport Bag”. (One per bag, please.) • Send it by mail as described below.

Mailing instructions Once the swab has been collected place it in the envelope addressed to S. Braithwaite, Alberta Provincial Laboratory for Public Health. Send it to the lab by Canada Post (the address is on the envelope along with correct postage). Note: The swab can be stored at room temperature. All swabs can be mailed in the same envelope. Please send your signed consent forms in the envelope addressed to B. Coleman, c/o The Ontario Well Water Study. If you have questions or comments please call Brenda Coleman at 519-631-3159 ext. 265. Thank you very much for your help with our study.

Appendices

137

Appendix L

Variables Derived from Personal and Household Questionnaires Variable Data source Item Derivation Human carriage of antimicrobial-resistant E. coli

-Study laboratory analyses (screening and NARMS testing)

- Rectal swab isolate contaminated with E. coli resistant to one or more antimicrobial agents in 2004 NARMS panel

Use of untreated water contaminated with antimicrobial-resistant E. coli

-Public health laboratory analysis -Study laboratory analyses (screening and NARMS testing) -Initial consent -Household questionnaire

1-Public health laboratory bacteriological analysis of water 2-Laboratory analyses: at least one water sample contaminated with E. coli resistant to one or more antimicrobial agents in 2002 or 2004 NARMS panel for enteric bacteria 3-Initial consent (eligibility for “B” controls: Has this water tested positive for E. coli in the past year? -H: date of household interview -H25: Do you currently treat the water you use for dinking? By treating I mean boiling, adding chlorine, or some other treatment to remove bacteria and other contaminants. -H25a: How do you treat it? -H25b: When did you start treating it?

-Water not contaminated or treated: no E. coli contamination for one year or longer1,3 -OR- contaminated with E. coli susceptible to all antibiotics1,2 –OR- contaminated with antimicrobial resistant E. coli and treated* for one year or longer** -Contaminated: contaminated with antimicrobial resistant E. coli1,2 and not treated*

*Treated: “yes” to H25 and H26 = water boiled, treated with chlorine, UV, or ozone

**One year from date of interview (H: date of household interview and H25b)

Days: water collection to interview

-MOHLTC-SWU database -Personal questionnaire

- Date recorded as water collection date - P: Date of personal interview

Days between date of water collection and date of personal interview

Use bottled water -Personal questionnaire

- P15: Do you regularly use bottled water at home? By regularly, I mean most days of the week. - P14: While at home, approx. how many 8 oz. (240 mL) glasses of water do you drink every day? This would also include water in hot drinks like tea or coffee and for cold drinks like orange juice or Kool Aid. - P16: Including water used for hot and cold drinks, about how many 8 oz. glasses of bottled water do you drink every day?

-Tap water only: “no” to P15 -Bottled water only: “yes” to P15 and P14=P16 -Both tap and bottled water: “yes” to P15 and P14>P16

Age -Personal questionnaire

- P1: What month and year were you born?

-Years: date of interview less date of birth (day=1)

Sex -Personal questionnaire

- P2: Respondent sex As stated

Farming property -Household questionnaire

- H32: Would you describe your property as being… farm, non-farm rural, village or hamlet, small town, or other

-Farm (as stated) -Non-farm: non-farm rural, village or hamlet, small town, or other

Education (household)

-Household questionnaire

- H45: What is the highest level of education that has been attained by any adult in the household?

As stated Not stated: refused + don’t know

Appendices

138

Variable Data source Item Derivation Income (household)

-Household questionnaire

- H46: What is your best estimate of the total off-farm income, before taxes and deductions, of all household members combined, from all sources, in 2005? - H47: What is your best estimate of the net income from your farm, before taxes, in 2005?

-Household income: H46 + H47. -Don’t know or refused in either item was treated as don’t know / refused for derived variable. -Not stated: refused + don’t know

Household size -Household questionnaire

- H4: Including yourself, how many adults, that is, people 20 years and older, currently live in your home? - H5: How many youths 12 to 19 live here? - H6: How many children 4 to 11 years? - H7: How many children under 4 years of age live in your household?

Household size: H4 + H5 + H6 + H7 -Don’t know or refused on any item was treated as refused for derived variable

Lab region -MOHLTC-SWU database

(Note: not all water samples are tested in the laboratory region of the water source)

As recorded in database *not eligible if water source outside study area

Antibiotic use -Personal questionnaire

- P7: How about antibiotics like penicillin, tetracycline, gentamicin, and other prescriptions for infections. Have you taken any antibiotics in the past three months?

As stated

Hospitalization -Personal questionnaire

- P5: Have you been hospitalized, for at least one night, in the past 12 months?

As stated

Travel outside Canada

-Personal questionnaire

- P18: Have you travelled outside Canada in the past 12 months?

As stated

Child in day care -Household questionnaire

- H7: How many children under 4 years of age live in your household? - H7b: Do any children in your household go to a day care centre? (defined as child in care 1 or more days/week and centre with 5 or more children)

Child in day care: yes to H7 + H7b

Livestock, dog, or cat contact


- P27: Over the past 3 months, how often have you been in direct contact with any of these animals, meaning actually touching them ore their manure? Would you say you were in contact several times per week, several times a month, less often, or not at all with … Cattle = Dairy cattle and/or beef cattle Horses Pigs Sheep or goats = Sheep and/or goats Poultry = Chickens and/or turkeys Dogs Cats

Yes: several times per week, several times per month, or less often No: not at all Not stated: refused or not stated

Raw red meat contact


- P25: Have you touched raw beef, pork, or lamb with your bare hands in the past 3 months?

As stated

Raw poultry contact


- P26: Over the past 3 months, have you touched raw poultry with your bare hands? This would include chicken, turkey, or other poultry.

As stated

P: Personal questionnaire H: Household questionnaire MOHLTC-SWU: Ministry of Health and Long-Term Care (Ontario) – Safe Water Unit NARMS: National Antimicrobial Resistance Monitoring System

Appendices

139

the role of drinking water as a source of …

Documents